Manuel Gomes1, Edmond S-W Ng1, Richard Grieve1, Richard Nixon2, James Carpenter3, Simon G Thompson4. 1. Department of Health Services Research and Policy, London School of Hygiene and Tropical Medicine, London, UK (MG, ESWN, RG) 2. Modeling and Simulation Group, Novartis Pharma AG, Basel, Switzerland (RN) 3. Department of Medical Statistics, London School of Hygiene and Tropical Medicine, London, UK (JC) 4. MRC Biostatistics Unit, Cambridge, UK (SGT)
Abstract
AIM: Cost-effectiveness analyses (CEAs) may use data from cluster randomized trials (CRTs), where the unit of randomization is the cluster, not the individual. However, most studies use analytical methods that ignore clustering. This article compares alternative statistical methods for accommodating clustering in CEAs of CRTs. METHODS: Our simulation study compared the performance of statistical methods for CEAs of CRTs with 2 treatment arms. The study considered a method that ignored clustering--seemingly unrelated regression (SUR) without a robust standard error (SE)--and 4 methods that recognized clustering--SUR and generalized estimating equations (GEEs), both with robust SE, a "2-stage" nonparametric bootstrap (TSB) with shrinkage correction, and a multilevel model (MLM). The base case assumed CRTs with moderate numbers of balanced clusters (20 per arm) and normally distributed costs. Other scenarios included CRTs with few clusters, imbalanced cluster sizes, and skewed costs. Performance was reported as bias, root mean squared error (rMSE), and confidence interval (CI) coverage for estimating incremental net benefits (INBs). We also compared the methods in a case study. RESULTS: Each method reported low levels of bias. Without the robust SE, SUR gave poor CI coverage (base case: 0.89 v. nominal level: 0.95). The MLM and TSB performed well in each scenario (CI coverage, 0.92-0.95). With few clusters, the GEE and SUR (with robust SE) had coverage below 0.90. In the case study, the mean INBs were similar across all methods, but ignoring clustering underestimated statistical uncertainty and the value of further research. CONCLUSIONS: MLMs and the TSB are appropriate analytical methods for CEAs of CRTs with the characteristics described. SUR and GEE are not recommended for studies with few clusters.
AIM: Cost-effectiveness analyses (CEAs) may use data from cluster randomized trials (CRTs), where the unit of randomization is the cluster, not the individual. However, most studies use analytical methods that ignore clustering. This article compares alternative statistical methods for accommodating clustering in CEAs of CRTs. METHODS: Our simulation study compared the performance of statistical methods for CEAs of CRTs with 2 treatment arms. The study considered a method that ignored clustering--seemingly unrelated regression (SUR) without a robust standard error (SE)--and 4 methods that recognized clustering--SUR and generalized estimating equations (GEEs), both with robust SE, a "2-stage" nonparametric bootstrap (TSB) with shrinkage correction, and a multilevel model (MLM). The base case assumed CRTs with moderate numbers of balanced clusters (20 per arm) and normally distributed costs. Other scenarios included CRTs with few clusters, imbalanced cluster sizes, and skewed costs. Performance was reported as bias, root mean squared error (rMSE), and confidence interval (CI) coverage for estimating incremental net benefits (INBs). We also compared the methods in a case study. RESULTS: Each method reported low levels of bias. Without the robust SE, SUR gave poor CI coverage (base case: 0.89 v. nominal level: 0.95). The MLM and TSB performed well in each scenario (CI coverage, 0.92-0.95). With few clusters, the GEE and SUR (with robust SE) had coverage below 0.90. In the case study, the mean INBs were similar across all methods, but ignoring clustering underestimated statistical uncertainty and the value of further research. CONCLUSIONS: MLMs and the TSB are appropriate analytical methods for CEAs of CRTs with the characteristics described. SUR and GEE are not recommended for studies with few clusters.
Authors: S L Bellamy; R Gibberd; L Hancock; P Howley; B Kennedy; N Klar; S Lipsitz; L Ryan Journal: Stat Methods Med Res Date: 2000-04 Impact factor: 3.021
Authors: Susan Ahern; Lee-Ann Burke; Brendan McElroy; Paul Corcoran; Elaine M McMahon; Helen Keeley; Vladimir Carli; Camilla Wasserman; Christina W Hoven; Marco Sarchiapone; Alan Apter; Judit Balazs; Raphaela Banzer; Julio Bobes; Romuald Brunner; Doina Cosman; Christian Haring; Michael Kaess; Jean-Pierre Kahn; Agnes Kereszteny; Vita Postuvan; Pilar A Sáiz; Peeter Varnik; Danuta Wasserman Journal: Eur Child Adolesc Psychiatry Date: 2018-02-14 Impact factor: 4.785
Authors: Elizabeth C Rhodes; Eeshwar K Chandrasekar; Shivani A Patel; K M Venkat Narayan; Thomas V Joshua; Lovoria B Williams; Lucy Marion; Mohammed K Ali Journal: Diabetes Res Clin Pract Date: 2018-09-28 Impact factor: 5.602
Authors: Christopher C Butler; Nick A Francis; Emma Thomas-Jones; Mirella Longo; Mandy Wootton; Carl Llor; Paul Little; Michael Moore; Janine Bates; Timothy Pickles; Nigel Kirby; David Gillespie; Kate Rumsby; Curt Brugman; Micaela Gal; Kerenza Hood; Theo Verheij Journal: Br J Gen Pract Date: 2018-02-26 Impact factor: 5.386
Authors: Howard Ring; James Howlett; Mark Pennington; Christopher Smith; Marcus Redley; Caroline Murphy; Roxanne Hook; Adam Platt; Nakita Gilbert; Elizabeth Jones; Joanna Kelly; Angela Pullen; Adrian Mander; Cam Donaldson; Simon Rowe; James Wason; Fiona Irvine Journal: Health Technol Assess Date: 2018-02 Impact factor: 4.014
Authors: Gary L O'Brien; Denis O'Mahony; Paddy Gillespie; Mark Mulcahy; Valerie Walshe; Marie N O'Connor; David O'Sullivan; James Gallagher; Stephen Byrne Journal: Drugs Aging Date: 2018-08 Impact factor: 3.923