Background: Reactive case detection (RACD) is an active case finding strategy where households and neighbours of a passively identified case (index case) are screened to identify and treat additional malaria infections with the goal of gathering surveillance information and potentially reducing further transmission. Although it is widely considered a key strategy in low burden settings, little is known about the costs and the cost-effectiveness of different diagnostic methods used for RACD. The aims of this study were to measure the cost of conducting RACD and compare the cost- effectiveness of microscopy to the more sensitive diagnostic method loop-mediated isothermal amplification (LAMP). Methods: The study was conducted in RACD surveillance sites in five sub-districts in Aceh Besar, Indonesia. The cost inputs and yield of implementing RACD with microscopy and/or LAMP were collected prospectively over a 20 months study period between May 2014 and December 2015. Costs and cost-effectiveness (USD) of the different strategies were examined. The main cost measures were cost per RACD event, per person screened, per population at risk (PAR); defined as total population in each sub-district, and per infection found. The main cost-effectiveness meas- ure was incremental cost-effectiveness ratio (ICER), expressed as cost per malaria infection detected by LAMP versus microscopy. The effects of varying test positivity rate or diagnostic yield on cost per infection identified and ICER were also assessed. Results: Among 1495 household members and neighbours screened in 36 RACD events, two infections were detected by microscopy and confirmed by LAMP, and four infections were missed by microscopy but detected by LAMP. The average total cost of conducting RACD using microscopy and LAMP was $1178 per event with LAMP- specific consumables and personnel being the main cost drivers. The average cost of screening one individual during RACD was $11, with an additional cost of diagnostics at $0.62 and $16 per person for microscopy and LAMP, respec- tively. As a public health intervention, RACD using both diagnostics cost an average of $0.42 per PAR per year. Com- paring RACD using microscopy only versus RACD using LAMP only, the cost per infection found was $8930 and $6915, respectively. To add LAMP as an additional intervention accompanying RACD would cost $9 per individual screened annually in this setting. The ICER was estimated to be $5907 per additional malaria infection detected by LAMP versus *Correspondence: Michelle.Hsiang@UTSouthwestern.edu Malaria Elimination Initiative, Global Health Group, University of California, San Francisco (UCSF), San Francisco, USA Full list of author information is available at the end of the article © The Author(s) 2018. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creat iveco mmons .org/ publi cdoma in/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Zelman et al. Malar J (2018) 17:220 Page 2 of 12 microscopy. Cost per infection identified and ICER declined with increasing test positivity rate and increasing diagnos- tic yield. Conclusions: This study provides the first estimates on the cost and cost-effectiveness of RACD from a low transmis- sion setting. Costs per individual screened were high, though costs per PAR were low. Compared to microscopy, the use of LAMP in RACD was more costly but more cost-effective for the detection of infections, with diminishing returns observed when findings were extrapolated to scenarios with higher prevalence of infection using more sensitive diagnostics. As malaria programmes consider active case detection and the integration of more sensitive diagnostics, these findings may inform strategic and budgetary planning. To enable national malaria programmes to assess Background the feasibility of implementing RACD in their low bur- Malaria transmission in low endemic areas tends to den setting, this study aimed to: (1) estimate the cost of cluster geographically and temporally. Also, in low implementing RACD using data from the malaria-elim- transmission settings, a higher proportion of infected inating district of Aceh Besar, Indonesia; and (2) com- individuals is asymptomatic, and therefore will not be pare the cost and cost-effectiveness of using a standard detected by passive surveillance which occurs through diagnostic (microscopy) to more sensitive LAMP for health facilities. Reactive case detection (RACD), or identifying infections during RACD. This information active case finding among households and neighbours can help guide strategic and budgetary planning for local of passively identified index cases, is a strategy to malaria control programmes, and also help to inform the address this challenge. With the goal of RACD being the research, development, and implementation of highly identification and treatment of asymptomatic or other sensitive diagnostics for malaria. infections that would not otherwise present through the passive surveillance system, it is considered a key Methods malaria strategy for gathering surveillance informa- Study design tion  and may reduce transmission and facilitate The study was a prospective economic analysis of costs the achievement of malaria elimination [2, 3]. RACD and cost-effectiveness. The main cost measures were cost is widely implemented , yet it is operationally chal- per RACD event, per person screened, per population lenging requiring significant human resources, com - at risk (PAR; defined as total population in each sub-dis - modities, and time for an “on-call” team to conduct trict), and per infection found. The main cost-effective - screenings in villages, often traveling long distances to ness measure was incremental cost-effectiveness ratio reach remote locations. There are also limitations with (ICER), expressed as cost per malaria infection detected the standard diagnostics used, microscopy or rapid by LAMP versus microscopy. diagnostic test (RDT), to detect low-density infections. Highly sensitive diagnostics are available but the costs Study location and cost-effectiveness of using them is unclear. Further, Aceh Besar District is located in the western part of Aceh the costs of conducting RACD in general have not been Province on Sumatra Island, Indonesia. Malaria trans- systematically documented [5, 6]. mission occurs year-round with the higher transmission The detection limit of microscopy or RDTs is typically occurring from January to July. In the past, Plasmodium 100 parasites/µL, and in low endemic settings, a high falciparum and Plasmodium vivax were reported as the proportion of asymptomatic infections fall below this main malaria species, but Plasmodium knowlesi has also threshold [7, 8]. Outside of use in research settings, more been reported from the area in 2016 . Due to intensi- sensitive detection methods such as polymerase chain fied efforts in case management, vector control, and sur - reaction (PCR) are impractical due to high cost, sophis- veillance, Aceh Besar District has successfully reduced ticated training, resources required, and long turnaround annual malaria incidence from 2.6 cases per 1000 popu- time (several hours). Another molecular method called lation in 2006 to 0.2 per 1000 in 2014 (population in loop-mediated isothermal amplification (LAMP) pro - 2014 was 371,412 ). It is one of 114 low endemic vides the sensitivity of PCR with fewer requirements. districts categorized by the Indonesian government as Testing is not point-of-care and requires use of a labora- ‘eliminating’, with a goal to interrupt transmission by tory, but the assay is simple, does not require sophisti- 2020 [13, 14]. As part of the malaria elimination strat- cated equipment, and can be performed in half a day to egy, the District Health Office in 2010 initiated RACD, one full day [9, 10]. However, the costs and cost effective - which is locally referred to as ‘contact survey’. After a ness of using LAMP in RACD are not clear. Zelman et al. Malar J (2018) 17:220 Page 3 of 12 microscopy-confirmed malaria case is diagnosed at the smear. If positive, an additional 100 high-powered fields health facility, designated surveillance staff visit the vil - in the thin smear were examined to determine species lage of the index case to perform malaria testing among . Cross-check of the malaria positive slides for spe- household members and neighbours. Based on WHO cies identifications were done by the district level micros - and national guidelines, RACD is performed using copist. A provincial expert level-certified microscopist microscopy and all subjects residing within 500 m of the did quality assurance (QA) for all positive slides, 10% of index case are targeted. randomly selected negatives and to resolve any discrep- ancies in results between the clinic and district-level Study population microscopists . Index cases were enrolled from four sub-district level All DBS samples collected were dried overnight then health facility sites that reported 78% of all reported cases stored in sealed plastic bags with desiccant. Extraction in Aceh Besar in 2013: Indrapuri, Kuta Cot Glie, Lhoong, of DNA and LAMP testing on all samples were per- and Saree. An additional health facility Lhoknga was also formed at the Aceh Provincial Health Laboratory. DNA included purposively due to high case burden prior to was extracted from DBS using the Saponin/Chelex-100 2013. RACD was conducted in the villages where index method . Using 15 µL of chelex-extracted DNA cases resided, and the people reached during RACD were solution, Pan-LAMP testing followed by Pf-LAMP spe- enrolled for this study. The study was conducted over a cific testing for Pan-LAMP positive samples was per - period of 20 months from May 2014 to December 2015. formed using a commercial Loopamp detection kit [17, 18] in accordance to manufacturer’s instructions (Eiken Health facility and field procedures Chemical, Japan) (Limit of detection (LOD) between 1 All subjects presenting to health facilities with suspected and 5 parasites/µL). As previously detailed, using DNA malaria were assessed for malaria infection by micros- chelex-extracted from a second DBS, Pan-LAMP posi- copy of blood smears. Microscopy confirmed cases, also tive and 10% of Pan-LAMP negative samples underwent referred to as an index case, had a subsequent venous further molecular testing for QA at the Eijkman Institute blood draw to generate a dried blood spot (DBS) for for Molecular Biology using nested PCR methods (LOD LAMP and PCR. Within one to 7 days of the index case between 0.1 and 1 parasites/µL) targeting the cytochrome report, RACD was conducted among individuals residing b gene followed by AluI enzyme digestion for species in households located within 500 m of the index cases. identification and 18S rDNA nested PCR, with positive During an RACD event, blood was collected from indi- samples undergoing further P. knowlesi-specific PCR viduals by finger prick to prepare slide blood smears and testing . DBS. The RACD team consisted of five people: a micros - copist and surveillance officer from the health facility, Costing data collection two community health workers from the village (a mid- Costs were collected for RACD programme set-up, train- wife plus a village malaria worker or village leader), and ing, coordination, and all activities beginning from when the study field coordinator. Up to two return visits were an index case was reported from a health facility (e.g. conducted to include any missing residents. The coverage preparing supplies and contacting the index case house- goal of RACD in each of the target area was to recruit a hold) through to sample processing, analysis, and quality minimum of 40 subjects or at least 80% of the residents assurance conducted after the RACD event. Time spent within each of the closest five households. Additional by personnel on microscopy or LAMP, including the details of index case enrolment and RACD have been quality assurance by subsequent microscopy and PCR, previously described . respectively, was allocated to diagnostic-specific costs as appropriate. Follow-up costs including a return visit to Laboratory testing provide treatment to subjects that were initially micros- The microscopists collected blood smears in the field and copy-negative but subsequently found to be LAMP-posi- transported them in closed slide boxes for subsequent tive were not included. examination at the health facility. Slides were fixed and Data on costs were collected by a local field study stained with 3% Giemsa. A slide was determined malaria coordinator through review of expenditure records. All positive if at least one clear form of any malaria parasite costs were converted from the local currency to US dol- species was found after examination of the whole spread lars (USD) using the average mid-year exchange rate for of the thick smear. Parasite densities were measured by 2015 (1 USD = 13,432 Indonesian Rupiah, from Oanda. counting the number of asexual parasites per 200 or 500 com). These data were inputted to a costing tool devel - white blood cells (WBC) and calculating parasites/µL oped in Microsoft Excel 2010. Costs were organized by assuming a WBC count of 8000 parasites/µL in the thick location at which the expenses occur (health facility, Zelman et al. Malar J (2018) 17:220 Page 4 of 12 provincial laboratory, and national laboratory); the type LAMP kits, laboratory reagents, centrifuge tubes, etc. Per of diagnostic test used (microscopy or LAMP); and the unit costs of these consumables were multiplied by the month in which the cost was incurred. Each cost input amounts utilized during the study period to estimate the was grouped into one of five major categories. total cost of consumables. Personnel Analysis Personnel costs covered project staff salaries for train - The main outcome measure was number of infections ings, coordination, field activities and laboratory work. detected by LAMP versus microscopy in RACD. Other Project staff included one full time field coordinator, measures included number of RACD events and number a part time field team in each of the five sub-districts of individuals screened per RACD event. including one surveillance officer and a microscopist, To estimate the costs associated with microscopy ver- three part-time laboratory technicians at the provincial sus LAMP, RACD costs were separated into ‘general and national levels, and one laboratory technician at RACD costs’ and ‘diagnostic-specific costs’. General the national level. Most of the personnel had duties out- RACD costs covered routine activities related to RACD side the scope of this study, and only supported RACD such as the field visit, including time for preparation, as index cases presented to health facilities. Thus, spe - travel, and screening households. Diagnostic specific cific time contributions of all personnel involved in the costs for microscopy and LAMP were calculated by iso- RACD activities were logged and later used for approxi- lating any personnel, training, consumables, processing mating the cost of human resources. Time required for time, services from general RACD costs. As a key activity initial diagnosis of index cases at health facilities was not for an elimination surveillance programme, particularly included. for programmes considering the use of highly sensitive diagnostics, QA was included for both microscopy- and Trainings LAMP-specific costs . Trainings for microscopy and Costs associated with any trainings conducted for this microscopy QA were conducted in conjunction with the study were captured and include costs for training sup- surveillance trainings and included in the general RACD plies, room rental, and participant per diems. training costs. Cost proportions for each input category were compared for microscopy-specific, LAMP-specific Services and general RACD costs. The cost of services included utilities such as internet, Cost per RACD event and cost per individual screened communication, courier services, and vehicle rental/fuel were calculated by dividing the total cost of conduct- for sample transportation used in conducting RACD. ing RACD using microscopy and/or LAMP during the Services utilized prior to the study period but for project study period by the total number of RACD events and setup were included. screened individuals. Cost per population at risk (PAR; defined as total population in each sub-district) was cal - Capital culated at the sub-district level. Cost per PAR per year Capital costs were initial investments to set up RACD was calculated by dividing the total cost of RACD using and included the cost of owned motor vehicles used for both diagnostics by the population and by the number travel and transportation; electronic devices such as com- of months the study was conducted (20 months). The puters, computer accessories, printers, tablets, refrig- resulting montly cost per PAR, was then multiplied by erators; and laboratory equipment such as microscopes, 12 (months in a year) to determine the cost per PAR per centrifuges, heat blocks, and PCR machines. If the capi- year. Costs per infection identified were calculated as tal was used for activities other than RACD, a time use cost of RACD using microscopy and/or LAMP, divided percentage was assigned. The cost of capital was valued by the number of infections detected by microscopy and/ for the study duration (20 months) after accounting for or LAMP, respectively. The total cost for the study period depreciation based on the useful life years of each capital, and annual recurrent programmatic cost of the RACD and discounted using a rate of 3%. The remaining values programme were also calculated. The annual recurrent of capital at the end of the study period were subtracted. programmatic cost was approximated by excluding any The cost of office space and furniture were not included capital non-recurrent costs from the total cost. due to limitations in data availability. To compare the cost-effectiveness of the different diag - nostic methods, the incremental cost-effectiveness ratio Consumables (ICER), defined by the difference in cost between two Consumables included field and laboratory supplies such interventions divided by the difference in their effect, as slides, lancets, alcohol swabs, plastic bags, filter paper, was estimated by comparing the cost and outcomes Zelman et al. Malar J (2018) 17:220 Page 5 of 12 associated with LAMP and microscopy. As general enrolled in the study. On average, about 42 individuals RACD costs were applied to both methods, the differ - were screened or tested per RACD event. Of these indi- ences in costs were due to diagnostic-specific costs and viduals tested during RACD, a total of eight additional respective yield. infections were identified. Three of these additional cases To explore how costs might vary across sites with dif- were initially found to be microscopy positive, but one ferent index case burdens or prevalence of infection in was determined to be a false positive as confirmed by RACD , cost measures were compared across the LAMP. An additional five individuals, originally found sub-districts. The cost drivers of RACD across the study to be microscopy negative, were positive when tested sites were also assessed. To consider how changes or dif- by LAMP. Thus, seven additional infections found via ferences in diagnostic sensitivity or prevalence of infec- RACD were confirmed LAMP positives. When the seven tion in RACD might affect findings, the costs and ICER LAMP positives underwent PCR for quality assurance, for different scenarios was estimated based on the costs one was negative. In total, six PCR-confirmed cases were and outcomes observed in this study. identified, with three each from Saree and Lhoong study sites, and were classified as three P. vivax, two P. falcipa - Results rum, and one P. knowlesi. The remaining three sites (Kuta Enrollment and outcome measures Cot Glie, Indrapuri, and Lhoknga) identified no addi - Study enrollment and laboratory reports have previously tional infections as a result of RACD. A summary of the been reported , but are shown in Fig. 1 for reference. outcome measures (number of RACD events, subjects In brief, a total of 36 eligible positive index cases were screened, and infections identified) for the overall study identified in the health facilities triggering 36 RACD and by sub-district is shown in Table 1. events. Of the 1638 eligible individuals residing in the screening radius of the index cases, 1495 (91%) were Fig. 1 Enrollment in reactive case detection (RACD) with laboratory results Zelman et al. Malar J (2018) 17:220 Page 6 of 12 Table 1 Summary of outcome measures, total costs, and cost effectiveness of RACD using microscopy and LAMP by sub- district over study period Sub‑ district Population Outcome measures Total cost Average costs at risk No. No. No. infections Total Per PAR Per RACD Per Per of RACD of subjects identified (%) per year event individual infection events screened screened identified Lhoong 9592 14 511 3 (0.6%) $12,703 $0.79 $907 $25 $4234 Saree 11,346 12 527 3 (0.6%) $12,748 $0.67 $1062 $24 $4249 Kuta Cot Glie 10,560 8 377 0 (0%) $9657 $0.55 $1207 $26 – Indrapuri 14,052 2 80 0 (0%) $4391 $0.19 $2196 $55 – Lhoknga 15,659 0 0 0 (0%) $2918 $0.11 – – – Total 61,209 36 1495 6 (0.4%) $42,418 $0.42 $1178 $28 $7070 Population at risk is defined as total population of sub-district LAMP loop-mediated isothermal amplification, RACD reactive case detection Sub-district average cost per PAR Overall cost of RACD using microscopy and LAMP $0.62 per individual screened (Table 2). LAMP-specific The total cost of conducting RACD using microscopy and costs, which includes LAMP-specific capital and con - LAMP during the study period was $42,418 (Table 1). sumables (kits, DBS, reagents, and laboratory equipment On average, the cost per RACD event was estimated to for QA) as well as trainings and personnel time to pro- be $1178 and the cost per individual screened was $28. cess and run the blood samples for LAMP with quality The average cost per infection identified was $7070. As a assurance, totaled $24,557 over the course of the study. public health intervention, the annual cost per PAR was This is equivalent to $16 per individual screened. Con - estimated at $0.42 averaged across the sub-districts. sumables, followed by training and personnel costs were the main cost drivers for LAMP (Fig. 2). Compared to a Cost and cost‑effectiveness by microscopy versus LAMP box of 100 microscopy slides, which were purchased for Microscopy-specific costs, which consists of consuma - $4.50 (equivalent to $0.05 per test), Pan-LAMP tests cost bles, such as slides and Giemsa, personnel time used to $3.75 each and Pf-LAMP tests cost $11.97 each. Reagents process and read the slides, microscopes, and staff train - used for PCR for LAMP QA also constitute to a major ing, was calculated at $929 in total, which translated to share of consumables. General RACD costs, which could Table 2 Summary of costs by outcome measure General Microscopy‑ LAMP‑ Overall Overall Overall Annual Annual Annual recur‑ RACD specific specific RACD RACD RACD recurrent recurrent rent LAMP‑ costs costs costs (microscopy (microscopy (LAMP only) general microscopy‑ specific costs and LAMP) only) RACD costs specific costs Total costs $16,931 $929 $24,557 $42,418 $17,860 $41,489 $6957 $527 $12,871 (36 RACD events) Average cost $470 $26 $682 $1178 $496 $1152 $193 $14.65 $358 per RACD event Average $11 $0.62 $16 $28 $12 $28 $5 $0.35 $9 cost per individual screened Average $2822 $465 $4093 $7070 $8930 $6915 $1159 $264 $2145 cost per infection identified RACD reactive case detection, LAMP loop-mediated isothermal amplification Zelman et al. Malar J (2018) 17:220 Page 7 of 12 Fig. 2 Cost proportions by input category for general RACD, microscopy-specific, and LAMP-specific activities over study period not be attributed to either diagnostic specifically, totaled breakdown of all capital costs can be found in the annex $16,931, or $11 per individual screened. Table 2 also pre- (Additional file 1: Figure S1). Of all capital costs, 61% was sents costs for overall RACD costs with microscopy and from laboratory equipment mainly for LAMP and LAMP LAMP, RACD with microscopy only, and RACD with QA followed by tablets and accessories (33%). Eight of LAMP only over the course of the study period, as well the ten highest capital costs, ranging from $82 to $1542, as annualized recurrent costs for RACD, microscopy, and were attributed to LAMP or LAMP QA and included LAMP specific costs. The annual recurrent costs, or costs laboratory equipment such as a heat block, Gel doc sys- which exclude one-time purchase capital costs, are indic- tem, a thermomixer, a UV sterilization cabinet, a refriger- ative of the costs that the programme can expect for each ator, a centrifuge, and a PCR machine, though all of these year that the intervention is implemented. For example, a items except the heat block and pipette were only used programme that has all the necessary capital equipment for RACD between 15 and 75% of the time. Usage of the for LAMP, such as a heat block and PCR machine, can gel doc system, thermomixer, UV Sterilization cabinet, expect to budget $9 per person screened by LAMP, in and PCR machines are not crucial for LAMP detection addition to the annual recurrent cost of RACD. itself, but are important for LAMP QA and represent a Cost shares by input category were compared between large share of the capital cost. Two of the top ten high the diagnostic methods (LAMP and microscopy), as value capital were attributed to general RACD costs and well as to general RACD costs (Fig. 2). Personnel costs included seven tablets used for data collection and study accounted for the major share for general RACD and laptops. The remaining capital items used for RACD had microscopy specific costs, at 41 and 51%, respectively. attributed costs ranging from $1.75 to $74. Additional Consumables accounted for the largest share of LAMP file 1: Tables S1, S2 provide a more detailed breakdown of specific costs at 51%, and second largest share for micros - input costs by location and cost category. copy at 44%. Training cost was the second largest as a Two infections were identified among all individuals share of cost for LAMP and general RACD at 20% each. tested with microscopy. The resulting cost per additional Four trainings were held between September 2013 and infection identified by RACD using microscopy was esti - January 2015. Of the four, one training was LAMP-spe- mated to be $8930, where the microscopy-specific cost cific laboratory training. The remaining three focused accounted for only about 5% (or $465 per infection). Six on general RACD protocols, and included refreshers on infections were identified via LAMP through RACD at blood collection techniques for microscopy and LAMP. a cost of $6915 each (LAMP-specific cost of $4093 per Personnel time and training were also the main cost driv- infection). The ICER of detecting an infection through ers for RACD across each study site. LAMP compared to microscopy was estimated to be The share of capital costs for LAMP-specific and gen - $5907 per infection. eral RACD activities was at 13% each. A more detailed Zelman et al. Malar J (2018) 17:220 Page 8 of 12 Costs by sub‑district Eec ff t of infection prevalence on costs Of the five cost categories, LAMP-specific consumables and cost‑effectiveness were the largest cost driver at 34% of the total expense For settings with lower or higher prevalence of infec- for all sub-districts. Personnel and trainings accounted tion compared to Aceh Besar, the costs using LAMP or for the next largest shares of the expense at an average of another equivalent test with 3:1 diagnostic yield relative 27 and 20%, respectively. Capital and services accounted to microscopy were estimated (Fig. 3, solid line). At 0.4% for an average of 13 and 7%, respectively, across the sub- prevalence of infection, the cost per infection identified districts. As Lhoknga had no index cases, no services or is $7070, and declines to $1767 when the prevalence consumables costs were incurred, which resulted in a is 1.6%. Cost declines begin to plateau thereafter. The larger proportional share of training and capital costs (57 impact of prevalence of infection on ICER was also esti- and 37%, respectively). mated (Fig. 3, dotted line). As the calculation for ICER The distribution of total costs for RACD using micros - is based on cost per infection, ICER also declines with copy and LAMP varied across the study sites (Table 1). increasing prevalence of infection. At 0.4% prevalence of In Lhoong and Saree, which had the highest number of infection, ICER was $5907, and declines to $1477 when RACD events and individuals screened, the total cost was the prevalence of infection is 1.6%. The ICER decline the highest of the five study sites at $12,703 and $12,748 slows and begins to plateau thereafter. respectively, and the cost per individual screened was between $25 and $24 per person screened each. With Eec ff t of diagnostic yield on costs and cost‑effectiveness eight RACD events in Kuta Cot Glie, total cost was lower Although LAMP detected threefold more infections than than Saree and Lhoong, with cost per individual screened microscopy, our sample size was small. Other diagnos- being similar ($26). With two RACD events, Indrapuri tics, existing or in development, may provide a different had the highest cost per individual screened at $55. yield relative to microscopy . Assuming prevalence Lhoknga had no RACD events and thus no cost per indi- of microscopy-detectable infection to be 0.13% and base vidual screened, but the site maintained a cost of $4391, costs similar to LAMP, the impact of diagnostic yield on primarily due to personnel (including training) and a costs per infection identified and ICER was estimated small amount of capital requirements to be prepared (Fig. 4). When the yield is three fold relative to micros- to launch RACD if needed. The cost per PAR per year copy, the cost per infection identified is $7070 (as in our by sub-district ranged from $0.11 in Lhoknga to $.79 in study). When the yield increases to fivefold, the costs Lhoong, and averaged $0.42 across the total population decreases to $4242, with the curve nearing a plateau of the five sub-districts (Table 1). when the yield reaches tenfold, where costs are $2121 per infection identified. The ICER follows the same trend. At Fig. 3 Estimated costs per infection identified in RACD and incremental cost-effectiveness ratio (ICER) of infections identified in using LAMP versus microscopy, by prevalence of LAMP-detectable infection in RACD, assuming same general and per unit costs and 33.3% sensitivity of microscopy compared to LAMP as observed in study Zelman et al. Malar J (2018) 17:220 Page 9 of 12 Fig. 4 Estimated costs per infection identified in RACD and incremental cost-effectiveness ratio (ICER) by yield of diagnostic test to detect infec- tions relative to microscopy, assuming same general and per unit costs as LAMP and same 0.13% prevalence of microscopy-detectable infections as observed in study a yield of three fold relative to microscopy, the ICER is additional malaria infection detected. As malaria pro- $5907 and it continues to reduce to $2954 when the yield grammes consider active case detection and the integra- is fivefold and $1313 when the yield is tenfold relative to tion of highly sensitive diagnostics, these findings may microscopy. The same trends would apply to a situation inform strategic and budgetary planning. with a higher proportion of low-density infection (e.g. as Due to the intensified interventions required, it is well- malaria transmission declines). established that malaria elimination compared to con- trol is more costly but that longer term benefits make Discussion the pursuit worthwhile [21, 22]. A systematic review on RACD is a strategy widely implemented to gather surveil- the costs and cost-effectiveness of control interventions lance information in low transmission settings and it may by White et al.  found the median cost of preventa- contribute to reductions or interruption of transmission. tive interventions (insecticide-treated bed nets, indoor However, the costs of doing RACD have not been well residual spraying, intermittent preventative treatment) evaluated. This study investigated the cost of conducting was $0.60–$6.70 per person and the median costs for RACD and the cost-effectiveness of different diagnos - case management services (diagnosis, treatment) ranged tic methods in identifying additional malaria infections. from $4.32–$30.26 per patient, with the most expensive During the study period of 20 months, 1495 individuals service being treatment of severe malaria. Our finding were screened in 36 RACD events yielding six additional that RACD with microscopy and/or LAMP cost $12– cases by microscopy or LAMP, at an estimated total $28 per individual screened is consistent with previous cost of $42,418. The average cost of conducting RACD reports on the higher cost for elimination versus control using microscopy and LAMP was $1178 per event, with interventions. The high cost of $7070 per infection iden - the main cost driver being personnel. The average cost tified also raises the issue of the value of RACD as a low of RACD for each individual was $11 for general costs, yield activity where case detection rates are generally less with an additional $0.62 and $16 per person for micros- than 2% even with highly sensitive diagnostics . How- copy and LAMP, respectively. As a public health inter- ever, the value of an intervention should be considered vention, RACD using both diagnostics cost an average of in a broader context. As a public health intervention, the $0.42 per PAR per year. LAMP was more costly but more annual cost of RACD using both diagnostics was low, at cost-effective for the detection of infections mainly due an average of $0.42 per PAR at the sub-district level. to higher diagnostic sensitivity. For RACD using micros- Despite the high costs, RACD using LAMP versus copy only, the cost per infection found was $8930 com- microscopy led to a lower cost per infection identified pared to $6915 for RACD using LAMP only. The ICER (23%) and the ICER of $5907 per additional infection for RACD using LAMP versus microscopy was $5907 per detected favors the use of LAMP or a LAMP-equivalent Zelman et al. Malar J (2018) 17:220 Page 10 of 12 assay in RACD. Based on our analyses of cost inputs, sensitive histidine-rich protein 2-based rapid diagnostic there would opportunities to mitigate costs. Person- test  could be useful in P. falciparum predominant nel and LAMP-specific consumables, and not capi - settings and has lower costs due to a discounted price tal, training, or services, accounted for the highest cost and does not require laboratory work nor travel back to shares . Personnel costs could be minimized by bet- the households to inform on results). In a more remote ter integrating staff in the health system. LAMP-specific and underserved setting, there may be higher costs asso- consumable costs could be decreased if prices were nego- ciated with establishing laboratory infrastructure to con- tiated or subsidized, or a less expensive highly sensitive duct LAMP. A new diagnostic may have higher costs than diagnostic could be used. In addition, the higher sensitiv- those required for LAMP. These scenarios would increase ity of LAMP made it possible to detect P. knowlesi in the costs per infection identified and reduce cost-effective - study area for the first time. This important benefit may ness. Also, the estimations assume very low parasite den- outweigh the high costs of RACD using LAMP. sities and do not take into account that in other settings, To explore how costs might vary across sites with dif- due to endemicity or species (e.g. P. falciparum presents ferent index case burdens or prevalence of infection in with higher parasite densities than P. vivax), parasite den- RACD, cost measures were compared across the sub-dis- sities may be different and thus impact the sensitivities tricts. In sub-districts with more RACD events (such as and yield of microscopy and LAMP. More empiric data Saree and Lhoong) overall costs were higher than in sub- from other settings along with use of more sophisticated districts with fewer cases, and consequently the cost per modeling techniques could improve generalizability. individual screened was lower. Despite having zero cases Additionally, while microscopy and LAMP both require during the study period, Lhoknga still needed to main- time to acquire results, processing time for LAMP may tain costs for training “on-call” local health facility per- be longer resulting in a delay in treatment and therefore sonnel as well as capital in the event an index case were transmission-blocking, which were not taken into con- to present and RACD would need to be conducted. sideration in this analysis. Finally, this study only assessed To further inform how costs and cost-effectiveness costs from service provider’s perspective, including cost might vary with differences in prevalence of infection or sharing and in-kind donations, and costs are specific to the sensitivity of the diagnostic used, estimated projec- this situation. In other contexts, absolute costs may be tions show how costs per infection identified along with different. Future studies can build upon these findings ICER decrease with increases in prevalence of infec- to quantify costs and benefits of conducting RACD by tion and yield of a diagnostic to detect infections, rela- including the broader costs to society when additional tive to microscopy (Figs. 3 and 4). As seen in the Fig. 3, infections are not detected and may result in further the curve began to plateau when the test positive rate transmission. or prevalence of infection in RACD increased above This study had several strengths. Firstly, a detailed and 1.6% suggesting that the highest relative cost-savings prospective collection of RACD costs was carried out. will be realized in low transmission settings. The curve Retrospective data collection can introduce bias and also began to plateau when the yield of the diagnostic limit the granularity of the costing data. Importantly, this test exceed fivefold–tenfold relative to microscopy, sug - study fills a critical gap on the economics of focal screen gesting diminishing returns with use of more sensitive and treat. Previous studies have documented a frame- diagnostics  when the prevalence is already low. As work for evaluating the costs of RACD , conducted more sensitive diagnostics are being developed or come cost analyses of mass screen and treat , or identi- to market, these considerations can help to inform deci- fied potentials for operational efficiencies , but only sion-making on investments and strategic planning. one other study has measured the actual costs and that There were some limitations of this study. Estimation study was from a higher transmission setting where the of the personnel costs relied on self-reported time allo- RACD test positivity rate was 19% . To the best of the cations. Further, due to the few cases in this low trans- authors’ knowledge, this study is the first on RACD or mission setting, limitations exist for the generalizability focal screening and treatment to report costs from a low of the study and precision of the ICER estimate. Extrapo- transmission setting, costs using a highly sensitive diag- lations to consider the influence of prevalence of infec - nostic, and cost-effectiveness of using a highly sensitive tion and diagnostic yield were performed, but assumed versus standard diagnostic in active case detection. other fixed factors (epidemiological, or cost-related). In In summary, in the low transmission setting of Aceh real world implementation, several relevant factors could Besar, RACD costs per individual screened were found change and affect costs. For example, cost-effectiveness to be high, though costs per PAR were low. Compared to of LAMP could be improved with discounted prices for microscopy, the use of LAMP in RACD was more costly LAMP or other less costly diagnostics (e.g. a new highly but more cost-effective for the detection of infections, Zelman et al. Malar J (2018) 17:220 Page 11 of 12 References with diminishing returns observed when findings were 1. WHO. A framework for malaria elimination. Geneva: World Health Organi- extrapolated to scenarios with higher prevalence, or zation; 2017. using more sensitive diagnostics when prevalence is very 2. WHO. Disease surveillance for malaria control: an operational manual. Geneva: World Health Organization; 2012. low. These findings can inform strategic and budgetary 3. Moonen B, Cohen JM, Snow RW, Slutsker L, Drakeley C, Smith DL, et al. decisions faced by the many countries that are pursuing Operational strategies to achieve and maintain malaria elimination. malaria elimination and considering active case detection Lancet. 2010;376:1592–603. 4. Galappaththy GN, Gueye CS, Sanders K, Rundi C, Vestergaard L, Cotter C, with new diagnostics [1, 30, 31]. et al. Reactive surveillance methods used for malaria elimination in Asia and the Pacific: results from a 12 country survey. Malar J. 2012;11(Suppl Additional file 1):P90. 5. Moonen B, Cohen JM, Tatem AJ, Cohen J, Hay SI, Sabot O, Smith DL. A framework for assessing the feasibility of malaria elimination. Malar J. Additional file 1: Figure S1. Cost proportion breakdown of capital costs. 2010;9:322. Table S1. Detailed summary of input costs (USD) by location. Table S2. 6. Sturrock HJ, Hsiang MS, Cohen JM, Smith DL, Greenhouse B, Bousema Detailed summary of input costs (USD) by cost category. Table S3. T, et al. Targeting asymptomatic malaria infections: active surveillance in Detailed list of top ten capital costs, discounted annually at 3% from year control and elimination. PLoS Med. 2013;10:e1001467. of purchase (USD). 7. Okell LC, Ghani AC, Lyons E, Drakeley CJ. Submicroscopic infection in Plasmodium falciparum-endemic populations: a systematic review and meta-analysis. J Infect Dis. 2009;200:1509–17. Authors’ contributions 8. Wu L, van den Hoogen LL, Slater H, Walker PG, Ghani AC, Drakeley CJ, MSH, RB, and BZ conceived the paper and prepared the manuscript. BZ coor- et al. Comparison of diagnostics for the detection of asymptomatic dinated inputs to the draft. RB and BZ conceived the design and conducted Plasmodium falciparum infections to inform control and elimination the analysis. FC and CC contributed to study design. FC managed sample strategies. Nature. 2015;528:S86–93. analysis. FC and IZ managed the local data collection. All authors read and 9. Hsiang MS, Greenhouse B, Rosenthal PJ, et al. Reply to Goyal et al. J Infect approved the final manuscript. Dis. 2015;211:1687. 10. Britton S, Cheng Q, McCarthy JS. Novel molecular diagnostic tools Author details for malaria elimination: a review of options from the point of view of Malaria Elimination Initiative, Global Health Group, University of California, high-throughput and applicability in resource limited settings. Malar J. San Francisco (UCSF), San Francisco, USA. Malaria Pathogenesis Unit, Eijkman 2016;15:88. Institute for Molecular Biology, Jakarta, Indonesia. Department of Medicine, 11. Herdiana H, Cotter C, Coutrier FN, Zarlinda I, Zelman BW, Tirta YK, et al. UCSF, San Francisco, USA. Paritrana Asia Foundation, Jakarta, Indonesia. Malaria risk factor assessment using active and passive surveillance data United Nations Children’s Fund (UNICEF), Aceh Field Office, Banda Aceh, from Aceh Besar, Indonesia, a low endemic, malaria elimination setting Indonesia. Department of Pediatrics, University of Texas Southwestern Medi- with Plasmodium knowlesi, Plasmodium vivax, and Plasmodium falciparum. cal Center, Dallas, USA. Department of Pediatrics, UCSF, San Francisco, USA. Malar J. 2016;15:468. 12. Monthly malaria report and findings: January–December 2014–2015. Acknowledgements Aceh Besar, Aceh Province, Indonesia: Aceh Besar District Health Office The authors extend their gratitude to the Indonesian Sub-directorate for (DHO); 2015. Malaria, particularly Dr. Maria Endang and Dr. Ferdinand Laihad, for their 13. National malaria control program strategic plan 2015–2019. Indonesia: contributions. Ministry of Health Republic of Indonesia; 2014. 14. Ekawati LL, Herdiana H, Sumiwi ME, Barussanah C, Ainun C, Sabri S, et al. Competing interests A comprehensive assessment of the malaria microscopy system of Aceh, The authors declare that they have no competing interests. Indonesia, in preparation for malaria elimination. Malar J. 2015;14:240. 15. WHO. Basic malaria microscopy. Part 1: learner’s guide. Geneva: World Availability of data and materials Health Organization; 2010. Data used for this manuscript is available upon request. 16. Plowe CV, Djimde A, Bouare M, Doumbo O, Wellems TE. Pyrimethamine and proguanil resistance-conferring mutations in Plasmodium falciparum Consent for publication dihydrofolate reductase: polymerase chain reaction methods for surveil- All authors have read and approved the final manuscript. lance in Africa. Am J Trop Med Hyg. 1995;52:565–8. 17. Aydin-Schmidt B, Xu W, Gonzalez IJ, Polley SD, Bell D, Shakely D, et al. Ethics approval and consent to participate Loop-mediated isothermal amplification (LAMP) accurately detects Not applicable. malaria DNA from filter paper blood samples of low density parasitae - mias. PLoS ONE. 2014;9:e103905. Funding 18. Hopkins H, Gonzalez IJ, Polley SD, Angutoko P, Ategeka J, Asiimwe C. Funding for this work comes from a Grant from the Bill & Melinda Gates Foun- Highly sensitive detection of malaria parasitemia in a malaria-endemic dation [A121292]. Funding was also provided by National Institute of Allergy setting: performance of a new loop-mediated isothermal amplification and Infectious Diseases at the National Institutes of Health [AI101012 to kit in a remote clinic in Uganda. J Infect Dis. 2013;208:645–52. M.S.H.] and the Horchow Family Fund at the University of Texas, Southwestern 19. Muriuki D, Hahn S, Hexom B, Allan R. Cross-sectional survey of malaria [5300375400 to M.S.H.]. The funders had no role in the design or writing of this prevalence in tsunami-affected districts of Aceh Province, Indonesia. Int J paper or the decision to submit the paper for publication. The authors were Emerg Med. 2012;5:11. not paid to write this manuscript. 20. Daily J. Malaria diagnostics technology and market landscape. 3rd ed. Geneva: UNITAID; 2016. 21. Sabot O, Cohen J, Hsiang M, Kahn J, Basu S, Tang L, et al. Malaria elimina- Publisher’s Note tion 4: costs and financial feasibility of malaria elimination. Lancet. Springer Nature remains neutral with regard to jurisdictional claims in pub- 2010;376:1604–15. lished maps and institutional affiliations. 22. Shretta R, Avanceña ALV, Hatefi A. The economics of malaria control and elimination: a systematic review. Malar J. 2016;15:593. Received: 18 October 2017 Accepted: 22 May 2018 23. White MT, Conteh L, Cibulskis R, Ghani AC. Costs and cost-effectiveness of malaria control interventions—a systematic review. Malar J. 2011;10:337. Zelman et al. Malar J (2018) 17:220 Page 12 of 12 24. van Eijk AM, Ramanathapuram L, Sutton PL, Kanagaraj D, Sri Lakshmi Priya 28. Silumbe K, Yukich JO, Hamainza B, Bennett A, Earle D, Kamuliwo M, et al. G, Ravishankaran S, et al. What is the value of reactive case detection in Costs and cost-effectiveness of a large-scale mass testing and treat - malaria control? A case-study in India and a systematic review. Malar J. ment intervention for malaria in Southern Province, Zambia. Malar J. 2016;15:67. 2015;14:211. 25. Larson BA, Ngoma T, Silumbe K, Rutagwera M-RI, Hamainza B, Winters 29. Searle KM, Hamapumbu H, Lubinda J, Shields TM, Pinchoff J, Kobayashi T, AM, et al. A framework for evaluating the costs of malaria elimination et al. Evaluation of the operational challenges in implementing reactive interventions: an application to reactive case detection in Southern screen-and-treat and implications of reactive case detection strategies for Province of Zambia, 2014. Malar J. 2016;15:408. malaria elimination in a region of low transmission in southern Zambia. 26. Imwong M, Hanchana S, Malleret B, Renia L, Day NP, Dondorp A, et al. Malar J. 2016;15:412. High-throughput ultrasensitive molecular techniques for quantifying 30. Newby G, Bennett A, Larson E, Cotter C, Shretta R, Phillips AA, et al. The low-density malaria parasitemias. J Clin Microbiol. 2014;52:3303–9. path to eradication: a progress report on the malaria-eliminating coun- 27. Das S, Jang IK, Barney B, Peck Rek JC, Arinaitwe E, et al. Performance of a tries. Lancet. 2016;387:1775–84. high-sensitivity rapid diagnostic test for Plasmodium falciparum malaria in 31. UNITAID. Malaria diagnostics technology and market landscape. 3rd ed. asymptomatic individuals from Uganda and Myanmar and naïve human Geneva: World Health Organization; 2016. challenge infections. Am J Trop Med Hyg. 2017;97:1540–50. Ready to submit your research ? Choose BMC and benefit from: fast, convenient online submission thorough peer review by experienced researchers in your ﬁeld rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions
Malaria Journal – Springer Journals
Published: Jun 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera