Introduction: Many health technologies exhibit some form of learning effect, and this represents a barrier to rigorous assessment. It has been shown that the statistical methods used are relatively crude. Methods to describe learning curves in fields outside medicine, for example, psychology and engineering, may be better.

Methods: To systematically search non–health technology assessment literature (for example, PsycLit and Econlit databases) to identify novel statistical techniques applied to learning curves.

Results: The search retrieved 9,431 abstracts for assessment, of which 18 used a statistical technique for analyzing learning effects that had not previously been identified in the clinical literature. The newly identified methods were combined with those previously used in health technology assessment, and categorized into four groups of increasing complexity: a) exploratory data analysis; b) simple data analysis; c) complex data analysis; and d) generic methods. All the complex structured data techniques for analyzing learning effects were identified in the nonclinical literature, and these emphasized the importance of estimating intra- and interindividual learning effects.

Conclusion: A good dividend of more sophisticated methods was obtained by searching in nonclinical fields. These methods now require formal testing on health technology data sets.

Objective: We studied the influence of the methodologic quality of individual trials on the outcome of a landmark meta-analysis on thrombolytic therapy in acute myocardial infarction. From each study we extracted the number of patients in both groups who died in hospital or during follow-up. Methodologic quality was assessed using the Delphi list. We first recalculated pooled odds ratios (ORs) and 95% confidence intervals (CIs), on the studies found and compared them with the original results of Yusuf et al. Next we incorporated the results of quality assessment in five different ways in the calculation of the pooled ORs: a) component analysis; b) visual plot; c) quality score as a threshold score; d) quality score as a weighting factor; and e) cumulative pooling.

Results and Conclusion: No correlation between quality scores and ORs was found. Studies with a proper description of the different quality components provided an estimate close to the true treatment effect. No major differences were found between the results of the five different methods of incorporating the quality scores into the final conclusion.

Objectives: The aim was to determine which generations of the evolving technology of magnetic resonance angiography (MRA) are currently of clinical relevance in two clinical applications. Our purpose was to plan a systematic review that would be valuable both to purchasers driven by cost-effectiveness and to practicing clinicians.

Methods: Information was gathered from a search of major bibliographic databases, from a short questionnaire sent to 500 U.K. vascular radiologists and vascular surgeons, and from local clinical experts. We asked which of the MRA techniques were currently used and, assuming availability, what would be their technique of choice.

Results: There were 206 published articles that satisfied preliminary inclusion criteria: 69 discussed 2D time of flight (TOF); 47, 3D TOF; and 38, contrast-enhanced techniques. There were 162 questionnaires returned (60 radiologists, 102 surgeons). Of the total respondents, 77/162 (48%) used MRA in the assessment of carotid artery stenosis; 47/77 (61%) used 2D TOF; 32/77 (42%), 3D TOF; and 26/77 (34%), contrast-enhanced techniques. Thirty-five of 162 (22%) respondents used MRA in the assessment of peripheral vascular disease (PVD); 15/35 (43%) used 2D TOF, 4/35 (11%) used 3D TOF, and 22/35 (63%) used contrast-enhanced techniques. For those wishing to use MRA, contrast-enhanced techniques were the method of choice.

Conclusions: The TOF methods that represent earlier generations of the technology remain clinically relevant, and will therefore be included in our systematic review. To ensure complete and relevant coverage in reviews of other evolving technologies, it would be advisable to obtain data for guidance in a similar way.

Objectives: Although psychiatric comorbidity is relevant for a number of diseases, it is often ignored in technology assessment. This study examines the service use rate in mental healthcare facilities and related costs for stroke patients discharged from the University Hospital Maastricht between 1987 and 1995.

Methods: Through anonymous record linkage, the medical registration of the hospital and the registration of the Maastricht Mental Health Case Register were linked.

Results: Linkage succeeded for 16% of the 2,020 stroke patients, indicating that these patients used mental health services during a 10-year period around the stroke (±5 years). Of the users' group, 88% had a mental healthcare contact following stroke. Regression analysis shows that age, length of hospital stay, and mental healthcare contact before stroke are associated with mental healthcare use after stroke. It is remarkable in that there is already an increase in the consumption of mental health care in the prodromal phase just before the stroke occurred. When comparing costs before and after stroke, the outpatient costs increased on average by \epsfbox{cj108-4-euro.eps}42.64, semi-institutionalized costs increased on average by \epsfbox{cj108-4-euro.eps}208.10, and intramural costs by \epsfbox{cj108-4-euro.eps}1,189.21. The total increase in costs is \epsfbox{cj108-4-euro.eps}1,439.95. For all mental healthcare facilities, the increase in costs is significant.

Conclusions: No study so far has revealed the total costs of mental healthcare facilities following stroke. Extrapolating these costs to the Netherlands illustrates that stroke patients have a high psychiatric comorbidity, inducing about 1.3% of total mental healthcare costs.

Objectives: The great interest focused on home care technologies during the last decade resulted from its potential to cut costs. However, the reallocation of costs between healthcare providers and social welfare providers, and the indirect costs of informal care, are not as frequent topics of discussion. The aim of this paper is to discuss different models for estimating the costs of informal care in the home care setting in economic appraisals.

Methods: The outcome of using different models for estimating indirect costs was illustrated using empirical data regarding the time spent by informal caregivers in providing care in a group of home care patients (n = 59). The models used comprise different interpretations of the traditional human capital approach and the friction cost model.

Results and Conclusions: Informal care is an important component in home care. The inclusion of indirect costs of informal care in economic appraisals will have implications for the cost-effectiveness of home care, since it will raise costs depending on the model used for estimating indirect costs. In this study we have shown that indirect costs estimated by the friction cost model only amount to 18% to 44% of the cost when the human capital approach is used. The results indicate that, regardless of the method used to estimate indirect costs, the cost of informal care in evaluations of home care programs is often underestimated due to the exclusion of indirect costs.

Objective: To characterize the net cost and quality of life effects associated with hip protector use among the elderly relative to no intervention.

Methods: We developed two deterministic hypothetical cohorts of 500,000 65-year-old women or men and followed them over the remainder of their lifetimes. Data inputs were collected from secondary sources. Net costs are expressed in 1999 U.S. dollars, whereas net effectiveness is expressed in lives saved and in quality-adjusted life-years (QALYs) gained.

Results: Hip protector use results in net cost savings for both elderly women and men. Women over age 65 and men over age 85 also gain QALYs through hip protector use.

Conclusions: Using the estimates available in the literature, our analysis indicates that use of hip protectors among women is associated with large cost savings and QALY gains even when accounting for the inconvenience of using the protectors. Among men, hip protectors are also associated with cost savings (although of smaller magnitude), but there are net losses of QALYs because of the inconveniences associated with the protectors.

Objective: Familial adenomatous polyposis (FAP) is a well-known hereditary colorectal cancer-predisposing syndrome. Genetic testing for colorectal cancer risk is now part of standard medical practice, but very little is known about the economic impact of this technology. The aim of this study was to assess, from a healthcare system perspective, the direct costs of two strategies for screening at-risk relatives of FAP patients: clinical screening versus genetic testing for FAP.

Methods: A systematic review of the literature was carried out. Additional information was gathered from experts in research and clinical laboratories and in hospital departments. A decision tree was constructed to compare per-person and per-family costs of the two strategies for screening at-risk relatives of FAP patients. Sensitivity analysis was performed to assess the stability of the model across the full range of plausible values for all key parameters.

Results: According to the decision analysis, with FAP screening starting at puberty, the average screening costs are $3,181 and $2,259 (Canadian dollars), respectively, for the clinical screening and the genetic testing strategies. Genetic screening is cost saving up to a first screening age of 36. Sensitivity analysis shows that the results of the baseline analysis hold across a variety of assumptions concerning the parameter values.

Conclusions: The genetic testing strategy is cost saving relative to the clinical screening alternative. Apart from its lower costs, it is associated with many other benefits. Accordingly, under predefined conditions, predictive genetic testing seems to be the optimal screening strategy for FAP.

Objectives: To evaluate the cost-effectiveness of reusing hemodialyzers for patients with kidney failure on dialysis employing either a heated citric acid or formaldehyde sterilization method, in comparison to the standard practice of single-use dialysis.

Methods: A meta-analysis of all relevant studies was performed to determine whether hemodialyzer reuse was associated with an increased relative risk of mortality or hospitalization. A decision tree was constructed to model the effect of three different dialysis strategies (single-use dialysis, heated citric acid, and formaldehyde dialyzer reuse) on the costs and quality-adjusted life expectancy of “typical” hemodialysis patients. The cost of heated citric acid reuse was estimated from a center experienced with the technique. The cost of end-stage renal disease (ESRD) care, survival data, and patient utilities were estimated from published sources.

Results: There was evidence of a higher relative risk of hospitalization (but not mortality) for hemodialyzer reuse compared with single-use dialysis. Depending on the assumptions used, the cost savings that could be expected by switching from single-use dialysis to heated citric acid reuse were small, ranging from CAN $0–739 per patient per year.

Conclusions: ESRD programs can incorporate the results of this study based on their individual situations to determine whether hemodialyzer reuse is appropriate in their setting.

Objectives: Decision analytic modeling allows a rational, feasible, scientific, and timely approach to measure the efficiency of new medical technologies in health care by using the best available evidence of different sources to produce detailed estimates of the clinical and economic consequences of different healthcare interventions. The aim of this article has been twofold: first, to state the benefit, usefulness, and limitations of decision analytic modeling; and second, to propose a list of steps that should be considered when carrying out this sort of analysis in order to increase its credibility and validity for medical decision making.

Methods: After having read the published literature about decision analytic models, a checklist was created, listing the paramount items and points that should always be followed when carrying out disease modeling.

Results: A checklist with 13 items and steps to be taken was drawn up to ensure that all important methodologic aspects are considered when performing and assessing decision analytic modeling, thereby constituting a standard guide to good modeling practices. In addition, a set of limitations inherent to the use of decision analytic models is also contemplated, along with a list of recommendations increasing their validity and utility.

Conclusions: Health economic assessments will continue to be an important tool for the correct allocation of resources, and disease analytic modeling will be a potentially invaluable tool in assisting decision makers in this task.

Interest is growing in the application of standard statistical inferential techniques to the calculation of cost-effectiveness ratios (CER), but individual level data will not be available in many cases because it is very difficult to undertake prospective controlled trials of many public health interventions. We propose the application of probabilistic uncertainty analysis using Monte Carlo simulations, in combination with nonparametric bootstrapping techniques where appropriate. This paper also discusses how decision makers should interpret the CER of interventions where uncertainty intervals overlap. We show how the incorporation of uncertainty around costs and effects of interventions into a stochastic league table provides additional information to decision makers for priority setting. Stochastic league tables inform decision makers about the probability that a specific intervention would be included in the optimal mix of interventions for different resource levels, given the uncertainty surrounding the interventions.

Objective: The role of classical and Bayesian statistical approaches remains in dispute in health services research and policy. The goal of this study was to determine if results differ when both analytic techniques are used with the same data set.

Design: We searched MEDLINE and related databases for English-language articles published January 1, 1978 through August 31, 1999. We combined Bayesian and classical statistics search terms and their variants with randomized control trials (RCTs) and meta-analyses.

Results: Searches found 18 studies in 14 publications that met all review criteria—nine RCTs, eight meta-analyses, and one epidemiologic estimate. Statistical analyses using both methods agreed in five RCTs , four meta-analyses, and for the epidemiologic estimates. For four RCTs where results disagreed, classical analysis found the experimental intervention was efficacious compared with the control, and Bayesian reanalysis concluded the intervention was not proven efficacious. Classical meta-analyses of the four studies where results disagreed concluded the experimental intervention was not better than the control; Bayesian reanalysis concluded it was efficacious.

Conclusion: Classical and Bayesian methods in this review exhibited important divergence of results. Disagreement on many fundamental beliefs between classical and Bayesian statistics means continuing debate. One way to resolve this debate is for proponents of each technique to decide together the circumstances for use of each method and analytic framework. If the experts do not agree on the methodologic requirements, other decision makers likely will force their own views.

Objectives: Different cost-effectiveness analyses have been presented for interferon beta-1b (IFNB) in the treatment of secondary progressive multiple sclerosis (SPMS). All studies have used modeling techniques since any effect on progression of disability achieved during a clinical trial will last beyond the trial. Different approaches to extrapolation have been taken, but generally they have been based on disease progression and relapse rates in clinical trials. The problem with this approach is that the population in clinical trials is a selected group of patients, which has the potential to bias results. A better method for extrapolation is to use epidemiologic data. The objective of this study is to incorporate natural history data for MS into a previously presented cost-utility model and to compare the two methods of extrapolation.

Methods: Clinical trial data were used to model disease progression during the first 3 years in the model. To extrapolate beyond the trial (10 years), data on progression of disability were available from a geographically based epidemiologic study of the natural history of MS in Canada. There were 568 patients who had converted to SPMS during the follow-up that were included in the data set. Mean costs and utilities for each Markov state were calculated from a population-based cross-sectional study in Sweden.

Results: The extrapolation using clinical trial data appears to have underestimated the progression of disability in the long term and thus also the potential benefit of treatment. Using the epidemiologic data, the incremental cost per QALY is SEK 257,000 (US $25,700; US $1 = SEK 10; November 2000) when all costs (direct, informal care, and indirect) are included (discounted 3%). This compares to SEK 342,000 in the previous model. The lower cost-effectiveness ratio is mostly due to a larger QALY gain with treatment than in the previous model (0.217 compared with 0.162).

Conclusion: Cost-effectiveness analysis in SPMS requires that the effect of treatment beyond clinical trials be included. The method of extrapolation clearly affects the results, and when available, epidemiologic data should be used. Using the longitudinal data from Canada, the cost-utility ratios for IFNB-1b in the treatment of SPMS appear well within the acceptable range.

This paper investigates the structural reliability of a discrete choice experiment within health technology assessment. Two versions of a discrete choice experiment, in the form of a self-completion questionnaire, were randomly administered to two samples of women who had recently given birth as part of an exercise to determine women's preferences for alternative modes of intrapartum care. In the first questionnaire, two of the attributes had only their highest and lowest levels included, while in the second questionnaire all three levels for these two attributes were included. The levels included for all other attributes remained the same throughout both questionnaires. The evidence relating to the structural reliability of the discrete choice experiment in this context was mixed. The results indicated that the relative importance of the two attributes in which the levels were varied increased as the number of levels for these attributes increased. However, the relative importance of the attributes in which the levels were not varied remained relatively stable throughout. The results provide evidence in support of a psychological effect whereby respondents place more importance upon specific attributes as the number of levels for these attributes increases. It is recommended that further research of both a qualitative and quantitative nature should be undertaken to assess the potential importance (or otherwise) of a psychological effect relating to the number of levels chosen for attributes within discrete choice experiments in health technology assessment.

Alcohol-related problems are common in society. Large population studies from the United States have shown that 10% to 15% of all men and approximately 5% of all women suffer from chronic alcohol dependency. In these groups, approximately one-fourth are in a phase of active abuse.