Speech and Language Disorders in Children: Implications for the Social Security Administration's Supplemental Security Income Program (2016)

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 D Description of Methods Used to Calculate Trends in National Survey Data METHODS This report includes analyses of three datasets: (1) the 2008 panel of the Survey of Income and Program Participation (SIPP), (2) the 2007 and 2011-2012 iterations of the National Survey of Children’s Health (NSCH), and (3) the 2005-2006 and 2009-2010 iterations of the National Survey of Children with Special Health Care Needs (NS-CSHCN). Survey of Income and Program Participation The SIPP is a nationally representative survey of the noninstitutional- ized population of the United States. The 2008 panel was administered to household units over 16 waves occurring every 4 months. In addition to core questions asked at each wave, additional sets of questions referred to as topic modules were asked at 12 of the waves. The Wave 6 topic module was administered from May to August 2010, and it contained questions about the functional limitations and disabilities of adults (defined by the SIPP as those aged 15 years or older) and children aged 6-14 in the surveyed families. Wave 6 of the 2008 SIPP included 88,164 people in 34,850 house- holds. Of these, there were 14,784 children aged 6-17 in 8,836 families. Children aged 6-17 were identified as having a speech disorder if a posi- tive response was recorded for the question asking whether he or she had “any difficulty having [his or her] speech understood” or if “speech prob- lems” was identified as a condition underlying any difficulty with activities of daily living. In addition, because those aged 15-17 were treated as adults 261 

262 SPEECH AND LANGUAGE DISORDERS IN CHILDREN in the SIPP and additional questions were asked of them, children of these ages were identified as having a speech disorder if “speech problems” was identified as a condition underlying difficulty that they had with working or as a reason for their having fair or poor health. The SIPP provides various weighting variables that take into account its complex design (including clustering) and the probability of someone being in the sample. These allow for relatively direct estimates of how many people and households across the Unites States each surveyed person and household represents. Person-weights were used to calculate descriptive sta- tistics for individual-level variables, such as race and gender. For household- level variables, such as the percentage of households experiencing low food security, household-weights were used. Because child Supplemental Security Income (SSI) receipt was recorded at the household level in the SIPP, it was treated as a household-level variable. Child SSI receipt for those aged 15-17 was sometimes labeled in the survey as individual adult SSI receipt given the SIPP’s treatment of this age group as adults. These were counted as instances of child SSI and added to household-level calculations. For variance estimations, the committee used Fay’s modified balanced repeat replication method. To prevent respondents from small geographic areas from being identifiable, the SIPP alters its primary sample units (PSUs) by combining them into larger variance strata and then splitting each stra- tum into two variance units. Fay’s method is recommended by the SIPP for variance estimation because it is able to account for both halves of the strata that are generated. Resulting variance estimates are nominally con- servative. STATA (version 13.0) was used to calculate point and variance estimates using the weight and variance estimation variables in the SIPP. National Survey of Children’s Health The NSCH is nationally representative of noninstitutionalized children aged 0-17 in the United States. The survey is administered through dial- ing of random telephone numbers. One child in each household including children is then selected at random, and the survey is administered to the adult in the household who is most knowledgeable about the child’s health and health care. The 2007 NSCH included data on 91,642 children with completed surveys, and the 2011-2012 NSCH included 95,677. Children with speech disorders were identified in the survey through a series of questions. Respondents were first asked general questions about whether the child had any functional limitations or if he or she needed or received any medicines, services, therapies, or treatments. Those with posi- tive responses to any of these questions were then asked about a series of diagnoses and whether the child had received any of them. Included in this list for those aged 2-17 was the diagnosis of any speech or other language 

APPENDIX D 263 problem. Finally, for each diagnosis with an affirmative response, it was determined whether the child currently had the diagnosis. The committee used this final step to determine whether an individual child currently had a speech disorder. The NSCH provided one weight for each child in the survey based on the probability of being sampled. Numerous adjustments were then made, including adjustments for nonresponse bias, multiple land and/or cell phone lines, and subsampling of children within the household. This allowed for weighted point estimates. For weighted variance estimates, the NSCH directly provided stratum and PSU variables. These were applied using STATA for the analysis. National Survey of Children with Special Health Care Needs The NS-CSHCN is nationally representative of noninstitutionalized children aged 0-17 with one or more special health care needs in the United States. It was administered by the same organization (the Data Resource Center for Child and Adolescent Health, or DRC) as the NSCH and used similar procedures. Respondents were selected using the same methods outlined above, with the added step of having interviewers assess whether there were children in the household with special health care needs. This allowed for an estimate of the proportion of children in the country with special health care needs, in addition to detailed information about this population. The 2005-2006 NS-CSHCN included data on 40,804 children with special health care needs with completed surveys, and the 2009-2010 NS-CSHCN included 40,242 children. Regarding speech disorders in children aged 1-17, respondents were asked whether the child, compared with others his or her age, experienced “a lot, a little, or no difficulty with . . . speaking, communicating, or being understood.” Children who had either a little or a great deal of difficulty in this area were identified as having a speech disorder. Weights for point and variance estimates were constructed and pro- vided similarly to those for the NSCH, although the NS-CSHCN provided additional weights related to estimating the proportion of children nation- ally who had special health care needs. To develop estimates of the propor- tion of children in the United States with speech disorders, the proportion of children with special health care needs who had a speech disorder was multiplied by the proportion of children in the general population who had special health care needs.