The use of percentage change from baseline as an outcome in a controlled trial is statistically inefficient: a simulation study
A J Vickers, A J Vickers
Abstract
Background: Many randomized trials involve measuring a continuous outcome - such as pain, body weight or blood pressure - at baseline and after treatment. In this paper, I compare four possibilities for how such trials can be analyzed: post-treatment; change between baseline and post-treatment; percentage change between baseline and post-treatment and analysis of covariance (ANCOVA) with baseline score as a covariate. The statistical power of each method was determined for a hypothetical randomized trial under a range of correlations between baseline and post-treatment scores.
Results: ANCOVA has the highest statistical power. Change from baseline has acceptable power when correlation between baseline and post-treatment scores is high;when correlation is low, analyzing only post-treatment scores has reasonable power. Percentage change from baseline has the lowest statistical power and was highly sensitive to changes in variance. Theoretical considerations suggest that percentage change from baseline will also fail to protect from bias in the case of baseline imbalance and will lead to an excess of trials with non-normally distributed outcome data.
Conclusions: Percentage change from baseline should not be used in statistical analysis. Trialists wishing to report this statistic should use another method, such as ANCOVA, and convert the results to a percentage change by using mean baseline scores.
Figures
References
- Glorioso N, Troffa C, Filigheddu F, Dettori F, Soro A, Parpaglia PP, Collatina S, Pahor M. Effect of the HMG-CoA reductase inhibitors on blood pressure in patients with essential hypertension and primary hypercholesterolemia. Hypertension. 1999;34:1281–1286.
- Messier SP, Loeser RF, Mitchell MN, Valle G, Morgan TP, Rejeski WJ, Ettinger WH. Exercise and weight loss in obese older adults with knee osteoarthritis: a preliminary study. J Am Geriatr Soc. 2000; 48:1062–1072.
- Kleinhenz J, Streitberger K, Windeler J, Gussbacher A, Mavridis G, Martin E. Randomised clinical trial comparing the effects of acupuncture and a newly designed placebo needle in rotator cuff tendinitis. Pain. 1999;83:235–241. doi: 10.1016/S0304-3959(99)00107-4.
- Matthews JNS, Altman D, Campbell MJ, Royston P. Analysis of serial measurements in medical research. BMJ. 1990;300:230–235.
- Omar RZ, Wright EM, Turner RM, Thompson SG. Analysing repeated measurements data: a practical comparison of methods. Stat Med. 1999;18:1587–1603. doi: 10.1002/(SICI)1097-0258(19990715)18:13<1587::AID-SIM141>;2-Z.
- Harris CW., Ed Problems in measuring change. Madison: University of Wisconsin Press. 1967.
- Frison L, Pocock SJ. Repeated measures in clinical trials: analysis using mean summary statistics and its implications for design. Stat Med. 1992;11:1685–1704.
- S Senn. Statistical Issues in Drug Development. Chichester: John Wiley. 1997.
- Loprinzi CL, Michalak JC, Quella SK, O'Fallon JR, Hatfield AK, Nelimark RA, Dose AM, Fischer T, Johnson C, Klatt NE, et al. Megestrol acetate for the prevention of hot flashes. N Engi J Med. 1994;331:347–352. doi: 10.1056/NEJM199408113310602.
- Anderson H, Hopwood P, Stephens RJ, Thatcher N, Cottier B, Nicholson M, Milroy R, Maughan TS, Falk SJ, Bond MG, Burt PA, Connolly CK, Mclllmurray MB, Carmichael J. Gemcitabine plus best supportive care (BSC) vs BSC in inoperable non-small cell lung cancer - a randomized trial with quality of life as the primary outcome. UK NSCLC Gemcitabine Group. Non-Small Cell Lung Cancer. Br J Cancer. 2000;83:447–453. doi: 10.1054/bjoc.2000.1307.
- Crouse JR, 3rd, Morgan T, Terry JG, Ellis J, Vitolins M, Burke GL. A randomized trial comparing the effect of casein with that of soy protein containing varying amounts of isoflavones on plasma concentrations of lipids and lipoproteins. Arch Intern Med. 1999;159:A2070–2076. doi: 10.1001/archinte.159.17.2070.
Source: PubMed