Dataset Information

Multivariate meta-analysis for non-linear and other multi-parameter associations.

ABSTRACT: In this paper, we formalize the application of multivariate meta-analysis and meta-regression to synthesize estimates of multi-parameter associations obtained from different studies. This modelling approach extends the standard two-stage analysis used to combine results across different sub-groups or populations. The most straightforward application is for the meta-analysis of non-linear relationships, described for example by regression coefficients of splines or other functions, but the methodology easily generalizes to any setting where complex associations are described by multiple correlated parameters. The modelling framework of multivariate meta-analysis is implemented in the package mvmeta within the statistical environment R. As an illustrative example, we propose a two-stage analysis for investigating the non-linear exposure-response relationship between temperature and non-accidental mortality using time-series data from multiple cities. Multivariate meta-analysis represents a useful analytical tool for studying complex associations through a two-stage procedure.

SUBMITTER: Gasparrini A

PROVIDER: S-EPMC3546395 | biostudies-literature | 2012 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Multivariate meta-analysis for non-linear and other multi-parameter associations.

Gasparrini A A Armstrong B B Kenward M G MG

Statistics in medicine 20120716 29

In this paper, we formalize the application of multivariate meta-analysis and meta-regression to synthesize estimates of multi-parameter associations obtained from different studies. This modelling approach extends the standard two-stage analysis used to combine results across different sub-groups or populations. The most straightforward application is for the meta-analysis of non-linear relationships, described for example by regression coefficients of splines or other functions, but the method ...[more]

PMID: 22807043

Similar Datasets

Project description:OBJECTIVE:Delayed response to clinical deterioration as a result of intermittent vital sign monitoring is a cause of preventable morbidity and mortality. This review focuses on the clinical impact of multi-parameter continuous non-invasive monitoring of vital signs (CoNiM) in non-intensive care unit patients. DESIGN:Systematic review and meta-analysis of primary studies. Embase, MEDLINE, HMIC, PsycINFO and Cochrane were searched from April 1964 to 18 June 2019 with no language restriction. SETTING:The search was limited to hospitalised, non-intensive care unit adult patients who had two or more vital signs continuously monitored. PARTICIPANTS:All primary studies that evaluated the clinical impact of using multi-parameter CoNiM in adult hospital wards outside of the intensive care unit. MAIN OUTCOME MEASURES:Clinical impact of multi-parameter CoNiM. RESULTS:This systematic review identified 14 relevant studies from 3846 search results. Five studies were classified as Group A - associations found between measured vital signs and clinical parameters. Nine studies were classified as Group B - comparison between clinical outcomes of patients with and without multi-parameter CoNiM. Vital signs data from CoNiM were found to associate with type of presenting complaint, level of renal function and incidence of major clinical events. CoNiM also assisted in diagnosis by differentiating between patients with acute heart failure, stroke and sepsis (with sub-clustering of septic patients). In the meta-analysis, patients on multi-parameter CoNiM had a 39% decrease in risk of mortality (risk ratio [RR] 0.61; 95% confidence interval [95% CI] -0.39-0.95) when compared to patients with regular intermittent monitoring. There was a trend of reduced intensive care unit transfer (RR 0.86; 95% CI -0.67-1.11) and reduced rapid response team activation (RR 0.61; 95% CI 0.26-1.43). A trend towards reduced hospital length of stay was also found using weighted mean difference (WMD -3.32 days; 95% CI -8.82-2.19 days). CONCLUSION:There is evidence of clinical benefit in implementing CoNiM in non-intensive care unit patients. This review supports the use of multi-parameter CoNiM outside of intensive care unit with further large-scale RCTs required to further affirm clinical impact.

Project description:BackgroundOverweight/obesity remains a major risk factor for non-communicable diseases and their associated morbidities and mortalities. Yet, limited studies have comprehensively examined factors contributing to the rural-urban disparities in overweight/obesity among women in sub-Saharan Africa. Thus, our study sought to decompose the rural-urban disparities in overweight/obesity among women in sub-Saharan Africa (SSA) using nationally representative datasets.MethodsWe performed a cross-sectional analysis of data from the Demographic and Health Surveys of 23 sub-Saharan African countries conducted from 2015 to 2022. A sample of 177,329 women was included in the analysis. Percentages with confidence intervals (CIs) were used to summarize the prevalence of overweight/obesity per rural-urban strata and pooled level. A multivariate non-linear decomposition analysis was used to identify the factors contributing to the rural-urban disparities in overweight/obesity. The results were presented using coefficients and percentages.ResultsThe pooled prevalence of overweight/obesity among the women was higher in urban areas (38.9%; 95% CI = 38.2-39.6) than rural areas (19.1%; 95% CI = 18.7-19.6). This pattern was observed in all the countries surveyed, except in South Africa, where women in rural areas (53.1%; 95% CI = 50.0-56.4) had a higher prevalence of overweight/obesity than those in urban areas (46.0%; 95% CI = 43.2-48.9). Approximately 54% of the rural-urban disparities in overweight/obesity was attributable to the differences in the women's characteristics or explanatory variables. More than half of the rural-urban disparities in overweight/obesity would be reduced if the disparities in women's characteristics were levelled. Among the women's characteristics, frequency of watching television (29.03%), wealth index (26.59%), and level of education (9.40%) explained approximately 65% of the rural-urban differences in overweight/obesity.ConclusionThe prevalence of overweight/obesity among women in SSA remains high and skewed towards women in urban areas. Increased frequency of watching television, high wealth index, and higher educational attainment contributed largely to the rural-urban disparities in overweight/obesity among women in SSA. Thus, interventions aimed at reducing overweight/obesity among women in SSA could be targeted at reducing the frequency of television watching as well as promoting physical activities among wealthy women and those with higher education, particularly in urban areas.

Dataset Information

Multivariate meta-analysis for non-linear and other multi-parameter associations.

Publications

Multivariate meta-analysis for non-linear and other multi-parameter associations.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets