Laboratory medicine, a branch of medicine and a component of the healthcare delivery system that provides diagnostic services such as laboratory investigations, the results of which are critical to clinicians' disease management decisions, is unfortunately affected by unnecessary laboratory investigations that are not required but repeatedly requested due to irregular documentation by clinicians. ^{1 – 3} These unnecessary laboratory test requests could indicate a request made for the wrong patient, at the wrong time, or for the wrong disease condition. ⁴ A B-type natriuretic peptide (NT-proBNP) test, for example, is useful as a diagnostic test for congestive heart failure patients. However, evidence suggests that while there is no guideline for repeated NT-proBNP measurements in the management of heart failure, it should be measured only once except in certain pragmatic situations such as clinical deterioration. As a result, requesting this test repeatedly, such as on a daily basis, is not the best course of action for determining the severity or stability of a heart condition. ⁴ A rippling effect of this is that it leads to an increase in patient discomfort, duplication of laboratory investigations and monetary costs associated with the use of kits, staff and time. ^{3 – 5} Consequently, countries have taken decisive action to review, develop robust and multifaceted strategies to tackle waste associated with repeated laboratory investigation and adopt those that work best in their environment. ^{4 , 6 – 8} These initiatives are a response to the recognition of the need for an urgent reduction in unnecessary laboratory test requests and costs. ^{9 , 10}

In 2010, the South African National Department of Health (NDoH) and its partner - National Health Laboratory Services (NHLS), collaborated to develop a sustainable and affordable Electronic Gate Keeping (EGK) system. The EGK is a demand management system, defined as a full electronic stop (system-based intervention) to reduce overutilization on some laboratory test orders. ¹¹ It is a Laboratory Interface System (LIS) used to approve or reject laboratory tests based on the protocols for patient management devised by hospital management; the NHLS translates the rules into LIS. Upon a rejection of a test, an SMS alert is sent to the requesting medical officer notifying them that a test has been rejected. The SMS reads “Test automatically rejected on the basis of hospital electronic gate keeping rules”. If it is clinically necessary for the test to be performed, the rejection can only be reversed by providing a valid approval number to the laboratory, this must be done before the stability of the test is exceeded.

The primary application of this demand management is to ensure appropriate laboratory requesting and testing; for the right patient at the right time. For example, requesting a prostate specific antigen test for a male patient over 40 years for screening every six months if the previous result was normal, is appropriate and necessary testing. Since the inception of EGK in South Africa from 2011 to date, only two published articles have evaluated the EGK system. The two previous studies by Smit et al. ⁴ in 2015 and Pema et al. ⁸ in 2017 were conducted in urban hospital settings in the Western Cape and Gauteng provinces respectively. However, there was constrasting findings on the cost effectiveness of EGK. Therefore, this study builds on the previous studies' work and aims to provide evidence on the cost-effectiveness of the EGK system on laboratory tests requested at a rural academic hospital in the Eastern Cape province of South Africa.

This study purposed to determine the cost savings obtained from tests rejected by the EGK system; a demand management tool (National Health Laboratory Service 2012). This is the first study that investigated the cost effectiveness of EGK, particularly in a rural setting in South Africa. In this paper, the majority of the 28 types of tests audited were low cost tests, i.e. were above USD 1.50 but less than USD 5.30 per test; and the minority of the tests studied that were high cost tests, i.e. were above USD 7.50 but less than USD 11.00 per test. We also found that the test that generated the highest savings was a low cost c-reactive protein (CRP) test; followed by free thyroxine (FT4) and hepatitis A IgM (HAM) which are both in the category of high cost tests. The top 15 tests that were rejected and thereby generated the highest cost savings were tests that are usually requested in a bundle with other accompanying tests; except for the CRP. In particular, the FT4 which generated the second highest cost savings in our study is a test that is often requested in a bundle, i.e. a thyroid function test, comprising a thyroid stimulating hormone (TSH) and tri-iodothyronine (FT3). Similarly, HAM generated the third highest cost savings because this test is usually requested in a bundle as part of a hepatitis screen, which is unnecessary most times, not unless it is clearly marked and requested on the laboratory form.

When we compare our study findings with that from the only two other studies that investigated the same intervention in South Africa, ^{4 , 8} we found that the tests audited were quite similar; i.e. were basic, routine, frequently requested tests and as Konger et al. ¹⁴ would say, ‘these tests were unnecessary and overused’. Furtheron, while we studied 28 tests, Smit et al. ⁴ and Pema et al. ⁸ studied 26 and 31 tests respectively. Our study revealed test requests of 448,028, while the number of requested tests by Smit et al. ⁴ was 68,480 and Pema et al. ⁸ was 1,493,965. The difference in the number of tests requested between our study and these studies could be due to the hospital facility's population coverage, which is much larger than our study setting, which serves a rural district with just over 5 million. Furthermore, the meagre number of test requests in the study by Smit et al. ⁴ compared to our study could be justified and linked to their short six-month study period; thus, we presume that if their study were more than six months, more tests would have been requested.

Based on the top three repertoire of tests requested in this study compared to other studies, creatinine, urea and full blood count were the three most requested in this study. On the other hand, Smit et al. ⁴ demonstrates calcium, magnesium and phosphate while Pema et al. ⁸ reports the top most requested tests as urea and electrolytes, calcium, magnesium and phosphate. Parallel to the number of tests requested between our study and the two other studies, our study presented a rejection rate of 3.9%, which was found to be insignificantly higher (0.72%) than the rejection rate reported by Pema et al. at 3.18%, but significantly lower (2.9%) than the rejection rate reported by Smit et al. at 6.7%. The varying degree of rejection rate, especially between our study and Smit et al., may be linked mainly to our higher number of test requests compared to theirs and/or possibly the criteria of the EGK test rules applied between the two hospital settings. Regarding the cost savings, our study found the top three tests that generated the most savings to be single tests, i.e. CRP, FT4 and HAM respectively, whereas Smit et al. reported their cost savings to be generated by the thyroid function panel, liver function panel and CRP respectively. The CRP was the common test between the two studies. The tests that generated the greatest cost savings as reported by Pema et al. were urea and electrolytes panel, thyroid function tests and glycated hemoglobin. There were no similarities regarding the tests that generated cost savings between these two studies. In contrast, while our greatest cost savings were obtained on single tests, Smit et al. and Pema et al. reported having obtained theirs mainly on test panels and or profiles.

Based on these findings, we suggest that requesting tests in a bundle is discouraged as this may contribute to unnecessary testing that may not be clinically indicated. Besides, another resultant effect of requesting tests in a bundle is wasted resources accompanied by increased expenditure and increased laboratory workload amongst other things. So we advocte for selective testing to be the most encouraged type of medical laboratory investigations. More importantlly we note that to use a single intervention method, like we did with the EGK system, to address unnecessary laboratory testing might not have a great impact and/or result; but rather is a multifaceted approach that should be used for better outcomes. Therefore this novel EGK sytem as an intervention should be used in conjunction with continuous education focusing on the importance of making mindful laboratory test requesting, selective testing, the costs of individual tests and the impact these have on expenditures when they are requested unnecessarily; following of recommended guidelines on the management of acute and chronic diseases. Lastly, additional studies are recommended to assess the cost effectiveness of introducing EGK at across all facilities nationwide.