Intended for healthcare professionals

Endgames Statistical Question

Meta-analyses I

BMJ 2011; 342 doi: https://doi.org/10.1136/bmj.d45 (Published 11 January 2011) Cite this as: BMJ 2011;342:d45
  1. Philip Sedgwick, senior lecturer in medical statistics
  1. 1Section of Medical and Healthcare Education, St George’s, University of London, Tooting, London, UK
  1. p.sedgwick{at}sgul.ac.uk

Researchers undertook a meta-analysis of the effectiveness of parenteral corticosteroids for relief of acute severe migraine headache in adults.1 Seven randomised controlled trials were identified in which single dose parenteral dexamethasone, administered alone or in combination with standard abortive therapy, was compared with placebo or any other standard treatment for acute migraine in adults. For each trial, the relative risk for recurrence of headache within 72 hours for the dexamethasone treatment arm compared with the placebo arm was obtained.

The results of the meta-analyses are presented in the figure. When combined with standard abortive therapy, single dose parental dexamethasone was more effective than placebo in reducing recurrence of headache within 72 hours (relative risk 0.74, 95% confidence interval 0.60 to 0.90).

Figure1

Forest plot of the effectiveness of dexamethasone compared with placebo in preventing the recurrence of acute severe migraine headache in adults

Which of the following statements, if any, are true?

  • a) Each trial produced an estimate of the population relative risk

  • b) The meta-analysis combined the results of the seven trials

  • c) The overall relative risk was more precise than any of those for the individual trials

  • d) The results of the meta-analysis could be generalised to a wider patient population than could any of the results for the individual trials

Answers

Answers a, b, c, and d are all true.

Seven independent controlled trials were identified for this meta-analysis, each of which investigated the effectiveness of single dose parenteral dexamethasone when administered alone or with standard abortive therapy. The recurrence of headache within 72 hours, categorised as “yes” or “no,” was a primary outcome in each trial. Relative risk was used to compare dexamethasone with placebo for the recurrence of headache. Each sample relative risk is an estimate of the population relative risk—that is, the relative risk that would be observed if all adults in the population with acute severe migraine headache were studied (a is true).

Sample size varied between the trials, whereas the samples differed with respect to their demographics and prognostic factors. Therefore, different sample estimates of the population relative risk were produced, making it difficult to definitively estimate the effects of the therapeutic regime. The purpose of the meta-analysis was to combine the results of the seven trials and achieve a single estimate of the population relative risk of recurrence of headache within 72 hours for dexamethasone compared with placebo (b is true). By combining the seven sample estimates of the population relative risk, the meta-analysis reduced the treatment evidence to a manageable quantity.

The results of the meta-analysis are presented in the forest plot. The seven trials are identified by their principal author on the left side of the forest plot. The estimated relative risk and 95% confidence interval for each trial are shown on the right. The total overall estimate of the population relative risk is presented in the line labelled “Total (95% CI).”

Because each sample incorporated a small proportion of individuals from the population, the sample relative risks are inaccurate estimates of the population parameter. As sample size increases and approaches the size of the population, accuracy of a sample estimate improves. The overall total effect was calculated by combining the samples and results from the seven trials. The overall estimate of the population relative risk therefore had increased precision compared with any of the individual sample estimates because it was based on a larger proportion of individuals from the population (c is true). In turn, the overall result from the meta-analysis could be generalised to a wider patient population (d is true).

Systematic reviews are a formal process of combining results from independent studies, published or unpublished, of the same health condition. Meta-analyses are a type of systematic review that focuses on quantitative results. Typically, meta-analyses include randomised controlled trials of comparable therapeutic regimes, although observational studies are sometimes incorporated. For each outcome common to the identified trials, a separate forest plot is constructed. Interpretation of forest plots will be discussed further in the next endgame.

Notes

Cite this as: BMJ 2011;342:d45

Footnotes

  • Competing interests: The author has completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declares: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.

References

View Abstract