Preparing Airports for Communicable Diseases on Arriving Flights (2017)

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.

13 This chapter presents results and observations from the online survey. Detailed aggregated survey data are presented in Appendices D-1 (Airports) and D-2 (Health Departments). Fifty airports responded to the data request, with 49 submitting complete survey responses and one responding by e-mail. When fewer than 49 airports responded to a particular question, the actual number (“n”) is noted. In chapters three and four, the survey questions, and the data appendix of this report, “international” refers to flights “arriving from outside the U.S. and Canada.” Types and Frequency oF arriving FlighTs Of the types of flights, not all carry passengers, but all eight types have flight crews. The implication of this is that airport communicable disease plans need to consider arriving passengers and flight crews. As noted, there is some variability in how these plans address both types of groups. More than 60% of the responding airports receive 1,001 or more international flights per month, which is not surprising given the heavy weighting in the synthesis sample toward large hubs, espe- cially those that serve international markets (20 of 50 responding airports are large hubs). (Note: The actual percentage is more accurately 63.1% because Memphis International Airport chose to exclude FedEx flights from its response to this question.) The number of international flights arriving per month is a major determinant of staffing size and facility requirements for processing incoming passengers and flight crews. Of the 49 survey respon- dents, 40 (81.6%) are a port of entry with U.S. Customs and Border Protection (CBP) or Canada Border Services Agency (CBSA) on site. The size of the CBP or CBSA staff on site ranged from zero (seven airports, 14.3%) to more than 30 (23 airports, 46.9%). Ten (20.4%) airports report having one to five such staffers; five (10.2%) airports report having six to ten staffers; and four airports (8.2%) report having 11 to 30 staffers. The physical size of CBP or CBSA spaces is an indicator of poten- tial crowding and flexibility for isolating incoming passengers. The space ranges from less than 500 square feet (four airports, 8.2%) to greater than 5,000 square feet (19 airports, 38.8%), with 17 airports falling in between; in addition, seven airports reported having no CBP/CBSA facility, and two were uncertain about the size of the area that would be used for passenger isolation. experiences wiTh communicable disease responses and preparaTions Table 1 shows the communicable diseases the surveyed airports and health departments have responded to or prepared to respond to. In the case of nearly every disease, and not surprisingly, the health departments have experienced or prepared for more kinds of communicable disease incidents than have the partner airports. how airporTs expecT To learn oF a poTenTial communicable disease issue on an arriving FlighT ICAO Document 4444 Paragraph 16.6 specifies standard procedures for the pilot on an international flight to notify authorities of a suspected communicable disease onboard the plane. The pilot is required to notify the air traffic control (ATC) sector in charge of the plane at the moment, and the chapter three survey resulTs

14 ATC sector notifies ATC at the destination and departure point (or other authority designated by the country) and the airline. ATC at the destination notifies the public health authority at the destination (in the United States, coordination is through the Domestic Events Network), the aircraft operator, and the arrival airport. The following hypothetical example, prepared with assistance from Portland International Airport and CDC, illustrates these dynamics: If a communicable illness is suspected on a flight from Tokyo to Portland, the pilot contacts the Anchorage Oceanic Air Traffic Service (ATS) Flight Information Region (FIR) with the required standard information (see Appendix B, this report) before arrival into the United States. In accordance with U.S. regulations, the pilot is also required to notify the CDC Quarantine Station of Jurisdiction, however, this is not part of ICAO 4444. The Anchorage Center would then communicate the information to the Seattle ATC Center and ATS in Japan. The Seattle ATC Center notifies the Domestic Events Network (DEN), the airline, and Portland International Airport. The DEN conveys the information to the Centers for Disease Control and Prevention (CDC) Quaran- tine Stations and other federal response agencies (DHS, CBP, etc.). As a redundant method of communication, the Seattle ATC may notify the Quarantine Station of jurisdiction, CBP, and EMS response partners directly. Airport Operations notify EMS and other response partners. Despite the clear statement of the notification process in ICAO Document 4444 Paragraph 16.6, there is great variability in how U.S. and Canadian airports receive notification of a potential communicable disease issue on an arriving flight. Table 2 summarizes the actual responses by 49 air ports, annotated by the extent to which they conform to the standard stated in ICAO Docu- ment 4444. The responses are scored according to how many of the three pathways described in ICAO Document 4444—direct from ATS, through the public health agency, and/or through the aircraft operator—appear to be included in the airports’ responses. The scoring is necessarily a rough TABlE 1 AIRPORT AND HEAlTH DEPARTMENT RESPONSE AND PREPARATION FOR COMMUNICABlE DISEASES (Since 2003) Disease Airports (n = 49) Health Departments (n = 37) n Percent n Percent Ebola 21 42.9 28 75.7 H1N1 influenza 20 40.8 24 64.9 SARS 20 40.8 16 43.2 Tuberculosis 19 38.8 22 59.5 Other influenza types (includes bird flu, swine flu, H5N1) 20 40.8 12 32.4 Measles 15 30.6 23 62.2 Zika 11 22.4 16 43.2 Norovirus 9 18.4 11 29.7 Dengue 5 10.2 6 16.2 Chikungunya 4 8.2 6 16.2 MERS-CoV 3 6.1 1 2.7 Foot and mouth disease 1 2.0 0 0.0 Mad cow disease 1 2.0 0 0.0 Pertussis (whooping cough) 1 2.0 0 0.0 Meningococcal meningitis 1 2.0 1 2.7 Varicella (chicken pox) 0 0.0 2 5.4 Anthrax 0 0.0 1 2.7 Cholera 0 0.0 1 2.7 Mumps 0 0.0 1 2.7 None of the above 17 34.7 2 5.4 Source: Smith and Greenberg data.

15 estimate. However, it is clear that every airport appears to know what the mechanism is for learning of a communicable disease on an arriving flight. No airport scored zero of three. Although most airports do not indicate receiving notice of a potential communicable disease by all three pathways implicit in ICAO Document 4444, every airport reports at least one established pathway for receiving notice. The survey did not address notification procedures for domestic flights. naTure oF airporT communicable disease or quaranTine plans Of the 49 airports responding to the survey, 43 (87.8%) have written communicable disease or quarantine plans. Thirty-eight (77.6%) of these plans are airport centered, and five (10.2%) apply to the airport but are maintained by the local public health preparedness coordinator. Five (10.2%) airports, all of which are reliever airports, do not have a written plan, and one airport reported being unsure whether it has a plan. The relationship between the communicable disease plans and other major airport plans varies. The survey did not ask if an airport has a stand-alone plan referenced in its AEP, but current practice is to have detailed plans separate from AEPs and referenced in the AEPs so that revisions to the detailed plans do not trigger reviews of entire AEPs, either by the U.S. FAA or Transport Canada. Twenty (40.8%) airports indicated that their communicable disease response plans are reflected in the airport’s crisis communications or emergency communications plan; 15 (30.6%) airports reported ICAO Model: Pilot notifies current ATS; ATS notifies destination and departure ATS, also notifies state health agency, aircraft operator (or designee), and destination airport. The destination airport ideally receives notice from the health agency, the aircraft operator, and directly from ATS. Aircraft operator’s designee for medical issues is often a company such as MedLink. Response n Match with ICAO Model A report from the pilot to ATC, then through the airline (possibly MedLink) and/or the national health agency (CDC/PHAC) 19 3/3 CDC/PHAC established process and internal protocols 8 1/3 CBP/CBSA established process and internal protocols 3 2/3 Pilot or Airline operations 2 1/3 911 Comm Center, DOH, CDC, CBP 1 1/3 Airline, public health, or emergency responder agency 1 2/3 Fire Department, airport Operations Center, airport Response Coordination Center 1 1/3 Interaction with CDC, U.S. Dept. of Health, MedLink, Air Carrier, Fire Dispatch (911) 1 2/3 It’s hoped that a flight crew’s discussion with an inflight medical consultant (such as MedLink) would give us advance notice; however, I’m afraid we may not learn of the issue until our first responders have already made contact with the ill passenger. 1 1/3 Local Health officials, city or county. CDC—Detroit may call with a heads-up 1 1/3 Notification from either the airline or United States Custom and Border Protection. 1 2/3 Pilot notifies the CDC or Local Health Department who in turn notifies the airport. 1 1/3 Through either the FAA Tower or Airline Operations 1 2/3 Tower communication 1 1/3 Typically, our dispatch center is notified by the airline, by an in-flight medical advice provider, or CBP. Atypically, we might also be notified by one of four county health departments. 1 2/3 Source: Smith and Greenberg data. TABlE 2 HOW AIRPORTS SAy THEy lEARN OF A COMMUNICABlE DISEASE ON AN ARRIvINg FlIgHT

16 their plans do not. An additional 14 (28.6%) airports responded that they were unsure. This reflects the trend uncovered by ACRP Synthesis 73 (Smith et al. 2016b) that showed airports increasingly seeking to integrate all their communications needs and tools into comprehensive crisis communica- tions plans; it also reflects what is increasingly a standard to develop strategic integrated plans. Most airport communicable disease response plans are not categorized as sensitive security infor- mation (SSI). SSI is a form of access control to sensitive documents. It corresponds to the classifi- cations of “For Official Use Only” or “Confidential.” Standards for what makes a portion of a plan SSI vary among airports and depend on the relationship between the TSA/Canadian Air Transport Security Authority (CATSA) security director and the airport’s management. That most of the plans are not SSI likely reflects their interest in the widest possible dissemination of the plans among stake- holders and the public. Thirty-two (65.3%) airport plans are not SSI, and 13 (26.5%) are SSI. Four (8.2%) reported not knowing if their airport’s plans were classified as SSI. Being able to communi- cate plans and procedures to stakeholders and the public during a health emergency can be critical to the success of the response and the protection of the airport’s reputation, as noted in the discussion of case examples in chapter four of this synthesis. Other survey questions looked at relationships to other major plans. Twenty-one (42.9%) of the airports participating in the current study report that U.S. CBP or CBSA has communicable disease/ quarantine plans at those airports, and four (8.2%) have no CBP/CBSA plan. Twenty (40.8%) air- ports responded they do not know the status of such plans, and four (8.2%) responded that the ques- tion was not applicable. Few communicable disease plans differentiate between how an airport handles flight crews who have been exposed to an infectious illness and how passengers are handled: 12 (24.5%) airports make this distinction, whereas 29 (59.2%) do not. Five (10.2%) airports indicated that they do not know, and three (6.1%) say it is not applicable (most often, these respondents were general aviation airports with no scheduled or charter passenger flights). Among the 32 U.S. health departments responding to this question, six (18.8%) said flight crews are treated differently, 18 (56.3%) said there was no difference, two (6.3%) responded not applicable, and six (18.8%) did not know. Among the five Canadian health agencies that responded, one responded that its plan differentiates the handling of flight crews from that of passengers, whereas three said there is no difference, and one said not applicable. Nearly three-quarters of the airport communicable disease plans (36 of 49 airports, 73.5% of sur- vey respondents) are based on the principles and procedures of the National Incident Management System (NIMS), whereas seven (14.3%) are not and six (12.2%) reported not knowing. Most of the airports that participated in this study use social media to communicate with the pub- lic during communicable disease incidents. In most cases, the messages appear to be worked out cooperatively between the airport and the local health department and/or quarantine station but are issued (and presumably monitored) by the airport. Precise numbers are not available for this item because the question was open ended, with respondents providing a wide variety of answers that do not support more detailed analysis. ACRP Synthesis Project A11-03 (S04-18), “Using Social Media to Inform Response and Recovery during Airport Emergencies” (expected publication in mid-2017), offers related information. When asked if their plan takes into account the possibility of social media posts by a passenger about disease on an inbound flight while the flight is en route, 19 (38.8%) airports said yes, 18 (36.7%) said no, four (8.2%) said not applicable, and eight (16.3%) reported not knowing the answer. mainTaining and reviewing plans given how rapidly the public health and communications environments are changing, it might be expected that airports would commit to regular reviews of and revisions to their plans. Thirty-one (63.3%) airports reported they perform regularly scheduled reviews and updates of their commu- nicable disease response plans. When the health departments were asked the same question about their department’s plan, the results were similar, with 56.3% responding that they perform regularly scheduled reviews and updates.

17 Training, drilling, and exercising plans Of the 44 airports that described their training programs for communicable disease response, nearly all of them (n = 42 or 95.5%) responded that they use exercises (most often tabletop exercises), training sessions, online training, or some combination of these activities. Several reported that they train and drill in conjunction with their annual review of their plans. Several of the larger airports noted they have regularly scheduled briefings with CDC/PHAC partners. Five (11.4%) airports cited after-action reviews (AARs) following exercises or real-world incidents as their most valuable training. When specifically asked how the airport evaluates learning from its communicable disease training, education, drilling, and exercising program, 17 (38.6%) airports cited AARs or AARs/improvement programs (IPs). Seven (15.9%) airports reported using formal evaluation surveys after training. The health departments were asked if they had trained jointly with their partner airport on com- municable disease response. Of the 37 health departments that participated in the survey, 21 (56.8%) responded yes, 11 (29.7%) responded no, two (5.4%) responded “not applicable,” and three (8.1%) did not know whether joint training had been conducted. The airports and health departments were asked at what intervals or on what occasions they exercise their plan. Respondents were allowed to include multiple answers in their responses. The results, reported in Table 3, show that the two occasions cited most often for having an exercise or TABlE 3 INTERvAlS AND OCCASIONS ON WHICH PlANS ACTUAlly ExERCISED Interval or Occasion Airports (n = 49) Health Departments (n = 37) n % n % When we become aware of a new threat 18 36.7 14 41.2 When new procedures are introduced 15 30.6 9 26.5 Annually 12 24.5 8 21.6 When regulatory requirements change 11 22.4 4 11.8 A real-world incident can substitute for 10 20.4 15 44.1 As part of after-action review/improvement 10 20.4 8 23.5 Every 2 years 10 20.4 4 11.8 To validate plan revisions 7 14.3 11 32.4 Every 6 months 3 6.1 0 0.0 Every 3 years 2 4.1 13 38.2 Every 3 months or more often 2 4.1 0 0.0 Upon request by stakeholder 1 2.0 4 11.8 Continuous 1 2.0 1 2.9 Monthly 1 2.0 0 0.0 Part of annual AEP review and update 1 2.0 0 0.0 Not applicable 8 16.3 2 5.9 Source: Smith and Greenberg data.

18 drill—upon learning of a new threat or when new procedures are introduced—are closely linked, and both probably are associated with risk-based emergency management and safety programs in recent years. Two serious global public health emergencies (Ebola, Zika) in the past 2 years have intensified this focus. The fixed time interval data may be confounded because a responding airport could mark “monthly” without marking every 3 months or 6 six months. What is meant by “continuous” exercise cannot be determined from the data. The airports and health departments were also asked what the ideal interval would be for drilling or exercising a communicable disease response plan. This was an open-ended question, so some respondents gave answers that included two or more components. There is a consensus among air- ports and health departments that annual exercises would be ideal. This would represent a significant increase in exercise frequency compared with what both groups report doing at present. Annual exer- cises occur only among one in four airports and slightly more than one in five health departments. communicable disease–relaTed services provided by healTh deparTmenTs To airporTs The surveys of U.S. and Canadian health departments asked what types of services they expected to provide at their partner airports related to communicable disease responses. Their responses are presented in Table 4. Although Canadian and U.S. responses have been combined, they are not strictly comparable. All the U.S. responses came from local health departments, whereas the Canadian responses came from one local health department, one provincial health department, and three PHAC regional offices. Despite this variability, the most valid message in these data is the range of services an airport can expect to receive from its public health partner. Also notable is the discrepancy between U.S. and Canadian health partner organizations regarding public communication during an emergency. All Canadian respondents indicated this was an area of leadership for them, whereas only 71.9% of U.S. TABlE 4 HEAlTH SERvICES PROvIDED TO AIRPORTS DURINg COMMUNICABlE DISEASE RESPONSES Service U.S. Health Department (n = 32) Canadian Health Department (n = 5) Combined (n = 37) n % n % n % Liaising with state/provincial health department/CDC/PHAC as needed to identify disease agent and arrange for laboratory testing 30 93.8 4 80.0 34 91.9 Investigating cases and collecting epidemiological information, including interviewing ill and exposed individuals (passengers, flight crew, airport staff, etc.) 29 90.6 5 100.0 34 91.9 Instituting control measures (isolation and quarantine or other measures necessary to control disease spread) 27 84.4 5 100.0 32 86.5 Providing guidance regarding treatment or prophylaxis that may be needed for ill/exposed individuals 27 84.4 2 40.0 29 78.4 Providing guidance regarding appropriate PPE and infection control measures 26 81.3 2 40.0 28 75.7 Leading any public information/messaging efforts in partnership with airport 23 71.9 5 100.0 28 75.7 Providing guidance on environmental cleaning measures/waste disposal 22 68.8 3 60.0 25 67.6 Collecting environmental samples 14 43.8 3 60.0 17 45.9 Source: Smith and Greenberg data.

19 health departments reported the same. The relationship between the airport and its health prepared- ness partners is explored further in the six case examples in chapter four. dealing wiTh a surge in demand For healTh services aT an airporT A surge is a sudden requirement for greatly enhanced staff size to carry out one or more aspects of an emergency response. The 37 responding health departments reported four basic patterns of dealing with a surge in demand for services and health personnel for a communicable disease response at an airport, and in some cases agencies used combinations of these approaches: 1. Reallocation of agency personnel, often nurses with additional specialized training for com- municable disease functions; 2. Augmentation from regional, state/provincial/territorial, or national resources (e.g., the disaster medical assistance team, regional EPI team, or regional health care coalition); 3. Hiring of temporary employees; or 4. Use of volunteers such as the Medical Reserve Corps. Surges in demand for health services, whether in the United States or Canada, typically necessi- tate a partnership approach. The airport plan would trigger district health plans and coalition plans. This in turn could trigger the infectious disease network at the state level. The coalition in partner- ship with public health and emergency management has a surge plan that would be activated, with communications taking place all the way up the line to the state department of health and including assistance from the state health care association. The third of these examples shows the benefit of spelling out the triggers and procedures for a surge situation in the airport’s communicable health disease plan. proTecTive measures provided To airporT employees Airport communicable disease response plans typically provide special mention of personal protec- tive equipment (PPE) for employees. Table 5 summarizes the types of PPE provided. There appears to be a relationship between airport size and the magnitude of a recent communicable disease inci- dent, on one hand, and the range of PPE provided at an airport, on the other. There appears to be an issue at many airports regarding who is responsible for the OSHA-required fit testing of personnel for respirators (l. Slepski, personal communication, April 6, 2017). TABlE 5 PROTECTIvE MEASURES PROvIDED TO AIRPORT EMPlOyEES Measure n % Training 40 81.6 Personal protective equipment 39 79.6 Equipment 21 42.9 Decontamination facilities 19 38.8 Other clothing 11 22.4 Vaccines 10 20.4 Countermeasures 7 14.3 Medical checks 7 14.3 Counseling 1 2.0 Not applicable 3 6.1 Don’t know 1 2.0 Source: Smith and Greenberg data.

20 lessons learned The airports and health departments were asked to identify the top three lessons they have learned con- cerning communicable diseases on arriving flights. The question was open ended, allowing respon- dents to discuss the widest possible variety of experiences. The verbatim or lightly edited lessons learned that were stated by the airports and local health departments are reproduced in Appendix E. The most important item revealed by comparing the lessons learned by airports and those learned by health departments is how much those lessons overlap. The two attention-getting communica- ble diseases—influenza A virus subtype H1N1 (often called swine flu) in 2009 and Ebola in 2014, but especially Ebola—that brought airports and public health agencies together to plan their joint responses probably account for the overlap. The most significant differences come from the airports’ concern with handling an aircraft on the ground compared with the health departments’ greater con- cern with resource constraints, legal questions, and restrictions or delays on access by public health officials to planes because of airport security procedures. Both groups are concerned by the gaps in authority and information sources regarding a passenger who becomes ill on a domestic flight. selF-esTimaTes oF preparedness level The airports and health departments self-reported their estimated level of preparedness to respond to a communicable disease on an arriving flight. When data from the 37 matched pairs of airports and health departments are the only data considered, the two estimates agree in 24 (64.9%) pairs, with “somewhat prepared” being the most common shared evaluation. In three (8.1%) pairs, the airport rated itself more prepared than its partner health department rated itself (Table 6). In ten (27.0%) pairs, the health department rated itself more prepared to respond at the airport than the airport rated itself. There was only one case in 37 in which the perception of readiness varied widely: one responded “very prepared,” and the other responded “not prepared.” The examination of answers to the follow-up question, “What is the one thing you feel your organization could or should do to enhance its preparedness to respond to a communicable disease on an arriving flight?” suggests that the two main causes of disagreements and the preponderance of “somewhat prepared” ratings are failure to exercise or drill together frequently and infrequent review and revision of plans. When only large hub airports in the surveyed pairs are considered (Table 7), there appears to be a shift toward a higher level of perceived preparedness, but the small sample size and subjective nature of the self-reported preparedness estimates limit the generalizability of this observation. TABlE 6 COMPARISON OF SElF-REPORTED PREPAREDNESS ESTIMATES Local Public Health Preparedness Coordinator Estimate of Its Preparedness to Respond at the Airport A ir po rt E st im at e of P re pa re dn es s Very Prepared Somewhat Prepared Not Prepared Very Prepared 9 9 — Somewhat Prepared 2 15 1 Not Prepared 1 — — Source: Smith and Greenberg data. TABlE 7 COMPARISON OF SElF-REPORTED PREPAREDNESS ESTIMATES: lARgE HUBS Health Department Estimate of Its Preparedness to Respond at the Airport A ir po rt E st im at e of P re pa re dn es s Very Prepared Somewhat Prepared Not Prepared Very Prepared 7 4 0 Somewhat Prepared 0 5 0 Not Prepared 0 0 0 n = 16. Source: Smith and Greenberg data.

21 Taking a similar look at airports that have commercial passenger service but are not large hubs— that is, medium hubs, small hubs, nonhub primaries, and secondary commercial service airports—the most frequently chosen response is “somewhat prepared” (Table 8). Two factors probably account for the difference from that seen for large hub airports’ estimated preparedness: (1) the airports that are not large hubs usually have much smaller airport staff sizes, and (2) they lack international flights that would tend to heighten awareness of communicable disease risks and procedures. Airport–health department pairs in both size categories report being prepared to some degree, implying that they are aware of the risks, procedures, and resources available to them. In addition, the members of each pair in the two groups agree on their preparedness estimates: 12 pairs of 16 (75%) for the large hubs, and 11 pairs of 14 (79%) for the other airports with passenger service. Airports in these categories, especially medium hub and small hub airports, are increasingly likely to receive international flights as low-cost, long-haul airlines establish intercontinental service to the United States and Canada using airports that have not previously had international flights (see for example Anna.aero 2017). A reliever airport is a specifically designated general aviation airport that serves to divert general aviation traffic from nearby commercial passenger airports. The seven reliever airports in the study and their local health department partners reported a somewhat different pattern (Table 9). Only one pair of seven (14%) agreed on the estimated level of preparedness. This may perhaps result from a lower level of interaction and collaboration between the airports and the health departments. In general, reliever airports are much lower profile in their communities than are airports with commer- cial service. Six among the seven reliever airports in the surveyed pairs were among the 20 busiest in terms of flight operations in the United States in 2009, ranking first, third, seventh, 11th, 14th, and 18th (National Business Aviation Association 2017). All seven airports routinely serve nonstop international flights by corporate and private jets. Examining the percentage of airports (not pairs) self-reporting themselves as “very prepared” shows that the large hub airports (ten of 16, 62%) and reliever airports (five of seven, 71%) are more similar than either is to the smaller airports with commercial passenger service (two of 14, 14%). TABlE 8 COMPARISON OF SElF-REPORTED PREPAREDNESS ESTIMATES: MEDIUM HUBS, SMAll HUBS, NON-HUB PRIMARy, AND SECONDARy COMMERCIAl SERvICE AIRPORTS Health Department Estimate of Its Preparedness to Respond at the Airport A ir po rt E st im at e of P re pa re dn es s Very Prepared Somewhat Prepared Not Prepared Very Prepared 1 1 0 Somewhat Prepared 1 10 1 Not Prepared 0 0 0 n = 14. Source: Smith and Greenberg data. TABlE 9 COMPARISON OF SElF-REPORTED PREPAREDNESS ESTIMATES: RElIEvER AIRPORTS Health Department Estimate of Its Preparedness to Respond at the Airport A ir po rt E st im at e of P re pa re dn es s Very Prepared Somewhat Prepared Not Prepared Very Prepared 1 4 0 Somewhat Prepared 1 0 0 Not Prepared 1 0 0 n = 7. Source: Smith and Greenberg data.