James M. Sinacore

A comparative study of seven measures of patient satisfaction.

The acceptability of satisfaction as a quality indicator is qualified by several well known measurement problems. This study examines the variability in satisfaction evaluations related to different measurement methods and the effect of response biases on reported satisfaction. Satisfaction evaluations using seven different, commonly used measures of patient satisfaction were obtained from the same sample of respondents. The seven measures were: 1) a global measure of satisfaction using a visual analogue scale; 2) a multidimensional measure of satisfaction based on the Patient Satisfaction Questionnaire using an evaluation response format (poor, fair, good, very good, excellent); 3) a two-item overall evaluation of quality using the evaluation response format; 4) a six-item attitude measure of general satisfaction using a five-point Likert agree-disagree response format; 5) a four-item attitude measure of satisfaction with physician, using the agree-disagree response format; 6) a four-item measure of behavioral intention; and 7) willingness-to-pay in dollars. The percentage of favorable evaluations of care ranged from 63% to 82% across six of the seven measures. Willingness-to-pay does not appear to be a valid measure of satisfaction. Correlations were highest between measures with similar response formats. Although an oppositional response bias was not found, a very substantial acquiescent response bias was detected. Acquiescence reduced the internal consistency of three multiple-item measures, the general and physician attitude and behavioral intention measures, to levels unacceptable even for group comparisons. Between highly and nonacquiescent respondents, levels of satisfaction were somewhat lower for the multidimensional measure of satisfaction and significantly lower for the two attitude satisfaction measures. Highly acquiescent respondents were older, less well educated, and in poorer health than nonacquiescent subjects. Results of satisfaction evaluations dependent on the measurement method used, and unreliability of measurement may be a significant problem in satisfaction measurement, especially for the oldest and most ill patients.

The importance of patient preferences in the measurement of health care satisfaction.

The idea that patients will be more satisfied with health care services that are delivered to meet their preferences is central to the concept of health care marketing. Health care providers increasingly use market segmentation and target marketing to optimize the fit between their services and the consumers who receive them. This study evaluates one model for incorporation of patient preferences into the measurement of satisfaction. Using multiple regression analysis, evaluations of three dimensions of health care satisfaction, interpersonal care, technical quality, access to care accounted for 63% of the variance in overall satisfaction. Inclusion of preferences, defined as importance ranks of each dimension, did not improve ability to predict satisfaction. Four preference segments were identified: interpersonal care seekers, access/quality seekers, access seekers and quality seekers. These four subgroups differed significantly on a number of sociodemographic, health status and health service use characteristics but no significant differences were found in satisfaction between preference segments. Patient satisfaction can best be measured as quality evaluations of dimensions without regard to preferences. In considering the merits of market segmentation and target marketing, alternative satisfaction models that link preferences to health care satisfaction or the possibility that preference targeting does not lead to greater satisfaction should be evaluated.

Predicting stroke inpatient rehabilitation outcome using a classification tree approach

A classification tree, a nonparametric statistical analysis, was used to develop decision rules to predict a favorable inpatient stroke rehabilitation outcome. Descriptive and functional status data collected on admission from 225 patients were the predictor variables. Favorable outcome was defined as having met three criteria: discharged to community, survival greater than 3 months postdischarge, and no more than minimal physical assistance required in functional activities on discharge. The classification tree correctly classified 88% of the sample using only four of the predictor variables (level of independence in Toilet Management, Bladder Management, and Toilet Transfer, and adequacy of Financial Resources). The cross validation error rate was 18%. The advantages of the classification tree approach over parametric methods are that it is desirable for ordinal data, it readily identifies the interactions among predictor variables, the results are easily communicated, and it provides additional insights into the factors that predict outcome.

Self Report and Performance-Based Hand Function Tests as Correlates of Dependency in the Elderly

Preventing or minimizing functional dependency in older adults rests, in part, upon the ability to predict who is at risk. The purpose of this study was to compare the ability of five tests of hand function to discriminate the degree of dependency in older adults. Seven hundred sixty four subjects were assessed for hand function on performance‐based (Williams Test of Hand Function, a test of Williams Board items only, Jebsen Test of Hand Function, grip strength), and self‐reported (Dexterity Scale of the Geriatrics‐Arthritis Impact Measurement Scale (GERI‐AIMS)) measures of hand function, and self‐reported multidimensional functional status (GERI‐AIMS). A trichotomous variable representing a continuum of dependency based upon living site (independent living, home‐bound, institutional) was used as the measure of dependency. Sixty‐two cases were dropped for incomplete data. Discriminant function analyses of the 702 subjects (age = 76.78 years, SD = 8.79) showed that basic demographic variables explain 40.8% of the variance in dependency; all hand function tests significantly correlated with dependency; the Williams Board correlated best (additional 12.5% variance explained). However, a multidimensional functional status measure explains substantially more variance in dependency (16.9%) after controlling for demographic variables and performance on the Williams Board. This comparison of methods and tests available for measuring hand function was made to provide criteria for selecting an instrument for a given setting.

Stroke Inpatient Rehabilitation: A Comparison across Age Groups

Objective: To examine and compare the inpatient stroke rehabilitation experience of older adults (≥75 years) with that of young adults (<65 years) and young‐old adults (65–74 years).

Assessing the semantic content of clinical case presentations: studies of reliability and concurrent validity.

No abstract available.

Increasing Compliance to Medical Treatment Regimens A Meta-Analysis of Program Evaluation

Research articles evaluating the effectiveness of programs to increase compliance with medical treatment regimens were quantitatively integrated to assess the impact of these programs on the behavior ofpatients. A total of58 studies involving two or more groups with 133 measures of compliance were identified and analyzed. The mean effect size was .47, indicating that the typicalprogram participant complied better than 68% of the members of the control groups. The advantage of the program groups dropped as the amount of lifestyle changes required by the treatment regimen increased. Overall, the most successful interventions involved improving the facility providing care and helping patients to incorporate the treatment regimen into their daily routine. It is suggested that publishedevaluationsofcomplianceprograms would be more useful and more likely to contribute to an accumulation of knowledge if more careful descriptions of the interventions, including costs estimates, were included in reports of the program evaluations.

Interrater reliability of the DSM-III-R with preschool children

Little attention has been paid to evaluating the use of DSM-III-R with preschool children. Children (N = 510) ages 2 to 5 years who were screened at the time of a pediatric visit were selected to participate in an evaluation which included questionnaires, a semistructured interview, developmental testing, and a play observation. Following the evaluation, two clinical child psychologists independently assigned DSM-III-R diagnoses. For each diagnostic category, kappa and Ycoefficients were calculated; Ycoefficients are less sensitive to base rates of disorders. For overall agreement, the weighted mean kappa (.61), and mean Y(.66) were moderately high. Overall agreement that the child had at least one of the disruptive disorders was substantial (kappa =.64; Y =.65);agreement that there was at least one of the emotional disorders was moderate for kappa (.54), but substantial for Y(.70). Kappa coefficients were higher for major categories of disorder than for specific disorders; however, Ycoefficients did not show a decline for specific disorders. Interrater reliability of DSM-III-R appears to be similar for preschoolers and older children.

Self‐Reported Functional Status Predicts Change in Level of Care in Independent Living Residents of a Continuing Care Retirement Community

To test the hypothesis that self‐reported functional status predicts change in level of care from independent to dependent in residents of a continuing care retirement community (CCRC).

Book Reviews : Multiple Regression: Testing and Interpreting Interactions, by Leona S. Aiken and Stephen G. West. Newbury Park, CA: Sage, 1991, 212 pp:

James M. Sinacore Co-Director, Biostatistical and Data Management Core, Multipurpose Arthritis and Musculoskeletal Diseases Center, Northwestern University Medical School, Ward Hall, Room 3-315, 303 East Chicago Avenue, Chicago, IL 60611 As a research and statistical consultant I have noted that there is a broad misunderstanding of the meaning, analysis, and interpretation of interactions. Unfortunately, some textbooks discourage the general researcher from learning more about interactions by placing too much emphasis on mathematical principles and equations. In contrast, however, this book by Leona Aiken and Stephen West is oriented toward applications and is a good addition