| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Nancy E. Schoenberg, PhD, Department of Behavioral Science, University of Kentucky, 125 College of Medicine Office Building, Lexington, KY 40536-0086. E-mail: nesch{at}uky.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionLimitationsReferences |
|---|
Key Words: Multiple morbidity • Medical expenditures • Financial status
The financial toll wrought by multiple morbidities is a significant challenge for many older adults. Multiple morbidity often increases older adults' economic vulnerability (Mueller, Schur, & O'Connell, 1997; Bierman & Clancy, 2000; Crystal, Johnson, Harman, Sambamoorthiu, & Kumar, 2000; Stroupe, Kinney, & Knieser, 2000; Tu, 2004). Because those individuals who endure the greatest frequency of multiple morbidities tend to be those with the fewest financial resources, multiple morbidity can pose a significant burden to older adults with lower incomes.
In this article, we explore the key financial issues related to the most commonly occurring multiple morbidities, focusing on how specific constellations of illnesses impact out-of-pocket health care expenditures. Exactly what is considered an out-of-pocket health care expenditure varies across studies. In general, health-related out-of-pocket expenses include the purchasing of prescription and nonprescription medications, medical supplies, and home-centered treatments not covered by Medicare; insurance copayments or deductibles; nutritional and mobility-related expenditures; and noncovered dental expenses. Most studies do not include health insurance premiums as out-of pocket expenditures.
Multiple Morbidity: An Increasing and Compelling Trend with Measurement Challenges
Between 29% and 73% of older adults (older than 65 years of age) in the United States have at least two co-occurring chronic conditions (Fillenbaum, Pieper, Cohen, Cornoni-Huntley, & Guralnik, 2000; Fortin, Bravo, Hudon, Vanasse, & Lapointe, 2005). A 1999 survey found that 48% of Medicare beneficiaries had at least three chronic conditions and 21% had five or more conditions (Anderson & Horvath, 2002), a percentage similar to Hwang and colleagues' analysis (Hwang, Weller, Ireys, & Anderson, 2001) of data from the Medical Expenditures Panel Survey. Even among middle-aged adults, Hoffman and colleagues estimated that 51% have multiple morbidities (Hoffman & Rice, 1996). In all, approximately 60 million Americans have multiple morbidities, a number that is expected to increase to 81 million by 2020 (Anderson & Horvath, 2002).
In addition to escalating rates of multiple morbidity, several factors have compelled researchers and clinicians to focus on this topic. First, those individuals with several simultaneously occurring illnesses tend to report poorer quality of life, greater disability, and increased loss of mobility than those without multiple morbidities (Fortin, Lapointe, Hudon, Vanasse, & Ntetu, 2004; Gijsen, Hoeymans, Schellevis, Ruwaard, Satariano, & van den Bos., 2001; Guralnik, 1996; Hoffman et al., 1996). The longitudinal Established Populations for Epidemiologic Studies of the Elderly project demonstrated that the relative risk of mobility loss between those with no chronic diseases and those with one, two, three, or four or more conditions was 1.4, 1.7, 2.5, and 2.9, respectively. (Guralnik et al., 1993). Over the long run, the type, number, duration, and severity of these chronic conditions has a direct relationship with the maintenance of independence and may affect longevity (Cornoni-Huntley, Foley, & Guralnik, 1991).
Adding to the toll that multiple morbidities take on the health and functioning of older adults, the management of multiple illnesses may lead to serious financial consequences for both elders and the nation as a whole. Stroupe and associates (Stroupe, Kinney, & Knieser, 2000) reported that having even one chronic illness decreased the probability of adequate insurance coverage by 10% and 25% among all individuals and single individuals, respectively. Moreover, managing multiple chronic conditions can increase patient use of specialty physicians, many of whom will not accept standard Medicare reimbursement, thereby increasing the potential for higher out-of-pocket payments (Starfield, Lemke, Herbert, Pavlovich, & Anderson, 2005).
Finally, multiple morbidity raises concerns about the sustainability of Medicare and other national health programs. In 2001, U.S. health expenditures amounted to $1,424.5 billion, or 14% of the gross domestic product. The perpetual growth of per capita health care expenditures ($205 in 1965 compared with $5,035 in 2001) reflects increasing medical costs, new technologies, and substantial end-of-life expenditures. There is little sign that such expenditures will decrease, as Medicare enrollee coverage increased from 19.1 million in 1966 to approximately 40.6 million in 2002, a growth of 113% (Centers for Medicare and Medicaid Services, 2006). Currently, 89% of Medicare's annual budget is consumed by those individuals with at least three multiple morbidities (Anderson & Horvath, 2002).
Nevertheless, despite the demographic forecasts of an escalating older population and health care expenditures, we currently lack research that examines common multiple morbidities (van den Akker, Buntix, Metsemakers, Roos, & Knottnerus 1998). The scant literature that does address correlates to multiple morbidity generally reports associations between number of co-occurring morbidity and older age, lower socioeconomic status, and female gender (Hwang et al., 2001; van den Akker et al., 1998; van den Akker, Buntinx, Metsemakers, Roos, & Knottnerus, 2001). Fillenbaum and colleagues (2000) also report a clustering of certain co-occurring conditions, namely hypertension, diabetes, and cardiovascular disease.
One of the challenges confronting multiple-morbidity research concerns measurement. Varying definitions and operationalizations and the absence of a gold standard make it difficult to pinpoint multiple-morbidity rates and risk factors. In a critical review of multiple-morbidity measurement methods, de Groot and colleagues identified 13 differing methods of measurement, including indices of definitively diagnosed conditions (disease counts), indices of the burden of multiple morbidities, and indices based on the severity of the most serious multiple morbidities (de Groot, Beckerman, Lankhorst, & Bouter, 2003). All of these approaches have limitations, such as inconsistent data source or sources (e.g., Medicare claims vs medical record review vs self-reported disease), how to "count" multiple morbidities (number of conditions vs disease severity vs disability vs outcome, including mortality), shortcomings in claims and hospital data (incomplete coding of diagnoses on claims forms, diagnoses caused by hospitalization rather than preexisting conditions), recall bias, time, and cost constraints in collecting multiple-morbidity data in population-based research, and changing international standards and codes from the International Statistical Classification of Diseases (Fleming, Pursley, Newman, Pavlov, & Chen, 2005; Wang & Walker, 2000). To improve these methods requires reaching a consensus as to what specific constellation of chronic conditions constitutes the most significant multiple morbidities, whether and how such a constellation should vary by outcome (mortality, disability, cost), and how to establish an accessible data source capable of providing valid and reliable estimates of multiple morbidity.
Out-of-Pocket Medical Expenditures Among Older Adults
Out-of-pocket health care expenditures are shaped by a range of factors including age, income, gender, marital status, and health insurance coverage in addition to the number, severity, and treatment modalities (e.g., self-care regimes) of illnesses. Results from the 1995 Medicare Current Beneficiary Survey indicated that out-of-pocket expenditures averaged 19% of total respondent income, with this percentage increasing with age (Crystal et al., 2000). A later study found that individuals older than 80 years of age spent five times the amount on out-of-pocket expenses than those in the youngest age categories (Hwang et al., 2001). Crystal and colleagues have suggested that this trend remains intact even when one considers only those individuals older than 65. Individuals between the ages of 75 and 79 had higher out-of-pocket expenses than those aged 65 to 69, whereas the oldest old (85 or older) were found to spend 22% of their income on medication-related costs alone. Income also makes a difference; those older adults in the lowest income quintile have been shown to spend up to 31% of their income on prescription drugs and other medical needs not covered by health insurance (Crystal et al.).
Health Insurance
Research conducted on out-of-pocket expenditures for multiple chronic conditions has found variations in spending according to the type of health insurance (Hwang et al., 2001). Among older adults, mean out-of-pocket spending was lowest for those insured by both Medicare and Medicaid ("dual eligibles"), who had only half the out-of-pocket burden as their Medicare-only cohorts. The combination of Medicare plus a private insurance plan only slightly reduced the out-of-pocket burden in those individuals older than 65, a result that persisted even as the number of multiple morbidities increased.
Medication Expenditures
Eighty percent of the Medicare population uses prescription drugs (Poisal, Murray, Chulis, & Cooper, 1999). As of mid-June 2006, the Centers for Medicare and Medicaid services (CMS) reported that over 38 million Medicare recipients had coverage through Medicare Part D, a former employer, or from other sources deemed to constitute creditable coverage. About 22.5 million people had coverage through a Medicare Part D plan, including stand-alone plans, Medicare Advantage prescription drug plans, or dual-eligible Medicare–Medicaid automatically enrolled beneficiaries. Over 10 million people had coverage from a former employer, including the federal government (FEHB) or military (TRICARE). Finally, about 5.4 million people had other creditable coverage, including coverage through the Veterans Affairs, an employer if the employed person was older than 65 and currently employed, and from those states whose prescription drug plans have not yet been joined with Medicare Part D (CMS, 2006).
Although it is likely that Part D will reduce prescription drug costs and out-of-pocket expenses in general, for whom and to what extent remain unresolved and, in the future, such costs must be reexamined. For example, Sambamoorthi and colleagues (Sambamoorthi, Shea, & Crystal, 2003) found that each chronic condition increased individuals' annual out-of-pocket health care expenditures on prescription drugs by $151. Even such modest increases have been shown to compromise patient ability to adhere to medication recommendations (Piette, Heisler, & Wagner, 2004) and are linked to suboptimal health outcomes (see Fonseca & Clara, 2000; Ku, 2003; Saver, Doescher, Jackson, & Fishman, 2004; Schafheutle, Hassell, Noyce, & Weiss, 2002). Copayment burdens reduce medication adherence (O'Brien 1989; Stuart & Zacker, 1999; Tamblyn et al., 2001), as difficulty affording medications has been shown to increase patient reliance on potentially adverse strategies such as delaying or simply not filling prescriptions, or rationing what medication the patient is able to afford (Atella, Schafheutle, & Hassell, 2005; Klein, Turvey, & Wallace, 2004; Sharkey, Ory, & Brown, 2005).
Summary and Contribution
Despite the unprecedented growth in the older population and the concomitant increase in multiple morbidities, only now are we developing a corpus of research on multiple morbidity, particularly focused on those diseases that affect a majority of older adults. Notably absent in this research is a longitudinal and nationally representative perspective on one of the key challenges to self-management of multiple morbidity: mustering the financial resources to manage multiple morbidities over time. In this article, we provide results from analysis of The Health and Retirement Study (HRS) data on out-of-pocket health care expenditures among older adults with the most common constellations of chronic conditions. With an improved understanding of the impact of chronic conditions on the physical as well as the fiscal health of older populations, programs can be developed to pinpoint problems and improve the overall efficacy of health services.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionLimitationsReferences |
|---|
The HRS provides in-depth information about out-of-pocket expenses for health care services, as well as comprehensive information about disease prevalence and other health information. Moreover, the longitudinal nature of the data set captures changes in the impact of multiple chronic conditions over time. For this study, we included respondents if they (a) were age 65 or older in 1998 and, to ensure a longitudinal perspective, (b) were interviewed in both 1998 and 2002. Thus, our final data set consists of 8,180 respondents or 78.1% of the respondents aged 65 and older (10,472) who participated in 1998.
Measures: Multiple Morbidity and Out-of-Pocket Conceptualization and Measurement
Multiple Morbidity
The most common multiple-morbidity measure is a sum of the conditions present (Holman, Preen, Baynham, Finn, & Semmens, 2005; Rozzini et al., 2002). Although there is a growing body of literature that suggests disease counts or "severity of disease" index summaries are independent predictors of several types of health outcomes (mortality, quality of life, disability), to our knowledge there has been relatively little research undertaken to examine the specific effect of the most prevalent clusters of chronic diseases on these or other outcomes.
Some researchers (e.g., de Groot et al., 2003) have suggested that disease combinations have an additive impact on health outcome, whereas others have found that such combinations had a more synergistic relationship in that their combination leads to poorer outcomes than one would expect by adding or summing the number of diseases in a given subject. Another confounding methodological issue is the wide variation across studies as to what constitutes a chronic disease, which necessarily limits the scope and usefulness of a given study's findings for comparative purposes across similar studies.
Although the current epidemiological literature provides some evidence of which clusters are the most common (Caporali et al., 2005; Chen, Fryer, & Norris, 2005; Fultz, Ofstedal, Herzog, & Wallace, 2003), confounding variables (i.e., the chronological age range of the subjects, which chronic diseases are included in the index, etc.) complicate the reaching of definitive conclusions. Additionally, the methodological approach most often utilized has been to examine the role of comorbidity in relation to a single so-called index disease or disease cluster (see Caporali et al., 2005), resulting in comorbid clusters that are specifically tailored to the particular research being reported, but that are not necessarily reflective of the most common disease combinations in the general population.
The HRS provides an excellent opportunity to examine the most commonly occurring chronic conditions among older adults: high blood pressure, diabetes, cancer, lung disease, heart conditions, stroke, arthritis, and psychiatric problems including emotional or nervous conditions (Wallace & Herzog, 1995). The HRS places a high priority on these chronic conditions because of their public health significance; these diseases are most prevalent among middle-aged and older adults and "account for much of the morbidity and mortality among older persons in western societies" (Fisher, Faul, Weir, & Wallace, 2005).
Because considering all combinations (256 cases) out of the eight chronic conditions is daunting work and does not necessarily capture common patterns of and salient financial issues underlying multiple morbidity, we first identified the most common multiple-morbidity combinations within the data. Consistent with existing literature, these were, in descending order, hypertension and arthritis, hypertension and heart disease, diabetes with one other chronic illness, and the triad of hypertension, heart disease, and arthritis (Bayliss, Ellis, & Steiner, 2005; Caporali et al., 2005; Fultz et al., 2003; Holman et al., 2005; Walke, Gallo, Tinetti, & Fried, 2004).
First, we created a categorical variable measuring multiple morbidity. We identified the number of chronic conditions that older adults have, which ranged from one to five or more. Among those individuals with only one chronic illness, arthritis and high blood pressure were the most frequently occurring. We categorized those individuals with one condition were into three groups: those with arthritis only, those with high blood pressure only, and those with only one condition but not arthritis or hypertension. Next, we identified 28 different combinations of two conditions out of the eight. Among them, we categorized those individuals with two conditions into six groups on the basis of the most commonly occurring clusters: high blood pressure + arthritis, high blood pressure + heart disease, high blood pressure + diabetes, arthritis + heart disease, arthritis + cancer, or "others," representing those clusters with fewer frequencies. For those individuals with three conditions (47 different combinations), we divided them into two groups: those with high blood pressure + arthritis + heart disease and all other triadic multiple-morbidity combinations. Ultimately, we used a categorical variable for multiple morbidity with high blood pressure only as a reference group (we selected high blood pressure because it is the most common morbidity that tends to be associated with other conditions such as diabetes and heart disease): none, arthritis, all other single conditions, high blood pressure + arthritis, high blood pressure + heart disease, high blood pressure + diabetes, arthritis + heart disease, arthritis + cancer, all other combinations of two conditions, high blood pressure + arthritis + heart disease, and all other combinations of three, four, or five conditions.
In addition to examining the effects of changes in these conditions, we created 28 categorical variables to measure changes in multiple-morbidity patterns over the 4-year period from 1998 to 2002. We then estimated the effects of these changes in multiple morbidity with a reference group of those individuals who had no chronic conditions in both 1998 and 2002.
The data revealed two major groups: those with and those without changes between 1998 and 2002. This first group included respondents who reported no chronic conditions in both years or who reported the same number of chronic conditions in both years. Then, we classified those with changes on the basis of the type of change, such as respondents who had no chronic conditions in 1998 but one condition in 2002. Among these respondents, we identified those who, in 2002, had specific chronic conditions such as high blood pressure, heart disease, or arthritis. In order to maintain adequate power to detect differences as Cohen (1988) suggested, we collapsed in an "any others" category those persons with other types of changes and less than a 30-cell size (e.g., cancer). Likewise, we identified other types of changes in multiple morbidity, such as those persons who had no conditions in 1998 but two new conditions (see Tables 3 and 7 for the 28 types of changes).
|
|
For this study, we asked respondents to report out-of-pocket expenses for the past 2 years on (a) hospital and nursing home stays, (b) outpatient services, such as doctor or dental visits and outpatient surgery, (c) home care or other community-based services, and (d) prescription medications. We measured out-of-pocket expenditures by summing the dollar amount of these four categories of out-of-pocket expenses not covered by health insurance over the past 2 years. We measure changes in out-of-pocket expenses (either positive or negative) as the difference of the 2002 out-of-pocket costs less the 1998 expenses.
Additional variables
Differences in health insurance status are one of the main causes of variation in health care utilization and out-of-pocket expenditures (Trevino, Moyer, Valdez, & Stroup-Benham, 1991). Although most elders receive Medicare benefits, significant variation exists in their supplementary health insurance holdings. Thus, we added a categorical variable indicating whether a respondent had additional health insurance, such as (a) Medicare + Medicaid; (b) Medicare + other supplementary plans, including employer-sponsored plans or other supplementary plans (e.g., Medigap); and (c) Medicare only with no additional health insurance. We used this final category as the reference group.
We included the number of adult children who may provide informal care that could assist in medical management and reduce out-of-pocket expenditures. We also included age (years), gender (female), geographic region (rural), education (school years), log of household income, and log of total assets. We converted all dollar figures for out-of-pocket medical expenses, income, and assets in 1998 into 2002 dollars, using the current methods version of the Consumer Price Index for all urban consumers (U.S. Department of Labor, 2005).
Analysis
We used two empirical models to estimate the association and effects of multiple morbidity on out-of-pocket expenses. First, we used a two-part regression model for estimating the association between multiple morbidity and out-of-pocket expenses in 1998 and 2002. Because some respondents (i.e., 11.8% in 1998 or 10.3% in 2002) reported spending nothing on out-of-pocket medical expenses, a regression using the ordinary least squares (OLS) method would introduce biased results. We used the two-part regression model to correct the problem with nonspenders by separating behavior into two stages: decision to spend and then the level of spending (Duan, Manning, Morris, & Newhouse, 1983). For the first part, we used logistic regression to determine the relationship between multiple morbidity and the likelihood of having positive out-of-pocket medical expenses. For the second part, we used OLS linear regression to estimate the association between multiple morbidity and the log amount of out-of-pocket expenses among those individuals with positive expenses. Next, we used the results of the two parts to estimate average out-of-pocket spending across multiple-morbidity status, including those persons with or without spending out-of-pocket expenses. That is, we used this formula based on the research of Duan and colleagues: predicted probability of having any out-of-pocket expenditure (from the results of the first part) x predicted out-of-pocket x predicted residuals (from the results of the second part). To enhance interpretation, we retransformed the log amount of out-of-pocket spending to the dollar amount (Langa et al., 2004; Manning 1998).
Second, we used another estimation model to regress changes in out-of-pocket expenses, 2002 less 1998, on changes in multiple morbidity between 1998 and 2002. Although the first model tells us how out-of-pocket expenses are associated with multiple morbidity in each year, it does not provide information on how multiple morbidity affects the increase in out-of-pocket expenses over the 4-year period. This dynamic model allows us to estimate the effects of multiple morbidity on the increase (or decrease) in out-of-pocket expenses over the 4-year period.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionLimitationsReferences |
|---|
|
|
Table 4 demonstrates unadjusted out-of-pocket medical expenditures. Out-of-pocket expenditures averaged $2,164 in 1998, increasing by 73.2% to $3,748 in 2002. As would be expected, mean out-of-pocket spending grew as the number of conditions increased, ranging from $1,191 (none), $1,716 (one), $2,367 (two), $2,520 (three), $3,428 (four), and $3,103 (five or more) in 1998. The mean out-of-pocket expenses varied across different types of conditions and different illness combinations. In 1998, among those persons with one condition, elders diagnosed with heart, stroke, or lung disease had higher out-of-pocket expenditures than those with cancer, psychiatric problems, or arthritis. By 2002, lung disease, diabetes, and stroke had eclipsed other single morbidities as the most costly. Among those persons with two conditions in 1998, the combination of high blood pressure + diabetes was associated with the highest out-of-pocket expenditures, a pattern repeated in 2002.
|
Table 5 presents the results of the two-part regression analysis for out-of-pocket expenses. The likelihood ratio or F-test results showed that the coefficients of our models were not zero, and they indicated that the models for the first and second parts in 1998 and 2002 were significant at the.01 level. Because the 1998 and 2002 results were very similar except for the magnitude of multiple morbidity and out-of-pocket expenditures, we describe the results of 1998.
|
After we controlled for other factors, we found that there were differences in out-of-pocket expenditure level among the multiple-morbidity combinations. Having heart disease or diabetes in addition to high blood pressure was associated with a 39% higher out-of-pocket spending compared with having high blood pressure alone, while having arthritis in addition to high blood pressure was associated with a 16% increased out-of-pocket spending. Having a multiple-morbidity constellation of high blood pressure, heart disease, and arthritis was associated with a 55% increased out-of-pocket spending compared with having high blood pressure alone, whereas for all others combinations of three conditions, there was 47% more out-of-pocket spending.
Additional multiple morbidities were associated with an increase in out-of-pocket expenses in each year (e.g., in 2002, 47% more with three conditions, 65% more with four, and 90% more with five multiple conditions). Particularly noteworthy is the large increase in out-of-pocket expenditures between 1998 and 2002 for those respondents with five or more illnesses.
Several other variables had an effect on out-of-pocket expenditures. Those persons who were older, female, lived in a rural area, or had higher levels of education and wealth had higher levels of out-of-pocket spending. Elders with only Medicare paid the highest out-of-pocket expenses, followed by those with supplemental health insurance and then individuals with Medicaid.
To ensure that these findings take into account issues such as attrition and nursing home admissions, we conducted two additional analyses to check the sensitivity of our results. First, we used the Chow test to assess the effect of attrition, mainly caused by death, on our results. We conducted an equality test of coefficients of multiple-morbidity variables between a model (Model A) estimated with respondents who, in 1998, were 65 years of age or older (10,472) and another model (Model B) with respondents who had participated in both 1998 and 2002 HRS surveys (8,180). We jointly tested the following hypothesis: All coefficients of multiple-morbidity variables in Model A are equal to those in Model B. The results showed that we cannot reject the hypothesis at the 5% significance level, that is, 
2(13) = 16.51, p > 2 =.2225; this suggests no differences between the two models and, consequently, that attrition effects are not problematic in our study.
Although the HRS is not designed to collect data from respondents who are in nursing facilities, it does include data (from proxy respondents such as adult children) on those who have entered a nursing home since the initial interview. We added two nursing home-related variables, (a) currently living in a nursing home and (b) used a nursing home in the past 2 years, to our existing model. Our analyses indicate that only 112 (1.4%) respondents out of 8,180 were residing in a nursing home and that nursing home use is positively associated with out-of-pocket health care expenditures. However, the magnitude and direction of multiple-morbidity variables are very stable regardless of nursing home use.
Table 6 presents adjusted out-of-pocket expenditures across multiple morbidities based on the results of the two-part regressions. The adjusted out-of-pocket expenditures were similar to the unadjusted ones in their overall pattern, but they differed in their magnitude. In both years, greater number of chronic diseases contributed to higher out-of-pocket expenses. For example, in 1998, an older adult with high blood pressure could expect to spend $1,743 in out-of-pocket expenses; with the additional chronic condition of arthritis, heart disease, or diabetes, the elder would spend an extra $420 (24% more than high blood pressure alone), $526 (30%), or $684 (39%), and with a constellation of high blood pressure, heart disease, and arthritis, the elder would spend an additional $1,209 (69%). Over time, this spending appears even more dramatic. From 1998 to 2002, the sample's annual out-of-pocket spending increased by 20%, 22%, 17%, and 17%, respectively, for elders with two, three, four, and five of the most common disease constellations.
|
It appears that there is a period effect going from 1998 to 2002 in that the out-of-pocket expenditures increased even for those individuals with no changes in multiple morbidity (e.g., $314 for one chronic condition; $1,994 for those with five or more multiple morbidities). Transitioning from no chronic conditions to at least one results in a fairly substantial increase in out-of-pocket expenses, from $1,498 for the development of arthritis, $1,654 for heart disease, and $2,121 for a diagnosis of high blood pressure. Adding a second illness results in a more modest increase (<$350) except in the case of a new cancer diagnosis. The addition of two or more new diseases is particularly burdensome; for example, going from one condition to three or four multiple morbidities is associated with $3,551 more out-of-pocket spending than going from two conditions to four or five multiple morbidities. This latter pattern is associated with $3,586 in additional out-of-pocket spending.
In summary, our findings indicate that, in general, individuals face growing rates of multiple morbidity as they age, that the number of chronic conditions is significantly associated with increases in out-of-pocket health care expenditures, and that among the most common constellations of diseases, out-of-pocket expenses are escalating over time. Results indicate that the percentage of older adults with two or more chronic diseases has increased from 58% to 70% over a 4-year period. Each additional condition confers a significant increase in out-of-pocket expenses. Compare, for example, those elders who in 2002 had no chronic conditions ($2,101 in out-of-pocket spending), one condition ($3,480), two conditions ($4,270), three conditions ($5,345), four conditions ($6,281), and, interestingly breaking the pattern, six conditions ($5,609). Consistent with this spending trend, adjusted out-of-pocket expenditures from 1998 to 2002 increased by 103.9%. While being female was positively associated with greater out-of-pocket spending, supplementary health insurance and Medicaid provided a mitigating influence on these expenses. These findings are also confirmed with dynamic models of change in multiple morbidity. Every increase in multiple morbidities increases out-of-pocket spending, ranging from $287 (no conditions in 1998 to one condition in 2002) to $3,686 (from two conditions in 1998 to four or five conditions in 2002).
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionLimitationsReferences |
|---|
First, health problems often multiply at the same time that income sources traditionally decrease or become stagnant, resulting in increased expenditures on supplementary insurance premiums, copayment burdens, and other out-of-pocket expenses. This study suggests that escalating health care costs coupled with declining health status leaves those elders with the greatest financial and health vulnerability facing increasing demands on their resources. In 2002, about 70% of study participants suffered from multiple morbidities with annually increasing out-of-pocket expenses of 26%. Between 1998 and 2002, then, the average out-of-pocket expenditures for multiple morbidity increased by nearly 104%. This increase far outpaced the inflation rate (10.4%) during that same period (2005). These costs do not include what elders might spend on indirect expenses such as home modification in the face of disabilities. Given that about 40% of those people older than 65 years of age and 60% of people aged 90 or older modified their home to accommodate disabilities (Kutty, 1999), the total burden from multiple morbidity actually may be much higher than out-of-pocket expenses for specific medical needs.
Our findings substantiate the concern that, despite the federal guarantee of health insurance to those aged 65 and older, the health and financial security of older Americans is increasingly vulnerable over time (Wiener & Illston, 2001). Although nearly all Americans older than 65 years of age have medical insurance from Medicare, a common misperception is that such coverage eliminates medical care spending for older adults. In point of fact, in 1996, Medicare covered just over half (56%) of medical expenditures for noninstitutionalized older adults, whereas private insurance covered an additional 20%, out-of-pocket spending comprised 15%, Medicaid covered 4%, and other sources, both private and public, paid for the remaining 5% of medical costs (Cohen et al., 2000).
Although 15% of out-of-pocket spending may not sound prohibitive, out-of-pocket expenses comprise a larger percentage of income for poorer elders. Furthermore, as researchers have demonstrated, not only do lower income elders experience higher disease burdens, particularly common constellations of multiple morbidities (Schoenborn, Vickerie, & Powell-Griner, 2006), but they also arrive at older age without sufficient resources to buffer such expenses (Dannefer, 2003). Because we focused on the most common chronic disease constellations and their financial implications, we have avoided skewed results commonly associated with average expenditure counts, such as when a few people may account for a large portion of out-of-pocket spending (Goldman & Zissimopoulis, 2003).
Another group of individuals who may be particularly vulnerable to financial challenges from out-of-pocket spending consists of those who are not quite poor enough to be eligible for Medicaid but who lack resources to purchase supplemental insurance. Existing research has documented that large out-of-pocket expenditures tend to reduce access to needed health services and force decisions among such essentials as housing, transportation, and food, ultimately resulting in a compromised quality of life and health status (Altman, Cooper, & Cunningham, 1999).This concern is similar to the new Medicare prescription drug plan's "donut hole" problem (after one's prescription medication plan has paid 75% of the amount from $251 through $2,250, the individual will be responsible for all prescription drug costs until the catastrophic benefit level of $5,100). This lack of coverage will disproportionately affect those with several prescriptions who are on small, fixed incomes but who are not indigent.
A second issue that warrants discussion involves the implications of our findings for local service providers and health policy makers. Older adults on fixed incomes who face large increases in out-of-pocket expenditures must choose between disease management and other essential issues such as dietary quality, safe housing, and recreation, perhaps fostering a greater reliance on community-based services. We might expect a swell of demand, for example, on congregate meal services or home heating assistance if people allocate their disposable income for medications or specialist visits rather than food or heating bills. This is particularly salient for those individuals with two, three, and four co-occurring conditions, who endured 41%, 85%, and 100% increases in expenditures, respectively, over the 4 years.
|
Limitations |
|---|
|
TOPAbstractMethodsResultsDiscussionLimitationsReferences |
|---|
Conclusion
Goldman and Zissimopoulis (2003) conclude their article on older adult's out-of-pocket spending by highlighting several "disturbing trends," most notably the erosion of managed care, declining drug coverage, and cuts in retiree benefits. In the 4 years since their article's publication, additional concerns have arisen, including the dismantling of many states' Medicaid programs and the yet unknown, but presumably significant, issue of the new Medicare prescription drug plan. Although many assume that Medicare recipients (and certainly those who are dually eligible) pay few out-of-pocket expenses for their prescription drugs, the ever-increasing cost of medications has led to a popular mandate for a prescription drug plan. Only time will tell whether the new prescription drug plans will adequately address the burdens of multiple morbidity.
New approaches, both financial and medical, have to be implemented and evaluated in order to address this daunting trend of multiple morbidity. Such approaches include a more proactive role for medical management that increases chronic disease coordination among primary care physicians (Starfield et al., 2005). Chronic care models that enhance the active involvement of patients in their own health care and the more efficient functioning of preventive and chronic health care delivery increasingly are recognized as crucial components of managing current health trends. Despite the development of new and better coordinated models of care, we need to continue to fill the knowledge deficit on how older adults manage their multiple morbidities—fiscally and physically. Without such insights, we will be stymied to develop effective interventions (Bayliss et al., 2003).
|
Footnotes |
|---|
1 Department of Behavioral Science, University of Kentucky, Lexington.
2 Department of Family Studies, University of Kentucky, Lexington.
3 Department of Sociology, University of Kentucky, Lexington.
4 Department of Epidemiology, University of Kentucky, Lexington.
Decision Editor: Linda S. Noelker, PhD
Received for publication September 27, 2006. Accepted for publication March 15, 2007.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionLimitationsReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. M. Madden, A. J. Graves, F. Zhang, A. S. Adams, B. A. Briesacher, D. Ross-Degnan, J. H. Gurwitz, M. Pierre-Jacques, D. G. Safran, G. S. Adler, et al. Cost-Related Medication Nonadherence and Spending on Basic Needs Following Implementation of Medicare Part D JAMA, April 23, 2008; 299(16): 1922 - 1928. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
