Systematic review of methods for individual patient data meta- analysis with binary outcomes

Pooled Cohort Profile: ReCoDID Consortium's Harmonized Acute Febrile Illness Arbovirus Meta-Cohort

Infectious disease (ID) cohorts are key to advancing public health surveillance, public policies, and pandemic responses. Unfortunately, ID cohorts often lack funding to store and share clinical-epidemiological (CE) data and high-dimensional laboratory (HDL) data long term, which is evident when the link between these data elements is not kept up to date. This becomes particularly apparent when smaller cohorts fail to successfully address the initial scientific objectives due to limited case numbers, which also limits the potential to pool these studies to monitor long-term cross-disease interactions within and across populations. CE data from 9 arbovirus (arthropod-borne viruses) cohorts in Latin America were retrospectively harmonized using the Maelstrom Research methodology and standardized to Clinical Data Interchange Standards Consortium (CDISC). We created a harmonized and standardized meta-cohort that contains CE and HDL data from 9 arbovirus studies from Latin America. To facilitate advancements in cross-population inference and reuse of cohort data, the Reconciliation of Cohort Data for Infectious Diseases (ReCoDID) Consortium harmonized and standardized CE and HDL from 9 arbovirus cohorts into 1 meta-cohort. Interested parties will be able to access data dictionaries that include information on variables across the data sets via Bio Studies. After consultation with each cohort, linked harmonized and curated human cohort data (CE and HDL) will be made accessible through the European Genome-phenome Archive platform to data users after their requests are evaluated by the ReCoDID Data Access Committee. This meta-cohort can facilitate various joint research projects (eg, on immunological interactions between sequential flavivirus infections and for the evaluation of potential biomarkers for severe arboviral disease).

Efficacy of iron supplementation on physical capacity in non-anaemic iron-deficient individuals: protocol for an individual patient data meta-analysis

BACKGROUND

A deficiency in iron stores is associated with various adverse health complications, which, if left untreated, can progress to states of anaemia, whereby there is significant detriment to an individual's work capacity and quality of life due to compromised erythropoiesis. The most common methods employed to treat an iron deficiency include oral iron supplementation and, in persistent and/or unresponsive cases, intravenous iron therapy. The efficacy of these treatments, particularly in states of iron deficiency without anaemia, is equivocal. Indeed, both randomised control trials and aggregate data meta-analyses have produced conflicting evidence. Therefore, this study aims to assess the efficacy of both oral and intravenous iron supplementation on physical capacity, quality of life, and fatigue scores in iron-deficient non-anaemic individuals using individual patient data (IPD) meta-analysis techniques.

METHODS

All potential studies, irrespective of design, will be sourced through systematic searches on the following databases: Cochrane Central Register of Controlled Trials, MEDLINE Ovid, Embase Ovid, Web of Science: Science Citation Index Expanded, Web of Science: Conference Proceedings Citation Index-Science, ClinicalTrials.gov, and World Health Organization (WHO) International Clinical Trials Registry Platform. Individual patient data from all available trials will be included and subsequently analysed in a two-stage approach. Predetermined subgroup and sensitivity analyses will be employed to further explain results.

DISCUSSION

The significance of this IPD meta-analysis is one of consolidating a clear consensus to better inform iron-deficient individuals of the physiological response associated with iron supplementation. The IPD approach, to the best of our knowledge, is novel for this research topic. As such, the findings will significantly contribute to the current body of evidence.

SYSTEMATIC REVIEW REGISTRATION

PROSPERO CRD42020191739.

Heterogeneity in meta-analysis: a comprehensive overview

In recent years, meta-analysis has evolved to a critically important field of Statistics, and has significant applications in Medicine and Health Sciences. In this work we briefly present existing methodologies to conduct meta-analysis along with any discussion and recent developments accompanying them. Undoubtedly, studies brought together in a systematic review will differ in one way or another. This yields a considerable amount of variability, any kind of which may be termed heterogeneity. To this end, reports of meta-analyses commonly present a statistical test of heterogeneity when attempting to establish whether the included studies are indeed similar in terms of the reported output or not. We intend to provide an overview of the topic, discuss the potential sources of heterogeneity commonly met in the literature and provide useful guidelines on how to address this issue and to detect heterogeneity. Moreover, we review the recent developments in the Bayesian approach along with the various graphical tools and statistical software that are currently available to the analyst. In addition, we discuss sensitivity analysis issues and other approaches of understanding the causes of heterogeneity. Finally, we explore heterogeneity in meta-analysis for time to event data in a nutshell, pointing out its unique characteristics.

Association between long-term use of calcium channel blockers (CCB) and the risk of breast cancer: a retrospective longitudinal observational study protocol

INTRODUCTION

Calcium channel blockers (CCB), a commonly prescribed antihypertensive (AHT) medicine, may be associated with increased risk of breast cancer. The proposed study aims to examine whether long-term CCB use is associated with the development of breast cancer and to characterise the dose-response nature of any identified association, to inform future hypertension management.

METHODS AND ANALYSIS

The study will use data from 2 of Australia's largest cohort studies; the Australian Longitudinal Study on Women's Health, and the 45 and Up Study, combined with the Rotterdam Study. Eligible women will be those with diagnosed hypertension, no history of breast cancer and no prior CCB use at start of follow-up (2004-2009). Cumulative dose-duration exposure to CCB and other AHT medicines will be captured at the earliest date of: the outcome (a diagnosis of invasive breast cancer); a competing risk event (eg, bilateral mastectomy without a diagnosis of breast cancer, death prior to any diagnosis of breast cancer) or end of follow-up (censoring event). Fine and Gray competing risks regression will be used to assess the association between CCB use and development of breast cancer using a generalised propensity score to adjust for baseline covariates. Time-varying covariates related to interaction with health services will also be included in the model. Data will be harmonised across cohorts to achieve identical protocols and a two-step random effects individual patient-level meta-analysis will be used.

ETHICS AND DISSEMINATION

Ethical approval was obtained from the following Human research Ethics Committees: Curtin University (ref No. HRE2022-0335), NSW Population and Health Services Research Ethics Committee (2022/ETH01392/2022.31), ACT Research Ethics and Governance Office approval under National Mutual Acceptance for multijurisdictional data linkage research (2022.STE.00208). Results of the proposed study will be published in high-impact journals and presented at key scientific meetings.

TRIAL REGISTRATION NUMBER

NCT05972785.

Methods for Integrating Trials and Non-experimental Data to Examine Treatment Effect Heterogeneity

Estimating treatment effects conditional on observed covariates can improve the ability to tailor treatments to particular individuals. Doing so effectively requires dealing with potential confounding, and also enough data to adequately estimate effect moderation. A recent influx of work has looked into estimating treatment effect heterogeneity using data from multiple randomized controlled trials and/or observational datasets. With many new methods available for assessing treatment effect heterogeneity using multiple studies, it is important to understand which methods are best used in which setting, how the methods compare to one another, and what needs to be done to continue progress in this field. This paper reviews these methods broken down by data setting: aggregate-level data, federated learning, and individual participant-level data. We define the conditional average treatment effect and discuss differences between parametric and nonparametric estimators, and we list key assumptions, both those that are required within a single study and those that are necessary for data combination. After describing existing approaches, we compare and contrast them and reveal open areas for future research. This review demonstrates that there are many possible approaches for estimating treatment effect heterogeneity through the combination of datasets, but that there is substantial work to be done to compare these methods through case studies and simulations, extend them to different settings, and refine them to account for various challenges present in real data.

Outcomes of endovascular embolization for Vein of Galen malformations: An individual participant data meta-analysis

Introduction

Understanding outcomes after Vein of Galen malformation (VOGM) embolization has been limited by small sample size in reported series and predominantly single center studies. To address these limitations, we perform an individual-participant meta-analysis (IPMA) to identify risk factors associated with all-cause mortality and clinical outcome after VOGM endovascular embolization.

Methods

We performed a systematic review and IPMA of VOGM endovascular outcomes according to PRISMA guidelines. Individual patient characteristics including demographic, intra/post-operative adverse events, treatment efficacy (partial or complete occlusion), and clinical outcome were collected. Mixed-effects logistic regression with random effects modeling and Bonferroni correction was used (p ≤ 0.003 threshold for statistical significance). The primary and secondary outcomes were all-cause mortality and poor clinical outcome (moderate/severe developmental delay or permanent disabling injury), respectively. Data are expressed as (mean ± standard deviation (SD)) or (odds ratio (OR), 95% confidence interval (CI), I 2, p-value).

Results

Thirty-five studies totaling 307 participants quantifying outcomes after endovascular embolization for VOGM were included. Follow up time was 42 (±57) months. Our analysis contained 42% neonates (<1 month) at first embolization, 45% infants (1 month ≤2 years), and 13% children (>2 years). Complete occlusion was reported in 48% of participants. Overall all-cause mortality was 16%. Overall, good clinical outcome was achieved in 68% of participants. First embolization as a neonate [OR = 6.93; 95% CI (1.99-24.08); I 2 < 0.01; p < 0.001] and incomplete embolization [OR = 10.87; 95% CI (1.86-63.55); I 2 < 0.01; p < 0.001] were associated with mortality. First embolization as a neonate [OR = 3.24; 95% CI (1.47-7.15); I 2 < 0.01; p < 0.001], incomplete embolization [OR = 5.26; 95% CI (2.06-13.43); I 2 < 0.01; p < 0.001], and heart failure at presentation [OR = 3.10; 95% CI (1.03-9.33); I 2 < 0.01; p = 0.002] were associated with poor clinical outcomes. Sex, angioarchitecture of lesion, embolization approach (transvenous vs. transarterial), and single or multistage embolization were not associated with mortality or clinical outcome.

Conclusions

We identify incomplete VOGM embolization independently associated with mortality and poor clinical outcome. While this study provides the highest level of evidence for VOGM embolization to date, prospective multicenter studies are needed to understand the optimal treatment strategies, outcomes, and natural history after VOGM embolization.

Harmonization-Information Trade-Offs for Sharing Individual Participant Data in Biomedicine

Biomedical practice is evidence-based. Peer-reviewed papers are the primary medium to present evidence and data-supported results to drive clinical practice. However, it could be argued that scientific literature does not contain data, but rather narratives about and summaries of data. Meta-analyses of published literature may produce biased conclusions due to the lack of transparency in data collection, publication bias, and inaccessibility to the data underlying a publication ('dark data'). Co-analysis of pooled data at the level of individual research participants can offer higher levels of evidence, but this requires that researchers share raw individual participant data (IPD). FAIR (findable, accessible, interoperable, and reusable) data governance principles aim to guide data lifecycle management by providing a framework for actionable data sharing. Here we discuss the implications of FAIR for data harmonization, an essential step for pooling data for IPD analysis. We describe the harmonization-information trade-off, which states that the level of granularity in harmonizing data determines the amount of information lost. Finally, we discuss a framework for managing the trade-off and the levels of harmonization. In the coming era of funder mandates for data sharing, research communities that effectively manage data harmonization will be empowered to harness big data and advanced analytics such as machine learning and artificial intelligence tools, leading to stunning new discoveries that augment our understanding of diseases and their treatments. By elevating scientific data to the status of a first-class citizen of the scientific enterprise, there is strong potential for biomedicine to transition from a narrative publication product orientation to a modern data-driven enterprise where data itself is viewed as a primary work product of biomedical research.

Reducing antibiotic use in uncomplicated urinary tract infections in adult women: a systematic review and individual participant data meta-analysis

BACKGROUND

Randomised controlled trials (RCTs) investigated analgesics, herbal formulations, delayed prescription of antibiotics and placebo to prevent overprescription of antibiotics in women with uncomplicated urinary tract infections (uUTI).

OBJECTIVES

To estimate the effect of these strategies and to identify symptoms, signs or other factors that indicate a benefit from these strategies.

DATA SOURCES

MEDLINE, EMBASE, Web of Science, LILACS, Cochrane Database of Systematic Reviews and of Controlled Trials, and ClinicalTrials.

STUDY ELIGIBILITY CRITERIA, PARTICIPANTS AND INTERVENTIONS

RCTs investigating any strategies to reduce antibiotics versus immediate antibiotics in adult women with uUTI in primary care.

DATA SYNTHESIS

We extracted individual participant data (IPD) if available, otherwise aggregate data (AD). Bayesian random-effects meta-analysis of the AD was used for pairwise comparisons. Candidate moderators and prognostic indicators of treatment effects were investigated using generalised linear mixed models based on IPD.

RESULTS

We analysed IPD of 3524 patients from eight RCTs and AD of 78 patients. Non-antibiotic strategies increased the rates of incomplete recovery (odds ratio [OR] 3.0; 95% credible interval [CI] 1.7-5.5; Bayesian p-value p_B=0.0017; τ=0.6), subsequent antibiotic treatment (OR 3.5 [95% CI 2.1, 5.8; p_B=0.0003) and pyelonephritis (OR 5.6; 95% CI 2.3, 13.9; p_B=0.0003). Conversely, they decreased overall antibiotic use by 63%. In patients positive for urinary erythrocytes and urine culture were at increased risk for incomplete recovery (OR 4.7; 95% CI 2.1-10.8; p_B =0.0010), but no difference was apparent where both were negative (OR 0.8; 95% CI 0.3-2.0; p_B =0.667). In patients treated with using non-antibiotic strategies, urinary erythrocytes and positive urine culture were independent prognostic indicators for subsequent antibiotic treatment and pyelonephritis.

CONCLUSIONS AND RELEVANCE

Compared to immediate antibiotics, non-antibiotic strategies reduce overall antibiotic use but result in poorer clinical outcomes. The presence of erythrocytes and tests to confirm bacteria in urine could be used to target antibiotic prescribing.

Bayesian heterogeneity in a meta-analysis with two studies and binary data

The meta-analysis of two trials is valuable in many practical situations, such as studies of rare and/or orphan diseases focussed on a single intervention. In this context, additional concerns, like small sample size and/or heterogeneity in the results obtained, might make standard frequentist and Bayesian techniques inappropriate. In a meta-analysis, moreover, the presence of between-sample heterogeneity adds model uncertainty, which must be taken into consideration when drawing inferences. We suggest that the most appropriate way to measure this heterogeneity is by clustering the samples and then determining the posterior probability of the cluster models. The meta-inference is obtained as a mixture of all the meta-inferences for the cluster models, where the mixing distribution is the posterior model probability. We present a simple two-component form of Bayesian model averaging that is unaffected by characteristics such as small study size or zero-cell counts, and which is capable of incorporating uncertainties into the estimation process. Illustrative examples are given and analysed, using real sparse binomial data.

Recreational Physical Activity and Risk of Incident Knee Osteoarthritis: An International Meta-Analysis of Individual Participant-Level Data

OBJECTIVE

The effect of physical activity on the risk of developing knee osteoarthritis (OA) is unclear. We undertook this study to examine the relationship between recreational physical activity and incident knee OA outcomes using comparable physical activity and OA definitions.

METHODS

Data were acquired from 6 global, community-based cohorts of participants with and those without knee OA. Eligible participants had no evidence of knee OA or rheumatoid arthritis at baseline. Participants were followed up for 5-12 years for incident outcomes including the following: 1) radiographic knee OA (Kellgren-Lawrence [K/L] grade ≥2), 2) painful radiographic knee OA (radiographic OA with knee pain), and 3) OA-related knee pain. Self-reported recreational physical activity included sports and walking/cycling activities and was quantified at baseline as metabolic equivalents of task (METs) in days per week. Risk ratios (RRs) were calculated and pooled using individual participant data meta-analysis. Secondary analysis assessed the association between physical activity, defined as time (hours per week) spent in recreational physical activity and incident knee OA outcomes.

RESULTS

Based on a total of 5,065 participants, pooled RR estimates for the association of MET days per week with painful radiographic OA (RR 1.02 [95% confidence interval (95% CI) 0.93-1.12]), radiographic OA (RR 1.00 [95% CI 0.94-1.07]), and OA-related knee pain (RR 1.00 [95% CI 0.96-1.04]) were not significant. Similarly, the analysis of hours per week spent in physical activity also showed no significant associations with all outcomes.

CONCLUSION

Our findings suggest that whole-body, physiologic energy expenditure during recreational activities and time spent in physical activity were not associated with incident knee OA outcomes.

Prevalence Estimates of Amyloid Abnormality Across the Alzheimer Disease Clinical Spectrum

IMPORTANCE

One characteristic histopathological event in Alzheimer disease (AD) is cerebral amyloid aggregation, which can be detected by biomarkers in cerebrospinal fluid (CSF) and on positron emission tomography (PET) scans. Prevalence estimates of amyloid pathology are important for health care planning and clinical trial design.

OBJECTIVE

To estimate the prevalence of amyloid abnormality in persons with normal cognition, subjective cognitive decline, mild cognitive impairment, or clinical AD dementia and to examine the potential implications of cutoff methods, biomarker modality (CSF or PET), age, sex, APOE genotype, educational level, geographical region, and dementia severity for these estimates.

DESIGN, SETTING, AND PARTICIPANTS

This cross-sectional, individual-participant pooled study included participants from 85 Amyloid Biomarker Study cohorts. Data collection was performed from January 1, 2013, to December 31, 2020. Participants had normal cognition, subjective cognitive decline, mild cognitive impairment, or clinical AD dementia. Normal cognition and subjective cognitive decline were defined by normal scores on cognitive tests, with the presence of cognitive complaints defining subjective cognitive decline. Mild cognitive impairment and clinical AD dementia were diagnosed according to published criteria.

EXPOSURES

Alzheimer disease biomarkers detected on PET or in CSF.

MAIN OUTCOMES AND MEASURES

Amyloid measurements were dichotomized as normal or abnormal using cohort-provided cutoffs for CSF or PET or by visual reading for PET. Adjusted data-driven cutoffs for abnormal amyloid were calculated using gaussian mixture modeling. Prevalence of amyloid abnormality was estimated according to age, sex, cognitive status, biomarker modality, APOE carrier status, educational level, geographical location, and dementia severity using generalized estimating equations.

RESULTS

Among the 19 097 participants (mean [SD] age, 69.1 [9.8] years; 10 148 women [53.1%]) included, 10 139 (53.1%) underwent an amyloid PET scan and 8958 (46.9%) had an amyloid CSF measurement. Using cohort-provided cutoffs, amyloid abnormality prevalences were similar to 2015 estimates for individuals without dementia and were similar across PET- and CSF-based estimates (24%; 95% CI, 21%-28%) in participants with normal cognition, 27% (95% CI, 21%-33%) in participants with subjective cognitive decline, and 51% (95% CI, 46%-56%) in participants with mild cognitive impairment, whereas for clinical AD dementia the estimates were higher for PET than CSF (87% vs 79%; mean difference, 8%; 95% CI, 0%-16%; P = .04). Gaussian mixture modeling-based cutoffs for amyloid measures on PET scans were similar to cohort-provided cutoffs and were not adjusted. Adjusted CSF cutoffs resulted in a 10% higher amyloid abnormality prevalence than PET-based estimates in persons with normal cognition (mean difference, 9%; 95% CI, 3%-15%; P = .004), subjective cognitive decline (9%; 95% CI, 3%-15%; P = .005), and mild cognitive impairment (10%; 95% CI, 3%-17%; P = .004), whereas the estimates were comparable in persons with clinical AD dementia (mean difference, 4%; 95% CI, -2% to 9%; P = .18).

CONCLUSIONS AND RELEVANCE

This study found that CSF-based estimates using adjusted data-driven cutoffs were up to 10% higher than PET-based estimates in people without dementia, whereas the results were similar among people with dementia. This finding suggests that preclinical and prodromal AD may be more prevalent than previously estimated, which has important implications for clinical trial recruitment strategies and health care planning policies.

Efficacy and safety of sodium-glucose cotransporter 2 inhibitors initiation in patients with acute heart failure, with and without type 2 diabetes: a systematic review and meta-analysis

Background

There is uncertainty and limited data regarding initiation of sodium-glucose cotransporter 2 (SGLT2) inhibitors among patients hospitalized with acute heart failure (AHF). This systematic review and meta-analysis aim to establish the efficacy and safety of SGLT2 inhibitors initiated in patients hospitalized for AHF.

Methods

PubMed/Medline, Embase, and Cochrane library were searched using the following terms: (“sglt2" and "acute heart failure") and (“sglt2" and "worsening heart failure") from inception till November 15th, 2021 for randomized controlled trials (RCTs) comparing the efficacy and safety of initiating an SGLT2 inhibitor compared with placebo in patients with AHF. Major cardiovascular and diabetes scientific meetings in 2021 were also searched for relevant studies. Prespecified efficacy outcomes were all-cause mortality, rehospitalization for heart failure, and improvement in Kansas City Cardiomyopathy Questionnaire (KCCQ) scale score. Prespecified safety outcomes were acute kidney injury (AKI), hypotension, and hypoglycemia. Random effects odds ratio (OR) and mean difference with 95% confidence intervals (CIs) were calculated.

Results

Three RCTs with a total of 1831 patients were included. Initiation of SGLT2 inhibitors in patients with AHF reduced the risk of rehospitalization for heart failure (OR 0.52; 95% CI [0.42, 0.65]) and improved Kansas City Cardiomyopathy Questionnaire scores (mean difference 4.12; 95% CI [0.1.89, 6.53]). There was no statistically significant effect for initiation of SGLT2 inhibitors in patients with AHF on all-cause mortality (OR 0.70; 95% CI [0.46, 1.08]). Initiation of SGLT2 inhibitors in patients with AHF did not increase the acute kidney injury (OR 0.76; 95% CI [0.50, 1.16]), hypotension (OR 1.17; 95% CI [0.80, 1.71]), or hypoglycemia (OR 1.51; 95% CI [0.86, 2.65]).

Conclusion

Initiation of SGLT2 inhibitors in patients hospitalized for AHF during hospitalization or early post-discharge (within 3 days) reduces the risk of rehospitalization for heart failure and improves patient-reported outcomes with no excess risk of adverse effects.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12933-022-01455-2.

Community-based differentiated service delivery models incorporating multi-month dispensing of antiretroviral treatment for newly stable people living with HIV receiving single annual clinical visits: a pooled analysis of two cluster-randomized trials in southern Africa

Introduction

Differentiated service delivery (DSD) models for HIV treatment decrease health facility visit frequency and limit healthcare facility‐based exposure to severe acute respiratory syndrome coronavirus 2. However, two important evidence gaps include understanding DSD effectiveness amongst clients commencing DSD within 12 months of antiretroviral treatment (ART) initiation and amongst clients receiving only single annual clinical consultations. To investigate these, we pooled data from two cluster‐randomized trials investigating community‐based DSD in Zimbabwe and Lesotho.

Methods

Individual‐level participant data of newly stable adults enrolled between 6 and 12 months after ART initiation were pooled. Both trials (conducted between August 2017 and July 2019) had three arms: Standard‐of‐care three‐monthly ART provision at healthcare facilities (SoC, control); ART provided three‐monthly in community ART groups (CAGs) (3MC) and ART provided six‐monthly in either CAGs or at community‐distribution points (6MC). Clinical visits were three‐monthly in SoC and annually in intervention arms. The primary outcome was retention in care and secondary outcomes were viral suppression (VS) and number of unscheduled facility visits 12 months after enrolment. Individual‐level regression analyses were conducted by intention‐to‐treat specifying for clustering and adjusted for country.

Results and Discussion

A total of 599 participants were included; 212 (35.4%), 128 (21.4%) and 259 (43.2%) in SoC, 3MC and 6MC, respectively. Few participants aged <25 years were included (n = 32). After 12 months, 198 (93.4%), 123 (96.1%) and 248 (95.8%) were retained in SoC, 3MC and 6MC, respectively. Retention in 3MC was superior versus SoC, adjusted risk difference (aRD) = 4.6% (95% CI: 0.7%−8.5%). Retention in 6MC was non‐inferior versus SoC, aRD = 1.7% (95% CI: −2.5%−5.9%) (prespecified non‐inferiority aRD margin −3.25%). VS was similar between arms, 99.3, 98.6 and 98.1% in SoC, 3MC and 6MC, respectively. Adjusted risk ratio's for VS were 0.98 (95% CI: 0.92−1.03) for 3MC versus SoC, and 0.98 (CI: 0.95−1.00) for 6MC versus SoC. Unscheduled clinic visits were not increased in intervention arms: incidence rate ratio = 0.53 (CI: 0.16−1.80) for 3MC versus SoC; and 0.82 (CI: 0.25−2.79) for 6MC versus SoC.

Conclusions

Community‐based DSD incorporating three‐ and six‐monthly ART refills and single annual clinical visits were at least non‐inferior to standard facility‐based care amongst newly stable ART clients aged ≥25 years.

ClinicalTrials.gov: NCT03238846 & NCT03438370

Adverse childhood experiences and substance misuse in young people in India: results from the multisite cVEDA cohort

BACKGROUND

Adverse childhood experiences (ACEs) increases vulnerability to externalising disorders such as substance misuse. The study aims to determine the prevalence of ACEs and its association with substance misuse.

METHODS

Data from the Consortium on Vulnerability to Externalising Disorders and Addictions (cVEDA) in India was used (n = 9010). ACEs were evaluated using the World Health Organisation (WHO) Adverse Childhood Experiences International Questionnaire whilst substance misuse was assessed using the WHO Alcohol, Smoking and Substance Involvement Screening Test. A random-effects, two-stage individual patient data meta-analysis explained the associations between ACEs and substance misuse with adjustments for confounders such as sex and family structure.

RESULTS

1 in 2 participants reported child maltreatment ACEs and family level ACEs. Except for sexual abuse, males report more of every individual childhood adversity and are more likely to report misusing substances compared with females (87.3% vs. 12.7%). In adolescents, family level ACEs (adj OR 4.2, 95% CI 1.5-11.7) and collective level ACEs (adj OR 6.6, 95% CI 1.4-31.1) show associations with substance misuse whilst in young adults, child level ACEs such as maltreatment show similar strong associations (adj OR 2.0, 95% CI 1.1-3.5).

CONCLUSION

ACEs such as abuse and domestic violence are strongly associated with substance misuse, most commonly tobacco, in adolescent and young adult males in India. The results suggest enhancing current ACE resilience programmes and 'trauma-informed' approaches to tackling longer-term impact of ACEs in India.

FUNDING

Newton Bhabha Grant jointly funded by the Medical Research Council, UK (MR/N000390/1) and the Indian Council of Medical Research (ICMR/MRC-UK/3/M/2015-NCD-I).

Optimal interpregnancy interval in autism spectrum disorder: A multi-national study of a modifiable risk factor

It is biologically plausible that risk of autism spectrum disorder (ASD) is elevated by both short and long interpregnancy intervals (IPI). We conducted a retrospective cohort study of singleton, non-nulliparous live births, 1998-2007 in Denmark, Finland, and Sweden (N = 925,523 births). Optimal IPI was defined as the IPI at which minimum risk was observed. Generalized additive models were used to estimate relative risks (RR) of ASD and 95% Confidence Intervals (CI). Population impact fractions (PIF) for ASD were estimated under scenarios for shifts in the IPI distribution. We observed that the association between ASD (N = 9302) and IPI was U-shaped for all countries. ASD risk was lowest (optimal IPI) at 35 months for all countries combined, and at 30, 33, and 39 months in Denmark, Finland, and Sweden, respectively. Fully adjusted RRs at IPIs of 6, 12, and 60 months were 1.41 (95% CI: 1.08, 1.85), 1.26 (95% CI: 1.02, 1.56), and 1.24 (95% CI: 0.98, 1.58) compared to an IPI of 35 months. Under the most conservative scenario PIFs ranged from 5% (95% CI: 1%-8%) in Denmark to 9% (95% CI: 6%-12%) in Sweden. The minimum ASD risk followed IPIs of 30-39 months across three countries. These results reflect both direct IPI effects and other, closely related social and biological pathways. If our results reflect biologically causal effects, increasing optimal IPIs and reducing their indications, such as unintended pregnancy and delayed age at first pregnancy has the potential to prevent a salient proportion of ASD cases. LAY SUMMARY: Waiting 35 months to conceive again after giving birth resulted in the least risk of autism. Shorter and longer intervals resulted in risks that were up to 50% and 85% higher, respectively. About 5% to 9% of autism cases might be avoided by optimizing birth spacing.

Metformin extended-release versus metformin immediate-release for adults with type 2 diabetes mellitus: A systematic review and meta-analysis of randomized controlled trials

AIM

To compare the efficacy and tolerability of metformin extended-release (MXR) and the conventional metformin immediate-release (MIR) formulations in adults with type 2 diabetes mellitus (T2DM) METHODS: PubMed, the Cochrane Library, ClinicalTrials.gov and other sources were searched until 19 March 2021 for randomized controlled trials (RCTs) that compared equal daily doses of MXR and MIR in adults with T2DM. Random-effects model meta-analysis was performed to obtain pooled mean difference (MD) of change from baseline for continuous outcomes and risk ratio (RR) for dichotomous outcomes. Primary outcomes considered were HbA1c and key gastrointestinal (GI) symptoms (abdominal discomfort or pain, diarrhea, dyspepsia, and nausea & vomiting).

RESULTS

Nine RCTs that randomized a total of 2609 adults revealed that MIR was statistically associated with better HbA1c lowering (MD 0.09% [95% confidence interval or CI, 0.01%, 0.17%]), MXR only reduced dyspepsia (RR 0.58 [95% CI, 0.34, 0.98]), and both formulations were associated with similar cumulative incidence of other key GI symptoms.

CONCLUSIONS

MXR was associated with statistically worse but likely clinically similar HbA1c lowering and minimal improvement of GI intolerance compared to MIR. Protocol Registration: PROSPERO (CRD42019148008).

68Ga-PSMA PET/CT and mpMRI for primary lymph node staging of intermediate to high-risk prostate cancer: a systematic review and meta-analysis of diagnostic test accuracy

Purpose

To evaluate the diagnostic accuracy of Gallium-68 prostate-specific membrane antigen positron emission tomography-computed tomography (68Ga-PSMA PET/CT) compared with multiparametric magnetic resonance imaging (mpMRI) for detection of metastatic lymph nodes in intermediate to high-risk prostate cancer (PCa).

Methods

PRISMA-compliant systematic review updated to September 2020 was performed to identify studies that evaluated the diagnostic performance of 68Ga-PSMA PET/CT and mpMRI for detection of metastatic lymph nodes in the same cohort of PCa patients using histopathologic examination as a reference standard. The quality of each study was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) instrument. STATA version 16.0 was used to obtain the pooled estimates of diagnostic accuracy for per-patient and per-lesion analyses. Heterogeneity in the accuracy estimates was explored by reviewing the generated forest plots, summary receiver operator characteristic (SROC) curves, hierarchical SROC plots, chi-squared test, heterogeneity index, and Spearman’s correlation coefficients.

Results

Six studies, which included 476 patients, met the eligibility criteria for per-patient analysis and four of these studies, reporting data from 4859 dissected lymph nodes, were included in the per-lesion analysis. In the per-patient analysis (N = 6), the pooled sensitivity and specificity for 68Ga-PSMA PET/CT were 0.69 and 0.93, and for mpMRI the pooled sensitivity and specificity were 0.37 and 0.95. In the per-lesion analysis (N = 4), the pooled sensitivity and specificity for 68Ga-PSMA PET/CT were 0.58 and 0.99, and for mpMRI the pooled sensitivity and specificity were 0.44 and 0.99. There was high heterogeneity and a threshold effect in outcomes. A sensitivity analysis demonstrated that the pooled estimates were stable when excluding studies with patient selection concerns, whereas the variances of the pooled estimates became significant, and the characteristics of heterogeneity changed when excluding studies with concerns about index imaging tests.

Conclusion

Both imaging techniques have high specificity for the detection of nodal metastases of PCa. 68Ga-PSMA PET/CT has the advantage of being more sensitive and making it possible to detect distant metastases during the same examination. These modalities may play a complementary role in the diagnosis of PCa. Given the paucity of data and methodological limitations of the included studies, large scale trials are necessary to confirm their clinical values.

Effectiveness of collaborative care in reducing suicidal ideation: An individual participant data meta-analysis

To assess whether CC is more effective at reducing suicidal ideation in people with depression compared with usual care, and whether study and patient factors moderate treatment effects.

METHOD

We searched Medline, Embase, PubMed, PsycINFO, CINAHL, CENTRAL from inception to March 2020 for Randomised Controlled Trials (RCTs) that compared the effectiveness of CC with usual care in depressed adults, and reported changes in suicidal ideation at 4 to 6 months post-randomisation. Mixed-effects models accounted for clustering of participants within trials and heterogeneity across trials. This study is registered with PROSPERO, CRD42020201747.

RESULTS

We extracted data from 28 RCTs (11,165 patients) of 83 eligible studies. We observed a small significant clinical improvement of CC on suicidal ideation, compared with usual care (SMD, -0.11 [95%CI, -0.15 to -0.08]; I², 0·47% [95%CI 0.04% to 4.90%]). CC interventions with a recognised psychological treatment were associated with small reductions in suicidal ideation (SMD, -0.15 [95%CI -0.19 to -0.11]). CC was more effective for reducing suicidal ideation among patients aged over 65 years (SMD, - 0.18 [95%CI -0.25 to -0.11]).

CONCLUSION

Primary care based CC with an embedded psychological intervention is the most effective CC framework for reducing suicidal ideation and older patients may benefit the most.

Knee osteoarthritis and time-to all-cause mortality in six community-based cohorts: an international meta-analysis of individual participant-level data

BACKGROUND

Osteoarthritis (OA) is a chronic joint disease, with increasing global burden of disability and healthcare utilisation. Recent meta-analyses have shown a range of effects of OA on mortality, reflecting different OA definitions and study methods. We seek to overcome limitations introduced when using aggregate results by gathering individual participant-level data (IPD) from international observational studies and standardising methods to determine the association of knee OA with mortality in the general population.

METHODS

Seven community-based cohorts were identified containing knee OA-related pain, radiographs, and time-to-mortality, six of which were available for analysis. A two-stage IPD meta-analysis framework was applied: (1) Cox proportional hazard models assessed time-to-mortality of participants with radiographic OA (ROA), OA-related pain (POA), and a combination of pain and ROA (PROA) against pain and ROA-free participants; (2) hazard ratios (HR) were then pooled using the Hartung-Knapp modification for random-effects meta-analysis.

FINDINGS

10,723 participants in six cohorts from four countries were included in the analyses. Multivariable models (adjusting for age, sex, race, BMI, smoking, alcohol consumption, cardiovascular disease, and diabetes) showed a pooled HR, compared to pain and ROA-free participants, of 1.03 (0.83, 1.28) for ROA, 1.35 (1.12, 1.63) for POA, and 1.37 (1.22, 1.54) for PROA.

DISCUSSION

Participants with POA or PROA had a 35-37% increased association with reduced time-to-mortality, independent of confounders. ROA showed no association with mortality, suggesting that OA-related knee pain may be driving the association with time-to-mortality.

FUNDING

Versus Arthritis Centre for Sport, Exercise and Osteoarthritis and Osteoarthritis Research Society International.

Hypofractionated radiotherapy in locally advanced bladder cancer: an individual patient data meta-analysis of the BC2001 and BCON trials

BACKGROUND

Two radiotherapy fractionation schedules are used to treat locally advanced bladder cancer: 64 Gy in 32 fractions over 6·5 weeks and a hypofractionated schedule of 55 Gy in 20 fractions over 4 weeks. Long-term outcomes of these schedules in several cohort studies and case series suggest that response, survival, and toxicity are similar, but no direct comparison has been published. The present study aimed to assess the non-inferiority of 55 Gy in 20 fractions to 64 Gy in 32 fractions in terms of invasive locoregional control and late toxicity in patients with locally advanced bladder cancer.

METHODS

We did a meta-analysis of individual patient data from patients (age ≥18 years) with locally advanced bladder cancer (T1G3 [high-grade non-muscle invasive] or T2-T4, N0M0) enrolled in two multicentre, randomised, controlled, phase 3 trials done in the UK: BC2001 (NCT00024349; assessing addition of chemotherapy to radiotherapy) and BCON (NCT00033436; assessing hypoxia-modifying therapy combined with radiotherapy). In each trial, the fractionation schedule was chosen according to local standard practice. Co-primary endpoints were invasive locoregional control (non-inferiority margin hazard ratio [HR]=1·25); and late bladder or rectum toxicity, assessed with the Late Effects Normal Tissue Task Force-Subjective, Objective, Management, Analytic tool (non-inferiority margin for absolute risk difference [RD]=10%). If non-inferiority was met for invasive locoregional control, superiority could be considered if the 95% CI for the treatment effect excluded the null effect (HR=1). One-stage individual patient data meta-analysis models for the time-to-event and binary outcomes were used, accounting for trial differences, within-centre correlation, randomised treatment received, baseline variable imbalances, and potential confounding from relevant prognostic factors.

FINDINGS

782 patients with known fractionation schedules (456 from the BC2001 trial and 326 from the BCON trial; 376 (48%) received 64 Gy in 32 fractions and 406 (52%) received 55 Gy in 20 fractions) were included in our meta-analysis. Median follow-up was 120 months (IQR 99-159). Patients who received 55 Gy in 20 fractions had a lower risk of invasive locoregional recurrence than those who received 64 Gy in 32 fractions (adjusted HR 0·71 [95% CI 0·52-0·96]). Both schedules had similar toxicity profiles (adjusted RD -3·37% [95% CI -11·85 to 5·10]).

INTERPRETATION

A hypofractionated schedule of 55 Gy in 20 fractions is non-inferior to 64 Gy in 32 fractions with regard to both invasive locoregional control and toxicity, and is superior with regard to invasive locoregional control. 55 Gy in 20 fractions should be adopted as a standard of care for bladder preservation in patients with locally advanced bladder cancer.

FUNDING

Cancer Research UK.

Bayesian Meta-Analysis for Binary Data and Prior Distribution on Models

In meta-analysis, the structure of the between-sample heterogeneity plays a crucial role in estimating the meta-parameter. A Bayesian meta-analysis for binary data has recently been proposed that measures this heterogeneity by clustering the samples and then determining the posterior probability of the cluster models through model selection. The meta-parameter is then estimated using Bayesian model averaging techniques. Although an objective Bayesian meta-analysis is proposed for each type of heterogeneity, we concentrate the attention of this paper on priors over the models. We consider four alternative priors which are motivated by reasonable but different assumptions. A frequentist validation with simulated data has been carried out to analyze the properties of each prior distribution for a set of different number of studies and sample sizes. The results show the importance of choosing an adequate model prior as the posterior probabilities for the models are very sensitive to it. The hierarchical Poisson prior and the hierarchical uniform prior show a good performance when the real model is the homogeneity, or when the sample sizes are high enough. However, the uniform prior can detect the true model when it is an intermediate model (neither homogeneity nor heterogeneity) even for small sample sizes and few studies. An illustrative example with real data is also given, showing the sensitivity of the estimation of the meta-parameter to the model prior.

A brief introduction of meta-analyses in clinical practice and research

With the explosive growth of medical information, it is almost impossible for healthcare providers to review and evaluate all relevant evidence to make the best clinical decisions. Meta‐analyses, which summarize all existing evidence and quantitatively synthesize individual studies, have become the best available evidence for informing clinical practice. This article introduces the common methods, steps, principles, strengths and limitations of meta‐analyses and aims to help healthcare providers and researchers obtain a basic understanding of meta‐analyses in clinical practice and research.

Comparing methods for estimating patient-specific treatment effects in individual patient data meta-analysis

Meta‐analysis of individual patient data (IPD) is increasingly used to synthesize data from multiple trials. IPD meta‐analysis offers several advantages over meta‐analyzing aggregate data, including the capacity to individualize treatment recommendations. Trials usually collect information on many patient characteristics. Some of these covariates may strongly interact with treatment (and thus be associated with treatment effect modification) while others may have little effect. It is currently unclear whether a systematic approach to the selection of treatment‐covariate interactions in an IPD meta‐analysis can lead to better estimates of patient‐specific treatment effects. We aimed to answer this question by comparing in simulations the standard approach to IPD meta‐analysis (no variable selection, all treatment‐covariate interactions included in the model) with six alternative methods: stepwise regression, and five regression methods that perform shrinkage on treatment‐covariate interactions, that is, least absolute shrinkage and selection operator (LASSO), ridge, adaptive LASSO, Bayesian LASSO, and stochastic search variable selection. Exploring a range of scenarios, we found that shrinkage methods performed well for both continuous and dichotomous outcomes, for a variety of settings. In most scenarios, these methods gave lower mean squared error of the patient‐specific treatment effect as compared with the standard approach and stepwise regression. We illustrate the application of these methods in two datasets from cardiology and psychiatry. We recommend that future IPD meta‐analysis that aim to estimate patient‐specific treatment effects using multiple effect modifiers should use shrinkage methods, whereas stepwise regression should be avoided.

Altered immune phenotype and DNA methylation in panic disorder

Background

Multiple studies have related psychiatric disorders and immune alterations. Panic disorder (PD) has been linked with changes in leukocytes distributions in several small studies using different methods for immune characterization. Additionally, alterations in the methylation of repetitive DNA elements, such as LINE-1, have been associated with mental disorders. Here, we use peripheral blood DNA methylation data from two studies and an updated DNA methylation deconvolution library to investigate the relation of leukocyte proportions and methylation status of repetitive elements in 133 patients with panic disorder compared with 118 controls.

Methods and results

We used DNA methylation data to deconvolute leukocyte cell-type proportions and to infer LINE-1 element methylation comparing PD cases and controls. We also identified differentially methylated CpGs associated with PD using an epigenome-wide association study approach (EWAS), with models adjusting for sex, age, and cell-type proportions. Individuals with PD had a lower proportion of CD8T cells (OR: 0.86, 95% CI: 0.78–0.96, P-adj = 0.030) when adjusting for age, sex, and study compared with controls. Also, PD cases had significantly lower LINE-1 repetitive element methylation than controls (P < 0.001). The EWAS identified 61 differentially methylated CpGs (58 hypo- and 3 hypermethylated) in PD (Bonferroni adjusted P < 1.33 × 10^–7).

Conclusions

These results suggest that those with panic disorder have changes to their immune system and dysregulation of repeat elements relative to controls.

Approaches for combining primary care electronic health record data from multiple sources: a systematic review of observational studies

OBJECTIVE

To identify observational studies which used data from more than one primary care electronic health record (EHR) database, and summarise key characteristics including: objective and rationale for using multiple data sources; methods used to manage, analyse and (where applicable) combine data; and approaches used to assess and report heterogeneity between data sources.

DESIGN

A systematic review of published studies.

DATA SOURCES

Pubmed and Embase databases were searched using list of named primary care EHR databases; supplementary hand searches of reference list of studies were retained after initial screening.

STUDY SELECTION

Observational studies published between January 2000 and May 2018 were selected, which included at least two different primary care EHR databases.

RESULTS

6054 studies were identified from database and hand searches, and 109 were included in the final review, the majority published between 2014 and 2018. Included studies used 38 different primary care EHR data sources. Forty-seven studies (44%) were descriptive or methodological. Of 62 analytical studies, 22 (36%) presented separate results from each database, with no attempt to combine them; 29 (48%) combined individual patient data in a one-stage meta-analysis and 21 (34%) combined estimates from each database using two-stage meta-analysis. Discussion and exploration of heterogeneity was inconsistent across studies.

CONCLUSIONS

Comparing patterns and trends in different populations, or in different primary care EHR databases from the same populations, is important and a common objective for multi-database studies. When combining results from several databases using meta-analysis, provision of separate results from each database is helpful for interpretation. We found that these were often missing, particularly for studies using one-stage approaches, which also often lacked details of any statistical adjustment for heterogeneity and/or clustering. For two-stage meta-analysis, a clear rationale should be provided for choice of fixed effect and/or random effects or other models.

INFOMATAS multi-center systematic review and meta-analysis individual patient data of dynamic cerebral autoregulation in ischemic stroke

Rationale

Disturbances in dynamic cerebral autoregulation after ischemic stroke may have important implications for prognosis. Recent meta-analyses have been hampered by heterogeneity and small samples.

Aim and/or hypothesis

The aim of study is to undertake an individual patient data meta-analysis (IPD-MA) of dynamic cerebral autoregulation changes post-ischemic stroke and to determine a predictive model for outcome in ischemic stroke using information combined from dynamic cerebral autoregulation, clinical history, and neuroimaging.

Sample size estimates

To detect a change of 2% between categories in modified Rankin scale requires a sample size of ∼1500 patients with moderate to severe stroke, and a change of 1 in autoregulation index requires a sample size of 45 healthy individuals (powered at 80%, α = 0.05). Pooled estimates of mean and standard deviation derived from this study will be used to inform sample size calculations for adequately powered future dynamic cerebral autoregulation studies in ischemic stroke.

Methods and design

This is an IPD-MA as part of an international, multi-center collaboration (INFOMATAS) with three phases. Firstly, univariate analyses will be constructed for primary (modified Rankin scale) and secondary outcomes, with key co-variates and dynamic cerebral autoregulation parameters. Participants clustering from within studies will be accounted for with random effects. Secondly, dynamic cerebral autoregulation variables will be validated for diagnostic and prognostic accuracy in ischemic stroke using summary receiver operating characteristic curve analysis. Finally, the prognostic accuracy will be determined for four different models combining clinical history, neuroimaging, and dynamic cerebral autoregulation parameters.

Study outcome(s)

The outcomes for this study are to determine the relationship between clinical outcome, dynamic cerebral autoregulation changes, and baseline patient demographics, to determine the diagnostic and prognostic accuracy of dynamic cerebral autoregulation parameters, and to develop a prognostic model using dynamic cerebral autoregulation in ischemic stroke.

Discussion

This is the first international collaboration to use IPD-MA to determine prognostic models of dynamic cerebral autoregulation for patients with ischemic stroke.

Estimating treatment importance in multidrug-resistant tuberculosis using Targeted Learning: An observational individual patient data network meta-analysis

Persons with multidrug-resistant tuberculosis (MDR-TB) have a disease resulting from a strain of tuberculosis (TB) that does not respond to at least isoniazid and rifampicin, the two most effective anti-TB drugs. MDR-TB is always treated with multiple antimicrobial agents. Our data consist of individual patient data from 31 international observational studies with varying prescription practices, access to medications, and distributions of antibiotic resistance. In this study, we develop identifiability criteria for the estimation of a global treatment importance metric in the context where not all medications are observed in all studies. With stronger causal assumptions, this treatment importance metric can be interpreted as the effect of adding a medication to the existing treatments. We then use this metric to rank 15 observed antimicrobial agents in terms of their estimated add-on value. Using the concept of transportability, we propose an implementation of targeted maximum likelihood estimation, a doubly robust and locally efficient plug-in estimator, to estimate the treatment importance metric. A clustered sandwich estimator is adopted to compute variance estimates and produce confidence intervals. Simulation studies are conducted to assess the performance of our estimator, verify the double robustness property, and assess the appropriateness of the variance estimation approach.

An overview of systematic reviews of complementary and alternative therapies for infantile colic

BACKGROUND

Infantile colic is a distressing condition characterised by excessive crying in the first few months of life. The aim of this research was to update the synthesis of evidence of complementary and alternative medicine (CAM) research literature on infantile colic and establish what evidence is currently available.

METHODS

Medline, Embase and AMED (via Ovid), Web of Science and Central via Cochrane library were searched from their inception to September 2018. Google Scholar and OpenGrey were searched for grey literature and PROSPERO for ongoing reviews. Published systematic reviews that included randomised controlled trials (RCTs) of infants aged up to 1 year, diagnosed with infantile colic using standard diagnostic criteria, were eligible. Reviews of RCTs that assessed the effectiveness of any individual CAM therapy were included. Three reviewers were involved in data extraction and quality assessment using the AMSTAR-2 scale and risk of bias using the ROBIS tool.

RESULTS

Sixteen systematic reviews were identified. Probiotics, fennel extract and spinal manipulation show promise to alleviate symptoms of colic, although some concerns remain. Acupuncture and soy are currently not recommended. The majority of the reviews were assessed as having high or unclear risk of bias and low confidence in the findings.

CONCLUSION

There is clearly a need for larger and more methodologically sound RCTs to be conducted on the effectiveness of some CAM therapies for IC. Particular focus on probiotics in non-breastfed infants is pertinent.

SYSTEMATIC REVIEW REGISTRATION

PROSPERO: CRD42018092966.

Disaggregating the predictive effects of impaired psychosocial functioning on future DSM-5 eating disorder onset in high-risk female adolescents

OBJECTIVE

Impaired psychosocial functioning previously emerged as the only risk factor to predict future onset of each of the four Diagnostic and Statistical Manual of Mental Disorder (5th ed.) (DSM-5) eating disorders. The goal of this follow-up report was to refine understanding of this newly identified risk factor.

METHOD

Combining data from women at risk for eating disorders because of body image concerns (N = 1,153, mean age = 18.5 years, SD = 4.2), we investigated which subdomain(s) and individual item(s) of psychosocial functioning (friends, family, school, and work) at baseline predicted onset of any eating disorder, using Cox regression (CRA) and classification tree analysis (CTA).

RESULTS

Psychosocial impairment with friends, family, and at school, but not at work, significantly increased risk for disorder onset over 3-year follow-up in univariate models. A one-unit increase in each domain raw score was associated with a 107, 22, and 43% increased hazard of eating disorder onset, respectively. Multivariate CRA found friends functioning, with a 92% increased hazard of disorder onset, contributed the strongest unique effect. CTA suggested that loneliness was the most potent risk factor with a threefold increased onset risk (eating disorder incidence for high vs. low scorers was 27 and 8%). Three friends functioning items and one school functioning item produced additional CTA branches.

DISCUSSION

Results refine understanding of the relation of psychosocial impairment to future onset of eating disorders, suggesting that peer functioning is the most critical. Data imply it would be useful to target young women with impaired psychosocial functioning in prevention programs.

Effect of recent seasonal influenza vaccination on serum antibody responses to candidate pandemic influenza A/H5N1 vaccines: A meta-analysis

Recent studies have suggested that among those receiving seasonal influenza vaccine (SIV), reduced immunogenicity is observed in recently vaccinated (RV; within the past season or 2) persons when compared with those not recently vaccinated (NRV). We performed a meta-analysis to assess the effect of recent immunization with SIV on serum H5 hemagglutination inhibition (HAI) antibody responses after influenza A/H5N1 vaccination using data from a series of randomized controlled trials. The primary outcome was seroconversion measured by HAI assays following receipt of 2 doses of H5N1 vaccine. The geometric mean titer (GMT) of serum HAI antibody after vaccination was the secondary outcome. Analyses were performed using propensity score (PS) matching. The PS for each individual in the meta-analysis cohort was calculated using logistic regression and covariates included age, gender, race, antigen dose, adjuvant, statin use and vaccine manufacturer. 2015 subjects enrolled in 7 clinical trials were eligible for inclusion in the meta-analysis cohort; among these, 915 (45%) were RV. 901 RV subjects were matched (1:1) with replacement to a subject who was NRV. Subjects who received SIV within the previous season were significantly less likely to seroconvert following H5N1 vaccination (adjusted odds ratio 0.76; 95%CI 0.60-0.96; p = 0.024), and the GMT was 18% higher among NRV subjects (GM ratio of HAI antibody 1.18; 95%CI 1.04-1.33; p = 0.008). Further work is needed to better define the effects of, and mechanisms contributing to, reduced immune responses to H5N1 vaccine among RV subjects.

An Empirical Comparison of Meta- and Mega-Analysis With Data From the ENIGMA Obsessive-Compulsive Disorder Working Group

Objective: Brain imaging communities focusing on different diseases have increasingly started to collaborate and to pool data to perform well-powered meta- and mega-analyses. Some methodologists claim that a one-stage individual-participant data (IPD) mega-analysis can be superior to a two-stage aggregated data meta-analysis, since more detailed computations can be performed in a mega-analysis. Before definitive conclusions regarding the performance of either method can be drawn, it is necessary to critically evaluate the methodology of, and results obtained by, meta- and mega-analyses.

Methods: Here, we compare the inverse variance weighted random-effect meta-analysis model with a multiple linear regression mega-analysis model, as well as with a linear mixed-effects random-intercept mega-analysis model, using data from 38 cohorts including 3,665 participants of the ENIGMA-OCD consortium. We assessed the effect sizes and standard errors, and the fit of the models, to evaluate the performance of the different methods.

Results: The mega-analytical models showed lower standard errors and narrower confidence intervals than the meta-analysis. Similar standard errors and confidence intervals were found for the linear regression and linear mixed-effects random-intercept models. Moreover, the linear mixed-effects random-intercept models showed better fit indices compared to linear regression mega-analytical models.

Conclusions: Our findings indicate that results obtained by meta- and mega-analysis differ, in favor of the latter. In multi-center studies with a moderate amount of variation between cohorts, a linear mixed-effects random-intercept mega-analytical framework appears to be the better approach to investigate structural neuroimaging data.

Outcomes of oxygen saturation targeting during delivery room stabilisation of preterm infants

OBJECTIVE

To determine the association between SpO₂ at 5 min and preterm infant outcomes.

DESIGN

Data from 768 infants <32 weeks gestation from 8 randomised controlled trials (RCTs) of lower (≤0.3) versus higher (≥0.6) initial inspiratory fractions of oxygen (FiO₂) for resuscitation, were examined.

SETTING

Individual patient analysis of 8 RCTs INTERVENTIONS: Lower (≤0.3) versus higher (≥0.6) oxygen resuscitation strategies targeted to specific predefined SpO₂ before 10 min of age.

PATIENTS

Infants <32 weeks gestation.

MAIN OUTCOME MEASURES

Relationship between SpO₂ at 5 min, death and intraventricular haemorrhage (IVH) >grade 3.

RESULTS

5 min SpO₂ data were obtained from 706 (92%) infants. Only 159 (23%) infants met SpO₂ study targets and 323 (46%) did not reach SpO₂80%. Pooled data showed decreased likelihood of reaching SpO₂80% if resuscitation was initiated with FiO₂ <0.3 (OR 2.63, 95% CI 1.21 to 5.74, p<0.05). SpO₂ <80% was associated with lower heart rates (mean difference -8.37, 95% CI -15.73 to -1.01, *p<0.05) and after accounting for confounders, with IVH (OR 2.04, 95% CI 1.01 to 4.11, p<0.05). Bradycardia (heart rate <100 bpm) at 5 min increased risk of death (OR 4.57, 95% CI 1.62 to 13.98, p<0.05). Taking into account confounders including gestation, birth weight and 5 min bradycardia, risk of death was significantly increased with time taken to reach SpO₂80%.

CONCLUSION

Not reaching SpO₂80% at 5 min is associated with adverse outcomes, including IVH. Whether this is because of infant illness or the amount of oxygen that is administered during stabilisation is uncertain and needs to be examined in randomised trials.

Association of longitudinal alcohol consumption trajectories with coronary heart disease: a meta-analysis of six cohort studies using individual participant data

BACKGROUND

Studies have shown that alcohol intake trajectories differ in their associations with biomarkers of cardiovascular functioning, but it remains unclear if they also differ in their relationship to actual coronary heart disease (CHD) incidence. Using multiple longitudinal cohort studies, we evaluated the association between long-term alcohol consumption trajectories and CHD.

METHODS

Data were drawn from six cohorts (five British and one French). The combined analytic sample comprised 35,132 individuals (62.1% male; individual cohorts ranging from 869 to 14,247 participants) of whom 4.9% experienced an incident (fatal or non-fatal) CHD event. Alcohol intake across three assessment periods of each cohort was used to determine participants' intake trajectories over approximately 10 years. Time to onset for (i) incident CHD and (ii) fatal CHD was established using surveys and linked medical record data. A meta-analysis of individual participant data was employed to estimate the intake trajectories' association with CHD onset, adjusting for demographic and clinical characteristics.

RESULTS

Compared to consistently moderate drinkers (males: 1-168 g ethanol/week; females: 1-112 g ethanol/week), inconsistently moderate drinkers had a significantly greater risk of incident CHD [hazard ratio (HR) = 1.18, 95% confidence interval (CI) = 1.02-1.37]. An elevated risk of incident CHD was also found for former drinkers (HR = 1.31, 95% CI = 1.13-1.52) and consistent non-drinkers (HR = 1.47, 95% CI = 1.21-1.78), although, after sex stratification, the latter effect was only evident for females. When examining fatal CHD outcomes alone, only former drinkers had a significantly elevated risk, though hazard ratios for consistent non-drinkers were near identical. No evidence of elevated CHD risk was found for consistently heavy drinkers, and a weak association with fatal CHD for inconsistently heavy drinkers was attenuated following adjustment for confounding factors.

CONCLUSIONS

Using prospectively recorded alcohol data, this study has shown how instability in drinking behaviours over time is associated with risk of CHD. As well as individuals who abstain from drinking (long term or more recently), those who are inconsistently moderate in their alcohol intake have a higher risk of experiencing CHD. This finding suggests that policies and interventions specifically encouraging consistency in adherence to lower-risk drinking guidelines could have public health benefits in reducing the population burden of CHD. The absence of an effect amongst heavy drinkers should be interpreted with caution given the known wider health risks associated with such intake.

TRIAL REGISTRATION

ClinicalTrials.gov, NCT03133689 .

Transfusion of packed red blood cells at the end of shelf life is associated with increased risk of mortality - a pooled patient data analysis of 16 observational trials

Observational studies address packed red blood cell effects at the end of shelf life and have larger sample sizes compared to randomized control trials. Meta-analyses combining data from observational studies have been complicated by differences in aggregate transfused packed red blood cell age and outcome reporting. This study abrogated these issues by taking a pooled patient data approach. Observational studies reporting packed red blood cell age and clinical outcomes were identified and patient-level data sets were sought from investigators. Odds ratios and 95% confidence intervals for binary outcomes were calculated for each study, with mean packed red blood cell age or maximum packed red blood cell age acting as independent variables. The relationship between mean packed red blood cell age and hospital length of stay for each paper was analyzed using zero-inflated Poisson regression. Random effects models combined paper-level effect estimates. Extremes analyses were completed by comparing patients transfused with mean packed red blood cell aged less than ten days to those transfused with mean packed red blood cell aged at least 30 days. sixteen datasets were available for pooled patient data analysis. Mean packed red blood cell age of at least 30 days was associated with an increased risk of in-hospital mortality compared to mean packed red blood cell of less than ten days (odds ratio: 3.25, 95% confidence interval: 1.27–8.29). Packed red blood cell age was not correlated to increased risks of nosocomial infection or prolonged length of hospital stay.

Microbial Preparations (Probiotics) for the Prevention of Clostridium difficile Infection in Adults and Children: An Individual Patient Data Meta-analysis of 6,851 Participants

OBJECTIVETo determine whether probiotic prophylaxes reduce the odds of Clostridium difficile infection (CDI) in adults and children.DESIGNIndividual participant data (IPD) meta-analysis of randomized controlled trials (RCTs), adjusting for risk factors.METHODSWe searched 6 databases and 11 grey literature sources from inception to April 2016. We identified 32 RCTs (n=8,713); among them, 18 RCTs provided IPD (n=6,851 participants) comparing probiotic prophylaxis to placebo or no treatment (standard care). One reviewer prepared the IPD, and 2 reviewers extracted data, rated study quality, and graded evidence quality.RESULTSProbiotics reduced CDI odds in the unadjusted model (n=6,645; odds ratio [OR] 0.37; 95% confidence interval [CI], 0.25-0.55) and the adjusted model (n=5,074; OR, 0.35; 95% CI, 0.23-0.55). Using 2 or more antibiotics increased the odds of CDI (OR, 2.20; 95% CI, 1.11-4.37), whereas age, sex, hospitalization status, and high-risk antibiotic exposure did not. Adjusted subgroup analyses suggested that, compared to no probiotics, multispecies probiotics were more beneficial than single-species probiotics, as was using probiotics in clinical settings where the CDI risk is ≥5%. Of 18 studies, 14 reported adverse events. In 11 of these 14 studies, the adverse events were retained in the adjusted model. Odds for serious adverse events were similar for both groups in the unadjusted analyses (n=4,990; OR, 1.06; 95% CI, 0.89-1.26) and adjusted analyses (n=4,718; OR, 1.06; 95% CI, 0.89-1.28). Missing outcome data for CDI ranged from 0% to 25.8%. Our analyses were robust to a sensitivity analysis for missingness.CONCLUSIONSModerate quality (ie, certainty) evidence suggests that probiotic prophylaxis may be a useful and safe CDI prevention strategy, particularly among participants taking 2 or more antibiotics and in hospital settings where the risk of CDI is ≥5%.TRIAL REGISTRATIONPROSPERO 2015 identifier: CRD42015015701Infect Control Hosp Epidemiol 2018;771-781.

Apathy and depressive symptoms in older people and incident myocardial infarction, stroke, and mortality: a systematic review and meta-analysis of individual participant data

Background

Previous findings suggest that apathy symptoms independently of depressive symptoms measured using the Geriatric Depression Scale (GDS) are associated with cardiovascular disease (CVD) in older individuals.

Aims

To study whether apathy and depressive symptoms in older people are associated with future CVD, stroke, and mortality using individual patient-data meta-analysis.

Methods

Medline, Embase, and PsycInfo databases up to September 3, 2013, were systematically searched without language restrictions. We sought prospective studies with older (mean age ≥65 years) community-dwelling populations in which the GDS was employed and subsequent stroke and/or CVD were recorded to provide individual participant data. Apathy symptoms were defined as the three apathy-related subitems of the GDS, with depressive symptoms the remaining items. We used myocardial infarction (MI), stroke, and all-cause mortality as main outcomes. Analyses were adjusted for age, sex, and MI/stroke history. An adaptation of the Newcastle–Ottawa scale was used to evaluate bias. Hazard ratios were calculated using one-stage random-effect Cox regression models.

Results

Of the 52 eligible studies, 21 (40.4%) were included, comprising 47,625 older people (mean age [standard deviation] 74 [7.4] years), over a median follow-up of 8.8 years. Participants with apathy symptoms had a 21% higher risk of MI (95% confidence interval [CI] 1.08–1.36), a 37% higher risk of stroke (95% CI 1.18–1.59), and a 47% higher risk of all-cause mortality (95% CI 1.38–1.56). Participants with depressive symptoms had a comparably higher risk of stroke (HR 1.36, 95% CI 1.18–1.56) and all-cause mortality (HR 1.44, 95% CI 1.35–1.53), but not of MI (HR 1.08, 95% CI 0.91–1.29). Associations for isolated apathy and isolated depressive symptoms were comparable. Sensitivity analyses according to risk of bias yielded similar results.

Conclusion

Our findings stress the clinical importance of recognizing apathy independently of depressive symptoms, and could help physicians identify persons at increased risk of vascular disease.

Patterns of association of chronic medical conditions and major depression

AIMS

Age and sex-related patterns of association between medical conditions and major depressive episodes (MDE) are important for understanding disease burden, anticipating clinical needs and for formulating etiological hypotheses. General population estimates are especially valuable because they are not distorted by help-seeking behaviours. However, even large population surveys often deliver inadequate precision to adequately describe such patterns. In this study, data from a set of national surveys were pooled to increase precision, supporting more precise characterisation of these associations.

METHODS

The data were from a series of Canadian national surveys. These surveys used comparable sampling strategies and assessment methods for MDE. Chronic medical conditions were assessed using items asking about professionally diagnosed medical conditions. Individual-level meta-analysis methods were used to generate unadjusted, stratified and adjusted prevalence odds ratios for 11 chronic medical conditions. Random effects models were used in the meta-analysis. A procedure incorporating rescaled replicate bootstrap weights was used to produce 95% confidence intervals.

RESULTS

Overall, conditions characterised by pain and inflammation tended to show stronger associations with MDE. The meta-analysis uncovered two previously undescribed patterns of association. Effect modification by age was observed in varying degrees for most conditions. This effect was most prominent for high blood pressure and cancer. Stronger associations were found in younger age categories. Migraine was an exception: the strength of association increased with age, especially in men. Second, especially for conditions predominantly affecting older age groups (arthritis, diabetes, back pain, cataracts, effects of stroke and heart disease) confounding by age was evident. For each condition, age adjustment resulted in strengthening of the associations. In addition to migraine, two conditions displayed distinctive patterns of association. Age adjusted odds ratios for thyroid disease reflected a weak association that was only significant in women. In epilepsy, a similar strength of association was found irrespective of age or sex.

CONCLUSIONS

The prevalence of MDE is elevated in association with most chronic conditions, but especially those characterised by inflammation and pain. Effect modification by age may reflect greater challenges or difficulties encountered by young people attempting to cope with these conditions. This pattern, however, does not apply to migraine or epilepsy. Neurobiological changes associated with these conditions may offset coping-related effects, such that the association does not weaken with age. Prominent confounding by age for several conditions suggests that age adjustments are necessary in order to avoid underestimating the strength of these associations.

Prognostic Value of High-Sensitivity Troponin T in Chronic Heart Failure

Background:

Most patients with chronic heart failure have detectable troponin concentrations when evaluated by high-sensitivity assays. The prognostic relevance of this finding has not been clearly established so far. We aimed to assess high-sensitivity troponin assay for risk stratification in chronic heart failure through a meta-analysis approach.

Methods:

Medline, EMBASE, Cochrane Library, and Scopus were searched in April 2017 by 2 independent authors. The terms were “troponin” AND “heart failure” OR “cardiac failure” OR “cardiac dysfunction” OR “cardiac insufficiency” OR “left ventricular dysfunction.” Inclusion criteria were English language, clinical stability, use of a high-sensitivity troponin assay, follow-up studies, and availability of individual patient data after request to authors. Data retrieved from articles and provided by authors were used in agreement with the PRISMA statement. The end points were all-cause death, cardiovascular death, and hospitalization for cardiovascular cause.

Results:

Ten studies were included, reporting data on 11 cohorts and 9289 patients (age 66±12 years, 77% men, 60% ischemic heart failure, 85% with left ventricular ejection fraction <40%). High-sensitivity troponin T data were available for all patients, whereas only 209 patients also had high-sensitivity troponin I assayed. When added to a prognostic model including established risk markers (sex, age, ischemic versus nonischemic etiology, left ventricular ejection fraction, estimated glomerular filtration rate, and N-terminal fraction of pro-B-type natriuretic peptide), high-sensitivity troponin T remained independently associated with all-cause mortality (hazard ratio, 1.48; 95% confidence interval, 1.41–1.55), cardiovascular mortality (hazard ratio, 1.40; 95% confidence interval, 1.33–1.48), and cardiovascular hospitalization (hazard ratio, 1.42; 95% confidence interval, 1.36–1.49), over a median 2.4-year follow-up (all P <0.001). High-sensitivity troponin T significantly improved risk prediction when added to a prognostic model including the variables above. It also displayed an independent prognostic value for all outcomes in almost all population subgroups. The area under the curve–derived 18 ng/L cutoff yielded independent prognostic value for the 3 end points in both men and women, patients with either ischemic or nonischemic etiology, and across categories of renal dysfunction.

Conclusions:

In chronic heart failure, high-sensitivity troponin T is a strong and independent predictor of all-cause and cardiovascular mortality, and of hospitalization for cardiovascular causes, as well. This biomarker then represents an additional tool for prognostic stratification.

Major depression and secondhand smoke exposure

BACKGROUND

Epidemiological studies have consistently linked smoking to poor mental health. Among non-smokers, some studies have also reported associations between secondhand smoke exposure and psychological symptoms. However, an association between secondhand smoke exposure and depressive disorders has not been well established.

METHODS

This analysis used cross-sectional data from a series of 10 population surveys conducted in Canada between 2003 and 2013. The surveys targeted the Canadian household population, included a brief structured interview for past year major depressive episode (MDE) and included items assessing secondhand smoke exposure. We used two-stage individual-level random-effects meta-regression to synthesize results from these surveys.

RESULTS

Over the study interval, about 20% of non-smokers reported substantial exposure to secondhand smoke. In this group, the pooled annual prevalence of MDE was 6.1% (95% CI 5.3-6.9) compared to 4.0% (95% CI 3.7-4.3) in non-smokers without secondhand smoke exposure. The crude odds ratio was 1.5 (95% CI 1.4-1.7). With adjustment for a set of potential confounding variables the odds ratio was unchanged, 1.4 (95% CI 1.2 - 1.6).

CONCLUSIONS

These results provide additional support for public health measures aimed at reducing secondhand smoke exposure. A causal connection between secondhand smoke exposure and MDEs cannot be confirmed due to the cross-sectional nature of the data. Longitudinal studies are needed to establish temporal sequencing.

Major depression prevention effects for a cognitive-behavioral adolescent indicated prevention group intervention across four trials

Major depressive disorder (MDD) in young people is a leading cause of disability but most depressed youth are not treated, emphasizing the need for effective prevention. Our goal is to synthesize MDD onset prevention effects for the Blues Program, a brief cognitive-behavioral (CB) indicated prevention group, by merging data from four trials (three of which included CB bibliotherapy) and conducting an individual patient data (IPD) meta-analysis. Data were available from 766 high school/college students (M age = 16.4, SD = 2.3; 60% female, 64% White). CB group resulted in significantly lower MDD incidence rates relative to brochure control that persisted to 6-month follow-up; CB group also was associated with a lower 2-year MDD incidence rate relative to bibliotherapy but heterogeneity across trials was detected. Contrasts between bibliotherapy and brochure control were nonsignificant. For significant contrasts, the number needed to treat (NNT) by CB group to prevent one MDD onset relative to brochure or bibliotherapy ranged from 10 to 21. A brief CB group depression prevention intervention for at-risk adolescent is achieving meaningful effects compared to both active and minimal controls but outcomes need to be improved, perhaps by better screening or augmentations to produce more persistent intervention effects.

Quantifying heterogeneity in individual participant data meta-analysis with binary outcomes

BACKGROUND

In meta-analyses (MA), effect estimates that are pooled together will often be heterogeneous. Determining how substantial heterogeneity is is an important aspect of MA.

METHOD

We consider how best to quantify heterogeneity in the context of individual participant data meta-analysis (IPD-MA) of binary data. Both two- and one-stage approaches are evaluated via simulation study. We consider conventional I ² and R ² statistics estimated via a two-stage approach and R ² estimated via a one-stage approach. We propose a simulation-based intraclass correlation coefficient (ICC) adapted from Goldstein et al. to estimate the I ², from the one-stage approach.

RESULTS

Results show that when there is no effect modification, the estimated I ² from the two-stage model is underestimated, while in the one-stage model, it is overestimated. In the presence of effect modification, the estimated I ² from the one-stage model has better performance than that from the two-stage model when the prevalence of the outcome is high. The I ² from the two-stage model is less sensitive to the strength of effect modification when the number of studies is large and prevalence is low.

CONCLUSIONS

The simulation-based I ² based on a one-stage approach has better performance than the conventional I ² based on a two-stage approach when there is strong effect modification with high prevalence.

Major depression and non-specific distress following smoking cessation in the Canadian general population

BACKGROUND

Outcome data from smoking cessation trials indicate that improvement in mental health occurs after smoking cessation. This suggests that smoking cessation should be a priority for mental health services. However, participants in such trials may not be representative of the general population. This study investigates changes in mental health following smoking cessation in a set of general population samples.

METHODS

Data from a library of cross-sectional surveys conducted by Statistics Canada between 2001 and 2013 were included in this analysis. Survey estimates were pooled in order to increase precision. Associations between smoking (and smoking cessation), major depressive episodes (MDE) and non-specific distress (assessed using the K-6 scale) were evaluated using meta-analysis and meta-regression techniques.

RESULTS

The annual prevalence of major depression was higher in daily (11.0%) than in never smokers (4.4%). The prevalence in former daily smokers was 5.1%. The prevalence of MDE and distress was elevated in those recently quitting but returned to baseline levels within one year.

CONCLUSIONS

After smoking cessation, indicators of mental health improve over time, especially in the first year. The findings support the idea that smoking cessation should be a part of the management of common mood and anxiety disorders. However, due to its observational nature this study in itself cannot confirm causality, sustained abstinence may be an effect of improved mental health rather than its cause.

LIMITATIONS

The cross-sectional nature of the constituent surveys does not allow causal inference. No biological measures (e.g. cotinine) were available.

Malaria, malnutrition, and birthweight: A meta-analysis using individual participant data

BACKGROUND

Four studies previously indicated that the effect of malaria infection during pregnancy on the risk of low birthweight (LBW; <2,500 g) may depend upon maternal nutritional status. We investigated this dependence further using a large, diverse study population.

METHODS AND FINDINGS

We evaluated the interaction between maternal malaria infection and maternal anthropometric status on the risk of LBW using pooled data from 14,633 pregnancies from 13 studies (6 cohort studies and 7 randomized controlled trials) conducted in Africa and the Western Pacific from 1996-2015. Studies were identified by the Maternal Malaria and Malnutrition (M3) initiative using a convenience sampling approach and were eligible for pooling given adequate ethical approval and availability of essential variables. Study-specific adjusted effect estimates were calculated using inverse probability of treatment-weighted linear and log-binomial regression models and pooled using a random-effects model. The adjusted risk of delivering a baby with LBW was 8.8% among women with malaria infection at antenatal enrollment compared to 7.7% among uninfected women (adjusted risk ratio [aRR] 1.14 [95% confidence interval (CI): 0.91, 1.42]; N = 13,613), 10.5% among women with malaria infection at delivery compared to 7.9% among uninfected women (aRR 1.32 [95% CI: 1.08, 1.62]; N = 11,826), and 15.3% among women with low mid-upper arm circumference (MUAC <23 cm) at enrollment compared to 9.5% among women with MUAC ≥ 23 cm (aRR 1.60 [95% CI: 1.36, 1.87]; N = 9,008). The risk of delivering a baby with LBW was 17.8% among women with both malaria infection and low MUAC at enrollment compared to 8.4% among uninfected women with MUAC ≥ 23 cm (joint aRR 2.13 [95% CI: 1.21, 3.73]; N = 8,152). There was no evidence of synergism (i.e., excess risk due to interaction) between malaria infection and MUAC on the multiplicative (p = 0.5) or additive scale (p = 0.9). Results were similar using body mass index (BMI) as an anthropometric indicator of nutritional status. Meta-regression results indicated that there may be multiplicative interaction between malaria infection at enrollment and low MUAC within studies conducted in Africa; however, this finding was not consistent on the additive scale, when accounting for multiple comparisons, or when using other definitions of malaria and malnutrition. The major limitations of the study included availability of only 2 cross-sectional measurements of malaria and the limited availability of ultrasound-based pregnancy dating to assess impacts on preterm birth and fetal growth in all studies.

CONCLUSIONS

Pregnant women with malnutrition and malaria infection are at increased risk of LBW compared to women with only 1 risk factor or none, but malaria and malnutrition do not act synergistically.

Automatic identification of variables in epidemiological datasets using logic regression

Background

For an individual participant data (IPD) meta-analysis, multiple datasets must be transformed in a consistent format, e.g. using uniform variable names. When large numbers of datasets have to be processed, this can be a time-consuming and error-prone task. Automated or semi-automated identification of variables can help to reduce the workload and improve the data quality. For semi-automation high sensitivity in the recognition of matching variables is particularly important, because it allows creating software which for a target variable presents a choice of source variables, from which a user can choose the matching one, with only low risk of having missed a correct source variable.

Methods

For each variable in a set of target variables, a number of simple rules were manually created. With logic regression, an optimal Boolean combination of these rules was searched for every target variable, using a random subset of a large database of epidemiological and clinical cohort data (construction subset). In a second subset of this database (validation subset), this optimal combination rules were validated.

Results

In the construction sample, 41 target variables were allocated on average with a positive predictive value (PPV) of 34%, and a negative predictive value (NPV) of 95%. In the validation sample, PPV was 33%, whereas NPV remained at 94%. In the construction sample, PPV was 50% or less in 63% of all variables, in the validation sample in 71% of all variables.

Conclusions

We demonstrated that the application of logic regression in a complex data management task in large epidemiological IPD meta-analyses is feasible. However, the performance of the algorithm is poor, which may require backup strategies.

Electronic supplementary material

The online version of this article (doi:10.1186/s12911-017-0429-1) contains supplementary material, which is available to authorized users.