Cost-Benefit Analysis of Providing Non-Emergency Medical Transportation (2005)

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.

Chapter 3: The Transportation-Disadvantaged Population and Access to Healthcare As stated in the Introduction, the overall goal of this study is to determine if the net benefits of providing NEMT to transportation-disadvantaged persons who currently lack access to NEMT exceed the costs of providing this service, including both transportation and healthcare costs. This determination, however, is complicated by one very important consideration: transportation-disadvantaged status alone is not sufficient to indicate that someone misses medical care due to a lack of NEMT. To miss care, one must first need care, recognize the need for care, attempt to schedule it, etc.; thus, at one extreme, a healthy person may never miss care for any reason, despite being transportation disadvantaged. Furthermore, even a clearly transportation-disadvantaged person may, perhaps through tremendous effort, succeed in obtaining needed NEMT on the one day that he or she needs it. In this instance, such a person, despite being transportation disadvantaged, is not someone who missed care due to a lack of access to NEMT. Thus, our target population for this study—transportation-disadvantaged persons who miss non-emergency medical care due to a lack of access to NEMT—lies at the intersection of two populations: (1) those who are transportation disadvantaged and (2) those who miss non-emergency medical care. This intersection is shown graphically in Figure 3-1. Figure 3-1: Identification of Target Population for This Study Transportation -disadvantaged Population Population That Misses Non- emergency Medical Care Transportation-disadvantaged persons who lack access to healthcare due to lack of transportation Estimates of the size of the two major components in the Venn Diagram in the figure suggest that they are of the same order of magnitude, but that the population that Final Report 16

misses medical care is larger. According to the 2001 MEPS, 31.63 million Americans missed non-emergency medical care for some reason or the other in a year. While estimating the total size of the transportation-disadvantaged population is more problematic and highly dependent on the definition of transportation disadvantaged used, the 2001 National Household Travel Survey (NHTS) shows that 14.5 million Americans live in households with zero vehicles, one possible definition of being transportation disadvantaged. Similarly, a study completed the Bureau of Transportation Statistics in 2002 revealed that 15.5 million Americans report having “difficulty getting the transportation needed” (U.S. DOT, Bureau of Transportation Statistics, 2003b). Of course, not all persons who miss medical care do so for transportation-related reasons, and not all transportation-disadvantaged persons miss medical care. While the intersection of these two populations is the primary focus of this study, some attention to the broader populations from which this intersection derives is warranted. These larger populations are at risk of falling into the intersection at any given point in time—that is, a transportation-disadvantaged person who was healthy may become unhealthy and then lack access to NEMT, or a person who misses care due to a non-transportation reason (e.g., lack of health insurance) may overcome his or her primary barrier but then become transportation disadvantaged and miss care for that reason. The issue of persistence arises in that different transportation- disadvantaged persons actually miss healthcare due to a lack of access to NEMT in a given year, but a larger population is at risk of missing care (and does miss care) for transportation-related reasons over time. This is shown graphically in Figure 3-2. Note that this phenomenon is well documented in other policy arenas, such as welfare and welfare reform (see, for example, Danziger and Gottschalk, 2004). In the health domain, it is widely understood that spending is concentrated on a small subset of individuals over a given period of time. For example, the sickest 10 percent of enrollees in health plans account for 70 percent of spending in any one year (Robinson and Yegian, 2004). Even with the prevalence of chronic conditions, while high-cost individuals incur a disproportionate share of spending in subsequent years, the correlation is far from perfect. One plan found that its sickest 5 percent of enrollees accounted for 45 percent of costs in one year but only 18 percent in the next. Finally, because of this persistence, or lack thereof, the benefits of disease management can be overstated. That is, if study participants for a disease management intervention are selected on the basis of high costs in a previous period, their costs would be expected to fall in subsequent periods, regardless of the effectiveness of a specific disease management protocol simply as a result of regression to the mean (CBO, 2004). Given that the primary reasons that people miss care are related to lack of insurance and healthcare cost, identifying those who miss care due to transportation barriers (lack of access to NEMT) is a challenging task, complicated by the multiple possible definitions of transportation disadvantaged. Fortunately, several sources for information regarding non-emergency medical services for both preventive services and treatment of chronic health conditions are maintained at the federal level that allow direct estimation of the number of people who miss medical care precisely due to problems accessing transportation. Two resources that have been used extensively in the health economics and health services research fields and that are largely representative of the non-institutionalized population of the U.S. are the National Health Interview Survey (NHIS, produced by the U.S. Department of Health and Final Report 17

Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics), and the Medical Expenditure Panel Survey (MEPS, produced by the U.S. Department of Health and Human Services, Agency for Healthcare Research and Quality). To identify the medical conditions that are most critical to the population that has difficulty meeting its transportation needs—that is, the most common conditions for those who have recently experienced problems with access to non-emergency care – we analyzed data from both the 2001 and 2002 National Health Interview Surveys. These data are used because they are the most recent available and because the NHIS is the most comprehensive and nationally representative attempt to understand the state of the nation’s health. Indeed, for each year, this survey has a sample size of more than 90,000 persons and covers a wide array of health-related issues. The 2002 NHIS data were released in December 2003, but further enhancements are forthcoming. In particular, the variable indicating urban/rural status has not yet been released (see Section 3.0), and some indications are that it might not ever be. Additionally, we analyzed the 2001 Medical Expenditure Panel Survey. The MEPS contains healthcare utilization and expenditure information at the individual level for more than 30,000 individuals, is nationally representative, and can be linked with the NHIS data (discussed below). Full-year data for 2001 were released on April 2004; while this project had access to 2002 MEPS data, the key variable to indicate transportation difficulties vis à vis ambulatory health utilization was altered and made less relevant to the research at hand. 3.1 Estimating the Size of the Transportation-Disadvantaged Population That Lacks Access to NEMT As made clear in the Introduction and Literature Review (Chapters 1 and 2, respectively), definitions of the transportation-disadvantaged population vary at the conceptual level, complicating estimates of the size of this population. Research on this issue demonstrates a range of factors affecting the transportation disadvantaged, and a corresponding range of estimates of the size of this population. At the empirical level, however, these complications can and must be reduced to specific variables to operationalize the concept for data queries and subsequent analyses. Because the NHIS and MEPS data sets are used to establish the medical conditions that are most critical to transportation-disadvantaged persons, we first discuss how the NHIS and MEPS can be used to identify this target population. 3.1.1 NHIS Perspective on the Transportation-Disadvantaged Population With its focus on healthcare and healthcare outcomes, the NHIS can be used to measure the size of the transportation-disadvantaged population in terms of lack of access to care. Thus, the NHIS takes an approach that meshes well with the goals and objectives of this TCRP project. Specifically, the NHIS contains the following question. There are many reasons people delay getting medical care. Have you delayed getting care for any of the following reasons in the PAST 12 MONTHS … because you didn’t have transportation? Final Report 18

For the 2002 NHIS, the weighted results to this question for the adult sample show that 1.33 percent of adults responded affirmatively as shown in Table 3-1. A separate analysis (not shown) demonstrates remarkable consistency in this ratio over the five- year period of 1998 – 2002. The four other reasons given in the survey for delaying care are: 1) Couldn’t get through on the telephone; 2) Couldn’t get an appointment soon enough; 3) Once there, had to wait too long to see the doctor; and 4) The clinic/doctor’s office wasn’t open when you could get there. Clearly, the last one of these four could include transportation-related elements, but the data are not detailed enough to parse these out from other reasons (such as could not get off work in time or the like). Table 3-1: Adults Reporting Lack of Transportation to Medical Care from 2002 NHIS Response Weighted Frequency Percentage Yes (lack of access) 2,745,947 1.33 No 201,250,000 97.78 Refused/NA/Don’t know 1,827,604 0.89 Total 205,830,000 100.00 The same question was asked about children, and the weighted results from the 2002 NHIS for the child sample show that 1.31 percent of children received delayed medical care due to lack of transportation. This result is shown in Table 3-2. Table 3-2: Children Reported to Lack Transportation to Medical Care from 2002 NHIS Response Weighted Frequency Percentage Yes (lack of access) 956,584 1.31 No 71,615,707 97.78 Refused/NA/Don’t know 397,651 0.55 Total 72,969,942 100.00 Combining the results for adults and children, the 2002 NHIS indicates that 3,702,531 individuals received delayed medical care in the past year due to transportation difficulties (2,745,947 adults plus 956,584 children), or approximately 3.7 million people. Importantly, these numbers derive from respondents directly linking a transportation rationale to delayed care over a specific one-year period. Thus, however the notion of transportation disadvantage is conceptualized, this highly rigorous, nationally representative sample reflects exactly the question of interest for this study—medical care missed due to lack of transportation. 3.1.2 MEPS Perspective on the Transportation-Disadvantaged Population The MEPS also investigates barriers to care, including transportation. For the 2001 study, 11.1 percent of respondents answered “yes” to the following baseline question. Anyone have difficulty obtaining care? Final Report 19

After responding “yes” to this question, respondents are given a list of 14 items from which they are asked to select the main reason for experiencing difficulty. Of these reasons, three are germane to transportation access: (1) medical care too far away; (2) cannot drive/no car/no public transportation; and (3) too expensive to get there. (Other reasons are: 1) Could not afford care; 2) Insurance company would not approve/cover/pay; 3) Pre-existing condition; 4) Insurance company required referral – could not get; 5) Doctor refused family insurance plan; 6) Different language; 7) Could not get time off work; 8) Don’t know where to go to get care; 9) Was refused services; 10) Did not have time or took too long; and 11) Other.) It is not surprising that access is compromised by lack of health insurance or low income. In addition to this main reason, the MEPS also gives respondents an opportunity to cite a secondary reason. Summing these responses (but only counting respondents once if they selected a transportation-related reason for both their main and secondary reason) produces a weighted estimate of 1.21 percent, approximately 3.5 million people, who cite a transportation-related reason (main or secondary response or both) to explain why they had difficulty obtaining care. While some of these reasons clearly overlap, and one could argue that this number is biased either upwardly (secondary reasons matter less than the main reason and should be discounted) or downwardly (reasons past first or second are not asked about and if asked would boost the estimate), this estimate closely matches the estimate derived from the NHIS and thus appears to be a reasonable baseline for identifying the population that misses medical care due to a lack of access to NEMT. 3.1.3 Conclusions Regarding NHIS, MEPS, and This Study Together, these two nationally representative perspectives on the intersection of delaying care and having difficulty obtaining care produces a consistent result – that is, approximately 3.6 million people in a given year most likely miss non-emergency medical trips due to transportation problems. Needless to say, the number of people missing trips does not equate to the number of missed trips, but it does provide a lower bound – by definition, all of these roughly 3.6 million people had to miss at least one round trip for non-emergency care. On average, Americans have more than three visits per year either to primary care offices, surgical specialty offices, medical specialty offices, or outpatient departments (Burt and Schappert, 2004). Additional visit information is provided in Chapter 4 to help bound an estimate of the number of missed trips associated with the study’s target population. As will be discussed further in Chapter 7, the cost-effectiveness approach used in this study conservatively assumes that all NEM trips for the target population are missed and must be accounted for in determining the cost-effectiveness of improved access. Thus, a positive result is highly encouraging, because it means that the medical care is cost-effective even if all the transportation needed to obtain it has to be funded. In the remainder of this chapter, we review this estimate of 3.6 million persons who lack adequate access to NEMT and compare it to other possible estimates to validate the reasonableness of this estimate. In Chapter 4, this target population is described in more detail, including its demographic, socio-economic, and medical characteristics. Final Report 20

3.2 Additional Estimates of the Size of the Transportation- Disadvantaged Population That Lacks Access to NEMT As discussed earlier in this report, an unambiguous estimate of the size of the transportation-disadvantaged population does not exist, and even the definition of transportation disadvantaged varies. The variation in estimates grows further if one attempts to incorporate missed trips that derive from the transportation-disadvantaged designation. Our literature review indicates that a direct, nationally representative estimate of the number of trips missed because of a lack of transportation is not currently available. In short, health-related data lack sufficient detail on transportation to directly measure the number of missed trips, and transportation data lack sufficient detail on health conditions to address utilization. Furthermore, these studies are done using entirely independent samples, making linking of data sets precarious at best. Hence, we focused on establishing and validating the estimate of the population that misses non-emergency medical care due to transportation-related reasons using a variety of available data sources. Checking it against other approaches, including use of two more data sources, the following estimates were analyzed and/or calculated: • In 2002, the Bureau of Transportation Statistics (BTS) conducted a survey to investigate the transportation status of Americans with disabilities and to compare their status with Americans without disabilities. In this study (the 2002 National Transportation Availability and Use Survey), the BTS achieved an overall sample size of 5,000 persons, roughly half of whom had disabilities (persons with disabilities were over-sampled to allow for more powerful statistical comparison of this population with the non-disabled population). This study indicated that 3.5 million Americans never leave their homes. Of these, 1.9 million have disabilities. Of these 1.9 million, 528,000 “experience transportation difficulties.” In other words, nearly 1.4 million people with disabilities who never leave the home do not report experiencing transportation difficulties. Presumably, then, they could obtain needed transportation if their other problems could be overcome. Thus, these 528,000 persons can be seen as constituting the lowest estimate of the transportation-disadvantaged population that misses medical care due to a lack of access to NEMT, because with available transportation, these individuals likely would have made at least one medical trip during the year related to their disability or otherwise; this leads to a low extreme estimate of 528,000 persons in the target population. Not surprisingly, the homebound population with disabilities of 1.9 million tends to be older (average age is 66) and is more severely disabled (58 percent report their disability as severe) than the population at large. As a result, many of the 528,000 who experience transportation difficulties likely are missing more than one medical trip per year. • The same BTS study also revealed that 57 percent of persons with disabilities who never leave the home need specialized assistance or equipment to travel outside the home, compared with only 22 percent of people with disabilities who travel outside the home at least once a week. Also, those with disabilities who never leave home have more trouble getting transportation (29 percent) than those with disabilities who leave at least once a week (11 percent). Similarly, 22.9 percent of those with disabilities need specialized Final Report 21

equipment or assistance to travel outside the home, compared with just 0.66% of those without disabilities. • As discussed earlier in this chapter, the 2001 MEPS indicates approximately 3.5 million people who cite a transportation-related reason to explain why they had difficulty obtaining medical care. • As also discussed above, the 2002 NHIS indicates approximately 3.7 million people (adults and children) who delayed getting medical care in the past year due to transportation difficulties. With a sufficiently short reference period, “delay” and “missed” are equivalent. • Again, based on the 2002 National Transportation Availability and Use Survey, the data show that approximately 12.19 percent of those with disabilities either have difficulty obtaining transportation or cannot get transportation they need for any purpose. For those who are not disabled, this value is 3.3 percent. Because approximately 23 percent of the nation’s 290 million people experience disabilities according to criteria used in this BTS study, there are a total of 15.5 million persons who cannot obtain the transportation that they need (regardless of trip purpose): 0.23 * 290 million * 0.1219 = 8.13 million people with disabilities and 0.77 * 290 million * 0.033 = 7.37 million non-disabled persons, for a total of 15.5 million. Clearly, an unknown subset of this 15.5 million people is unable to undertake medical trips because of a lack of transportation, and the data are not sufficiently refined to permit this further estimation. • The 2002 National Household Travel Survey (NHTS), also conducted by BTS, revealed that 9 percent of Americans over the age of 14 years have a “travel-affecting medical condition.” Furthermore, the NHTS clearly demonstrates that this population takes fewer trips per day than the population that does not have such a medical condition (2.8 vs. 4.4 trips per day), documenting a trip gap across trip purposes, but it does not directly point to missed trips for medical care. To summarize, the estimates of the number of persons, highlighted above, who miss medical care due to a lack of transportation—the target population for this study— are: • 528,000 (BTS) • 3,500,000 (BTS) • 3,500,000 (MEPS) • 3,700,000 (NHIS) • Fewer than 15.5 million (BTS). Because of the fairly close consistency of the MEPS and NHIS estimates, and their explicit intersection of the health and transportation domains, we have most confidence with an estimate of roughly 3.6 million people who miss at least one trip in a year due to a lack of non-emergency medical transportation. Of course, this is an estimate of the number of people who miss care, not an estimate of the number of missed visits. Importantly, even if a different estimate of the size of the transportation-disadvantaged population that misses non-emergency medical care due Final Report 22

to lack of transportation someday proves correct, only the estimates derived from the NHIS and MEPS are associated with data related to medical conditions and healthcare use; thus, only for the 3.6 million is the data rich enough to permit the cost comparisons required for this study. Nonetheless, despite confidence in the estimate of the size of the target population, there are important limitations. First, any sampling biases arising from NHIS and MEPS are also present in this estimate. For example, if non-respondents (typically poor, homeless, and in poor health) are more likely to match the target population than otherwise, then our estimate is downwardly biased. Notably, NHIS and MEPS ignore some sectors of the population such as Native Americans living on reservations who are part of the Indian Health Service, and thus these are not represented in our estimate. These individuals are known to have worse health than the general public (Barnes et al., 2005). If these unrepresented population sectors have a higher percentage of transportation-disadvantaged persons who lack access to NEMT than is present in the rest of the population, then again our estimate is low. Second, because people fall into and out of the target population over time, the population at risk of missing non-emergency medical care due to a lack of transportation certainly is higher than 3.6 million. As illustrated graphically in Figure 3-2, some transportation-disadvantaged persons will succeed in obtaining transportation to medical care on the day that they need it, and some may simply not recognize a need for medical care, even if it exists. Figure 3-2: Transportation-Disadvantaged Population at Risk of Missing Non-Emergency Care Although the available data suggest that this intersection currently contains about 3.6 million Americans in a given year, the at-risk population is larger. Furthermore, as the total population grows and continues to age, the size of this intersection can be expected to grow. Transportation- Disadvantaged Persons Transportation- disadvantaged persons who missed non- emergency medical care Transportation-disadvantaged persons who found transportation from a source that is not always available – a friend, acquaintance, family member, etc. Transportation-disadvantaged persons who found transportation from a source that is not always available – a friend, acquaintance, family member, etc. Transportation-disadvantaged persons who should be in a disease-management program or should be receiving preventive care. Transportation-disadvantaged persons who should be in a disease-management program or should be receiving preventive care. Those Who Miss Non-Emergency Medical Care Primary reasons for missing care include lack of insurance or funds to pay for care, time conflicts with appointment, refusal to seek care, etc. Final Report 23