Overriding concerns: the role of electronic medical record-based best practice alerts in reducing unnecessary laboratory testing.
Multiple investigators have studied strategies to reduce unnecessary testing that include provider education, audit and feedback, incentives or penalties, and system-based interventions such as electronic order entry and clinical decision support systems (2). These interventions have met with varying success, with reported changes in use ranging from a 98% reduction to a 28% increase in testing (2). Techniques that use multimodality strategies to reduce laboratory overutilization seem to be most effective (3). With the widespread adoption of the electronic health record (EHR)  in recent years, computerized provider order entry (CPOE) systems have emerged as attractive targets for intervention, with the potential for interrupting unwarranted services before implementation. However, studies have shown mixed outcomes and cost savings with this approach (4-6).
Cardiac troponin testing has drawn particular scrutiny for potential overuse because of the high proportion of hospitalized patients in whom this test is performed, as well as the frequent ordering of serial tests. In a recent study by Makam and Nguyen (7), cardiac biomarkers were measured in 16.9% of all emergency department visits in the US over a 2-year period, despite the absence of acute coronary syndrome (ACS) symptoms in nearly one-third of those tested. Although cardiac troponins are recognized for their sensitivity and specificity in detecting myocardial injury, they are not specific for ACS (8). Thus, the advantage of a high negative predictive value in ruling out myocardial infarction (MI) is offset by a high clinical false-positive rate due to non-ACS troponin increases, a balance that is heavily influenced by the pretest probability of ACS in the tested population (7-9). The implications of unnecessary cardiac troponin testing therefore include not only the cost of the assay itself, but also the costs (financial and otherwise) of downstream cardiac testing and patient care to follow up on any false-positive results (7). Already a source of major concern and frustration for practicing cardiologists, the "troponin consult" for non-ACS troponin increases is likely to become even more common with increasing adoption of high-sensitivity troponin assays (9).
In this context, we read with great interest the study by Love et al. (10) in this issue of Clinical Chemistry. The investigators examined the effect of a 2-pronged strategy that combined a standard electronic order set for serial troponin testing with an electronic medical record best practice alert (BPA) to reduce unnecessary cardiac troponin I (cTnI) testing at their institution. Troponin ordering was consolidated, such that a single cTnI order placed through the CPOE system elicited a linked set of 4 serial cTnI tests at 3-h intervals. The BPA was triggered when CPOE was used to order cTnI concentrations on patients for whom cTnI results were available within the preceding 30 days. This BPA notified the user that the patient had recent cTnI results available and advised that troponin testing was often inappropriately overused. The BPA also included a dropdown list to query for the reason cTnI testing was believed to be indicated, with choices for ACS, demand ischemia, non-ACS myocardial necrosis, or other, the latter of which prompted users to enter a free-text comment.
The overall result of the study was null, with no significant reduction in the rate of cTnI testing following implementation of the interventional strategy. Overall, 97% of BPA triggers were acknowledged and overridden by providers, with 65% of orders accompanied by selection of a reason from the dropdown list. Of those choosing "other, " most providers gave no reason for ordering repeat testing. In this sense, the BPA did not provide a true "hard stop, " since providers could continue ordering tests without providing a valid reason for doing so, effectively bypassing the alert. The investigators acknowledge this as a weakness in the execution of their BPA. Interestingly, the majority of BPAs were triggered before the initial set of cTnI tests had been completed, suggesting that providers were either indifferent to ordering redundant tests or were not aware that the initial order created a set of 4 timed serial tests.
These data provide further evidence for the high rate of unnecessary cardiac troponin testing, as 92% of the BPAs were triggered for patients who ultimately did not have a final International Classification of Diseases, Revision 9 diagnosis related to ACS. Whereas this may speak to the utility of cTnI in ruling out ACS, it also suggests that ordering providers use a "check it just in case" mentality for patients with low probability of ACS (7). Those patients with false-positive test results are subsequently subjected to the costs and risks of additional downstream testing, as discussed above.
At the other end of the disease spectrum--patients who ruled in for ACS--redundant cTnI testing was also quite common, with a mean of 6 cTnI tests performed per patient. Although peak biomarker concentrations, or the area under the biomarker time-activity curve, were historically used as markers of infarct size, the value of serial troponin measurements once MI has been confirmed is questionable in the modern era. Because performance of echocardiography is now a quality measure after MI, and thus is done nearly universally in this setting, there is little value added from additional cTnI testing once MI has been confirmed, since it is unlikely to significantly impact further decision-making.
Another point to highlight from this study is the apparent ineffectiveness of an electronic BPA in reducing unwarranted testing. The vast majority of BPAs were overridden and tests were ordered without regard for the information provided in the BPA. Anecdotally, we have had a similar experience with BPAs at our institution: they are largely disregarded by users, a trend that is reflected frequently in the literature (6, 11-13). The reasons for such high rates of overriding the BPA are not clear. In this study, >2 alerts were generated per patient encounter, possibly contributing to what some authors have described as "alert fatigue, " in which providers are inundated with BPAs and other alerts such that each individual alert becomes less meaningful (6, 12). To pre vent alert fatigue, strategies should be considered that reduce the frequency of alert generation and tailor alerts to the most important indications. Alternatively, providers may exhibit clinical inertia whereby their predetermined decision to order a test or therapy is unlikely to be altered by an electronic alert. Future studies that use test-reduction strategies that are more difficult to override or ignore may result in more success.
As we consider other EHR-based approaches to reduce troponin overtesting in the future, insight may be gained by examining some successes of BPAs in altering provider behavior in other settings. Examples include prompting users to prescribe influenza vaccination (14) and reducing redundant laboratory testing for B-type natriuretic peptide (4). These interventions suggest several important points regarding the potential success for BPAs in achieving their goal. First, BPAs may be more effective in eliciting an action, such as ordering influenza vaccination, than preventing one, such as unnecessary testing. In this sense, the BPA may be better suited for preventing errors of omission than errors of commission, particularly if the alert incorporates a hard stop that requires the provider to take an action. Second, the type of injudicious testing that a BPA attempts to discourage will in part determine its success. In the case of serial B-type natriuretic peptide testing, there is little risk to the patient of not ordering another test and marginal benefit from the information it would provide, making it easier for providers to forego the repeat test. On the other hand, providers may consider that the hazard of potentially missing MI far outweighs any possible benefit from deferring troponin testing, even if the pretest probability of ACS is low.
These are still relatively early days in the era of EHRs and CPOE, and although these technologies offer tremendous potential to enhance value-based healthcare, this promise has yet to be fully realized. Using the EHR to alter provider behavior will seemingly depend on the type of behavior targeted, the clinical context in which this occurs, and how strict the intervention is. The study by Love et al. is a welcome early attempt to harness the EHR to address the important problem of troponin overtesting. Lessons learned from their experiences, both positive and negative, will be of considerable value for others who take on this important problem.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contribution to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:
Employment or Leadership: None declared.
Consultant or Advisory Role: J.A. de Lemos, Abbott Diagnostics, Diadexus.
Stock Ownership: None declared.
Honoraria: None declared.
Research Funding: None declared.
Expert Testimony: None declared.
Patents: None declared.
(1.) van Walraven C, Naylor CD. Do we know what inappropriate laboratory utilization is? A systematic review of laboratory clinical audits. JAMA 1998; 280:550-8.
(2.) Kobewka DM, Ronksley PE, McKay JA, Forster AJ, van Walraven C. Influence of educational, audit and feedback, system based, and incentive and penalty interventions to reduce laboratory test utilization: a systematic review. Clin Chem Lab Med 2015; 53: 157-83.
(3.) Solomon DH, Hashimoto H, Daltroy L, Liang MH. Techniques to improve physicians' use of diagnostic tests: a new conceptual framework. JAMA 1998; 280:2020-7.
(4.) Levick DL, Stern G, Meyerhoefer CD, Levick A, Pucklavage D. Reducing unnecessary testing in a CPOE system through implementation of a targeted CDS intervention. BMC Med Inform Decis 2013; 13:43.
(5.) Palen TE, Price DW, Snyder AJ, Shetterly SM. Computerized alert reduced d-dimer testing in the elderly.AmJManag Care 2010; 16:e267-75.
(6.) Ranji SR, Rennke S, Wachter RM. Computerised provider order entry combined with clinical decision support systems to improve medication safety: a narrative review. BMJ Qual Saf 2014; 23:773-80.
(7.) Makam AN, Nguyen OK. Use of cardiac biomarker testing in the emergency department. JAMA Intern Med 2015; 175:67-75.
(8.) de Lemos JA. Increasingly sensitive assays for cardiac troponins: a review.JAMA2013; 309:2262-9.
(9.) de Lemos JA, Morrow DA, deFilippi CR. Highly sensitive troponin assays and the cardiology community: a love/hate relationship? Clin Chem 2011; 57:826-9.
(10.) Love SA, McKinney ZJ, Sandoval Y, Smith SW, Kohler R, Murakami MM, Apple FS. Electronic medical record-based performance improvement project to document and reduce excessive cardiac troponin testing. Clin Chem 2015; 61:498504.
(11.) Bryant AD, Fletcher GS, Payne TH. Drug interaction alert override rates in the meaningful use era: no evidence of progress. Appl Clin Inform 2014; 5:802-13.
(12.) Isaac T, Weissman JS, Davis RB, Massagli M, Cyrulik A, Sands DZ, et al. Overrides of medication alerts in ambulatory care. Arch Intern Med 2009; 169:305-11.
(13.) van der Sijs H, Aarts J, Vulto A, Berg M. Overriding of drug safety alerts in computerized physician order entry. J Am Med Inform Assoc 2006; 13:138-47.
(14.) KlattTE, Hopp E. Effect of a best-practice alerton the rate of influenza vaccination of pregnant women. Obstet Gynecol 2012; 119:301-5.
David M. Lofthus,  Jay Y. Gadgil,  and James A. de Lemos  *
 Cardiology Division, The University of Texas Southwestern Medical School, Dallas, TX.
* Address correspondence to this author at: Cardiology Division, The University of Texas Southwestern Medical School, 5323 Harry Hines Blvd, Dallas, TX 75390. Fax 214-6452480; firstname.lastname@example.org.
Received December 18, 2014; accepted December 23, 2014.
Previously published online at DOI: 10.1373/clinchem.2014.236406
 Nonstandard abbreviations: EHR, electronic health record; CPOE, computerized provider order entry; ACS, acute coronary syndrome; MI, myocardial infarction; BPA, best practice alert; cTnI, cardiac troponin I.
|Printer friendly Cite/link Email Feedback|
|Author:||Lofthus, David M.; Gadgil, Jay Y.; de Lemos, James A.|
|Article Type:||Author abstract|
|Date:||Mar 1, 2015|
|Previous Article:||Comprehensive metabolomic profiling of type 2 diabetes.|
|Next Article:||Hot cells, raise your hands.|