Diagnostic stewardship aims to influence the process of ordering, performing, and reporting diagnostic tests to improve patient care and reduce waste. Reference Morgan, Malani and Diekema1 The diagnostic stewardship approaches with the highest reported success rates include computerized clinical decision support (CCDS) interventions and real-time evaluations. Reference Boly, Reske and Kwon2
Unnecessary repeat testing is a good target for diagnostic stewardship. Duplicate testing is valid in some instances but is often unnecessary. Similar to the experience of others, we have found that duplicate testing is often performed because the second ordering clinician is unaware of a previous order or is too busy to check the pending order list. Reference Procop, Yerian, Wyllie, Harrison and Kottke-Marchant3 Navigation away from the ordering screen to check for pending orders requires additional steps that may be time consuming. Duplicate laboratory tests have several potential adverse effects. Pain or discomfort may occur with repeated specimen collection. Iatrogenic anemia may occur from repeated phlebotomies, which in turn may affect wound healing and lead to infection. Increased healthcare costs can accrue through specimen collection, transport, testing, storage, and resulting waste disposal. The clinical response to the result may include unnecessary treatment or further investigation. Reference Procop, Yerian, Wyllie, Harrison and Kottke-Marchant3 Overdiagnosis of healthcare-associated infections may lead to inappropriate antimicrobial use and risks of antimicrobial resistance and adverse drug effects as well as their associated unnecessary costs. Reference Madden, Weinstein and Sifri4
Repeat testing is associated with low diagnostic yield and with increased false-positive results. For example, Clostridioides difficile studies revealed that testing repeated within 48 hours has low diagnostic value. Reference Boly, Reske and Kwon2 Additionally, in a study prompted by a pseudo-outbreak at an academic center, repeated testing led to an increase in false-positive results due to the decrease in prevalence: the positive predictive values of the second and third C. difficile tests were 30% and 4%, respectively. Reference Litvin, Reske and Mayfield5 Repeated C. difficile polymerase chain reaction (PCR) testing has been associated with increased cost to the patient and healthcare system. Reference Boly, Reske and Kwon2,Reference Mohan, McDermott and Parchuri6–Reference Jun Huang, Sivadas, Spitzer, Spitzer, Aldrete Sdel and Kraft9 Furthermore, test stewardship driven by electronic medical record (EMR)–based CCDS decreased the frequency of C. difficile testing, improved test fidelity, and decreased the number of patients potentially overtreated with antibiotics for C. difficile colonization. Reference Fleming, Hess and Albert10
Electronic test ordering provides opportunities for pathology stewardship and the implementation of duplicate order alerts to decrease repeat testing. Few studies have measured the impact of reduction of duplicate orders on tests other than C. difficile. Reference Hueth, Prinzi and Timbrook11–Reference Luo, Spradley and Banaei13 In one such study, an intervention to block duplicate orders demonstrated avoidance of 11,790 tests in 2 years, with cost savings of US$183,586. The CCDS intervention consisted of an immediate electronic notification alert that a same-day duplicate test was being ordered and informed the provider that repeat testing is not warranted more than once per day. Reference Procop, Yerian, Wyllie, Harrison and Kottke-Marchant3 In another study, implementation of a new order set in the electronic health record that required practitioners to choose an indication for the type of urine study, led to a 40% reduction in urine cultures performed, a decrease in antibiotic days of therapy, and substantial financial savings. Reference Watson, Trautner and Russo14
The cost of infectious waste disposal is much higher than that of noninfectious waste. For example, in the United States, it costs US$0.79 per kilogram to dispose of infectious waste, which represents a 560% cost premium over the typical noninfectious waste disposal cost of US$0.12 per kilogram. It has been estimated that as much as 23% of overall infectious medical waste comes from laboratories. Reference Windfeld and Brooks15 In a recent study of the entire New South Wales health system, pathology and diagnostic imaging together accounted for ∼9% of its healthcare footprint comprising total emissions, water use, and waste footprints. Reference Malik, Padget and Carter16 A recent Australian study was conducted to estimate the carbon footprint of 5 common pathology tests. These researchers concluded that reducing unnecessary testing would be the most effective approach to reducing the carbon footprint of pathology testing. Reference McAlister, Barratt, Bell and McGain17
Some electronic test-ordering systems allow design of customized duplicate order alerts with CCDS tailored to the test type. In our study, we implemented test-specific duplicate order alerts based on various repeat-test time frames (ranging from 3 days to 2 years), in which the alert message suggested time frames and clinical indications for appropriate retesting. We measured the impact of these customized alerts on ordering over a 6-month period. We studied the environmental impact of repeat testing given that climate change influences health outcomes and that requesters may be motivated to order tests prudently if they are more aware of the environmental impacts.
Methods
Monash pathology laboratory services Monash Health, the largest public healthcare network in Melbourne, Australia, with 2,150 inpatient beds and 3 emergency departments serving more than one-quarter of Melbourne’s population. The laboratory also serves outpatient clinics and accepts pathology samples referred by local general practitioners.
Duplicate order alerts were implemented on October 15, 2020, for inpatients and Emergency Department patients at Monash Health. Data were extracted from the electronic medical record retrospectively from February 17, 2022, to August 17, 2022. We chose to study the 2 most commonly ordered microbiology tests (urine culture and fecal PCR) and serology tests [ie, Epstein-Barr virus (EBV) and cytomegalovirus (CMV) serology] and the 2 most commonly ordered expensive microbiology tests (hepatitis C and hepatitis B PCR).
We performed a budget impact analysis 18 of direct financial cost savings to the hospital of implementing duplicate order alerts. We used the Australian Medicare Benefits schedule fee to determine cost savings for each test: 19 urine culture (item 69333; AU$20.55 or US$13.21), fecal PCR which includes targets for C. difficile toxin (item 69496; AU$43.05 or $US27.67), hepatitis C PCR (item 69488; AU$180.25 or US$115.86), hepatitis B PCR (item 69482; AU$152.10 or US$97.77), EBV serology (item 69474; AU$28.65 or US$18.42), and CMV serology (item 69387; AU$29 or $US18.64).
We used specimen container volume to measure the waste disposal savings by weight: urine container (Sarstedt multipurpose container, 70 mL, 75.9922.721, 55 × 44 mm, 12 g), feces tube (Sarstedt 70 mL, 80.9924.027 55 × 44 mm, 13.5 g), hepatitis C and hepatitis B PCR (BD Vacutainer Plasma Preparation tube PPT, 5 mL, 362791, 13 ×100 mm, 8 g) and EBV/CMV serology (BD Vacutainer SST, 5 mL, 367954, 13 × 100 mm, 8 g). 20,21 We estimated the weight of each urine or feces sample by assuming that each container with 70 mL water weighed 70 grams. The mean weight of 1 mL blood is 1.06 g, and we estimated that the weight of a full PPT or SST tube containing 5 mL blood was 13.3 g including the weight of the empty tube (Supplementary Table 1 online). 22
We estimated greenhouse gas footprint, measured in carbon dioxide equivalent (CO2e) emissions. The CO2e attributed to sample collection consumables for EBV and CMV serology tests was based on previous calculations of 89 g per test for swabs, gloves, vacutainer holders, and collection tubes and specimen bags for Gold Top tubes. We estimated equivalent emissions produced by distance driven for which 1 g CO2e emissions produced equivalent emissions as 6.5 m travelled in a standard car. Reference McAlister, Barratt, Bell and McGain17
To determine the direct cost to the hospital of implementation of the duplicate order alerts, we reviewed information technology staff logs to calculate the time required for staff to configure the alerts and code behind them and to validate the functioning of the alerts.
We grouped tests based on appropriate time frames for repeat testing. These time frames (ranging from 3 days to 2 years) were based on national quality assurance and treatment guidelines, and Australian Medicare reimbursement time frames. They were agreed upon by medical microbiologists and senior scientific staff. Alerts were designed to trigger based on these test time frames. For each test group, we also formulated a customized comment to be included as an alert to guide the clinician’s decision making (Table 1). Examples of 2 different duplicate order alerts are shown in Figure 1 and Figure 2.
Table 1. Test Groups, Minimum Retest Intervals and Alert Text
Note. Retesting intervals were based on the following references (see Methods).
a Royal College of Pathologists of Australasia (RCPA) Quality Assurance Program Survey Reports.
b Reference 19 . Medicare Benefits Schedule. Australian Government, Department of Health and Aged Care. Published July 2022. Accessed 8th February 2023. https://testingportal.ashm.org.au/treatment-of-chronic-hepatitis-b-virus-infection/. Accessed February 8, 2023. https://www.ashm.org.au/resources/australian-recommendations-for-the-management-of-hepatitis-c-virus-infection-a-consensus-statement/. Accessed February 8, 2023.
c https://ranzcog.edu.au/wp-content/uploads/2022/05/Maternal-Group-B-Streptococcus-in-Pregnancy-Screening-and-Management.pdf.
Figure 1. Example of duplicate order alert for urine microscopy and culture.
Figure 2. Example of duplicate order alert for Hepatitis B PCR.
Ethics approval for the study was obtained from the Monash Health Research Office (no. RES-23-0000136Q-94743).
Results
Pathology ordering practice was studied over the 6-month period from February 17, 2022, to August 17, 2022, which was 16 months after the duplicate alerts were implemented. For these 6 tests, 4,020 duplicate requesters proceeded and 3,602 tests were cancelled, resulting in a 47% reduction in test ordering (Table 2).
Table 2. Direct Cost Savings and Direct Costs to the Hospital
Note. MCS, microscopy, culture, and sensitivity; PCR, polymerase chain reaction; EBT, Epstein-Barr virus; CMV, cytomegalovirus.
The 3,602 cancelled tests resulted in a financial cost saving of AU$100,274. The cost of waste saved by weight was calculated to be 280 kg during this 6-month period (209 kg for urine pots, 67.7kg for feces pots, and 3.2 kg for hepatitis C PCR/hepatitis B PCR/EBV serology and CMV serology tubes). The 199 cancelled EBV and CMV serology tests resulted in savings of 17,711 g of CO2e emissions for sample collection consumables, equivalent to driving 115 km in a standard car (Table 2).
The labor required for information technology staff to build and validate the alerts for the 6 tests studied was calculated to be 40 hours. The time required for the medical microbiologists to determine the retesting time frames and alert comments was not included as this work was considered as part of routine work and was partly performed for the test-rejection process that was in place prior to the duplicate alert implementation.
Discussion
Customized alerts issued at the time of test ordering can have an enormous impact on reducing cost, waste, and inappropriate testing. Our customized alerts were only triggered if the test was ordered within a prespecified time frame determined according to the test type. Alerts also contained information about appropriate reordering frequency or indications for retesting based on test type. Our alerts also stated when the previous test was requested. Because our alerts are not merely a simple general alert but are tailored to the test type, they are also a form of decision support.
Concern has been raised that “alert fatigue” may decrease the effectiveness of these types of interventions over time. However, we examined a period of 16 months after alert implementation and found that the duplicate alerts remained very powerful because approximately half of duplicate requests were cancelled after the alert was triggered.
The cost of pathology testing includes the financial cost to the laboratory as well as the cost to the wider community when government funding is used to reimburse the healthcare institution. We used Australian Medicare reimbursements to measure costs that essentially represent cost to the public rather than laboratory costs to perform the test. A large proportion of healthcare costs in Australia are for pathology services. During the 2020–21 financial year, 18.4% of all Medicare spending (AU$5.36 billion or US$3.45 billion) was for pathology services. 23 A 15-year meta-analysis concluded that 12%–44% of laboratory tests ordered globally are not clinically indicated. Reference Zhi, Ding and Theisen-Toupal24 Thus, reducing unnecessary testing can yield significant savings.
The cost of pathology waste not only includes the financial cost to the healthcare institution of waste collection and disposal but also includes the cost to the greater community of the environmental impact of loss of land use, loss of biodiversity, and contamination of waterways. In a recent Australian study, the carbon footprint of common pathology tests was dominated by those of sample collection. Reducing unnecessary testing was thought to be the most effective approach to reducing the carbon footprint of pathology testing. Reference McAlister, Barratt, Bell and McGain17 We used calculations from this study to estimate the carbon footprint of sample collection consumables for EBV and CMV serology tests because sample collection consumables were determined to be the main sources of CO2e emissions. However, the total carbon footprint savings of our alert implementation is likely to be much higher if other components of the test cycle (eg, tube and reagent manufacture, pneumatic tube system use, testing) are included and if the other tests are included particularly bearing in mind that the volume of the urine and feces pots is much greater than that of the blood containers.
Many laboratories without the facilities for electronic alerts at the point of test ordering have other processes in place to reduce duplicate testing. These processes are usually activated once a specimen has been received in the laboratory, which results in additional labor for the preanalytical steps of specimen collection, transport, access, and storage, as well as laboratory staff training regarding details of test rejection, including the time to inform clinicians of test rejection. Prior to implementation of duplicate order alerts, our laboratory had in place a process in which EBV and CMV serology testing was not performed for requests within 7 days of a previous test or for patients with IgG previously detected. Instead, a report was issued to alert the requester that the test was not performed. Because this process was manual, it was inconsistently applied, particularly due to the extra workload on staff to look up previous results, to issue the correct report, and to store the sample compared to the workload of simply doing the test. Additionally, the workload of receiving a call from the (often frustrated) requester and retrieving a stored specimen if repeat testing was clinically indicated was an extra barrier for adherence to this test nonperformance protocol. Significant further cost savings are likely when reduction of repeat testing is implemented prior to a sample being received in the laboratory. Due to their complexity, we did not measure costs savings associated with these preanalytical steps. Test cancellation by clinicians at the point of ordering is a more efficient and robust method of reducing duplicate testing and allows the valuable resource of laboratory staff to be used more wisely.
Calls to reduce inappropriate diagnostic testing have suggested that computerized ordering systems can improve ordering practice by providing real-time feedback and guidance on test appropriateness. Reference Hammett and Harris25 In addition, when the decision to cancel a test is made by the clinical team rather than by laboratory staff, it is more likely to be accepted because the process is less authoritarian or confrontational for reducing waste. Behavioral approaches to guide decision making through the strategic placement of choice architecture while still maintaining prescriber autonomy are called “nudging strategies” and have been used successfully in altering the antibiotic selection of the prescriber. Reference Langford, Leung and Haj26 Information in the alert text is also an opportunity to educate requesting clinicians about appropriate reordering time frames and indications. This process is likely to promote a more collegial “team effort” approach to appropriate test ordering. Furthermore, because pathology test ordering has been shown to change over the course of a medical practitioner’s career and may increase significantly during the first 2 years of clinical practice, Reference Magin, Tapley and Morgan27 it may be particularly beneficial to utilize these educational activities to foster good ordering practices early in a clinician’s career during training in the hospital system.
Choosing Wisely Australia is part of a global healthcare initiative to promote a national dialogue on unnecessary tests, treatments, and procedures. Our pathology program at Monash Health was the champion organization of the inaugural Choosing Wisely Awards in 2022, and our EMR alerts were part of this testing optimization project. 28
For our study, we only calculated cost savings of duplicate alert implementation for the 4 most ordered microbiology tests and the 2 most commonly ordered expensive tests. Because we did not calculate cost savings for the other 77 microbiology tests for which duplicate alerts were implemented, the total cost savings of alert implementation will likely be much greater. Currently in our institution, EMR is only implemented for inpatients. We expect that broader implementation of EMR with duplicate alerts to the outpatient setting will maximize the benefit. This added benefit may particularly affect the less frequently ordered but more expensive tests such as hepatitis B and hepatitis C PCR tests, which are ordered in greater number through the outpatient setting. The 2023 guidelines from the Society for Healthcare Epidemiology of America Diagnostic Stewardship Task Force mention restriction of repeat testing for C. difficile and pneumonia multiplex panel, but we have found that restriction of duplicate testing for many more tests can be a very powerful tool for diagnostic stewardship. Reference Fabre, Davis and Diekema29 Our study was a single-center study with 1 laboratory serving 3 teaching hospital campuses; further studies are needed in other laboratories to elucidate wider applicability in other situations.
We calculated the direct cost to the hospital of information technology staff labor to implement the duplicate alerts. This cost is only an initial financial outlay without any requirement for ongoing maintenance costs. To estimate the cost of waste by weight we estimated container volume to measure the waste disposal savings, but we acknowledge that this estimation may be affected by the fact that most containers are not completely filled on specimen submission and this is only an approximate estimate. However, this novel but important measurement has not been previously studied. We did not quantify indirect costs to the hospital such as potential harm of test nonperformance; for example, the yield of repeated C. difficile testing is low but not zero. However, our alerts are “soft stops” (ie, the requester still has the option to the request the test after reading the alert), so this risk is low.
Another limitation of our study was that cost for EBV and CMV serology calculations have assumed that only IgG and IgM are tested, which underestimates the true cost savings. For example, algorithms for testing in some laboratories states that if IgM is detected, CMV IgG avidity or EBV nuclear antigen IgG antibody are performed, with a recommendation for repeat serology. Furthermore, potential further cost savings of cascade testing, particularly in the case of false-positive IgM results, are not included here. Lastly, we did not measure the impact of alerts on clinician workload, nor did we assess the magnitude of the benefits of reducing repeat testing on patient health outcomes, comfort, and satisfaction. Although these aspects are more difficult to measure, they are nonetheless important to consider when implementing CCDS interventions.
In conclusion, we conclude that customized electronic duplicate order alerts issued at the time of request ordering can have an enormous impact on reducing the detrimental economic, health, and environmental effects of unnecessary testing. Further studies are required to ascertain the wider effects of such initiatives beyond financial gains. Institutional support for sophisticated digital solutions to optimize pathology testing can facilitate this endeavor.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/ice.2023.198
Acknowledgments
We thank Dr Samar Ojaimi (Infectious Diseases Physician, Immuno-Pathologist and Clinical Immunologist at Monash Health) for reviewing this manuscript.
Financial support
No financial support was provided relevant to this article.
Conflicts of interest
All authors report no conflicts of interest relevant to this article.
Open access
