|
TRENDSTrends In Avoidable Hospitalizations, 1980–1998
A lack of timely and effective ambulatory care can result in a greater number of hospitalizations, especially for certain conditions and among vulnerable groups.1 Communities where people perceived poor access to care had higher rates of hospitalization for certain chronic conditions.2 Preventive care was linked to a reduced probability of avoidable hospitalization for children on Medicaid, and continuity of care with a provider decreased hospitalizations for a Medicaid population of children and adults.3 Also, persons living in counties designated as primary care shortage areas were found to have more avoidable hospitalizations.4 In its 1993 report, Access to Health Care in America, the Institute of Medicine (IOM) recommended that avoidable hospitalizations be used to monitor access to health care services at the national level over time, tracking whether conditions for obtaining care were improving or getting worse, especially for vulnerable population groups.5 However, few trend analyses have been undertaken. One exception was a study of hospitalizations in New York City, which showed that admission rates for avoidable conditions were rising in low-income areas from 1982 through 1993.6 Other research found increasing rates of avoidable hospitalization for uninsured and Medicaid children in selected states from 1990 to 1995.7 Here we undertake the first examination of national trends in avoidable hospitalizations among patients of all ages. We present data for 1980–1998, during which period sweeping changes took place in the health care system. These trends should provide evidence about how these changes have affected access to care for the population as a whole and for various subgroups.
We examined twelve principal or first-listed diagnoses (see Exhibit 5
Data for this study are from the National Hospital Discharge Survey (NHDS), which has been conducted by the National Center for Health Statistics (NCHS) annually since 1965. The data are collected from a sample of inpatient records obtained from a national probability sample of nonfederal, short-stay hospitals (defined as those with an average length-of-stay of fewer than thirty days). Since 1988 all general hospitals and children’s general hospitals have been included in the NHDS sampling frame. Federal, military, and Department of Veterans Affairs (VA) hospitals are excluded, as are institutional hospital units such as prison and college infirmaries. During 1980–1998 an average of 445 hospitals participated in the survey each year; the average response rate was 88 percent. These hospitals provided data on an average of 242,000 discharges per year. Sampled cases were weighted to produce national statistics using multistage estimation procedures. Descriptions of the estimation processes and other aspects of the survey’s design and operation have been published elsewhere.9 Age-specific rates for avoidable hospital conditions were calculated for each year from 1980 through 1998 by sex, race, and region. Annual age-specific rates for each avoidable condition, and percentages of avoidable conditions by expected principal source of payment, were also computed. Standard errors for these statistics were obtained using SUDAAN software.10 Trends were tested using a weighted least squares regression method developed for complex data sets.11 Although the exhibits show data only for selected years, estimates for each year from 1980 through 1998were used in the tests for trends. Differences between rates for selected groups were tested using the two-sided t-test with a critical value of 1.96 (.05 level of significance).
The number of hospitalizations for avoidable conditions increased from 2.2 million in 1980 to 3.7 million in 1998—from 5.9 percent to 11.5 percent of all hospitalizations. The rate per 10,000 population was 99.2 in 1980 but reached 133.8 in 1998 (Exhibit 1  ). In contrast, hospitalizations for other conditions declined during this period. The discharge rate per 10,000 excluding avoidable conditions declined from 1,577.6 in 1980 to 1,031.5 in 1998.
Age and sex. The rate of avoidable hospitalization did not change significantly for patients under age sixty-five as a group (Exhibit 2 ). The rate decreased for children under age 15 but increased for the 15–44 and 45–64 age groups.
Rates rose sharply for the group age 65 and older and for the age groups 65–74, 75–84, and 85 and older. The differences between the 1980 and 1998 rates grew larger with age: The 1998 rate was 34 percent higher than the 1980 rate for the 65–74 group, 51 percent higher for the 75–84 group, and 61 percent higher for the group age 85 and older. Rates of avoidable hospitalization rose for males and females. Within the groups under age sixty-five and age sixty-five and older, rates were not significantly different for males and females in 1980 or 1998.
Race and geography.
Rates of avoidable hospitalization increased for black and white patients in general and for both racial groups age sixty-five and older. In neither 1980 nor 1998 did the rates for the elderly differ significantly by race. In contrast, race differences were seen in the trends for the group under age sixty-five: The rate for white patients decreased, while the rate for black patients rose (Exhibit 3
In all four geographic regions of the country, avoidable hospitalizations rose among the elderly. For patients under age sixty-five, however, the rate increased only in the Northeast.
Source of payment.
Hospitalizations for avoidable conditions increased across payment categories (Exhibit 4
Specific conditions. Rates increased for pneumonia, congestive heart failure, and cellulitis, three of the four most frequently occurring avoidable conditions (Exhibit 5  ).12 Other conditions with increasing rates were ruptured appendix and hypokalemia (potassium deficiency). Hospitalization rates decreased for asthma, perforated or bleeding ulcer, and pyelonephritis (inflammation of the kidney). The hospitalization rate did not change overall for diabetes with ketoacidosis or coma or for malignant hypertension. (Trends for immunizable conditions and gangrene could not be tested because of small sample sizes.) Both the groups under age sixty-five and age sixty-five and older had rising rates of hospitalization for congestive heart failure and cellulitis, decreasing rates for pyelonephritis, and no change in rates for malignant hypertension. The trends for other conditions differed somewhat by age. For example, the hospitalization rate for pneumonia increased for the elderly but did not change for the group under age sixty-five; the asthma rate decreased for the elderly but did not change for those under age sixty-five; and the rate for perforated or bleeding ulcer did not change for the elderly but decreased for the group under age sixty-five. NHDS measures of the severity of illness presented mixed trends. From 1980 through 1998 the rate of hospital deaths per 100 avoidable discharges decreased for the elderly but not for those under age sixty-five. Average lengths-of-stay declined for patients under age sixty-five and age sixty-five and older with avoidable conditions during this period. However, the proportion of patients with avoidable conditions discharged to long-term care facilities increased, as did the average number of diagnoses per avoidable discharge and the proportion of all patients with selected serious comorbidities.
The increase in avoidable hospitalizations seen here is consistent with the 1982–1993 trend reported for New York City.13 The NHDS data show that the rise is a national phenomenon and that it continued into the late 1990s. As in New York City, the national data show increasing disparities, with avoidable hospitalizations growing more rapidly for vulnerable groups. In New York City, disparities were increasing between low-income and higher-income areas for patients under age sixty-five. The national data show a widening gap in avoidable hospitalization rates between black and white populations under age sixty-five and large increases in such rates for those age sixty-five and older. Racial differences. The reasons for the increasing racial disparity in rates of avoidable hospitalization are not clear, although numerous studies have found racial differences in the provision of health care.14 The lack of significant racial differences in rates of avoidable hospitalization for the elderly, who are almost all covered by Medicare, suggests that insurance coverage should be considered as a possible factor in the differences among younger populations. Researchers need to examine whether racial differences in levels of insurance or the adequacy of coverage are creating barriers to ambulatory care that lead to more avoidable hospitalizations. Growing proportions of both the black and white populations enrolled in managed care during the past two decades, but further research needs to explore whether their experiences have been the same. The black population is more likely to be covered by Medicaid, which often has mandatory managed care plans, and concerns have been raised about these plans’ quality of care for minorities.15 Elderly. Avoidable hospitalization trends for the population age sixty-five and older have not been described before. The increases for this group are especially large, indicating that further research should be undertaken to evaluate the factors involved. One area to study is the effects on access to care of out-of-pocket costs, which have grown more rapidly for the elderly than for younger persons.16 It is also important to investigate whether a relationship exists between coverage of prescription drugs and avoidable hospitalizations. Elderly patients who lack drug coverage may use less medication and suffer deterioration or complications that lead to hospitalization. The elderly have enrolled in managed care at a slower rate than the younger population has, suggesting that managed care should be examined as a factor in avoidable hospitalizations. The emphasis of health maintenance organizations (HMOs) on prevention and ambulatory care may have resulted in fewer avoidable hospitalizations for younger patients. In addition, older patients may not have benefited from minimal out-of-pocket payments and from drug coverage associated with HMOs, which could improve access to care. Other factors. Other factors, such as the level of illness in the population and changes in treatment patterns, should be considered to explain the rise in avoidable hospitalizations. The aging of the U.S. population would be expected to raise the level of chronic health problems. Even with good-quality ambulatory care, it would probably be more difficult for increasingly frail elderly patients to avoid hospitalization. This is consistent with the strong positive relationship observed between aging and rising rates of avoidable hospitalization. The health status of the black population is likely to be lower than that of the white population, but that is probably the case both for the group under age sixty-five, which had significant racial differences in rates of avoidable hospitalization, and for the elderly, which did not. Possible changes in physician practice patterns also should be examined. For example, physicians could have lowered the threshold for hospitalization for frail elderly or minority patients, especially when there were complicating conditions. However, one would have expected to see not only more avoidable hospitalizations but also more hospitalizations for other conditions as well, and other hospitalizations declined greatly during the past two decades. The trends for specific avoidable conditions raise questions about whether there have been changes in the severity, incidence, or treatment of some conditions. Like the hospitalization rate, the death rate for pneumonia has been rising for the elderly, suggesting increased difficulty treating this condition.17 Hospitalizations for congestive heart failure may have risen because the decrease in deaths from acute myocardial infarction has resulted in a larger population with chronic heart problems.18 The decline in hospitalizations for perforated or bleeding ulcer for patients under age sixty-five may be related to the adoption of antimicrobial therapy, although the absence of a decline among the elderly needs further investigation. Children. The decrease in avoidable hospitalizations for children may indicate progress in improving children’s health care. It confirms at the national level the decline in avoidable hospitalizations for children seen in New York City. A study of children in nineteen states also documented fewer avoidable hospitalizations for privately insured children, although not for uninsured and Medicaid children.19 At the national level, the decrease was mainly due to a decline in children’s hospitalizations for pneumonia. Vaccination coverage levels have been increasing, which should have resulted in decreases in illnesses such as measles that can lead to pneumonia.20 Data limitations. Because this study was based on hospital discharge data, it was limited to information about hospitalizations rather than individual patients. It was not possible to link inpatient episodes directly with ambulatory care to examine relationships between the two. It also was not possible to obtain information about patients who had limited access to inpatient care as well as ambulatory care. Multiple hospitalizations for avoidable conditions by the same patient cannot be separately identified. The NHDS does not include indicators of socioeconomic status, and race and ethnicity are underreported, which limits the analysis of these important variables. The underreporting of race increased when the survey was redesigned in 1988, and white patients have been found to be underreported to a greater extent than patients of other races.21 However, the racial differences in avoidable hospitalizations for the population under age sixty-five were too large to be explained solely by the underreporting of white patients. Despite its limitations, our study can make an important contribution to monitoring access to health care over time and across social groups. The patterns revealed should help to stimulate debate about the impact of past policy initiatives to control costs and improve health care delivery, which may suggest new approaches to improving health care for underserved groups and the population as a whole.
The authors are health statisticians in the Hospital Care Statistics Branch, National Center for Health Statistics (NCHS), in Hyattsville, Maryland.
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Study Results
Discussion And Policy...
