BIDMC indicates Beth Israel Deaconess Medical Center.
A, The clinician’s response to the intervention is depicted according to the number tests ordered for patients older than 55 years and for patients simultaneously tested for viral hepatitis (B or C). B, This panel details the secular trend regarding testing for liver enzyme perturbations and shows that while the number of ceruloplasmin tests decreased, the number of α1-antitrypsin tests did not.
Tapper EB, Sengupta N, Lai M, Horowitz G. A Decision Support Tool to Reduce Overtesting for Ceruloplasmin and Improve Adherence With Clinical Guidelines. JAMA Intern Med. 2015;175(9):1561-1562. doi:10.1001/jamainternmed.2015.2062
Nondirected testing, simultaneously assessing a multitude of diseases, is a specific form of overtesting. This pattern of testing is expensive and generates false-positive test results.1 It is also common, particularly for the evaluation of liver disease.2,3 For example, elevated liver enzymes affect 7.9% of the US population, whereas Wilson disease, an inborn error of copper metabolism associated with liver injury, is routinely assessed along with viral hepatitis despite affecting only 0.003% of the US population.3- 5
Guidelines suggest testing for Wilson disease with ceruloplasmin blood levels after excluding common liver diseases and, given the rarity of late-onset Wilson disease, recommends rarely testing patients older than 55 years.4 We conducted a prospective study to evaluate the effect of a decision support tool on ceruloplasmin use by measuring use rates 7 months before and 7 months after implementing an electronic pop-up in electronic medical record system at Beth Israel Deaconess Medical Center. The study reviewed records from October 1, 2013, through November 27, 2014.
All ceruloplasmin orders that had been placed by all physicians associated with the academic medical center were reviewed. When clinicians ordered the ceruloplasmin test after implementation, the decision support tool (Figure 1) pops up on the computer screen to confirm the physicians’ intent. The primary outcome was the number of ceruloplasmin orders made at the same time as orders for viral hepatitis. The secondary outcome was the number of orders made for patients older than 55 years. Ceruloplasmin values lower than 20 mg/dL were considered positive for Wilson disease.4 Diagnoses were determined from confirmatory testing of urine or hepatic copper levels. α1-Antitrypsin levels are often ordered to evaluate liver disease.2 As such, we were able to compare trends in ceruloplasmin use with α1-antitrypsin use. Outcomes were assessed using rate ratios (RRs) with time denominators specific to the length of observation 211 days before and 210 days after implementation. This study was approved by the medical center’s institutional review board.
Ceruloplasmin was ordered 448 times (mean times per day, 2.12) before implementation and 219 times (mean times per day, 1.04) after implementation (RR, 0.49; 95% CI, 0.42 - 0.58; P < .001). There was no reduction in the rate of α1-antitrypsin orders: 449 before and 418 after. Compared with α1-antitrypsin orders, the RR of ceruloplasmin orders decreased from 1.00 (95% CI, 0.88-1.14; P >.99) before implementation to 0.53 (95% CI, 0.45-0.62; P < .001) after implementation (Figure 2).
Simultaneous tests for viral hepatitis decreased from 407 before implementation to 185 after implementation (RR, 0.46; 95% CI, 0.38-0.54; P < .001). The number of patients older than 55 years whose blood samples were tested for ceruloplasmin decreased from 158 to 61 (RR, 0.39; 95% CI, 0.29-0.52; P < .001). For comparison, in this subpopulation, α1-antitrypsin tests were ordered at rates of 0.88 patients per day before and 0.84 patients per day after implementation. The RR for α1-antitrypsin and ceruloplasmin orders made simultaneously decreased from 0.85 (95% CI, 0.69-1.06; P = .14) before implementation to 0.34 (95% CI, 0.26-0.46; P < .001) after implementation. No new diagnoses of Wilson disease were made during the study period; all results that had tested positive for ceruloplasmin (regardless of study phase) had negative confirmatory urine or hepatitic test results.
In this prospective trial, the decision support tool was associated with a reduction in ceruloplasmin overutilization. The findings from our study could inform programs to reduce the number of orders for nondirected tests in other common clinical scenarios, including antibody tests for rheumatotic or infectious diseases or routine blood tests for routine daily blood tests for inpatients.
Our results may not be generalizable to centers with other approaches to liver disease. Further study is needed to extend this intervention into systematic changes including reflex testing of rare conditions after common diseases have been excluded or restricting testing options for clinicians.
In summary, a simple decision support tool that interrupts workflow can reduce overtesting. A pop-up screen intervention can be used for other conditions for which nondirected testing is common.
Corresponding Author: Elliot B. Tapper, MD, Division of Gastroenterology, Beth Israel Deaconess Medical Center, 330 Brookline Ave, Boston, MA 02215 (firstname.lastname@example.org).
Published Online: June 1, 2015. doi:10.1001/jamainternmed.2015.2062.
Author Contributions: Dr Tapper had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Tapper.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Tapper.
Critical revision of the manuscript for important intellectual content: Sengupta, Lai, Horowitz.
Statistical analysis: Tapper.
Study supervision: Tapper.
Conflict of Interest Disclosures: None reported.
Additional Contributions: We thank J. Thomas Lamont, MD, for his guidance and mentorship, for which he received no compensation.