What to add to nothing? Use and avoidance of continuity corrections in meta-analysis of sparse data.

By M. Sweeting, A. Sutton, Paul C. Lambert. Statistics in Medicine, 2004

1,624 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/b6f348c6b3bb229cc04896dfbad5acb8b6d4fad3

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

Citations in the corpus

Random effects meta‐analysis of event outcome in the framework of the generalized linear mixed model with applications in sparse data (2010, methodology): “[4] investigated through extensive simulation studies the performance of 14 different methods, among them several different continuity corrections and other methods such as the ones mentioned above.”
A comparison of heterogeneity variance estimators in simulated random‐effects meta‐analyses (2018, methodology): “5 was applied whenever simulated studies with a binary outcome contained zero events, but other methods with a better performance are available.(44) This choice may have affected the results in scenarios where the event is rare (ie, 0.”
Anti-TNF antibody therapy in rheumatoid arthritis and the risk of serious infections and malignancies: systematic review and meta-analysis of rare harmful effects in randomized controlled trials. (2006, methodology): “Our protocol called for a fixedeffects meta-analysis because of its superior performance when pooling trials with few or no events compared with the random-effects model, with results expressed as odds ratios (ORs) and associated 95% confidence intervals (CIs).(19) Based on the OR for each individua”
Conducting quantitative synthesis when comparing medical interventions: AHRQ and the Effective Health Care Program. (2011, methodology): “These studies are usually excluded from the analysis as they do not provide information on the direction and magnitude of the effect size [25,38].”
Consistency and inconsistency in network meta-analysis: model estimation using multivariate meta-regression (2012, methodology): “In the case of binary outcomes, zero counts in observed treatment arms require additional treatment; the standard procedure is to add 1/2 to all cells in observed treatment arms for any trial with a zero count (Sweeting et al., 2004).”
Comparative Effectiveness and Safety of Medications for Type 2 Diabetes: An Update Including New Drugs and 2-Drug Combinations (2011, methodology): “For the dichotomous outcomes of congestive heart failure and ischemic heart disease, we calculated pooled fixed-effects odds ratios by using the treatment-arm continuity correction (reciprocal of the sample size in the opposite treatment group in cells with 0 events) (17).”
Methods to estimate the between‐study variance and its uncertainty in meta‐analysis† (2015, methodology): “…and instead either use the FE model and the Mantel–Haenszel method for unbalanced group sizes or Peto’s method for balanced group sizes (Bradburn et al., 2007; Sweeting et al., 2004), or to switch to models/methods based on exact distributional assumptions (Kuss, 2014; Stijnen et al., 2010).”
Systematic Review and Meta-analysis Supplements for Primary and Secondary Prevention: Mortality in Randomized Trials of Antioxidant (2009, neutral): “We did not include trials with 0 events in both intervention groups.(27,28) Because the number of such trials was large, we performed exploratory analysis adding an imagined trial with 1 death and 20 000 participants in each group.”
Much ado about nothing: a comparison of the performance of meta‐analytical methods with rare events (2007, methodology): “[8] simulated meta-analyses with both rare events (baseline event rate=1 per cent) and substantial heterogeneity but found that the estimated among-study variance was zero in excess of 95 per cent of the simulations.”