Techniques for estimating health care costs with censored data: An overview for the health services researcher

Division of Cardiology, Schulich Heart Centre and Department of Medicine, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Ontario, Canada.
ClinicoEconomics and Outcomes Research 06/2012; 4(1):145-55. DOI: 10.2147/CEOR.S31552
Source: PubMed


The aim of this study was to review statistical techniques for estimating the mean population cost using health care cost data that, because of the inability to achieve complete follow-up until death, are right censored. The target audience is health service researchers without an advanced statistical background.
Data were sourced from longitudinal heart failure costs from Ontario, Canada, and administrative databases were used for estimating costs. The dataset consisted of 43,888 patients, with follow-up periods ranging from 1 to 1538 days (mean 576 days). The study was designed so that mean health care costs over 1080 days of follow-up were calculated using naïve estimators such as full-sample and uncensored case estimators. Reweighted estimators - specifically, the inverse probability weighted estimator - were calculated, as was phase-based costing. Costs were adjusted to 2008 Canadian dollars using the Bank of Canada consumer price index (
Over the restricted follow-up of 1080 days, 32% of patients were censored. The full-sample estimator was found to underestimate mean cost ($30,420) compared with the reweighted estimators ($36,490). The phase-based costing estimate of $37,237 was similar to that of the simple reweighted estimator.
The authors recommend against the use of full-sample or uncensored case estimators when censored data are present. In the presence of heavy censoring, phase-based costing is an attractive alternative approach.

Download full-text


Available from: Harindra C Wijeysundera, Oct 03, 2015
1 Follower
26 Reads
  • [Show abstract] [Hide abstract]
    ABSTRACT: Although multidisciplinary heart failure (HF) clinics are efficacious, it is not known how patient factors or HF clinic structural indicators and process measures have an impact on the cumulative health care costs. In this retrospective cohort study using administrative databases in Ontario, Canada, we identified 1216 HF patients discharged alive after an acute care hospitalization in 2006 and treated at a HF clinic. The primary outcome was the cumulative 1-year health care costs. A hierarchical generalized linear model with a logarithmic link and gamma distribution was developed to determine patient-level and clinic-level predictors of cost. The mean 1-year cost was $27,809 (range, $69 to $343,743). There was a 7-fold variation in the mean costs by clinic, from $14,670 to $96,524. Delays in being seen at a HF clinic were a significant patient-level predictor of costs (rate ratio 1.0015 per day; P<0.001). Being treated at a clinic with >3 physicians was associated with lower costs (rate ratio 0.78; P=0.035). Unmeasured patient-level differences accounted for 97.4% of the between-patient variations in cost. The between-clinic variation in costs decreased by 16.3% when patient-level factors were accounted for; it decreased by a further 49.8% when clinic-level factors were added. From a policy perspective, the wide spectrum of HF clinic structure translates to inefficient care. Greater guidance as to the type of patient seen at a HF clinic, the timeliness of the initial visit, and the most appropriate structure of the HF clinics may potentially result in more cost-effective care.
    Medical care 03/2014; 52(3):272-9. DOI:10.1097/MLR.0000000000000071 · 3.23 Impact Factor