Data points from week 1, 2013, were omitted from the Figure because they were amalgamated with week 53, 2012, creating very high values (QMC = 2829; NCH = 1601). These raw data were included in all statistical analyses. CRP indicates C-reactive protein; NHC, Nottingham City Hospital; and QMC, Queen’s Medical Centre.
Fogarty AW, Sturrock N, Premji K, Prinsloo P. Hospital Clinicians’ Responsiveness to Assay Cost Feedback: A Prospective Blinded Controlled Intervention Study. JAMA Intern Med. 2013;173(17):1654-1655. doi:10.1001/jamainternmed.2013.8211
In developed countries, the costs of providing medical care are consistently increasing1 and projected to continue to do so.2 Some components of health care may be regarded as discretional, and this is reflected in the large variation in hospital costs that was not associated with clinical outcomes3 and the conclusions of a systematic review that non–evidence-based variables influence physicians’ test ordering.4 In the United Kingdom, health care is provided by the National Health Service (NHS), with hospital physicians requesting diagnostic tests based on their perception of good clinical practice. The service is free at source, and physicians are not routinely informed of the individual cost of their activities. We tested the hypothesis that hospital clinicians are sensitive to feedback of assay cost in a blinded prospective controlled study.
The study was based in Nottingham University Hospitals NHS Trust. The study population consisted of all inpatients and outpatients who attended either Nottingham City Hospital (NCH) (intervention site) or Queen’s Medical Centre (QMC) (control site). The intervention consisted of a message stating “Cost per test £1.00; total NUH [Nottingham University Hospitals] spent on C-reactive protein [CRP] assays in 2010 was £200 914” that was added to the CRP reports. The intervention was introduced at the NCH site but not at QMC. The total number of blood CRP assays requested were obtained for the 52 weeks before and after the intervention. We also collected data over the same period on the number of complete blood cell count requests to generate an independent measure of clinical activity in the 2 centers. The primary comparison used time series analysis of the difference in weekly frequency of the total number of blood CRP assays requested in the 52 weeks before compared with the 52 weeks after the intervention was introduced at NCH vs QMC. More details are available in the eMethods in the Supplement.
At NCH there was a significant decrease in the number of assays after the intervention was implemented, with a mean decrease in demand of 382 CRP assays per week (95% CI, 319 to 444) (P < .001), representing a proportionate decrease from baseline of 32% (95% CI, 27% to 37%). No such change was observed in the weekly number of CRP assays at QMC (Table).
Comparison of the difference in the weekly number of CRP assays requested at NCH vs QMC demonstrated a significant increase in the difference between the 2 sites of 338 CRP assays per week (95% CI, 279 to 396) (P < .001) after the intervention was implemented (Figure), representing a 99% (95% CI, 82% to 116%) change compared with the baseline weekly difference. These differences were a consequence of a decrease in demand in inpatients at NCH (increase in weekly difference between sites of 386 assays [95% CI, 339 to 433]).
To our knowledge, this is the first study that has investigated the implementation of feedback of the cost of an assay to the requesting clinicians on subsequent demand at the institutional level. Our data clearly demonstrate that after the introduction of the intervention, the demand for CRP blood assays decreased.
The strengths of our study design include that it was a prospective blinded controlled study with the specific hypothesis of testing whether feedback of the cost of CRP assays will modify subsequent demand. The data are captured by information technology systems and hence are correct and unbiased. Stratifying for requests for blood CRP assays identified that the impact of price feedback is effective on demand for inpatients but not for outpatients. This is important, as the more junior physicians who are in training posts request most of the inpatient blood tests in the United Kingdom, and our data suggest that this group may be more susceptible to feedback of cost.
The main weakness in our data is that the configuration of clinical services across sites within Nottingham University Hospitals NHS Trust are consistently being modified.
Because our results remain significant in comparisons before and after the intervention, both internally within the NCH and also in comparison with the QMC (control site), and because the size of change in demand for CRP assay is large (with a 32% reduction, which is closely temporarily associated with the intervention [Figure]), the associations are likely to be a causal consequence of providing the cost of the CRP assay back to the requesting clinician.
This study has significant potential implications for health care delivery. As we remain challenged with the task of providing health care in the face of rising costs and demand,5 the possibility of introducing feedback of the cost of discretionary diagnostic tests to the requesting clinician is an area that deserves further evaluation, ideally under controlled circumstances, considering any inadvertent adverse effects that may occur as well as the potential economic benefits.
Corresponding Author: Andrew W. Fogarty, DM, Division of Epidemiology and Public Health, University of Nottingham, Clinical Sciences Building, City Hospital, Nottingham NG5 1PB, England (firstname.lastname@example.org).
Published Online: July 29, 2013. doi:10.1001/jamainternmed.2013.8211.
Author Contributions: Dr Fogarty had full access to all of the data and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Fogarty, Sturrock.
Acquisition of data: Premji.
Analysis and interpretation of data: Fogarty, Prinsloo.
Drafting of the manuscript: Fogarty.
Critical revision of the manuscript for important intellectual content: Sturrock, Premji, Prinsloo.
Statistical analysis: Fogarty.
Administrative, technical, and material support: Sturrock, Premji, Prinsloo.
Conflicts of Interest Disclosures: None reported.
Additional Information: This work was initiated in collaboration with Tony Hitch, PhD, who sadly died in April 2012.