How much can we learn about missing data?: an exploration of a clinical trial in psychiatry

When a randomized controlled trial has missing outcome data, any analysis is based on untestable assumptions, e.g. that the data are missing at random, or less commonly on other assumptions about the missing data mechanism. Given such assumptions, there is an extensive literature on suitable methods of analysis. However, little is known about what assumptions are appropriate. We use two sources of ancillary data to explore the missing data mechanism in a trial of adherence therapy in patients with schizophrenia: carer-reported (proxy) outcomes and the number of contact attempts. This requires additional assumptions to be made whose plausibility we discuss. Proxy outcomes are found to be unhelpful in this trial because they are insufficiently associated with patient outcome and because the ancillary assumptions are implausible. The number of attempts required to achieve a follow-up interview is helpful and suggests that these data are unlikely to depart far from being missing at random. We also perform sensitivity analyses to departures from missingness at random, based on the investigators’ prior beliefs elicited at the start of the trial. Wider use of techniques such as these will help to inform the choice of suitable assumptions for the analysis of randomized controlled trials.

[1]  Roderick J. A. Little,et al.  Statistical Analysis with Missing Data: Little/Statistical Analysis with Missing Data , 2002 .

[2]  Daniel O Scharfstein,et al.  Incorporating prior beliefs about selection bias into the analysis of randomized trials with missing outcomes. , 2003, Biostatistics.

[3]  Geert Molenberghs,et al.  Analyzing incomplete longitudinal clinical trial data. , 2004, Biostatistics.

[4]  Ian R White,et al.  Adjusting for partially missing baseline measurements in randomized trials , 2005, Statistics in medicine.

[5]  S G Thompson,et al.  Methods for summarizing the risk associations of quantitative variables in epidemiologic studies in a consistent form. , 1996, American journal of epidemiology.

[6]  Matthew Hotopf,et al.  Using number of failed contact attempts to adjust for non‐ignorable non‐response , 2006 .

[7]  N. C. Schaeffer,et al.  Institute for Research on Poverty Discussion Paper no. 1024-93 Using Survey Participants to Estimate the Impact of Nonparticipation , 1993 .

[8]  J. Ware.  SF-36 health survey: Manual and interpretation guide , 2003 .

[9]  London School of Hygiene and Tropical Medicine , 1938, Nature.

[10]  Andrew Gelman,et al.  General methods for monitoring convergence of iterative simulations , 1998 .

[11]  Roderick J. A. Little,et al.  Modeling the Drop-Out Mechanism in Repeated-Measures Studies , 1995 .

[12]  Christian Genest,et al.  Combining Probability Distributions: A Critique and an Annotated Bibliography , 1986 .

[13]  A. David,et al.  Adherence therapy for people with schizophrenia , 2006, British Journal of Psychiatry.

[14]  Rong Huang,et al.  The role of proxy information in missing data analysis , 2005, Statistical methods in medical research.

[15]  Juha M. Alho,et al.  Adjusting for nonresponse bias using logistic regression , 1990 .

[16]  C H Brown,et al.  Protecting against nonrandomly missing data in longitudinal studies. , 1990, Biometrics.

[17]  M. Kenward,et al.  Every missingness not at random model has a missingness at random counterpart with equal fit , 2008 .

[18]  Kypros Kypri,et al.  Assessment of nonresponse bias in an internet survey of alcohol use. , 2004, Alcoholism, clinical and experimental research.

[19]  Nicole A. Lazar,et al.  Statistical Analysis With Missing Data , 2003, Technometrics.

[20]  Andrew Thomas,et al.  WinBUGS - A Bayesian modelling framework: Concepts, structure, and extensibility , 2000, Stat. Comput..

[21]  Ian R White,et al.  Are missing outcome data adequately handled? A review of published randomized controlled trials in major medical journals , 2004, Clinical trials.

[22]  Geert Molenberghs,et al.  Sensitivity analysis for incomplete contingency tables: the Slovenian plebiscite case , 2001 .

[23]  M. Kenward,et al.  Informative dropout in longitudinal data analysis (with discussion) , 1994 .

[24]  S. Sutton,et al.  Effectiveness of individually tailored smoking cessation advice letters as an adjunct to telephone counselling and generic self-help materials: randomized controlled trial. , 2007, Addiction.

[25]  G. Molenberghs,et al.  Protective Estimation of Longitudinal Categorical Data With Nonrandom Dropout , 1997 .

[26]  M. Kenward,et al.  Informative Drop‐Out in Longitudinal Data Analysis , 1994 .

[27]  Jeremy E. Oakley,et al.  Uncertain Judgements: Eliciting Experts' Probabilities , 2006 .

[28]  Joseph G. Ibrahim,et al.  Using auxiliary data for parameter estimation with non‐ignorably missing outcomes , 2001 .

[29]  James M. Robins,et al.  Semiparametric Regression for Repeated Outcomes With Nonignorable Nonresponse , 1998 .

[30]  Jonathan J. Forster,et al.  Model‐based inference for categorical survey data subject to non‐ignorable non‐response , 1998 .

[31]  P. Campanelli,et al.  Exploring survey non-response: the effect of attrition on a follow-up of the 1984-85 health and life style survey , 1996 .

[32]  D J Spiegelhalter,et al.  The CHART trials: Bayesian design and monitoring in practice. CHART Steering Committee. , 1994, Statistics in medicine.

[33]  I. White,et al.  Eliciting and using expert opinions about dropout bias in randomized controlled trials , 2007, Clinical trials.

[34]  M. Kenward.  Selection models for repeated measurements with non-random dropout: an illustration of sensitivity. , 1998, Statistics in medicine.