File(s) under permanent embargo
Estimating the sample mean and standard deviation from commonly reported quantiles in meta-analysis
journal contribution
posted on 2020-09-01, 00:00 authored by S McGrath, X F Zhao, R Steele, B D Thombs, A Benedetti, B Levis, K E Riehm, N Saadat, A W Levis, M Azar, D B Rice, Y Sun, A Krishnan, C He, Y Wu, P M Bhandari, D Neupane, M Imran, J Boruff, P Cuijpers, S Gilbody, J P A Ioannidis, L A Kloda, D McMillan, S B Patten, I Shrier, R C Ziegelstein, D H Akena, B Arroll, L Ayalon, H R Baradaran, M Baron, A Beraldi, C H Bombardier, P Butterworth, G Carter, M H Chagas, J C N Chan, R Cholera, N Chowdhary, K Clover, Y Conwell, J M de Man-van Ginkel, J Delgadillo, J R Fann, F H Fischer, B Fischler, D Fung, B Gelaye, F Goodyear-Smith, C G Greeno, B J Hall, P A Harrison, M Harter, U Hegerl, L Hides, S E Hobfoll, M Hudson, T Hyphantis, M Inagaki, K Ismail, N Jetté, M E Khamseh, K M Kiely, Y Kwan, F Lamers, S I Liu, M Lotrakul, S R Loureiro, B Löwe, L Marsh, A McGuire, S M Sidik, T N Munhoz, K Muramatsu, F L Osório, V Patel, B W Pence, P Persoons, A Picardi, K Reuter, A G Rooney, I S Santos, J Shaaban, A Sidebottom, A Simning, L Stafford, S C Sung, P L Lynnette Tan, Alyna TurnerAlyna Turner, C M van der Feltz-Cornelis, H C van Weert, P A Vöhringer, J White, M A Whooley, K Winkley, M Yamada, Y ZhangResearchers increasingly use meta-analysis to synthesize the results of several studies in order to estimate a common effect. When the outcome variable is continuous, standard meta-analytic approaches assume that the primary studies report the sample mean and standard deviation of the outcome. However, when the outcome is skewed, authors sometimes summarize the data by reporting the sample median and one or both of (i) the minimum and maximum values and (ii) the first and third quartiles, but do not report the mean or standard deviation. To include these studies in meta-analysis, several methods have been developed to estimate the sample mean and standard deviation from the reported summary data. A major limitation of these widely used methods is that they assume that the outcome distribution is normal, which is unlikely to be tenable for studies reporting medians. We propose two novel approaches to estimate the sample mean and standard deviation when data are suspected to be non-normal. Our simulation results and empirical assessments show that the proposed methods often perform better than the existing methods when applied to non-normal data.
History
Journal
Statistical methods in medical researchVolume
29Issue
9Pagination
2520 - 2537Publisher
SageLocation
London, Eng.Publisher DOI
ISSN
0962-2802eISSN
1477-0334Language
engPublication classification
C1 Refereed article in a scholarly journalUsage metrics
Categories
Keywords
Science & TechnologyLife Sciences & BiomedicinePhysical SciencesHealth Care Sciences & ServicesMathematical & Computational BiologyMedical InformaticsStatistics & ProbabilityMathematicsMeta-analysismedianfirst quartilethird quartileminimum valuemaximum valuePATIENT HEALTH QUESTIONNAIREDEPRESSIONTHERAPYDEPRESsion Screening Data (DEPRESSD) CollaborationStatistics
Licence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC