Andrew J. Hogan

Measuring hospital performance. The development and validation of risk-adjusted indexes of mortality, readmissions, and complications.

In this study we used information from discharge abstracts to develop three different risk-adjusted measures of hospital performance: a Risk-Adjusted Mortality Index, a Risk-Adjusted Readmissions Index, and a Risk-Adjusted Complications Index. The adjustments have face validity, and appear to account for much of the variation across hospitals in the rates of these adverse events. The indexes are stable over time, and are not biased with respect to hospital size, ownership, or teaching status. All three indexes appear to have construct validity when tested against the changes in hospital care that occurred when PPS was introduced.

Cost-effectiveness analysis in heart disease, part III: Ischemia, congestive heart failure, and arrhythmias

Cost-effectiveness analyses were reviewed in the following diagnostic and treatment categories: acute myocardial infarction (MI) and diagnostic strategies for coronary artery disease (CAD), coronary artery bypass graft (CABG) surgery, percutaneous transluminal coronary angioplasty (PTCA), congestive heart failure (CHF), and arrhythmias. In the case of acute MI, coronary care units, as presently used, are rather expensive but could be made much more efficient with more effective triage and resource utilization; reperfusion via thrombolysis is cost-effective, as are beta-blockers and angiotensin-converting enzyme (ACE) inhibitors post-MI in appropriate patients. Cost-effectiveness of CAD screening tests depends strongly on the prevalence of disease in the population studied. Cost-effectiveness of CABG surgery depends on targeting; eg, it is highly effective for such conditions as left-main and three-vessel disease but not for lesser disease. PTCA appears to be cost-effective in situations where there is clinical consensus for its use, eg, severe ischemia and one-vessel disease, but requires further analysis based on randomized data; coronary stents also appear to be cost-effective. In preliminary analysis, ACE inhibition for CHF dominates, ie, saves both money and lives. Cardiac transplant appears to be cost-effective but requires further study. For arrhythmias, implantable cardioverter defibrillators are cost-effective, especially the transvenous device, in life-threatening situations; radiofrequency ablation is also cost-effective in patients with Wolff-Parkinson-White syndrome apart from asymptomatic individuals; and pacemakers have not been analyzed except in the case of biofascicular block, where results were variable depending on the situation and preceding tests.

Cost-Effectiveness Analysis in Heart Disease, Part II: Preventive Therapies

Cost-effectiveness analysis of preventive therapies are reviewed in the following categories: lipid lowering, hypertension, smoking cessation, exercise, and anticoagulation. From review of 8 analyses, cost-effectiveness of primary prevention via cholesterol lowering drugs is generally expensive, whereas that of secondary prevention generally is favorable. However, targeting by age, coexisting risk factors, and gender strongly influence results that are also sensitive to drug costs. Treatment of hypertension (5 analyses) is cost-effective in virtually all patient populations and circumstances and for a wide variety of drugs. It is more so with coexisting risk. Issues relating to compliance and drug costs are important. Smoking cessation (4 analyses) is highly cost-effective and worthwhile. However, data on recidivism are incomplete, and cessation may be more difficult to achieve in the general population versus study patients. In one analysis, an exercise program was found to be cost-effective in prevention of coronary heart disease. Anticoagulants have been analyzed in various circumstances. Their cost-effectiveness is favorable for prosthetic valves, although sensitive to imprecision in monitoring. It is also favorable for mitral stenosis in the presence of atrial fibrillation but not normal sinus rhythm. Cost-effectiveness of heparinization for prosthetic valve patients undergoing surgery is rather variable and depends on type of surgery (major versus minor) and type of valve. Many topics in anticoagulant therapy remain to be explored from a cost-effectiveness point of view.

Evaluating managerial efficiency of Veterans Administration medical centers using Data Envelopment Analysis.

This study applied the methodology of Data Envelopment Analysis (DEA) to the set of VA medical centers to evaluate their relative managerial efficiencies. Each VAMC was viewed as a producer of multiple outputs and a consumer of multiple inputs. DEA uses linear programming to identify resources that were underutilized and services that were inefficiently produced. Managerial strategies based on the dual variables were constructed to indicate the manner in which inefficient VAMCs may be made efficient. The analysis showed that relative inefficiency existed in about one third of the VAMCs nationwide. Elimination of this inefficiency would save the VA over

Confidence Intervals for Cost - Effectiveness Ratios

300 million annually on personnel, equipment, drugs, and supplies, without reducing the level of services provided. A subsequent analysis of co-variance revealed that VAMCs affiliated with a university were generally less efficient than those without such an affiliation. A similar finding was obtained for larger VAMCs relative to smaller medical centers. In neither case, however, should these results be construed to imply that VAMCs should terminate their university affiliations or that VAMCs should be made smaller since factors other than relative efficiency are clearly as or more important in such decisions.

Changes in rates of unscheduled hospital readmissions and changes in efficiency following the introduction of the Medicare prospective payment system. An analysis using risk-adjusted data.

The problem of variability in computed cost-effectiveness ratios (CERs) is usually addressed by performing sensitivity analyses to determine the effects on these ratios of plausible ranges of values of input parameters. However, the sampling variation that exists in these estimated parameters can be utilized to obtain confidence intervals for cost-effectiveness ratios. As cost-effectiveness analysis becomes more widely used, new techniques need to be de veloped for establishing when a difference in strategies evaluated is meaningful. A first step is to establish the precision of the CER itself. The authors estimate the precision of a CER in the context of a statistical model in which the primary outcome is survival, with cost and effectiveness defined in terms of the underlying survival distribution (S). Effectiveness (α) is measured by life expectancy, restricted to a finite time horizon and discounted at a fixed rate r, α = √ e-rtS(t)dt. Cumulative cost (β) per patient is regarded as resource utilization and incurred randomly over time depending on the survival experience of the patient, p = √ e - rtS(t)dC(t), where C(t) is the total potential resources utilized up to time t. Average cost effectiveness (ACE) of a single strategy is β/α, and when comparing two strategies, the CER is Δβ/Δα, the ratio of the incremental cost to the difference in mean survival. Utilizing the sampling distribution of the Kaplan-Meier estimate of S yields standard errors and confidence intervals for ACE and CER. The technique is applied to survival data from 218 previously studied patients to assess 95% confidence intervals for the CER and ACE of the implantable cardioverter defibrillator as compared with electrophysiology-guided therapy. Key words: cost-effectiveness analysis; sensitivity analysis; Kaplan-Meier estimate; av erage cost-effectiveness; cardioverter defibrillator; survival data. (Med Decis Making 1995;15:254-263)

Activities of daily living as quantitative indicators of nursing effort.

The purpose of this study was to analyze changes in rates of unscheduled readmissions and changes in technical efficiency following the introduction of the Medicare Prospective Payment System (PPS). We developed the RiskAdjusted Readmissions Index (RARI), which allowed us to make comparisons in rates of unanticipated readmissions across hospitals and over time. Data envelopment analysis (DEA), a linear programming technique, was used to measure changes in technical efficiency by comparing the inputs used and the outputs produced across a cohort of hospitals, while adjusting for changes over time in case mix and case complexity. Rates of unscheduled readmissions and efficiency scores were computed for a sample of 245 hospitals for each year. Although both readmission rates and efficiency scores increased for most hospitals, there was no evidence that those hospitals that experienced the greatest increases in efficiency had the largest increases in their rates of unscheduled readmissions.

Measuring the costs of primary care education in the ambulatory setting.

Functional assessments of elderly or disabled people requiring long-term care have been used by clinicians for many years, and functional assessment instruments are now being used as indicators of required nursing care and its cost. The authors examine the ability of functional assessment items and instruments to measure accurately the variation in nursing care used by nursing home patients, with analysis of 290 patients. Nursing times, measured for each patient by nurse category (registered and all other) and type of care (skilled and personal) measure resource consumption. Activities of daily living (ADLs): eating, bathing, dressing, toileting, transferring, and continence are used to measure functional abilities on a four-point scale: independent, supervised, assisted or helped, and dependent, as well as two derived scales: Katzs Index and Resource Utilization Groups. The four-point measurement scales for ADLs are found to be necessary as indicators of nursing time required by patients. As a consequence, the three-point scales used for ADLs in the Long-term Care Minimum Data Set are not adequate, at least in nursing homes, for resource allocation. The relationship of nursing times with individual ADLs is nonlinear, so linear statistical techniques such as principal components, canonical correlations, or linear regression are inappropriate to produce patient classification systems based on ADLs. Individual ADLs do not explain use of registered nursing care time as well as they do care time by other nursing staff. Therefore, resource allocation and staffing for registered nurses must be done separately from other nursing personnel, using indicators other than ADLs.

Understanding the costs of ambulatory care training.

In 1995, the authors obtained cost, operations, and educational activity data from 98 ambulatory care sites across the United States in which primary care teaching was occurring and compared those data with the corresponding data from 84 ambulatory care sites where no teaching was going on. The teaching sites in the sample were found to have 24–36% higher operating costs than the non-teaching sites. This overall difference in costs is approximately the same difference in costs earlier estimated for university teaching hospitals compared with non-teaching hospitals. These costs are shared by all involved in the ambulatory education process: sponsors, sites, and faculty. In a related finding, the authors discovered that 30–50% of all ambulatory care sites thought not to be involved in education are in fact teaching at a high level of involvement. Further research into not only the costs but the value of education in the clinical setting is encouraged. The authors also hope that the publication of this report will encourage accrediting bodies and professional organizations to improve the information available about ambulatory care training in general.

Evaluation of the Effectiveness of Following Up Laboratory Reports of Elevated Blood Leads in Adults

While patient care has been shifting to the ambulatory setting, the education of health care professionals has remained essentially hospital-based. One factor discouraging the movement of training into community-based ambulatory settings is the lack of understanding of what the costs of such training are and how these costs might be offset. The authors describe a model for ambulatory care training that makes it easier to generalize about to quantify its educational costs. Since ambulatory care training does not exist in a vacuum separate from inpatient education, the model is compatible with the way hospital-based education costs are derived. Thus, the models elements can be integrated with comparable hospital-based training cost elements in a straightforward way to allow a total-costing approach. The model is built around two major sets of variables affecting cost. The first comprises three types of costs--direct, indirect, and infrastructure--and the second consists of factors related to the training site and factors related to the educational activities of the training. The model is constructed to show the various major ways these two sets of variables can influence training costs. With direct Medicare funding for some ambulatory-setting-based education pending, and with other regulatory and market dynamics already in play, it is important that educators, managers, and policymakers understand how costs, the characteristics of the training, and the characteristics of the setting interact. This model should assist them. Without generalizable cost estimates, realistic reimbursement policies and financial incentives cannot be formulated, either in the broad public policy context or in simple direct negotiations between sites and sponsors.