| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Division of Occupational and Environmental Medicine, 2 Division of Rheumatology, Dept of Medicine, University of California and 3 MGC Data Services, San Francisco, USA
CORRESPONDENCE: E. Yelin, UCSF Box 0920, San Francisco, California, 94143-0920, USA. Fax: 1 4154769030. E-mail: yelin2@itsa.ucsf.edu
Keywords: cost of illness, economics, respiratory conditions
Received: September 16, 2001
Accepted September 19, 2001
This study was supported by Grant NHLBI R01 HL56438.
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionAcknowledgementsReferences |
|---|
The study data were derived from the 1996 Medical Expenditure Panel Survey, a national sample of 21,571 persons. Of the 21,571, 1,027 reported one or more respiratory condition. After weighting, the individuals may represent about 12.1 million persons in the USA. All medical care expenditures of these individuals were tabulated, stratified by comorbidity status, and then compared to those among persons with nonrespiratory conditions or with no conditions. Regression techniques were then used to estimate the increment of healthcare expenditures attributable to the respiratory conditions.
From a national total of $45.3 billion, medical care expenditures averaged $3,753 among persons with respiratory conditions. Hospital stays comprised the largest component (45%). The per capita increment in total expenditures attributable to respiratory conditions ranged from $1,003–2,588, from a national total ranging from $12.1–31.3 billion.
The total medical care expenditure of persons with respiratory conditions was estimated to be $45.3 billion, of which $12.1–31.3 billion represents an increment in expenditures associated with the conditions themselves.
Cost of illness studies are a common method of documenting the impact of medical conditions. In recognition of the increased prevalence, severity, and mortality of chronic respiratory conditions in recent years 1, 2, there have been a large number of studies documenting the costs of specific upper and lower respiratory tract conditions, especially asthma and chronic obstructive pulmonary disease (COPD) 3–28. Although many of the studies in the literature are based on clinical 12, 16, 19, 26 or local 20, 26, population-based data, a few integrate data from several national surveys 4, 27–29. Several studies make national estimates from the 1987 National Medical Expenditures Survey, a national, population-based survey 10, 11, 15. However, none of the studies used a national, population-based survey to estimate medical care expenditures for the entire respiratory condition category, including chronic bronchitis, emphysema, asthma, and COPD, as well as several less prevalent conditions. Moreover, prior national, population-based studies of specific disease entities 4, 10, 11, 15, 27–29 may be outdated because they are based on data that precede much of the growth in the prevalence of asthma and chronic bronchitis, two of the most common respiratory conditions.
The present study was designed to present more contemporary national estimates of medical care expenditures for the entire respiratory condition category. Prior studies 13, 30, 31 predate the development of methods to estimate the increment in expenditures specifically attributable to a condition, methods which can provide more conservative estimates of the economic impact of respiratory conditions.
The specific goals of the present study were to: 1) provide estimates of all medical care expenditures on behalf of persons with chronic respiratory conditions in the USA in 1996; and 2) estimate the increment in expenditures specifically attributable to the respiratory conditions among such persons.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionAcknowledgementsReferences |
|---|
MEPS-H data were collected through six rounds of interviews over a 2.5-yr period. The first three interviews, covering expenditures over an entire year, provided the data used in the analysis for this paper. The interviews were used to collect information on health status, healthcare utilization and expenditures, as well as basic demographic information. The health status section elicited data on the specific medical conditions each respondent self-reported. These were then coded to three-digit levels using the International Classifications of Diseases-ninth revision (ICD-9) system. Such self-reports may not perfectly conform to diagnoses made by physicians. The utilization and expenditure sections elicited information on healthcare episodes since the prior interview. The frequency of interviews was designed to improve the reliability of responses.
In the MEPS, expenditure data derive from a combination of the MEPS-H interviews and information provided by insurance plans. Expenditures in MEPS are defined as the actual expenditures for the medical care services used, regardless of the source of payment 32, 33. In studies on medical care expenditures, the analyst studies the actual exchange of money which contrasts with studies on the costs of illness, in which costs are tabulated even if uncompensated care is provided on the assumption that resources are being consumed regardless of payment. Because MEPS is based on expenditures rather than costs, there are healthcare encounters for which no expenditures are made.
In an entirely fee-for-service system, all expenditures among respondents could be tracked. However, in many forms of managed care, charges are not rendered when services are provided and, hence, there are no expenditures specific to medical care encounters. Accordingly, in such instances, MEPS-H imputes expenditures based on the charges incurred within the fee-for-service sector for similar services provided to similar individuals.
Analyses
Data partitions
In the analyses for this report, estimates of the expenditures of persons with respiratory conditions are presented. The specific conditions included were ICD-9 codes 491 (chronic bronchitis), 492 (emphysema), 493 (asthma), 494 (bronchiectasis), 496 (chronic airway obstructive disease, not elsewhere classified), 500 (coal worker's pneumoconiosis), and 501 (asbestosis). The entire MEPS-H data file was then partitioned into the following condition groups on the basis of ICD-9 codes: persons with only respiratory conditions, persons with both respiratory and nonrespiratory chronic conditions, persons with one nonrespiratory condition, persons with two or more nonrespiratory chronic conditions, and persons with no chronic conditions. Owing to the sample size of the MEPS, some respiratory conditions were not reported by any MEPS respondent. Others, such as COPD, were reported by relatively few respondents, precluding reliable estimates of their national economic impact. For this report, chronic conditions were defined by the protocol devised by Hoffman et al. 34.
General considerations
Because MEPS-H is based on a two-stage cluster sample rather than a true random sample of the noninstitutionalized population, it was necessary to weight the data to make inferences for the USA population. In MEPS-H, the sampling weights also take into account nonresponses in the households targeted for inclusion and omission among respondents, after completion of the first interview 35. Software was used to account for the cluster-sampling design in the calculation of the se of parameters.
Description of utilization and expenditures
Inititially, the sizes of the five condition groups were enumerated (persons with respiratory conditions, persons with and without nonrespiratory conditions, persons with one, or more than one nonrespiratory chronic condition, and persons with no chronic conditions). The frequency with which each major category of healthcare was used by persons in the condition groups was then shown, including ambulatory visits to physicians and nonphysicians, prescription medications, home-health days (days in which health providers assist in daily activities), and hospital admissions. Subsequently, medical care expenditures of persons in the condition groups (and within the respiratory condition group, for those with the most prevalent specific diseases, including chronic bronchitis, emphysema, and asthma) were estimated by category of health services, and the distribution of total healthcare expenditures among persons with all forms of respiratory disease was shown. In the foregoing analysis, all expenditures among persons in the condition groups were tabulated, regardless of whether or not the condition in question accounted for the expenditures. The results indicate those estimates with low statistical reliability (estimates with a relative standard error of >30%).
Analysis of increment in healthcare expenditures
In order to assess the incremental contribution of respiratory conditions to healthcare expenditures, a series of regressions were estimated separately for persons with and without respiratory conditions. Because the two demographic variables used in these regressions (education and marital status) are not applicable to children, data on these characteristics were obtained from the adults in each child's family. In addition, since missing values for any of the independent variables in a regression will cause observations to be deleted, the data was subset only to those observations with values present. This resulted in two and 311 observations being deleted from the respiratory and nonrespiratory categories, respectively. The characteristics of persons with respiratory conditions were then substituted into the regression models developed for those without respiratory conditions. This technique allowed simulation of the level of expenditures that persons with respiratory conditions would experience in the absence of these conditions. The increment was then calculated as the difference between the simulated amount and the predicted expenditures from the respiratory group 36. To make these calculations with respect to ambulatory care, in-patient, and prescription drug expenditures, the two-stage method outlined by Duan et al. 37 was followed. Duan et al. 37 developed this method because many persons have relatively low health expenditures, or none, while a small proportion have very high expenditures, primarily due to hospital admissions. In this method, logistic regression is used to estimate the probability that an individual has any expenditures, followed by ordinary least squares regression to estimate the level of expenditure among those with expenditures.
The incremental contribution of respiratory conditions to total expenditures was estimated by a four-stage model, using separate logistic procedures to predict the probability of any hospital and medical expenditures. Separate ordinary least squares procedures were estimated to predict the level of total costs (including ambulatory and in-patient care, prescription drugs, and a residual category that included services such as home healthcare and medical devices) among persons with and without hospitalizations. In the ordinary least squares regressions, a log transformation was used to account for the skewed distribution of expenditures. In both the logistic and ordinary least squares procedures, the dependent variable was regressed on indicator variables for the presence or absence of respiratory conditions and the following major chronic conditions: hypertension, other forms of heart disease, stroke, other neurological conditions, diabetes, cancer, musculoskeletal conditions, and mental illness.
In addition to the model including only the indicator variables for conditions, a separate model was estimated, which included the condition variables and a count of chronic conditions. In the latter model, the parameter estimates indicated the magnitude of the effect of a condition after taking into account the extent of comorbidity. Models were also estimated which, in addition to the condition indicator variables, controlled for demographic characteristics (age, by categories; sex; White versus non-White race; Hispanic status; marital status; level of formal education), and overall health status (one item measurements of perceived physical and mental well-being 38, 39 singly and combined). The parameter estimate for the respiratory condition variable in the latter models indicated the magnitude of the effect of that condition on expenditures, after taking into account the difference between persons with and without respiratory conditions in demographic characteristics and health status.
The mean expenditures controlling for the covariates described earlier were calculated by exponentiating the predicted values for each observation, multiplying the result by a "smearing" coefficient (the sum of the exponentiated residuals divided by the sample size pooled from the respondents with and without respiratory conditions), and then averaging the observations.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionAcknowledgementsReferences |
|---|
12.1 million persons (4.5% of the population) with at least one respiratory condition. Of these, it was estimated that there were 9.7 million (3.6% of the entire population and 80.2% of all persons with respiratory conditions) with one or more nonrespiratory conditions (table 1
). In addition, it was estimated that there were 163.5 million persons (60.8%) with nonrespiratory chronic conditions in the absence of respiratory conditions and another 93.3 million (34.7%) with no chronic conditions. Among the 12.1 million persons estimated to have one or more respiratory conditions, 10.4 million were estimated to have asthma (individuals, however, could report more than one respiratory condition).
|
tabulates medical care utilization by condition group. A significantly larger proportion of persons with respiratory conditions reported ambulatory physician visits, prescription medications filled, and hospital discharges than persons with nonrespiratory conditions or those with no chronic conditions. In addition, a significantly larger proportion of persons with respiratory conditions reported ambulatory nonphysician visits and home healthcare days than those with no chronic conditions but such persons did not differ from those with nonrespiratory conditions in the proportion using medical care of this type. Among those using each kind of service, persons with respiratory conditions had higher utilization rates of ambulatory physician visits and prescription medications filled than those with nonrespiratory conditions. They also had higher utilization rates of each kind of service, other than home healthcare, than persons with no chronic conditions.
|
). Persons with respiratory and nonrespiratory conditions reported larger average total expenditures than persons with two or more nonrespiratory conditions ($4,465 versus $3,443), but persons who reported only having respiratory conditions had average total expenditures less than two-thirds of those of persons with one nonrespiratory condition ($843 versus $1,321). Among all persons with respiratory conditions, hospital stays accounted for 45% of total expenditures. Other large components included physician visits (18%) and prescriptions (17%). Total expenditures averaged $2,973 among all persons with asthma.
|
0.6% of the Gross Domestic Product for the USA in 1996 40. However, all but $2.0 billion of the expenditures were from persons with both respiratory and nonrespiratory conditions.
Table 4 shows the distribution of medical care expenditures among persons with respiratory and nonrespiratory chronic conditions and among those with no chronic conditions. Among persons who only had respiratory conditions, median annual medical care expenditures were only $189, and even at the 75th percentile, these expenditures only reached $452. Among persons with both respiratory and nonrespiratory chronic conditions, expenditure levels were much higher: median expenditures were $1,308 and expenditures at the 75th percentile were $4,253. Median medical care expenditures among persons with one or two or more nonrespiratory chronic conditions were $184 and $977, respectively. The latter figure is considerably lower than the $1,308 expenditure of persons with respiratory and nonrespiratory conditions.
|
presents the results of several sets of analyses intended to calculate the magnitude of the increment in medical care expenditures directly attributable to respiratory conditions. The results of the first set of analyses present estimates of the magnitude of the increment in ambulatory and in-patient care, and prescription drugs and total expenditures attributable to respiratory conditions, when taking into account the other conditions that the MEPS respondents reported. Respiratory conditions were responsible for annual per capita increments of $239 in ambulatory care expenditures, $358 in prescription drugs, $616 in in-patient expenditures, and $1,583 in overall expenditures (the latter figure is smaller than the sum of the prior three because persons with respiratory conditions have a negative increment in the residual category, indicating lower expenditures).
|
$19.2 billion a year. The $1,583 figure represented more than two-fifths of the average total expenditures of $3,753 among persons with respiratory conditions (latter datum from table 3
). The annual per capita increment in total expenditures actually increased after controlling for the total number of chronic conditions ($1,862) and demographic characteristics ($2,588), but was slightly smaller when controlling for health status ($1,003). After controlling for all of these sets of variables, the increment was $2,579 per capita. When the latter estimate was multiplied by the estimated number of persons with respiratory conditions, the total increment amounted to $31.2 billion. However, even when the smallest estimate of the per capita increment ($1,003) was multiplied by the estimated number of persons with respiratory conditions, the total increment remained a substantial $12.1 billion. |
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionAcknowledgementsReferences |
|---|
In the second set of analyses, the authors estimated the increment in expenditures attributable to respiratory conditions. When adjusting only for the other medical conditions, respondents reported an annual per capita increment of $1,583, or $19.2 billion overall. After adjustment for the other specific chronic conditions reported, the total number of conditions the respondent had, demographic characteristics and health status, the per capita increment was $2,579 and, when summed across all persons with respiratory conditions, amounted to $31.2 billion nationally. Even the smallest estimate of the increment amounted to $12.1 billion on a national basis. With respect to specific components, the estimated increment in in-patient costs ($496–1,473, depending upon the model) indicated higher usage of hospitals among persons with respiratory conditions than would be expected on the basis of their other characteristics.
The present study may have improved upon previous estimates of the national economic impact of respiratory conditions because it combined a systematic, community-based sampling frame with the prospective tracking of expenditures and applied the same methodology to all conditions within the respiratory disease rubric. It indicates that healthcare expenditures on behalf of persons with respiratory conditions have a substantial impact on the nation's economy, and that the increment specifically attributable to these conditions, albeit a smaller amount, nevertheless raises total expenditures among persons with respiratory conditions, substantially higher than the figure expected of such persons in the absence of these conditions. Because the present study tabulated only those medical care expenditures associated with self-reported conditions and omitted indirect costs altogether, it may underestimate the total economic impact of respiratory disease in the nation. The results may be specific to the USA because of the nature of the healthcare system in this nation, with a large proportion of the population without health insurance, and a large proportion of those with insurance in managed care plans. Therefore, despite the methodological advantages of MEPS, including population-based sampling and prospective monitoring of expenditures, results cannot be compared directly with studies from other nations.
The present study's estimates of expenditures for specific conditions would appear to be greater than prior studies. For example, Smith et al. 10 calculated that the direct cost of asthma was $5.1 billion in 1994. In contrast, in the present study it was found that expenditures of persons with asthma amounted to $30.8 billion. Thus, the pandemic of asthma and other respiratory conditions of increasing prevalence and severity may be causing an increase in costs, although methodological improvements implemented in MEPS may also account for part of the increase. Because MEPS is designed to be an ongoing survey, it will be possible to track changes in the expenditures associated with specific conditions over time, to determine whether the increases described here continue or are a one-time artefact of the methodological innovations in MEPS. It should be pointed out, however, that although the estimates of total healthcare costs from MEPS are lower than in other sources, such as the National Health Accounts, once differences in the scope of the expenses and in the populations covered by the two sources are taken into account, the estimates from MEPS are only slightly lower 41.
The data on the distribution of costs previously presented shows that relatively few individuals with respiratory conditions incur high levels of expenditures. Indeed, individuals with the highest 5% of expenditures account for 45% of total expenditures of respiratory conditions. Thus, interventions that can reduce the frequency of high expenditure levels, such as increased utilization of asthma action plans and more effective self-management strategies 42, can alter the national economic impact of respiratory conditions profoundly, by preventing hospital admissions and the use of emergency departments. In the interim, respiratory conditions will continue to represent a substantial drain on the nation's economy.
|
Acknowledgements |
|---|
|
TOPAbstractMethodsResultsDiscussionAcknowledgementsReferences |
|---|
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionAcknowledgementsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. L. Nahin Identifying and pursuing research priorities at the National Center for Complementary and Alternative Medicine FASEB J, August 1, 2005; 19(10): 1209 - 1215. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. H. Howard and J. E. McGowan Jr Initial and Follow-up Costs by Treatment Outcome for Children With Respiratory Infections Pediatrics, May 1, 2004; 113(5): 1352 - 1356. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. Cohen and N. A. Krauss Spending And Service Use Among People With The Fifteen Most Costly Medical Conditions, 1997 Health Aff., March 1, 2003; 22(2): 129 - 138. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
