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.

Final Report 84 Chapter 8: A Spreadsheet Tool for Regional and Local Analysis As shown in Chapter 7, the cost-effectiveness of increased access to NEMT varies by condition and is sensitive to the costs of transportation and healthcare. These latter two, of course, also vary regionally and locally throughout the U.S. Thus, while our analysis aims to be nationally representative, it may not reflect costs and benefits for a specific locale with significantly different rates of missed NEM trips or significantly different healthcare costs, or both. To allow local transportation and social service agencies (and other interested parties) to conduct their own analyses tailored to the local demographic and socio-economic environment, we developed a spreadsheet tool (available from the description of TCRP Project B-27 on the TRB website: (http://www4.trb.org/trb/crp.nsf/All+Projects/TCRP+B-27) that calculates condition- based cost-effectiveness that can be varied according to locally determined inputs. In many cases, we would expect these inputs to be more reliable than those used for national analyses, simply because there should be lower local variation in both transportation and healthcare costs, and the rate of missed trips may be better measured and understood at the local level. The intent of this spreadsheet tool is to allow local transportation agencies to tailor variables like the cost per trip (generally highlighted in yellow) to reflect actual data for your community. This report reviews the cost-effectiveness of providing NEMT for a number of conditions prevalent in the transportation-disadvantaged population. The spreadsheet tool accumulates the expected local savings and costs from these conditions to provide an overview of non-emergency medical transportation for a community. 8.1 Regional and Local Analysis The analysis performed and presented so far in this report examines the cost- effectiveness of receiving transportation to appropriate medical care using data estimated at the U.S. national level. In reality, local and regional costs for medical and transportation services are affected by the disease mix, the percent of the population living in urban versus rural settings, the local cost of transportation, the availability of reliable mass transit, the mode by which the transportation must be made, and local healthcare costs. Likewise, regional policies for providing NEMT for the transportation disadvantaged can have a significant effect on the cost and access to services. Depending on local costs, disease prevalence, and the other aforementioned characteristics, use of region-specific data could show that providing NEMT is more or less cost-effective in terms of the projected benefits compared to the total costs. To allow state, regional, and local agencies to forecast the effects of providing NEMT for the transportation-disadvantaged individuals who currently lack such access in a region, we developed an Excel spreadsheet tool (or model). This model embeds cost-effectiveness analysis (CEA) computations and methods previously discussed in this report, while still allowing users to specify select components of the analysis and to stratify inputs according to various relevant dimensions (e.g., urban v. rural). The model calculates the cost-effectiveness of providing needed NEMT for regions and locales according to local healthcare and transportation costs and disease prevalence estimates, as available. As an intermediate objective, the model strives to

produce a conservative estimate of the healthcare savings while trying to be realistic regarding the health treatment options and transportation costs within a region. With these inputs, model users can analyze the expected number of missed appointments for a region, the number of NEMT trips needed by type of transportation service (ambulatory, wheelchair, and stretcher), the estimated annual costs of transportation services for these trips, and the estimated annual healthcare costs for missed medical care. 8.2 Use of Local Knowledge and Data with Spreadsheet Tool The total population of the region under study drives calculations within the model. As described in Chapters 3 and 4 of this report, the ratios of the percent of the population lacking access to NEMT, the percent of the population in rural areas, the type of transportation service required for these trips, and the prevalence of chronic and other conditions are used to derive the expected number of missed trips per year. These ratios were developed using national databases, including the National Health Interview Survey and the Medical Expenditure Panel Survey. The transportation costs were estimated using data from the National Transit Database, information and data from the CTAA, and actual data from a panel of 20 transit providers, as described in Chapter 5. The tool allows local transportation agencies to tailor variables such as the region’s population, percent of the population living in areas designated as rural, the type of transportation used, and the cost per trip. These inputs are highlighted in yellow in the spreadsheet and should be changed to reflect actual data for your region. There are other inputs highlighted in green that represent national values that are generally harder to obtain at the regional or local level. Users of the spreadsheet tool are encouraged to collect data and use factors that more closely estimate regional data where they initially see national data in the spreadsheet. Using only a region’s population will provide a projection of the cost-effectiveness of providing NEMT to the transportation-disadvantaged lacking access to it, but local factors can also have a great influence on the model’s outcome; therefore, use of local factors is highly recommended. For example, while the city of El Paso, Texas and the entire state of Wyoming have populations that are roughly comparable in size, the health status of the people in these two regions is quite different. In El Paso, the prevalence of diagnosed diabetes is higher (7.4 percent) than it is in Wyoming (5.8 percent) (Sedillos, 2005; Kaiser Family Foundation, 2005). Such disparities clearly affect the outcomes of the model. Similarly, the density of the population and access to public transportation would also affect both the likelihood of missing a trip to the doctor’s office and the cost of transportation. Therefore, one would expect that using only the population size to estimate the cost-effectiveness of providing NEMT to the transportation-disadvantaged would lead to less than accurate results at the regional or local level. Not all of the factors used in the model are easily obtained for a region, and for some measures national averages will need to be used as a proxy for the regional data. National averages are highlighted in green in the spreadsheet and can be changed if accurate data for a region can be found. Upon careful review of this model, Final Report 85

transportation agencies, social service agencies, and other users are encouraged to use validated data for their regions as key inputs to the model. As is true for all models, this model provides results that are only as accurate as the data, ratios, and values that are used in the model. National ratios of prevalence or access to NEMT and average costs may not reflect actual data in a region. 8.3 Sensitivity Analysis with Spreadsheet Tool As stated above, an intermediate objective in developing this model was to develop a conservative estimate of the savings, and to be realistic regarding the healthcare and transportation costs within a region. Thus, the model includes a number of factors to account for how medical care is provided and the ability of transportation- disadvantaged people to obtain care. Some people miss appointments because they choose not to seek medical care, even when their physician has recommended it. The likelihood of making a trip takes into account those who will not seek care. This would increase the cost of care as a condition moves from well to poorly managed. While missing treatment would increase the overall cost of care as the CEAs show, providing transportation would not necessarily increase the likelihood of individuals actually receiving care. Individuals with multiple medical conditions (co-morbidities) also reduce the number of NEM trips required, because physicians can treat a patient for more than one condition during one medical visit. The data clearly show that transportation- disadvantaged persons are more likely to have more than one chronic condition than is the general population. Therefore, to include this issue in the model, the calculated number of missed trips is reduced to account for patients who see their doctor for more than one medical condition. A third factor that has a direct effect on the cost of providing transportation is mode substitution. Many people who are transportation disadvantaged nevertheless make it to their doctor’s office for non-emergency medical care using friends, family, or neighbors. Some of the trips supplied through these means could be converted to other means once additional NEMT is established in a region. If made via the newly supplied services, these trips would add to overall transportation costs but would not reduce healthcare costs, because the patient would have obtained care anyway. This substitution of services could be considered a form of induced demand. Although estimating the amount of mode substitution that would occur if additional transportation services are put into place is difficult to do, induced demand is a well recognized phenomenon in various human services policy domains. For example, in the long-term care arena, it has been known for some time that formal services (paid care) induce demand by substituting for informal services (unpaid care). The estimates of this induced demand, however, can vary anywhere from 8 to 43 percent, especially in the first year of a new program (Dale et al., 2003). If local data suggest that induced demand will not occur, then users of the tool can set this value to zero. In estimating the benefits of providing NEMT, the model follows the analysis presented in this report by including a compliance factor – the willingness of patients to follow the recommendations of their providers. If patients use transportation to access healthcare then fail to follow the provider’s requests, then all the costs of Final Report 86

transportation and healthcare are expended but few if any benefits are realized. The compliance factor for chronic conditions is generally higher than the compliance factor for preventive care. Therefore, the model expects a lower estimated benefit for preventive care when compliance is taken into account. As explained in Chapter 6 and in Appendix C, it is critically important to include health-related quality of life when evaluating healthcare interventions. Quality of life combines with life expectancy to comprise the quality adjusted life year (QALY) construct. The model enables the user to explicitly vary this parameter. This is very important for a study of individuals suffering from chronic conditions, who by virtue of enhanced transportation, can transition from poorly to well managed care and, in the process, dramatically improver their health-related quality of life. 8.4 Use and Application of Spreadsheet Tool The Non-Emergency Medical Transportation (NEMT) Cost-Effectiveness Model takes users through the following fours steps: Step 1: Estimating the Number of NEMT Trips Needed per Year Step 2: Estimating the Cost of Providing NEMT Step 3: Estimating the Benefits of Providing NEMT Step 4: Completing the Cost-Effectiveness Analysis Numbers and factors highlighted in yellow in the model should be modified to reflect known regional values. Numbers in the model that are highlighted in green come from professionally reviewed research reports and most likely should not be changed unless validated and trusted data for a given region indicates that a different value is more accurate. The primary driver of the yearly number of NEMT trips begins with the population of the region, the percent of the population that is lacking access to NEMT, and the prevalence of chronic diseases. In Step 1, a projection of the number of trips is made by estimating the number of people without access to transportation for NEM care, the chronic diseases and preventive treatment that this population should seek, and the average number of visits that a patient should have given the specified condition. As described previously, the likelihood of making a scheduled trip and the possibility that the visit covers more than one medical condition are included. With the number of visits estimated, the next step is to forecast the cost of providing NEMT, both the direct transportation cost and the cost of the resulting medical care. Rural-to-total-population ratios and type of transportation (ambulatory, wheelchair, and stretcher) are used to forecast the number of NEMT trips by type. A higher percentage of the population living in rural areas tends to increase the trip length and percent of trips that have only one rider. This increases the transportation costs. Urban areas generally have more transportation options and public transportation can be used for some trips. Regions where the population is older and sicker will likely require more transportation using wheelchairs and stretchers, also increasing the total cost of providing NEMT. There is a certain percentage of the population that will realize that paratransit for NEMT is available and will use the service instead of Final Report 87

using friends or family that they have traditionally used. This mode substitution for services will increase the cost of transportation but will not affect the healthcare costs, because the patient would have seen his or her doctor anyway. Step 3 of the model estimates the medical benefits of providing NEMT. This is based on the cost-effectiveness analysis of providing well managed care compared to poorly managed care; the underlining principle is that well managed care occurs when a patient seeks and receives care and complies with their doctor’s orders. The costs of well and poorly managed care are derived from both literature and through an analysis of the National Health Interview Survey and the Medical Panel Expenditure Survey. Condition specific cost data can be obtained for regions using these data sets with a good deal of effort. The outcome may not differ significantly from the national norms presented in the model. Compliance factors by condition have also been carefully researched. Patient behavior toward compliance is not expected to significantly differ by region and region specific compliance, especially for those needing NEMT, has not as yet been studied and reported. The final step, Step 4, is to review the results in their entirety. In Step 4, the model presents the estimated number of missed trips, the transportation and medical costs, and the QALY adjusted benefits of providing NEMT. Comparing the benefits to the costs gives the cost-effectiveness ratio of providing NEMT. Ratios above one indicate that the benefits outweigh the cost of providing care. The net cost- effectiveness forecasts the yearly cost savings or, if negative, the additional cost of providing NEMT over the medical savings generated. Final Report 88