Objectives: Analysis of and understanding the utilization patterns of diagnostic imaging technologies is important for planning health systems especially in middle- and low-income countries. Almost all published studies have been focused on utilization trends or utilization rates of magnetic resonance imaging (MRI) over time and little has been published about utilization patterns of MRI. This study aims to identify and describe the utilization pattern of MRI in Iran.

Methods: Data was collected from referrals and MRI reports of the population covered by one of the largest insurance organizations (Social Security Organization [SSO]) with more than 27 million beneficiaries in the Country. We focused on patients who had undergone an MRI examination during 1 month (October 13 to November 12), 2005. The data collected consisted of 1,656 referrals for MR imaging and 1,547 MRI reports.

Results: No clear association was found between the number of referrals per 100,000 population and the number of MRI machines per 100,000 population (r = 0.30; p = .13). Neurosurgeons, orthopedists, and neurologists contributed to more than 88 percent of MRI utilization. The Spinal column (55 percent), brain (25.5 percent), and knee (11.0 percent) were the most common parts of the body scanned by MRI.

Conclusions: Having identified some of the most frequent users of MRI technology, any plan for modifying or improving MRI use should focus mainly on these specialties. International evidence on using the technology for examining brain and spine indicate that MRI is mainly used in diagnosing the central nervous system (CNS) diseases.

Objectives: Of the fourteen counties and two municipalities that until recently were responsible for healthcare provision in Denmark, five introduced mammography screening (MS) programs. The objective of this research is to explain this decision-making variation and to gain insight into priority setting processes in health-care provision at the county level in Denmark.

Methods: Literature on priority setting in health care was used to derive seven explanatory factors for comparing decision making on MS between four selected counties, of which two had implemented MS. The relative importance of each explanatory factor in each county was determined by analyzing policy documents, supplemented with interviews of selected stakeholders. The results were combined and compared at the county level.

Results: Evidence of effectiveness of MS was considered satisfactory and ethical issues related to MS were perceived relatively unproblematic only in those counties that introduced MS. Lack of resources, that is, radiologists, was an additional important factor for counties not implementing MS. Local opinion leaders have played a stimulating role, whereas advisory policy documents at the central government level and even legislation have had a minor impact.

Conclusions: The four counties have based their decision making on the introduction of MS on different combinations of a limited number of factors that have been differentially weighted. The pattern of relevant factors in both counties not introducing MS is rather similar. The study elucidates the role of complementary factors to evidence in decision making. Of interest, recent public sector reforms have resulted in the decision to have MS implemented nationwide.

Objectives: The aim of this study was to assess the cost-effectiveness of screening for open-angle glaucoma (OAG) in the United Kingdom, given that OAG is an important cause of blindness worldwide.

Methods: A Markov model was developed to estimate lifetime costs and benefits of a cohort of patients facing, alternatively, screening or current opportunistic case finding strategies. Strategies, varying in how screening would be organized (e.g., invitation for assessment by a glaucoma-trained optometrist [GO] or for simple test assessment by a technician) were developed, and allowed for the progression of OAG and treatment effects. Data inputs were obtained from systematic reviews. Deterministic and probabilistic sensitivity analyses were performed.

Results: Screening was more likely to be cost-effective as prevalence increased, for 40 year olds compared with 60 or 75 year olds, when the re-screening interval was greater (10 years), and for the technician strategy compared with the GO strategy. For each age cohort and at prevalence levels of ≤1 percent, the likelihood that either screening strategy would be more cost-effective than current practice was small. For those 40 years of age, “technician screening” compared with current practice has an incremental cost-effectiveness ratio (ICER) that society might be willing to pay when prevalence is 6 percent to 10 percent and at over 10 percent for 60 year olds. In the United Kingdom, the age specific prevalence of OAG is much lower. Screening by GO, at any age or prevalence level, was not associated with an ICER < £30,000.

Conclusions: Population screening for OAG is unlikely to be cost-effective but could be for specific subgroups at higher risk.

Objectives: The aim of this study was to evaluate the costs and effectiveness associated with no screening, Helicobacter pylori serology screening, and the 13C-urea breath test (UBT) for gastric cancer in the Chinese population.

Methods: A Markov model simulation was carried out in Singaporean Chinese at 40 years of age (n = 478,500) from the perspective of public healthcare providers. The main outcome measures were costs, number of gastric cancer cases prevented, life-years saved, quality-adjusted life-years (QALYs) gained from the screening age to death, and incremental cost-effectiveness ratios (ICERs), which were compared among the three strategies. The uncertainty surrounding ICERs was addressed by scenario analyses and probabilistic sensitivity analysis using Monte Carlo simulation.

Results: The ICER of serology screening versus no screening was $25,881 per QALY gained (95 percent confidence interval (95 percent CI), $5,700 to $120,000). The ICER of UBT versus no screening was $53,602 per QALY gained (95 percent CI, $16,000 to $230,000). ICER of UBT versus serology screening was $470,000 per QALY gained, for which almost all random samples of the ICERs distributed above $50,000 per QALY.

Conclusions: It cannot be confidently concluded that either H pylori screening was a cost-effective strategy compared with no screening in all Chinese at the age of 40 years. Nevertheless, serology screening has demonstrated much more potential to be a cost-effective strategy, especially in the population with higher gastric cancer prevalence.

Objectives: Medical device companies are under growing pressure to provide health-economic evaluations of their products. Cost-effectiveness analyses are commonly undertaken as a one-off exercise at the late stage of development of new technologies; however, the benefits of an iterative use of economic evaluation during the development process of new products have been acknowledged in the literature. Furthermore, the use of Bayesian methods within health technology assessment has been shown to be of particular value in the dynamic framework of technology appraisal when new information becomes available in the life cycle of technologies.

Methods: In this study, we set out a methodology to adapt these methods for their application to directly support investment decisions in a commercial setting from early stages of the development of new medical devices.

Results and Conclusions: Starting with relatively simple analysis from the very early development phase and proceeding to greater depth of analysis at later stages, a Bayesian approach facilitates the incorporation of all available evidence and would help companies to make better informed choices at each decision point.

Objectives: Positron emission tomography (PET) is an innovative imaging tool. Associated with computed tomography (CT), it allows a better definition for the tumor volume for radiotherapy, compared with CT only. The aim of this study was to assess the effects of PET on resource allocation (costs and savings) and on the choice of the following treatment in radiotherapy.

Methods: In 2004 and 2005, 209 patients were enrolled (97 patients with Hodgkin's disease and 112 with non-small cell lung cancer) in a national study conducted in eight hospitals. Two treatment decisions made on the basis of CT only or CT associated with PET, were compared in a prospective study where each subject was his/her own control. The direct medical cost of using PET was assessed by microcosting, using data collected from specific questionnaires. The costs of new tests and the costs and savings associated with changes in the chosen treatment were calculated on the basis of reimbursement rates.

Results: The mean cost of using PET was approximately €800 per patient (50 percent for the radionuclide 18F-FDG [2-[18F]fluoro-2-deoxy-D-glucose]). Radiotherapy treatments were modified for 10 percent of patients with Hodgkin's disease versus 40 percent of patients with lung cancer. Overall, the use of PET induced both increases and decreases in the mean cost per patient: the net effect was a €425 and €931 cost increase in lung cancer and Hodgkin's disease, respectively.

Conclusions: The use of PET for radiotherapy decision making seems more valuable for lung cancer than for Hodgkin's disease, both in terms of costs and changes in radiotherapy treatment. This result might help policy makers for prioritization.

Objectives: The aims of this study were to estimate preferences and willingness-to-pay (WTP) for genetic screening for CYP2D6 polymorphisms among a group of former and currently depressed patients.

Methods: A Web-based discrete choice questionnaire was sent to 89 respondents, age 18–65. Four attributes were included: (i) shifts in antidepressant medication before symptom relief, (ii) time with antidepressant medication without symptom relief, (iii) time with antidepressant medication without symptoms but with adverse side-effects, (iv) cost of genetic screening. We used a switching model with two scenarios, one representing patients’ own treatment history and the other a treatment scenario with genetic screening.

Results: In a main-effects model involving the four attributes all coefficients had the expected sign, indicating that as the number of shifts, price or time without symptom relief, and/or dosage-adjustments increased, the likelihood of choosing the screening test decreased. Price and number of shifts in medicine were significant. Marginal WTP for 5 percent probability of a reduction of one in antidepressant shifts was DKK2,599 (€350).

Conclusions: Patients value reductions in shifts in antidepressants and price when choosing between genetic screening and no screening. They do not focus on how the reductions are provided, nor do they value the genetic information the test provides irrespective of its effect on outcome. Given, that the test is able to provide a reduction of one shift in the number of antidepressant shifts with a probability of 5 percent, WTP for the test exceeds its cost.

Objectives: To catalogue and comparatively assess the quality of Clinical Practice Guidelines (CPG) for ischemic stroke taking into account format and development methodology.

Methods: We performed a comprehensive, systematic bibliographic search of CPGs addressing the management of ischemic stroke. We designed a sensitive strategy, using methodological filters in the following databases: Medline, IME and Lilacs, National Guidelines Clearinghouse, National electronic Library for Health, NICE, Guidelines International Network (GIN), Canadian Medical Association Infobase, development groups such as Scottish Intercollegiate Guidelines Network (SIGN), New Zealand Guidelines Group (NZGG), Agency for Healthcare Research and Quality (AHRQ), Ministry of Health Singapore, Institute for Clinical Systems Improvement (ICSI); and scientific societies: American Heart Association, American Medical Association, Royal College of Physicians London. We included all CPGs published in English, French, Italian, Portuguese, or Spanish from 1999 to 2005 and excluded those CPGs whose scope was primary prevention and rehabilitation from ischemic stroke. Four researchers independently assessed the structure and methodologies followed in drafting the CPGs using the Changing Professional Practice (CPP) and Appraisal of Guidelines Research & Evaluation (AGREE) instruments.

Results: We retrieved 117 documents; following application of exclusion criteria, twenty-seven CPGs were appraised. With regard to methodological quality (using the AGREE instrument), the domains that scored highest were “Scope and purpose” and “Clarity and presentation.” The lowest scoring domains were “Stakeholder involvement,” “Rigor of development,” and “Applicability.” Most guidelines received an overall score of “would not recommend” (77.8 percent). Finally, based on the CPP instrument, most of the CPGs evaluated were aimed at secondary care and did not provide updating procedures.

Conclusions: The overall quality of the CPGs published for ischemic stroke management did not have minimum methodological quality. Quality improvement has been observed in more recent CPGs and may be due to the publication of new tools such as the AGREE or CPP instruments, as well as international initiatives for CPG improvement.

Objectives: The aim of this prospective study was to perform a cost and outcome comparison between two alternative operative techniques (osteosynthesis and hemiarthroplasty) used in the treatment of elderly patients with unstable trochanteric hip fracture.

Materials and Methods: One hundred seventy-three trochanteric hip fracture patients were followed-up for 1 year after surgery. For each operative technique, hospital treatment's cost per patient was computed. Mortality and complication rate in-hospital and at specific time points after surgery were used as outcome measures. Patients' functional level before and after hip fracture was estimated according to their mobility and ability to perform basic and instrumental activities of daily living.

Results: The cost for patients undergoing osteosynthesis reached €1,931 per case, whereas for those treated with hemiarthroplasty reached €3,719 per case (2001 rates). There was no statistically significant difference regarding in-hospital mortality and complication rate, as well as mortality and complication rate 1 year after surgery, between the two patient groups.

Conclusions: The quite similar performance of the two operative techniques suggests that cost could be the key factor for choosing between them. However, it is critical that many more randomized studies, with larger sample sizes and wider follow-up time periods should be conducted.

Background: Economic evaluation as an integral part of health technology assessment is today mostly applied to established technologies. Evaluating healthcare innovations in their early states of development has recently attracted attention. Although it offers several benefits, it also holds methodological challenges.

Objectives: The aim of our study was to investigate the possible contributions of economic evaluation to industry's decision making early in product development and to confront the results with the actual use of early data in economic assessments.

Methods: We conducted a literature research to detect methodological contributions as well as economic evaluations that used data from early phases of product development.

Results: Economic analysis can be beneficially used in early phases of product development for various purposes including early market assessment, R&D portfolio management, and first estimations of pricing and reimbursement scenarios. Analytical tools available for these purposes have been identified. Numerous empirical works were detected, but most do not disclose any concrete decision context and could not be directly matched with the suggested applications.

Conclusions: Industry can benefit from starting economic evaluation early in product development in several ways. Empirical evidence suggests that there is still potential left unused.

Objectives: This report is a scoping review of the literature with the objective of identifying definitions, conceptual models and frameworks, as well as the methods and range of perspectives, for determining appropriateness in the context of healthcare delivery.

Methods: To lay groundwork for future, intervention-specific research on appropriateness, this work was carried out as a scoping review of published literature since 1966. Two reviewers, with two screens using inclusion/exclusion criteria based on the objective, focused the research and articles chosen for review.

Results: The first screen examined 2,829 abstracts/titles, with the second screen examining 124 full articles, leaving 37 articles deemed highly relevant for data extraction and interpretation. Appropriateness is defined largely in terms of net clinical benefit to the average patient and varies by service and setting. The most widely used method to assess appropriateness of healthcare services is the RAND/UCLA Model. There are many related concepts such as medical necessity and small-areas variation.

Conclusions: A broader approach to determining appropriateness for healthcare interventions is possible and would involve clinical, patient and societal perspectives.

Objectives: The aims of this study were, first, to define the main advantages and disadvantages of using bibliographic search protocols; second, to define a series of criteria that could aid in prioritizing the information resources to be consulted for each research project; and third, to rank these criteria.

Methods: First, a survey was e-mailed to the Spanish Health Technology Assessment Agencies (AUnETS) group, with the aim of evaluating the usefulness of using bibliographic search protocols. Second, a consensus group meeting with the information specialists of the AUnETS group was organized, where SWOT analysis technique (strengths, weaknesses, opportunities, and threats) was used, also to discuss the utility of using search protocols. Third, the same group designed a final prioritization criteria questionnaire intended for Health Technology Assessment International's Information Resources Group (HTAi IRG), based on a draft version written by the information specialist from the Basque Office for HTA. Finally, this questionnaire was e-mailed to the HTAi IRG experts, and their responses were analyzed.

Results: Some of the advantages defined were systematization of the searches and transparency and repeatability of the process. The perceived disadvantages were inflexibility to be adapted to some situations, inability to establish time frameworks and the difficulty of incorporating experts’ opinions into closed protocols. Five areas of prioritization criteria were defined: study topic; characteristics of the database and other information resources; database interface; characteristics of the organization; kind of research output for which the information is intended. A ranked list of prioritization criteria was established based on the responses of the HTAi IRG group.

Conclusions: The information specialists consulted agreed that search protocols are useful tools for guiding systematic searches. The twenty-one prioritization criteria defined will be used by the information specialists for determining, which databases to consult.

Objectives: The aim of this study was to estimate the potential cost-effectiveness and expected value of perfect information of a recently derived clinical prediction rule for patients presenting to emergency departments with chest discomfort.

Methods: A decision analytic model was constructed to compare the Early Disposition Prediction Rule (EDPR) with the current standard of care. Results were used to calculate the potential cost-effectiveness of the EDPR, as well as the Value of Information in conducting further research. Study subjects were adults presenting with chest discomfort to two urban emergency departments in Vancouver, British Columbia, Canada. The clinical prediction rule identifies patients who are eligible for early discharge within 3 hours of presentation to the emergency department. The outcome measure used was inappropriate emergency department discharge of patients with acute coronary syndrome (ACS).

Results: The incremental cost-effectiveness ratio of the EDPR in comparison to usual care was (negative) $2,999 per inappropriate ACS discharge prevented, indicating a potential cost-savings in introducing the intervention. The expected value of perfect information was $16.3 million in the first year of implementation, suggesting a high benefit from conducting further research to validate the decision rule.

Conclusions: The EDPR is likely to be cost-effective; however, given the high degree of uncertainty in the estimates of costs and patient outcomes, further research is required to inform the decision to implement the intervention. The potential health and monetary benefits of this clinical prediction rule outweigh the costs of doing further research.

Objectives: There are very few descriptions on how familial hypercholesterolemia (FH), a genetic disorder with an increased risk of coronary heart disease (CHD), affects older patients. The aim of this study was to analyze the health-related quality of life (HRQoL) of elderly patients with FH.

Methods: All FH patients with the same FH North Karelia-mutation aged 65 and over living in North Karelia province in Eastern Finland were asked to participate in the study in 2003. They were compared with age-standardized controls from the Finnish general population. A total of forty-three elderly FH patients were identified and thirty-seven of them (aged 65 to 84 years, M/F 10/27) agreed to participate. Most of them (27/37) had been diagnosed with CHD. All but one of these FH patients had been using statins for approximately 15 years. HRQoL was assessed with the RAND-36 (SF-36) and 15D instruments.

Results: Despite the clear cardiovascular morbidity and low socioeconomic status, these FH patients enjoyed a similar HRQoL as the age-standardized controls in the general population.

Conclusions: Lifelong hypercholesterolemia was not associated with poorer HRQoL in elderly survivors of FH. These elderly FH individuals were characterized by their healthy lifestyle and long-term statin treatment.

Objectives: A treatment pathway model was developed to examine the costs and benefits of the current bowel cancer service in England and to evaluate potential alternatives in service provision. To use the pathway model, various parameters and probability distributions had to be specified. They could not all be determined from empirical evidence and, instead, expert opinion was elicited in the form of statistical quantities that gave the required information. The purpose of this study is to describe the procedures used to quantify expert opinion and note examples of good practice contained in the case study.

Methods: The required information was identified and preparatory discussion with four experts refined the questions they would be asked. In individual elicitation sessions they quantified their opinions, mainly in the form of point and interval estimates for specified variables. New methods have been developed for quantifying expert opinion and these were implemented in specialized software that uses interactive graphics. This software was used to elicit opinion about quantities related to measurable covariates.

Results: Assessments for thirty-four quantities were elicited and available checks supported their validity. Eight points of good practice in eliciting and using expert judgment were evident. Parameters and probability distributions needed for the pathway model were determined from the elicited assessments. Simulation results from the pathway model were used to inform policy on bowel cancer service provision.

Conclusions: The study illustrates that quantifying and using expert judgment can be acceptable in real problems of practical importance. For full benefit to be gained from expert knowledge, elicitation must be conducted carefully and should be reported in detail.

Objectives: Although increasing effort is being devoted to developing strategies to increase knowledge transfer and the uptake of health technology assessment (HTA) by various stakeholders, very little is known about the utilization and dissemination of HTA findings by patient and consumer organizations. The goal of this study is to understand how and why patient and consumer organizations use HTA findings within their organizations, and what factors influence how and when they communicate their findings to members or other organizations.

Methods: We examined the use and dissemination of four controversial HTA reports by sixteen patient and consumer organizations in Ontario and Quebec. We gathered data from semistructured interviews conducted between December 2006 and April 2007.

Results: Although HTA findings are often used by the patient and consumer organizations, key differences were observed in exactly how the four HTA reports were used. Three types of use (instrumental, conceptual, and symbolic) are reported and illustrated. We highlight the importance of the organization's mission and knowledge base in explaining the types of use observed.

Conclusions: We contend that the use and dissemination of HTA reports by specific groups could help in widening the debate around controversial health technologies. The implications and opportunities for HTA agencies relate to the following: (i) identification of “lay” organizations that could help in disseminating results; (ii) acknowledgement of a “lay” audience for HTA findings; (iii) strategic inclusion of advocacy groups during the assessment process for highly controversial technologies; and (iv) contribution of these organizations to the push efforts of knowledge transfer.

Objectives: The Turkish healthcare system is currently undergoing reform, and efficient use of resources has become a key factor in determining the allocation of resources. The objective of this study was to analyze strengths, weaknesses, opportunities, and threats (SWOT) in the development of a health technology assessment (HTA) program in Turkey.

Methods: A SWOT analysis was performed using a literature review and interviews with key people in the Turkish Ministry of Health and Ministry of Labor and Social Security.

Results: Regarding recent reforms in health care, investments for information network and databank are the strengths, but the traditional “expert-based” decision making, poor availability of data, and poor quality of data could be seen as some of the weaknesses. Another major weakness is lack of general awareness of HTA. Increasing demand for transparency in decision making, demand for evidence, and demand for credibility by decision makers are some of the opportunities, and current healthcare reforms, i.e., restructuring of healthcare and general health insurance, could also be seen as major opportunities. These opportunities unfortunately could be threatened by lack of funding, and resources are challenged by large, recent national investments.

Conclusions: There is a good opportunity for Turkey to use the skills in HTA currently being developed through activities in Europe and the Americas to assist in the development of a much more cost-effective and transparent healthcare system in Turkey.

Objectives: The goal of meeting the desire of older adults to remain independent in their home setting while controlling healthcare costs has led to the conceptualization of “smart homes.” A smart home is a residence equipped with technology that enhances safety of residents and monitors their health conditions. The study aim is to assess older adults’ perceptions of specific smart home technologies (i.e., a bed sensor, gait monitor, stove sensor, motion sensor, and video sensor).

Methods: The study setting is TigerPlace, a retirement community designed according to the Aging in Place model. Focus group sessions with fourteen residents were conducted to assess perceived advantages and concerns associated with specific applications, and preferences for recipients of sensor-generated information pertaining to residents’ activity levels, sleep patterns and potential emergencies. Sessions were audio-taped; tapes were transcribed, and a content analysis was performed.

Results: A total of fourteen older adults over the age of 65 participated in three focus group sessions Most applications were perceived as useful, and participants would agree to their installation in their own home. Preference for specific sensors related to sensors’ appearance and residents’ own level of frailty and perceived need. Specific concerns about privacy were raised.

Conclusions: The findings indicate an overall positive attitude toward sensor technologies for nonobtrusive monitoring. Researchers and practitioners are called upon to address ethical and technical challenges in this emerging domain.

Objectives: To compare four contingent valuation elicitation methods as a means to estimate the value of a pneumococcal vaccine in Bangladesh and to test if the elicitation methods are subject to bias and if they produce valid responses.

Methods: Three hundred sixty-one households with at least one child under 5 years of age were recruited in Dhaka, Bangladesh. Subjects were cluster-randomized to various elicitation methods: open-ended, dichotomous choice (at one of two asking prices), payment card (one of two cards with differing ranges). The dichotomous choice method was then followed up with a bidding game methodology, with the dichotomous choice price acting as the starting price for the bidding game. Analysis focused on summary statistics, demand curve estimation and multivariate regression analysis to test for validity and bias.

Results: Thirty-one households refused to participate, leaving a total of 330 participating households (a 91.4 percent response rate). Willingness to pay estimates varied significantly across the methods (p < .001), with average estimates varying between $2.34 and $18 (US). The open-ended elicitation method was found to produce highly inflated values that were insensitive to construct validity tests. The dichotomous choice method produced quantity (demand) estimates rather than value estimates, and there was some evidence of yea saying. The payment card elicitation method was found to be affected by range bias. The bidding game elicitation method was found to be less sensitive to starting point bias and yea-saying.

Conclusions: Different elicitation format do give rise to different demand curves; however, this may be partially due to the fact that they do not measure the same outcome. For example, the dichotomous choice format produces a demand curve, while the payment card, open-ended and bidding game produce inverse demand curves. All formats are prone to multiple biases. When choosing an elicitation format, it is important to first consider the purpose and use of the data. Each elicitation method has strengths and weaknesses and can be used for different purposes in technology assessment.

Objectives: Individual patient data (IPD) meta-analyses have been proposed as a major improvement in meta-analytic methods to study subgroup effects. Subgroup effects of conventional and IPD meta-analyses using identical data have not been compared. Our objective is to compare such subgroup effects using the data of six trials (n = 1,643) on the effectiveness of antibiotics in children with acute otitis media (AOM).

Methods: Effects (relative risks, risk differences [RD], and their confidence intervals [CI]) of antibiotics in subgroups of children with AOM resulting from (i) conventional meta-analysis using summary statistics derived from published data (CMA), (ii) two-stage approach to IPD meta-analysis using summary statistics derived from IPD (IPDMA-2), and (iii) one-stage approach to IPD meta-analysis where IPD is pooled into a single data set (IPDMA-1) were compared.

Results: In the conventional meta-analysis, only two of the six studies were included, because only these reported on relevant subgroup effects. The conventional meta-analysis showed larger (age < 2 years) or smaller (age ≥ 2 years) subgroup effects and wider CIs than both IPD meta-analyses (age < 2 years: RDCMA -21 percent, RDIPDMA-1 -16 percent, RDIPDMA-2 -15 percent; age ≥2 years: RDCMA -5 percent, RDIPDMA-1 -11 percent, RDIPDMA-2 -11 percent). The most important reason for these discrepant results is that the two studies included in the conventional meta-analysis reported outcomes that were different both from each other and from the IPD meta-analyses.

Conclusions: This empirical example shows that conventional meta-analyses do not allow proper subgroup analyses, whereas IPD meta-analyses produce more accurate subgroup effects. We also found no differences between the one- and two-stage meta-analytic approaches.