The aim of this paper is to demonstrate the use of conjoint analysis (CA) in health services research. Conjoint analysis is first explained, with emphasis on the history of the technique, followed by an explanation of how to carry out such a study and how the results from such a study can be used. The technique is demonstrated with reference to a study that looks at the benefits of in vitro fertilization. It is shown how CA can be used to estimate the relative importance of attributes, the trade-offs individuals make between these attributes, willingness to pay if cost is included as an attribute, and utility or benefit scores for different ways of providing a service. The paper then considers the potential advantages of CA over other, more commonly used benefit assessment instruments. Finally, there is discussion of the issues raised in the design and analysis of CA studies. It is concluded that these issues must be addressed before the technique becomes an established instrument for technology assessment.

Objectives: To develop and evaluate a new model of continuing medical education (CME) for general practitioners (GPs). The study is part of the joint European Drug Education Project (DEP). This paper presents the Swedish part regarding the design of the evaluation study, the educational methodology, and the participants' evaluation.

Methods: An educational model was developed. Two peer group discussions (facilitated by a GP/pharmacist team), including individual feedback on the GPs' judgments of written simulated cases and prescribing, were main components. The model was tested in a parallel randomized controlled study including 36 GP groups, allocated to education on asthma or urinary tract infections. Background and outcome data were knowledge and attitudes (K/A) assessed by a questionnaire and prescribing practices for actual and written simulated cases. The GPs' evaluation of the model was captured through a questionnaire.

Results: All 36 groups completed the program. The mean participation rate in the group discussions was 75%. The response rates were 82–98% regarding background, 60–83% regarding outcome data (K/A questionnaire and written cases), and 80% regarding the evaluation questionnaire. Prescribing data were captured for 99% of the GPs. Both group discussions were considered important by 84–89%. Eighty-seven percent wished to take part in similar CME activities for other conditions. About 80% reported that their purpose in participating had been fulfilled.

Conclusions: It was feasible to evaluate the developed educational model by using a two-armed parallel study design. The model was well received by the participants.

Objectives: To examine the distribution of interventional cardiac catheterization laboratories, their case load, the time trends, and the regional variation of percutaneous transluminal cutaneous angioplasty (PTCA) utilization in Italy.

Methods: Analysis of data was provided by the annual reports of the Italian Group of Studies and Interventional Cardiology over the period from 1983 to 1996.

Results: The number of PTCA facilities and their use steadily increased, mainly in the North. In 1996 the utilization rate was 34 per 100,000 population, but only 60% of labs performed 200 or more procedures.

Conclusions: Dramatic time trends and regional variations often took place without an epidemiology and technology assessment-based planning process.

Objectives: The purpose of this review was to examine whether studies from the medical literature focusing on efficiency of diagnostic facilities reported economic evaluation methods appropriately, following guidelines for conducting and reporting economic evaluations.

Methods: A Medline search was conducted, and studies that concerned a diagnostic technology and fulfilled the Drummond criteria were selected for methodological review. The reliability of selection and methodological review based on the abstracts was determined by scoring a random sample of both abstracts and full articles. lnterrater reliability was determined by scoring a random sample of abstracts by both authors. Kappa values were calculated. Nine methodological aspects were reviewed: study design, the type of economic evaluation, the comparison made, the study's perspective, the cost-effectiveness ratio used, the definition of cost-effective, the types of costs analyzed, the cost calculation method, and the use of sensitivity analysis.

Results: Two hundred fifty studies published between 1992 and 1997 were found regarding efficiency of diagnosticfacilities; 134 studiesfulfilled the Drummond criteriaand were selected for methodological review. Kappavalues showed reliability of selection and methodological review and interrater reliability. The existing literatue on the economic evaluation of diagnostic facilities does not adhere well to guidelines for economic evaluation. In 95%, no perspective was mentioned, in 50% of the cases no ratio was given, in 82% the cost calculation method was not mentioned, and in 66% no sensitivity analysis was reported.

Conclusions: Our review suggests that to improve the quality of reporting economic evaluations, editorial boards could issue and enforce guidelines for standard reporting of such studies.

Objectives: To examine the influence of a picture archiving and communication system (PACS) on the length of stay (LOS) for patients receiving total hip replacement (THR) or total knee replacement (TKR) procedures.

Methods: A before-and-after design was used. Data were collected on all THR and TKR procedures at Hammersmith Hospital from 1993–96. A regression approach was used to examine the influence of PACS on LOS. Factors such as patient age, sex, and physician were controlled for.

Results: Type of admission and discharge, month of procedure, complications, and number of procedures all significantly influenced LOS for patients undergoing THR. For patients receiving TKR, age, sex, admission, prosthetic complications, number of procedures, and PACS significantly influenced LOS.

Conclusions: While this study shows an apparent reduction of 25% in the average LOS for TKR patients at the time PACS was introduced, this is unlikely to be a true PACS effect and no similar reduction in LOS was shown for THR patients.

Recently, many disease management programs, especially for patients with chronic diseases, have emerged. This paper discusses the potential benefits and disadvantages of disease management, on the basis of an extensive literature review. Disease management is an innovative technology in health care management, which is diffusing throughout the health care system without critical evaluation. Evidence on its effectiveness and costs is still very scarce, while the legal, ethical, organizational, and social implications of this practice have not been analyzed seriously. Before disease management is implemented on a broader scale in different European settings, first, empirical evidence about its alleged benefits and cost-effectiveness should be collected.

Objectives: To examine the economic efficiency of current cochlear implant technology under Australian conditions in profoundly deaf adults, partially deafened adults, and children

Methods:> Cost—utility study, with weights based on judgments from persons experienced with the technology, and cost data from Australian sources.

Results: Quality—of— improvements due to functional consequences of hearing improvement were greater than those due to amelioration of hearing disability. Costs in Australian dollars per QALY (15—year assessment) ranged from $5,070—$11,100 for children, $11,790—$38,150 for profoundly deaf adults, and $14,410—$41,000 for partially deaf adults.

Conclusions: Results suggest cochlear implantation is acceptable value for money when compared with other health programs to which resources are committed in Australia.

Objective: To develop a simulation model to project costs, life expectancy, and cost-effectiveness in discounted dollars per quality-adjusted life-year (QALY) saved for clinical strategies to prevent Mycobacterium avium complex (MAC) in patients with AIDS.

Methods: We used natural history data from the Multicenter AIDS Cohort Study, efficacy and toxicity data from randomized clinical trials, and cost data from the AIDS Cost and Services Utilization Survey. The model permits timing of prophylaxis to be stratified by CD4 count (201–300, 101–200, 51–100, and ≤ 50/mm3), and allows combinations of prophylaxis, crossover to second- and third-line agents for toxicity, and consideration of adherence, resistance, and quality of life.

Results: The model projects that the average HIV-infected patient with a beginning CD4 count between 201 and 300/mm3 has total lifetime costs of approximately $43,150 and a quality-adjusted life expectancy of 42.35 months. If azithromycin prophylaxis for M. avium complex is begun after the CD4 declines to 50/mm3, costs and quality-adjusted survival increase to approximately $44,040 and 42.78 months, respectively, for an incremental cost-effectiveness ratio of $25,000/QALY compared with no M. avium complex prophylaxis. Other prophylaxis options (i.e., rifabutin, clarithromycin, and combination therapies) either cost more but offer shorter survival, or have cost-effectiveness ratios above $260,000/QALY. Sensitivity analysis reveals that, for reasonable assumptions about quality of life, risk of infection, prophylaxis cost, adherence, and resistance, azithromycin remains the most cost-effective prophylaxis option.

Conclusions: Azithromycin prophylaxis, begun after the CD4 count has declined to 50/mm3, is the most cost-effective M. avium complex prophylaxis strategy. Consistent with new United States Public Health Service guidelines, it should be the first-line prophylaxis option.

Objectives: To assess the economics of patient-controlled analgesia (PCA) treatment versus regular intramuscular (IM) injections of opioid analgesia for pain management after hysterectomy.

Methods: Cost-minimization analysis was used based on the comparable pain control results achieved in the two treatment groups. Observations were taken of treatment-related events with personnel (mostly nursing) time implications during the trial. Times were then associated with these events in an independent study of personnel activity. Costs were linked by using average wage rates for the various personnel for the Montreal area during the time of the study. Drug and material costs were hospital acquisition costs for all items. The cost of the PCA pump itself was not included in the analysis. Several analyses were performed to test the sensitivity of the results to various assumptions.

Results: The results for total costs of the two therapies generally showed PCA to be more costly than regular IM injections despite no costs of the pump being included in the analyses. These results were robust with respect to changes in assumptions. Even when intentionally biasing the analysis against IM therapy, it was difficult to obtain results that favored PCA.

Conclusions: Based upon the institutions and assumptions in this analysis, PCA offers no cost advantages over regular IM therapy in the pain management after hysterectomy. Regular IM injections provided less costly analgesia.

Objectives: We explore the policy implications of probabilistic sensitivity analysis in cost-effectiveness analysis by applying simulation methods to a decision model.

Methods: We present the multiway sensitivity analysis results of a study of the cost-effectiveness of vaccination against pneumococcal bacteremia in the elderly. We then execute a probabilistic sensitivity analysis of the cost-effectiveness ratio by specifying posterior distributions for the uncertain parameters in our decision analysis model. In order to estimate probability intervals, we rank the numerical values of the simulated incremental cost-effectiveness ratios (ICERs) to take into account preferences along the cost-effectiveness plane.

Results: The 95% probability intervals for the ICER were generally much narrower than the difference between the best case and worst case results from a multiway sensitivity analysis. Although the multiway sensitivity analysis had indicated that, in the worst case, vaccination in the 85 and older age group was not acceptable from a policy standpoint, probabilistic methods indicated that the cost-effectiveness of vaccination was below $50,000 per quality-adjusted life-year in greater than 92% of the simulations and below $100,000 in greater than 95% of the simulations.

Conclusions: Probabilistic methods can supplement multiway sensitivity analyses to provide a more comprehensive picture of the uncertainty associated with cost-effectiveness ratios and thereby inform policy decisions.

Objectives: Knowledge of resource use and costs can be useful when evaluating existing services or planning new services. This study investigates the use of health care and social services during the first year after a stroke. Total costs are calculated, costs are compared across subgroups of patients, and resource items of major importance for the total costs are identified.

Methods: The study is based on a database comprising data on all stroke patients admitted to a university hospital in Copenhagen, Denmark, over a 1-year period, 1994–95. Patients were followed for 1 year after the stroke, and data on resource use during and after hospitalization were collected prospectively at interviews. This paper focuses on a subset of 385 patients who were admitted because of cerebral infarct or unspecified stroke.

Results: The mean cost, based on all patients, of health care and social services during the first year was 142,900 DKK (US $25,500). The hospital care until the first discharge, including acute care and rehabilitation, cost 101,600 Danish krones (DKK) (US $18,100), i.e., 71% of the total cost. Major resource items after discharge were nursing homes, readmissions, outpatient rehabilitation, and home help. The cost during the first year varied with a number of factors, with the most important being survival and degree of disability. Patients who survived the acute phase and who had severe disability (Barthel Activities of Daily Living [ADL] Index: 0–9) 7–10 days after admission had a total cost during the first year that was five times as high as patients with no disability (Barthel ADL Index: 20).

Conclusion: Costs of health care and social services during the first year after a stroke vary considerably. Disability as measured with the Barthel ADL Index is a stronger predictor of costs than Scandinavian Stroke Scale scores and other clinical and demographic variables.

This paper examines the rationality of the concepts underlying evidence—based medicineand health technology assessment (HTA), which are part of a new current aimed at promoting the use of the results of scientific studies for decision making in health care. It describes the different approaches and purposes of this worldwide movement, in relation to clinical decision making, through a summarized set of specific HTA case studies from Catalonia, Spain. The examples illustrate how the systematic process of HTA can help in several types of uncertainties related to clinical decision making.

In this paper, we discuss the complex relationship between health technology assessment (HTA) and the regulation of medical devices and procedures. The relationship is first examined through a conceptual framework describing the itinerary from research to three levels of policy making: micro (standards of medical practice), meso (institutional rules), and macro (health policies). Four reports from the Quebec Health Technology Assessment Council (CÉTS) are used to illustrate how HTA activities can influence the regulatory mechanisms operating at each decision-making level. We then discuss the skillful balancing act required from HTA agencies to constantly negotiate the right distance from the regulatory process at which to operate. We propose that HTA agencies should not be incorporated into any regulatory, auditing, or monitoring process. Finally, the relationship between health technology assessment and health care reform is discussed. It is suggested that HTA activities will contribute most during the data-driven preparation and consolidation phases of a reform process. The fast pace of events and the political turmoil characteristic of the implementation phase provide a less receptive environment for HTA contributions.

Objective: To assess the impact of guidelines on drug use issued by a consensus conference on polycythemia vera held in Paris in June 1993. 32Phosphorus (32P) was recommended for patients over 70 and/or at risk, whereas pipobroman and hydroxyurea were recommended for patients under 70.

Methods: A questionnaire was sent to all 119 departments of nuclear medicine in France 1 year after the conference to find out whether and how often they measured plasma volume and red cell mass (the recommended diagnostic tests for polycythemia vera). Time-series analyses were performed on sales of 32P, pipobroman (both virtually exclusively prescribed for polycythemia), and hydroxyurea over a 4-year span (January 1992–December 1995).

Results: The average number of plasma volume determinations per year did not change significantly after the conference (22 ± 26 before vs 21 ± 25 after). 32P and pipobroman sales were stable until July 1993, when 32P sales decreased while pipobroman sales rose steadily. Hydroxyurea sales increased over the whole period with no change in trend after the guidelines were published.

Conclusions: The guidelines apparently influenced clinical practice since sales of drugs that are specifically used to treat polycythemia vera showed clear changes in trend after publication of the guidelines. This type of study seems to be an effective way of assessing the impact of consensus conferences.

Objective: To compare different methods to estimate the confidence interval of the incremental cost-effectiveness ratio (ICER).

Methods: The adequacy of Fieller intervals and three methods for calculating bootstrap intervals are compared based on a simulation of 10,000 trials, using data from one trial.

Results: Both Fieller and bootstrap methods lead to unsatisfactory results when the difference in effectiveness is approximately zero. Where this difference is significant, the four methods for calculating confidence intervals for ICER do not give very different results, but Fieller's interval performs best.

Conclusions: Since Fieller's confidence limits are relatively easy to compute compared with bootstrap simulations, we recommend using this method.