A review of statistical estimators for risk-adjusted length of stay: analysis of the Australian and new Zealand intensive care adult patient data-base, 2008–2009

Abstract

BackgroundFor the analysis of length-of-stay (LOS) data, which is characteristically right-skewed, a number of statistical estimators have been proposed as alternatives to the traditional ordinary least squares (OLS) regression with log dependent variable.MethodsUsing a cohort of patients identified in the Australian and New Zealand Intensive Care Society Adult Patient Database, 2008–2009, 12 different methods were used for estimation of intensive care (ICU) length of stay. These encompassed risk-adjusted regression analysis of firstly: log LOS using OLS, linear mixed model [LMM], treatment effects, skew-normal and skew-t models; and secondly: unmodified (raw) LOS via OLS, generalised linear models [GLMs] with log-link and 4 different distributions [Poisson, gamma, negative binomial and inverse-Gaussian], extended estimating equations [EEE] and a finite mixture model including a gamma distribution. A fixed covariate list and ICU-site clustering with robust variance were utilised for model fitting with split-sample determination (80%) and validation (20%) data sets, and model simulation was undertaken to establish over-fitting (Copas test). Indices of model specification using Bayesian information criterion [BIC: lower values preferred] and residual analysis as well as predictive performance (R2, concordance correlation coefficient (CCC), mean absolute error [MAE]) were established for each estimator.ResultsThe data-set consisted of 111663 patients from 131 ICUs; with mean(SD) age 60.6(18.8) years, 43.0% were female, 40.7% were mechanically ventilated and ICU mortality was 7.8%. ICU length-of-stay was 3.4(5.1) (median 1.8, range (0.17-60)) days and demonstrated marked kurtosis and right skew (29.4 and 4.4 respectively). BIC showed considerable spread, from a maximum of 509801 (OLS-raw scale) to a minimum of 210286 (LMM). R2 ranged from 0.22 (LMM) to 0.17 and the CCC from 0.334 (LMM) to 0.149, with MAE 2.2-2.4. Superior residual behaviour was established for the log-scale estimators. There was a general tendency for over-prediction (negative residuals) and for over-fitting, the exception being the GLM negative binomial estimator. The mean-variance function was best approximated by a quadratic function, consistent with log-scale estimation; the link function was estimated (EEE) as 0.152(0.019, 0.285), consistent with a fractional-root function.ConclusionsFor ICU length of stay, log-scale estimation, in particular the LMM, appeared to be the most consistently performing estimator(s). Neither the GLM variants nor the skew-regression estimators dominated.

Individual and network characteristic associated with hospital-acquired Middle East Respiratory Syndrome coronavirus

“…This might be closely linked to the second outcome, because the group of hospital-acquired infections included patients who had already been hospitalized for other health issues. The length of hospital stay has been investigated from various perspectives both medical and economical [39][40][41][42]. Our result is in accordance with Glance et al [43], who showed that the length of hospital stay, associated costs and mortality rate of hospital-acquired infections were significantly higher for trauma patients.…”

Associations of Polyethylenimine-Coated AN69ST Membrane in Continuous Renal Replacement Therapy with the Intensive Care Outcomes: Observations from a Claims Database from Japan

“…For the secondary outcome of ICU-LOS, we used 2 approaches. First, we used generalized linear modeling assuming a gamma distribution of ICU-LOS utilizing a log link, adjusting for ICU mortality as a measure of illness severity [28,29,30]. Because ICU death is a competing risk to ICU discharge, we also used competing risks regression based on the Fine and Gray's proportional subhazards model, modeling ICU-LOS accounting for the competing outcome of ICU death [31].…”

Tracheostomy and infection prolong length of stay in hospital after surgery for head and neck cancer: a population based study

“…LOS distribution was markedly skewed to the right and was modeled as count data by using negative binomial regression. 13 A minority of patients underwent >1 surgery; therefore, robust standard errors were calculated for incident rate ratios (IRR) to account for clustering on these patients. Univariate IRRs were computed for each variable.…”