Abstract
Public hospital spending consumes a large share of government expenditure in many countries. The large cost variability observed between hospitals and also between patients in the same hospital has fueled the belief that consumption of a significant portion of this funding may result in no clinical benefit to patients, thus representing waste. Accurate identification of the main hospital cost drivers and relating them quantitatively to the observed cost variability is a necessary step towards identifying and reducing waste. This study identifies prime cost drivers in a typical, mid-sized Australian hospital and classifies them as sources of cost variability that are either warranted or not warranted—and therefore contributing to waste. An essential step is dimension reduction using Principal Component Analysis to pre-process the data by separating out the low value ‘noise’ from otherwise valuable information. Crucially, the study then adjusts for possible co-linearity of different cost drivers by the use of the sparse group lasso technique. This ensures reliability of the findings and represents a novel and powerful approach to analysing hospital costs. Our statistical model included 32 potential cost predictors with a sample size of over 50,000 hospital admissions. The proportion of cost variability potentially not clinically warranted was estimated at 33.7%. Given the financial footprint involved, once the findings are extrapolated nationwide, this estimation has far-reaching significance for health funding policy.
Highlights
We propose that predictable cost drivers contain information independent of the care provided in hospital and explain warranted variability of patient-level cost of care in hospital
The predictable cost drivers accounted for 66.3% of the total patient-level cost variability, leaving 33.7% potentially associated with cost drivers not predictable a-priori
To illustrate the significance of using the sparse group lasso technique to eliminate the artifacts of co-linearity, Table 3 provides a comparison with results that would be obtained if unpenalised simple linear regression were used instead, as is common practice in hospital cost predictor analysis [24]
Summary
The study has been approved by a Human Research and Ethics Committee (HREC), which is the relevant Institutional Review Board. The Human Research and Ethics Committees in Australia grant approvals in accordance with the Australian National Statement on Ethical Conduct in Human Research (2007), which in turn fulfills the obligations under the Declaration of Helsinki. As the study is only a statistical analysis of large data, with negligible risk of patients being identified, the above-mentioned ethics committee has granted a waiver of patient consent. This waver has been subsequently endorsed and approved by the Queensland Department of Health. The current dates of the ethical approval are from 09 March 2016–09 March 2019
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
