A “placement of death” approach for studies of treatment effects on ICU length of stay

Length of stay in the intensive care unit (ICU) is a common outcome measure in randomized trials of ICU interventions. Because many patients die in the ICU, it is difficult to disentangle treatment effects on length of stay from effects on mortality; conventional analyses depend on assumptions that are often unstated and hard to interpret or check. We adapt a proposal from Rosenbaum that addresses concerns about selection bias and makes its assumptions explicit. A composite outcome is constructed that equals ICU length of stay if the patient was discharged alive and indicates death otherwise. Given any preference ordering that compares death with possible lengths of stay, we can estimate the intervention’s effects on the composite outcome distribution. Sensitivity analyses can show results for different preference orderings. We discuss methods for constructing approximate confidence intervals for treatment effects on quantiles of the outcome distribution or on proportions of patients with outcomes preferable to various cutoffs. Strengths and weaknesses of possible primary significance tests (including the Wilcoxon–Mann–Whitney rank sum test and a heteroskedasticity-robust variant due to Brunner and Munzel) are reviewed. An illustrative example reanalyzes a randomized trial of an ICU staffing intervention.

[1]  R J Carroll,et al.  On design considerations and randomization-based inference for community intervention trials. , 1996, Statistics in medicine.

[2]  Anthony C. Davison,et al.  Bootstrap Methods and Their Application , 1998 .

[3]  Dylan S. Small,et al.  Outcomes and statistical power in adult critical care randomized trials. , 2014, American journal of respiratory and critical care medicine.

[4]  P A Lachenbruch,et al.  Comparisons of two‐part models with competitors , 2001, Statistics in medicine.

[5]  I. White,et al.  Choice of test for comparing two groups, with particular application to skewed outcomes , 2003, Statistics in medicine.

[6]  Andrew D. Ho.  A Nonparametric Framework for Comparing Trends and Gaps Across Tests , 2009 .

[7]  E. Lehmann.  Elements of large-sample theory , 1998 .

[8]  D. Rubin.  Causal Inference Using Potential Outcomes , 2005 .

[9]  Terence P. Speed,et al.  Introductory Remarks on Neyman (1923) , 1990 .

[10]  Rob J Hyndman,et al.  Sample Quantiles in Statistical Packages , 1996 .

[11]  K. McGraw,et al.  A common language effect size statistic. , 1992 .

[12]  Emma Sutton.  Rejoinder , 2010 .

[13]  Joseph P. Romano,et al.  Asymptotically valid and exact permutation tests based on two-sample U-statistics , 2016 .

[14]  Dylan S. Small,et al.  A randomized trial of nighttime physician staffing in an intensive care unit. , 2013, The New England journal of medicine.

[15]  Philip B. Stark,et al.  Statistical Models and Causal Inference: A Dialogue with the Social Sciences , 2009 .

[16]  A. Hallstrom,et al.  A modified Wilcoxon test for non‐negative distributions with a clump of zeros , 2009, Statistics in medicine.

[17]  D. Rubin.  Causal Inference Through Potential Outcomes and Principal Stratification: Application to Studies with “Censoring” Due to Death , 2006, math/0612783.

[18]  Xuefeng Li,et al.  Confidence intervals for two sample binomial distribution , 2005 .

[19]  S. Ratcliffe,et al.  An Empirical Derivation of the Optimal Time Interval for Defining ICU Readmissions , 2013, Medical care.

[20]  E. Korn.  Comment: Causal Inference in the Medical Area , 2006, math/0612785.

[21]  Discussion of ‘Parametric versus nonparametrics: two alternative methodologies’ , 2009 .

[22]  J. Hahn.  Bootstrapping Quantile Regression Estimators , 1995, Econometric Theory.

[23]  Comment: The Place of Death in the Quality of Life. , 2006, math/0612786.

[24]  Huifang Zhao,et al.  Hospital mortality, length of stay, and preventable complications among critically ill patients before and after tele-ICU reengineering of critical care processes. , 2011, JAMA.

[25]  M. Joffe.  The International Journal of Biostatistics Principal Stratification and Attribution Prohibition : Good Ideas Taken Too Far , 2012 .

[26]  P. Holland.  Statistics and Causal Inference , 1985 .

[27]  Paul W. Holland,et al.  Two Measures of Change in the Gaps Between the CDFs of Test-Score Distributions , 2002 .

[28]  E. Brunner,et al.  The Nonparametric Behrens‐Fisher Problem: Asymptotic Theory and a Small‐Sample Approximation , 2000 .

[29]  D. Rubin.  Estimating causal effects of treatments in randomized and nonrandomized studies. , 1974 .

[30]  John W. Pratt,et al.  Obustness of Some Procedures for the Two-Sample Location Problem , 1964 .

[31]  Jean Dickinson Gibbons,et al.  Nonparametric Statistical Inference , 1972, International Encyclopedia of Statistical Science.

[32]  Edgar Brunner,et al.  A studentized permutation test for the non-parametric Behrens-Fisher problem , 2007, Comput. Stat. Data Anal..

[33]  Paul R Rosenbaum,et al.  Confidence Intervals for Uncommon but Dramatic Responses to Treatment , 2007, Biometrics.

[34]  Peter A Lachenbruch,et al.  Analysis of data with excess zeros , 2002, Statistical methods in medical research.

[35]  P. Rosenbaum.  Design of Observational Studies , 2009, Springer Series in Statistics.

[36]  Robert Tibshirani,et al.  An Introduction to the Bootstrap , 1994 .

[37]  Charles S. Reichardt,et al.  Justifying the use and increasing the power of a t test for a randomized experiment with a convenience sample. , 1999 .

[38]  D. Moher,et al.  CONSORT 2010 explanation and elaboration: Updated guidelines for reporting parallel group randomised trialsqqq , 2018 .

[39]  Xiao-Hua Zhou,et al.  Inferences about population means of health care costs , 2002, Statistical methods in medical research.

[40]  M. Fay,et al.  Wilcoxon-Mann-Whitney or t-test? On assumptions for hypothesis tests and multiple interpretations of decision rules. , 2010, Statistics surveys.

[41]  Alan Agresti,et al.  Simple and Effective Confidence Intervals for Proportions and Differences of Proportions Result from Adding Two Successes and Two Failures , 2000 .

[42]  Gary King,et al.  The essential role of pair matching in cluster-randomized experiments, with application to the Mexican Universal Health Insurance Evaluation , 2009, 0910.3752.

[43]  W. Lin,et al.  Agnostic notes on regression adjustments to experimental data: Reexamining Freedman's critique , 2012, 1208.2301.

[44]  J. Felson,et al.  Statistical Models and Causal Inference: A Dialogue with the Social Sciences , 2011 .

[45]  Michael Proschan,et al.  Chop‐Lump Tests for Vaccine Trials , 2009, Biometrics.

[46]  M. Gibson,et al.  Beyond ANOVA: Basics of Applied Statistics. , 1986 .

[47]  Subhabrata Chakraborti,et al.  Nonparametric Statistical Inference , 2011, International Encyclopedia of Statistical Science.

[48]  Sangeeta Mehta,et al.  Daily sedation interruption in mechanically ventilated critically ill patients cared for with a sedation protocol: a randomized controlled trial. , 2012, JAMA.

[49]  E. L. Lehmann,et al.  Parametric versus nonparametrics: two alternative methodologies , 2009 .

[50]  D. Rubin.  [On the Application of Probability Theory to Agricultural Experiments. Essay on Principles. Section 9.] Comment: Neyman (1923) and Causal Inference in Experiments and Observational Studies , 1990 .

[51]  Philip B. Stark,et al.  Statistical Models and Causal Inference: Survival Analysis: An Epidemiological Hazard? , 2009 .

[52]  Frank C Curriero,et al.  Mixture models for quantitative HIV RNA data , 2002, Statistical methods in medical research.

[53]  B. M. Brown,et al.  Kruskal–Wallis, Multiple Comparisons and Efron Dice , 2002 .

[54]  D. Rubin,et al.  Principal Stratification in Causal Inference , 2002, Biometrics.

[55]  E. Brunner,et al.  Wilcoxon–Mann–Whitney test for stratified samples and Efron's paradox dice , 2007 .