Cost-effectiveness of point-of-care devices for detection of anemia in community settings in India
Sutapa Bandyopadhyay Neogi, Denny John, Jyoti Prakash Sharma, Rakhee Kar, Sitanshu Sekhar Kar, Maitreyee Bhattacharya, Kartavya Tiwari, Renu Saxena · Clinical Epidemiology and Global Health · 2022
Anemia is a public health problem affecting 46.3% of the population in India.1Didzun O. De Neve J.-W. Awasthi A. et al.Anaemia among men in India: a nationally representative cross-sectional study.Lancet Global Health. 2019; 7: e1685-e1694Google Scholar It affects 58% children under five years of age, 54% adolescent girls of 15–19 years of age, 53% women of reproductive age, 50% pregnant mothers, and 29% adolescent boys in the country.2IIPS: National Family Health Survey 4. in: India Fact Sheet. Sciences IIoP. Mumbai, India2015-16Google Scholar Detection of anemia is important for institution of appropriate management. Several methods are available for diagnosis of anemia especially in outreach and field settings where well-equipped laboratories are unavailable. The devices have been tested in diverse settings and have yielded varied levels of accuracy.3Chaudhary R. Dubey A. Sonker A. Techniques used for the screening of hemoglobin levels in blood donors: current insights and future directions.Hematol Res Rev. 2017; 8: 75-88Google Scholar, 4Dubey A. Murti S. Validation of a hemoglobinometer for use in outdoor blood donation camps.Asian J Transfus Sci. 2016; 10Google Scholar, 5Gamal M. Abdelhamid B. Zakaria D. et al.Evaluation of noninvasive hemoglobin monitoring in trauma patients with low hemoglobin levels.Shock: Inj Inflamm Sepsis: Lab Clin Approaches. 2018; 49: 150-153Google Scholar, 6Neogi S.B. Negandhi H. Kar R. et al.Diagnostic accuracy of haemoglobin colour strip (HCS-HLL), a digital haemoglobinometer (TrueHb) and a non-invasive device (TouchHb) for screening patients with anaemia.J clin Pathol. 2016; 69: 164-170Google Scholar, 7Parker M. Han Z. Abu-Haydar E. et al.An evaluation of hemoglobin measurement tools and their accuracy and reliability when screening for child anemia in Rwanda: a randomized study.PLoS One. 2018; 13 (e0187663-e0187663)Google Scholar In the absence of a concrete evidence, there is a need to examine the devices that have the potential to be included in public health programs. Given the fact that most of the screening happens in community and outreach settings where provision for a laboratory support seems difficult, the device ought to be tested in field settings with health workers (Auxillary Nurse Midwives or ANMs) as end users. Currently four such devices have been identified that have the potential to be used in public health settings, namely, digital hemoglobinometers (True Hb and Hemocue), and non-invasive devices, i.e. Masimo Pulse Oximetry test and Spectroscopic spectroscopic device test.8Neogi S. Negandhi H. Sharma J. Ray S. Saxena R. Diagnostic efficacy of digital hemoglobinometer (TrueHb), HemoCue and non invasive devices for screening patients for anemia in the field settings-a proposal.Indian J Community Health. 2018; 30 (Supp): 86-88Google Scholar Identification of a suitable device for public health services warrants analysis of parameters other than the diagnostic accuracy. Some such parameters include cost of devices, cost effectiveness, user-friendliness and operational factors. There is a need to evaluate them in order to identify the most cost effective device suitable for use in Indian and similar settings. A multicentric study across four sites was undertaken to assess the diagnostic accuracy (effectiveness) of these devices (described elsewhere).9Neogi S.B. Sharma J. Pandey S. et al.Diagnostic accuracy of point-of-care devices for detection of anemia in community settings in India.BMC Health Serv Res. 2020; 20: 468Google Scholar The objective of the current study is to estimate the cost effectiveness of the devices by evaluating the incremental cost for each case detected from the health systems perspective. The secondary objective is to assess the user friendliness and operational issues of the devices for detection of anemia in field settings in India. An analysis of the performance of some such devices was undertaken recently. Four devices (index tests) namely, True Hb Digital Hemoglobinometer, HemoCue, Masimo's non-invasive device and Spectroscopic Spectroscopic devices were examined for diagnostic accuracy on an adequate sample of adult population attending field practice areas of Medical colleges in India: Jawaharlal Institute of Postgraduate Medical Education and Research (JIPMER), Puducherry, and Calcutta Medical College, Kolkata, West Bengal. To test the accuracy in different weather conditions, the devices were additionally tested at All India Institute of Medical Sciences (AIIMS) Jodhpur and Rural hospital, Reckong Peo (Himachal Pradesh). The index tests were compared against autoanalyzer (gold standard). Accuracy was expressed by sensitivity, specificity, likelihood ratios, predictive values and area under the curve (AUC) and levels of agreement. According to this study, HemoCue and TrueHb performed better than Massimo and spectroscopic devices. Detection of anemia by technicians was similar between TrueHb and HemoCue (AUC 0.92 v/s 0.90, p > 0.05). Community workers performed better with Hemocue for detecting anemia compared to TrueHb (AUC 0.92 v/s 0.90, p < 0.05). For detection of severe anemia, accuracy of TrueHb was significantly better with technicians (AUC 0.91 v/s 0.70; p < 0.05) and community workers (AUC 0.91 v/s 0.73; p < 0.05). HemoCue showed a bias or mean difference (95%CI) of 0.47 g/dl (0.42, 0.52), and 0.92 g/dl (0.82, 1.03) for severe anemia. For TrueHb, it was −0.28 g/dl (−0.37, −0.20) for all readings, and 0.06 g/dl (−0.52, 0.63) for severe anemia.9Neogi S.B. Sharma J. Pandey S. et al.Diagnostic accuracy of point-of-care devices for detection of anemia in community settings in India.BMC Health Serv Res. 2020; 20: 468Google Scholar This study on cost effectiveness was undertaken alongside the diagnostic accuracy study with cross-sectional design. The data collection for the study was conducted between August 2018 and March 2019. The protocol of the cost-effectiveness analysis approach has been published elsewhere.10Neogi S. John D. Sharma J. et al.Cost-effectiveness of invasive devices versus non-invasive devices for screening of anemia in field settings in India: a study protocol [version 1; peer review: 1 not approved].F1000Research. 2019; 8Google Scholar The costs were captured through primary data collection using a micro-costing (bottom up) methodology from the study sites. At the first step, every input consumed in the detection of anemia using the point-of-care devices in an individual was identified. Secondly, the quantity of each resource item, such as human resource (ANM time for each test), equipment (device, charger, adapter), accessories (micro cuvettes/strips), consumables (used in test), non-medical (items in facility rooms), and capital space (i.e. the space used for conducting the tests) was identified. At the third step, unit costs were calculated for each of these resource items. Inputs from device manufacturers were used to cost the equipment and accessories, while procurement costs were considered for consumables and no-medical costs items. Unit cost for ANM was calculated by dividing the total salary per month divided by the time taken by ANM for each test (i.e. 26 working days with 7 h per day was time considered per month for receiving full salary). Other resource costs was calculated by multiplying the per unit costs with the total units consumed. Rental value for the space used for testing was captured using reference rental values prevalent in the area. Since the study sites were government hospitals, shared costs, and administrative costs were not captured as the administration requirement for receiving this data was time-consuming, hence these costs were not included in the total cost analysis. Data on costs and sensitivity of the devices for anemia detection were collected simultaneously at each study site. Effectiveness is defined here as the case detection rate or sensitivity of the devices to detect anemia and severe anemia. The details are available elsewhere.10Neogi S. John D. Sharma J. et al.Cost-effectiveness of invasive devices versus non-invasive devices for screening of anemia in field settings in India: a study protocol [version 1; peer review: 1 not approved].F1000Research. 2019; 8Google Scholar Cost effectiveness analysis was performed to assess the relative efficiency of the invasive and non-invasive methods for detecting anemia by health workers. The Incremental Cost Effectiveness Ratios (ICERs) were calculated from costs (reported in this study) and effectiveness (reported elsewhere).9Neogi S.B. Sharma J. Pandey S. et al.Diagnostic accuracy of point-of-care devices for detection of anemia in community settings in India.BMC Health Serv Res. 2020; 20: 468Google Scholar The ICER was calculated as: [(Mean cost per test) device A - (Mean cost per test) device B]/[(Detection rate) device A – (Detection rate) device B]. The accuracy of the devices was compared against the autoanalyzer (gold standard) in the main study. However, the utility of autoanalyzer and point of care devices are very different and hence comparison for cost effective analysis was difficult. In the absence of a suitable comparator, for the purpose of calculations, we considered HemoCue against which all other devices were compared. Since the data for both costs and effectiveness were for an annual year, the costs were not discounted as per standard guidelines for economic evaluations.11Husereau D. Drummond M. Petrou S. et al.Consolidated health economic evaluation reporting standards (CHEERS) statement.BMJ Br Med J (Clin Res Ed). 2013; 346: f1049Google Scholar Sensitivity analysis was conducted with lower and upper bounds of sensitivity of diagnostic parameters across all medical devices. Scenario analysis for detection of anemia in an annual year was also conducted with the assumption that an ANM can screen patients 2 days/week for 1 year (approximately 50 weeks), and assuming 10 min for each test, will be able to examine 6 patients in 1 h or 42 patients per day (assuming 7 h of work per day). The scenario also assumes that the ANMs will be conducting testing for anemia with no outside work. The analysis was done from a quasi-health system perspective with a time horizon of one year. The outcome measure was ‘correct diagnosis’ that referred to the sensitivity of the devices i.e. ability of the devices to identify true positives. Given the fact that severe anemia is a priority of the national program, cost effectiveness was calculated separately for anemia and severe anemia. Reporting of study findings followed CHEERS guidelines.11Husereau D. Drummond M. Petrou S. et al.Consolidated health economic evaluation reporting standards (CHEERS) statement.BMJ Br Med J (Clin Res Ed). 2013; 346: f1049Google Scholar Although cost utility analysis would have been a better estimate, this was not deemed appropriate in the current study. Hence, it was not included. Operational issues while using a device also aid in taking a decision on the feasibility of its use in field settings. We therefore explored different aspects like ease of use, portability, objectivity in readings, average time taken to perform each test, expertise required and strip/cuvette wastage. These were gathered from specifications mentioned in product inserts, from the manufacturers and research staff involved in the study. None of the patients had any immediate complications following needle pricks although we could not record any long term effects due to logistic reasons. Laboratory technicians and ANMs independently asked to rate each of the following criteria: ease of use, efficiency in daylight, scope of subjective errors, portability, convenience to patient, interpretation of hemoglobin (Hb) results, need for power/battery, average time taken for performing one test, and expertise required, on a scale of 1 (very poor) to 5 (excellent) for each device method. The total score across each of these parameters for laboratory technicians and ANMs were used to identify user friendliness for each method (maximum score for each method). These were subjective assessments to understand the perspectives of the users (4 per study site, 16 health workers in total). The individual cost of every device is shown in Table 1. The cost of conducting each test was minimum with Massimo (INR 104.8) in rural areas and with Spectroscopic (INR 140.0) in urban and maximum with TrueHb (INR 137.0 in rural and 177.4 in urban areas). Costs in urban areas are more than in rural areas. Additional cost to the health system for conducting each test in rural areas was found to be INR 24.4 with HemoCue while it was INR 38.7 with TrueHb. (Table 1).Table 1Costs of resources for each test for measuring Hemoglobin (in INR).ComponentData sourcesHemocueTrue HbSpectroscopic deviceMasimoRuralUrbanRuralUrbanRuralUrbanRuralUrbanHuman resource (ANM)Health systems data16.7 (13.9%)17.7 (10.7%)17.1 (12.5%)17.5 (9.9%)18.1 (16.5%)18.1 (12.9%)16.6 (15.8%)18.3 (12.6%)Equipment (device, charger, adapter)Device manufacturers0.2 (0.1%)0.2 (0.1%)0.004 (0%)0.004 (0%)0.05 (0.04%)0.05 (0.04%)3.4 (3.2%)3.4 (2.3%)Accessories (Micro cuvettes/strips)Device manufacturers11.9 (9.9%)11.9 (7.2%)26.4 (19.3%)26.4 (14.9%)0.0 (0%)0.0 (0%)0.0 (0%)0.0 (0%)Consumables (items used in the test)Collected from hospital records12.3 (10.2%)19.7 (12.0%)12.3 (9.0%)19.7 (11.1%)6.1 (5.6%)4.3 (3.1%)6.1 (5.8%)4.3 (3.0%)Non-medical (items in facility rooms)Collected from hospital records78.8 (65.5%)113.4 (69.0%)80.8 (59.0%)112.3 (63.3%)85.4 (77.6%)116.4 (82.9%)78.3 (74.7%)117.2 (81.0%)Capital space (rental)Collected from hospital records0.4 (0.3%)1.6 (0.9%)0.4 (0.3%)1.5 (0.9%)0.5 (0.4%)1.6 (1.1%)0.4 (0.4%)1.6 (1.1%)Total cost/test (INR)120.3164.3137.0177.4110.1140.4104.8144.7 Open table in a new tab The cost of consumables of testing devices contributed highest among Hemocue (12.1% of total cost), and True Hb (11.5% of total) in urban areas. The non-medical items in the examination room (such as furniture-examination table chairs, cupboards, curtains, wash basins, soaps, buckets, light, fan etc) contributed to the majority of the total costs for each device. The cost calculated per case detected and effectiveness results (case detection rate or sensitivity) were calculated to compute the ICER for every device. HemoCue was found to be more cost effective than True Hb for detection of anemia while TrueHb is more cost effective for detection of severe anemia for both urban and rural areas. Since the users in the study were ANMs and we are primarily looking at rural areas where such point of care devices will be better utilized, we present the results of rural areas (Table 2). Sensitivity analysis for upper bound of sensitivity values of diagnostic tests showed higher costs per detection rate.Table 2Incremental cost effectiveness ratios (ICER) of devices for detection of anemia and severe anemia (costs in INR).ComponentsHemoCueTrueHbSpectroscopic deviceMassimoAnemia Base caseReference−464.330.465.1 Lower bound of sensitivityReference−407.728.060.0 Upper bound of sensitivityReference−557.134.073.3Severe anemia Base caseReference41.554.152.2 Lower bound of sensitivityReference46.246.458.9 Upper bound of sensitivityReference45.896.853.4 Open table in a new tab Table 3 provides results of scenario analysis across all anemia and severe anemia. HemoCue will detect 1869 cases per year per ANM with a cost of INR 224840 per device year, with a wastage of INR 27789.30, for all anemia. In comparison TrueHb will identify 1806 cases with an cost of INR 247422 per device per year, with a wastage of INR 402278. With regards to severe anemia, TrueHb will detected 73 cases with cost of detection at INR 11508 per year. While HemoCue will be able to detect only 39 cases of severe anemia with cost of detection at INR 10105 per year.Table 3Scenario analysis for detection of anemia in year using point of care devices.Haemo CueTrue HbSpectroscopic deviceMasimoDetection of anemiaTotal Cost Per unit Test (INR)120.3137110.1104.8Total number of tests in 1 year per ANM (approx)aThe total numbers of tests an ANM is required to perform was estimated from a population of 7000–8000 that she has to cater to. The approximate number of pregnant women/@ crude birth rate of 20/1000 results in 150–180 pregnant women and Hb estimation at every visit makes it (1000 tests/yr) plus women of population and of population children plus some wastage. that an ANM can perform for days a the year, the approximate number of tests per year was to be number of cases of anemia number of cases of anemia on sensitivity detected per number of anemia cases detected per for detection of anemia using the device per year cost for the system for detection of anemia per year to the system of severe anemiaTotal Cost Per unit Test (INR)120.3137110.1104.8Total number of tests in 1 year per number of cases of severe anemia on available number of cases of severe anemia detected per number of cases detected per for detection of severe anemia using the device per year to the system cost for the system for detection of anemia per year The total numbers of tests an ANM is required to perform was estimated from a population of 7000–8000 that she has to cater to. The approximate number of pregnant women/@ crude birth rate of 20/1000 results in 150–180 pregnant women and Hb estimation at every visit makes it (1000 tests/yr) plus women of population and of population children plus some wastage. that an ANM can perform for days a the year, the approximate number of tests per year was to be Open table in a new tab While a device for use, parameters were taken Although accuracy across the of the in maximum other also this the four devices, HemoCue was the by the health workers followed by TrueHb, Massimo and Spectroscopic device (Table HemoCue better in of user wastage per test, of the equipment and the accessories, and number of tests that can be done in field settings with the other True Hb HemoCue with regards to of the of for its use and and feasibility to use it in community settings to use (Table HemoCue showed a bias or mean difference of 0.92 g/dl (0.82, 1.03) for severe anemia.9Neogi S.B. Sharma J. Pandey S. et al.Diagnostic accuracy of point-of-care devices for detection of anemia in community settings in India.BMC Health Serv Res. 2020; 20: 468Google Scholar by 1 the of anemia and this can of user-friendliness of the devices by health of in to for time taken for performing one of Open table in a new tab Table of digital hemoglobinometers and detection rate or in Hb compared to standard for all detection rate or in Hb compared to standard in severe at different weather Hb of the (in of use by health workers and of in for wastage per test per of the of of in of in be no of tests in field settings and and under - Open table in a new tab - The analysis of the four devices that HemoCue is more cost effective than TrueHb for detection of anemia True Hb is more cost effective for severe anemia. The HemoCue was found to be more user than TrueHb as mentioned the health workers. With regards to the operational TrueHb HemoCue for most factors. is expressed not only in of of using diagnostic evaluation of diagnostic tests is different from that of or B. to cost-effectiveness of diagnostic a of health Med Res 2017; Scholar very not evaluation of M. A. R. evaluation for devices and or Health J Res. H. B. accuracy to outcome and cost-effectiveness of diagnostic tests and an Med Res 2013; Scholar In outcome of anemia is to from diagnosis that be efficacy of the device on the users and costs of the devices on the numbers of users. There are on the of use of to these there is no to these with of diagnostic M. decision a approach to medical equipment 20: Scholar In a similar study from HemoCue was compared to different methods for estimation of A. J. and costs of different haemoglobin methods for use in in clin Pathol. Scholar on user friendliness and HemoCue was for use in While the in that study used accuracy for detection of anemia, we have used accuracy for anemia and severe anemia in the requirement of the health system in the and accessories contributed to the cost of an individual test in while in case non-medical costs in health had the The that be while a device for a national are accuracy across different field in the absence of and A. et effectiveness of medical devices to in 2017; Scholar We could assess the user-friendliness with health workers as the end users. We also the accuracy across different field However, the devices to were new and the manufacturers mentioned that of digital hemoglobinometers was not We are not the requirement of and therefore this a We were not able to the full cost due to data on costs of costs and cost of the the study was done on adult patients for of accuracy. The national primarily with anemia in pregnant women and future economic at all done be able to calculated for population of pregnant women and This study both cost and effectiveness data from a primary study conducted in a low and i.e. which to this analysis. This is important from higher to on most effective of A. et effectiveness of medical devices to in 2017; Scholar Diagnostic accuracy to with the of the and hence it is in a where it is to be used in Diagnostic test accuracy with for Clin Scholar is an that on the of the between all and to an for a et the of point-of-care diagnostic 2017; 8: Scholar To on the diagnostic accuracy of the devices across different field conditions, costs and detection for anemia and severe anemia, digital hemoglobinometers performed better than non-invasive devices. the of digital hemoglobinometers we found that for detection of anemia, Hemocue is most cost However, for detection of severe anemia is to and other TrueHb was found to be more cost With the of a will be required for of of the devices and to the