National Academies Press: OpenBook

Assessing Changing Food Consumption Patterns (1981)

Chapter: Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators

« Previous: Appendix B: Background Papers for Workshop on Evaluation of Methods for Obtaining Food Consumption Data
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 205
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 206
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 207
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 208
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 209
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 210
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 211
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 212
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 213
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 214
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 215
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 216
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 217
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 218
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 219
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 220
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 221
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 222
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 223
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 224
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 225
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 226
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 227
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 228
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 229
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 230
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 231
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 232
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 233
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 234
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 235
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 236
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 237
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 238
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 239
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 240
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 241
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 242
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 243
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 244
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 245
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 246
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 247
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 248
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 249
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 250
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 251
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 252
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 253
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 254
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 255
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 256
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 257
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 258
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 259
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 260
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 261
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 262
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 263
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 264
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 265
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 266
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 267
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 268
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 269
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 270
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 271
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 272
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 273
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 274
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 275
Suggested Citation:"Appendix C: Background Papers for Workshop on Nutrition and Health Status Indicators." National Research Council. 1981. Assessing Changing Food Consumption Patterns. Washington, DC: The National Academies Press. doi: 10.17226/380.
×
Page 276

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

APPENDIX Background Papers for C Workshop on Nutrition and Health Status Indicators

Can Nutritional Status Be Determined from Food Consumption or Other Measures? A. E. SCHAEFER I vividly recall the frustrations we had in developing and editing the Manual for Nutrition Surveys by the ICNND in 1957 followed by a second edition in 1963. In the early days of the ICNND'S country nutrition surveys, there was immediate recognition of the need for a procedural manual that would: 1. Establish uniformity in methods, techniques, procedures, and guidelines for conducting surveys in population groups in order to make meaningful comparisons of results within and between countries. It was recognized that in some instances more refined and precise methods were known; however, for fieldwork as in population surveys, methods were selected that had the advantage of practicability, simplicity, and reproduc- ibility. The interpretative guidelines for the biochemical data were de- veloped for use of the methods cited. 2. Outline and define responsibilities of each segment of investigation (clinical examination, anthropometric, dietary, food production and proces- sing). 3. Provide a guide for interpretation of the dietary, biochemical and clinical data. 4. Serve as a working reference of major facts essential for appraisal of nutritional status and to aid in interpretation of findings in order to draft practical recommendations to alleviate those nutritional problems discov- ered. 5. Assist in training personnel in nutrition appraisal techniques. 207

208 APPRAISAL METHODS A. E. SCHAEFER The general objectives of a nutrition survey are to define dietary intake at the time of the study; describe dietary patterns of consumption and food prep- aration practices; assess evidence concerning the prevalence of clinical, physical, biochemical, and anthropometric measurements indicative of nu- tritional status; and identify programs for improvement. This combined ap- praisal (dietary, biochemical, physical, anthropometric) gives more mean- ingful information than any single type of study. Biochemical studies pro- vide means for estimating the proportion of the population in various broad zones of nutriture and provide an estimate of tissue reserves of the various nutrients. Dietary studies constitute an essential part of nutritional appraisal providing essential information for solution of problems uncovered. In a population survey, additional information should be collected, such as in- fant and child-feeding practices, school and institutional feeding, economy of food use, and the unique nutrient contribution of special foods. Of the various methods of appraisal, biochemical and anthropometric measurements provide the key hard-core data, applicable to an individual or population group, for diagnosis of nutritional status and provide basic data to follow improvement or further deterioration. Dietary data is restricted in use; it does not define nutritional status. It is not a diagnostic measure. All of us appreciate the limitations of a 1-day, or 3-day, 7-day, or 30-day, etc., dietary assessment be it determined by recall or food weighings. Of concern is the perpetual argument reference the validity and value of a 24-hour dietary recall versus the 1- to 3- to 7-day home record + food weighings. Should we even waste time in doing it? COMPARISON OF DIETARY INTAKE METHODS Central America, INCAP-ICNND Survey Dietary surveys in the six Central American countries included a compari- son of the following methodology: In the case of families, food consumption was estimated by a 3-day daily record, a 24-hour recall with two different approaches (interviews in the home and interviews at the clinic), and a 1-day direct food weighing. For preschool children, a 1-day direct weighing and a 3-day daily record were used. The population sample for Nicaragua included a total of 125 families randomly divided in two groups. The 24-hour recall at home and the daily 3-day record were applied to the first group of families, while the 24-hour recall at the clinic and a 1-day direct weighing were applied to the second group of families (Table 1~. A comparison of the data obtained on the family studies revealed small differences that were not statistically significant. On the other hand, com-

Can Nutritional Status Be Determined from Food Consumption? 209 TABLE 1 Comparison of Nutrient Intake by Different Methods in Families in Rural Nicaragua 24-Hour Recall In Home In Clinic 3-Day Daily 1-Day Direct Record Weighing Calories x 1,928 1,857 1,916 1,950 SD 702 639 683 674 Protein, g x 66.1 61.1 60.6 62.8 SD 26.1 23.2 22.4 24.2 Calcium, mg x 725 790 773 755 SD 455 572 457 438 Iron, mg x 17.8 16.6 18.7 17.9 SD 8.1 7.7 7.9 10.5 Retinol, mcg x 150 98 169 247 SD 569 86 570 1,270 Beta carotene, mcg x 522 481 578 527 SD 591 525 630 642 Thiamin, mg x p.87 0.82 0.92 0.91 SD 0.38 0.30 0.40 0.42 Riboflavin, mg x 0.95 0.93 0.97 0.93 SD 0.55 0.52 0.58 0.52 Niacin, mg x 10.90 9.89 10.26 9.97 SD 5.28 3.47 4.01 3.74 Vitamin C, mg x 75 59 71 57 SD 80 57 133 70 parison of the data of the preschool children revealed significant differences when comparing the 1-day direct weighing with the 3-day daily record for calcium, retinal, and riboflavin. For retinal and riboflavin, the differences were highly significant with the lower estimates found by the 1-day direct weighing (Table 2~. Examples of Use of Data Central America Nutrition Surveys, 1965-67. In a study of over 3,800 families (approximately 20,000 people) in the six Central American countries, dietary intake studies were made on a sam- ple of families on which detailed biochemical and hematological investiga- tions were carried out. Two methods of obtaining quantitative data on fam- ily food consumption were used: one by means of a 24-hour recall, using the

210 A. E. SCHAEFER TABLE 2 Comparison of Nutrient Intake by Various Methods in Preschool Children in Rural Nicaragua 1-Day Direct 3-Day Daily Weighing Record 1,089 < 142 Calories x SD 1,100 433 Protein,g x 31.9 34.0 SD 13.7 15.8 Calcium, mg x 413 693 SD 294a 606a Iron, mg x 9.0 8.1 SD 6.0 4.0 Retinal, mcg x 47 130 SD 496 1686 Beta carotene x 167 300 SD 171a 287a Thiamin, mg x 0.44 0.50 SD 0.24 0.21 Riboflavin, mg x 0.47 0.94 SD 0.33b 0.80b Vitamin C, mg x 28 32 SD 36 33 a Significant at 5 percent probability. Significant at 1 percent probability. NOTE: X = mean; SD = standard deviation. housewife's own account of the previous day's meals; the other a 3-day record consisting of the actual recording in the home of the foods prepared and consumed. The 3-day record method was applied to a random portion of the families surveyed by the 24-hour recall. At the end of the survey in each location, typical family diets were calculated and food composites were locally prepared for subsequent chemi- cal analysis. Individual food intakes of preschool children were recorded by the 24-hour recall in most of the countries. The value and kinds of use of dietary data are illustrated in Tables 3-7. The intake of vitamin A revealed a dramatic deficit when compared to the recommended levels of vitamin A. This finding was supported by the

Can Nutritional Status Be Determined from Food Consumption? 211 TABLE 3 Consumption of Selected Foods per Person per Day in Rural Areas (g, Edible Portion) Guate- El Sal- Hon- Nica- Costa Pana- Foods malaa vador auras ragua Rica ma Milk productsb 84 190 194 243 193 73 Eggs 13 10 13 12 15 11 Meat, poultry, and fish 44 37 41 58 40 90 Beans end oily seeds 54 59 56 72 57 20 Vegetables 66 53 51 27 66 25 Fruits 14 17 40 41 7 50 Bananas and plantains 20 16 43 72 47 99 Starchy roots and tubers 14 13 22 33 46 82 Cereal products Rice 16 27 29 54 100 186 Corn tortillas, etch 359 352 224 139 41 32 Wheat bread 36 26 12 28 54 37 Wheat flour and pastes 4 0 8 7 12 10 Others 2 6 5 16 0 0 Sugar 52 41 39 58 89 51 Fats and oil 4 15 16 19 19 26 Number of families 203 293 331 355 456 361 aIn the report on Guatemala (INCAP v-25), the figures for food consumption were those ob- tained by the 3-day-record method. In order to conform to the data in the reports for the other five countries, the dietary data for Guatemala used throughout this report are based on the 24-hour-recall method. bIn terms of liquid milk. CIn terms of grain. TABLE 4 Percent Contribution of Selected Foods to Calorie Intake of Rural Families Guate- El Sal- Hon- Nica- Costa Pana- Foods male vador auras ragua Rica ma Milk products 2.8 5.6 6.9 7.3 7.3 2.7 Eggs 1.0 0.7 1.1 0.9 1.3 0.8 Meat, poultry, and fish 3.9 3.3 5.0 5.4 3.8 7.2 Beans and oily seeds 9.5 9.7 10.7 13.0 10.7 3.9 Vegetables 1.4 0.9 1.4 0.6 2.2 1.1 Fruits 0.6 0.6 1.8 1.5 0.4 4.1 Bananas and plantains 1.1 1.0 2.9 4.8 3.2 5.9 Starchy roots and tubers 0.6 0.7 1.8 2.2 2.6 4.9 Cereal products 65.1 61.8 49.5 41.5 39.3 47.4 Sugar 9.4 7.4 8.2 11.5 17.9 8.8 Fats and oil 2.6 6.9 8.8 9.3 9.8 11.7

212 A. E. SCHAEFER TABLE 5 Average Intake of Calories and Specific Nutrients per Person per Day in Rural Families Guate- E1 Sal- Hon- Nica- Costa Pana- mala vador auras ragua Rica ma Calories 2,117 2,146 1,832 1,986 1,894 2,089 Total protein, g 68.0 67.9 58.0 64.4 53.6 60.1 Animal protein, g 15.4 17.3 18.5 23.4 18.5 26.6 Vitamin A., flu 2,420 900 1,280 1,700 1,800 1,830 Riboflavin, mg 0.80 0.78 0.79 0.93 0.84 0.69 Iron, mg 17.9 11.6 15.5 18.2 15.4 14.3 TABLE 6 Percent Adequacy of Average Intake of Calories and Specific Nutrients in Rural Families Guate- E1 Sal- Hon- Nica- Costa Pana- mala vador auras ragua Rica ma Calories 109 109 89 96 91 104 Protein 133 128 108 115 98 112 Vitamin A 68 24 34 44 49 49 Riboflavin 70 66 64 75 68 58 Iron 178 114 152 175 150 141 TABLE 7 Percent Distribution of Rural Families by Adequacy of Vitamin A Intake Country No. of Families <25 Percent Adequacy 25-49 50-74 75-99 ~ 100 Guatemala 200 44 22 10 6 17 E1 Salvador 278 69 19 7 3 2 Honduras 323 57 26 9 2 6 Nicaragua 331 45 23 13 8 11 Costa Rica 414 44 26 11 7 12 Panama 352 42 32 13 5 8 biochemical evidence. Likewise when the data were analyzed by "low," "medium," and "high" socioeconomic index (Figure 1), the serum vita- min A concentration or urinary riboflavin values were lower in the poorer . . socioeconomic groups. The incidence of clinical symptoms specifically related to nutrient de- ficiencies was not as high as the occurrence of biochemical values in the "deficient" or "high risk" range. This is to be expected.

Can Nutritional Status Be Determinedfrom Food Consumption? 213 800 700 con ~5 <., 600 - 500 400 LL o m 200 100 I Interpretation Nicaragua - :-:-:- ·--:-: :-:-. ·:-:-: .:~ .: ·--:-: :-:-:- :.:.:-. :-:-:- ·:-:-: :-:-:- ·:-:-: Am: ·--:-: ·-.-.:.: :::::: :::::: ·:-:- . ........ :.:-.:-. :-:-:- ·:-:-. ·-:-:- .:.:, :-:-:- :--.:- ::~ .~.- .~ :-:-:- ·:-:-: ·.:.:-.: ·:-:-: :-:-:- ·:-:-: ·---:- . _ _ _ _ Costa R ice Panama _., ·..... :-:-:- ·:-:-: ....... ·:-:- ,........ ·:-:-: :- -:- ·:-:-: ·.:-:.: ·:-:-: :: ,.. .:-.:-: ·:-:-: :-:-:- ..... ~. .:: :-:-:- :::..-. :-.: ·:-:-: :-:-:- ·:-:-: :-.:-: ·:-:-: :-:-:- .......... ·:-:-: :-:-:- ·..... ·:-:-: :-:-:- *,: ·:-:-: ·:-:- ·.:-:-. .-.- ·.:-.:. ·:-:- .~ Socioeconomic Index Low Medium High r l Acceptable 1 Low . Def icient FIGURE 1 Nutritional study in Central America and Panama: Urinary excretion of riboflavin in relation to the socioeconomic index of the families in the rural areas of Central America, 1965-1967, males 0-4 yr. Dietary studies revealed an even higher incidence of inadequate nutrient intakes than was evident from the biochemical and clinical studies. The results of the biochemical analyses and dietary studies (with excep- tion of vitamin A and riboflavin) indicated that the average consumption of essential nutrients in each country would have been enough if the distribu- tion of the foods had been equitable. Important differences in the average consumption of various foods per person per day was observed between countries. Likewise, these differences were observed between rural and urban families within each country. The basic dietary data correlated with biochemical estimates for those groups at "risk" or at "moderate risk" developing "malnutrition." No attempt to

214 A. E. SCHAEFER directly relate individual dietary intake to biochemical data was done nor should it have been done. For defining the problems and potential solutions, it made no difference whatever which set of dietary intake data were used. Bangladesh Nutrition Survey Correlation of Diet and Biochemical Assessment In the 1960-62 nutrition survey conducted in Bangladesh (then East Pakis- tan) (Pakistan, 1969), in which over 25,000 individuals were studied over a 2-year cycle, seasonal variations of intake of leafy green and yellow vegeta- bles and fruits correlated with seasonal variations in serum vitamin A levels and the prevalence of keratomalacia. During the season when leafy green vegetables and yellow fruits were in short supply with an average daily consumption of only 12 g per day, 42 percent of the people had serum vitamin A levels of less than 20 mcg per 100 ml (Figure 2~. Likewise, the percent of total cases seen at the Dacca medical college eye clinic with keratomalacia increases from 0.5 to 3.0 percent during the period of low intake of leafy green and yellow vegetables and fruits. A similar type correlation was noted on the influence of pulse consump- tion on serum albumin levels (Figure 31. When intake reached a low level of 12 g of pulses per day, 58 percent of the population had "low" serum albumin levels of less than 3.5 g per 100 ml, whereas during periods of the year when pulse consumption reached 40 g per day, only 28 percent had "low" serum albumin levels. Ecuador ICNND Nutrition Survey The Ecuador nutrition survey conducted in 1959 employed three methods of obtaining food consumption data: 24-hour recall, 24-hour recipe food weighing, and food composite chemical analysis. Comparisons of the nu- trients consumed in two major regions (coastal and sierra) are given in Tables 8 and 9. The amounts of food consumed per person per day as determined by the direct recipe weighing method in the home and.the 24- hour recall were in remarkably close agreement. The data illustrated that the results obtained from the recall questionnaire can be considered essentially as valid as the data obtained from direct weighings. The same number of personnel were required to obtain data from 28 families by the recipe weighing method as it took to obtain data from 341 families by the recall method. Food composites structured after the data obtained by the recipe method were analyzed by two laboratories. The study revealed that in general the dietary recall method slightly un-

Can Nutritional Status Be Determinedfrom Food Consumption? 215 ,,, 50 _ UJ O ~ ~ 40 UJ Z 30 2 o ~ IL ~ Z O ~ ~ 20 ~ O z ~ I= ~ ~ ~ 10 UJ ~ O INTAKE (%) 50 ~ 1- - ~1 At' if\ ~ Jan Mar May (Fete) Apr Jun Mar May Jul _ __ ___% % " Def icient or `' Low" Serum Vitamin A's 4° ~30 20 _ 10 Intake of Leafy Green Vegetables and Ripe Fruit (ems) I I I I . Jul Sep Nov Jan Aug (Oct) Dee ( Feb) Sep Nov Jan Mar O FIGURE 2 East Pakistan Nutrition Survey. Seasonal correlation of intake of leafy green vegetables and ripe fruit with serum vitamin A levels, 3-monthly averages. 60 1 an J 2 5o z J 1- Z 40 m car J O ~ ~ ~ 30 0 ~ LL Z Oh 2~ ~ ~ v cay ~ 0 ~ 'a ~ 1 0 i, _` O I I I I 1 1 1 1 1 1 1 1 Dec ( Feb) Apr Jun Aug (Oct.) Dec Jan Mar May Jul Sep Nov Jan (Fete) Apr Jun Aug (Oct) Dec (Fete) ~ , ~-'_~` . ' ~ = %"Deficient" or "Low" Serum Albumin Levels Intake of Pulses (Gms) PULSE INTAKE (Gms) 60 50 40 _ 30 _ 20 10 O FIGURE 3 East Pakistan Nutrition Survey. Seasonal correlation of pulse intake with serum albumin levels, 3-monthly averages.

216 A. E. SCHAEFER TABLE 8 ICNND Ecuador: Comparison of Average Daily Nutrient Intake (1959~: Sierra Civilians 24-Hour Food NCR Calc. 24-Hour Recipe Composite Allowance, Recall, Food Weigh, Chemical, 213 Families 213 Families 15 Families 15 Families Calories 2,057 2,068 1,783 1,935 Protein, g 55 58 59 69 Calcium, g 1.04 0.49 0.47 0.61 Iron, mg 11.0 22.3 17.1 32.5 Vitamin A, flu 4,287 4,245 1,453 3,450 Thiamin, mg 1.18 1.17 1.0 0.82 Riboflavin, mg 1.64 0.81 0.91 1.68 Niacin, mg 16.0 22.2 20.8 20.3 Vitamin C, mg 71.0 112.0a 77.0a 54 aNo cooking losses calculated. TABLE 9 IcNND Ecuador: Comparison of Average Daily Nutrient Intake (19591: Coastal Civilians 24-Hour Food NCR 24-Hour Recipe Composite Allowance, Recall, Food Weigh, Chemical, 116 Families 116 Families 13 Families 13 Families Calories 1,878 1,791 1,769 1,940 Protein, g 53.3 58.2 57.2 63.3 Calcium, g 1.04 0.32 0.32 0.60 Iron, mg 11.0 14.0 14.1 21.0 Vitamin A, flu 4,042 4,554 3,524 2,614 Thiamin, mg 1.13 0.68a 0.68a 0.39 Riboflavin, mg 1.57 0.69 0.65 0.73 Niacin, mg 14.0 11.5 19.4 15.1 Vitamin C, mg 62.0 97.0a 86.0a 47 aNo cooking losses calculated. derestimates the food consumed by 10-15 percent. However, when larger numbers of families were studied, such as 341 families versus 28, the values for the recall method were equal or greater than the results obtained by the home recipe weighing method. The food composite analysis method, which does require food intake data, provides the opportunity to obtain data on nutrient losses during cooking and on many nutrients for which food nutrient composition data were not available (vitamin B6, folic acid, zinc, sodium, potassium, etch. Chemical analyses of food composites in general include local water and

Can Nutritional Status Be Determined from Food Consumption? 217 local food preparation. In virtually all cases, higher iron and calcium values were found. DIETARY METHODS IN EPIDEMIOLOGIC STUDIES Morgan et al. (1978), as well as unpublished data by Margaret Moore and co-workers at Louisiana State University and Tulane, conclude that, ". . . Dietary history properly taken for research purposes serves a very useful purpose and is of greater value than is generally appreciated." Tests for reliability showed no significant differences in the values obtained within 2 years for a dietary history and 7-day food records. Morgan et al. employed three methods: 24-hour recall, 4-day dietary record, and diet history for 2 months, plus a recall for 2 months dating back 8 months. There was a high correlation between the current diet history intake and that measured 8 months previously. Current history data correlated with the 24-hour recall and the 4-day record. When all three methods were directed to the same period, they measured the same things. Although each method may be satis- factory for comparing group values, there is doubt as to which, if any, of the three methods is satisfactory for interpreting individual values. The authors concluded that the 24-hour recall and even the 4-day diary are inherently less reliable for individual estimates of usual intake and that when such estimates are required, the diet history should be used. SUMMARY The constraints of dietary survey are known by anyone familiar with human biology. Why one would expect a high degree of sensitivity is beyond comprehension. If one begins with crude data on nutrient composition, then why try to over interpret the data? For example, we continually see intense discussions and conclusions on precise intakes of fat, protein, carbohydrate, and fiber. The data in the data bank still represent carbohydrate values on the basis of "crude fiber determination" and in agricultural biochemistry terminology we talk about "proximate analysis." These are the first con- straints of dietary intake data, unless one is willing to really invest dollars instead of talk in obtaining better than "proximate crude analysis data" to define such key nutrients as carbohydrates, fat, protein, crude fiber, and resultant calories. On the positive side, one can interpret dietary intake data on population groups. An example of dietary intake correlating with biochemical assess- ment is illustrated by Figure 4. From a practical slant, undoubtedly the most important findings relative to identifying diet pattems, habits, and food availability are data that define

218 A. E. SCHAEFER 400 a, 350 . _ . _ a, cat - z o - LL X LU ~ 150 al Z 100 - ~ 50 300 250 200 o 0 0 OMilitary Data · Civilian Data AS ~ O · O I /. l 0 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 THIAMINE INTAKE (mg 1,000 calories day) FIGURE 4 Relationship between thiamine intake and thiamine urinary excretion in adults of 18 countries. Data from civilian and military groups surveyed by the Interdepartmental Committee on Nutrition for National Defense of the United States of America. From: FAG Nutrition Meetings Report Series No. 41. foods or food groups with estimates of quantitative intake. Note Table 10, which illustrates major variations in foods consumed between countries (Chile and Ecuador) and within regions of a country. For planning purposes from a health-economic-social standpoint, it is imperative to have such data on foods consumed.

Can Nutritional Status Be Determined from Food Consumption ? 2 19 TABLE 10 Representative Diet Patterns Chile Ecuador Ecuador (Lowlands) (Highlands) Type of diet wheat rice corn No. offamilies 278 116 213 Wheat products Rice Corn Beef and mutton Pork Poultry Fish Milk products Eggs Fats and oils Leafy green and g consumed per person per day 351 35 39 23 122 79 2 116 8 1 66 42 9 10 33 107 27 22 124 38 267 2 11 yellow vegetables 97 57 38 Potatoes 153 125 275 Tomatoes 16 30 6 Other vegetables 46 20 21 Pulses 27 20 39 Citrus fruit 5 21 13 Bananas 6 99 19 Other 96 SECOND Survey Reports- 1960, Chile; 1959, Ecuador. REFERENCES AND BIBLIOGRAPHY Archivos Latino Americanos de Nutricion, Vol. XXIII, Sept. 1973, pp. 325-344. Marina Flares, Maria Teresa Menchu', and Miguel A. Guzman. Evalucion dietetica de familias y preescolares mediante la aplicacion de diferentes me'todos y tecnicas Area rural de Nicaragua. Ecuador. Nutrition Survey, 1959. Report by the ICNND, July 1960. Published by NIAMD, HEW, Bethesda, Md. Manual for Nutrition Surveys, 2d ed. 1963. Interdepartmental Committee on Nutrition for National Defense, National Institutes of Health, Bethesda, Md. U.S. Government Printing Office, Washington, D.C. Morgan, R. W., et al., 1978. A comparison of dietary methods in epidemiologic studies. Am. J. Epidemiol., 107(6):488-498. Pakistan. Nutrition Survey, March 1962-January 1964. 1969. Report by the Ministry of Health, Gov't. Pakistan, University of Dacca. Nutrition Section of NIH, Department of HEW.

Relating Food Intake, Demographic, Attitudes and Behavior Data to Measures of Nutritional Status HOWARD G. SCHUTZ Other than for major deficiencies in intake, the relationship between dietary intake and health status has not been clearly delineated. Certainly for the objectives of the FDA, there is not available any generally accepted method for determining the possible relationships between consumption of food ingredients and health status. There are three major problem areas associated with this task. First are the myriad of difficulties associated with the collection of reliable and valid data concerning individual food intake patterns. These difficulties have been discussed in a previous workshop. Second are the problems associated with the collection of meaningful measures of health status. These will be discus- sed at the present workshop. Among the questions to be answered are: (1) What is the value of self-reported health status? (2) What kind of biochemi- cal data should be collected and for what period of time? (3) What non- biochemical attitude behavior and demographic measures are appropriate to collect? Third is the problem of how we relate our measures of food intake and other variables to the health status measures. The latter issue is the one that I would like to address. I believe the assumption that relationships can and should be studied between single measures of intake and nonintake and health status is simplistic. There is every reason to believe that there are additive and perhaps interactive effects among nutrients that contribute to particular health conditions. Also, it is quite likely that measures of attitude and behavior, as well as demographic characteristics, can play an important role alone and in combination with intake to influence health measures. What appears to me to be one approach to the problem of determining relationships is to relate measures of intake and other variables to health 220

Measures of Nutritional Status status utilizing a multivariate approach. Using the statistical method of multiple regression, the intake, demographic, attitude, and behavior vari- ables become the independent variables and the particular measure of health status becomes the dependent variable. This approach in its simplest form yields the linear additive relationships between the predictor variables, i.e., intake, etc., and the dependent variables, a health status measure. By utilizing squared and cross-multiplied terms, one can also include nonlinear effects and interactions. As an aid in reducing and simplifying the independent and sometimes the dependent variable, I have used a clustering technique such as factor analysis first in order to select variables that represent fundamental dimen- sions and that are not related to one another statistically. Four examples of the general approach are noted. The first of three papers (Schulz et al., 1977) is a paper on predicting purchase and use that does not measure health status but does demonstrate the multivariate approach to the general areas of food. The second (Baird and Schutz, 1976) is a paper that demonstrates the approach applied to predicting judged dietary adequacy for four ethnic groups using cognitive structure based on food-appropriateness information as independent variables. The third paper (Baird and Schutz, 1980) demonstrates the successful prediction of a number of measures of nutritional status by demographic and attitudinal variables. These data were for a sample of 100 Californians. The fourth example consists of summary tables showing the results of a multivariate analysis of intake, demographic attitude and behavior, and nutritional status data collected (N varies from 349 to 435) on female heads of households in five Western states. These data are from an experiment station bulletin that is in press. Table 1 gives the variables studied and Tables 2-11 give stepwise multiple regression results for nine measures of health status. Three regressions are given, one with only environmental variables (demographic D, one with environmental plus attitude and be- havior variables (IIa), and the third with environmental, attitude, and be- havior, plus dietary intake (III). Examination of the tables reveals that statistically significant predictions were possible in all cases but that the variance accounted for varied from about 1 percent to 70 percent. I view the results in this last set of tables as very encouraging, since measures of intake were only 3-day diet records, biochemical measures were single occasion measures, and most importantly five different sets of experimenters col- lected the data. More reliable and valid data collection techniques should result in the ability to predict higher amounts of variance in the health status variables. 221

222 REFERENCES HOWARD G. SCHUTZ Baird, P. C., and H. G. Schutz. 1976. The marketing concept applied to "selling" good nutrition. J. Nutr. Educ. 8:13-17. Baird, P. C., and H. G. Schutz. 1980. Dietary and nutritional status: Life style correlates. J. Am. Diet. Assoc. 76:228-235. Schutz, H. G., S. M. Moore, and M. H. Ricker. 1977. Predicting food purchase and use by multivariate attitudinal analysis. Food Tech. 31:85-92. (See tables overleap andfollowing pages) '1

Measures of Nutritional Status 223 TABLE 1 List of Vanables. (*Indicates Dummy Vanable, 1 = Yes, 0 = No) 1. NUTRITIONAL STATUS MEASURES a. Anthropometric and Clinical Measures (dependent variables, regressions I, IIa, III) 1. Triceps skinfold thickness, mm 2. Diastolic blood pressure, mm Hg b. Biochemical Measures (dependent variables, regressions I, IIa, III) 3. Total serum protein, g/dl 4. Iron binding capacity, ,ug/dl 5. Serum albumin, g/dl 6. Plasma vitamin A, ,ug/dl 7. Serum iron, ,ug/dl 8. Hematocrit, % 9. Serum vitamin C, mg/dl 10. Iron transferrin saturation, % Dietary Intake (mean intake from food and supplements/day) (dependent variables, regressions I, IIa; independent variables, regression III) 11. Carbohydrate intake, g 12. Viamin C intake, mg 13. Niacin intake, mg 14. Vitamin A intake, flu 15. Protein intake, g 16. Thiamin intake, mg 17. Riboflavin intake, mg 18. Iron intake, mg 19. Energy intake, kcal 20. Calcium intake, mg 21. Phosphorus intake, mg 22. Fat intake, g 2. FOOD ACCEPTANCE MEASURES (dependent variables, regressions I, IIb; independent variables, regression IIa) 23. Frequency scale (freq scale) average frequency of serving 174 foods (high scores = low frequency) 24. Like-dislike scale (like/dislike) 25. "Finicky'' scale (finicky) 26. Vitamin usage number of vitamin pills taken daily (vit pill user) 3. ENVIRONMENTAL FACTORS (independent variables) a. Selected Physical, Socioeconomic, and Social Environmental Factors (selected before the study commenced) (regressions I, IIa&b, III) 27. Age (yr), varies from (18-57 yr) *28. Ethnic group (nonwhite) (reference = caucasian) 29. Income- annual family income (income) 1. Less than $3,000 2. 3,000-5,999 3. 6,000-8,999 4. 9,000-11,999 5. 12,000-24,999 6. Over 25,000

224 TABLE 1 Continued HOWARD G. SCHUTZ 3 30. Number of persons living in household (# persons), including the respondent (varies from 1 -over 1 1) (household size) 1. Highest level of education completed by (educ) lead female 1. 5 grades or less 2. 6-8 grades 3. 9-11 grades 4. High school graduate 5. Some technical training or college 6. Junior/eommunity college graduate 7. College graduate 8. Graduate school 32. Rural/urban size of community 1. Small rural community, <1,000 people 2. Rural community, 1,000-2,500 3. Small community, 2,500- 10,000 4. Small urban community, 10,000-25,000 5. Large urban community, 25,000-50,000 6. Metropolis, 50,000 and over *33. State Arizona *34. State Hawaii ~ *35. State Nevada ~ Reference = California *36. State Washington ,1 b. Other Factors Hypothesized to Affect Nutritional Status andlor Food Acceptance (in- dependent variables, regressions IIa, b, III) i. Activity factors *37. Employed (yes = 1, no = 0) 38. Job activity level 1. Sedentary 2. Mildly active 3. Very active 39. DPAS = daily physical activity score. Total average hourly kilocalorie expenditures for each physical activity (swimming, exercising, walking, dancing, bowling, tennis, other specific listed by subject. If none, score was average hourly energy expenditure for sedentary office work. ii. Reasons for serving food to the family. Number of times respondents mentioned reasons that fall into the following categories. Code 40. Sensory reason 41. Social reason 42. Education reason 43. Psychological reason 44. Emotional reason 45. Religious reason 46. Health reason 47. Preparation reason 48. Miscellaneous reason 49. Price reason 50. Availability reason 2 s 6 7 9 10 8.01, 8.02, 8.11 8.03, 8.08, 8.09

Measures of Nutritional Status TABLE 1 Continued 51. "Do (grow/hunt) yourself" 8.06, 8.08, 8.10 reason iii. Consumer behavior factors *52. Don't buy any food more frequently than others (don't buy) *53. Have no regular pattern of shopping for groceries (no reg shop) *54. Shop on the weekends (shop w/e) *55. Shop in chain supermarket (shop supermarket) *56. Shop where get the lowest price (shop low price) *57. Shop for maximum quality at minimum price (shop min/max) *58. Not eligible for food stamps (ineligible FS) *59. Would buy more meat if had more money (more meat) *60. Would buy fewer sweets and snacks if had less money (fewer sweets/snacks) *61. Have adequate kitchen storage (storage OK) *62. Get food information from newspapers and magazines (info pop press) *63. Shop when convenient (shop convenience) *64. Has family members on special diet (special diet) 225

226 . ~ - .= Ct . ~ C~ C~ ~: ._ E~ - ~: ._ ~: V) C~ C) ._ s~ l C~ au - C~ ._ ~: 3 m E~ o . ca 0C Ct S o . C~ C) ~4 C) .o U. C~ ~4 _' ~ 1 | tl I , oo U. Ct 'L) s~ 1 S 1 _' ~: Ct C~ o o cn c~ ~ ~ _ ~ oo ~ ~ ~ ~ oo _ ~ ~ ~r ~ _ _ ~ ~ _ _ o . . . . . . * o o o o o o m I + ++ + I i ~ ~ ~ 00 ~ 00 1 ~ ~ ~ ° ° — o ~ 1 ~ 1 + + + + 1 ~ ~ ° ~: Ct P~ C) ~ Ct O O ~ a, o o _ c a~ _ O o~ o _ (U .0 (D _ ._ Ct C) _ ~ O O ~ (_) ~ ~ C~ ~ — 0N 00 O _ _ _ _ . . . . * O O O O m I + ++ ~ ox 0 ~ _ ~ ~ 0 ~ _ o 0 0 0 d · · · · ~ ~ 0 0 0 0 + o+ ~ ~ O cr, ~ 1 — ~ ~ ~ u~ 00 _ _____ . . . . . * 0 0 0 0 0 m I I + I + 00 ~ ~ _ _ _ ~ ~ ~ ~ ~ _ t_ _ - ) ~ — ~ 0 d 0 0 ~ ~ m I I + I + o ~ O ~ 00 0 0 ~ _ _ _ _ . . . O 0 0 1 1 + ~ ~ _ 00 _ _ O — . . . . O 0 0 0 + + + 1 1 ~ ~ O O e G E 3 c 0 c 0 5 U - y, cy a 0 ~t _ ~ <, ~, 0 ') _ ~ ° e. C) Z —) m} C~l —} - ) ~ — — Z ~ ~ ~ ~ o 0 V P" C~ 1 ~ Ct ~O ._ 1 ~_ ._ C~ 1 t5

227 . ~ a~ - Ct ._ s~ C~ Cq 50 o o - m .o U. Ct .o l s~ 1 1 1 ~ 1 - 4 1 ~ 1 _ 1 Ct 1 C) ~ ._ 1 ._ - C) ._ S~ 4_ _ ~ +1 00 ~ 11 m ~ ~ F" - - o . C~ U) C) 04 - o . U. 0~ - .o C~ ~q 0 C) C C) ~ C > c ~ ~ ~ ~D — ~ ~ ~ oo Ct O O C) ~ o~ Oo _ (: ~: Ct O ~ ~ ~ 00 00 C~ O _ 00 ~ X ~ ~ ~ O ~ _ _ t_ ~ ~ _ _ . . . . . . . . O O O O O O O O *m + + I + I I I + V) 00 ~ ~ U~ ~ O ~ ~ 00 ~ _ ~ ~ _ t— ~ 0 ~o U~ ~ ~ ~ ~ ~ °. O ~ ~ — m + + I + ~ 0 0 - O O ~ _ ~ ~ 0 ~ ~ ~ ~ -o ~ ~ CQ V, ~ m + + I ~ ~ m + + ~ ~ ,, 0 ~ ~, C~ C~ 1 — S~ ~ ~ ~ O O C~ ~ ~ o o _ ( 0 _, . _ _ ._ . ~ I (~ 0 0 ~ V C) C~ z z 00 o - * 0 m + _ ~ _ ~ ~ ° 0 ~ 0 ~ _ m + _ ', 00 ~ ~ ~ U~ 0 — ~ _ 0 ~ ~ ~ _ _ 0 C~ — 0 0 0 . . . . . . . . O 0 0 0 0 0 0 0 + + 1 + + + 1 + ~: o .= ;^ ._ ~ ._ X =. ~ ~ ,,4 D E ~ - `, - z ~ c,, ~ D C ~ ~D ~ c: oo ~ \0 ~ ~ o ~ ~ O ~ - , ~ _ _ _ ~ ~ Z ~ '; ~ 1 O C: ~ Ct O4 1 I ~ O O .= C) V) o o V Ct Ct C) ._ ._ ~Q

228 Ct l ~ ~ .o 1 ,,, ~ o~ 1 ~ 1 ~ o o 11 Ct a' . ~ 1 Ct 1 - g 1 ·= 1 ~ ct 1 os 1 I. ~ - 1 b4 1 Ct o o ~q C~? ~ o o _ [, _ ~ ', _ o ~ ~ ~ t— ~ o ~ ', ~ ~ V) * · ~ ~ o o m- I I I + — 0 ~ 0 0 x ~n 00 U~ . . . . . O U) ~ — O m+ I I I + ~ — ~ ~ _ ~ ~ ~ 00 U~ ~ ~ ~ ~ ~ ~ 0 oo ~ cr~ oo . . . — _ 0 0 1 1 1 1 0 ~ O ~ O O . . . . O ~ 0 0 1 1 1 1 ~ ~ 00 1 ~ o ~o ~ ~t ~.O [_ - ~ —~ . U~ ~ ~ ~ O ~ ~ ~ 0 00 ~ ~ 00 Ce 0 0 * _ ~ ~ 0 0 _ _ O O cn ~ ~ ~ m + ~ ~ ~ + ~ ~ ~ ~ — 0N 0\ 0 0 . ~ ~ ~ 00 V) X ~ ~ . . O O O V~ ~ _ O mo m + I I I + ~ j _ o I ~ O ~ O ~ O O . . . . O ~ 0 0 1 1 1 1 ~ ~ 4.) ~ _ O l 0 0 0 molm + ~ - ~ ~ - ~ v~ ~ ~ - . . . . . * o o o o o m + I I I + - \0 0 oo ~ ~n — . . . O 0 0 1 1 1 1 ~ — 1 ~ ~D ~ ~ ~ oo — ~ I_ . _ o oo ~ ~ U~ ~ ~ o G) ~ ~ O O O O — O O o o o o o o o o o o 1 ~ ~ 1 I_ + 1 1 1 + 1 + 1 1 1 1 &= I C.) ·— ~ ,_ C:~O o ~ ~t ~ ~ oo Z 1 ~) ~ ~) ~ C-l o ~o ~ ~

229 l ~ o ~ ~ ~ ~ ~ 1 _ _ _ ~ _ _ ~o ~ ~ ~ ~ ~ 1 ~ _ ~ ~ ~ ~ ~ 1 t~-) ~ ~' d d — O . . . . . . . o o o o o o o + + + 1 1 1 1 1 _ o ~ ~ ~ c~ 1 ~ ~t ~ o o o o o oo o _ o o o~ ~o o! ~ ~ o ~, o C~ o C~ ~ ~ U~ ~ _____ _ 1 — — ~— ~ 1~ ~C) ~ 1 t~, ~. ~ ~ _ o . . . . . . . o o o o o o o + + + ~ ~ ~ ~ 1 o o U~ V~ ~ oo ~ ~ ~ o ~ ~ ~ o o o. o. ~o X ~ ~ o o o o o~ o~ o~ C~ ~ o ~, o o o . . . ~o o oo ~ ~ ~ ~ ~ Q ~ r~ O ~ o o ~ o . . . . . . . o o o o O o O + 1 1 1 1 1 1 V) - l l _ j c: ~ 1 o O e 2 ~ ~, ° V j ~e a ~ · 0 ~—U O ;) Y . _ ~ ~ a °O ~ 1 ~ ~ O — ~ Z O ~ ~ ~ —

230 C~ P~ _' +1 ~o 11 . ~ Ct ._ Ct 1 - ~: o C~ U~ a~ 04 Ct - —^ 1 =° ~ 1 ~ ~ ~ C) ~, 1 m ._ Ct ~: ._ ._ m _. C) ._ o ._ m m E~ C) ~ Ct O O _ _ C: o~ o _ ~ ~ C - .) Ct _ O V) C~ . o Oo _ c o U. 50 = 0 0 V~ V) ~ C) ~ {oL, o _ (: ~ ._.O C~ C~ _ O O C) ~ C) eD Ct (L) ·— Ct ~ >> C) Ct z \0 C~ 00 ~ 00 0N ~ O ~ ~ 00 ~ 00 O O . . . . . . . ~ ~ O O O O O 1 1 1 + + + 1 _ 00 X X ~ O . . . . . . 00 ~ ~ ~ 00 ~4 \0 — 1 1 1 + + + *m ~ ~ ~ ~ ~ 00 0 V) ~ ~ 00 0\ ~ _ . . . . . 0 0 0 0 0 m I I I + + C~ 00 00 o 1 00 ~ O 0 ~ _ . . O 0 1 1 0 C~ ~ C~ . . ~ U~ 1 1 v0 U~ ~ O O _ _ _ _ . . . . O 0 0 0 + 1 1 1 0 ~ X _ _ ~ 00 00 ~ C~ ~ ~ u~ ~ ~ ~ _ ~ ~ ~D . . . . . . . . . C~ O ~ U~ \0 ~ O ~ ~ ~ ~ ~ 00 * ~ _ m I I I + + + I m_~' ~D _ ~ C~ t— ~ _ _ . . . . . * _ O O 0 0 m I I I + + \0 — ~ ~ _ 00 0 . . . . . C~ ~ 00 ~D ~ ~ - m I I I + + _ oo ~D O _ - , C~ O O — . . . . . . . O 0 0 0 0 0 0 ,_ 1 + 1 + 1 + + e ~, 3 ~ C C u ' ~ o ° z ~ ~ CQ ~ CQ ~ ~ ~ ~ 00 ~ O _ _

231 C~ ~ ~ _ _ _ _ oo o ~ o ~ oo ~ _ _ _ o o o . . . . . . o o o o o o + + 1 1 1 1 _ ~ ~ ~o ~ ~ t_ t~, ~ ~ ~o o o ~ o o C~ ~ ~ ~ o oo + + 1 1 1 1 ~ ~ — o oo o o + X oo ~ ~ 0 ~i 0 + ~ ~,, ~ ~ ~ G ~ ~ U~ 00 ~ ~ 0\ O ~ O O _ O CO ~ ~ — — O ~ — ~ _ — O O — O O . . . . . . O O O O O O 1 1 + + 1 +, :: ~ a O c 2 e ° "2 c, c 0 `< ° ~ ~ 2; — ~ ~ °O ~o z u :E [L, =, o o o V ce c~ ._ 04 ._

232 1 1 of JO tl _, ~: c~ II c: o ._ c~ oo ct :: . ~ .-o n ~ ~4 . - ~: ct Pt ~ O4~ ~ _ 0 c~ ~ ~ o o ~: ~: _ 0 c~ v~ ~ o~ Oo _ (-, _ e=e ~ D i9 u~ r.~ ~ ~ $ v' ~ e - m ~ ~ u, 1 o c: c.; ~ o~ 0 ~ ~ o ._ ~ o ~D - ~ ~ o ~° -= oo oo . . . * 0 0 0 m I I I 0 0 ~ 0 . . . O 0 0 + + 1 0 x 0 ~ 0 ~ ~ ~ cn 0 0 0 o. 0 ~ ~ 0 m Ol ° ° ° ° ° ~ ° ~ * 0 m I v~ ~ ~ x oo 0 0 . . . . O 0 0 0 1 1 + + 0 oo ~ c 0 ~ x - , 0 0 0 ~ ~ ~ 0 ~ o~ ~ m 1° ° ° °+ °+ ° ~ 0 . . O ~ 0 - oo x 0 0 . . . O 0 0 + + 1 5 00 ~ ~0 _% _ ~ ~ O ~ "a ~ au ~ ~ O ¢, _ e a `: ~ c °u~ oo 0 x E ~ c ~ ~ ~ ,~ c~, - , _ Z ~ ~ ~ ~ - o° V° ct _ c: ce 04 ._ v' 5

233 Y~ Ct Pt o +1 V~ cr 11 ~: e~ a~ . ~ - D .= Ct . ~ - ~4 .s 4 - Ct U' P" 1 C~ U. Ct Ct C) ·_4 o .~ m m o . C~ ca 00 Pt Ct O O ~ ~ o o _ c ~ y ~: Ct Cd O O U. V' ~ Ct - o t4 o" o _ (, o C~ ca 04 C) :: ~ ~ D 'D ·- 1 _ ~ ~ O O V~ V~ ~ O C_) c: ~ .0 C, ~ ._ k ~ o o a~ z zo _ t— X o _ . . . o o o + + + ~ ~o X ~ ~ - , o X ~ o ~ U~ m + + + :5 0 0 ~ ~ _ ~ ~ — 00 0 _ . . . * O O O m I + + ~ \0 ~ X ~ ~ ~ ~ ~D 0 X U~ O ~ U~ m + + + ~. 0 0 ~0 ,~ ~: ~: _ ~ ~ 00 0 _ . . . * 0 0 0 m + + + U~ 00 0 ~ - ) 0 X ~ 0 ~ 0 m + + + ° 0 0 ~ O ~4 ~ C~ . . . O O .o + + + ~ _ ~ ' ' e O Y ' Z ~ ~ ~ oo. o v - C~ - :: Ct C) a~ .= 0o ._ V' 1 ~

234 _ ~4 . Al In of 11 Ct . ~ - ._ Ct . ~ - - ~4 o ~ 1 Cal Al Cal - C) .s US Cal o ._ m 00 m - o . Cal 0 - 0 Cal Cal 0c :: o C~q C~ o4 C) ~: 1 ~ 1 ~ ~: c~ ~D ~ Ct a., ~ C~ I ~ ~ ~: C~ _ ~ ~ U~ O ~ X ~ - ) ~ ~ ~ ~ ~ ~ ~ O Ct O O * O O O O O mmo m + ~ + ~ ~ 1 0 ~ ~ \0 .° m + I + 0 ~ ~ ~~~ oo , _ ~ ~ ~ U~ 00 ~ 0 ~ ~ —~ _ O—~ ~ _ 0 0 * 0 0 0 0 0 0 m~o m + ~ + + ~ ~ —t—- - 0 _ \0 ~ a° ~ ~ 00 ~~) O~ ~$ (~} (~N ~ _ _ o~ o 0 0 ~ ~:\ 0 . ~— ~ r— ~ ~ m + ~ + + ~ ~ ~ ~ o 00 ~' ~: _ ~ o O cq C~ O ~ — .° _ . _ _ ._ ~ C~ U~ 0 _ r~ O _ _ . . . * 0 0 0 m+ I + \0 ~ 0 0 ~ ~q ~ C 0 _ ') cr~ oo ~ ~ c~ ~ O _ ~ . . . . m + I + 0 ~ O ~ O 0 ~ r~ 00 ~ C~ ~ ~ V~ 0 0 0 0 0 0 . . . . . . . . O 0 0 0 0 0 0 0 + 1 + + 1 1 1 1 C - ~I~ c ° ° ' -^ ° C ~ 3 c = c~ 2 c ~ c~$ ~ ,, ~ ~ ° ~ ~ ~ 1 ~) ~~ _ <~ ~s) 4~~ ~ oo 3 ~ O ~ ~ ~ C~ C~ 1 ~ ~ ~ - 1 ~ ~ — ~ Z V ~ C _ _ Ct Ct C~ C) ._ ._ `: b0 0 ._ ._ U) U) o o V

~: e~ ~ 1 ~ ~ o . - c~ c) o4 o c~ - ~ o +1 c~ 11 . ~ oo ~9 e~ · c~ . ~ . ~ o ce l s~ u, c~ a~ - '~ 1 ·s | ,= 1 c~ o · - m cr 1 ~ 1 ~; E~ ~: . U. ca C) .o C~ C'2 04 C) j ~ _ 1 c' ~ e~ (L) ·— Ct >~ ~ =4 ~ ~ _ O CQ cn ~ o Oo - a~ ~ 0= C~ V) 00 C~ ~ ~ _ ~ _ O ~ ~ ~ ~ ~ _ _ _ _ _ _ _ _ . . . . . . . * O O O O O O O m I I I + + I I 00 ~ ~ 00 V) O ~, ~.o r— V-} _ 0 0 t— ~ _ 0 0 0 0 ~ ~ ~ · · C~ C~ ~ —) 0 0 0 0 0 0 0 m I I I + + I I ~ ~ 0 ~ _ rY Ct _ 00 _ _ _ . . 0 0 *m I I 00 ~ _ _ 0 0 1 + ~D — c~ CJ~` ,o 1 . ~ ~o oo ~ ~ \0 ~ 1 c~ < - oo ~ _ 0 ~ ao 0 1 o~ co O O O O ~ ~ ~ | cn ~ m I I I + ~ ~ o ~ -~- 1 oo oo ~ 1 1 O O * 0 0 0 0 ~ vi ~ ~ m I I I I I ~ 0\ ~ ~ ~ O O _ O O ~- O ~ 0 ~ ~ ~: O . _ C) Ct _ ._ C) ~ O O | ! ~1 . _ ~ o o _ _ _ . . . o o o , 1 1 1 ._ 3 c ~: . ° Ct o ._ ~ o ~ ._ o ._ — ~ Pt oo — ~ a~ C~ o C~ Ct C) ON O ~ _ _ o ~ ~ oo o~ _ _ _ o o . . . . . o o o o o 1 1 + 1 1 o cn Ct 1 I . ·_ . _ o . _ oo ~ ~ — C) C~ o x o~ o o 1 C) C5 ._ Ct ._ _ ~ _ _ 235 a~ C) ~ — ~ .= ._ o C) ~ , z =~ oo o ~ o o ·V, o V V · ~ O 5 _ ~ e : C,) Ct: c) X ~ V)

236 o tl . 11 - . ~ - Ct ._ . ~ ~4 ._ Ct ._ V, U. U) Ct e~ C) .s C) o ._ m o m _. o . C~ C) o4 o . U. C~ S~ ~4 C) ~. o . U' ca - Ct Ct ~: ~ 0= V~ U o o Ct ~ O O I O O - Ct ~ _ _ _ 0- ~ _ _ ~ _ * O O O m+ ++ C~ _ ~ .0 ~ O 0 0 _ 0 0 0 m+ ++ ~ oO =_ * m _ 0 00 O _ _ _ _ 0 . . . . . O 0 0 0 0 + + + 1 1 00 ~ ~ ~ O _ 0 ~ ~ — 0 0 0 0 0 . . . . . O 0 0 0 0 m + + + I I _ C~ 1 _ _ ~ 0 * _ _ m 0 0 0 00 ~ . . . . ~ _ O ~ 00 o~ ~ ~o ~ o ~ ~ c~ ~ ! m 0 0 -° ~ ° ~ ~ :: ~ .—o c) ~ . - c~ c) o o ce z s~ oo ~ o ~ — —) —) a ) - c;} - - - o o ~ . . . . . . . . o o o o o o o o + 1 + + + 1 1 + ~ w ° V v r ~i ~.' D ~ ~ ~, ~ <,, ~ ~ ° _, ° ~ O c ' a ~ ~ _ \~) ~ ~ ~ C ~ c C ~ ~ 41) D. 'c ~

237 rS~ _ +1 .o . ~ 11 . ~ o - Ct ._ Ct . ~ o ._ - r~ Ct ·_ S~ cq C~ E~ ~: ~0 1 C~ e~ ·_ - Ct - o ._ o ._ U. U. 04 e~ C) .s o ·_ m ~ _ C) ~D m =.= Ct E~ ~ ;>. cs 1 ~ I C 0= 0 ~ V) ~ 0. ~ o o _ (U a~ ~ Ct O O o~ o _ O ~: C~ Pt ~ O O O ~ ~ — .o _ . _ {._ O O O O _ ~ _ . . * O O m I + 0 o 0 ~ 0 X 0 _ O m I + I OX _ . . * 0 0 m I + . . ~ _ m I + 00 * 0 m I ·0 00 0 _ 00 ~ ~ 00 . . . . ~ ~ 0 ~ ~ ~0 0\ 0 — ~ ~ ~ ~ x . . . . ~ ~ o ~ ~ oo ~ ~ ~ oo ~ v~ . . . o oo o ~ ~ m I _ ~ _ 00 ~ 00 0\ \0 ~ 0 0 . . . O 0 0 1 + 1 r~ c~ ~ ~ . _ a C - ~ ~ ·~ 0 ° f3 ,~ ~ O O e ~ . 0 _ 0 0 . . O 0 V V _ _ C;: Ct _ _ C: Ct `: ~ 0C 04 ._ ._ ~ V) ~ .0

Methods of Relating Food Consumption to Nu~idonal Status and Health: Dietary Methodology VICTOR M. HAWTHORNE S UM MA RY New national intiatives in promoting health and preventing disease in nutri- tion suggest need for awareness of the limitations of dietary methodology and its use in epidemiological studies designed to assess efficiency and effectiveness, determine "normal" values, compare rates, validate tech- niques and decide priorities. Use of the four main quantitative methods of collecting dietary data are reviewed in relation to their use in cardiovascular disease, cancer, diabetes and obesity. It is concluded that in epidemiological studies of nutrition, reliance can be placed on a combination of estimated current record, 24-hour recall and diet history adapted in frequency and extent to meet the design needs of indi- vidual studies. Quantitative assessments can be improved by use of new specifically designed food tables to derive nutrient values; and qualitative data collected from food frequency studies through a scoring system restricted to specific foods. The need is postulated for a nutritional health information system incor- porating facilities for medical record linkage allowing outcomes to be as- sessed against baseline values. INTRODUCTION Nutrition is the sum of the processes by which man uses natural and artificial substances ingested for the growth, maintenance, and repair of vital proces- ses. Nutrition is related to metabolism, in itself the sum of all processes 238

Relating Food Consumption to Nutritional Status and Health 239 concerned with the anabolism and catabolism of protoplasm. Food, and more generally diet, cannot be considered in isolation from nutrition, which is the ultimate expression, for better or for worse, of what has or has not been taken into the body. The aims of this review are to examine some adquacies and deficiencies in the use of dietary data as a means of assessing the extent to which nutrition is a determinant of cardiovascular disease, cancer, and allied chronic condi- tions; and to appraise at present the best use of these methods to monitor the effects of nutritional change on human health. Epidemiologically the spectrum of nutrition corresponds to health and disease. Modified by heredity and subject to environment, nutrition depends on the same demographic factors. Studies must therefore be population- based and encompass optimal and suboptimal as well as frank, malnutri- tional states. Indeed, as in mild hypertension, intervention in suboptimal nutrition may offer the best interventive strategy. Epidemiology uses whole populations as denominators to provide the statistical basis for determining "normal" values, to compare rates in different populations, to validate diagnostic techniques, to implement field trials, to decide management priorities, and to coordinate the provision of appropriate, multidisciplinary skills in medicine, biochemistry, anthropology, physiology, sociology, de- mography, and biostatistics. GENERAL METHODS Literature purporting to describe the relationship between food intake pat- terns and chronic disease is as extensive as it is inconclusive, but describes a number of techniques common to the acquisition and use of dietary data. Dietary intakes of households and individuals have been measured since 1900, but studies designed to relate diet and nutrition to health and disease were uncommon until the 1930's. Burk and Pao's review (1976) of the methodology for large-scale surveys of household and individual diets criti- cally evaluates the four main quantitative methods of collecting data: the weighed record, the estimated record, 24-hour recall, and dietary history. Employing an extensive range of validating criteria, they conclude that no one method was consistently advantageous over all others and no selection could be made independently of study objectives. Marr (1973) comments that, although details of weights or standard mea- sures of food can be recorded at the time food is eaten, the demand on the participants adversely affects response. Recall makes least demand, but has the disadvantage of placing too great reliance on memory. Current record methods, on the other hand, avoid inaccuracies due to memory, but the study procedure itself may influence eating habits.

240 VICTOR M. HAWTHORNE Both recall and record methods collect details of foods eaten, and tables of food composition must be used to derive nutrient values. The basis for calculation of these tables is suspect due to variation in conditions affecting growth and processing of food not provided for in the tables, a deficiency that can be overcome when comparing groups by using only the extremes of the distributions. Selective response in certain age and socioeconomic groups are another source of bias for which provision must be made. Day- to-day and week-to-week variation of diet affect the precision of most methods and influence considerations establishing the shortest period over which recall can be used to estimate dietary intake. The greater within- person and day-to-day variation, compared with between-person variation, the longer should be the period of dietary recording. The type of nutrient introduces a further element of variation. The measurement of nutritional status is in effect a measure of outcome. Dietary data must be combined with biochemical and clinical measurements with all their implications for cost in response and resources. As a general guide food intake can be used as a valid index of daily energy expenditure, but estimates must extend over a minimum period of a week. CARDIOVASCULAR METHOD The relationship of diet to cardiovascular disease was reviewed by Rose and Blackburn (1968) in the late 1960's. Like others, they fail to identify any single approach to assessing dietary intake but list the main methods in common use for epidemiological studies of ischemic heart disease. These again include direct measurement of food intake in which all food eaten by an individual or family is weighed or measured by a dietician in the home. This method is limited by time and cost. Results may not reflect seasonal variation and may be biased by the investigative process. The same reserva- tions apply to the chemical analysis of food intake by collection of an approximate aliquot of each dish or food eaten. Twenty-four-hour recall is described as simple, direct, and inexpensive. A tendency to underestimate intake may be remedied by adjusting the number of days covered to meet the objectives of the study. Food diaries over 1 week with entries made at mealtimes and interviewer review either with or without weighing had already been adapted for postal use (Keen and Rose, 1958~. Response problems, however, restricted use to literate populations. Dietary history can be recorded on questionnaires (Burke, 1947) using a method incorporating cross-checks and with 24-hour recall as a further check. Apart from being time-consuming and expensive in staff and re- sources, the method needs careful work-up. Only usual diet is measured,

Relating Food Consumption to Nutritional Status and Health 241 past changes are excluded and much depends on the ability of the participant to define and recall usual food intake. Abbreviated methods of collecting dietary data cheaply in epidemiologi- cal studies had not been attempted in the sixties, but have now been assessed (Hankin, 19781. Instead of quantitating nutrient intake, frequency of intake of specified food can be used to develop a scoring system. This method has been used with 24-hour recall in the Tecumseh Population Study (Metzner et al., 1977; Nichols et al ., 1976a,b). Epidemiological studies have been used to arrive at mean values for population groups using these methods and to study the effects of dietary intervention on biological indices such as cholesterol. In these studies diet has to be assessed over the course of a year using statistical models like those of Becker and his colleagues (Becker et al., 1960), who studied food intake in a time series approach. No satisfactory method of measuring long- term diet retrospectively has yet been established, and long-term prospective designs with interval sampling are recommended. In assessing nutrients of special interests to cardiovascular disease, stan- dard food table procedures have been inadequate. New chemical analyses now being made for food tables in respect of special terms of nutrients may enhance the value of food tables in future studies. Rose and Blackburn (1968) comment on the relation of data collection to computer processing. They list the staging of data collection from tabulation of the quantity of food eaten in multiples of a predetermined unit of volume or weight dependent on the food tables to be used; reporting of food intake for a predetermined period; grouping of foods into combinations of similar nutritive value, perhaps using abbreviated food tables for this purpose; and the recording of other variables such as number of meals per day and meals away from home. As already mentioned, where nutrition is to be included as a form of outcome the data must be associated with biochemical, clinical, psychological, and sociological variables. Computer programs were already available from the Heart Disease Control Program of the U.S. Public Health Service and the Coronary Prevention Evaluation Program. The role of diet and various dietary factors in ischemic heart disease has been reviewed recently by a working group on ischemic heart disease of the Chief Scientist Office of the Scottish Home and Health Department (Chief Scientist Organization, 1979~. Commenting on differences in Scottish food consumption compared with the average for Great Britain, the working party emphasized the need for analysis of the diets of randomly selected small groups of the Scottish population in areas of contrasting rates of ischemic heart disease mortality and socioeconomic opportunity.They regarded diet- ary analysis methodology as of critical importance and proposed utilizing

242 VICTOR M. HAWTHORNE the more recent tables on food consumption that include new items such as vegetable fiber and polyunsaturated fatty acid as a means of producing more accurate dietary studies by computer. They recommended prospective re- cording of a week's food and drink intake supported by appropriate labora- tory procedures, including such techniques as the analysis of adipose tissue samples to permit dietary fat saturation level assessment. The review of nutritional assessment in health programs that emerged from the Proceedings of the Conference on Nutritional Assessment (1972), sponsored by the American Public Health Association in 1972, refers to pathological evidence of frank atherosclerosis in the arteries of young Americans in Korea and Vietnam as indicating the need for earlier identifi- cation of nutritional factors that may be disease precursors. In underlining the need for nutritional studies in cardiovascular disease, obesity, stroke, diabetes, and cancer, they draw attention to the possibility of certain foods and food additives being related to cancer of the gastrointestinal tract. Nu- trition is also related to gout, hyperlipidemia, various anemias, intestinal enzyme deficiencies, alcoholism, arthritis, and accidents. They delineate four major nutritional classifications to which nutrition may contribute: those with an excessive intake of calories or certain nu- trients (usually carbohydrates); those with deficiencies of essential nutrients or calories; those with special nutritional problems, such as those occurring in diseases like ulcerative colitis, regional ileitis, alcoholism; and those with nonnutritional problems. They point to the need for guidelines to determine the extent of malnutrition (in itself an outcome) and its outcomes and for a broad system of dietary, clinical, and biochemical data collection. At a basic level, information on an on-going epidemiological approach to cardiovascular disease involving nutrition is still available in the Tecumseh Study (Metzner et al., 1977; Nichols et al., 1976a,b), where stratified random samples of approximately 4,000 males and females between the ages of 19 and 25 years in a general population who had been clinically examined in 1960 had a dietary assessment in 1967. A card sort of 110 foods determined frequency, and a subsample in 2,000 participants completed 24-hour dietary recall. Adiposity was related to diet and food frequency to cholesterol. The study is prospective and may be expected to yield cumulative outcome data based on mortality. More frequent food intake appeared to be associated with less obesity. CANCER METHODS Morgan and his colleagues (Morgan et al., 1 978) evaluated three methods of estimating group and individual diet consumption in a case control study of diet in breast cancer. Their review of previous work in which nutritional

Relating Food Consumption to Nutritional Status and Health 243 methods were applied to epidemiological problems excluded the collection of nutrient data by methods requiring weighing and laboratory analysis as impractical for survey research. They restricted consideration to methods useful for field investigations aimed at identifying the dietary characteristics of large groups. As in general methods of cardiovascular disease, the methods included the food frequency interview and inquiries about usual intake in terms of frequency of consumption of various food items, a method considered by Graham and his colleagues (Graham et al., 1967) in a study of diet and gastric cancer as the most feasible annroach riving a mean reliability of 81 percent. Or r ~ =- - ---= Use of the quantitative research dietary history developed by Burke to record the subject's usual pattern of eating, as used in the Framingham Study where the interview was not structured, was considered to have failed to elicit all the required information. The Israel Ischaemic Heart Disease Study (Abramson et al., 1963) also was not supportive of this method. The diet record of foods and beverages consumed by the respondents was directly related to current intake, but was restricted to use by literate people. Three days was regarded as the minimum time for a fair picture of food intake of the individual. The diet recall interview for the 24-hour period was believed to give fairly reliable data on current food consumption of a group. In their evaluation, the authors compare the efficacy of 24-hour recall, detailed quantitative diet history directed towards the most recent 2-month period after diagnosis and the 2-month period 6 months before diagnosis, and a 4-day diet diary. They found in this case control study of diet and breast cancer a high degree of correlation between the estimates of food consumption for the controls using each of the three methods. The highest estimate was obtained from diet history, with a slightly higher estimate in the period 6 months before diagnosis than in the current period. The lowest estimate was in 24-hour recall. They concluded that all three methods were applicable to case control studies, but diet history was preferred when cur- rent food intake may be influenced by a disease frequently the case in cancer. A closer examination of the methodology of nutritional assessment in relation to cancer is provided by a review of diet and breast cancer by Jean Hankin and her colleague (Hankin and Rawlings, 1978~. Their review em- phasizes the need for methodology to encompass the usual demographic factors and relate diet to outcome in high and low geographic areas of prevalence. The need to assess the possiblity of tumor initiation or inhibition in relation to diet, nutritional status, and hormonal pattern for and during teenage years is proposed, as well as the need to include genetic and en- vironmental factors in studies of families who share genes as well as com-

244 VICTOR M. HAWTHORNE mon environmental factors, including food intake. The excess of estrogen in obese women is considered a factor, although there has been no consistency between obesity measures and breast cancer. Weight and height individually were risk factors in prospective breast cancer study in women aged 55 to 74 years, but the association was weaker using the Quetelet's index (weight per height) (Marr, 1973~. Height only was found constant for all risk attributable to body size. Geographical variations are dealt with by correlation of breast cancer incidence with dietary and other environmental variables on a per capita basis, principally in relation to fat intake. In one study sugar was positively correlated and starch negatively correlated in per capita nutrient intakes in 10 countries with cancer mortality in single and married women. The per capita food and nutrient intakes were reported by the Food and Agricultural Organi- zation (FAO) of the United Nations and were based on estimates of domestic food production, imports and exports, military supplies, and changes in stocks and reserves. Standard food composition values were then applied. These food estimates only reflect the amounts available per person and not the quantities consumed. The distribution of food by sex and age within the country are not described. Studies of cancer rates in ethnic groups are helpful in formulating hypoth- eses. Breast cancer in first- and second-generation Japanese women living in San Francisco compared with Japan incriminate increased intakes of fat and its nutritional correlates. Stocks (1970) examined the regional variation of breast cancer death rates in England and Wales and the higher rates in the South compared with the North of Wales using per capita food intakes from annual national food surveys and cancer mortality. He found positive as- sociations between milk, butter, and cheese and negative associations be- tween other fats such as margarine. Nutrient intakes were not reported. Food frequency was collected from 100,000 Seventh Day Adventists in California and followed for 4 years by Philips (1975~: 77 breast cancer cases were each matched by age, sex, and race with three controls from the study population. Comparison of the case-control dietary data showed that the frequencies of eating 5 food items were associated with breast cancer: fried potatoes, hard fat (butter, margarine, or shortening) for frying, all fried foods, dairy prod- ucts except milk, and white bread. Relative risks ranged from 2.6 to 1.6. It was considered difficult, however, to draw inferences from qualitative food frequencies. Quantitative data were used in a case control study in Canada where Miller (1977) collected data on usual intakes of participants. Mean nutrient intakes in 309 matched pairs showed difference in calories and total and saturated fats significantly higher among cases than controls. In mice it has been suggested that tumorigenesis inhibited through a 30

Relating Food Consumption to Nutritional Status and Health 245 percent reduction in calories is due to lack of energy (Tannenbaum, 1942), or possibly through pituitary insufficiency leading to a decrease of estrogen production and lower sensitivity leading to a decrease of estrogen production and lower sensitivity of breast cancer to estrogenic stimuli (Huseby et al., 1945~. In animal experiments, a one-third reduction in calories was necessary for tumor inhibition, but it is possible that inadequate calories and undernu- trition in many developing countries of Asia and Africa might explain their low incidence of breast cancer. Hill and his colleagues found persons on high-fat diets had higher propor- tions of anaerobic bacteria in the intestinal microflora and secreted more biliary steroids than those in low-fat intakes (Hill et al., 1971~. These bacteria are able to synthesize estrogens such as estradiol and estrone, po- tentially carcinogenic to the breast. In animal studies the inhibitory effects of diets low in protein or selected amino acids have been considered to be related to changes in the im- munological system (Alcantara and Speckman, 1976~. It has been hypothesized that cellular immunity is a major defense against cancer, whereas humeral (antibody) immunity enhances tumorigenesis by depress- ing cellular immune responses. It is suggested that a low-protein intake might suppress antibody production, but have no effect on cellular immu- nity. There has been little evidence that different amounts of vitamins and minerals influence the incidence of mammary tumors. Morris (1947), on the other hand, observed that a deficiency of riboflavin decreased the number and growth of tumors in mice, whereas a supplement increased incidence. This has been confirmed in other studies, but the mechanism is unknown. MacMahon and his colleagues have reviewed hypotheses relating breast cancer to viruses, genetics, exogenous hormones, and radiation. Diet or nutritional status might function as a modifier of mammary tumorigenesis. Sections of the review (Hankin and Rawlings, 1978) consider antigens, estrogens, and prolactin. It is concluded that human breast cancer is mul- tifactorial and that it seems highly probable that dietary factors are related to the disease in several ways. Metabolism of nutrients, synthesis and activity of several hormones, and the role of fatty tissue may be common areas of interaction that might favor mammary carcinogenesis. Dietary studies in animals should not be extrapolated to humans. The National Cancer Institute, National Institutes of Health, Diet, Nutri- tion and Cancer Program report (1977) identified diet as an etiologic factor distinct from dietary contaminants and from genetic and environmental fac- tors as playing a prominent role in the causation of certain major forms of cancer on the basis of the correlation between incidence and dietary habits in numerous studies of migrant populations, high- and low-cancer incidence populations, and special population groups and by studies of trends in

246 VICTOR M. HAWTHORNE cancer incidence and diet habits in the United States, Japan, and other countries. The report concludes that the most likely clue to the etiology of colon and breast cancer is diet-dependent, because these conditions do not appear to be affected appreciably by environmental polluters, food contam- inants, smoking, or occupational exposure. It is also considered that the epidemiological evidence is reinforced because many of the relationships have been found to apply to cardiovascular disease (Stamler et al., 1972; Dayton and Pearce, 19691. This hypothesis is supported by experimental evidence of dietary factors in which calorie restriction in animals generally inhibits dietary fat and promotes tumorigenesis in the pregenerative stage of hormone-dependent tumors (Carroll, 1975~. Modan and his colleagues in Israel report a higher consumption of starches among gastric cancer patients and a lower fiber consumption among colon cancer patients (Modan et al., 1975~. World population studies are used by Lowenfels and Anderson (1977) to positively correlate dietary factors specifically with colon and gastric cancers. They describe total calorie intake, nutritional excess or deficit, exposure to car- cinogens, and consumption of alcohol as contributory factors. Comparison of spouses of Japanese women with breast cancer was used to isolate en- vironmental from dietary factors prospectively. The spouses of breast cancer cases consumed more beef or meat, butter/margarine/cheese, corn, and wieners than other men (Gutsch and Holler, 19754. DIABETES METHODS The role of diet in diabetes is uncertain in a situation in which treatment does not appear to prolong life expectancy. The epidemology of diabetes in Europe is reviewed by Gutsch and Hol- ler (1975), whose most important contribution has been extension of stan- darization recommendations to facilitate international comparisons, such as the collaborative studies of the relationship of glucose intolerance to car- diovascular mortality involving several countries reported by the Stamlers at the VIII World Congress of Cardiology at Tokyo, Japan, in September 1978. Studies in North America would seem to indicate that the main role of diet in therapy was calorie restriction to reduce obesity in itself proposed to- gether with genetic factors as the major risk factors, especially in the elderly (West, 19731. The reviewer reports studies whose methodologies compare nutrition and dietary patterns in Japan, Israel, and Africa and support a relationship with diet in his own studies in 11 countries and in a review of published data in New World aboriginals: Indians, Eskimos, Polynesians, and Micronesians. A major conclusion from a review of all available

Relating Food Consumption to Nutritional Status and Health 247 laboratory and epidemiological evidence identifies total calorie intake inde- pendent of source as the most important dietary factor in increasing the risk of diabetes. OB E SITY The risk of ischemic heart disease associated with obesity is small compared with other factors (Prevention of coronary heart disease, 19761. When fac- tors associated with obesity, like blood pressure, glucose intolerance or diabetes mellitus, physical activity, or raised plasma levels of triglycerides, cholesterol, and uric acid are held constant, the predictive value of obesity for ischemic heart disease is much reduced, suggesting that when obesity does increase morbidity or mortality it does so through these factors. Use of obesity as an index of dietary intake, nutritional health or disease, and disability status requires provision in analyses for hypertension, hyper- lipidaemia, and glucose intolerance, as well as morbidity from ischemic heart disease. Provision must be made for neuromuscular incoordination, accidents, respiratory and skin disorders, emotional problems, and also for a general increase in mortality. Measurement of obesity requires definition of criteria and standardization of technique. If obesity-associated conditions can be provided for, obesity or overweight might have possibilities as an abbreviated and, dependent of criteria, a simple index of nutritional status. The importance of birth weight as an accurate index of short-term outcomes, with important heredity and nutritional implications in the epidemiological assessment of health status, cannot be overemphasized. DISCUSSION There seems to be agreement that weighing of food is too much to ask of participants in epidemiological surveys, but within certain limits (Madden et al., 1976; Young et al., 1952) reliance can be placed on a combination of estimated current record, 24-hour recall, and diet history adapted in fre- quency and extent to meet the design requirements of individual studies. Quantitative assessments can be made, and the use of new food tables or tables prepared specifically by chemical analysis for individual studies can reduce error in deriving nutrient values. Qualitative data can be collected from food frequency studies, which may be restricted to specific foods in developing a scoring system. Cross-sectional epidemiological studies can be used to relate different "cut points" in the distribution of food or nutrient intake to the prevalence of outcomes like nutritional status measured by physical examination of

248 VICTOR M. HAWTHORNE indices like height and weight or biochemistry taken at the time of collection of dietary data. Relative risks can be calculated from these data, but a method much to be preferred would be prospective, in which outcomes in different percentiles of baseline dietary intake can be measured over time against incidence outcomes like weight gain or loss in nutrition, incidence of angina, or ECG abnormality in morbidity or cardiac infarction, stroke, or cancer in mortality. Only in prospective studies can outcomes against baseline dietary distributions be used to calculate relative and attributable risks for the individual and attributable risk for the community with the requisite degree of accuracy. When mortality incidence is required for chronic disease outcomes, the methodology requires Framingham (Kennel and Gordon, 1970) or Tecumseh-like (Metzner et al., 1977; Nichols et al., 1976) studies on a sufficiently large scale to produce enough deaths or major disabilities within a reasonable time interval. Sample size is increased in proportion to the need to adjust the statistical cells to provide for demographic factors like age and sex and confounding factors like smoking. Medical record linkage may offer a possible solution to the resource difficulties of conventional population studies. The application of validated, abbreviated dietary recall (Hankin et al., 1978) to random probability sam- ples such as those developed for statewide telephone inquiry could provide baseline dietary data in a sample of 2,500, which could characterize a population of 10,000,000. Deficiencies could be reviewed by deployment of the resource represented by the First Health and Nutrition Examination Survey (19721. Linkage to death certifications, cancer registration, and morbidity such as hospital admission data available from CPHA (Wylie and Slee, 1979) could provide appropriate outcomes. The larger the sample, the shorter the time to acquire an adequate number of outcomes to calculate risk. What seems to be needed in nutrition is a more comprehensive and inte- grated health intelligence service. An appropriate existing organization such as CDC might be encouraged to assume responsibility for a more active role in collection and surveillance of national statistics of health outcome in terms of mortality and morbidity having nutritional implications. Access to existing computer tapes can provide mortality and morbidity for relevant ICD chronic disease classifications. These could be analyzed for the usual demographic variables by geographic area down to HSA'S, thus provid- ing a means of identifying high- and low-risk groups for more detailed study by telephone or other methods outlined above. Equally, surveillance of various indices of infant mortality, together with birth certification data, including birth weight when studied in the same comprehensive way, could shorten the long observation periods needed to

Relating Food Consumption to Nutritional Status and Health 249 collect mortality and morbidity prospectively in adult populations. Fetal nutritional experience would seem to offer the only prospect of studying chronic disease within a realistic time cycle for most investigators. Although the permeability of the placental barrier has been adequately demonstrated over many years, the potential of fetal susceptibility and use of cordblood for surveillance has not yet been purposefully used. The best deployment for federal dietary research resources (First Health and Nutrition Examination Survey, 1974; Owen, 1978) would seem to be in prospective studies of high- and low-risk populations identified by the sur- veillance procedure using appropriate combinations of the methodologies described in this paper. The relatively ineffectual nature of past contributions in large part stems from failure to attempt more than cross-sectional esti- mates of population samples under study. There is urgent need to review these practices and use these resources prospectively in high- and low-risk samples. Increased precision in correlating diet and nutrition to health status can best be obtained by observation of the same individual over time. A small national resource could economically and effectively acquire such extra detailed data as might be needed to complement data already available from existing sources. Creation of an integrated nutritional health intelli- gence service might be the best method of relating diet and nutrition to outcome in chronic disease. The overriding need would seem to be integra- tion of responsibility. REFE RENCES Abramson, J. H., C. Some, and C. Kosovsky. 1963. Food frequency interview as an epidemiological tool. Am. J. Public Health 53:1093-1101. Alcantara, E. N., and E. W. Speckman. 1976. Diet, nutrition and cancer. Am. J. Clin. Nutr. 29:1035. Becker, B. G., B. P. Indik, and A. M. Beeuwkes. 1960. Diet intake methodologies. Techni- cal report. University of Michigan, Ann Arbor. Burk, C. M., and E. M. Pao. 1976. Methodology for large-scale surveys of household and individual diets. Home Economics Research Report No. 40. Agricultural Research Service, Untied States Department of Agriculture, Washington, D.C. Burke, B. S. 1947. The dietary history as a tool in research. J. Am. Diet. Assoc. 23:1041. Carrol, K. K. 1975. Experimental evidence of dietary factors and hormone-dependent cancer. Cancer Res. 35:3502-3506. Chief Scientist Organization. 1979. Report into coronary heart disease in Scotland. A report to the Chief Scientist by the Working Group on Ischaemic Heart Disease, Scottish Home and Health Department, Trinity Park House, Edinburgh, Scotland. pp. 8-9. Dayton, S., and M. L. Pearce. 1969. Prevention of coronary heart disease and other compli- cations of atherosclerosis by modified diet. Am. J. Med. 46:751. Diet, nutrition and cancer program. 1977. National Cancer Institute, National Institutes of Health Status Report.

250 VICTOR M. HAWTHORNE First Health and Nutrition Examination Survey, United States, 1971-1972. 1974. Dietary intake and biochemical findings. U.S. Department of Health, Education, and Welfare, Public Health Service, Health Resources Administration, National Center for Health Statis- tics, Rockville, Md. Graham, S., A. M. Lilienfeld, and J. E. Tidings. 1967. Dietary and purgation factors in the epidemiology of gastric cancer. Cancer 20:2224-2234. Gutsch H., and H. D. Holler. 1975. Diabetes epidemiology in Europe. Georg Thieme Pub- lishers, Stuttgart. Hankin, J. H., and M. S. Rawlings. 1978. Diet and breast cancer: A review. Am. J. Clin. Nutr. 31 :2005-2016. Hankin, J. H., V. Rawlings, and A. Nomura. 1978. Methods in nutrition. Am. J. Clin. Nutr. 31 :355-359. Hill, M. J., P. Goddard, and R. E. O. Williams. 1971. Gut bacteria and aetiology of cancer of the breast. Lancet 2:472. Huseby, R. A., Z. B. Ball, and M. B. Visscher. 1945. Further observation on the influence of simple calorie restriction on mammary cancer incidence and related phenomena in C3H mice. Cancer Res . 5:40. Kannel, W. B., and T. Gordon. 1970. The Framingham Study, an epidemiological investiga- tion of cardiovascular disease. Section 24 (circulated). Keen H., and G. A. Rose. 1958. Diet and arterial disease in a population sample. Br. Med. J. 1:1508. Lowenfels,,A. B.,andM. E. Anderson. 1977. Diet and cancer. Cancer39:1809-1814. Madden, J. P., S. J. Goodman, and H. A. Guthrie. 1976. Validity of the 24-hour recall. Research 68:143-147. Marr, J. W. 1973. Dietary survey methods: Individual and group aspects. Proc. R. Soc. Med. 66:639-641. Metzner, H. L., D. E. Lamphiear, N. C. Wheeler, and F. A. Larkin. 1977. The relationship between frequency of eating and adiposity in adult men and women in the Tecumseh Community Health Study. Am. J. Clin. Nutr. 30:712-715. Miller, A. B. 1977. Role of nutrition in the etiology of breast cancer. Cancer 39:2704. Modan, B., V. Barell, F. Lubin, and M. Modan. 1975. Dietary factors and cancer in Israel. Cancer Res. 35:3503-3506. Morgan, R. W., M. Jain, A. B. Miller, N. W. Choi, L. Mathews, L. Munen, J. D. Burch, J. Feather, G. R. Howe, and A. Kelly. 1978. A comparison of dietary methods in epidemiologic studies. Am. J. Epidmiol.6(107):488-498. Morris, H. P. 1947. Effects on the genesis and growth of tumors associated with vitamin intake. Ann. N.Y. Acad. Sci. 49:119. Nichols, A. B., C. Ravenscoft, D. E. Lamphiear, and L. D. Ostrander. 1976a. Independence of serum lipid levels and dietary habits: The Tecumseh Study. JAMA 17(236):1948-1953. Nichols, A. B., C. Ravenscoft, D. F. Lamphiear, and L. D. Ostrander. 1976b. Daily nutri- tional intake and serum lipid levels: The Tecumseh Study. Am. J. Clin. Nutr. 29: 1384- 1392. Owen, A. Y., 1978. Community nutrition in preventive health care services: A critical review of the literature. U.S. Department of Health, Education, and Welfare, U.S. Government Printing Office, Washington, D.C. Philips, R. L. 1975. Role of life-style and diet habits in risk of cancer among Seventh-Day Adventists. Cancer Res. 35:3513. Prevention of coronary heart disease. 1976. Report of a joint working party of the Royal College of Physicians of London and the British Cardiac Society. J. R. Coll. Phys. 3(10). Proc. of Conference on Nutritional Assessment. 1972. Nutritional assessment in health pro- grams. Am. J. Public Health 63:64.

Relating Food Consumption to Nutritional Status and Health 251 Rose, G. A., and H. Blackburn. 1968. Cardiovascular survey methods. WHO Monograph Series No. 56. WHO, Geneva. Stocks, P. 1970. Breast cancer anomalies. Br. J Cancer 24:633. Stamler, J., D. M. Berleson, and H. A. Lindberg. 1972. Pages 41-119 in R. W. Wissler and J. C. Geer, eds. The pathogenesis of arteriosclerosis. Williams and Willins, Baltimore. Tannenbaum, A. 1942. The genesis and growth of tumors. II. Effects of calorie restriction per se. Cancer Res. 2:460. West, K. M. 1973. Prevention and therapy of diabetes mellitus. Nutr. Rev. 7:193-197. Wylie, C. M., and V. N. Slee. 1979. Community health surveillance through hospital dis- charges. Univ. Mich. Med. Center J. 2(44):60-64. Young, C. M., G. C. Hagan, R. E. Tucker, and W. D. Foster. 1952. A comparison of dietary study methods. II. Dietary history vs. seven-day record vs. 24-hour recall. J. Am. Diet. Assoc. 28:218-221.

Anthropometry and Nutnhonal Stables FRANCIS E. JOHNSTON "Nutritional status" is a widely used term, usually employed in some fash- ion by almost anyone discussing the relationship between diet and health. Yet, despite its ubiquitous nature, definitions of nutritional status are dif- ficult to find. Some use it epidemiologically to estimate the prevalence of malnutrition in a population, while others use it clinically to refer to the nutritional well-being of an individual. On the one hand, there are those who focus upon dietary intake in evaluating nutritional status, while on the other, there are those who rely upon indicators of disease as evidence for distur- bances in status. I will use nutritional status to refer to the extent to which the dietary intake has been sufficient to satisfy the nutrient needs. I will also use the term interchangeably to indicate the evaluation of an individual as well as the assessment of an entire group. While I realize that this is a global definition, lacking perhaps in specificity, the choice of the definition is dictated by the topic being discussed, as well as by the orientation of this workshop. Clearly, such a broad definition demands a comprehensive set of techniques of assessment, in fact, a battery of techniques that will transcend any single discipline. While a single set of methods, e.g., biochemical, can yield a picture of nutritional status, and quite possibly one that is sufficient for the immediate purpose, the complete picture requires a broader ap- proach. Anthropometry, traditionally located within the history of physical an- thropology, provides one set of methods and techniques, though only one of many (see, e.g., Jelliffe, 1966; Christakis, 1973; Guthrie, 19751. If I appear to be overselling anthropometry, it is because my task for this workshop is to indicate its usefulness. I hope also to indicate the limitations on an- 252

Anthropometry and Nutritional Status thropometry as a technique, one of several that are utilized in research and in assessment. As a physical anthropologist, I do feel that anthropometry has not yet received its "just due" as a technique for assessing nutritional status. Part of that has come about because of the speckled history of anthropometry even within anthropology. But part of the blame must also rest upon our fascination, in Western culture, for methods developed in the experimental laboratory and in hospitals, and for the assumption that the methods de- veloped in those settings could be exported into the surrounding community, or, worse yet, the remote communities of the nation and the world. Thus, anthropometry is one of the basic techniques for the comprehensive assessment of nutritional status. The information provided by the careful measurement of the body cannot be obtained reliably by any other method. Likewise, the information provided by anthropometric data is valid within the constraints of the method, and, regardless of our wishes, we cannot push body measurements beyond those constraints, no matter how intently we wish or how elaborately we analyze our measurements. 253 ANTHROPOMETRY AS AN INDICATOR OF NUTRITIONAL STATUS The term nutritional anthropometry began first to appear after World War II, but seems to have had its first significant impact with the publication of Body Measurements and Human Nutrition, edited by Josef Brozek (1956), the proceedings of a conference sponsored by the National Research Coun- cil. In his introduction to that volume, Brozek began to establish the link that forms the foundation for my discussion, namely, that body morphology size, shape, and composition reflects the dietary history of an individual and is an indicator of the adequacy of that history, or, in other words, of nutritional status. Accepting that, one central task of the confer- ence was to determine a minimum number of measurements that could provide an estimate of nutritional status, selected from the countless number religiously measured by 250 years of anthropologists. This list has been updated in subsequent years as ensuing research has validated some, invali- dated others, and developed still newer ones. A number of additional con- ferences, seminars, and workshops have concerned themselves with this task, and their deliberations have been duly reported in various publications (World Health Organization, 1968; McKigney and Munroe, 19761. The Advantages of Anthropometry The use of anthropometry in assessing nutritional status offers certain ad- vantages. Some of these advantages are biological, i.e., they provide unique information, while others are practical.

254 FRANCIS E. JOHNSTON First of all, anthropometry provides information on past nutritional his- tory, information that is of value and that cannot be obtained with equal confidence using other techniques. In a more elegant way, we may say that one's nutritional history is embedded in his or her morphology and that this history is revealed through the analysis of measurements of that morphol- ogy. The analysis of body measurements can assist in the identification not only of severe states of malnutrition, but also of the mild-to-moderate mal- nutrition that we all suspect is so widespread, but that is more difficult to document than might be imagined. Next, anthropometry is a relatively "easy" technique to employ, since it is noninvasive and applicable to situations where large sample sizes are desirable. Anthropometric equipment is generally portable and relatively inexpensive, requiring no portable generators, refrigerators, centrifuges, or autoclaves. While considerably more technician training is required than many scientists believe, a few months of careful training and some system of monitoring continued technical performance will suffice to provide a body of data with acceptably low errors of measurement. The Limitations of Anthropometry Anthropometric techniques do not provide a panacea, and anyone using it must be aware of its limitations. They reduce its comprehensiveness as a means of evaluating nutritional status and affect the accuracy of the data obtained. Anthropometry provides only limited information on present nutritional status. While it is true that one's present status is a function of past nutri- tional intake, nonetheless, anthropometry is not able to detect short-term, nutrient-specific disturbances in nutritional status. Second, while body morphology is indeed sensitive to nutritional distur- bances, it is similarly altered by a variety of such disturbances; an- thropometric indicators are not specific. Nutritional status may be altered adversely by any number of factors, primary as well as secondary. Careful measurement of the body will reveal that something has indeed gone adrift, and, if the limiting nutrient is already suspected, anthropometry can quan- tify the degree of nutritional insult, especially if it involves protein or energy. But body measurements will be much less likely to identify a defi- ciency or excess of a more specific nutrient, a vitamin or mineral. Further- more, anthropometry may indicate simply that the morphology has been altered, but we may be uncertain whether the alteration is a direct result of malnutrition, or perhaps of some concomitant of malnutrition such as infec- tious disease, or in the case of overnutrition, of reduced energy expenditure. Finally, anthropometry, in the classic sense, deals only with the surface

Anthropometry and Nutritional Status 255 of the body. Even though techniques such as radiogrammetry and ul- trasonics may sometimes be subsumed under the general term "an- thropometry," such is seldom the case in nutritional surveys. While the noninvasive nature of anthropometry is a decided asset, it also means that the measurements made are indirect. Their validity rests upon research that relates morphological alterations to underlying physiological processes to an extent that yields statistically reliable results. Advantages and Limitations: One and the Same A comparison of the foregoing reveals that, in fact, the advantages of an- thropometry are its limitations. The reasons for adopting anthropometry in assessing nutritional status provide limitations upon its usefulness. The so- lution seems to be to judge the results of this trade-off, exploiting an- thropometric data fully when the trade-off is not detrimental. The setting in which the assessment is to be made, the extent of data to be collected, and the use to which the assessments are to be put, all determine the methods eventually chosen. Anthropometry seems especially appropriate when there is a suspected chronic imbalance of protein and energy. Such imbalances are known to distort the normal proportions of body tissues and to alter the patterns of physical growth, often quite drastically. These alterations become fixed, as it were, and may be identified during the adult years, long after growth has ceased. Thus, nutritional status changes may be evaluated from one genera- tion to the next, analyzing the phenomenon known as the secular trend (Marina, 1979~. The Rationale for Nutritional Anthropometry The initial step in establishing a rationale for nutritional anthropometry has been given above: Variations in the ratio between nutrient intake and nu- trient requirement are recorded in the morphology of the body, perhaps permanently. The greater the disparity, the greater the morphological alter- ations, reaching the extremes of obesity and marasmus. Taking this position to the extreme would require us to accept that all variation in body morphology results from variation in the ratio of intake to need. While there may be some extremists who hold this view, it seems to be an easy one to disprove. Nutritional factors are important determinants of our size, shape, and composition, but others may also be demonstrated: heredity, for example. Fortunately for our purposes, it seems that hereditary factors determining body morphology may be rather easily overridden by

256 FRANCIS E. JOHNSTON chronic nutritional disturbances. Mueller (1975), for example, has demon- strated that heritability estimates, derived from parent-offspring correla- tions, decrease as the prevalence of undernutrition increases. At the same time, different parameters show differential susceptibility. Body shape, or proportion, seems particularly resistant to nutritional variation (Tanner, 1978), and we would expect to find it altered only under the most stringent alterations of the diet. Another aspect of the rationale for anthropometry is that body measure- ment may indicate the presence of malnutrition, even though a particular dimension may not itself be affected by the malnutrition that is detected. In humans, malnutrition, either over or under, seldom occurs in splendid iso- lation, but rather exists as part of an ecological complex. Undernutrition is almost always found where there is poverty, infectious disease, and crowd- ing (Cravioto, 19701. Deficient fat and muscle stores, small body size, and delayed growth will characterize the inhabitants of such ecosystems, and, while we may not be able to specify how much of a given morphological deviation is directly produced by malnutrition, the greater the deviations from the normal, the greater the prevalence and the severity of the malnutri- tion. In other words, anthropometry may be, in many instances, not an effect, but an indicator. Thus, the rationale for anthropometry as one component of the assessment of nutritional status is rooted in these two postulates. First, alterations in the intake/need ratio will be recorded in morphology. Second, the inter- relationships of factors in a nutritional ecosystem will make anthropometry valuable as an indicator at the community level. The selection of dimensions and the analysis of data is, in nutritional anthropometry, done within the context of two different schemes that, while not mutually exclusive, are to some extent independent. These schemes are (1) the measurement of growth and (2) the measurement of body composi- tion. The Measurement of Growth Nutrient requirements, relative to body weight, are highest during the growing years, due to the additional nutrients needed to support body growth. Although reliable estimates of the requirements for growth are few, available evidence suggests that, during infancy, the energy necessary for growth may be as much as one-third that necessary for maintenance (Payne and Waterlow, 19771. There is no doubt that the cost of undernutrition is highest for infants and preschool children exceeding, so it would seem, the cost among adults of overnutrition. Consequently, the effects of malnutrition upon the growth

Anthropometry and Nutritional Status 257 process are marked and often dramatic, and the growth patterns of the children of a population provide perhaps the best evidence of the general level of health of that population (Eveleth and Tanner, 1976~. Van Duzen et al. (1976) have presented compelling evidence for this in their analysis of the increased growth among Navajo preschoolers in conjunction with the initiation of a feeding program. Not only did severe protein-energy malnu- trition drop dramatically in its incidence, so did the heights and weights of the children increase. Observations of individual children in response to dietary therapy or nutritional supplementation also provide evidence of the close relationship between growth and nutritional intake. Ashworth (1969) has described the sharp increases in the rates of growth of children from Jamaica who were recovering from PEM, while Lampl et al . (1978) documented a general, and statistically significant, increase in growth among chronically under- nourished Papua New Guinean children whose diets were supplemented with a skimmed milk powder, compared to their age peers whose diets were not supplemented. Because of the sensitivity of growth to nutritional deficiencies and exces- ses, the analysis of the growth records of children is widely used in evaluat- ing nutritional status. There is, by now, ample evidence of the growth- retarding effects of undernutrition, especially of protein and energy, man- ifested in small body size, reduced fat stores, and deficient stores of muscle, as well as delays in the rate of biological maturation. Overnutrition is as- sociated with increased deposition of adipose fat, accelerated rates of mat- uration, and possibly increased stature and lean body mass. Some researchers have suggested that decreased growth in stature and muscle mass is indicative of a deficiency of protein, while decreased fat stores reflect energy insufficiency. However, this is not well supported by the available evidence and, given the fact that energy seems to be the nutrient that, on a worldwide basis, is limiting (Behar, 1977), it seems premature to utilize nutritional anthropometry in such a specific way at present. Measurements of growth are especially well-suited for detecting children who have been subject to chronic mild-to-moderate malnutrition. An- thropometry can reveal this deficit, especially among children older than 12 months of age, when the cumulative effects of nutritional deficits become significantly registered in their morphology. The Measurement of Body Composition Three decades of intensive research have by now revealed that it is possible to separate analytically the mass of the human body into its component parts

258 FRANCIS E. JOHNSTON (see, e.g., Keys and Grande, 1973). The conceptual model used for analysis will vary from researcher to researcher, but, at a bare minimum, we may conceive of body mass as consisting of a fat component and a component called the lean body mass. Quantitatively body weight may be thought of as the sum of body fat plus the lean body mass (LBM), So that, by knowing weight, we need only determine one of the components in order to be able to solve for the other. Various techniques are available for estimating body composition in the laboratory. The most widely used of these techniques involve the determi- nation of one of the three variables: (1) the body density, (2) the volume of intracellular water, or (3) the mass of potassium, determined from counting the radioactive isotope, K40. From any one of these three variables, it is possible to compute a value for LBM or for fat as a percentage of weight. Since the mid-1920's, physical anthropologists have sought to measure more than linear dimensions and have developed instruments and techniques for estimating underlying masses of fat, muscle, and bone. These an- thropometrically determined values have been correlated with values esti- mated from laboratory procedures in attempts to validate anthropometry and to provide equations allowing the quantification of body composition (Brozek and Henschel, 1963; Malina, 1969~. Variations in the amounts and proportions of body fat and lean body mass are indicators of nutritional status. The fat depot is especially sensitive to deviations from optimal energy intake, while chronic undernutrition will be recorded in muscle wasting. NUTRITIONAL ANTHROPOMETRY A MINIMUM LIST To suggest a minimum list of measurements to be used in assessing nutri- tional status is dangerous and possibly foolhardy. Different investigators will favor different measurements: Some will insist upon a lengthy list, while others will utilize but one or two. Despite this, it seems that, over the years, a certain consensus has emerged, at least insofar as a minimum set of dimensions is concerned. In this section, I wish to present my interpretation of that list. But I wish also to present a second list, consisting of a few additional measurements that will be of use in certain situations. The first list is, in my opinion, the bare minimum without which an adequate assess- ment of nutritional status based upon anthropometry cannot be done. The second list includes measurements that will be of value if other kinds of analysis are to be done, especially relating possible causative factors in the environment to body morphology. The minimum list consists of the following dimensions:

Anthropometry and Nutritional Status 1. body length (from 2 years on, standing height is to be measured); 2. head circumference; 3. body weight; 4. mid upper arm circumference; 5. triceps skinfold; and 6. subscapular skinfold. The additions, referred to above as the second list, are as follows: 1. biacromial diameter (bony shoulder girdle); 2. bicristal diameter (bony pelvic girdle); 3 . sitting height; and 4. suprailiac skinfold. 259 The procedures outlined in the Staff Training Manual of Cycle III of the U.S. Health Examination Survey (HES), represent the most comprehensive collection of health data, including anthropometry, perhaps ever collected in the world, and the various mensurational procedures were developed care- fully and thoughtfully. However, other such sets of procedures are availa- ble, and, for those few measurements of the above lists not measured di- rectly by the HES (stature and supine length), one should consult the hand- book of the International Biological Program (IBP) (Weiner and Lourie, 1969~. It goes without saying that those taking the measurements must be well- trained, and even simple measurements such as height and weight cannot be approached naively. A big part of the problem of the validity of height and weight as measures of nutritional status has been the often-cavalier way they have been approached by clinicians, nurses, and public health workers. Anthropometry can give useful results, but only if the technicians involved are well-trained, utilize standard techniques, and undergo periodic training sessions to eliminate idiosyncratic factors, to reduce both systematic and random error, and to quantify the error of measurement. In addition to the above measurements, three derivations from the raw data are recommended. First is an evaluation of weight-for-height. Or, alternatively, one may calculate the index: weight-(height)2, known as Quetelet's index. Second is the calculation of the relative sitting height, sitting height/stature, a measure of the ratio of trunk (plus head) to leg length. Finally, the muscle mass of the body may be estimated by deriving the estimated circumference of the muscle of the upper arm, upper arm muscle circumference. Since both the upper arm circumference and the triceps skinfold are to be taken at the same level, the middle of the upper arm, and since the triceps skinfold is an estimate of the double layer of fat at

260 FRANCIS E. JOHNSTON that site (plus skin), if we asume that the arm and the underlying muscle are circular, the UAMC may be estimated as follows: UAMC = (Upper Arm Circumference) - (Triceps). While this may seem to involve a number of assumptions, as it does, it will be seen below that this measurement does have enough validity to warrant its use. Reference Standards One of the problems of nutritional anthropometry is that it requires standards against which to evaluate the data collected. After several years of discus- sion and debate, the majority of investigators now agree that it is impossible to specify true "norms" as standards. Instead it is now customary to use "reference standards" or "reference data" that provide a useful yardstick against which measurements may be compared. At least two schools of thought exist regarding the selection of reference standards. First, there are those persons who advocate the use of local standards, derived from ethnically similar groups, to minimize genetic fac- tors. Second, there are those who advocate the use of a single standard, drawn from some well-defined and accurately sampled population. Propo- nents of the single standard feel that the problem of genetic differences is minimized because of the advantages associated with one "universal" yardstick. My own sympathies are with the latter position, since the de- velopment of a large number of local standards, whatever local means, could very well lead to chaos. It is generally agreed that the single best set of reference standards are those of the HES. For children, these standards have been highly refined, using sophisticated statistical procedures, and are currently published and distributed by Ross Laboratories. Since they are based upon national proba- bility samples, and since they are as up-to-date as is possible, they are suitable not only for the United States, but also for reference use elsewhere. Unfortunately, only height and weight have been subjected to statistical smoothing, but the descriptive statistics are generally available for other measurements in the publications of the National Center for Health Statistics (NCHS). Data on adults are not as satisfactory. Cycle I of the HES examined 18- to 74-year-olds, and the measurements were analyzed and published in the NCHS series. However, the Durvey was conducted in the early 1960's, and any secular trend could create problems. The successor to the HES, HANES, has examined all age-groups, and the results are beginning to appear. As

Anthropometry and Nutritional Status 261 these results are published, we will have adequate data on the entire age range to be examined. THE VALIDITY OF NUTRITIONAL ANTHROPOMETRY The utility of anthropometry in assessing nutritional status must ultimately rest upon its ability to predict individuals who are nutritionally at-risk, or to determine the existence of some degree of malnutrition in a specific group. This may be done in a number of ways, and, in fact, there is now a large literature on anthropometry as an indicator of nutritional status. A review of that literature is well beyond the scope of this paper, but it is worthwhile to point to some examples of research that have attempted to validate nutritional anthropometry. We may begin with the consideration of the relationship between physical growth and malnutrition, and focus, for convenience, upon my own re- search. For about 5 years we have been analyzing the data collected by Joaquin Cravioto as part of his study of "The Land of the White Dust," in southern Mexico. We have collected additional data ourselves, and several publications have either appeared or are in press, resulting from our studies. In one such study, Scholl (1975) identified, from the cohort under study of over 280, a subgroup of 72 children who manifested failing growth between 6 months and 3 years of age. These children were hypothesized to be suffering from chronic undernutrition. Support for this came from our find- ing that, of the 19 children of the cohort subsequently hospitalized for severe protein-energy malnutrition, 14 came from this failing subgroup. Further- more, of the seven children who died, between 12 and 60 months, of nutrition-related causes, 6 came from this failing subgroup. In other words, failing physical growth was effective at identifying children who were nu- tritionally at-risk. Many other investigators have also established the link between reduced growth and malnutrition. In some instances, this link is used to identify individual children who may be singled out for therapy, while, in other cases, the measurement of growth may be used as an indicator of the exis- tence of malnutrition in a community or a socioeconomic group. Another aspect of the validity of anthropometry is its correlation with estimates of body composition determined physiologically. Since an- trhopometry is localized and noninvasive, estimates of body composition are necessarily indirect. Despite this, there does seem to be agreement that body measurements, given the proper analysis, can estimate underlying tissue masses. Certainly the measurement of the triceps skinfold is often used as the best indicator of obesity, and certainly is much better than an index of weight-for-height.

262 FRANCIS E. JOHNSTON TABLE 1 Partial Correlations (Age-Constant) of Anthropometry and Body Composition (Densitometry) in 12-17-Year-Olds Males Females Sample size 1 19 24 Arm muscle circ./~sM 0.778 0.629 Triceps skinfold/fat 0.838 0.741 Subscapular skinfold/fat 0.832 0.766 Table 1 presents some yet-unpublished data from a recent study of ours, in which we correlated, among 12- to 17-year-olds, anthropometric dimen- sions with LBM and body fat, estimated by densitometry. The correlations indicate that, among this sample, anthropometry can provide good estimates of these two variables, and, using more than simple bivariate correlations, considerable higher proportions of the variance in fat or LBM could be accounted for by body measurement. Again, research is still needed to establish the best use of the measurements, but the results seem quite prom- . . slng. To me there is little doubt that anthropometry provides useful measures of nutritional status, and, in certain situations, may be the method of choice. As I noted earlier, anthropometry is not the nutritional utopia that we all desire, but, used judiciously and wisely, it becomes an integral component. It is at its best when used in surveys and when used epidemiologically. When applied to individuals whose nutritional imbalance has not been chronic, it is not as accurate. For example, recently Young and Hill (1978) assessed changes in anthropometric dimensions of 54 adult patients under- going surgery. They found anthropometry to be insensitive as an indicator of protein depletion in individuals. However, a reading of their paper reveals flaws in their methodology, which may or may not have detracted from their findings. In any event, it seems that anthropometry is not as useful in such conditions as are other indicators. Anthropometry may be used to indicate parents who are at-risk of having children who will develop malnutrition or who may be especially suscepti- ble to nutritional deficiency. Our research in Mexico indicates that the body sizes of the parents are the best indicators of children who display chronic malnutrition. In another study, Lechtig et al. (1976), of the INCAP, have found maternal stature, head circumference, and housing quality to be sig- nificant biosocial indicators of women more likely to produce a low birth- -weight infant. The link to nutrition seems to be established, in this group, through the reduction in the incidence of low birth weights of women from the lowest socioeconomic groups (Lechtig et al., 1975~.

Anthropometry and Nutritional Status SUMMARY 263 Anthropometry is a much-abused yet time-honored technique. However, nutritional anthropometry is a dynamic methodology, in which measure- ments are being developed and applied to specific situations. The results are promising and suggest that body morphology is a sensitive indicator of malnutrition. Anthropometry cannot be employed haphazardly, uncritically, or without an understanding of the underlying theory relating body size and shape to body composition and growth and their alteration with malnutr~- tion. REFERENCES Ashworth, A. 1969. Growth rates in children recovering from protein-calorie malnutrition. Br. J. Nutr. 23:835-845. Behar, M. 1977. Protein-calorie deficits in developing countries. Ann. N.Y. Acad. Sci. 300:176-187. Brozek, J., ed. 1 956. Body measurements and human nutrition. Wayne State University Press, Detroit. Brozek, J., and A. Henschel, ed. 1963. Techniques for measuring body composition. National Academy of Sciences-National Research Council, Washington, D.C. Christakis, G., ed. 1973. Nutritional assessment in health programs. American Public Health Association, Washington, D.C. Cravioto, J. C. 1970. Complexity of factors involved in protein-calorie malnutrition. Bibl. Nutr. Dieta 14:7-22. Eveleth, P. B., and J. M. Tanner. 1976. Worldwide variation in human growth. Cambridge University Press, New York. Guthrie, H. A. 1975. Introductory nutrition. C. V. Mosby, St. Louis. Jelliffe, D. B. 1966. The assessment of the nutritional status of the community. WHO, Geneva. Keys, A. F., and F. Grande. 1973. Body weight, body composition, and calorie status. In R. S. Goo&art and M. E. Shils, eds. Modern nutrition in health and disease. Lea and Febiger, Philadelphia. Lampl, M., F. E. Johnston, and L. A. Malcolm. 1978. The effects of protein supplementation on the growth and skeletal maturation of New Guinean schoolchildren. Ann. Human Biol. 5:219-228. Lechtig, A., H. Delgado, R. E. Lasky, R. E. Klein, P. L. Engle, C. Yarbrough, and J-P. Habicht. 1975. Maternal nutrition and fetal growth in developing societies: Socioeconomic factors. Am. J. Dis. Child. 129:434-437. Lechtig, A., H. Delgado, C. Yarbrough, J-P. Habicht, R. Martorell, and R. E. Klein. 1976. A simple assessment of the risk of low birth weight to select women for nutritional interven- tion. Am. J. Obstet. Gynecol. 125:25-34. McKigney, J. I., and H. N. Munroe, ed. 1976. Nutrient requirements in adolescence. MIT Press, Cambridge, Mass. Malina, R. M. 1969. Quantification of fat, muscle, and bone in man. Clin. Orthoped. 65:9-38. Malina, R. M. 1979. Secular changes in size and maturity: Causes and effects. In A. F. Roche, ed. Monogr. Soc. Ref. Child Dev., in press.

264 FRANCIS E. JOHNSTON Mueller, W. H. 1975. Parent-child and sibling correlations and heritability of body measure- ment in a rural Colombian population. Ph.D. Dissertation, University of Texas, Austin. Payne, P. R., and J. C. Waterlow, 1977. Relative energy requirements for maintenance, growth, and physical activity. Lancet 2:210-211. Scholl, T. O. 1975. Body size in developing nations: Is bigger better? PhD. Dissertation. Temple University, Philadelphia. Tanner, J. M. 1978. Foetus into man: Physical growth from conception to maturity. Harvard University Press, Cambridge, Mass. Van Duzen, J., J. P. Carter, and R. V. Zwagg. 1976. Oritein and calorie malnutrition among preschool Navajo children: A follow-up. Am. J. Clin. Nutr. 29:657-662. Weiner, J. S., and J. A. Lurie. 1969. Human biology: A guide to field methods. IBP Handbook No. 9. F. A. Davis, Philadelphia. World Health Organization. 1968. Nutritional status of populations: A manual on an- thropometric appraisal of trends. wHo/Nutr/70.129. WHO, Geneva. Young, G. A., and G. L. Hill. 1978. Assessment of protein-calorie malnutrition in surgical patients from plasma proteins and anthropometric measurements. Am. J. Clin. Nutr. 31 :429-435.

Relation of Nubidonal Anemias to Food Consumption Patterns VICTOR HERBERT S UMM ARY The American diet is marginal in absorbable iron content for infants and children under age 5, for children at the growth spurt of puberty, and for women in the child-bearing years. Folate deficiency is common among people not eating each day one fresh or fresh-frozen uncooked fruit or vegetable or fruit juice. Vitamin BY deficiency is common on a dietary basis only among pure vegetarians. Iron-deficiency anemia and folate (folic acid) deficiency anemia are more common in women than men both because of the monthly loss of these two nutrients in blood during the menstrual years and the taking up of these nutrients by the fetus at the expense of the mother during pregnancy. During the menstrual years, because of the monthly blood loss, women have ap- proximately twice the daily iron need of men. For them, the American diet is marginal in iron, and approximately 40 to 50 percent of premenopausal women may have iron depletion and about 15 percent have iron-deficiency anemia. After age 50, only about 13 percent of women have iron depletion. Iron depletion is present in about 50 percent of infants and iron-deficiency anemia in about 25 percent. Iron depletion is present in about 10 percent of children at the growth spurt of puberty. Both the Committee on Maternal Nutrition (1970) and the Committee on Dietary Allowances (1974) of the Food and Nutrition Board, National Re- search Council recommend that folic acid and iron supplements should be taken throughout pregnancy. The Committee on Dietary Allowances rec- ommendation was that the recommended dietary allowance be doubled in 265

266 VICTOR HERBERT pregnancy, to 800 ,ug daily. Data presented at the 1975 Folate Workshop of the National Academy of Sciences suggest that such an amount could not easily be achieved without supplementation. Oral contraceptives reduce monthly blood loss, thereby reducing the frequency of iron deficiency. The question of whether the existing iron fortification of American flour should be increased (AMA Council on Foods and Nutrition, 1972: Wintrobe, 1973) has been decided in the negative on grounds of inadequate information to make an adequate risk:benefit assessment. In this connection, it should be noted that American breads are now mainly fortified with ferrous sulfate, with a trend to ferrum reductum, and American spaghettis and pastas with the less absorbable ferrous pyrophosphate (Dudley Titus, personal com- munication). Canada requires that more than 90 percent of ferrum reductum fortification be the more absorbable less than 10 micron particle size, but the United States has no size requirements. Several surveys have shown lower serum and red cell folate in women taking oral contraceptives, but a daily fresh uncooked vegetable, fruit, or fruit juice would probably prevent folate deficiency in this group as well as many other groups. INTRODUCTION Anemia is defined as a reduction below normal in the amount of red blood that occurs when the equilibrium between blood production and loss (through bleeding or destruction) is disturbed (Dorland's Illustrated Medi- cal Dictionary, 1974~. By World Health Organization criteria (1968; Baker and DeMaeyer, 1979), anemia is considered to exist when the nonpregnant adult female has a hemoglobin below 12 and the pregnant adult female a hemoglobin below 11 g/100 ml of venous blood (when at sea level; normal values are higher at higher altitudes). The observations of Scott et al. ( 1970) indicate that the hemoglobin concentration of healthy nonpregnant young women without iron deficiency will almost always be 12 g/100 ml or more, that at mid-pregnancy this value will practically always be at least 10 g/100 ml, but fairly often may be less than 11 g/100 ml, and late in pregnancy this figure will almost always be 10 g/100 ml or more, and most often 11 g/100 ml or more, in the absence of iron deficiency. The normal fall of hemoglo- bin during pregnancy is simply pregnancy hypervolemia (which increases both the plasma and red cell volume, with a greater increase in the former). Nutritional anemia is defined as a condition in which the hemoglobin content of the blood is lower than normal as a result of deficiency of one or more essential nutrients. To delineate a given anemia as nutritional, two criteria must be met: lack of the nutrient must produce, and providing the

Relation of Nutritional Anemias to Food Consumption Patterns 267 nutrient must correct, the anemia. By these two criteria, there are only three unequivocal nutritional anemias: those due to lack of iron, folate, or vitamin BE (Herbert, 19701. These three anemias reflect an important nutritional problem affecting large population groups, particularly the poverty-stricken and those under metabolic stress. Iron deficiency and folate deficiency are more common in women be- cause of two forms of metabolic stress peculiar to women: the monthly blood loss in premenopausal women and the drain on maternal nutrient stores imposed by pregnancy. The fetus will take from the mother whatever it needs in order to be born normal, even if this produces severe nutrient deficiency in the mother (Committee on Maternal Nutrition, 19701. Since anemia is a relatively late manifestation of nutritional deficiency, those patients diagnosed as having nutritional anemia are the "tip of the iceberg" part of a larger group suffering from nutrient depletion of more moderate degree, which is not yet manifest by unequivocal anemia. The metabolic stress of menstrual blood loss is increased by the use of some intrauterine contraceptive devices (Anonymous, 1974, 1975) and de- creased by the use of oral contraceptives (Anonymous, 1973~. In fact, ad- vertising of "unique vitamin-mineral formulas for the special needs of patients taking oral contraceptives" are misleading because it has not been established that there is any such special need (Anonymous, 1973; Sym- posium, 1975~. IRON- DEFI CIENCY ANEMIA Fairbanks et al. (1971) tabulated the approximate frequency of iron deple- tion as 50 percent of infants, 50 percent of premenopausal women, and 100 percent of pregnant women. They tabulated the frequency of iron-deficiency anemia (the end product of prolonged iron depletion) as 25 percent of in- fants, 0-5 percent of children (higher frequency was in economically de- prived children), 15 percent of premenopausal women, and 30 percent of pregnant women not receiving iron supplementation. About one-third to one-half of apparently healthy young American women have laboratory evidence of iron depletion (Monsen et al., 1967; Scott and Pritchard, 1967; Sturgeon and Shoden, 19711. Sturgeon and Sho- den (1971) found less than 5 mg of iron/100 g liver tissue in 40 percent of women age 20 to 50, indicative of iron depletion. This was true of only 13 percent of women over age 50 (and less than 10 percent of all men). It should be noted that iron depletion (loss of body iron stores) precedes anemia. While a majority of women who are anemic have iron deficiency, this is not always the case, so self-administration of medicinal iron may be

268 VICTOR HERBERT unwise, and blanket treatment of every anemic woman with iron, without ascertaining that she, in fact, has iron deficiency, can do positive harm (as, for example, in women with sickle-cell or other hemolytic anemias with increased iron stores, in whom the giving of iron may produce "iron over- load" syndrome). Nevertheless, the incidence of anemia in venous groups of pregnant American women has ranged from 10 to 60 percent, most of which could be prevented by prophylactic iron therapy (AMA Council on Foods and Nutrition, 19681. Menstrual loss of iron is the main source of the iron losses in nonpregnant women in the fertile age-group (Rybo, 1970; Fairbanks et al., 19711. The average menstrual blood loss is about 40 ml/cycle (Fairbanks et al., 1971), representing a loss of about 20 mg of iron per cycle. About 10 percent of women have menorrhagia, with a blood loss exceeding 80 ml/cycle (Hallberg et al., 1966), making them particularly susceptible to iron defi- ciency. The use of more than 12 pads during a menstrual period, of the damming up of blood behind tampons, often suggests excessive menstrual bleeding (More and Dubach, 1956: Fairbanks et al., 1971~. An understanding of the situation of American women with respect to iron balance is more clearly made by reference to Tables 1 and 2, which present respectively the estimated dietary iron requirements of Americans (Table 1) and the iron requirements of pregnant Amencan women (Table 2~. The adsorbability of iron from different food sources is highly variable, averag- ing out to about 10 percent of iron in the total diet being absorbed (Layrisse, 1975; Cook, 19781. Therefore, the amount of iron ingested must be 10-fold TABLE 1 Estimated Dietary Iron Requirements Absorbed Iron Requirement, mg/day Normal men and nonmenstruating women 0.5-1.0 5-10 Menstruating women 0.7 -2.0 7 -20 Pregnant women 2.0-4.8 20-48b Adolescents 1 . 0-2.0 10-20 Children 0.4-1.0 4-10 Infants 0.5-1.5 1.5 mg/kgC Dietary Iron Requirement, a mg/day a Assuming 10 percent absorption. b This amount of iron cannot be derived from diet and should be met by iron supplementation in the latter half of pregnancy. c To a maximum of 15 ma. SOURCE: After Council on Foods and Nutrition, 1968. Courtesy of the Journal of the American Medical Association.

Relation of Nutritional Anemias to Food Consumption Patterns 269 TABLE 2 Iron Requirements for Pregnancy Average, mg Range, mg External iron loss 170 15(}200 Expansion of red-blood-cell mass 450 20~600 Fetal iron 270 20~370 Iron in placenta and cord 90 3~170 Blood loss at delivery 150 9~310 Total requirements 980 58~1,340 Cost of pregnancyb 680 44~1,050 a Blood loss at delivery not included. b Expansion of red-blood-cell mass not included. SOURCE: After Council on Foods and Nutrition, 1968. Courtesy of the Journal of the American Medical Association. the daily requirement, as Table 1 indicates. Since the average American diet provides about 6 mg of iron per thousand kcal (Monsen et al., 1967), iron intake from dietary sources is borderline for teenage girls and women and may be inadequate for infants and pregnant women (AMA Council on Foods and Nutrition, 1968: Committee on Dietary Allowances, 19741. Nevertheless, a woman who has sufficient iron stores to provide for increase in hemoglobin mass during pregnancy and who breast feeds her infant for 6 months (thereby delaying the return of menstruation) will have her iron needs covered by an adequate intake of dietary iron (FAD/WHO Expert Group, 1970~. The diagnostic features of iron deficiency are summarized in Table 1. To this should be added the fact that 5 mg or less of iron/100 g of liver tissue is indicative of depletion of the storage pool to the extent that iron-deficiency anemia may be either present or anticipated with any further depletion (Sturgeon and Shoden, 19711. As stated in footnote b to Table 1, the amount of iron required to meet the needs of pregnancy should usually be met by iron supplementation in the latter half of pregnancy, since it cannot usually be derived from diet. The Committee on Maternal Nutrition (1970) recommends supplementation with 30-60 mg of iron daily (i.e., 15~300 mg of ferrous sulfate) during preg- nancy. The physician should use his judgment in this regard, based on knowledge of the patient, the dietary habits, the fact that iron deficiency is frequent in pregnant women, and his evaluation of the blood and iron status of the particular patient (Herbert, 1975a). He may routinely give iron (Wal- lerstein, 1973~. In general, oral ferrous sulfate, the least expensive iron preparation, is the drug of choice for treating iron deficiency. A detailed discussion of iron

270 VICTOR HERBERT therapy is presented elsewhere (Herbert, 1972; Herbert, 1975a). It is im- portant to remember that the duration of oral therapy for iron deficiency should be approximately 6 months, since a lesser duration will not adequately replete body iron stores. The physician must remember that the iron deficiency may have developed in association with menorrhagia; if that menorrhagia persists, iron therapy may have to persist as well so that the iron loss in blood does not again produce a negative iron balance. About 3 - percent of the iron in vegetable foods and 15-20 percent of the iron in animal foods is absorbed. Heme iron accounts for about one-third the iron in animal tissues, but the heme in a moderate portion of meat, fish, or poultry may supply up to one-third of the daily requirement because it is 5 to 10 times as absorbable as inorganic iron in food (Cook, 1978~. Nonheme iron absorption is enhanced by vitamin C and a "meat tissue factor" (not found in milk, eggs, or cheese); it is inhibited by eggs, bran, tea, EDTA, and calcium phosphate (Cook, 1978~. Thus, the amount of iron absorbed from a particular meal is dependent largely on its proportion of heme and nonheme iron, and its content of ascorbic acid and animal food, and iron deficiency, is much more frequent in vegetarians. The high frequency of iron deficiency in infants correlates with diets consisting largely of milk, so the maximum incidence is between 6 months and 2 years of age, and it is corrected by eating a mixed diet, so by age 5 it has fallen considerably. The 0.2-0.3 ,ug Fe/ml breast milk is adequate to the requirements of the exclusively breast-fed infant until birth weight is tripled. When breast feeding is discontinued before age 6 months, the use of formula rather than fresh cow's milk reduces the risk of intestinal blood loss. Cow's milk is higher in iron and lower in lactoferrin than human milk, but the iron in breast milk is better absorbed (Lockhart, 19791. Bothwell and Charlton note (1973) that use of the traditional iron pot instead of aluminum cookware to prepare their predominantly cereal diet (and traditional drinks) made iron deficiency anemia very infrequent in Bantu women, but made iron overload disease frequent in Bantu men. The striking effect of cooking in an iron skillet on iron content of food is illus- trated by them in Table 3. FOLATE-DEFICIENCY ANEMIA Studies carried out under the aegis of the World Health Organization (1968) and by Baker and DeMaeyer (1979) in various countries suggest that up to a third of all the pregnant women in the world have folate deficiency. In a recent study in a New York City municipal clinic, our group (Herbert et al., 1975) found tissue deficiency of folio acid, as measured by a red-cell folate

Relation of Nutritional Anemias to Food Consumption Patterns 271 TABLE 3 Effect of Cooking in Iron Skillet (Dutch Oven) on Fe Content of Foods Food Iron Content, mg/100 gm Cooling Time, min Glass Dish Dutch Oven Spaghetti sauce 180 3.0 87.5 Gravy 20 0.43 5.9 Potatoes, fried 30 0.45 3.8 Rice casserole 45 1.4 5.2 Beef hash 45 1.52 5.2 Apple butter 120 0.47 52.5 Scrambled eggs 3 1.7 4. 1 SOURCE: Bothwell and Charlton, 1973. level below 150 ng/ml, present in 16 percent of 110 sequential pregnant women at the time of their first prenatal visit to the clinic. A further 14 percent had red-cell folate levels in the range "suggestive but not conclusive for tissue folate depletion" (150-199 ng/ml). It was suggested that daily ingestion of one fresh or fresh-frozen uncooked fruit or vegetable or fruit juice could have prevented this folate deficiency. These studies add to a growing body of evidence that nutritional defi- ciency of folic acid is prevalent among Americans of poor economic status. Based on findings up until 1970, Pritchard, writing for the Committee on Maternal Nutrition of the Food and Nutrition Board (1970), recommended that folic acid supplements should be taken throughout pregnancy. Sub- sequently, the data of the Ten-State Nutrition Survey (1968-70) became available. Although that survey found that "the mean serum folate values were, with few exceptions, above the acceptable level of 6 ng/ml and the mean red cell folate values were in the acceptable range of 150-650 ng/ml," a more detailed evaluation indicates a real problem in fact exists, obscured by the use of mean values alone (Herbert et al., 19751. The mean values obscured the existence of substantial numbers of actual values sufficiently below the mean as to suggest widespread folio acid deficiency. This is indicated by the data from the Survey Director for Massachusetts of the Ten-State Nutrition Survey (Edozien, 19721. In Massachusetts, serum and red-cell folates were measured on most of the samples collected. Of all the 1,087 Massachusetts blood samples from females on which such estimations were made, 25.6 percent had red-cell folate values below 150 ng/ml. This includes 115 pregnant women, all of whom were receiving prenatal clinic care, in the economically poor

272 VICTOR HERBERT Roxbury-Dorchester area, and most of whom were receiving vitamin sup- plements, presumably containing folio acid. Among these pregnant women, 7.1 percent had red-cell folate values below 160 ng/ml and 7.1 percent had serum folate levels below 3 ng/ml. It is probably relevant that the over 10,000 individuals surveyed in Massachusetts were randomly selected from enumeration districts with the lowest average income (lowest quartile) ac- cording to the 1960 census (Ten-State Survey, 1968-70), and poverty and folate deficiency tend to run hand in hand (Herbert, 1968; Kahn et al., 1970~. As the Massachusetts report noted (Edozien, 1972), "The results suggest that the diets currently eaten by a large segment of the population cannot provide the allowance of folic acid recommended for optimal health and, therefore, that dietary deficiency of folic acid may pose a major nutri- tional problem. Considered together with the finding of a high prevalence of low plasma vitamin A levels, it would appear that these diets contain insuf- ficient amounts of green leafy vegetables which are major sources of both folic acid and provitamin A. Current processes for preservation, storage and preparation of foods may also destroy a high proportion of the folate in foods. " The hazard to mother and fetus of folate deficiency in the absence of frank anemia is unclear and has been extensively reviewed (Rothman, 1970~. Studies from South Africa suggest that folate supplements in this situation decrease the incidence of prematurity (Baumslag et al., 1970) and cause significant elevation of hemoglobin levels (Colman et al., 1975 a, b, c), suggesting that even mild deficiency may limit DNA synthesis. Prospective studies of the effects on the fetus are difficult to interpret because folate administration invariably starts only after the period of maximum fetal sus- ceptibility in the first trimester. However, animal experiments demonstrate a consistent teratogenic effect of folate deprivation from the time of concep- tion, dependent on the duration of the experiment (Herbert et al., 1975~. Thus, in the light of present knowledge it appears appropriate to correct folate deficiency in pregnancy. The implementation of this principle by improving the quality and quantity of available food is a long-term ideal limited by custom and economic circumstances. For this reason, the Joint FAD/WHO Expert Committee on Nutrition has recommended that food for- tification should be considered as an immediate possibility for the improve- ment of intake of any deficient nutrient (FAD/WHO Expert Committee, 19711. A series of studies indicate that fortification of staple foods with folic aicd is feasible, safe, effective, and in accordance with the recommenda- tions of the Expert Committee (Colman et al., 1974a,b,c, 1975a,b,c). With adequate fortification, possible hazards of folate deficiency in early preg- nancy would be averted. Until such fortification is practiced, administration of folic acid tablets, 200-400,ug/day, is appropriate for all pregnant women,

Relation of Nutritional Anemias to Food Consumption Patterns 273 with 300 ,ug probably adequate for any pregnant population group (Herbert, 1977). The evidence relating folate deficiency in pregnancy to presence in the offspring of mental retardation and other defects in central nervous system function and development is reviewed elsewhere (Herbert and Tisman, 1973~. This evidence is, as yet, far from conclusive but does constitute one more slight increment in the balance favoring the concept of daily folate supplementation throughout pregnancy. In 1975, there was a National Academy of Sciences "Workshop on Human Folate Requirements," whose proceedings were published in 1977. That workshop included papers presenting the latest information on dis- tribution of folates in food, food folate availability, results of several sur- veys to detect folate deficiency in certain American population groups, and reviews of the folic acid requirement in children, adults, and in situations of increased need. To briefly summarize the findings most pertinent to nutri- tional anemias: Measurement of serum and red-cell folate together constitute the best method for delineating the existence of folate deficiency; food folate availability is affected by various constituents present in different foodstuffs; pregnancy increases folate requirement. The minimal daily adult requirement for folic acid, which must be ab- sorbed from food to sustain normality, is in the range of 50 ,ug daily (Her- bert, 1968), and the Food and Nutrition Board (1974) recommends that the diets of adults contain 400 ,ug daily. This requirement appears to be ap- proximately doubled by pregnancy. Thus, if a woman is absorbing from her food in the range of 100 ,ug of folio acid daily from the start of pregnancy, she may not need supplementation, but assuming lesser stores than normal at the start of pregnancy, 20~300 ,ug of folate supplementation daily may be necessary (Herbert, 1975; Herbert, 1977~. The Food and Nutrition Board (1974) recommends a daily dietary intake of 800 ,ug during pregnancy and 600 ,ug during lactation. Although serum and red-cell folate may be lowered by the use of oral contraceptives (Smith et al., 1975; Prasad et al., 1975), it is not yet clear that folate supplementation is needed by women taking such products (Anonymous, 1973; Lindenbaum, 1975~. This folate need would probably be adequately met by one fresh uncooked vegetable, fruit, or fruit juice daily (Herbert, 1975b). VITAMIN Bi2 Although serum vitamin BE level falls in pregnancy (Cooper, 1973) and may also fall with the use of oral contraceptives (Wertalik et al., 1972; Smith et al., 1975), tissue levels of vitamin BE may remain normal and vitamin BE

274 VICTOR HERBERT deficiency anemia has not seemed to be a problem (Wertalik et al., 1972~. From evidence so far, vitamin BE deficiency anemia is rarely a dietary problem in the United States (Herbert, 1975b). In population groups where vitamin B,2 deficiency is common due to vegetarianism, such vitamin BE deficiency would be increased by the metabolic stress in pregnancy, including the fetal drain on maternal stores of about 0.3 ,ug B~2/day (Herbert, 1968; WHO, 1970) and by a mean of 0.3 ,ug B~2/day lost in breast milk during lactation (WHO, 1970~. It is for these reasons that the recommended dietary allowance (Food and Nutrition Board, 1974) for vitamin B ~2 was raised from the 3 ,ug for adults in general to 4 ,ug for pregnant or lactating females. Certain microorganisms and all animal protein are the sole source of dietary vitamin Bit. Thus, any diet devoid of animal protein, or not con- taining microorganisms that synthesize the vitamin, will eventually produce vitamin BE deficiency, unless the diet is supplemented with the vitamin. Certain seaweeds and legume nodules contain microorganisms that syn- thesize the vitamin; eating these foods and unwashed food contaminated with fecal microorganisms that synthesize the vitamin delays the onset of vitamin BE deficiency. REFERENCES AMA Council on Foods and Nutrition. 1968. Iron deficiency in the United States. JAMA 203:407-412. AMA Council on Foods and Nutrition. 1972. Iron in enriched wheat flour, farina, bread, buns and rolls. JAMA 220:855 - 859. Anonymous. 1973. Feminins and other vitamin-mineral supplements for women taking oral contraceptives. Med. Lett. 15:81-82. Anonymous. 1974. Topical and systematic contraceptive agents. Med. Lett. 16:37-40. Anonymous. 1975. Cu-7, a copper containing IUD. Med. Lett. 17:26-27. Baker, S. J., and E. M. DeMaeyer. 1979. Nutritional anemia: Its understanding and control with special reference to the work of the World Health Organization. Am. J. Clin. Nutr. 32:368-417. Baumslag, N., T. Edelstein, and J. Metz. 1970. Reduction of incidence of prematurity by folic acid supplementation in pregnancy. Br. Med. J. 1:16-17. Bothwell, T. H., and R. W. Charlton. 1973. Iron deficiency geographically speaking. In W. H. Crosby, ed. Iron. Medcom, New York. Colman, N., M. Barker, R. Green, and J. Metz. 1974a. Prevention of folate deficiency in pregnancy by food fortification. Am. J. Clin. Nutr. 27:339. Colman, N., J. V. Larsen, M. Barker, E. A. Barker, R. Green, and J. Metz. 1974b. Preven- tion of folate deficiency by food fortification. V. A pilot field trial of folic acid-fortified maize meal. S. Afr. Med. J. 48:1763. Colman, N., R. Green, K. Stevens, and J. Metz. 1974c. Prevention of folate deficiency by food fortification. VI. The antimegaloblastic effect of folic acid-fortified maize meal. S. Afr. Med. J. 48:1795. Colman, N., R. Green, and J. Metz. 1975a. Prevention of folate deficiency by food fortifica- tion. II. Absorption of folic acid from fortified staple foods. Arn. J. Clin. Nutr. 28:459-464.

Relation of Nutritional Anemias to Food Consumption Patterns 275 Colman, N., J. V. Larsen, M. Barker, E. A. Barker, R. Green, and J. Metz, 1975b. Preven- tion of folate deficiency by food fortification. III. Effect in pregnant subjects of varying amounts of added folic acid. Am. J. Clin. Nutr. 28:465~70. Colman, N., E. A. Barker,~M. Barker, R. Green, and J. Metz. 1975c. Prevention of folate deficiency by food fortification. IV. Identification of target groups in addition to pregnant women in an adult rural population. Am. J. Clin. Nutr. 28:471-476. Cook, J. D. 1978. Food iron availability. Food Nutr. News 49(3):1,4. (National Live Stock and Meat Board, Chicago) Cooper, B. 1973. Folate and vitamin BE in pregnancy. Clin. Haemat. 2:461~76. Committee on Dietary Allowances. 1974. Recommended dietary allowances. Food and Nutri- tion Board, National Research Council, National Academy of Sciences, Washington, D.C. Committee on Maternal Nutrition. 1970. Maternal nutrition and the course of pregnancy. Food and Nutrition Board, National Research Council, National Academy of Sciences, Washington, D.C. Dorland's illustrated medical dictionary, 25th ed. 1974. W. B. Saunders, Philadelphia. Edozien, J. C. 1972. National Nutrition Survey, Massachusetts, July 1969-June 1971. Report of the Survey Director to the Commissioner for Public Health, Commonwealth of Mas- sachusetts, Boston. (Kindly supplied to us by Derek Robinson, M.D., Director, Division of Community Operations, Department of Public Health, Commonwealth of Massachusetts.) Fairbanks, V. F., J. L. Fahey, and E. Beutler, eds. l 971. Clinical disorders of iron metabolism, 2d ed. Grune and Stratton, New York and London. FAD/WHO Expert Committee on Nutrition. 1971. Food fortification. WHO Tech. Rep. Ser. No. 477. FAD/WHO Expert Group. 1970. Requirements of ascorbic acid, vitamin D, vitamin Be, folate and iron. WHO Tech. Rep. Ser. No. 452. (Purchasable for $1.25 from WHO in Geneva or American Public Health Association, 1740 Broadway, New York, NY 10019. Available in English, French, Spanish, Russian, and Chinese.) Food and Nutrition Board, National Research Council. 1977. Folic acid: Biochemistry and physiology in relation to the human nutrition requirement. National Academy of Sciences, Washington, D.C. Hallberg, L., A.-M. Hogdahl, L. Nilsson, and G. Rybo. 1966. Menstrual blood loss, a population study: Variation at different ages and attempts to define normality. Herbert, V. 1968. Nutritional requirements of vitamin BE and folic acid. Am. J. Clin. Nutr. 21:743 -752. Herbert, V. 1970. Introduction to the nutritional anemias. Semin. Haemat. 7:2-5. Herbert, V. 1972. Oral iron therapy. In W. H. Crosby, ed. Iron. Medcom, New York. Herbert, V. 1975a. Drugs effective in iron-deficiency and other hypochromic anemias. Pages 1309-1323 in L. S. Goodman and A. Gilman, eds. The pharmacological basis of therapeu- tics, 5th ed. Macmillan, New York. Herbert, V. 1975b. Drugs effective in megaloblastic anemias. Vitamin BE and folic acid. Pages 132~1349 in L. S. Goodman and A. Gilman, eds. The pharmacological basis of therapeutics, 5th ed. Macmillan, New York. Herbert, V. 1977. Folic acid requirements in adults (including pregnant and lactating females). In Folic acid. Food and Nutrition Board, National Academy of Sciences, Washington, D. C. Herbert, V., and G. Tisman.1973. Effects of deficiencies of folic acid and vitamin B ~2 on central nervous system function and development. Pages 373-392 in G. Gaull, ed. Biology of brain dysfunction, vol. 1. Plenum Press, New York and London. Herbert, V., N. Colman, M. Spivack, E. Ocasio, V. Ghanta, K. Kimmel, L. Brenner, J. Freundlich, and J. Scott. 1975. Folic acid deficiency in the United States: Folate assays in a prenatal clinic. Am. J. Obstet. Gynecol. 123:175-179.

276 VICTOR HERBERT Kahn, S. B., S. Fein, S. Rigberg, and I. Brodsky. 1970. Correlation of folate metabolism and socioeconomic status in pregnancy and in patients taking oral contraceptives. Am. J. Obstet. Gynecol. 108:931-935. Layrisse, M. 1975. Iron nutriture. Pages 148- 154 in P. L. White and N. Selvey, eds. Pro- ceedings, Western Hemisphere Nutrition Congress IV. Publishing Sciences Group, Acton, Mass. Lindenbaum, J., N. Whitehead, and F. Reyner. 1975. Oral contraceptive hormones, folate metabolism, and the cervical epithelium. Am. J. Clin. Nutr. 28:346-353. Lockhart, J. D. 1979. Breast milk. Contemp. Nutr. 4(1):1-2 (General Mills, Minneapolis). Monsen, E. R., I. N. Kuhn, and C. A. Finch. 1967. Iron status of menstruating women. Am. J. Clin. Nutr. 20:842-849. Moore, C. V., and R. Dubach. 1956. Metabolism and requirements of iron in the human. JAMA 162:197-204. Prasad, A. S., K. Y. Lei, D. Oberleas, K. S. Moghissi, and J. C. Stryker. 1975. Effect of oral contraceptive agents on nutrients II. Vitamins. Am. J. Clin. Nutr. 28:385-391. Rothman, D 1970. Folicacidin pregnancy. Am. J. Obstet. Gynecol. 108:149-175. Rybo, G. 1970. Menstrual loss of iron. Pages 163- 171 in L. Hallberg, H.-G. Harwerth, and A. Vannotti, eds. Iron deficiency. Academic Press, London and New York. Scott, D. E., and J. A. Pritchard, 1967. Iron deficiency in healthy young college women. JAMA 199:897-900. Scott, D. E., J. A. Pritchard, A. S. Saltin, and J. M. Humphreys. 1970. Iron deficiency during pregnancy. Pages 491-503 in L. Hallberg, H.-G. Harwerth, and A. Vannotti, eds. Iron deficiency, Academic Press, London and New York. Smith, J. L., G. A. Goldsmith, and J. D. Lawrence.1975. Effects of oral contraceptive steroids on vitamin and lipid levels in serum. Am. J. Clin. Nutr. 28:377-384. Sturgeon, P., and A . Shoden. 1971. Total liver storage iron in normal populations of the U. S . A. Am. J. Clin. Nutr. 24:469-474. Symposium. 1975. Effects of oral contraceptive hormones on nutrient metabolism. Am. J. Clin. Nutr. 28:333-402, 521-560. Ten-State Nutrition Survey. 1968-70. U.S. Department of Health, Education, and Welfare, Center for Disease Control, Atlanta, Ga. DHEW Publ. No. (HSM) 72-8130 through 8134. Wallerstein, R. O. 1973. Iron metabolism and iron def~ciency during pregnancy. Clin. Haemat. 2:45~460. Wertalik, L. F., E. N. Metz, A. F. LoBuglio, and S. P. Balcerzak. 1972. Decreased serum B~ levels with oral contraceptive use. JAMA 221:1371-1374. WHO Scientific Group. 1968. Nutritional anaemias. WHO Tech. Rep. Ser. No. 405. (Purchasable for $ 1.00 from Publication Distribution and Sales, World Health Organization, 1211 Geneva 27, Switzerland.) Wintrobe, M. M. 1973. The proposed increase in the iron fortification of wheat products. Nutr. Today 8:18-20.

Next: Appendix D: Participants and Observers »
Assessing Changing Food Consumption Patterns Get This Book
×
Buy Paperback | $80.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF
  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!