Combining estimates of the odds ratio: the state of the art

Relationship between vaccine hesitancy and vaccination behaviors: Systematic review and meta-analysis of observational studies

BACKGROUND

Vaccination is crucial for prevention of infectious diseases, and identification of the impact of vaccine hesitancy on vaccination programs is crucial for early intervention and formulation of policies to alleviate vaccine hesitancy. The aim of this systematic review was to explore the relationship between vaccine hesitancy and negative vaccination behavior globally.

METHODS

We searched for observational studies in various databases. We conducted a meta-analysis using pooled odds ratios (OR) and 95 % confidence intervals (CI), performed meta regression and subgroup analysis to explore the role factors such as location and individual characteristics on the association between vaccine hesitancy and vaccination behavior.

RESULTS

A total of 46 articles were included in systematic analysis and 34 articles were included in the meta-analysis. The systematic analysis comprised 162,601 samples, whereas the meta-analysis included 147,554 samples. The meta-analysis showed that a higher rate of vaccine hesitancy was associated with an increased likelihood of adverse vaccination behaviors (all adverse behaviors: OR = 1.50, 95 % CI, 1.33-1.70, P < 0.001; unvaccinated: OR = 1.48, 95 % CI, 1.29-1.70, P < 0.001; vaccine delay: OR = 2.61, 95 % CI, 1.97-3.44, P < 0.001). The meta-regression results indicated that the heterogeneity observed was mainly from sample selection methods, age of vaccinees and the health status of participants. The results showed that parents of minor vaccinees or without high-risk health status had a higher association between vaccine hesitancy and vaccine uptake compared with populations exposed to higher health risks or adult vaccinees.

CONCLUSION

The findings provide evidence on the association between vaccine hesitancy and adverse vaccination behaviors. The results showed that these population-specific factors should be considered in future research, and during formulation of interventions and implementation of policies to improve vaccination uptake.

Meta-Analysis of Proportions

The meta-analysis of single proportions has become a popular application over the last two decades. Especially, systematic reviews of prevalence studies are conducted in various fields of science, including medicine, ecology, psychology, or social sciences. In this chapter, we illustrate meta-analysis methods to pool single proportions and to compare proportions from two groups. We introduce classic approaches based on the inverse variance method as well as generalized linear mixed models taking the binary structure of the data into account. The most common transformations of proportions and their back-transformations are described both for individual studies and in the meta-analysis setting.

The Association of Premorbid Metformin Exposure With Mortality and Organ Dysfunction in Sepsis: A Systematic Review and Meta-Analysis

Supplemental Digital Content is available in the text.

To examine the association between premorbid metformin exposure and mortality, hyperlactatemia, and organ dysfunction in sepsis.

A systematic review and meta-analysis of randomized controlled trials comparing 17-alpha-hydroxyprogesterone caproate versus placebo for the prevention of recurrent preterm birth

BACKGROUND

Preterm birth causes an increased risk for perinatal morbidity and mortality.

OBJECTIVE

To determine whether mid-trimester 17-alpha-hydroxyprogesterone caproate (17-OHPC) reduces the risk of recurrent preterm birth and adverse perinatal outcomes.

SEARCH STRATEGY

Systematic search to identify relevant studies published in different languages, registered after 2000, using appropriate MeSH terms.

SELECTION CRITERIA

Inclusion criteria were women between 16 and 26⁺⁶  weeks of pregnancy with history of preterm delivery in any pregnancy randomized to either 17-OHPC or placebo/no treatment.

DATA COLLECTION AND ANALYSIS

The number of preterm births and adverse outcomes in the 17-OHPC and placebo arms over the total number of patients in each randomized group were used to calculate the risk ratio (RR) by random-effects models using the Mantel-Haenszel method. Between-study heterogeneity was assessed using tau² , χ² (Cochrane Q), and I² statistics.

MAIN RESULTS

Four studies were included. There was a 29% (RR 0.71; 95% CI, 0.53-0.96; P=0.001), 26% (RR 0.74; 95% CI, 0.58-0.96; P=0.021), and 40% (RR 0.60; 95% CI, 0.42-0.85; P=0.004) reduction in recurrent preterm birth at <37, <35, and <32 weeks, respectively, in the 17-OHPC group compared with placebo. The reduction in neonatal death was 68% (RR 0.32; 95% CI, 0.15-0.66; P=0.002).

CONCLUSIONS

17-OHPC could reduce the risk of recurrent preterm birth at <37, <35, and <28 weeks and neonatal death.

PROSPERO

CDR42017082190.

Methods for evidence synthesis in the case of very few studies

In systematic reviews, meta-analyses are routinely applied to summarize the results of the relevant studies for a specific research question. If one can assume that in all studies the same true effect is estimated, the application of a meta-analysis with common effect (commonly referred to as fixed-effect meta-analysis) is adequate. If between-study heterogeneity is expected to be present, the method of choice is a meta-analysis with random effects. The widely used DerSimonian and Laird method for meta-analyses with random effects has been criticized due to its unfavorable statistical properties, especially in the case of very few studies. A working group of the Cochrane Collaboration recommended the use of the Knapp-Hartung method for meta-analyses with random effects. However, as heterogeneity cannot be reliably estimated if only very few studies are available, the Knapp-Hartung method, while correctly accounting for the corresponding uncertainty, has very low power. Our aim is to summarize possible methods to perform meaningful evidence syntheses in the situation with only very few (ie, 2-4) studies. Some general recommendations are provided on which method should be used when. Our recommendations are based on the existing literature on methods for meta-analysis with very few studies and consensus of the authors. The recommendations are illustrated by 2 examples coming from dossier assessments of the Institute for Quality and Efficiency in Health Care.

A Common Reference-Based Indirect Comparison Meta-Analysis of Buccal versus Intranasal Midazolam for Early Status Epilepticus

BACKGROUND

Intranasal and buccal midazolam have recently emerged as possible alternatives to intravenous or rectal diazepam or intravenous lorazepam in the treatment of early status epilepticus (SE). However, to date no randomized controlled trial (RCT) has directly compared intranasal midazolam with buccal midazolam.

OBJECTIVE

The aim of this study was to indirectly compare intranasal midazolam with buccal midazolam in the treatment of early SE using common reference-based indirect comparison meta-analyses.

METHODS

RCTs comparing intranasal or buccal midazolam versus either intravenous or rectal diazepam for early SE were systematically searched. Random-effects Mantel-Haenszel meta-analyses were performed to obtain odds ratios (ORs) for the efficacy and safety of intranasal or buccal midazolam versus either intravenous or rectal diazepam. Adjusted indirect comparisons were then made between intranasal and buccal midazolam using the obtained results.

RESULTS

Fifteen studies, with a total of 1662 seizures in 1331 patients (some studies included patients with more than one episode of SE) were included; 1303 patients were younger than 16 years. Indirect comparisons showed no difference between intranasal and buccal midazolam for seizure cessation (OR 0.98, 95% CI 0.32-3.01, comparator: intravenous diazepam; OR 0.87, 95% CI 0.46-1.64, comparator: rectal diazepam). For serious adverse effects, we found a large width and asymmetrical distribution of confidence intervals around the obtained OR of 2.81 (95% CI 0.39-20.12; comparator: rectal diazepam). No data were available for OR using intravenous diazepam as the comparator.

CONCLUSIONS

Indirect comparisons suggest that intranasal and buccal midazolam share similar efficacy in the treatment of early SE in children. Intranasal midazolam should be used with caution and under clinical monitoring of vital functions. RCTs directly comparing intranasal midazolam with buccal midazolam are required to confirm these findings.

Meta-analysis of rare and adverse event data

Potentially beneficial drugs may have side effects. Such adverse events may be serious but rare, making their effect difficult to quantify, or even their identification problematic within randomized controlled trials. Meta-analysis provides a means of combining such outcome data across a series of studies and provides a more powerful analysis than is possible from a single study. However, meta-analysis of adverse event data presents some unique methodological challenges. Procedural issues pertaining to study selection and the inclusion of both randomized and observational data are discussed. The statistical problem of combining rare outcomes is also considered at length. Model choice, continuity corrections, exact statistics, Bayesian methods and sensitivity analysis are all covered. The application of results from adverse event meta-analyses to clinical and economic decision models is also considered. A number of examples are presented to illustrate the key points.

Why most discovered true associations are inflated

Newly discovered true (non-null) associations often have inflated effects compared with the true effect sizes. I discuss here the main reasons for this inflation. First, theoretical considerations prove that when true discovery is claimed based on crossing a threshold of statistical significance and the discovery study is underpowered, the observed effects are expected to be inflated. This has been demonstrated in various fields ranging from early stopped clinical trials to genome-wide associations. Second, flexible analyses coupled with selective reporting may inflate the published discovered effects. The vibration ratio (the ratio of the largest vs. smallest effect on the same association approached with different analytic choices) can be very large. Third, effects may be inflated at the stage of interpretation due to diverse conflicts of interest. Discovered effects are not always inflated, and under some circumstances may be deflated-for example, in the setting of late discovery of associations in sequentially accumulated overpowered evidence, in some types of misclassification from measurement error, and in conflicts causing reverse biases. Finally, I discuss potential approaches to this problem. These include being cautious about newly discovered effect sizes, considering some rational down-adjustment, using analytical methods that correct for the anticipated inflation, ignoring the magnitude of the effect (if not necessary), conducting large studies in the discovery phase, using strict protocols for analyses, pursuing complete and transparent reporting of all results, placing emphasis on replication, and being fair with interpretation of results.

Meta-analyses of safety data: a comparison of exact versus asymptotic methods

The objectives of this study were to establish and describe a database of Cochrane and non-Cochrane meta-analyses of safety data and to determine under what conditions exact methods differ from asymptotic methods in meta-analyses of safety data. A sample of Cochrane (n = 500) and non-Cochrane (n = 200) systematic reviews was randomly selected and a database of safety meta-analyses established. Point estimates and confidence intervals for each meta-analysis were recalculated using exact methods and compared to the results of asymptotic methods. Cochrane reviews were nearly four times as likely as non-Cochrane reviews to contain meta-analyses of safety data (35% compared to 9%). More than 50% of safety meta-analyses contained an outcome with a rare event rate (<5%) and 30% contained at least one study with no events in one arm of the study. For rare event meta-analyses, exact point estimates differed substantially from asymptotic estimates 46% of the time, compared to 17% for those without rare events. Exact confidence intervals differed substantially from asymptotic ones 67% of the time compared to only 19% for those without rare events. The magnitude of differences was also correlated with the number of studies and the summary statistic used to combine the data. Asymptotic methods will not always be a good approximation for exact methods in safety meta-analyses. Event rates and number of studies should be closely examined when choosing the statistical method for combining rare event data.

WITHDRAWN: Transcutaneous electrical nerve stimulation and acupuncture-like transcutaneous electrical nerve stimulation for chronic low back pain

BACKGROUND

In view of the claims and counter-claims of the effectiveness of transcutaneous electrical nerve stimulation, it would seem appropriate to systematically review the literature.

OBJECTIVES

To determine the effectiveness of transcutaneous electrical nerve stimulation in reducing pain and improving range of movement in patients with chronic low back pain.

SEARCH STRATEGY

Electronic searches of EMBASE, MEDLINE, CISCOM, AMED for all studies of TENS in the English language, identifying those treating chronic low back pain and hand searching their references.

SELECTION CRITERIA

The inclusion criterion for studies included in this review, 6 of 68 identified, was comparisons of TENS/ALTENS versus placebo in patients with chronic low back pain.

DATA COLLECTION AND ANALYSIS

Outcome data on pain reduction, range of movement, functional status and work was extracted by two independent reviewers together with trial design qualities to construct a Quality Index.

MAIN RESULTS

The ratio of odds of improvement in pain for each comparison was calculated: TENS vs. placebo at 1.62 (95% CI 0.90, 2.68); ALTENS vs. placebo at 7.22 (95% CI 2.60, 20.01) and TENS/ALTENS vs. placebo at 2.11 (95% CI 1.32, 3.38) times that of placebo. An improvement in pain reduction was seen in 45.80% (CI 37.00%, 55.00%) of TENS; 86.70% (CI 80.00%, 93.00%) of ALTENS; 54.00% (CI 46.20%, 61.80%) of TENS/ ALTENS and 36.40% (95%CI 28.40%, 44.40%) of placebo subjects. The odds of improvement in range of movement on ALTENS vs. placebo was 6.61 times (95% CI 2.36, 18.55) that of placebo.

AUTHORS' CONCLUSIONS

Transcutaneous electrical nerve stimulation appears to reduce pain and improve the range of movement in chronic low back pain subjects. A definitive randomised controlled study of ALTENS, TENS, placebo/no treatment controls, of sufficient power, is needed to confirm these findings.

A test for publication bias in meta-analysis with sparse binary data

A new test for the detection of publication bias in meta-analysis with sparse binary data is proposed. The test statistic is based on observed and expected cell frequencies, and the variance of the observed cell frequencies. These quantities are utilized in a rank correlation test. Type I error rate and power of the test are evaluated in simulations; results are compared to those of two other commonly used test procedures. Sample sizes were generated according to findings in a survey of eight German medical journals. Simulation results indicate that, in contrast to existing test procedures, the new test holds the prescribed significance level when data are sparse. However, the power of all tests is low in many situations of practical importance.

Extreme between-study homogeneity in meta-analyses could offer useful insights

OBJECTIVES

Meta-analyses are routinely evaluated for the presence of large between-study heterogeneity. We examined whether it is also important to probe whether there is extreme between-study homogeneity.

STUDY DESIGN

We used heterogeneity tests with left-sided statistical significance for inference and developed a Monte Carlo simulation test for testing extreme homogeneity in risk ratios across studies, using the empiric distribution of the summary risk ratio and heterogeneity statistic. A left-sided P=0.01 threshold was set for claiming extreme homogeneity to minimize type I error.

RESULTS

Among 11,803 meta-analyses with binary contrasts from the Cochrane Library, 143 (1.21%) had left-sided P-value <0.01 for the asymptotic Q statistic and 1,004 (8.50%) had left-sided P-value <0.10. The frequency of extreme between-study homogeneity did not depend on the number of studies in the meta-analyses. We identified examples where extreme between-study homogeneity (left-sided P-value <0.01) could result from various possibilities beyond chance. These included inappropriate statistical inference (asymptotic vs. Monte Carlo), use of a specific effect metric, correlated data or stratification using strong predictors of outcome, and biases and potential fraud.

CONCLUSION

Extreme between-study homogeneity may provide useful insights about a meta-analysis and its constituent studies.

Quasi-empirical Bayes methodology for improving meta-analysis

This article addresses the problem of heterogeneity among various studies to be combined in a meta-analysis. We adopt quasi-empirical Bayes methodology to predict the odds ratios for each study. As a result, the predicted odds ratios are pulled toward the estimated common odds ratio of the various studies under consideration. With strong heterogeneity among the studies, we jointly consider the display of the 95% CIs of the ORs and a Dixon's test (1950) for "outliers" to exclude the "extreme" estimated ORs. We demonstrate the effectiveness of our methodology based on the data analyzed by Thompson and Pocock (1987) demonstrating the power of the new approach to meta-analysis to find statistical agreement in what looks like great disagreement via a chi-squared test. We believe our technique (i.e., minimum mean-square sense) will go a long way toward increasing the trustworthiness of meta-analysis.

Synthesizing medical evidence: systematic reviews and metaanalyses

Systematic reviews and metaanalyses have become increasingly popular ways of summarizing, and sometimes extending, existing medical knowledge. In this review the authors summarize current methods of performing meta-analyses, including the following: formulating a research question; performing a structured literature search and a search for trials not published in the formal medical literature; summarizing and, where appropriate, combining results from several trials; and reporting and presenting results. Topics such as cumulative and Bayesian metaanalysis and metaregression are also addressed. References to textbooks, articles, and Internet resources are also provided. The goal is to assist readers who wish to perform their own metaanalysis or to interpret critically a published example.

Language of publication restrictions in systematic reviews gave different results depending on whether the intervention was conventional or complementary

OBJECTIVE

To assess whether language of publication restrictions impact the estimates of an intervention's effectiveness, whether such impact is similar for conventional medicine and complementary medicine interventions, and whether the results are influenced by publication bias and statistical heterogeneity.

STUDY DESIGN AND SETTING

We set out to examine the extent to which including reports of randomized controlled trials (RCTs) in languages other than English (LOE) influences the results of systematic reviews, using a broad dataset of 42 language-inclusive systematic reviews, involving 662 RCTs, including both conventional medicine (CM) and complementary and alternative medicine (CAM) interventions.

RESULTS

For CM interventions, language-restricted systematic reviews, compared with language-inclusive ones, did not introduce biased results, in terms of estimates of intervention effectiveness (random effects ration of odds rations ROR=1.02; 95% CI=0.83-1.26). For CAM interventions, however, language-restricted systematic reviews resulted in a 63% smaller protective effect estimate than language-inclusive reviews (random effects ROR=1.63; 95% CI=1.03-2.60).

CONCLUSION

Language restrictions do not change the results of CM systematic reviews but do substantially alter the results of CAM systematic reviews. These findings are robust even after sensitivity analyses, and do not appear to be influenced by statistical heterogeneity and publication bias.

Meta-analysis and indirect comparisons of levetiracetam with other second-generation antiepileptic drugs in partial epilepsy

Few comparative clinical trials of newer antiepileptic drugs (AEDs) in patients with refractory partial epilepsy are available. Therefore, meta-analysis is a widely used and useful method for comparing them. Despite the limitations of indirect comparisons, and recognizing that these drugs were tested at different doses, such comparisons can be helpful to physicians making practical treatment decisions. The purposes of this study were to present newer meta-analysis results for add-on levetiracetam compared with placebo and to estimate its efficacy and tolerability compared with other new AEDs (gabapentin, lamotrigine, oxcarbazepine, tiagabine, topiramate, and zonisamide) in a meta-analysis using methods for making indirect comparisons. Randomized placebo-controlled clinical trials of add-on therapy with levetiracetam, gabapentin, lamotrigine, oxcarbazepine, tiagabine, topiramate, and zonisamide in patients with refractory partial epilepsy were identified in the Cochrane Library 2002. A fixed-effects model was used to estimate Mantel-Haenszel odds ratios for the responder rate (efficacy measure) and withdrawal rate (mainly tolerability measure) of levetiracetam and other new AEDs versus placebo. Because no head-to-head clinical trials comparing these new AEDs exist, adjusted indirect comparisons were then made between levetiracetam and each other AED using the meta-analysis results. At the doses tested, levetiracetam was more effective in terms of responder rate than gabapentin (odds ratio 2.64 with 95% CI 1.51-4.63) and lamotrigine (odds ratio 1.86 with 95% CI 1.04-3.34) and equally well tolerated. Levetiracetam had a significantly lower withdrawal rate than topiramate (odds ratio 0.52 with 95% CI 0.29-0.93) and oxcarbazepine (odds ratio 0.55 with 95% CI 0.33-0.92), with comparable efficacy. Although levetiracetam did not differ significantly from the other AEDs, numerical trends favoring levetiracetam were obtained in response rate and in withdrawal rate (tiagabine, zonisamide). Indirect comparisons based on meta-analysis suggest that add-on therapy with levetiracetam has a favorable responder and/or withdrawal rate relative to several AEDs in patients with partial epilepsy with doses used in clinical trials. These meta-analyses give only short-term efficacy and safety data. Comparative clinical trials and long-term studies of these agents are needed to confirm these findings.

Meta-analysis and structured literature review in radiology

The overall goal of a systematic review or meta-analysis is to combine results of previous studies to arrive at summary conclusions about a body of research. In radiology, systematic reviews or meta-analyses can be used to calculate a summary estimate of effect size of a treatment that used imaging data to assess outcomes in observational or randomized controlled clinical trials, estimate the clinical effectiveness of an imaging-guided therapy procedure, evaluate the summary diagnostic accuracy of an imaging test, or synthesize results of economic evaluations that used imaging data. This article outlines the general concepts of structured literature reviews and discusses the approaches for conducting a meta-analysis in radiology, emphasizing the methods available for data synthesis and handling heterogeneity between and among studies.

Proving non-inferiority or equivalence of two treatments with dichotomous endpoints using exact methods

Since the early work of RA Fisher, exact methods have been recognized as important tools in data analysis because they provide valid statistical inference even with small sample sizes, or with sparse or skewed data. With the recent advance of computational power and the availability of commercial software packages, exact methods have gained substantial popularity over the past two decades. However, most of these exact methods have been devoted to testing classical null hypotheses of no differences, and until recently little was known about exact methods dealing with non-inferiority or equivalence hypotheses. The presence of nuisance parameters in testing non-inferiority/equivalence hypotheses presents a special challenge for exact methods because of the intense computational requirement. In this paper, we review exact methods available for proving non-inferiority or equivalence of two treatments with a dichotomous endpoint. First, we present the general methodology for conducting exact tests for non-inferiority or equivalence; we then discuss several unconditional and conditional methods available for constructing hypothesis tests and confidence intervals based on three commonly used measures, namely, the difference, relative risk, and odds ratio of two independent proportions or rates. Finally, we illustrate with several examples the application of these exact methods in analysing and planning non-inferiority or equivalence trials.

Assessing the potential for bias in meta-analysis due to selective reporting of subgroup analyses within studies

Subgroup analysis is frequently used to investigate heterogeneity in meta-analysis. Subgroup data are not always available, and researchers should record what data were available for each trial. If data were not available, and it is known that the subgroup data were collected, the potential for selective reporting should be considered. Bias due to selective publishing of reports is widely recognized in meta-analysis. In contrast, selective reporting within studies is little discussed but potentially important. We explored this problem by evaluating the effect of potential bias in subgroup analysis due to within-study selective reporting with an existing meta-analysis. The review addressed malaria chemoprophylaxis in pregnancy. The conclusion in the original review, that benefit is limited to primigravidae, was based on subgroup analysis using the three trials out of five which reported on subgroups. We developed a method of sensitivity analysis that imputes data for the missing subgroups to assess the robustness of the results and the conclusions drawn. In this particular example, our analysis indicates that the estimate of effect reported in the review is most likely to overestimate the true effect and the conclusion that benefit is limited to primigravidae may be false.

Strategies for comparing treatments on a binary response with multi-centre data

This paper surveys methods for comparing treatments on a binary response when observations occur for several strata. A common application is multi-centre clinical trials, in which the strata refer to a sample of centres or sites of some type. Questions of interest include how one should summarize the difference between the treatments, how one should make inferential comparisons, how one should investigate whether treatment-by-centre interaction exists, how one should describe effects when interaction exists, whether one should treat centres and centre-specific treatment effects as fixed or random, and whether centres that have either 0 successes or 0 failures should contribute to the analysis. This article discusses these matters in the context of various strategies for analysing such data, in particular focusing on special problems presented by sparse data.

Transcutaneous electrical nerve stimulation and acupuncture-like transcutaneous electrical nerve stimulation for chronic low back pain

BACKGROUND

Transcutaneous electrical nerve stimulation (TENS), originally based on the gate-control theory of pain, is widely used for the treatment of chronic low back pain. Despite its wide use and theoretical rationale, there appears at first glance little scientific evidence to support its use. This Cochrane review examines the available evidence on TENS for the treatment of chronic back pain through an exhaustive search of the literature.

OBJECTIVES

Transcutaneous electrical nerve stimulation (TENS) and acupuncture-like transcutaneous electrical nerve stimulation (ALTENS) for chronic low back pain management have experienced a tremendous growth over the past 25 years. The objective of this review was to assess the effects of TENS and ALTENS for reducing pain and improving function in patients with chronic back pain.

SEARCH STRATEGY

We searched MEDLINE up to November 1997, EMBASE from 1985 to September 1995, Amed and Ciscom to January 1995, reference lists of the retrieved articles, proceedings of conferences and contacted investigators in the field.

SELECTION CRITERIA

Randomised trials comparing TENS or ALTENS therapy to placebo in patients with chronic low back pain.

DATA COLLECTION AND ANALYSIS

Two reviewers independently assessed trial quality and extracted data on pain reduction, range of movement, functional and work status.

MAIN RESULTS

Six trials were included. The trials included 288 participants with an average age range of 45 to 50 years and approximately equal numbers of women and men. The overall odds ratio for improvement in pain for each comparison was: TENS/ALTENS versus placebo 2.11 (95% confidence interval 1.32 to 3. 38), ALTENS versus placebo 7.22 (95% confidence interval 2.60 to 20.01) and TENS versus placebo 1.52 (95% confidence interval 0.90 to 2.58). The odds ration for improvement in range of motion on ALTENS versus placebo was 6.61 (95% confidence interval 2.36 to 18.55).

REVIEWER'S CONCLUSIONS

There is evidence from the limited data available that TENS/ALTENS reduces pain and improves range of motion in chronic back pain patients, at least in the short term. A large trial of ALTENS and TENS is needed to confirm these findings.

Systematic reviews and meta-analysis: a structured review of the methodological literature

OBJECTIVE

To systematically review methods for systematic review/meta-analysis in order to identify the different methodological and statistical methods that have been proposed. A summary of the main findings is presented here, with emphasis given to health services research topics.

METHODS

A thorough systematic search for methodological papers was carried out using a variety of methods, including the use of electronic databases. Approximately 1000 potentially relevant references were identified, a number of them from education, psychology and sociology.

RESULTS

After briefly reviewing the procedural methods required to carry out a review, and the basic statistical methods used to combine study estimates, less established methods are discussed. These include methods for dealing with publication bias, meta-regression, meta-analysis of individual patient data, the synthesis of non-randomized evidence alone and in combination with randomized studies. Bayesian modelling and economic evaluation through meta-analysis. Recommendations for meta-analytical practice are given; these are either distilled from previous guidelines, or constructed where there appears to be a broad consensus across the literature.

CONCLUSIONS

It is hoped that this review will provide a consistent and comprehensive, but concise, description of the methods available for synthesizing evidence, that it will promote better quality reviews of the results of health services research and identify specific areas which require methodological development.

Risperidone in the treatment of schizophrenia: a meta-analysis of randomized controlled trials

This study evaluates the relative effectiveness and side-effects of risperidone as compared with conventional neuroleptics in the treatment of schizophrenia, by meta-analysis of 11 double-blind, randomized controlled trials. The proportion of patients showing clinical improvement; use of medications for extrapyramidal side-effects (EPS); the treatment drop-out rates; and the changes in negative PANSS scores were measured. Compared with conventional neuroleptics, slightly more patients in the risperidone group showed clinical improvement [57 vs 52%; odds ratio 1.27, 95% confidence interval (CI): 1.04, 1.56]. The use of concomitant medications for EPS was significantly less in the risperidone group than in the conventional neuroleptic group (22.8 vs 38.4%; odds ratio 0.51, 95% CI: 0.41, 0.63). The overall drop-out rate was lower in the risperidone group than in other neuroleptic group (29.1 vs 33.9%; odds ratio 0.75, 95% CI: 0.61, 0.94). The difference in changes in negative PANSS score between the risperidone and the haloperidol group was -0.74 (95% CI: -1.50, 0.02). Weight gain and tachycardia are more common in patients treated with risperidone. Sensitivity analysis of different analytic approaches did not materially change the main estimates. It was concluded that the short-term efficacy of risperidone is comparable to other neuroleptics in the treatment of schizophrenia. It is associated with significantly fewer EPS than conventional neuroleptics (mainly haloperidol).