Medical Decision Making

Objective. This study establishes the least costly strategy for evaluation of rural women in need of colposcopy among 3 alternatives: telemedicine, local practitioners, and referral experts. Methods. Women in rural Georgia who needed colposcopy were examined by an expert colposcopist on site, by a local practitioner, and by a distant expert colposcopist linked by telemedicine. Independent determinations of biopsy intent were used to model the differing biopsy costs of the 3 methods. Record reviewdetermined the average total cost of telemedicine. Reports of average cost in year 2000 dollars from societal perspective include medical costs and pain and suffering due to additional biopsies and curettage, telemedicine costs, and costs of potential diagnostic delay for a 1-year time horizon. Results: From the societal perspective in the baseline case, the average cost per patient evaluated was $270 for patients seen by referral experts. The cost was $38 less (e.g., $232) for patients seen by local practitioners, and $35 more (e.g., $305) for patients seen by telemedicine. From the societal perspective, local practitioners were less costly than referral experts because of lower travel costs for patients, but from the medical perspective, their average cost was $32 higher than referral experts because they performed more biopsies and curettage procedures than experts. Telemedicine assistance would have lowered the number of biopsies performed by local practitioners, but as of year 2000 the costs of this technology could not be justified by the savings. Conclusion. From the societal perspective, local practitioners performing colposcopy are the least costly way to evaluate cervical abnormalities in rural patients with substantial time and travel costs.

Purpose. The purpose ofthis study was to explore prostate cancer patients’ perceptions, feelings, ideas, and experiences regarding making decisions to use (or not use) complementary/alternative medicine (CAM). Methods. Five focus groups were conducted with 29 men diagnosed with prostate cancer. Content analysis ofthe verbatim transcripts was used to identify key themes in the data. Findings. Decision making about CAM treatments appears to depend on both fixed (e.g., disease characteristics, demographic characteristics, and medical history) and flexible (e.g., perceptions of CAM and conventional-medicine, experiences with the health care system and health care practitioners, and perceptions about the need for control or action) decision factors. Conclusions. The participants in this study appeared more likely to be “pushed” toward using CAM by negative experiences with the health care system than to be “pulled” toward CAM by perceptions about its safety or congruence with their beliefs about health and illness.

Background. Quality of life is often best estimated by standard gamble techniques. However, these techniques usually require time-consuming and expensive interviews or computerdirected questionnaires. Paper Standard Gamble (PSG) is a paper questionnaire that has previously been shown to accurately represent standard gambles elicited by computer. The authors sought to demonstrate its test-retest reliability in comparison to other, paper-based measures of quality of life. Methods. The authors used a longitudinal cohort designwith duplicate assessments of quality of life by PSG, the Dermatology Life Quality Index, and the Mental and Physical Component Summary scores of the SF-12 in stable dermatology outpatients. Baseline measures were performed by mail 1 to 2 weeks before a scheduled dermatology clinic visit. Follow-up measures were performed in the waiting room before being seen by the dermatologist. The authors calculated the coefficient of variation and the Spearman rank correlation coefficient for each of the instruments. Results.74 patientswith stable skin conditions participated. The coefficient of variation of PSG (0.47%) was smaller than the other instruments (4.26%–5.22%); PSG’scorrelation washigher (0.97 v. 0.65–0.80). Conclusion. PSG, a 1-page paper questionnaire, is a reliable measure of patient utility suitable for use in postal surveys.

Purpose. Past research into health-related quality of life (HRQL) in diabetes using preference-based measures, such as the Health Utilities Index Mark 2 (HUI2) or the Health Utilities Index Mark 3 (HUI3), is sparse. Important differences between the HUI2 and HUI3 could lead to differences in their abilities to capture HRQL deficits in type 2 diabetes. This analysis compared the extent to which the HUI2 and HUI3 detect differences associated with varying levels of disease severity or advancement in type 2 diabetes. Methods. This analysis was conducted as part of using baseline, crosssectional data from a larger, prospective, controlled study of an intervention to improve care for individualswith type 2 diabetes in ruralcommunities in Alberta, Canada. The HUI2 and HUI3were self-administered to 372 community-dwelling individuals with type 2 diabetes. Results. Relative to HUI2 scores, larger differences in overall HUI3 scores were seen forcomparisons for individuals presumed to have more advanced disease. The pain attribute of the HUI3 categorized a larger proportion of individuals as moderately to severely impaired (41.5% v. 24.2%,P < 0.001), as did the emotion attribute (20.5% v. 7.7%,P < 0.001). For individuals with negative overall HUI3 scores, differences between overall HUI2 and HUI3 scores persisted after rescaling (mean difference = 0.33,P = 0.009). Conclusions. The greater range of possibl e scores on the HUI3, its relative ability to assess the utility of states worse than dead, and its relative superiority in discriminating moderate to severe impairment from mild or no impairment might favor its use over the HUI2 in type 2 diabetes.

Aim. To study the effect of sociodemographic and attitudinal determinants of physicians making end-of-life decisions (ELDs). Methods. The physicians having signed 489 consecutive death certificates in the city of Hasselt (Belgium) were sent an anonymous questionnaire regarding their ELDs and another on their attitudes toward voluntary euthanasia (EUTH) and physician-assisted suicide (PAS).Results.55% response rate. Nontreatment decisions occurred in 16.7% of all death cases; in 16%, there was potentially life-shortening use of drugs to alleviate pain and symptoms; in 4.8% of cases,death was deliberately induced by lethal drugs, including EUTH, PAS, and life termination without explicit request by the patient. In their attitudes toward EUTH and PAS, the 92 responding physicians clustered into 3 groups: positive and rule oriented, positive rule-adverse, and opposed. Cluster groupmembership, commitment to life stance, years of professional experience, and gender were each associated with specific ELD-making patterns.

Background. Significant disparities in health care utilization exist across gender and race. Little is known about the patient-specific factors that may contribute to this variation. This study examined variations in risk attitude across major sociodemographic groups.Methods. A survey elicited utility measures for health states under risk-insensitive and risksensitive conditions (time tradeoff and standard gamble methods, respectively). Risk attitude was modeled assuming constant proportional risk posture, thus the utility function used was a power function. A multivariable linear regression model was used to examine the relationship between risk attitude and sociodemographic factors.Results.Of the 62 study subjects, the mean age was 47.6 years, 47% were female, and 33% were African American. Overall, 37% of respondents-were decidedly risk averse, 37% moderately risk averse, 15% moderately risk seeking, and 11% decidedly risk seeking. Significant predictors of increasing risk aversion in multivariate modeling were white race (P < 0.01) and lower education (P < 0.05). Women also tended to be more risk averse (P = 0.07).Conclusions.This study found significant differences in risk attitude across race and educational status, with a smaller difference across gender. Further research is needed to validate these findings and clarify their contribution to racial and gender variations in health care utilization and their future role in decision and cost-effectiveness analyses.

Quality improvement programs must compete with other health care interventions for limited health care resources. The goal of the research presented here was to develop a model that portrays the mathematical relationship between the size of a quality deficit caused by the noncompliance of health professionals and the cost-effectiveness of a quality improvement program. The model allows the determination of the minimum size of a quality deficit for which it is worth introducinga quality improvement program. If a quality improvement program has already been implemented, the model can be used to define the quality threshold beyond which a reduction in quality becomes economically unattractive. An example consideringthe reduction of underuse in depression treatment demonstrates that an intervention with a favorable cost-effectiveness ratio may become economically unattractive once the costs for the implementation effort are considered.

Background. There is an increasing movement towards the release of hospital “report-cards.” However, there is a paucity of research into the abilities of the different methods to correctly classify hospitals as performance outliers.Objective.To examine the ability of risk-adjusted mortality rates computed using conventional logistic regression and random-effects logistic regression models to correctly identify hospitals that have higher than acceptable mortality.Research Design.Monte Carlo simulations.Measures.Sensitivity, specificity, and positive predictive value of a classification as a high-outlier for identifying hospitals with higher than acceptable mortality rates.Results.When the distribution of hospital-specific log-odds of death was normal, random-effects models had greater specificity and positive predictive value than fixed-effects models. However, fixed-effects models had greater sensitivity than random-effects models.Conclusions.Researchers and policy makers need to carefully consider the balance between false positives and false negatives when choosing statistical models for determining which hospitals have higher than acceptablemortality in performance profiling.