Health Affairs, 22, no. 5 (2003): 210-221 doi: 10.1377/hlthaff.22.5.210 © 2003 by Project HOPE	New Online   Obama's Success & Next Steps   3 Health Policy Options for President Obama   A Hybrid Proposal   Bipartisan Compromise

This Article Abstract Reprint (PDF) Submit a response to this article Alert me when this article is cited Alert me when Comments are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to Citation Manager Reprints & Permissions Citing Articles Citing Articles via ISI Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by Xu, K. T. Search for Related Content PubMed PubMed Citation Articles by Xu, K. T. Related Collections Pharmaceuticals Health Spending Elderly

Datawatch

Financial Disparities In Prescription Drug Use Between Elderly And Nonelderly Americans

Abstract

 
This study examines cross-sectional disparities in the financial burden of prescription drug use among U.S. elderly and nonelderly adult populations, using data from the 1998 Medical Expenditure Panel Survey. Out-of-pocket spending for prescriptions, copayment rates, and the proportion of family income spent on prescription drugs were examined to compare elderly people with working-age adults. Even after utilization or need was adjusted for, financial disparities were still observed between elderly and nonelderly adult populations. In particular, low-income elderly people were worse off than were nonelderly adults in the same poverty class and their elderly peers in other poverty classes.

With the recent political debate over outpatient drug coverage in Medicare, the problems of cost and availability of prescription drugs among the elderly have received widespread publicity. Current Medicare drug benefit proposals vary in their aims of targeted versus universal eligibility, subsidization mechanisms, and comprehensiveness.⁶ Also, various strategies for Medicare reform have been examined to see to what extent they can curtail drug costs and promote high-quality prescription drug coverage.⁷ The lack of research on elderly U.S. consumers’ access to prescription drugs has contributed to the lack of consensus regarding a detailed plan for Medicare reform. However, whichever approach Medicare eventually chooses, it is likely to be costly to society.⁸

Despite the differences among reform proposals, a common theme in all of them is reduction of financial disparities. However, it is difficult to find a quantitative measure of disparity in prescription drug use in the literature or in practice. In addition, only through comparison between two populations can disparities be established. So far, little research has demonstrated the degree of financial disparities between the elderly and a comparison population and how the disparities would change if a Medicare drug benefit were implemented. A reduction in the elderly population’s financial burden could spill over to the working-age adult population through cost shifting. In this scenario, although financial disparities between the elderly and nonelderly adult populations would diminish, the question arises as to whether the increased financial burden for the nonelderly adult population would be justifiable. Consequently, we would want to know whether the reform is "just right" or "overshooting."

Cross-sectional comparisons of the elderly population and working-age adults (ages 18–64) serve two main purposes. First, they indicate the financial burdens associated with prescription drug use across different populations and provide quantitative measures of disparity. Second, they establish benchmarks for achieving the potential reform goals (that is, exactly how much disparity needs to be eliminated). Examination of cross-sectional disparity indicators can reveal the overall impact of Medicare reform within the elderly population and across various populations.

This study examined financial disparity indicators of prescription drug use among U.S. elderly and nonelderly adult populations to provide benchmarks to guide policy making regarding the extent of Medicare prescription drug coverage that is needed. Based on the results of this study, the impact of various reform proposals can be projected by evaluating proposed changes against the indicators to identify which one is likely to be more successful in diminishing financial disparities in prescription drug use. Three indicators of financial disparities were used: patients’ out-of-pocket spending, patients’ copayment rates, and the proportion of family income spent out of pocket. To incorporate different philosophies upon which the judgment of disparities is based, these three indicators were evaluated in two different models: a market model or utilization-adjusted disparity, and an egalitarian model or need-adjusted disparity.

Study Data And Methods

Top Study Data And Methods Study Results Editor's Notes NOTES

 
Data. To provide nationally representative estimates, data were extracted from the 1998 Medical Expenditure Panel Survey (MEPS).⁹ MEPS was designed to provide nationally representative estimates of health care use, spending, sources of payment, and insurance coverage for the noninstitutionalized U.S. population, using a complex sampling design. The sampling frame for the household survey component (one of four components) was drawn from respondents in the National Health Interview Survey (NHIS); Hispanics and blacks were oversampled.¹⁰ Accordingly, all of the estimates provided in this study are nationally representative of the 1998 U.S. population. Person weights, primary sampling units, and strata used by MEPS were controlled for in all bivariate and multivariate estimations.

MEPS collected comprehensive information on the use of health care services, recording each medical care event a respondent had in 1998, including inpatient, outpatient, emergency room, office-based provider, home care, and prescription drug use. The prescription data file contained all new prescriptions, refills, and the quantity and dosage taken in 1998; it also contained information of payments made by various sources, including the patient/family and third-party payers.

After respondents who took at least one prescription drug in 1998 were identified, the following information related to prescription drug use was extracted from the prescription file for each user: (1) annual total out-of-pocket amount for prescriptions spent by the patient or the patient’s family; (2) out-of-pocket proportion (the proportion of annual total out-of-pocket drug spending compared with the annual total amount paid by all payers combined); and (3) proportion of annual family income paid out of pocket for prescription drugs.

Annual family income (from a total of eighteen sources) was derived from the MEPS population characteristic file. Income was not limited to wages or earnings from employment. Hence, income was applicable to both working and retired people. According to MEPS, the annual person-level incomes for nonelderly and elderly people in 1998 were $27,958 and $21,192, respectively.

Unadjusted/observed disparities. This study calculated differences between elderly and nonelderly adult populations in total out-of-pocket spending for prescriptions, out-of-pocket proportion, and proportion of family income spent out of pocket for prescriptions (referred to as "income proportion" below). To provide indicators that are more relevant to policy analyses, the sample was also stratified by federal poverty level in 1998.

With this stratification, elderly and nonelderly adults within a poverty category were compared on the three measures. Because these comparisons demonstrate disparities within a poverty group, the differences between nonelderly and elderly populations in a given poverty group represent the disparities associated with being elderly, or the disparities caused by age. In addition, elderly persons in each poverty group were compared with all adults. These comparisons established which subgroup of elderly people was considered most disadvantaged across the poverty subgroups. These comparisons indicate where disparities were caused by poverty differences within the elderly population.

Utilization-adjusted disparities. From a purely market-oriented philosophy, consumers who use more prescription drugs should pay more, regardless of need. Market distortions are beyond the scope of this study, so they are not addressed. The utilization-adjusted disparities are observed disparity indicators adjusted by utilization differences between elderly and nonelderly populations. That is, if an elderly consumer uses more prescription drugs, has more refills of one prescription, and gets more pills or tablets for the same prescription during each refill, then a higher out-of-pocket payment is expected, regardless of the person’s age or level of need.

Utilization levels of prescription drugs among the nonelderly and elderly populations were analyzed first. From the MEPS 1998 prescription file, the following variables were derived: (1) annual total number of different prescriptions taken in 1998 identified by National Drug Code (NDC); (2) annual average number of refills per NDC; and (3) average quantity dispensed per NDC (for example, number of tablets). Average number of refills was calculated based on each unique NDC in a patient’s prescription profile. For example, if a prescription was obtained in March and September with two and three refills, respectively, the annual number of refills for this NDC is five, even though these two prescriptions were for different conditions. A new prescription that was filled only once was considered one refill; a new prescription that was filled twice was considered two refills. Because of the way NDCs are constructed, multiple NDCs having the same ingredients but from different manufacturers could have been counted as different prescriptions. However, we were interested only in a patient’s aggregated use of prescriptions, and counting therapeutic equivalents separately did not double-count any prescriptions. Therefore, it is not necessary for this study to be exact on which prescriptions should be grouped as a single prescription.

First, the proportion of respondents among elderly and nonelderly populations who took prescriptions during 1998 was obtained. Next, the unadjusted odds ratio (elderly versus nonelderly) of taking prescription drugs was calculated. Among prescription drug users, the differences between the elderly and nonelderly adult populations in the total number of prescription drugs, refills per prescription, and quantity per prescription were calculated. Before addressing utilization-adjusted disparities, I calculated adjusted consumer utilization to establish how utilization levels vary across the two populations. The relative odds (elderly versus nonelderly) of taking any prescriptions in 1998 and the difference (elderly versus nonelderly) in the total number of prescriptions for users were adjusted by patients’ demographic and geographic characteristics, duration of insurance in 1998, type of insurance (excluding Medicare), general physical and mental health, and diseases/conditions. In addition, total spending for medical care, excluding drugs, was used as a proxy for the utilization intensity of other health services in 1998. The utilization-adjusted disparity indicators used were the differences between the elderly and nonelderly populations for patients’ out-of-pocket payment, out-of-pocket proportion, and income proportion, adjusted for number of prescriptions, average refills, average quantity, demographic and geographic characteristics, insurance, and total spending for medical care excluding drugs. Disparities due to age and poverty differences, as discussed before, were also calculated.

Need-adjusted disparities. Another philosophy regarding the distribution of medical resources across the U.S. population is that consumers who have greater health care need should receive more care and be able to afford it. In a purely egalitarian medical care system, no financial disparities should be observed. In such a system, depending on how financial disparity is defined, elderly and nonelderly adult prescription drug users should have the same out-of-pocket payments, out-of-pocket proportions, or income proportions, even if elderly people have a greater need for prescription drugs than nonelderly people have.

The MEPS data included information about both respondents’ general health and specific diseases/conditions in 1998. Respondents rated their health as excellent, very good, good, fair, or poor. Specific diseases/conditions were identified using the first three digits of the International Classification of Diseases, Ninth Revision (ICD-9) code and a list of 259 conditions grouped by the Clinical Classification Code that was developed by the Agency for Healthcare Research and Quality based on the full ICD-9 codes. In addition, certain diseases/conditions were classified as "priority" because of their prevalence, expense, or relevance to policy. To keep the model parsimonious, the following health and disease/condition variables were used in this study: general physical health, general mental health, cancer (of any body part), diabetes, emphysema, high cholesterol, HIV/AIDS, hypertension, ischemic heart diseases, stroke, arthritis, asthma, gall bladder diseases, stomach ulcers, back problems of any kind, Alzheimer’s disease and other dementias, depression and anxiety disorders, and any other diseases/conditions not on the priority list. Some priority conditions could have subcategories. For example, emphysema included emphysema, chronic obstructive pulmonary disease (COPD), chronic bronchitis, chronic obstructive bronchitis, and smoker’s cough.¹¹

Differences between the two populations in health status and prevalence rates of each disease/condition were calculated to establish the need differences between them. Need-adjusted disparities were the differences between the two populations in out-of-pocket payment, out-of-pocket proportion, and income proportion, adjusted for general physical and mental health status, diseases/conditions, demographic and geographic characteristics, and total spending for medical care excluding prescription drugs. Disparities due to age and poverty differences were also calculated.

Study Results

Top Study Data And Methods Study Results Editor's Notes NOTES

 
Unadjusted/observed disparities. Exhibit 1 shows the unadjusted means of and differences between elderly and nonelderly adult populations in annual out-of-pocket spending for prescription drugs, out-of-pocket proportion, and income proportion, respectively.¹² In most comparisons, elderly adults bore a greater financial burden for prescription drugs than nonelderly adults did.

The comparisons between elderly people in each poverty group and all adults show the unadjusted disparities due to poverty differences within the elderly population. The out-of-pocket proportion paid by poor elderly people was not significantly different (alpha =.05) from the average for nonelderly adults. There was no significant difference (alpha =.05) between the family income proportion variables for high-income elderly and nonelderly adults.

Utilization-adjusted disparities. Exhibit 2 reports differences in prescription drug use between elderly and nonelderly adults. The two right-hand columns are adjusted by patients’ demographic and geographic characteristics, duration of insurance in 1998, type of insurance (excluding Medicare), general physical and mental health, and diseases/conditions. About 88.25 percent of elderly persons used at least one prescription drug in 1998. Variations in the raw odds of taking prescription drugs and the total number of prescriptions were small across poverty levels. However, the differences among high-income elderly people when compared with high-income adults and all adults, respectively, in terms of the relative odds of using prescription drugs were smaller than other comparisons. Also, within the poor group, the difference between nonelderly and elderly people in the total number of prescriptions was smaller than in other income groups. It is interesting to note that after the adjustment factors were controlled for, there were no differences in the odds of using prescription drugs between nonelderly and elderly people in the poor and high-income categories. However, among the other three income categories, being poorer was associated with higher odds of using prescription drugs.

Need-adjusted disparities. Exhibit 4 shows the differences in health status between elderly and nonelderly adult populations. As expected, the elderly had worse general physical, general mental, and condition-related health than non-elderly adults had. Prevalence rates for asthma and depression in the two populations were similar. Health status comparisons were also made within each poverty group (data not shown). The associations between poverty categories and various disease prevalence rates were not uniform. For example, the near-poor had the largest difference between elderly and nonelderly adults in cancer prevalence, whereas the middle-income group had the largest difference between elderly and nonelderly adults in hypertension prevalence. Overall, however, the high-income group was better off than all of the other poverty groups in health status differences between nonelderly and elderly adult populations.

Discussion Identifying financial disparities in prescription drug use is critical in designing Medicare reform. Although it has been established that consumers age sixty-five and older are financially disadvantaged in access to and use of prescription drugs, little is known about how disadvantaged they are in comparison to other populations. This study quantifies the financial disadvantages elderly consumers face and contrasts elderly people with working-age adults. Furthermore, by stratifying the elderly population by poverty groups, the results of this study identify which subpopulations of elderly consumers are most in need of prescription drug coverage. This information should assist policymakers in targeting priority subpopulations for Medicare reform. Lastly, the findings can assist policymakers in determining whether a given reform will diminish disparities in U.S. prescription drug consumption.

Two contrasting philosophies, market versus egalitarian approaches, could result in different conclusions with regard to the degree of financial disparity. This study finds that regardless of the philosophy chosen, financial disadvantages—whether they be observed, utilization-adjusted, or need-adjusted—exist for elderly consumers. In particular, the most disadvantaged elderly subgroup is the low-income group, not the poor or the near-poor. In fact, under the current system and using these gauges, after the adjustment for consumer utilization or health need, elderly consumers in the middle- and high-income groups are as disadvantaged as their poor and near-poor elderly peers. It can be argued that public assistance programs such as Medicaid are successful in helping poor and near-poor elderly people obtain medications. In particular, in terms of out-of-pocket proportion, there were no observed or need-adjusted differences between non-elderly and elderly adults in the poor and near-poor categories.

Differentiating among reform proposals. These findings provide some guidance for differentiating among various proposals for Medicare reform. The applications based on the current study can be extended to comprehensive evaluations of different proposals. Although details of these applications are beyond the scope of this paper, the benchmarks of the disparity indicators can be readily used in reform debates. For example, to establish a specific proposal’s disparity reduction, an analyst can project the average out-of-pocket spending, out-of-pocket proportion, and income proportion based on the proposed reform, then compare the raw, utilization-adjusted, and need-adjusted disparities with the benchmarks. Another application is to set the targeted disparities first. By adjusting different parameters in the reform, a proposal can be modified to satisfy the targeted disparity indicators. Perhaps a more realistic approach is to reduce the disparities by stages. In addition, instead of reducing overall disparity, reforms could target specific income groups.

Study limitations. This study was intended to portray an overall picture of financial disparities resulting from prescription drug use between elderly and working-age adult populations. Consequently, there are limitations. First, establishing disparity estimates based on comparisons between elderly and nonelderly adult populations is assumed to be socially desirable and appropriate. Second, out-of-pocket expenditures for prescription drugs were not adjusted by premiums paid. Lower out-of-pocket amount, out-of-pocket proportion, and income proportion for adults could have been the results of higher premiums paid. If this were the case, the study would overestimate the financial disparities between nonelderly and elderly populations. Third, the estimated disparities could vary if patients were grouped by diseases/conditions, because the financial burdens of various diseases are different.¹³ Fourth, financial disparities were examined only at the aggregate level and across poverty groups. Elderly–nonelderly disparities based on other demographic and social structures, such as race, sex, and ethnicity, were not addressed. Lastly, the study investigated only filled prescriptions, not all written prescriptions. The disparities could be greater if elderly patients are less likely to fill needed prescriptions than their younger counterparts.

Elderly consumers overall are financially disadvantaged in out-of pocket spending for prescription drugs, out-of-pocket proportion, and come proportion, compared with working-age adults. The disadvantages persist even after utilization and health care need are adjusted for. In particular, the low-income elderly are worse off than nonelderly adults in the same income group and than their elderly peers in other income groups. Medicare reform to include drug coverage will be essential to diminish these financial disparities. Appropriate measures of financial disparities in drug use are needed to evaluate different reform proposals, so that disparities in the financial burden associated with prescription drugs can be lessened, if not eliminated altogether.

Editor's Notes

Top Study Data And Methods Study Results Editor's Notes NOTES

 
K. Tom Xu is an assistant professor of health economics in the School of Medicine, Texas Tech University Health Sciences Center, in Lubbock.

The author thanks two anonymous reviewers for their insightful comments and suggestions.

NOTES

Top Study Data And Methods Study Results Editor's Notes NOTES

 

1. See S.W.Letsch, "National Health Care Spending in 1991," Health Affairs (Spring 1993): 94–110; S.H.Long, "Prescription Drugs and the Elderly: Issues and Options," Health Affairs (Spring II 1994): 157–174; and J.A.Poisal and L. Murray, "Growing Differences between Medicare Beneficiaries with and without Drug Coverage," Health Affairs (Mar/Apr 2001): 74–85.
2. Poisaland Murray, "Growing Differences."
3. M.A.Laschober et al., "Trends in Medicare Supplemental Insurance and Prescription Drug Coverage, 1996–1999," 27 February 2002, www.healthaffairs.org/WebExclusives/Laschober_Web_Excl_022702 (15 June 2003).
4. Long, "Prescription Drugs and the Elderly."
5. M.K.Anglin and J.C. White, "Poverty, Health Care, and Problems of Prescription Medication: A Case Study," Substance Use and Misuse (14 December 1999): 2073–2093; T.J.Reutzel, "Pharmacists’ Estimates of How Often Patients Cannot Afford Their Prescriptions and What Pharmacists Can Do to Help," Journal of Social and Administrative Pharmacy 10, no. 3 (1993): 180–182; M.D.Schoen et al., "Impact of the Cost of Prescription Drugs on Clinical Outcomes in Indigent Patients with Heart Disease," Pharmacotherapy (12 December 2001): 1455–1463; and W.J.Strickland and C.M. Hanson, "Coping with the Cost of Prescription Drugs," Journal of Health Care for the Poor and Underserved 7, no. 1 (1996): 50–62.[Medline]
6. M.McClellan, I.D. Spatz, and S. Carney, "Designing a Medicare Prescription Drug Benefit: Issues, Obstacles, and Opportunities," Health Affairs (Mar/Apr 2000): 26–41.
7. H.A.Huskamp et al., "The Medicare Prescription Drug Benefit: How Will the Game Be Played?" Health Affairs (Mar/Apr 2000): 8–23.
8. S.Christensen and J. Wagner, "The Costs of a Medicare Prescription Drug Benefit," Health Affairs (Mar/Apr 2000): 212–218.
9. Agency for Healthcare Research and Quality, "Medical Expenditure Panel Survey," www.ahrq.gov/data/mepsix.htm (15 June 2003).
10. More details on data collection processes, survey design, and methodology are available in J.P.Vistnes and A.C. Monheit, Health Insurance Status of the U.S. Civilian Noninstitutionalized Population, 1996, MEPS Research Findings no. 1, Pub. no. 97-0030 (Rockville, Md.: AHRQ, 1997).
11. For a detailed listing, see AHRQ, "MEPS HC-027: 1998 Medical Conditions," December 2001, www.meps.ahrq.gov/Pubdoc/HC027/h27doc.pdf (15 June 2003).
12. The counts of prescribed medicine use were based on household reports, and the details of medication, spending, and payment data were obtained from pharmacies. Consequently, prescription drug use may be underreported. In addition, over-the-counter (OTC) drug use was not included in MEPS. These two factors could explain the differences in the findings regarding drug spending between the current study and data from the 1998 Medicare Current Beneficiary Survey (MCBS). However, if the probabilities of underreporting are similar between the elderly and nonelderly adult populations, the calculated elderly–nonelderly differences might not have been greatly influenced by the underreporting bias.
13. The use of different specifications of health care need could alter the magnitude of the estimated elderly–nonelderly disparities. As suggested by a study of gender differences, the more detailed the health measure, the smaller the gender difference. K.T.Xu and T.F. Borders, "Gender, Health, and Physician Visits among Adults in the United States," American Journal of Public Health (July 2003): 1076–1079. In the current study, an alternative specification of using 258 dummies for the Clinical Classification Codes was also tested. Similar to the conclusions drawn by Xu and Borders, elderly–nonelderly differences were smaller but still significant at alpha =.01. In another specification, where health is measured by only overall physical and mental health, elderly–nonelderly differences were greater. To keep the model parsimonious and to include somewhat detailed health conditions, we used the specification with intermediate complexity (overall health and selected priority conditions), following the conclusion from the Xu and Borders study.

                      What's this?