Search

Background: Hospital benchmarking is essential to quality improvement, but its usefulness depends on the ability of statistical models to adequately control for inter-hospital differences in patient mix. We explored whether the addition of injury-specific clinical variables to the current American College of Surgeons-Trauma Quality Improvement Program (TQIP) algorithm would improve model fit., Methods: We analyzed a prospective registry containing all adult patients who presented to a regional consortium of 14 trauma centers between 2010 and 2011 with severe traumatic brain injury (TBI). We used hierarchical logistic regression and stepwise forward selection to develop two novel risk-adjustment models. We then tested our novel models against the current TQIP model and ranked hospitals by their risk-adjusted mortality rates under each model to determine how model selection affects quality benchmarking., Results: Seven hundred thirty-four patients met inclusion criteria. Stepwise selection resulted in two distinct models: one that added three TBI-specific variables (pupil reactivity, cerebral edema, loss of basal cisterns) to the model specification currently used by TQIP and another that combined two TBI-specific variables (pupil reactivity, cerebral edema) with a three-variable subset of TQIP (age, Abbreviated Injury Scale score for the head region, Glasgow Coma Scale motor score). Both novel models outperformed TQIP. Although rankings remained largely unchanged across model configurations, several hospitals moved across quality terciles., Conclusion: The inclusion of injury-specific variables improves risk adjustment for patients with severe TBI. Trauma Quality Improvement Program should consider replacing several of its general patient characteristics with injury-specific clinical predictors to increase efficiency, reduce the risk of overfitting, and improve the accuracy of hospital benchmarking., Level of Evidence: Prognostic and epidemiological, level II.

Background: Little is known about hospital use of postacute care after surgery and whether it is related to measures of surgical quality., Research Design: We used data merged between a national surgery registry, Medicare inpatient claims, the Area Resource File, and the American Hospital Association Annual Survey (2005-2008). Using bivariate and multivariate analyses, we calculated hospital-level, risk-adjusted rates of postacute care use for both inpatient facilities (IF) and home health care (HHC), and examined the association of these rates with hospital quality measures, including mortality, complications, readmissions, and length of stay., Results: Of 112,620 patients treated at 217 hospitals, 18.6% were discharged to an IF, and 19.9% were discharged with HHC. Even after adjusting for differences in patient and hospital characteristics, hospitals varied widely in their use of both IF (mean, 20.3%; range, 2.7%-39.7%) and HHC (mean, 22.3%; range, 3.1%-57.8%). A hospital's risk-adjusted postoperative mortality rate or complication rate was not significantly associated with its use of postacute care, but higher 30-day readmission rates were associated with higher use of IF (24.1% vs. 21.2%, P=0.03). Hospitals with longer average length of stay used IF less frequently (19.4% vs. 24.4%, P<0.01)., Conclusions: Hospitals vary widely in their use of postacute care. Although hospital use of postacute care was not associated with risk-adjusted complication or mortality rates, hospitals with high readmission rates and shorter lengths of stay used inpatient postacute care more frequently. To reduce variations in care, better criteria are needed to identify which patients benefit most from these services.

Objective: To compare the classification of hospital statistical outlier status as better or worse performance than expected for postoperative complications using Medicare claims versus clinical registry data., Background: Controversy remains as to the most favorable data source for measuring postoperative complications for pay-for-performance and public reporting polices., Methods: Patient-level records (2005-2008) were linked between the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) and Medicare inpatient claims. Hospital statistical outlier status for better or worse performance than expected was assessed using each data source for superficial surgical site infection (SSI), deep/organ-space SSI, any SSI, urinary tract infection, pneumonia, sepsis, deep venous thrombosis, pulmonary embolism, venous thromboembolism, and myocardial infarction by developing hierarchical multivariable logistic regression models. Kappa statistics and correlation coefficients assessed agreement between the data sources., Results: A total of 192 hospitals with 110,987 surgical patients were included. Agreement on hospital rank for complication rates between Medicare claims and ACS-NSQIP was poor-to-moderate (weighted κ: 0.18-0.48). Of hospitals identified as statistical outliers for better or worse performance by Medicare claims, 26% were also identified as outliers by ACS-NSQIP. Of outliers identified by ACS-NSQIP, 16% were also identified as outliers by Medicare claims. Agreement between the data sources on hospital outlier status classification was uniformly poor (weighted κ: -0.02-0.34)., Conclusions: Despite using the same statistical methodology with each data source, classification of hospital outlier status as better or worse performance than expected for postoperative complications differed substantially between ACS-NSQIP and Medicare claims.

Context: Referral to high-volume hospitals has been recommended for operations with a demonstrated volume-outcome relationship. The characteristics of patients who receive care at low-volume hospitals may be different from those of patients who receive care at high-volume hospitals. These differences may limit their ability to access or receive care at a high-volume hospital., Objective: To identify patient characteristics associated with the use of high-volume hospitals, using California's Office of Statewide Health Planning and Development patient discharge database., Design, Setting, and Participants: Retrospective study of Californians receiving the following inpatient operations from 2000 through 2004: elective abdominal aortic aneurysm repair, coronary artery bypass grafting, carotid endarterectomy, esophageal cancer resection, hip fracture repair, lung cancer resection, cardiac valve replacement, coronary angioplasty, pancreatic cancer resection, and total knee replacement., Main Outcome Measures: Patient race/ethnicity and insurance status in high-volume (highest 20% of patients by mean annual volume) and in low-volume (lowest 20%) hospitals., Results: A total of 719,608 patients received 1 of the 10 operations. Overall, nonwhites, Medicaid patients, and uninsured patients were less likely to receive care at high-volume hospitals and more likely to receive care at low-volume hospitals when controlling for other patient-level characteristics. Blacks were significantly (P<.05) less likely than whites to receive care at high-volume hospitals for 6 of the 10 operations (relative risk [RR] range, 0.40-0.72), while Asians and Hispanics were significantly less likely to receive care at high-volume hospitals for 5 (RR range, 0.60-0.91) and 9 (RR range, 0.46-0.88), respectively. Medicaid patients were significantly less likely than Medicare patients to receive care at high-volume hospitals for 7 of the operations (RR range, 0.22-0.66), while uninsured patients were less likely to be treated at high-volume hospitals for 9 (RR range, 0.20-0.81)., Conclusions: There are substantial disparities in the characteristics of patients receiving care at high-volume hospitals. The interest in selective referral to high-volume hospitals should include explicit efforts to identify the patient and system factors required to reduce current inequities regarding their use.

Objective: To determine the appropriateness of use of percutaneous transluminal coronary angioplasty (PTCA) in New York State., Design: Retrospective randomized medical record., Setting: Fifteen randomly selected hospitals in New York State that provide PTCA., Patients: Random sample of 1306 patients undergoing PTCA in New York State in 1990., Main Outcome Measures: Percentage of patients who underwent PTCA for indications rated appropriate, uncertain, and inappropriate., Results: The majority of patients received PTCA for chronic stable angina, unstable angina, and in the post-myocardial infarction period (up to 3 weeks). Fifty-eight percent of PTCAs were rated appropriate; 38%, uncertain; and 4%, inappropriate. The inappropriate rate varied by hospital from 1% to 9% (P = .12); the uncertain rate, from 26% to 50% (P = .02); and the combined inappropriate and uncertain rate, from 29% to 57% (P < .001). There was no difference in appropriateness when the institutions were grouped by volume (fewer than 300 procedures annually or at least 300 procedures annually), location (upstate vs downstate), or by teaching status., Conclusions: Few PTCAs were performed for inappropriate indications in New York State. However, the large number of procedures performed for indications that were rated uncertain as to their net benefit requires further study and justification at both clinical and policy levels.

Objective: To determine the appropriateness of use of coronary artery bypass graft surgery in New York State., Design: Retrospective randomized medical record review., Setting: Fifteen randomly selected hospitals in New York State that provide coronary artery bypass graft surgery., Patients: Random sample of 1338 patients undergoing isolated coronary artery bypass graft surgery in New York State in 1990., Main Outcome Measures: Percentage of patients who had bypass surgery for appropriate, inappropriate, or uncertain indications; operative (30-day) mortality; and complications., Results: Nearly 91% of the bypass operations were rated appropriate; 7%, uncertain; and 2.4%, inappropriate. This low inappropriate rate differs substantially from the 14% rate found in a previous study of patients operated on in 1979, 1980, and 1982. The difference in rates was not due to more lenient criteria but to changes in practice, the most important being that the fraction of patients receiving coronary artery bypass grafts for one- and two-vessel disease fell from 51% to 24%. Individual hospital rates of inappropriateness (0% to 5%) did not vary significantly. Rates of appropriateness also did not vary by hospital location, volume, or teaching status. Operative mortality was 2.0%; 17% of patients suffered a complication. Complication rates varied significantly among hospitals (P < .01) and were higher in downstate hospitals., Conclusions: The rates of inappropriate and uncertain use of coronary artery bypass graft surgery in New York State were very low. Rates of inappropriate use did not vary significantly among hospitals, or according to region, volume of bypass operations performed, or teaching status.

Objective: To determine the appropriateness of use of coronary angiography in New York State., Design: Retrospective randomized medical record review., Setting: Fifteen randomly selected hospitals in New York State that provide coronary angiography., Patients: Random sample of 1335 patients undergoing coronary angiography in New York State in 1990., Main Outcome Measures: Percentage of patients who underwent coronary angiography for appropriate, uncertain, or inappropriate indications., Results: Approximately 76% of coronary angiographies were rated appropriate; 20%, uncertain; and 4%, inappropriate. Inappropriate use did not vary significantly between the elderly (ie, patients aged 65 years and older) and nonelderly, 4.7% and 3.9%, respectively. Although the rate of inappropriate use varied from 0% to 9% among hospitals, the difference was not significant. Rates of appropriateness did not vary by hospital location (upstate vs downstate), volume (fewer than 750 procedures annually or at least 750 procedures annually), teaching status, or whether revascularization was available at the hospital where angiography was performed., Conclusions: Although coronary angiography was used for few inappropriate indications in New York State, many procedures were performed for uncertain indications in which the benefit and risk were approximately equal or unknown.

Objective: To analyze whether important gender differences exist in the quality of hospital care provided to patients with four major medical conditions., Design: Bivariate and multivariate comparisons of clinically detailed sickness at admission, quality, utilization, and outcome measures., Setting: Acute care hospitals located in five states., Patients or Other Participants: A total of 11,242 patients 65 years or older who were hospitalized with one of four diseases: congestive heart failure, acute myocardial infarction, pneumonia, and cerebrovascular accident. We derived our data from the nationally representative sample used to study the quality of hospital care for Medicare patients before and after the implementation of the prospective payment system. A hierarchical (nested) cluster sampling design was used to draw disease-specific samples of patients hospitalized in 1981, 1982, 1985, or 1986 in one of 297 hospitals located in 30 areas within five states., Interventions: This was an observational study., Main Outcome Measures: Sickness at admission, process, use rates, length of stay, discharge status, discharge destination, and mortality., Results: Sex differences in sickness at admission varied by disease. There was some evidence that women received worse process of care, but the difference was very small. We found many similarities in the process and outcomes of care for male and female patients., Conclusions: After controlling for sickness at admission, age, and other important covariates, the in-hospital experiences of elderly men and women showed greater similarities than differences. The concern that sex bias enters into clinical decision making during hospitalization is eased, although not entirely eliminated.

Objective: To compare quality of care measured by explicit criteria, implicit review, and sickness-adjusted outcomes at different types of hospitals., Design: Further analysis of data retrospectively abstracted from medical records to evaluate the effects of prospective payment on quality of care for hospitalized Medicare patients., Setting: Hospitals in five states were sampled to represent the national Medicare admissions along many dimensions., Patients: A total of 14,008 elderly patients with one of the following five diseases: congestive heart failure, acute myocardial infarction, pneumonia, stroke, or hip fracture. These patients were randomly sampled from those with these diseases in 297 hospitals in two time periods, 1981 to 1982 and 1985 to 1986., Outcome Measures: Explicit criteria, implicit review, and mortality within 30 days of admission adjusted for sickness at admission., Results: Quality of care ratings for hospital types are similar using explicit criteria, implicit review, and outcomes adjusted for sickness at admission. Quality differences between types of hospitals were large, with the lowest group estimated to have four percentage points higher mortality than major teaching hospitals in a cohort of patients with average mortality of 16%. Quality varies from state to state, but teaching, larger, and more urban hospitals have better quality in general than nonteaching, small, and rural hospitals. Hospital quality persists over time, but small nonteaching hospitals narrowed the gap with better quality hospitals between 1981 and 1986., Conclusions: The different measures led to consistent and plausible relationships between quality and hospital characteristics. Thus, valid information about hospital quality can be obtained. We need to develop ways to use such information to improve care.

Objective: To determine how well one state's peer review organization (PRO) judged the quality of hospital care compared with an independent, credible judgment of quality of care., Design: Retrospective study comparing a PRO's review, including initial screening, physician review, and final judgments, with an independent "study judgment" based on blinded, structured, implicit reviews of hospital records., Setting: One state's medical and surgical Medicare hospitalizations during 1985 through 1987 audited randomly by the state's PRO., Sample: Stratified random sampling of records: 62 records that passed the PRO initial screening process and were not referred for PRO physician review; 50 records that failed the PRO screen and then were confirmed by PRO physicians to be "quality problems.", Intervention: None., Main Outcome Measure: A study judgment of below standard or standard or above based on the mean of overall ratings by five internists for records in medical diagnosis related groups (DRGs) and by five internists and five surgeons for surgical DRGs. Each step in the PRO review was evaluated for how many records passing or failing that step were judged standard or above or below standard in the study (positive and negative predictive value) and how well that step classified records that the study judged below standard or standard or above (sensitivity and specificity)., Results: An estimated 18% of records reviewed by the PRO were below standard according to the study judgment, compared with 6.3% quality problems according to the PRO's final judgment (difference, 12%; 95% confidence interval, 1 to 23). The PRO's initial screening process failed to detect and refer for PRO physician review two of three records that the study judged below standard. In addition, only one of three of the records that PRO physicians judged to be quality problems were judged below standard by the study judgment. Therefore, the PRO's final quality of care judgment and the study judgment agreed little more than expected by chance, especially about poor quality of care. Although the PRO correctly classified 95% of the records that the study judged standard or above, it detected only 11% of records judged below standard by the study., Conclusions: Most of all, this PRO review process would be improved by additional preliminary screens to identify the 67% of records that the study judged below standard but that passed its initial screening. The screening process also must be more accurate in order to be cost-effective, as it was only slightly better than random sampling at correctly identifying below standard care. More reproducible physician review is also needed and might be accomplished through improved reviewer selection and training, a structured review method, and more physician reviewers per record.

We measured quality of care before and after implementation of the prospective payment system. We developed a structured implicit review form and applied it to a sample of 1366 Medicare patients with congestive heart failure, acute myocardial infarction, pneumonia, cerebrovascular accident, or hip fracture who were hospitalized in 1981-1982 or 1985-1986. Very poor quality of care was associated with increased death rates 30 days after admission (17% with very good care died vs 30% with very poor care). The quality of medical care improved between 1981-1982 and 1985-1986 (from 25% receiving poor or very poor care to 12%), although more patients were judged to have been discharged too soon and in unstable condition (7% vs 4%). Except for discharge planning processes, the quality of hospital care has continued to improve for Medicare patients despite, or because of, the introduction of the prospective payment system with its accompanying professional review organization review.

In 1985, we began a 4-year evaluation of the effects of the diagnosis related groups-based prospective payment system on quality of care for hospitalized Medicare patients. This article provides an overview of the study's background, aims, design, and methods. We used a clinically detailed review of the medical record supplemented by data on postdischarge outcomes drawn from the files of the Health Care Financing Administration and fiscal intermediaries to (1) compare outcomes of care after adjustment for sickness at admission, (2) assess the process of in-hospital care and relationships between processes and outcomes, and (3) assess status at discharge for a nationally representative sample of patients hospitalized before and after prospective payment was implemented.

We developed explicit process criteria and scales for Medicare patients hospitalized with congestive heart failure, myocardial infarction, pneumonia, cerebrovascular accident, and hip fracture. We applied the process scales to a nationally representative sample of 14,012 patients hospitalized before and after the implementation of the diagnosis related group-based prospective payment system. For the four medical diseases, a better process of care resulted in lower mortality rates 30 days after admission. Patients in the upper quartile of process scores had a 30-day mortality rate 5% lower than that of patients in the lower quartile. The process of care improved after the introduction of the prospective payment system; eg, better nursing care after the introduction of the prospective payment system was associated with an expected decrease in 30-day mortality rates in pneumonia patients of 0.8 percentage points, and better physician cognitive performance was associated with an expected decrease in 30-day mortality rates of 0.4 percentage points. Overall, process improvements across all four medical conditions were associated with a 1 percentage point reduction in 30-day mortality rates after the introduction of the prospective payment system.

In this series we have described changes in the quality of care that have occurred in the treatment of hospitalized elderly Medicare patients with one of five conditions between 1981-1982 and 1985-1986. In this article we report on a mortality analysis, patient and hospital subgroup comparisons, and time series studies we have conducted in an attempt to determine whether changes in quality of care can be linked causally to the introduction of the prospective payment system. Based on these analyses we conclude that (1) mortality following hospitalization has been unaffected by the introduction of the prospective payment system, and improvements in in-hospital processes of care that began prior to the prospective payment system have continued after its introduction, but (2) the prospective payment system has increased the likelihood that a patient will be discharged home in an unstable condition. We recommend that efforts to correct this problem be intensified and that clinical monitoring of the impact of the prospective payment system continue as hospital cost-containment pressures intensify.

We developed disease-specific measures of sickness at admission based on medical record data to study mortality of Medicare patients with one of five conditions (congestive heart failure, acute myocardial infarction, cerebrovascular accident, pneumonia, and hip fracture). We collected an average of 73 sickness variables per disease, but our final sickness-at-admission scales use, on average, 19 variables. These scales are publicly available, and explain 25% of the variance in 30-day postadmission mortality for patients with acute myocardial infarction, pneumonia, or cerebrovascular accident. Sickness at admission increased following the introduction of the prospective payment system (PPS). For our five diseases combined, the 30-day mortality to be expected because of sickness at admission was 1.0% higher in the 1985-1986 period than in the 1981-1982 period (16.4% vs 15.4%), and the expected 180-day mortality was 1.6% higher (30.1% vs 28.5%). Studies of the effects of PPS on mortality must take this increase in sickness at admission into account.

We compared patient outcomes before and after the introduction of the diagnosis related groups (DRG)-based prospective payment system (PPS) in a nationally representative sample of 14,012 Medicare patients hospitalized in 1981 through 1982 and 1985 through 1986 with one of five diseases. For the five diseases combined; length of stay dropped 24% and in-hospital mortality declined from 16.1% to 12.6% after the PPS was introduced (P less than .05). Thirty-day mortality adjusted for sickness at admission was 1.1% lower than before (16.5% pre-PPS, 15.4% post-PPS; P less than .05), and 180-day adjusted mortality was essentially unchanged at 29.6% pre-vs 29.0% post-PPS (P less than .05). For patients admitted to the hospital from home, 4% more patients were not discharged home post-PPS than pre-PPS (P less than .05), and an additional 1% of patients had prolonged nursing home stays (P less than .05). The introduction of the PPS was not associated with a worsening of outcome for hospitalized Medicare patients. However, because our post-PPS data are from 1985 and 1986, we recommend that clinical monitoring be maintained to ensure that changes in prospective payment do not negatively affect patient outcome.

We used administrative (Part A Medicare) data to identify a representative sample of 1126 patients with congestive heart failure and 1150 with acute myocardial infarction in hospitals with significant unexpectedly high inpatient, age-sex-race-disease-specific death rates ("targeted") vs all other ("untargeted") hospitals in four states. Although death rates in targeted hospitals were 5.0 to 10.9 higher per 100 admissions than in untargeted hospitals, 56% to 82% of the excess could result from purely random variation. Differences in the quality of the process of care (based on a medical record review) could not explain the remaining statistically significant differences in mortality. Comparing targeted hospitals with subsets of untargeted ones, eg, those with lower than expected death rates, did not affect this conclusion. Severity of illness explained up to 2.8 excess deaths per 100 admissions for patients with myocardial infarction. Identifying hospitals that provide poor-quality care based on administrative data and single-year death rates is unlikely; targeting based on time periods greater than 1 year may be better.

If the quality of care provided by a hospital affects its death rate, then some deaths must be preventable. We have developed a new method to investigate this issue and have reviewed 182 deaths from 12 hospitals (6 high outliers and 6 low outliers for death rate) for three conditions (cerebrovascular accident, pneumonia, or myocardial infarction). The investigators prepared a dictated summary of each patient's hospital course. Then, at least three physicians reviewed each summary and independently judged whether the death could have been prevented. Using a majority rules criterion (at least two of three physicians agreed), we found that 27% of the deaths might have been prevented. Using a unanimity criterion (all three physicians independently agreed), we found a 14% rate of probably preventable deaths. Patients whose deaths were probably preventable were younger (74.7 compared with 78.6 years, P less than 0.05), less often demented (12% compared with 26%, P less than 0.05), and less severely ill (mean Acute Physiology and Chronic Health Evaluation score, 15.6 compared with 21.2; P less than 0.001) than patients whose deaths were nonpreventable. The physicians also listed causes for each probably preventable death; nine reasons encompassed almost all of them. For myocardial infarction, preventable deaths reflected errors in management. For cerebrovascular accident, however, deaths primarily reflected errors in diagnosis. The severity of illness can help a hospital retrospectively identify probably preventable deaths. In the group of patients who died, 42% of those with a low severity of illness had probably preventable deaths as compared to 11% admitted with a high severity of illness. We found that a significant number of hospital deaths might have been prevented. Our findings were based on a new method that needs further testing to substantiate its validity. These findings also need replication before they can be generalized to other hospitals.

Various potential measures of quality of care are being used to differentiate hospitals. Last year, on the basis of diagnostic and demographic data, the Health Care Financing Administration identified hospitals in which the actual death rate differed from the predicted rate. We have developed a similar model. To understand why there are high-outlier hospitals (in which the actual death rate is above the predicted one) and low-outlier hospitals (in which the actual death rate is below the predicted one), we reviewed 378 medical records from 12 outlier hospitals treating patients with one of three conditions: cerebrovascular accident, myocardial infarction, and pneumonia. After adjustment for the severity of illness, the death rate in the high outliers exceeded that predicted from the severity of illness alone by 3 to 10 percent, and in the low outliers, the actual death rate fell short of the severity-adjusted predictions by 10 to 15 percent (P less than 0.01). Reviews of the process of care using 125 criteria revealed no differences between the high and low outliers. However, detailed reviews by physicians of the records of patients who died during hospitalization revealed a higher rate of preventable deaths in the high outliers than in the low outliers. For the three conditions studied, we project that 5.7 percent of a standard cohort of patients admitted to the high-outlier hospitals would have preventable deaths, as compared with 3.2 percent of patients admitted to the low-outlier hospitals (P less than 0.05). A meaningful comparison of hospital death rates requires adjustment for severity of illness. Our findings indicate that high-outlier hospitals care for sicker patients. However, these same hospitals or their medical staffs may also provide poorer care. Our results need confirmation before death-rate models can be used to screen hospitals.

The quiescent interest in understanding the variations in the quality of hospital care has been revitalized recently with the government's release of hospital-specific mortality data. The authors reviewed all published studies that either named hospitals and gave their death rates or explored which of their characteristics explained the differences in their rates. The literature is only in fair condition. It is sparse and flawed, and, before the government's release, did not identify an individual hospital's mortality experience by name. Twenty-two studies were analyzed; only five (23%) met criteria for validity. Seventeen (77%) focused on in-hospital mortality. Data were collected for a short time, were not gathered uniformly, and came from a sample of local hospitals. Researchers identified and explained limitations in their studies and were particularly concerned with their inability to measure differences in patients that might affect death rates. Several characteristics of hospital care were found to be associated with lower inpatient mortality: communication among and commitment of staff, clinical experience and performing operations frequently, board certification, size, and teaching status. The authors urge caution in applying the literature's findings to evaluations of hospital quality and offer suggestions for researchers. Improved research is critically important in facilitating current policy discussions regarding the use of mortality as a measure of hospital quality.

Information about how appropriately procedures are performed is vital to the understanding of the impact of technology and to the success of efforts to channel its use appropriately. While the efficacy of coronary artery bypass surgery has been addressed in several large-scale, randomized trials, there is little information about how appropriately the procedure is actually being used in the community. We determined the appropriateness of coronary artery bypass surgeries performed in three randomly chosen hospitals in a western state. We determined appropriateness by comparing data obtained from a detailed medical record review with a list of 488 indications. This list, developed by a national panel of physicians, covered all possible reasons for performing the procedure. Three hundred eighty-six cases from the years 1979, 1980, and 1982 were examined. Fifty-six percent of the surgeries were performed for appropriate reasons, 30% for equivocal reasons, and 14% for inappropriate reasons. The percentage of appropriate surgeries varied by hospital, from 37% to 78%, but did not vary by patient age. Eliminating the performance of inappropriate procedures may lead to reductions in health care expenditures or to improved patient outcomes.

Increased economic pressure on hospitals has accelerated the need to develop a screening tool for identifying hospitals that potentially provide poor quality care. Based upon data from 93 hospitals and 205,000 admissions, we used a multiple regression model to adjust the hospitals crude death rate. The adjustment process used age, origin of patient from the emergency department or nursing home, and a hospital case mix index based on DRGs (diagnostic related groups). Before adjustment, hospital death rates ranged from 0.3 to 5.8 per 100 admissions. After adjustment, hospital death ratios ranged from 0.36 to 1.36 per 100 (actual death rate divided by predicted death rate). Eleven hospitals (12 per cent) were identified where the actual death rate exceeded the predicted death rate by more than two standard deviations. In nine hospitals (10 per cent), the predicted death rate exceeded the actual death rate by a similar statistical margin. The 11 hospitals with higher than predicted death rates may provide inadequate quality of care or have uniquely ill patient populations. The adjusted death rate model needs to be validated and generalized before it can be used routinely to screen hospitals. However, the remaining large differences in observed versus predicted death rates lead us to believe that important differences in hospital performance may exist.

We examined geographic variation in the rate of inappropriate hospitalization and the effect of cost sharing on that rate. The medical records of 1132 adults hospitalized in a randomized trial of health insurance plans were reviewed by two physicians who were blinded to the patients' insurance plan. They judged 23 percent of the admissions to be inappropriate and an additional 17 percent to have been avoidable by the use of ambulatory surgery. The percentage of inappropriate admissions varied among six sites (from 10 to 35 percent), but areas with low total admission rates did not necessarily have low proportions of inappropriate admissions. In plans with cost sharing for all services, 22 percent of admissions and 34 percent of hospital days were classified as inappropriate, as compared with 24 percent of admissions and 35 percent of hospital days in the plan under which care was free to the patient (these differences were not statistically significant). Our data show that a substantial fraction of hospitalization is potentially avoidable. Because cost sharing did not selectively reduce inappropriate hospitalization, it is important to develop other mechanisms to do so.

Using hospital discharge abstract data for fiscal year 1984 for all acute care hospitals treating Medicare patients (age greater than or equal to 65), we measured four mortality rates: inpatient deaths, deaths within 30 days after discharge, and deaths within two fixed periods following admission (30 days, and the 95th percentile length of stay for each condition). The metric of interest was the probability that a hospital would have as many deaths as it did (taking age, race, and sex into account). Differences among hospitals in inpatient death rates were large and significant (p less than .05) for 22 of 48 specific conditions studied and for all conditions together; among these 22 "high-variation" conditions, medical conditions accounted for far more deaths than did surgical conditions. We compared pairs of conditions in terms of hospital rankings by probability of observed numbers of inpatient deaths; we found relatively low correlations (Spearman correlation coefficients of 0.3 or lower) for most comparisons except between a few surgical conditions. When we compared different pairs of the four death measures on their rankings of hospitals by probabilities of the observed numbers of deaths, the correlations were moderate to high (Spearman correlation coefficients of 0.54 to 0.99). Hospitals with low probabilities of the number of observed deaths were not distributed randomly geographically; a small number of states had significantly more than their share of these hospitals (p less than .01). Information from hospital discharge abstract data is insufficient to determine the extent to which differences in severity of illness or quality of care account for this marked variability, so data on hospital death rates cannot now be used to draw inferences about quality of care. The magnitude of variability in death rates and the geographic clustering of facilities with low probabilities, however, both argue for further study of hospital death rates. These data may prove most useful as a screening mechanism to identify patterns of potentially poor quality of care. Careful choice of the mortality measure used is needed, however, to maximize the probability of identifying those hospitals, and only those hospitals, warranting more in-depth review.