nep-hea New Economics Papers
on Health Economics
Issue of 2008‒05‒10
two papers chosen by
Yong Yin
SUNY at Buffalo, USA

  1. Do obese patients stay longer in hospital? Estimating the health care costs of obesity. By Katharina Hauck; Bruce Hollingsworth
  2. An Empirical Model of Learning under Ambiguity: The Case of Clinical Trials By Fernandez, Jose

  1. By: Katharina Hauck (Department of Econometrics and Business Statistics, Monash University); Bruce Hollingsworth (Centre for Health Economics, Monash University)
    Abstract: Objective To determine if obese patients have longer average length of stay once they are admitted to hospital, across a range of specialties. This contributes to measuring the impact of obesity on health care resource use. Data Sources/Study Setting Administrative hospital data are used for the financial year 2005/06 covering all episodes of patient care (1.3 million) in 122 public hospitals in the state of Victoria, Australia. The data are collected as part of Diagnosis Related Group (DRG) case mix funding arrangements by the state government. Study Design Statistical analysis are undertaken using quantile regression analysis to determine differences in average length of stay within different specialties for two groups of patients, those classified as obese, and those not classified as obese. Quantile regression allows a comparison of differences between the length of stay of obese and non-obese patients across the whole distribution of length of stay of inpatients, in contrast to more commonly used statistical methods which use only the mean. We condition on a range of patient and hospital characteristics such as age, sex, socioeconomic status, medical complexity of patients, teaching status, size and location of hospitals. Data Collection/Extraction Methods Data on inpatient episodes with at least one overnight stay in hospital are used. We exclude episodes with missing information on one or more of the explanatory variables and we exclude specialties with less than 50 reported obese inpatients per financial year. The final sample consists of just over 460,000 observations. Principal Findings Large and significant differences in average length of stay are found between obese and non-obese patients for nearly all specialties. In some specialties, obese patients can stay up to 4 days longer. However, obesity does not necessarily lead to longer hospital stays. In a range of specialties, obese patients have shorter length of stay on average. In general, differences between obese and non-obese patients are more pronounced at greater levels of medical complexity. There is some evidence that differences may arise because obese patients are more likely to be treated medically rather than surgically, to be transferred to another hospital, thus shifting risks and costs, or to die from higher complication rates. Conclusions Our study sheds new light on the impact of obesity on health care costs. We demonstrate that an analysis across the whole spectrum of medical complexity provides much better estimates of resource use by obese patients than standard techniques. Future research should focus on differences in the way obese patients are managed in hospital. This will show where resource use is most intense, and help policy makers and hospital managers increase efficiency and quality of care for obese patients.
    Date: 2008–05
  2. By: Fernandez, Jose
    Abstract: In this paper, I present an empirical model of learning under ambiguity in the context of clinical trials. Patients are concern with learning the treatment effect of the experimental drug, but face the ambiguity of random group assignment. A two dimensional Bayesian model of learning is proposed to capture patientsbeliefs on the treatment effect and group assignment. These beliefs are then used to predict patient attrition in clinical trials. Patient learning is demonstrated to be slower when taking into account group ambiguity. In addition, the model corrects for attrition bias in the estimated treatment effect.
    Keywords: clinical trials; learning; Bayesian; structural model; treatment effect
    JEL: D8 C31 I1
    Date: 2008–04–03

This nep-hea issue is ©2008 by Yong Yin. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.