Research Briefs

Balas, E.A., Stockham, M.G., Mitchell, J.A., and others (1995, October). "The Columbia registry of information and utilization management trials." (AHCPR grant HS07715). Journal of the American Medical Informatics Association 2(5), pp. 307-315.

The authors describe the development of The Columbia Registry of Information and Utilization Management Trials. The Registry located, registered, and abstracted 600 reports from 24 countries on randomized controlled trials that examined management interventions—ranging from patient or physician education to telephone followup, patient or physician reminders, and home care services—their effects on the process and/or outcomes of patient care. Frequently reported effects included hospitalization rate, length of stay, and immunization and mortality rates. The registry has been used already for several meta-analyses, for example, assessing the clinical value of a physician reminder in increasing compliance with cervical cancer screening and tetanus immunizations. Conventional abstracts by investigators provide useful synopses but lack the detail and standardization provided by the Columbia Registry, according to the authors.

Barry, M.J., Fowler, F.J., Chang, Y., and others (1995, September). "The American Urological Association symptom index: Does mode of administration affect its psychometric properties?" (AHCPR grants HS06336 and HS08397). The Journal of Urology 154, pp. 1056-1059.

The American Urological Association (AUA) symptom index is a seven-item questionnaire designed to measure the severity of lower urinary tract symptoms among men with benign prostatic hyperplasia (BPH). It was designed to be self-administered and has been validated as such. However, some men who are visually impaired or illiterate cannot complete the index by themselves. This study compared self-administration and interviewer administration to 41 visually impaired or illiterate men. There was no significant difference between group mean scores. The researchers conclude that, although the AUA symptom index should be self-administered when possible, interviewer administration appears to be acceptable.

Carlson, B.L., Kemper, P., and Murtaugh, C.M. (1995). "Constructing a lifetime nursing home use data base from a sample of discharges." Journal of Economic and Social Measurement 21, pp. 187-211.

This paper, authored by researchers formerly with the Agency for Health Care Policy and Research's Center for Health Insurance and Expenditure Studies, describes the method used to select and re-weight discharge data from the 1985 National Nursing Home Survey so that the data represent a sample of decedents who used nursing homes at some point during their lives. The derived sample greatly expands the amount of information on lifetime nursing home use available to include information on the full distribution of use, its timing, and the source of payment at the beginning and end of episodes. An important part of the information is based on facility records, which are supplemented by information obtained from the decedent's next-of-kin. Analysis of the derived sample showed that 68 percent of decedents who used nursing homes were either discharged dead, or died on the day of discharge, and were represented directly in the discharge sample. Another 25 percent died within a year of discharge and were represented after a minor adjustment to the weights. Only the remaining 7 percent of decedents, who used a nursing home but not in the last year of life, were imperfectly represented in the derived sample. Estimates of lifetime nursing home use from this derived sample were similar to those from the Mortality Followback Survey. However, changes in both surveys could be made to improve the ability to address questions concerning lifetime nursing home use.

Colditz, G.A., Burdick, E., and Mosteller, F. (1995). "Heterogeneity in meta-analysis of data from epidemiologic studies: A commentary." (AHCPR grant HS05936). American Journal of Epidemiology 142(4), pp. 371-382.

Epidemiologic studies that appear to be similar in design can vary greatly in results (heterogeneity). The authors of this paper review the approaches taken to identify, deal with, and interpret heterogeneity in meta-analyses of epidemiologic data and suggest methods that may be used in future studies. They emphasize quantifying and reporting the magnitude of among-study variance and using meta-analysis to describe factors that contribute to variation among study results. They cite two examples showing that analysts sometimes identify heterogeneity and deal with it by excluding studies until a satisfactory degree of homogeneity is achieved. Researchers sometimes exclude 25 percent of the data and still generalize to the total population.

Cooper, J.K. (1995). "Accountability for clinical preventive services." Military Medicine 160(6), pp. 297-299.

In this paper, James K. Cooper, M.D., of AHCPR's Center for Primary Care Research, reviews civilian approaches to accountability for clinical preventive services and suggests a possible approach for military systems. He evaluates four civilian accountability models: DEMPAQ, a program developed for Medicare that collects data both from claims records and from office medical records; PROSPER, which uses patient questionnaires to collect information; HEDIS, which produces "report cards" on health care services and relies on administrative data generally available in managed care organizations; and Passports, a developmental program. None of these civilian models is directly applicable to today's military care system, according to Dr. Cooper. However, adding a preventive services module to the Composite Health Care System, a computer system widely used in military treatment facilities, is one way to provide accountability for clinical preventive services within military health care systems, and probably would lead to higher adherence to recommended standards.

Cooper, J.K. (1995). "Managed care and rural America." The Journal of Family Practice 41(2), pp. 115-117.

In this essay, James K. Cooper, M.D., of the Center for Primary Care Research, Agency for Health Care Policy and Research, discusses trends in rural health maintenance organization (HMO) growth and the forces stimulating this growth: efforts to preserve or increase availability of health care services, expansion of urban managed care organizations (MCOs) to adjacent rural communities, and pressure on States to control Medicaid costs. Dr. Cooper also points out the advantages and disadvantages of physician participation in rural managed care and discusses the legal issues involved in joining or establishing managed care networks in rural areas.

Dikmen, S.S., Machamer, J.E., Donovan, D.M., and others (1995). "Alcohol use before and after traumatic head injury." (AHCPR grants HS04146, HS05304, HS06497). Annals of Emergency Medicine 26(2), pp. 167-176.

University of Washington researchers studied 197 hospitalized adult head-injury survivors at a trauma center and followed them for 1 year. They found that pre-injury alcohol abuse was frequent, with 42 percent of the patients legally intoxicated while in the emergency department (ED). Forty-five percent of the head-injured patients reported two or more problems with alcohol prior to their injury, such as missing work or inability to stop drinking. About 25 percent of patients had been arrested for driving while intoxicated. Alcohol problems had decreased sharply 1 month after injury, but after 1 year, drinking was almost to the level of the year preceding injury. During the immediate post injury period, patients often lack access to alcohol (are in the hospital or living with family), are rethinking their lifestyle, and are advised to stop drinking to minimize the chance of posttrauma seizures. This period may be a natural window of opportunity in which to treat alcohol problems, according to the researchers. In this study, patients who received treatment for alcohol abuse after head trauma decreased the amount they drank per sitting from 6.4 to 3.0 drinks compared with 3.5 to 2.0 drinks for patients who did not receive treatment.

Farrow, D.C., Hunt, W.C., and Samet, J.M. (1995, September). "Biased comparisons of lung cancer survival across geographic areas: Effects of stage bias." (AHCPR grant HS06897). Epidemiology 6, pp. 558-560.

This paper illustrates the phenomenon of stage bias as it affects comparative analyses of lung cancer survival across geographic areas in the United States included in the Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute. The SEER Program collects demographic and clinical data on all incident cancers diagnosed in nine regions of the United States. The researchers examined the stage distributions of lung cancer cases (local, regional, metastatic, and all others), and calculated the proportion of local stage cases receiving surgery for each SEER area. They found that there was marked variability between the SEER areas in both 1- and 3-year survival following the diagnosis of local stage lung cancer. In SEER areas where more local-stage surgery was performed, the surgery provided an opportunity to identify regional and metastatic disease and remove those cases from the local stage group. These findings illustrate that patients within a stage category will not be homogeneous with respect to extent of disease, if staging is accomplished more aggressively (identified by surgery) in one geographic area than in another. As a result, stage-specific survival differences across geographic areas may be artifactual.

Fine, M.J., Hanusa, B.H., Lave, J.R., and others (1995). "Comparison of a disease-specific and a generic severity of illness measure for patients with community-acquired pneumonia." (AHCPR grant HS06468). Journal of General Internal Medicine 10, pp. 359-368.

This study compares the accuracy of the pneumonia severity of illness index (PSI) with a generic measure of severity of illness for identifying illness severity in patients with community-acquired pneumonia. The researchers retrospectively studied adult patients in 78 Medisgroups Comparative Database hospitals. Results showed that 14,199 patients had community-acquired pneumonia, and nearly 11 percent died during hospitalization. Compared with the generic severity measure, the PSI more accurately identified patients at extremely low risk of death (low outliers). Among the 11 low-outlier hospitals, six patients were classified by the generic severity measure alone, two by the PSI alone, and three by both systems. Among the six high-outlier hospitals, one patient was classified by the generic measure alone, three by the PSI alone, and two by both systems. The authors conclude that the PSI more accurately estimated hospital mortality and classified different hospital outliers for mortality than the generic severity of illness measure for patients with community-acquired pneumonia.

Friedman, B., and Elixhauser, A. (1995). "The changing distribution of a major surgical procedure across hospitals: Were supply shifts and disequilibrium important?" Health Economics 4, pp. 301-314.

In this paper, researchers with the Agency for Health Care Policy and Research's Center for Delivery System Research describe and analyze the changes in performance of total hip replacement (THR) among U.S. hospitals from 1980 to 1987. Using data for all hip replacement patients in a large sample of hospitals, the researchers found that hospitals where a large number of hip replacements were performed did not have a higher percentage of older, sicker, and lower income patients. Moreover, there was little evidence that hospitals responded to financial incentives inherent in the Medicare payment system after 1983 to select among hip replacement candidates in favor of those with below-average expected costs. The observed increased concentration over time of hip replacement procedures in hospitals with a high volume of THRs did suggest plausible demand shifts towards hospitals with a priori quality and cost advantages or hospitals that obtained those advantages with a high volume of patients.

Gold, M., and Wooldridge, J. (1995). "Surveying consumer satisfaction to assess managed-care quality: Current practices." (AHCPR contract 282-91-0027). Health Care Financing Review 16(4), pp. 155-173.

A growing number of managed care health plans are using consumer surveys to monitor and improve the quality of care they deliver and to promote informed consumer choice. However, there is no consensus on survey content or approach. This article synthesizes information about consumer satisfaction surveys conducted by managed care plans, government and other agencies, community groups, and purchasers of care. The authors discuss survey content, methods, and use of consumer survey information. The content of instruments appears to be better developed than the methods for using them. The researchers conclude that differences in the use of consumer surveys preclude one instrument or methodology from meeting all needs. They suggest that the effectiveness of plan-based surveys could be enhanced by increased information on alternative survey instruments and methods and new methodological studies.

Greene, V.L., Lovely, M.E., Miller, M.D., and Ondrich, J.I. (1995). "Reducing nursing home use through community long-term care: An optimization analysis." (AHCPR grant HS06757). Journal of Gerontology: Social Sciences 50B(4), pp. S259-S268.

The authors explore the capability of community long-term care (CLTC) services to reduce nursing home use when services are allocated strategically for this purpose. Using an actual CLTC clientele—the population of persons screened into the National Long-Term Care Channeling Demonstration--they noted the existing use of CLTC services and costs and nursing home use. The authors then simulated reallocation of the existing budget so that the CLTC service packages received by each individual minimized long-term nursing home use. A comparison of the two scenarios revealed that use of CLTC services could significantly reduce nursing home use without increasing total community expenditures, according to the authors.

Hennessy, S., Strom, B.L., Berlin, J.A., and Brennan, P.J. (1995). "Predicting cutaneous hypersensitivity reactions to cotrimoxazole in HIV-infected individuals receiving primary Pneumocystis carinii pneumonia prophylaxis." (NRSA Fellowship F32 HS00066). Journal of General Internal Medicine 10, pp. 380-386.

Cotrimoxazole is currently advocated by the U.S. Public Health Service as the preferred agent for Pneumocystis carinii pneumonia (PCP) prophylaxis for individuals infected with the human immunodeficiency virus (HIV), whose CD4 cell counts are below 200 or who experience symptoms. However, many HIV-infected patients are unable to tolerate the drug, most frequently because they develop a cutaneous hypersensitivity reaction that is characterized by rash, fever, and itching. In this study, the researchers measured the incidence of this reaction among 236 HIV-infected patients receiving cotrimoxazole for primary PCP prophylaxis. They found that 20 percent developed cutaneous hypersensitivity reactions, with 12.5 percent of these being severe, that is, resulting in hospital admission or systemic treatment with a corticosteroid. Being male, having a higher CD4 percentage, a history of syphilis, and having higher total protein have at least borderline associations with these reactions. However, clinical and laboratory variables do not appear sufficiently associated with the reactions to permit a clinically useful rule that would predict the reaction in these patients.

Kaluzny, A.D., Konrad, T.R., and McLaughlin, C.P. (1995). "Organizational strategies for implementing clinical guidelines." (AHCPR grant HS07286). Journal on Quality Improvement 21(7), pp. 347-351.

CQI (continuous quality improvement, also known as total quality management) and academic detailing represent significant organizational strategies for facilitating the implementation of clinical guidelines. CQI requires a systematic examination of the internal operations of the organization and focuses on identifying and implementing improvement in overall performance and is best suited for handling interactions within, not across, organizations. Academic detailing focuses on physician behavioral change, including the use of market research, to develop an understanding of motivational patterns of physicians' use, sociometric identification of key decisionmakers, and use of basic learning reinforcement techniques. The authors conclude that, without a managerial commitment and organizational strategy, guidelines may be perceived as a threat to the autonomy of both the clinician and the organization.

Morise, A.P., Diamond, G.A., Detrano, R., and others (1995). "Incremental value of exercise electrocardiography and thallium-201 testing in men and women for the presence and extent of coronary artery disease." (AHCPR grant HS06065). American Heart Journal 130, pp. 267-276.

This study demonstrates that there is incremental value in exercise testing for assessment of both the presence and extent of coronary artery disease in men and women. The researchers developed logistic algorithms and evaluated them in a separate set of 865 patients from four centers. They included pretest risk variables (age, sex, symptoms, diabetes, smoking, and cholesterol concentration); exercise electrocardiogram (ECG); and thallium-201 scintigram (to detect defect presence, reversibility, and intensity of hypoperfusion). They assessed the accuracy and incremental value by receiver operating characteristic (ROC) curve analysis. Incremental ROC curve areas for disease presence were pretest risk factors 0.75, post-exercise ECG 0.82, and post-thallium scintigram 0.85 and for disease extent were pretest risk factors 0.71, post-exercise ECG 0.76, and post-thallium scintigram 0.78.

Patrick, D.L., Deyo, R.A., Atlas, S.J., and others (1995). "Assessing health-related quality of life in patients with sciatica." (Back Pain PORT grant HS06344). Spine 20(17), pp. 1899-1909.

This study analyzes health-related quality-of-life (HRQOL) measures and other clinical and questionnaire data obtained from the Maine Lumbar Spine Study, a prospective study of 427 persons with low back problems. The SF-35 bodily pain questionnaire and the modified Roland measure demonstrated the greatest amount of change over the 3-month followup time and were the most highly associated with self-rated improvement in HRQOL. A high correlation between clinical findings or symptoms and the modified Roland measure, SF-36, and disability days, indicated a high degree of construct validity. The researchers conclude that these measures performed well in measuring the HRQOL of patients with sciatica.

Schwartz, M., Klimberg, R.K., Karp, M., and others (1995). "An integer programming model to limit hospital selection in studies with repeated sampling." (AHCPR grant HS06048). HSR: Health Services Research 30(2), pp. 359-376.

The authors describe an integer programming model that could be used by researchers to select a limited number of hospitals for medical record review when repeated sampling is required. The authors illustrate the model in the context of two studies, which share these common characteristics: hospitals are classified into categories, for example, high, medium, and low volume; the classification process is repeated several times, for example, for different medical conditions; medical records are selected separately for each iteration of the classification; and for budgetary and logistical reasons, reviews must be concentrated in a relatively small subset of hospitals. The researchers found the integer programming model to be useful for selecting a subset of hospitals at which more intensive reviews would be conducted. They caution, however, that limiting the number of hospitals at which records are reviewed may compromise the independence of the multiple analyses performed, since it ignores any overall "hospital effect."

Selby-Harrington, M.L., Donat, P.L.N., Quade, D., and Farel, A.M. (1995). "Facilitating random assignment in a community health education project." (AHCPR grant HS06507). Health Values 19(4), pp. 3-9.

The random assignment of individuals to intervention or control groups is a common practice in clinical research. However, barriers associated with the cost and complexity of random assignment often preclude its use in evaluating community health projects. This paper describes procedures that were developed to facilitate the use of random assignment in a community health project that tested educational interventions to encourage parents to obtain well-child care through the Medicaid program. A state-level Medicaid database was supplemented by staff members from social service offices and the research project to implement the procedures. A critical key to the specific procedures used in this project was the ability to generate a list of the target population and to arrange the list in random order.

Stump, T.E., Dexter, P.R., Tierney, W.M., and Wolinsky, F.D. (1995). "Measuring patient satisfaction with physicians among older and diseased adults in a primary care municipal outpatient setting." (AHCPR grant HS07632). Medical Care 33(9), pp. 958-972.

These authors examine the reliability and validity of three instruments reported to be suitable for measuring patient satisfaction with physicians among older and diseased adults in a primary care municipal outpatient setting. The first two of these instruments take a global approach to patient satisfaction: the physician-behavior subscales of the Physician Satisfaction Questionnaire (PSQ), and the American Board of Internal Medicine's (ABIM) questionnaire, which evaluates the relationship between patients and physicians in internal medicine residencies. In contrast, the third instrument focuses on satisfaction with the visit just made and consists of the nine items used in the Medical Outcomes Study (MOS) visit-specific questionnaire. The two general measures of patient satisfaction were found to be highly correlated with the visit-specific measure. In contrast, the ABIM was found to be shorter and to have a simpler and more pristine factor structure.

Tierney, W.M., Overhage, J.M., Takesue, B.Y., and others. (1995). "Computerizing guidelines to improve care and patient outcomes: The example of heart failure." Journal of the American Medical Informatics Association 2(5), pp. 316-322.

The researchers attempted to incorporate the AHCPR-sponsored guideline, Heart Failure: Evaluation and Care of Patients with Left-Ventricular Systolic Dysfunction, into a network of microcomputer workstations at an urban teaching hospital to facilitate use by physicians. The guideline was programmed into the workstation software, using an automated version of the CARE programming language. For all of the subsequent workstation studies, the guideline suggested orders for drug therapy, non-drug therapy, and diagnostic testing (with accompanying explanatory text) that the physician could accept with a single keystroke or mouse click. According to the authors, making the guideline useful via computer proved difficult for several reasons: one, the guideline frequently hinges on data that are not routinely stored in most electronic record systems in a useful format; two, heart failure usually is not an isolated phenomenon, and dealing with a patient's coexisting conditions often is the most difficult aspect of their care; three, the guideline does not cover many of the coexisting conditions or errors of commission (e.g., the inappropriate use of common drugs).

Tu, S.W., Eriksson, H., Gennari, J.H., and others (1995). "Ontology-based configuration of problem-solving methods and generation of knowledge-acquisition tools: Application of PROTEGE-II to protocol-based decision support." (AHCPR grant HS06330). Artificial Intelligence in Medicine 7, pp. 257-289.

PROTEGE-II is a suite of tools and a method for building knowledge-based systems and domain-specific knowledge-acquisition tools. In this paper, the authors demonstrate how PROTEGE-II can be applied to the task of providing protocol-based decision support for treating HIV-infected patients. The general goal of the PROTEGE-II approach is to produce systems and components that are reusable and easily maintained. While conceding that their evaluation of the PROTEGE-II system is still preliminary, the authors show that the goals of reusability and easy maintenance can be achieved. They discuss design decisions and the tradeoffs that have to be made in the development of the system.

Yeaton, W.H., Langenbrunner, J.C., Smyth, J.M., and Wortman, P.M. (1995, September). "Exploratory research synthesis: Methodological considerations for addressing limitations in data quality." (AHCPR grant HS06264). Evaluation & The Health Professions 18(3), pp. 283-303.

Exploratory meta-analysis or research synthesis has been advocated as a way to develop important hypotheses for further study. These investigators conducted an exploratory research synthesis on the carotid endarterectomy (CE) literature to illustrate this method. The CE literature is similar to that of many other new medical interventions because it contains numerous limitations to data quality. Exploratory research synthesis of such literature necessitates a number of methodological and statistical considerations to address these limitations, including the problems of missing data, appropriate unit of analysis, nonnormal distribution of outcomes, and lack of controlled studies. The authors discuss strengths and limitations of the exploratory research synthesis approach within the context of public policy decisions for assessing medical technologies.

AHCPR Publication No. 96-0032
Current as of November/December 1995