Cost analyses for malaria molecular diagnosis for research planners in India and beyond

ABSTRACT

Background

Malaria elimination mandates early and accurate diagnosis of infection. Although malaria diagnosis is programmatically dependent on microscopy/RDTs, molecular diagnosis has much better diagnostic accuracy. Higher cost of molecular diagnoses is a recognized challenge for use at the point of care. Because funding is always a recognized constraint, we performed financial cost-analyses of available molecular platforms for better utilization of available budget.

Methods

Two strategies were applied to deduce the cost per sample. Strategy 1 included recurring components (RC) in minimum pack size, and biologist’s time whereas strategy 2 included only RC and non-recurring components and costs are calculated for sample sizes (1–1,000,000) to infer the sample size effect.

Results

Spin column-based manual DNA extraction (US$ 3.93 per sample) is the lowest-cost method, followed by magnetic bead-based automated, semi-automated, and PCI-based manual method. Further, DNA extraction cost per sample via spin column-based manual method and semi-automated method decreases with an increase in sample size up to 10,000. Real-time PCRs are ~ 2-fold more economical than conventional PCR, regardless of sample size.

Conclusions

This study is the first for malaria to estimate systematic molecular diagnosis financial costs. Kit-based and automated methods may replace conventional DNA extraction and amplification methods for a frugal high-throughput diagnosis.

KEYWORDS:

• Automation
• cost analysis
• malaria
• molecular diagnosis
• India
• non-recurring cost
• recurring cost

1. Introduction

Malaria is a prehistoric and life-threatening vector-borne disease and its elimination is a global health intent, which could be achieved by an optimal interplay of prevention, precise diagnosis, treatment and effective vector control. Integrated initiatives by the Indian government and public awareness lead to a decrease in reported malaria cases by 27% between 2000 and 2015, but < 2% between 2015 and 2019 [1]. In 2019 India reported 6 million cases with a decline rate of 17.6% since 2000. But still, with the present decline rate, the goal of malaria elimination till 2030 is far-fetched, because of roadblocks like emerging resistance against drugs as well as insecticides, inadequate investments and improper tracing of asymptomatic/low density infections [2,3]. Failing to detect malaria cases will also impart a financial burden and divert the policymakers from visualizing the actual elimination path. Hence, accurate malaria diagnosis plays a crucial role in the elimination program.

Depending on the diagnostic setting, Rapid Diagnostic Test (RDT), microscopy and Polymerase Chain Reaction (PCR) are widely used for malaria diagnosis for decades [4,5]: RDT and microscopy at the point of patient care and PCR in research labs in most of the malaria endemic countries, including India. RDT and microscopy are benchmark techniques in malaria diagnosis as these are convenient and economical but lack high diagnostic accuracy (lowest limits of detection are 10 and 100 parasites per µL of blood for microscopy and RDT, respectively) and hence are not very reliable, particularly for the detection of mixed Plasmodium species infection and low-density infections [5]. On the other hand, PCR is the most sensitive method (lowest limit of detection ranges from 0.01 to 1 parasite per µL of blood depending on the type of PCR) and provides high diagnostic accuracy [6]. For achieving malaria elimination, we should not overlook the low-density and asymptomatic infections, hence PCR is preferred worldwide for the detection of Plasmodium parasites. However, the use of PCR currently is restricted to malaria diagnosis in research labs due to the technicalities and a higher cost.

Further, malaria surveillance has largely been acknowledged as an effective intervention in an elimination setting and regional or nationwide malaria molecular surveillance, in particular, is a recommended critical component [7–11]. Although the endpoints in malaria molecular surveillance vary, including monitoring parasite drug resistance, population structure, genetic diversity, multiplicity of infections, parasite transmission, and so on, accurate molecular detection of the infecting Plasmodium species is the starting point for all molecular surveillance platforms.

An optimal regional, sub-national or national malaria molecular surveillance requires a large number of samples to be tested continuously for a longer period of time. Considering that the global investments in malaria are already below the estimates for achieving the target of malaria elimination[5], it is even more difficult for developing countries to handle these financial constraints and still attain the intended goals of elimination in time. Hence, we did a systematic financial cost analysis, from sample collection to molecular diagnosis of infecting Plasmodium species using a simple Microsoft Excel® spreadsheet. The analyses will enable the researchers and policymakers to take informed decisions for choosing the most cost-effective and budget-friendly methods in malaria molecular diagnosis and allocate their funds accordingly to plan and improve large-scale molecular surveillance systems.

2. Methods

The Standard Operating Procedures (SOP) developed and used by laboratory researchers at the Indian Council of Medical Research – National Institute of Malaria Research (ICMR-NIMR), New Delhi, were utilized to pen down each and every component of recurring and non-recurring resources and their corresponding costs required in executing malaria molecular diagnosis. Cost statement for components of various methods/techniques was made by the last purchase done by the institute and/or surveying and choosing lowest price of a standard brand on Government of India e-Marketplace (GeM; https://gem.gov.in/), visiting the supplier’s brochures, visiting respective websites, or requesting latest quotation. Price for the smallest pack size available was noted for cost analysis so that work could be started with a minimum budget and less number of samples. Buffer cost (10%) was also added to the total amount of recurring cost to involve losses of components during execution. We chose similar global standard brands for recurring components (RCs), instead of native-make economic brands, to ensure global comparability of cost per sample for all molecular methods.

Cost analysis was performed in US$ using two different strategies. In strategy 1, a list of RC, their respective price and biologist’s time required for molecular diagnosis of malaria was casted for cost analysis (Tables 1–3). Given considerable price disparity between different standard brands, RC was further classified into standard RC (branded higher costs) and economized RCs (branded but economic versions for larger sample sizes). Time required by any personnel (field assistant/biologist) was noted based on the authors’ experience. In strategy 2, the cost analysis with respect to time was performed from sample size 1 to 10,00,000 (Table 4). The minimum (initial) budget required for one sample is represented as total cost, which incurred the smallest pack size of RC and one set of non-recurring components (NRC). Further, seven different sample sizes, that is, 1, 10, 100, and up to 10,00,000 (by taking the sample size ten times higher from the preceding one) were used to calculate total cost with an appropriate pack size of RC and only one set of NRC. For example, for 1, 10, and 100 samples, the SYBR™ Green PCR master mix of 1 mL pack size was acclaimed. However, for 1000 and 10,000 samples, the master mix cost of 10 mL and 2 × 50 mL pack size was noted, respectively. Further, the effect of increasing sample size on RC, NRC, and TC was assessed using dividend factor, which is calculated by dividing the cost at a particular sample size by the cost at immediately preceding sample size. If the value of dividend factor on cost and multiplication factor on sample size is same, it implies that the cost is dependent on sample size, else factors other than sample size are involved in determining cost per sample.

Table 1. Cost (US$) analysis for direct recurring components of sample collection and preparation.

Heparinized blood			Dried Blood Spot (DBS)
			Standard* Economised*			Standard* Economised*
RC	Qty	Cost per 30 samples	RC	Qty	Cost per 30 samples	RC	Qty	Cost per 30 samples
Alcoholized cotton swab	30	0.90	Alcoholised cotton swab	30	0.90	Alcoholised cotton swab	30	0.90
Lancet	30	1.08	Lancet	30	1.08	Lancet	30	1.08
GETP^a	1	0.50	GETP^a	1	0.88	GETP^a	1	0.88
Field assistant^b	1 hr	0.90	Field assistant^b	1 hr	0.90	Field assistant^b	1.5 hr	1.36
Heparinized Capillary tube	30	0.43	QIAcard™ FTA™ Micro card^c	30	109	Whatman® Filter paper^d	30	17.3
MCT^e (RNase DNase free)	30	0.90	MCT^e (RNase DNase free)	30	0.90	MCT^e (autoclavable)	30	0.36
			Punch machine^f	1	0.04	Punch machine^f	1	0.04
						Autoclave bags Autoclave tapes	1 1”	0.31 0.04
Total cost per 30 sample 10% Loss		4.72 0.47	Total cost per 30 sample 10% Loss		114 11	Total cost per 30 sample 10% Loss		22.26 2.22
			Biologist^g	0.5 hr	0.96	Biologist^g	0.5 hr	0.96
Cost per sample		0.17	Cost per sample		4.19	Cost per sample		0.85

*RC was further classified into standard RC (branded higher costs) and economized RCs (branded but economic versions for larger sample sizes).

^aGETP stands for gloves (Nitrile, medium size) 1 pair, ethanol 70% (American chemical society (ACS) grade) 30 mL (10 mL for whole blood collection), tissue paper 10 sheets and pen (one cryoware black marker) used for 10,000 MCTs (micro centrifuge tubes) for labeling, GETP usage is calculated based on the author experience while executing an experiment; ^bField assistant salary is calculated considering 1 sample collection takes 2 minutes using formula: cost/n sample = (total salary of biologist (US$) x time required per sample (hr) x total no. of sample)/(working days in a month x working hour per day); ^c QIAcard™ FTA™ Micro card^c card with 1 sample areas/card, with maximum volume 125 μL blood/sample area; ^dWhatman® Filter paper (grade 3) circles with diameter 185 mm; ^eEither microcentrifuge tubes (MCTs) could be used in sample preparation (autoclave bags and tapes will be used only with autoclavable MCTs); ^fOne punch machine is used for 4500 punches (maximum) only (as per author experience); ^gBiologist (JRF-Junior Research Fellow) salary is calculated for 30 minutes considering 1 punch/minute using formula: (cost of punch machine (US$) x total punches done)/maximum number of punches possible. All prices listed above are inclusive of 5% GST (Goods and Service Taxes) as per rule by Department of Scientific and Industrial Research (DSIR), Government of India. RC = recurring component; Qty = quantity.

Table 2. Cost (US$) analysis for recurring components of DNA extraction.

*Phenol chloroform isoamyl alcohol-based method			Spin column-based manual method			Spin column-based semi-automated (QIAcube®) method			Magnetic bead-based automated method
RC	Qty per 30 samples	Cost per 30 samples	RC	Qty per 30 samples	Cost per 30 samples	RC	Qty per 12 samples	Cost per 12 samples	RC	Qty per 96 samples	Cost per 96 samples
Tris buffer	24 mg	0.004	Reagent kit^d	30 reaction	99.26	Reagent kit^d	12 reaction	38.50	Reagent kit^e	96 reaction	322.53
EDTA	29 mg	0.01	Ethanol	30 ml	0.87	Ethanol	12 ml	0.35
NaCl	117 mg	0.01	Pipette Tips 200 μL^b	5	0.42	MCT	12	0.36	DTT (1 M)	576 μl	1.95
RNAse SDS	3 mg 2250 μl	1.56 0.34	Pipette Tips 1000 μL^b	35	2.94	Conductive filter tips 1000 μL	24	1.65	70% Ethanol^f Reagent Reservoir	4.2 ml 3	0.12 1.63
Proteinase K^a	4.5 mg	67	MCT	30	0.90				Pipette Tips 200 μL^b	5	0.42
Centrifuge tube (50 mL) Phenol	3 68 ml	0.85 54	GETP^c	1	0.50	Conductive filter tips 200 μL	12	0.90	Integra voyager Pipette tips 300 μL	96	13.80
Chloroform	65 ml	7.09				Rotor adapter	12	1.80	GETP^c	1	0.50
Isoamyl alcohol	3 ml	0.31				Pipette tips 200 μL	5	0.42
Sodium acetate Ethanol Pipette Tips 1000 μL^b	73 mg 22 ml 10	0.01 0.64 0.84				Pipette tips 1000 μL GETP^c	13 1	1.09 0.50
Pipette Tips 10 μL^b	30	2.52
MCT	93	2.80
GETP^c	1	0.50
Total cost per 30 samples 10% loss		138 13.8	Total cost per 30 samples 10% loss		102 10.2	Total cost per 12 samples 10% loss		45.58 4.6	Total cost per 96 samples 10% loss		341 34
Biologist^g	4 hr	7.69	Biologist^g	3 hr	5.77	Biologist^g	2.5 hr	4.81	Biologist^g	2 hr	3.85
Cost per sample		5.32	Cost per sample		3.93	Cost per sample		4.58			3.94

^aProteinase K noted here is of (>30units/mg) 10 mg/mL; bPipette Tips mentioned here are racked, filter and of RNase and DNase free (autoclavable MCTs could be used, price should be inclusive of autoclave bags and tapes); cGETP stands for Gloves (Nitrile, medium size) 1 pair, Ethanol 70% (American chemical society (ACS) grade) 10 mL for cleaning work surface, Tissue paper 10 sheets and Pen (one cryoware black marker) used for 10,000 MCT’s (micro centrifuge tubes) for labeling; dQIAamp DNA mini kit of 50 reactions price used; eCost for reagent kit QIAcard™ DNA Blood 400 Kit H96 is mentioned; fWashing required for equipment FTA™360 (once in a month) via 70% ethanol is calculated by assuming that machine will be working 8 hr/day and 20 days working in a month; gBiologist (Junior Research Fellow) time dedicated to any procedure (for salary calculation) is computed as per author experience while executing above experiments using formula: cost/n sample = (total salary of biologist (US$) x time required per sample (hr) x total no. of sample)/(working days in a month x working hour per day), *Reagents used in PCI method are for 75–100 μl blood which is equivalent to 3 punches of 3 mm diameter DBS as sample. All above prices listed above are inclusive of 5% GST (Goods and Service Taxes) as per rule by Department of Scientific and Industrial Research (DSIR), Government of India. RC = recurring component; Qty = quantity; EDTA = ethylenediaminetetraacetic acid; NaCl = sodium chloride; SDS = Sodium dodecyl sulfate; MCT = microcentrifuge tubes; DTT = dithiothreitol.

Table 3. Cost (US$) analysis for recurring components of PCR.

Conventional PCR (C-PCR)								Real-Time PCR (RT-PCR)
RC	Qty	Cost per 96 samples	Electrophoresis*			Qty	Cost per 30 samples	RT-PCRSG RC	Qty	Cost per 96 samples	RT-PCRTM RC	Qty	Cost per 96 samples
PCR plate with adhesive sealer/PCR strip with cap/tube^a Taq polymerase	1	5.9/21.8/17	Agarose 1.5% DNA marker 50 bp			1.5 gm 10 μl	14.21 12.56	PCR plate with adhesive sealer SYBR green mix (2x) Primers (20 μM)	1 960 μl	5.9 128.6	PCR plate with adhesive sealer	1	5.9
	72 U	87.66	Tris Boric acid			111 gm 5.35 gm	1.35 0.43		31 μl	0.19	Probe 10μM Primers (20μM)	38 μl 31 μl	26.03 0.19
dNTP (10 mM)	24 μL	65.69	EDTA			0.29 gm	0.18	Pipette Tips 20 μL^c Pipette Tips 200 μL^c	97	8.16	Master mix	960 μl	117
NFW Primers (20 μM)^b Pipette tips 2–10 μL^c	250 μL 31 μL 98	21.28 0.26 8.41	Sybr safe Loading dye			10 μL 30 μL	13.72 0.2		2	0.17	Pipette Tips 20μL^c Pipette Tips 200 μL^c	97 3	8.16 0.25
Pipette tips 1000 μL^c	2	0.17	Pipette Tips 20 μL			33	2.78	Pipette Tips 1000 μL^c	2	0.17	Pipette Tips 1000 μL^c	2	0.17
GETP	1	0.50	Parafilm (4 x 300 inch)			3”	0.06	NFW GETP	250μl 1	21.28 0.5	NFW GETP	250 μl 1	21.28 0.5
Total cost per 96 samples Total cost per 30 samples (electrophoresis)		190/206/201					45.27	Total cost per 96 sample		164.4	Total cost per 96 sample		180
10% loss		19/21/20					4.53	10% Loss		16.44	10% Loss		18
Biologist^d	3 hr	8.76						Biologist^d	2 hr	5.84	Biologist^d	2 hr	5.84
Cost per sample		2.27/2.45/2.4					1.66	Cost per sample		1.95	Cost per sample		2.12
Cost/sample with electrophoresis						3.93/4.11/4.06

^aEither PCR 96 well Plate or 8 well strip or single tube (0.2 mL) could be used based on availability/sample size in any variant of PCR, bPrimer price is calculated by considering 25 bp in length (300 μl volume with 100 μM concentration), cPipette Tips used in conventional PCR are of autoclavable and in Real time PCR are of filtered and racked, but either Pipette Tips could be used in above PCR methods, dBiologist salary is calculated using formula: cost/n sample = (total salary of biologist (US$) x time required per sample (hr) x total no. of sample)/(working days in a month x working hour per day), *Electrophoresis unit has combs with 15 wells (of standard size) each, 2 combs can be placed vertically with equal distance in one large gel, one well reserved for DNA marker in each row, thus in total 28 wells could be utilized at the same time. All above prices listed above are inclusive of 5% GST (Goods and Service Taxes) as per rule by Department of Scientific and Industrial Research (DSIR), Government of India. RC = recurring component; Qty = quantity; dNTP = deoxynucleoside triphosphate; NFW = nuclease-free water; EDTA = ethylenediaminetetraacetic acid.

Table 4. Effect of sample size (n) on cost (US$) and time (in months) on DNA extraction and PCR methods.

									Time (months)	*Addition of 5 and 10 sets of NRC at 100,000 and 1 million sample size, respectively
Sample size (n)	NRC cost	NRC cost per sample	RC cost	RC cost per sample	Total cost = NRC cost + RC cost	Total cost per sample	Dividend factor on total cost	NRC cost per total cost (%)	Systematic utilization	NRC cost*	Total cost*	Total cost per sample*
A. Phenol chloroform isoamyl alcohol (PCI) method
1	6,622	6,622.80	1,220	1,220	7,842	7,842	-	84	1
10	6,622	662.18	1,220	122	7,842	784	10	84	1
100	6,622	66.22	1,281	12.81	7,903	79.03	9.9	84	1
1,000	6,622	6.62	3,300	3.30	9,922	9.92	8.0	67	1
10,000	6,622	0.66	13,822	1.38	20,444	2.04	4.9	32	6
100,000	6,622	0.07	1,22,945	1.23	1,29,567	1.30	1.6	5	56	33,109	1,56,054	1.56
1,000,000	6,622	0.01	10,36,290	1.04	10,42,912	1.04	1.2	0.6	556	66,218	11,02,508	1.10
B. Spin column-based manual (SCM) method
1	7,030	7030	359	359	7,390	7,390	-	95	1
10	7,030	703	359	36	7,390	739	10	95	1
100	7,030	70.30	777	7.77	7,808	78.08	9.5	90	1
1,000	7,030	7.03	2,641	2.64	9,671	9.67	8.1	74	1
10,000	7,030	0.70	25,193	2.52	32,224	3.22	3.0	23	4
100,000	7,030	0.07	2,51,122	2.51	258,152	2.58	1.2	3	37	35,152	2,86,274	2.9
1,000,000	7,030	0.01	25,10,405	2.51	25,17,436	2.52	1.0	0.3	370	70,304	25,80,709	2.6
C. Spin column-based semi-automated (SCS) method
1	11,340	11340	542	542	11,882	11882		96	1
10	11,340	1134	542	54	11,882	1188	10	96	1
100	11,340	113.40	963	9.63	12,303	123	9.7	93	1
1,000	11,340	11.34	3063	3.06	14,403	14.40	8.5	80	1
10,000	11,340	1.13	29290	2.93	40,630	4.06	3.5	29	8
100,000	11,340	0.11	291737	2.92	303,077	3.03	1.3	4	83	56,700	3,48,437	3.5
1,000,000	11,340	0.01	2916697	2.92	2,928,037	2.93	1.0	0.4	833	1,13,400	30,30,097	3.0
D. Magnetic bead-based automated (MBA) method
1	69,734	69,734	3,501	3,501	73,235	73,235		95	1
10	69,734	6,973	3,501	350	73,235	7324	10	95	1
100	69,734	697.34	3,501	35	73,236	732.36	10.0	95	1
1,000	69,734	69.73	6,869	6.87	76,604	76.60	9.6	91	1
10,000	69,734	6.97	37,722	3.77	1,07,457	10.75	7.1	64	1
100,000	69,734	0.70	3,57,606	3.6	4,27,341	4.27	2.5	16	9	3,48,672	7,06,278	7.06
1,000,000	69,734	0.07	35,46,862	3.6	36,16,597	3.62	1.2	1.9	87	6,97,344	42,44,206	4.24
E. Nested PCR (N-PCR)
1	11,694	11,694	2,388	2,388	14,082	14,082		82	0.1
10	11,694	1,169	2,388	239	14,082	1,408	10	82	0.1
100	11,694	117	2,480	25	14,174	142	10	81	0.1
1,000	11,694	11.69	5,318	5.32	17,011	17.01	8	67	0.3
10,000	11,694	1.17	32,849	3.28	44,543	4.45	4	24	3
100,000	11,694	0.12	3,21,459	3.21	3,33,153	3.33	1	3	26	58,470	3,79,929	3.80
1,000,000	11,694	0.01	32,14,592	3.21	32,26,286	3.23	1	0.3	260	1,16,939	33,31,531	3.33
F. SYBR Green dye-based RT-PCR (RT-PCRSG)
1	21,495	21,495	795	795	22,290	22,290		96	0.1
10	21,495	2,150	795	80	22,290	2,229	10	96	0.1
100	21,495	215	796	7.96	22,291	223	10	96	0.1
1,000	21,495	21	1,931	1.93	23,426	23.43	9.5	92	0.1
10,000	21,495	2.15	11,868	1.19	33,363	3.34	7.0	64	1
100,000	21,495	0.21	1,16,876	1.17	1,38,371	1.38	2.4	15	9	1,07,475	2,24,351	2.24
1,000,000	21,495	0.02	11,68,759	1.17	11,90,254	1.19	1.2	2	87	2,14,950	13,83,709	1.38
G. Hydrolysis probe-based RT-PCR (RT-PCRHP)
1	21,495	21,495	1,191	1,191	22,686	22,686		95	0.1
10	21,495	2,150	1,191	119	22,686	2,269	10	95	0.1
100	21,495	215	1,191	12	22,686	227	10	95	0.1
1,000	21,495	21.50	5,439	5.44	26,934	26.93	8	79	0.1
10,000	21,495	2.15	14,449	1.44	35,944	3.59	7	59	1
100,000	21,495	0.21	1,39,138	1.39	1,60,633	1.61	2	13	9	1,07,475	2,46,613	2.47
1,000,000	21,495	0.02	13,69,324	1.37	13,90,819	1.39	1	1	87	2,14,950	15,84,274	1.58

Every successive sample size (n) is increasing 10 times from the preceding one. The effect of increased sample size on RC, NRC and TC was assessed using dividend factor, which is calculated by dividing the cost at a particular sample size by the cost at immediately preceding sample size. If the value of dividend factor on cost and multiplication factor on sample size is same, it implies that cost is dependent on sample size, else factors other than sample size are involved in determining cost per sample. NRC = non-recurring component; RC = recurring component. NRC includes 4A. Refrigerated microcentrifuge, weighing balance, heat block, vortex, and pipettes. 4B Refrigerated microcentrifuge, incubator, heat block, vortex, and pipettes. 4C. Semi automated DNA extractor (QIAcube, Qiagen, Germany), incubator, microcentrifuge, heat block, vortex, and pipettes. 4D. Automated DNA extractor (Chemagic™ 360, PerkinElmer, U.S.A.), automated shaker, microcentrifuge, vortex, Voyager™ Adjustable Tip Spacing Pipette (Integra Biosciences) and pipette 1000 μL 4E Thermal cycler, casting unit, power unit, flask, microwave, oven mitts, gel documentation system, glass bottle, beaker, autoclave 95 L, measuring cylinder, weighing balance, microcentrifuge, vortex, −20°C bucket (benchtop gel-based) and pipettes. 4F and 4 G real-time thermal cycler, microcentrifuge, vortex, −20°C bucket (benchtop gel-based) and pipettes. NRC cost used for the study are listed in the supplementary material, S4.

2.1. Sample collection and preparation

For malaria diagnosis, samples could be obtained as whole blood (heparinized) or as dried blood spot (DBS) on filter paper/FTA® card. All constituents of RCs required for blood collection into heparinized tubes or onto sterile paper as DBS were included in the cost evaluation. Before DNA extraction from DBS samples, pre-processing (punching) is required. Thus, a biologist was needed only for DBS samples, not for collected whole blood sample.

2.2. DNA extraction

For DNA extraction via phenol chloroform isoamyl alcohol (DE-PCI) method, only standard protocol out of various modified protocols available in the literature was used for cost comparison analysis [12] Commercially available kits were used in spin column-based DNA extraction (DE-SCM) and magnetic bead-based (DE-MBA) method and were compared on multiple grounds like components required, time consumption, rate of sample processing per day, expertise required and cost per sample. Equipment present at authors laboratory at ICMR-NIMR were used as a guide for selecting maximum capacity of samples, that could be processed in one batch (for example, available 30 places in centrifuge were used as maximum (100%) capacity for manual methods, 12 and 96 places available at semi-automated spin column-based DNA extractor and automated magnetic bead-based DNA extractor (Chemagic™ 360, PerkinElmer, U.S.A.) were used as their full capacity for one batch, respectively. DNA quantification cost was estimated for spectrophotometry and fluorometry methods.

2.3. Polymerase chain reactions

Nested PCR (N-PCR), SYBR™ Green dye-based Real-Time PCR (RT-PCR_SG) and Hydrolysis probe-based Real-Time PCR (RT-PCR_HP) methods were compared. Reagents concentration like Taq polymerase, dNTPs, oligonucleotides and protocol required for cost and time calculation respectively, for 20 μL RT-PCR reaction volume, was availed from SOP used by authors’ lab [13].

2.4. Personnel hiring

Manpower salary was calculated based on the ICMR guidelines for recruitment of staff for research projects [14,15]. Hourly wages were calculated for field assistant, junior research fellow (JRF), senior research fellow (SRF) and research associate-I (RA-I), using their respective monthly salaries, which are US$ 216.32 US$ 461.96, US$ 521.57 and US$ 700.40, respectively, considering a fixed number of days (30 days) of a month with 8 hrs working per day. Personnel cost is computed with the formula - (total salary of a biologist (US$) x time required per sample (hr) x total no. of the sample)/(days in a month x working hour per day).

2.5. Cost analysis

Microsoft Excel® 2010 was used for unit sample cost computation for methods like sample collection and processing, DNA extraction, quantification, and amplification.

3. Results

We identified constant parameters for a step-wise cost analysis and interpretation for malaria molecular diagnostics methods using Microsoft Excel® spreadsheet, provided as supplementary material, S1. Collection of blood samples as heparinized whole blood (US$ 0.17 per sample) is economically beneficial in comparison to the blood collection as DBS (Table 1), for the samples to be used within short period of time (up to 1–2 months). Use of Whatman® FTA® card for DBS sample is 25 fold expensive (US$ 4.19 per sample) as compared to heparinized whole blood method. But, DBS sample with economized recurring components could be used to reduce the overall cost (US$ 0.85 per sample) of sample collection and preparation.

3.1. DNA extraction via spin column-based manual (DE-SCM) method and real-time PCR via SYBR green chemistry (RT-PCR_SG) are the lowest costly methods

Both DE-PCI and DE-SCM method require similar equipment to perform the experiment. However, cost analysis depicted that DE-SCM method has the lowest cost (US$ 3.93 per sample) and is 1.4 fold lower priced than DE-PCI method (US$ 5.32 per sample). But, in DE-PCI method a long list of reagents is required, for which solutions are needed to be prepared before proceeding an experiment. Thus, making the whole process laborious, error-prone and time consuming (requires 4 hr per 30 samples, after excluding solution preparation time). On the other hand, kit-based methods don’t require reagent preparation and thus reduce error-rate and time consumption (3 hr per 30 samples). Semi-automated DNA extraction method by QIAcube®, Qiagen, Germany evaluated here showed relatively higher cost (US$ 4.58 per sample) as compared to other kit-based DNA extraction methods since only 12 samples per batch could be processed at any given time (Table 2). DNA extraction by SCM and magnetic bead-based automated (MBA) method are equally economical (US$ 3.94 per sample), when sample size is low and minimum pack size kits are used (Table 2).

Although the RT-PCR_SG method incurs the lowest cost (US$ 1.95 per sample), but to increase the specificity of assay, RT-PCR_HP method can be used without imparting any major burden, as its more budget friendly (US$ 2.12 per sample) than the conventional one (US$ 3.94 per sample). The overall cost may vary depending on the use of PCR plates/strips/individual tubes. The cost of PCR strips is highest followed by PCR plates and tubes. Conventional PCR cost is more than the Real-Time PCR, as it also includes the cost of electrophoresis and gel documentation, for visualization of results (Table 3).

3.2. Number of samples drives the cost per sample

Total cost for DNA extraction via magnetic bead automated (DE-MBA) method at sample size (n) one, is tenfold higher than the lowest cost DNA extraction via spin column-based manual (DE-SCM) method, but, at higher sample size of 100,000, this cost difference is decreased to 1.6 fold (Table 4-D). Further, to get insight for DE-SCM method, its RC cost was evaluated and compared with other DNA extraction methods and depicted that unit sample cost decreases with increase in sample size (from 1 to 1000), but at higher sample size of ≥ 10,000, its cost stalls at US$ 2.5 in DE-SCM. A similar trend was observed in DE-SCS method, where at sample size of 10,000, unit sample cost stalls at US$ 2.9. However, in DE-MBA method, at sample size of 100,000, unit sample cost stabilizes (US$ 3.6) and does not decrease further. In contrast, DE-PCI method RC cost per sample continues decreasing with increase in sample size from 100,000 to 1 million. It is also interpreted that at sample size of 10,000, DE-PCI method becomes least expensive (US$ 1.38) in comparison to other DNA extraction methods.

Cost analysis in Table 4E–G, depicted that nested PCR NRC costs are 2-fold economical (using similar brands) than RT-PCR NRC, however, if only RC cost is compared, the former method, at any sample size, is the most expensive method for DNA amplification while RT-PCR_SG has the lowest cost followed by RT-PCR_HP. To get a better picture of cost analysis for DNA amplification methods, total cost (NRC+RC) per sample was calculated and interpreted that at smaller sample size (1 to 1000), nested PCR has the lowest cost, thereafter, at sample size ≥ 10,000, RT-PCR_SG method is the least expensive followed by RT-PCR_HP method.

For advanced methods like DE-MBA and RT-PCR, equipment cost contributes 59–94% of total cost up to sample size of ≤ 10,000, but at sample size of ≥ 1,00,000 recurring cost takes over equipment cost. On the contrary, cost analysis on methods like DE-PCI, DE- SCM, DE-SCS, C-PCR and N-PCR shows that at sample size of 10,000, NRC cost is only 21–36% of total cost (Table 4).

3.3. High-throughput is better achieved by advanced methods like DE-MBA and RT-PCR

Automated DE-MBA method requires minimum time for DNA extraction among four methods described in the current study (Table 4). Also, the method is simple, more reliable (reducing dispensing errors and chances of contamination) and attracts lower cost (US$ 3.62 per sample) if, high throughput is required. A systematic approach for calculation of time is used. Kit-based methods require incubation time for the sample followed by running time on machine for DNA extraction/DNA amplification. Sample incubation preparation time and running time of machine for the first batch of samples could be utilized in the preparation/incubation of second batch of samples for systematic utilization of biologist time. First batch of 96 samples requires two hours for automated DNA extraction, then; using the aforementioned strategy, in every subsequent hour next batch of 96 samples could be processed on that same day. Thus, approx. 672 samples could be processed in one day (working 8 hrs.), reduces net cost per sample by automated DNA extractor. It is also interpreted that DE-MBA and RT-PCR methods can process equal number of samples in a given time.

Though, it is a user discretion to work with 1 set of NRC and biologists in a study with fixed sample size, additional set of NRC and biologist is essential at a higher sample size to complete the study within stipulated time. For sample size of 100,000, DE-SCM method requires 37 months to complete the DNA extraction. So, to match the speed and to overcome the budget constraints of DE-MBA method, DE-SCM method with four sets of NRC and biologist may be used.

Molecular diagnosis of malaria per sample is a multi-step process which requires sample preparation, DNA extraction, quantification and in the last, analysis by PCR. After estimating cost of individual laboratory technique/method, complete cost per sample is calculated for every possible combinations of molecular diagnosis method (Supplementary material, S2) and identified that sample processing via heparinized whole blood, DNA extraction using SCM method and final analysis by RT-PCR_SG is the lowest cost method (US$ 6.04 per sample) out of all possible combinations. In contrast, DNA extraction by PCI method from DBS FTA™ card followed by conventional PCR costs US$ 13.53 per sample, which is 2.5-fold higher than former calculated economical cost.

The spectrophotometry (US$ 0.08 per sample) method is much less expensive than the fluorometry method and does not require any assay kit for DNA quantification (Supplementary material, S3). Thus, it makes the former method the first choice for researchers to complete a project more budget-friendly. On the contrary, the fluorometry method requires costly assay kits, which showed high variation in the price with the type of dye: SYBR™ Green (US$ 1.76 per sample) or PicoGreen (US$ 3.66 per sample) and pack size change. The smaller the pack size of the assay kit higher will be the cost, i.e. a 100 assay pack size of SYBR™ Green assay kit costs around US$ 43 per 30 samples, but a 500 assay pack size costs US$ 28.4 per 30 samples, which is 1.5 fold costlier than the larger one.

4. Discussion

Malaria molecular (PCR-based) diagnosis is a multi-step process which requires sample collection and preparation, DNA extraction, quantification (optional), and finally, analysis by PCR. The present study systematically divides the resources into RC and NRC required to execute every step of molecular diagnosis. Two different strategies are applied to estimate malaria molecular diagnostic costs. Strategy 1 envisages the cost per sample with only RCs with available smallest pack size of reagents. Research planners may utilize Tables 1–3 to detail the RCs requirement for sample collection, DNA extraction, and PCR methods. Strategy 2 provides a vision to researchers, policymakers and funding agencies for identifying the appropriate method vis-à-vis sample size and time for DNA extraction and PCR. Cost per sample estimation using strategy 2 involves both RC and NRC costs (Table 4).

Blood sample may be collected either as heparinized whole blood or as DBS on FTA® (Flinders Technology Associates) card/Whatman® filter paper. Cost analysis depicted that heparinized whole blood sample collection is much less expensive than the DBS sample collection. However, the DBS sample collection provides a reliable method for the transport and storage of DNA by stabilizing and immobilizing nucleic acid on card matrix [16]. Usage of Whatman® filter paper over FTA® card for DBS sample collection can reduce the budget for sample collection. The DBS samples placed in plastic zip bags with silica as desiccant at −20°C may be stored up to 2 years without loss of sensitivity in amplification of DNA [17]. Storage condition and method of DNA extraction greatly affects sensitivity of DNA amplification [18–20]

Both DNA extraction and PCR have been revolutionized from a lengthy and laborious process to a time-efficient and automated process which effectively reduce manual handling of samples, thus, minimize errors while dispensing and chances of contamination. Four methods for DNA extraction are compared and interpreted that DE-SCM is the cheapest among them. Automation reduces the cost per sample with increase in sample size as a fixed equipment cost spreads over to higher sample size. It is also found that RCs with higher pack size have low price because mass production enables manufacturers to save labor cost and time to produce large quantity reagents. In RT-PCR_HP method, Excel® MGB Probe with 50,000 pmol (largest pack size available) instead of 6000 pmol (smallest pack size) reduces the cost per reaction from US$ 2.1 per sample to US$ 1.97 per sample. Difference of US$ 0.13 per sample may help in saving of US$ 13000 at sample size 100,000. Similarly, Whatman® Green PCR Master Mix available in a larger pack size of 50 mL decreases cost per sample to US$ 1.61 from US$ 1.93 (for smallest pack size of 1 mL). Besides the pack size, it is noteworthy that the prices are highly likely to vary with the brand of recurring and non-recurring components mentioned in the Tables 1–3. If conventional PCR reagents cost of an economical brand/company is compared with RT-PCR method, it costs near to RT-PCR_HP method. Further, diagnostics based on multiplex PCR which involves simultaneous study on multiple targets in one reaction/tube, further reduce cost per sample. But, due to considerable debate on sensitivity of monoplex PCR and multiplex PCR in malaria diagnostics [21,22] and to maintain the uniformity in results, cost analysis in the present study is performed with monoplex PCR only.

Traditional methods are utilized in budget constraint laboratories where advanced equipment facilities are not present. N-PCR is predominately used for DNA amplification [23]. Here, we depicted that N-PCR is expensive and time-consuming methodology over RT-PCR at any sample size. So, RT-PCR may be promoted in the settings where surveillance and research is recommended. Alternatively, the employment of central lab facilities or outsourcing could be one possible solution to work with advanced methods [24]. In addition, automated DNA extraction coupled with RT-PCR allow quick detection of sub-microscopic infection, enabling precise and early treatment. Thus, it may quench the parasite transmission to other individuals and henceforth, reduce malaria and financial burden.

In molecular diagnosis, the sample/DNA used as calibrator/standard curve generation for copy number calculation requires nucleic acid quantification, which may be performed either by spectrophotometry or fluorometry. Although, fluorometers are less-expensive than spectrophotometers (NanoDrop™), fluorometric assay are relatively expensive. PicoGreen dye present in fluorometric assay kit is ultrasensitive nucleic acid stain and provides sensitivity up to 5 ng/ml for DNA quantification [25,26]. However, field/patient samples DNA amplification by PCR may be performed without DNA quantification. Thus, DNA quantification cost is not included in the complete malaria molecular diagnosis test per sample.

As per authors’ best knowledge after the insight into the published data, the presented study is the first such comprehensive cost-analysis for standard/established malaria molecular diagnostic methods. Sazed et al., reported similar results in very brief for cost-comparison of PCR methods describing nested PCR as an expensive and RT-PCR_SG as most economic among three methods [27]. However, this study didn’t report the detailed strategy for cost computation. In another prospective study a molecular diagnostic method, loop mediated isothermal amplification (LAMP) versus microscopy cost was estimated [28]. Surprisingly, LAMP was found more cost-effective. But, knowing the fact that LAMP is not an established/standard method for malaria molecular diagnostics test, we have not included LAMP cost analysis in our study. Although molecular biology techniques evaluated in the present study are framed with malaria diagnosis only but may be extrapolated to any other research/clinical diagnosis where molecular techniques are required. A few reports on molecular diagnostics in the field of cancer were also reported [29,30], albeit, detailed information on every component of the diagnostic methods was missing which are covered in the present study. In addition to individual component costs, we have also shown the change in the cost of specific diagnostic methods with respect to sample size. Shah et. al included labor cost, equipment, and consumables costs for comparing the cost of tuberculosis diagnosis but lack the perspective of sample size [31].

Besides, the present study provides cost analyses for malaria molecular diagnosis, there is limitation in approach of choosing author’s laboratory-specific SOP for a set of experiments. Personnel hiring wages are computed against only three posts (JRF, SRF, and RA) and didn’t consider other positions like project assistant (PA), project officer (PO), scientist, and administrative staff required to handle the project [8]. Above all, equipment or consumables cost have dynamic range in the market, thus, the prices mentioned in this study are highly likely to vary with the change in brand, pack size of the same brand, but within the same brand, and catalog number. In the current study, prices of all components are valid till 2022 and are subject to change thereafter. The price mentioned in the study are the list price that can be further discounted. Besides discounts offered by manufacturer, bulk purchase also offers best negotiated prices which can reduce the cost further. Change in price is subject to inflation rate in a country which varies from country to country. Also, components/facilities which includes infrastructure (building, electric and electronic devices with accessories, electricity and water), manpower (administrative staff, security guard and cleaning staff), telecommunication gadgets and facility, and biomedical waste management, were not included in the cost-analysis study, assuming that these facilities are already present where the experiment is supposed to be performed. Additionally, to and fro sample/personnel transportation charges are not included in the study, considering it as fickle.

5. Conclusions

This study provides a guide for malaria molecular diagnosis method adoption for researchers, policymakers and funding agencies. Three main facets were recognized as determinants of the cost in the present study: First, type of method/technique adopted for malaria diagnostics, second is number of samples to be processed (indirectly, reagent pack size), and third is capacity utilization. This study identified and concluded that heparinized whole blood sample processed with DE-SCM method and DNA amplification via RT-PCR_SG provides 2.5-fold economical diagnostic test for Plasmodium infection in comparison to conventional methods. To achieve high throughput and high sensitivity automated DNA extraction coupled with RT-PCR provides quick detection of low-density, mixed infection and neglected Plasmodium species.

Declaration of interest

Some brand/company names are used in the current study, which the authors’ laboratory has used in scientific research. The authors declare that these brands/companies are not involved with this manuscript in any way including the study design, the data analysis and interpretation, and writing. Their inclusion is not intended to be promotional. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Glossary

Recurring Components (RC) refer to the goods that are either consumed (completely used-up) in any given experiment/assay and/or those that are for one-time use only (example reagents, tissue paper, gloves etc.).

Non-Recurring Components (NRC) refer to the goods that are not fully consumed in any given experiment/assay, and are available for the next experiment. In other words, these components are purchased once (one-time-charges) and used-up multiple times like equipments, pipette, micro tube racks etc.

Capacity-Utilization (CU) refers to the extent to which available productive space of equipment is used in any given assay at specific time. For example, in a centrifuge with 30 places for microcentifuge tubes (MCT’s) have maximum CU of 30.

Capacity-Utilization dependent Components (CUdC) refers to resources (mainly RC’s) whose quantity utilized is dependent on the sample size steered in a given experiment. For example, in one 20 μl qPCR reaction, 10 μl master mix (2×) is used and thus, in ten qPCR reactions 100 μl master mix (2×) is consumed.

Capacity-Utilization Independent Components (CUiC) refers to resources whose quantity utilized in one experiment is independent on the sample size steered. For example, one pair of gloves is required in any experiment irrespective of sample size.

Systematic-utilization refers to the management of time, while executing an experiment, for its maximum utilization.

Acknowledgments

The authors thank the Director of the Indian Council of Medical Research-National Institute of Malaria Research (ICMR-NIMR), New Delhi for providing administrative support and infrastructure.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/14737159.2024.2356172

Additional information

Funding

V Panwar, S Bansal and C Chauhan were supported by the project fellowships under the project “Establishing a pan-India landscape of human Plasmodium infections from dried blood spots collected under National Family Health Survey – 5” from the ICMR-NIMR.