Meta Analysis Nursing Literature

Nursing Literature and Meta Analysis

What is Meta-Analysis,Experimental and Control Groups,Approaches for Meta Analysis,Meta Analysis and Sample Size, Nonindependences and Its Sources,Specific Assumptions,Meta Analysis in Nursing Literature,Benefits of Meta Analysis in Nursing Research.

What is Meta-Analysis

    Meta-analysis is a quantitative approach that permits the synthesis and integration of results from multiple individual studies focused on a specific research question. Meta analysis was first introduced in 1976 by Glass, who referred to it as an analysis of analyses. 

    A meta-analysis is a rigorous alternative to the traditional narrative review of the literature. It involves the application of the research process to a collection of studies in a specific area. The individual studies are considered the sample. 

    The findings from each study are transformed into a common statistic called an effect size. An effect size is a measure of the magnitude of the experimental effect on outcome variables.

Experimental and Control Groups

    Once the results from each study have been converted to a common metric, these findings can be pooled together and synthesized. The most common effect size indicator is r, which is the Pearson product moment correlation. 

    Another effect size indicator is the d index. Cohen's d is the difference between the means of the experimental and control groups divided by the standard deviation. Cohen (1988) has provided guidelines for interpreting the magnitude of both the r and d effect size indicators. 

    For the r index, Cohen has defined small, medium, and large effect sizes as.10, .30, and .50 or more, respectively. For the d indicator an effect size of 2 is considered small. 5 is medium, and 8 or more is large.

Approaches for Meta Analysis

    Approaches are available to examine and reduce bias from operating within a meta  analysis. Some ways that biased conclusions can occur in a meta-analysis are effects of a bias toward publishing positive but not negative results, giving each study an equal weight in the meta analysis despite the fact they differ in sample size or quality, inclusion of multiple tests of a hypothesis from an individual study, and not ensuring an acceptable level of agreement or reliability among raters in coding the study characteristics.

    The possibility that unknown, unpublished studies may exist, whose results fail to support the pattern illustrated by the published findings, is referred to as the file drawer problem (Rosenthal, 1979). The conclusions of the meta-analysis can be distorted if the retrieval of studies yielded only published studies in which a publication bias in favor of significant results may occur. 

    R. Rosenthal developed a technique to assess the magnitude of the file drawer problem by calculating the minimum number of unpublished studies with nonsignificant results that would be necessary to change the conclusion reached by the meta-analysis.

 Meta Analysis and Sample Size

    It can be argued that not all studies synthesized in a meta-analysis should be given equal weight. Some studies may be poorly designed and have small unrepresentative samples, whereas other studies use randomized control group designs with large sample sizes. 

    To remedy this problem, studies can be evaluated and assigned a quality score. The meta-analysis can then be calculated with studies weighted by their quality scores.

Non Independence and Its Sources

    A source of non-independence in a meta- analysis can result from using multiple hypothesis tests based on multiple variable measurements obtained from a single study (Strube & Hartman, 1983). 

    One suggested remedy when selecting findings obtained from multiple measures of the hypothesis tests located within a single study is to collapse the various findings into a single, global hypothesis test.

Specific Assumptions

    One assumption that should be met before specific studies are quantitatively combined in one meta-analysis is that each study provides sample estimates of the effect sizes that are representative of the population effect size. Homogeneity tests can be calculated to identify any outlier studies. If outliers are identified, they can be removed.

Meta Analysis in Nursing Literature

    Meta-analysis first appeared in the nursing literature in 1982, when O'Flynn published her article describing meta-analysis in the "Methodology Corner of Nursing Research. A meta-analysis of the effects of psychoeducational interventions on length of postsurgical hospital stay (Devine & Cook, 1983) was the first meta-study analysis published in nursing. 

    Since then, meta-analyses have been conducted and published in a wide variety of areas, such as patient outcomes of nurse-practitioners and nurse-midwives, job satisfaction and turnover among nurses, relationship between postpartum depression and maternal infant interaction, effects of educational interventions in diabetes care, quality of life in cardiac patients, and non-nutritive sucking in preterm infants.     

    The outcome of this quantitative approach for reviewing the literature has tremendous potential for a practice-based discipline such as nursing. One example of a meta-analysis that has consequences for nursing practice integrated the research on predictors of postpartum depression. 

    C. T. Beck's (1996) meta- analysis of 44 studies helped to clarify which variables were significantly related to postpartum depression; there had been conflicting findings reported in the literature. 

    The following eight variables were revealed to be significant predictors: prenatal depression, history of previous depression, social support, life stress, child care stress, maternity blues, marital satisfaction, and prenatal anxiety. 

    An instrument based on the findings of this meta-analysis can be designed to help detect women at risk for developing postpartum depression.

Benefits of Meta Analysis in Nursing Research

    Meta-analysis of the abundance of research being conducted can benefit nursing practice. Not only will the use of meta-analysis further knowledge development in the discipline of nursing, but it also can help nurses in the clinical setting to decide whether to apply research findings to their practice based on the size of the difference an intervention makes. 

    Meta analysis can resolve issues in nursing where there are multiple studies with conflicting findings. In addition, meta-analysis highlights gaps in nursing research for future studies.