Showing papers in "Medical Care in 1986"

Concurrent and predictive validity of a self reported measure of medication adherence

Abstract: Adherence to the medical regimen continues to rank as a major clinical problem in the management of patients with essential hypertension, as in other conditions treated with drugs and life-style modification. This article reviews the psychometric properties and tests the concurrent and predictive va

Use of medical care in the Rand Health Insurance Experiment. Diagnosis- and service-specific analyses in a randomized controlled trial.

Abstract: * From The Behavioral Sciences Department, The Rand Corporation, Washington, DC. t From The System Sciences Department, The Rand Corporation, Santa Monica, California. t From the Departments of Medicine and Public Health, UCLA Health Sciences Center, Los Angeles, California. ? From the Health Policy Institute, Boston University, Boston, Massachusetts. "From The Economics and Statistics Department, The Rand Corporation, Santa Monica, California. # From The Computer Sciences Department, The Rand Corporation, Santa Monica, California. Supported by the Health Insurance Study Grant 016B80 from the U.S. Department of Health and Human Services, Washington, DC. The views expressed are those of the authors and do not necessarily represent those of the United States Government or The Rand Corporation. Address correspondence to: Kathleen N. Lohr, The Rand Corporation, 2100 M Street, NW, Washington, DC 20037.

Delayed feedback of physician performance versus immediate reminders to perform preventive care. Effects on physician compliance.

Abstract: In an academic general medicine clinic, we performed a randomized, controlled trial to compare (1) the effects of supplying monthly feedback reports of compliance with preventive care protocols by 135 internal medicine house staff with (2) the effects of specific reminders given to them at the time

Telephone surveys in public health research.

Abstract: The last 10 years have seen increasing use of telephone surveys in public health research. This paper reviews issues of sampling, data quality, questionnaire development, scheduling of interviewers, respondent burden, interviewer effects, and the use of the computer in telephone interviewing. Throughout, the authors focus on findings from recent research, with particular emphasis on those studies suggesting new advances or protocols for conducting telephone health surveys. The findings of this review suggest four conclusions. First, telephone interviews can be highly recommended for follow-up interviews in panel surveys that use an initial face-to-face interview. Second, telephone surveys can be recommended as a viable alternative to costly face-to-face surveys in cross-sectional studies of the general population. Third, when the focus of the survey is on subgroups of the population that have both low telephone coverage and higher rates of nonresponse (e.g., low income and low education respondents), telephone interviews should be used more cautiously. In these situations, a dual sampling frame approach (using a combination of face-to-face and telephone interviewing) may be considered. Finally, computer-assisted telephone interviewing (CATI) represents one of the most important and innovative technologic advances in health survey research in recent years. The advantages of CATI in improving survey management are noteworthy and ideally suited for moderate- to large-sample surveys. CATI also provides an attractive (and largely untapped) resource for testing and refining other methodologic protocols in survey research.

Physician and hospital factors associated with mortality of surgical patients.

Abstract: Recent studies have found an inverse relationship between hospital-specific mortality rates for selected conditions and the number of patients hospitalized with these conditions. These studies have not examined whether this inverse relationship is a result primarily of the nature and volume of services provided to patients by individual physicians or whether it reflects special characteristics of high-volume hospitals. This study examines these issues, using data that link characteristics of primary surgeons to the discharge abstract records of patients. The study analyzes variation in hospital mortality associated with: the total volume of specific surgical procedures performed in the hospital, the volume of these procedures performed by the patient's primary surgeon, physician board certification, and other factors including patient severity of illness, patient age, hospital control, teaching status, size, and location. The findings confirm the inverse relationship found in other studies between patient mortality and the total volume of specific surgical procedures performed in the hospital. Physician board certification and hospital's medical school affiliation also are found to be associated with lower patient mortality rates. However, there is no statistical relationship between the volume of services provided by individual surgeons and outcome, suggesting that the volume-outcome relationship reflects hospital rather than physician characteristics.

Medical and psychiatric determinants of outpatient medical utilization.

Abstract: Ninety-two general medical outpatients were surveyed with an interview, questionnaires, and a medical record review to investigate the relationships among psychiatric disorder (depression and hypochondriasis), somatic symptoms, medical morbidity, and the utilization of ambulatory medical services Medical utilization correlated with the number of somatic symptoms reported (r = 049, P = 00001), depressive symptoms (r = 034, P = 0001), and the number of medical diagnoses in the medical record Somatic symptoms were not significantly correlated with the number of medical diagnoses, but were related to hypochondriacal attitudes (r = 052, P = 00001) and depression (r = 051, P = 00001) In stepwise multiple regressions, the number of medical diagnoses accounted for 33% of the variance in medical utilization Somatic symptoms were the second most powerful predictor, increasing R2 to 0469 The next best predictors were two hypochondriacal attitudes and the presence of a major psychiatric diagnosis in the medical record This five-step model explained 56% of the variance Somatic symptoms are thus powerful determinants of medical utilization, even after controlling for medical morbidity Depression, disease fear, and bodily preoccupation are also important predictors of utilization Somatic symptoms are a final common pathway through which emotional disturbance, psychiatric disorder, and organ pathology all express themselves, and which prompt patients to visit doctors

Economic and policy analysis of university-based drug "detailing".

Abstract: The cost-effectiveness of quality assurance programs is often poorly documented, especially for innovative approaches. The authors analyzed the economic effects of an experimental educational outreach program designed to reduce inappropriate drug prescribing, based on a four-state randomized controlled trial (N = 435 physicians). Primary care physicians randomized into the face-to-face group were offered two individualized educational sessions with clinical pharmacists, lasting an average of 18 minutes each, concerning optimal use of three drug groups that are often used inappropriately. After the program, expenditures for target drugs prescribed by these physicians to Medicaid patients decreased by 13%, compared with controls (P = 0.002); this effect was stable over three quarters. Implementation of this program for 10,000 physicians would lead to projected drug savings (to Medicaid only) of $2,050,000, compared with resource costs of $940,000. Net savings remain high, even after adjustment for use of substitution medications. Although there was a ninefold difference in average preintervention prescribing levels between the highest and lowest thirds of the sample, all groups reduced target drug expenditures at the same rate. Targeting of higher-volume prescribers would thus further raise the observed benefit-to-cost ratio from approximately 1.8 to at least 3.0. Net benefits would also increase further if non-Medicaid savings were added, or if the analysis included quality-of-care considerations. Although print materials alone may be marginally cost-effective, print plus face-to-face approaches offer greater net benefits. The authors conclude that a program of brief, face-to-face "detailing" visits conducted by academic rather than commercial sources can be a highly cost-effective method for improving drug therapy decisions. Such an approach makes possible the enhancement of physicians' clinical expertise without relying on restriction of drug choices.

Selecting categories of patients for regionalization. Implications of the relationship between volume and outcome.

Abstract: A growing number of researchers have demonstrated an inverse relation between the number of patients treated with specific diagnoses or procedures in a hospital and subsequent adverse outcomes. Such findings support the notion that policies should be explored to concentrate patients in selected hospitals to reduce preventable patient mortality or morbidity. The authors used data from 15 diagnoses and procedures demonstrating an inverse relation between volume and mortality to explore the different implications of regionalization policies across categories of patients. In some instances, concentrating patients in hospitals with high volumes of such patients could avert more than 60% of all deaths. For some procedures or diagnoses, however, such mortality savings are either medically infeasible because of the emergency nature of the problem or logistically impossible because of the extent of regionalization implied.

Predicting and understanding influenza vaccination behavior. Alternatives to the health belief model.

Abstract: Recent reviews indicate that attempts to validate the Health Belief Model (HBM) have produced only modest support for it and that the HBM may be incomplete in its organization and development as an expectancy-value model. As an alternative, the Fishbein and Triandis models are discussed. The author investigated the utilities of the Fishbein and Triandis models as predictors of behavioral intention and behavior, using obtaining versus not obtaining an influenza vaccination as the health behavior. Eligible participants were patients at high risk for flu complications, registered at the Seattle VA Medical Center's Medical Comprehensive Care Unit (MCCU) clinic. A random sample of 439 patients was selected to participate in a two-wave longitudinal survey in which the Fishbein and Triandis model components were assessed at the beginning of the 1983 flu season, and a measure of flu shot-getting behavior was obtained at the end of the season. Both models accounted for significant and substantial proportions of variance in intentions and behavior. Comparisons of these models revealed that the three-component Triandis Intention model and the overall Triandis model were superior to the Fishbein model in predicting intention and behavior. The potential use of this model for developing interventions is discussed.

Sociodemographic factors and the use of outpatient mental health services.

Abstract: What are the effects of sociodemographic factors on the use of outpatient mental health services when different demographic groups have identical health insurance coverage? The authors answer this question using data from the Rand Health Insurance Experiment. Health insurance was randomly assigned t

Weights for scoring the quality of well-being instrument among rheumatoid arthritics. A comparison to general population weights.

Abstract: The importance of measuring health outcomes such as functional status and quality of life has increased with the greater emphasis on efficiency and on judgements of clinical effectiveness of therapies for patients with chronic disease. One measure of health status, the quality of well-being (QWB), has received significant attention as a health policy model because it quantifies health on a scale ranging from “zero” (death) to “one” (optimal health). The scale is based on weights (values) that were derived by having several thousand individuals in the general population rate scenarios in which a patient is described in terms of mobility, physical activity, social activity, and major symptom or problem. The present study was undertaken to determine if a disease-specific population composed of patients with moderate and moderately severe rheumatoid arthritis who were participating in a national multicenter trial of a new oral therapeutic agent, would rank scenarios similarly to the general population sample. In this study, close agreement was found between the weights obtained from the general population sample and the weights obtained from the sample of rheumatoid arthritic patients (R. = 0.937). The investigators believe that the study supports the use of the original general population weights and suggest that the index may be used for populations with a specific condition as well as for general populations.

Reliability and validity of the Severity of Illness Index.

Abstract: The authors discuss the objectives and definition of the Severity of Illness Index, which has been developed and refined at The Johns Hopkins University over the past 5 years. In addition, the training program for raters, the method used to ascertain reliability, and data from reliability testing in

The association between mother's and children's use of physician services.

Abstract: Using data from the National Health Interview Survey on approximately 30,000 children, maternal and child ambulatory care utilization patterns are compared. The results indicate that maternal physician utilization is closely associated with child utilization, as measured by presence or absence of a

Issues in the use of kappa to estimate reliability.

Abstract: Selected issues regarding the application of the kappa statistic in assessing the reliability of measures to classify patients are discussed. Data are used to illustrate how the methods used to estimate reliability can influence the resulting coefficient. Suggestions for a more comprehensive approac

"Risk" factors affecting readmission of the elderly into the health care system.

Abstract: One hundred one patients, 70 years and older, who were discharged to the community from an acute-care hospital were followed for 1 year to isolate risk factors affecting the probability of readmission. A total of five interviews were conducted with each patient. Postdischarge outcome at any point in

Parent Satisfaction With Children's Medical Care: Development, Field Test, and Validation of a Questionnaire

Abstract: Research on both adult patients and parents of pediatric patients has demonstrated that satisfaction with medical encounters predicts such important outcomes as compliance with medical regimen. The authors developed a questionnaire to measure parent satisfaction with children's medical encounters, administered it to 104 parents of pediatric patients (field trial 1), and revised it. The revised Parent Medical Interview Satisfaction Scale (P-MISS) was then tested on a new sample of parents whose medical visits were videotaped (field trial 2). On field trial 2, the P-MISS showed a high alpha reliability (0.95). The four factor-based subscales identified by field trial 1 showed high alpha reliabilities on field trial 2: physician communication with the parent (0.81); physician communication with the child (0.93); distress relief (0.85); and adherence intent (0.86). With the exception of the distress relief subscale, the subscales appear to measure distinct dimensions of satisfaction. Objective ratings of physicians' interpersonal skills to parents during medical interviews correlated significantly with parents' total satisfaction scores as well as with all four satisfaction subscale scores, providing preliminary evidence of the construct validity of the P-MISS.

Classifying Function for Health Outcome and Quality-of-life Evaluation: Self- Versus Interviewer Modes

Abstract: Validity assessment and the underreporting of dysfunction have been major problems in health-related quality-of-life measurement, including collecting data for analysis by the General Health Policy Model, using the Quality of Well-being scale (QWB). This analysis compares the results of self- versus interviewer modes of measurement and short, direct-answer questions versus probing algorithms in the QWB. The comparisons are made in terms of 1) correlations; 2) aggregate frequencies; 3) individual subject classifications; and 4) the actual state, established using evidence from multiple sources. Despite extremely high correlations between QWB scores from the two modes (greater than 0.98), the lowest interviewer mode sensitivity (0.86) and predictive value dysfunctional (0.91) were substantially superior to the highest self-classification characteristics (0.66 and 0.73). In the populations studied, specificities and predictive values functional were equivalent (greater than 0.94) for the two modes. The probe pattern of the interviewer mode was also less susceptible to false reports of dysfunction. These results are consistent with the underreporting of dysfunction noted by several major investigations of health status measurement. The authors conclude that interviewer-administered instruments using question algorithms are necessary if health-related quality of life is to be measured with sufficient reliability and validity to evaluate major clinical trials and follow-up studies.

My medical care is better than yours. Social desirability and patient satisfaction ratings

Abstract: It is well-documented that ratings of medical care received personally (personal referent) yield more favorable responses than ratings of care received by people in general (general referent). Hence general items are useful in achieving greater variation in responses to satisfaction surveys. However, the validity of general items relative to personal items is being debated currently. It has been hypothesized that bias due to socially desirable response set (SDRS) would be greatest for items with a personal referent. To test this hypothesis, the authors compared both kinds of satisfaction ratings for adults (N = 3,918) who scored high and low on SDRS during Rand's Health Insurance Experiment. Across sites and years of the experiment, the rating item with a personal referent was consistently biased upward for those manifesting SDRS. The rating item with a general referent was not. Further, the correlation between SDRS and the difference between ratings on the personal and general referent items was statistically significant, suggesting that more favorable ratings of medical care received personally compared with ratings of care received by people in general are in part due to SDRS bias. Results are discussed in terms of implications for constructing a valid satisfaction survey.

Predicting posttrauma functional disability for individuals without severe brain injury.

Abstract: The utility of the Abbreviated Injury Scale (AIS), the most widely used anatomic scale for rating severity of injuries, and its derivative for assessing the combined effect of multiple injuries, the Injury Severity Score (ISS), were tested for their ability to predict functional disability at time of discharge from the hospital and 6 months after discharge. The ISS has been shown to correlate well with mortality and length of stay, but the relationship to levels of subsequent disability has not been examined. Five hundred and ninety-seven patients (aged 16-45 years) were interviewed at time of discharge and 6 months after discharge to ascertain functional disability along three dimensions: activities of daily living (ADL), instrumental activities of daily living (IADL), and mobility. The authors report on the relationship between severity and functional disability at time of discharge and 6 months after discharge for a subset of 473 patients who did not suffer a severe brain injury. The results show that the relationship between ISS and status at discharge and 6 months after discharge is not monotonically increasing, as it is with mortality and length of stay (LOS). Rather, the proportion of people with severe injuries who report limitations is lower than for those with moderately severe injuries as defined by the ISS. Further, it is shown that the AIS of the most severe extremity and spinal cord injury carry considerably more weight when predicting functional status at discharge and 6 months after discharge than do the AIS scores of injuries to any other body region, although the relative explanatory power of each type of injury varies with the nature of the functional disability and the time interval between the initial insult and assessment.

Determinants of primary medical practice in adult cancer prevention.

Abstract: The authors conducted a study of primary care physicians in the province of Quebec to ascertain their patterns of preventive practice with respect to cancer in four anatomic sites: breast, cervix, colon-rectum, and lung. They further explored the data set to elicit the determinants of the patterns of preventive practice. Scales were constructed encompassing practice behaviors for each type of cancer, continuing education intensity, knowledge, and belief. The content of these scales was delineated through factor analysis and their reliability assessed using Cronbach's alpha. Other variables were also considered in the conceptual model. Bivariate analysis and multivariate techniques were used. The models tested contained many significant interaction terms. A limited number of the first-order interactions was explored for each of the dependent variables. Different patterns emerged for each cancer type. Mode of reimbursement, continuing education, gender of physician, provider-related barriers to prevention, and knowledge were found to be the major predictors of prevention scores for the cancers studied, but their relative importance varied according to each cancer. The importance of better understanding the determinants of physician behaviors is emphasized and the existence of several possible explanatory models suggested.

Maximizing compliance with hemoccult screening for colon cancer in clinical practice.

Abstract: A factorial design randomized controlled trial to test several clinically feasible strategies primary-care practitioners may use in routine practice to increase patient participation in occult blood testing for colorectal cancer is reported. Three compliance-enhancing intervention strategies (physician/nurse talk, and/or reminder postcard, and/or reminder phone call) were introduced. Patient health beliefs were examined as compliance predictors. High compliance levels were seen in all intervention groups, with a mean of 89% compared with 68% in controls. An interactive talk by the physician or nurse increased compliance by 12-13%. The reminder postcard was the most effective single intervention. It increased compliance by 24-25%, achieving 92.7% overall compliance, and appeared to be cost-effective. Patient health beliefs were of minimal value in predicting compliance in this study.

Centralization, certification, and monitoring. Readmissions and complications after surgery.

Abstract: Research on adverse outcomes following common surgical procedures has suggested the importance of hospital and surgeon variables. Policy directions depend on which factors are important in influencing patient outcomes and what sorts of policies are feasible. Focusing on where a given procedure is performed highlights a concern for centralization; emphasizing who should perform a particular operation implies physician certification. Finally, monitoring involves identifying particular hospitals that appear to have relatively poor (or relatively good) results. This paper analyzes patient, surgeon, and hospital characteristics associated with serious postdischarge complications of hysterectomy, cholecystectomy, and prostatectomy in patients age 25 and over in Manitoba, Canada, following surgery during 1974 through 1976. The three procedures differ markedly in the ease of prediction of the probability of complications and in the predictive importance of patient, hospital, and physician variables. The predictors worked fairly well for cholecystectomy, somewhat less well for hysterectomy, and not well at all for prostatectomy. Hospital variables were not generally important in the multiple logistic regressions. After controlling for case mix and type of surgery, physician surgical experience was found to account for relatively large differences (almost two to one) in the probability of patient complications following cholecystectomy. Cholecystectomy might be a candidate for certification because of the epidemiology of the operation. As of the mid-1970s, a substantial proportion of the cholecystectomies were being performed by physicians with comparatively little ongoing experience with this type of procedure. Moreover, a monitoring perspective identified one hospital with a significantly higher postcholecystectomy complication rate, even after physician experience was taken into account. Both identifying which procedures should be attended to and focusing on problems following surgery are important beyond Manitoba and highly relevant to such American requirements as Peer Review Organizations. Methods of increasing the efficiency of using claims data for quality assurance studies are outlined.

Medical Care Utilization by Hispanic Children: How Does it Differ From Black and White Peers?

Abstract: Factors associated with the utilization of medical care by Hispanic, black Non-Hispanic, and white Non-Hispanic children were analyzed using multivariate regression techniques on a Health Interview Survey sample. The findings indicate that Medicaid coverage was the sharpest enhancer of entry into ca

The lifetime risks and costs of nursing home use among the elderly.

Abstract: In this paper, we estimate the risk of an individual of entering a nursing home throughout the aging process. We then estimate the expected lifetime costs of nursing home use both for an individual and for society as a whole. The model is based on double-decrement life-table analysis. Data are taken from a 1977 survey of 4,400 Medicare beneficiaries. At age 65, the upper bound for the lifetime risk of entering a nursing home is 43.1%. The risk of entering a nursing home increases with age until around age 80. At about age 85, the risk begins to decline significantly. At almost all ages, the lifetime risk of entry for females is twice that of males. The expected lifetime costs of nursing home care across all ages are between $10,500 and $13,600. These costs are distributed very unequally. Only 13% of the elderly account for 90% of all nursing home expenditures. Given current life expectancy, the expected annual cost per person over age 65 is between $532 and $760. In the year 2000, the expected annual average costs of nursing home care per elderly person will range from $450 to $650. The decline in the average annual cost per person reflects shifts in the age structure and increased life expectancy. These figures need not represent an unmanageable burden on society's resources. Figures presented here help establish the feasibility and desirability of long-term care risk-sharing arrangements among the elderly, like long-term care insurance, life care communities, and other models.

Variation in resource use within diagnosis-related groups. The effect of severity of illness and physician practice.

Abstract: The diagnosis-related group (DRG)-based Medicare prospective payment system pays hospitals a fixed amount for the care of similar patients. The DRG definitions serve as the modifier of payment for Medicare patients. The dependence on these patient definitions raises many questions, among them the reason(s) for observed resource variability within a DRG. Various severity-of-illness measures have been shown to account for some of the resource variability noted within the DRGs. Most severity-of-illness studies to date, however, have not attempted to assess the effect of other known sources of resource variation, such as differing physician practice patterns. The authors examined the ability of severity of illness, as defined by disease staging, and physician practice variation to explain residual intra-DRG variability in length of stay. They demonstrate that physician practice variation accounts for more variance reduction than does severity of illness.

Cost-effectiveness of the North Karelia Hypertension Program. 1972-1977.

Abstract: The North Karelia Hypertension Program was initiated in 1972 as part of the North Karelia Project. This article examines the costs and effects of the first 5 years of the project. There were 288 fewer than expected fatal strokes and myocardial infarctions during the period; 134 of these are attributable to the hypertension program. The costs of the hypertension program totalled $5.16 million. Drugs consumed 86% of this cost. Using the zero discount rate as an outside boundary, the program is expected to increase old age pension costs by $2.5 million for the 5 years but will decrease earnings losses by $7 million for the same period. With earnings excluded, the cost per quality-adjusted life-year gained is $3,612 at zero discount and $5,830 at 10% discount. Hypertension care is more cost-effective than many of the treatments applied after the appearance of coronary heart disease symptoms but would be much more cost-effective if hypertension could be treated as effectively without medications or if the costs of medications could be reduced.

Cost and outcome of care for patients with chronic obstructive lung disease. Analysis by physician specialty.

Abstract: We studied the effect of physician specialization and board certification on costs and outcome of health care for a group of 213 patients with chronic lung disease followed prospectively for a year. Linear, semilogarithmic, and logistic regressions were used to control for differences in pulmonary function, functional ability, and sociodemographic characteristics. The cost of health services during the year was estimated from the total charges incurred. Patient's pulmonary function, functional ability, number of medical conditions, and insurance status were significant predictors of total cost. Combinations of these variables were important determinants of institutional days, outcome health status, and survival. Physician specialization and board certification were not significant descriptors of total costs or outcomes, although large variances limited the power of these findings. We conclude that differences in characteristics of primary care physicians do not appear to affect significantly the total cost or outcome of care for patients with moderate to severe chronic lung disease.

Influence of simple computerized feedback on prescription charges in an ambulatory clinic. A randomized clinical trial.

Abstract: Pharmaceuticals account for a significant portion of health care costs and are an important target for attempts at cost reductions. While many techniques have been shown effective, most are resource-intensive, have demonstrated fatigue after the intervention is ceased, and have been directed at specific items rather than total charges. The authors designed a computerized program to feed back prescription charges. The intervention is easy to execute, inexpensive, and can be maintained indefinitely. The intervention was performed in a randomized, prospective, controlled trial with the medical residents of a large county hospital. The goal was to reduce total prescribing charges and produce a meaningful financial result. The intervention reduced the mean charge for a prescription by 6.7% (P less than 0.025), but with a long latent period and minimal impact on resident knowledge of drug charges. Significant differences were seen only at the end of the study. The program was viewed positively by the residents. The low cost of the intervention yielded a benefit-to-cost ratio in excess of 50:1. Because of computerization and ongoing patient and resident randomization at the study hospital, added costs of this randomized trial in terms of computer time and research assistance were less than $1,000.

Measures of continuity of care. A register-based correlation study.

Abstract: In an empirical study using data from a health center in Sweden, correlation coefficients were computed among nine different measures of continuity of care, five of them visit-based and four individual-based. Generally, the correlations were high. This may be due, in part, to the similar behavior of the measures for people making few visits. The correlations were also quite high, however, when the sample was restricted to people with many visits. Several measures display a significant dependence on utilization level. The results suggest that, for general purposes, the measure COC should be preferred among the individual-based measures and fraction-of-care continuity among the visit-based measures. On grounds of flexibility and ease of interpretation, the authors recommend fraction-of-care measures.