Benchmark admissions identify areas for improvement
Benchmark admissions identify areas for improvement
By Ronald Lagoe, PhD Hospital Executive Council, Syracuse, NY
Although other forms of care continue to evolve, acute hospitals remain the largest single component of the U.S. health care system. During 1995, hospitals accounted for 40% of all health care expenditures in the nation. Hospital utilization is produced by admission or discharge rates and lengths of stay. Admission and discharge rates are especially significant because they identify the relationships between the sizes of populations and the use of acute care by these groups. All additional influences on hospital utilization, such as lengths of stay and patient migration, are derived from population-based admission and discharge rates. For that reason, most acute care expenditures depend on hospital admission rates.
Population
This study compared general hospital admission/discharge rates per 1,000 resident population in metropolitan counties in the United States. The approach was based on statewide hospital discharge data bases that require acute care facilities to submit data by legislative mandate. The importance of those mandated databases cannot be overstated. Only through this approach could resident hospital utilization in communities be identified. The approach employed comparisons with hospital utilization in Syracuse, NY; however, this method could be applied to any metropolitan area for which the study data were available.
Using these resources, the study identified resident hospital discharge rates by age level for metropolitan communities. That approach resulted in a target population including these areas:
o Northeastern states
Albany, NY (Albany County)
Buffalo, NY (Erie County)
Harrisburg, PA (Dauphin County)
Hartford, CT (Hartford County)
New York, NY (Bronx, Kings, New York, Richmond, Queens Counties)
Pittsburgh, PA (Allegheny County)
Providence, RI (Providence County)
Rochester, NY (Monroe County)
Syracuse, NY (Onondaga County)
Worcester, MA (Worcester County)
o Southern states
Baltimore (Baltimore City and County)
Charleston, SC (Charleston, Greenville Counties)
Columbia, SC (Richland County)
Jacksonville, FL (Duval County)
Orlando, FL (Orange County)
Richmond, VA (Richmond, Chesterfield, Hanover, Henrico Counties)
o Midwestern states
Des Moines, IA (Polk, Warren Counties)
Indianapolis (Marion County)
Madison, WI (Dane County)
Milwaukee, WI (Milwaukee County)
Minneapolis, St. Paul (seven-county metro area)
St. Louis (St. Louis City and County)
Wichita, KS (Sedgwick County)
o Mountain states
Albuquerque, NM (Bernalillo County)
Phoenix (Maricopa County)
Salt Lake City (Salt Lake County)
o Pacific Coast states
Portland, OR (Multnomah County)
Sacramento, CA (Sacramento County)
San Diego (San Diego County)
Seattle (King County)
The development of the sample population included almost all major metropolitan areas in the nation for which appropriate data could be collected. Communities located in the following states were not included in the study because they had state legislation for hospital data collection, but the process had not yet been fully implemented, or because they had no state legislation. In some of the latter group, voluntary databases existed that did not include all resident discharges and, therefore, could not be used to characterize hospital use by populations.
Method
The analysis of hospital admission/discharge rates for the study's metro areas focused on nonfederal hospital medical, surgical, pediatric, and obstetric inpatients. To collect data for that population, normal newborns (DRG 391) were excluded because they did not count as admis sions. Psych iatric and alcohol/substance abuse patients (DRGs 424-438) also were excluded because state hospital databases did not generally include those cases treated at private or state facilities.
For this population, resident hospital discharges by metropolitan area were collected for each of the two most recent calendar years. Resident discharge data and population data were collected for the following age levels. These categories have been widely used in health planning because they delineate changes in hospital admission rates:
· 0 - 19 years
· 20 - 44 years
· 45 - 64 years
· 65 - 74 years
· 75 - 84 years
· 85 years and over
· all ages
Analysis of the study focused on calculation of admission/discharge rates for each age level for each community. Consistent with health planning approaches, these rates were calculated per 1,000 resident population.
The discharge rates for each community and year adjusted to the distribution of the Syracuse population also were identified. This statistic was developed by applying admission and discharge rates by age level to the distribution of the Syracuse population for the relevant year.
Results
Using the population definition previously developed, hospital admissions/discharges for each of the metropolitan areas were identified. That information was produced by age level, for all ages combined, and adjusted to the age distribution of the Syracuse population.
Consistent with other research, the data demonstrated that hospital admission rates per capita increased with age. Rates for adults ages 44 to 65 were two or three times those of children and adolescents ages 0 to 19 years. Rates for the frail elderly (those 85 years and over) were four or five times those of adults 44 to 65 years.
To compare the experiences of the metropolitan areas comprising the two year mean adjusted admission rates were ranked from highest to lowest. (See Table 1, p. 86.) The data demonstrate that admission rates were highest in sample communities located in the southern United States. This group accounted for five of the 10 highest rates identified. The data also demonstrated that communities on the Pacific coast produced the lowest admission rates. All four of those communities were ranked among the 10 lowest rates identified.
That information demonstrates that a considerable range of hospital admission rates existed among the metropolitan areas represented in the sample. The mean rate for the two-year period was highest for Baltimore (141.9 per 1,000 population) and lowest for San Diego (81.8 per 1,000 population). Admission rates tended to be higher in communities located in the southern United States, where mean levels for the two-year period ranged from 107.6 to 141.9 per 1,000 population, and lowest on the Pacific coast, where rates ranged from 81.8 to 95.3 per 1,000 population.
The data in Table 1 demonstrated that the admission rate for Syracuse was among the lowest identified. The purpose of the study, however, focused on areas within which this performance could be improved.
Analyze data for benchmarks
The study also identified changes in hospital discharge rates during the most recent two-year period for which data were available. The data demonstrated that hospital admission rates in the sample communities generally declined but did not vary greatly during the study period. Rates declined in 20 of the 30 communities, with the largest reduction (10.6%) occurring in Rochester, NY. Rates increased in the remaining 10 areas. On an age-specific basis, the largest reductions in hospital admissions occurred among non-elderly populations.
The next phase of the study focused on benchmarking of hospital admission rates in communities with the most efficient performance, such as San Diego and Seattle, compared with Syracuse, NY. The initial phase of this project involved major diagnostic categories. (Relevant data are summarized in Table 2, pp. 87-88.)
Those data demonstrated that resident hospital discharge rates per 10,000 population for San Diego and Seattle were similar for almost all of the major diagnostic categories evaluated. The aggregate data in Table 1 had indicated that the aggregate admission rates for those communities differed by 3.1%. Admission rates for individual major diagnostic categories also were relatively small. For example, the rates for neurosurgery varied by about 10%, for respiratory medicine by 3%, and for cardiology (circulatory medicine) by 5.6%. The data identified a few areas with higher variations such as orthopedic surgery, for which the San Diego rate was 20% lower than the Seattle rate, and gynecology surgery, for which the San Diego rate was about one-third lower than the Seattle rate.
That information, combined with the data in Table 1, suggested that San Diego and Seattle successfully managed their hospital admissions to extremely efficient levels. As a result, the levels for major diagnostic categories could be used as benchmarks to evaluate the experiences of other communities. Those admission rates took the form of a benchmark range for each category.
Make meaningful comparisons
The information in Table 2 also describes the evaluation of hospital admissions in Syracuse with respect to the benchmark ranges. In each case, the Syracuse rate was compared with the benchmark range, and the Syracuse experience was identified as higher than, within, or lower than the range. To focus on the actual impact of admission rates on the use of health care, differences between Syracuse and the benchmark range were quantified in absolute numbers rather than as percentages.
The differences between Syracuse and the benchmarks identified in Table 2 suggest that admission rates in Syracuse were slightly higher than those in San Diego and Seattle for most Diagnostic Categories. At the same time, they demonstrate that the largest differences involved respiratory medicine (circulatory medicine), endocrine medicine, circulatory surgery, and neurology (nervous system). The total admissions per 10,000 population by which the Syracuse experience exceeded the midpoints of the San Diego/Seattle experience for those categories (48.8 admissions per 10,000 population) amounted to more than 58% of the difference between the aggregate admission rate for Syracuse and the midpoint for the benchmark communities.
Additional development of admission rate benchmarks and follow-up evaluation involved individual types of surgical procedures and medical diagnoses within major diagnostic categories. This effort focused on areas of care for which the Syracuse experience differed from the benchmark ranges. (Examples of those data are summarized in Table 3, pp. 89-90.)
Comparisons expose potential problems
The data in Table 3 are organized by collections of diagnosis related groups, identified as diagnostic categories. These were developed by aggregating DRGs with the same procedure or diagnosis content but different comorbidity or age ranges. As in the case of the data for major diagnostic categories, San Diego and Seattle produced similar admission rates for the majority of the categories. The rates contributed to relatively narrow benchmark ranges.
Comparison of the Syracuse experience with the benchmark ranges by diagnostic category helped identify the specific sources of a number of the differences in admission rates by major diagnostic category. For example, the data in Table 3 demonstrated that the largest difference between Syracuse and the benchmark range, which involves circulatory medicine, appeared to be generated by higher admission rates in Syracuse for congestive heart failure, acute myocardial infarction treated medically, and syncope.
Analysis leads to follow-up
The data in Table 3 also indicated that the higher admission rate in Syracuse in endocrine medicine was produced by nutritional and miscellaneous and metabolic disorders, and the higher admission rate for circulatory surgery was generated by vascular procedures and pacemaker implants.
Follow-up evaluation of those data will help determine the sources of the variations in admissions. For example, the higher rates for congestive heart failure and nutritional and metabolic disorders may have resulted from readmissions. At the same time, additional potential for ambulatory procedures may exist in other categories.
Subscribe Now for Access
You have reached your article limit for the month. We hope you found our articles both enjoyable and insightful. For information on new subscriptions, product trials, alternative billing arrangements or group and site discounts please call 800-688-2421. We look forward to having you as a long-term member of the Relias Media community.