Skip to main content

Cost effectiveness of intermittent screening followed by treatment versus intermittent preventive treatment during pregnancy in West Africa: analysis and modelling of results from a non-inferiority trial

Abstract

Background

Emergence of high-grade sulfadoxine-pyrimethamine (SP) resistance in parts of Africa has led to growing concerns about the efficacy of intermittent preventive treatment of malaria during pregnancy (IPTp) with SP. The incremental cost-effectiveness of intermittent screening and treatment (ISTp) with artemether-lumefantrine (AL) as an alternative strategy to IPTp-SP was estimated followed by a simulation of the effects on cost-effectiveness of decreasing efficacy of IPTp-SP due to SP resistance. The analysis was based on results from a multi-centre, non-inferiority trial conducted in West Africa.

Methods

A decision tree model was analysed from a health provider perspective. Model parameters for all trial countries with appropriate ranges and distributions were used in a probabilistic sensitivity analysis. Simulations were performed in hypothetical cohorts of 1000 pregnant women who received either ISTp-AL or IPTp-SP. In addition a cost-consequences analysis was conducted. Trial estimates were used to calculate disability-adjusted-life-years (DALYs) for low birth weight and severe/moderate anaemia (both shown to be non-inferior for ISTp-AL) and clinical malaria (inferior for ISTp-AL). Cost estimates were obtained from observational studies, health facility costings and public procurement databases. Results were calculated as incremental cost per DALY averted. Finally, the cost-effectiveness changes with decreasing SP efficacy were explored by simulation.

Results

Relative to IPTp-SP, delivering ISTp-AL to 1000 pregnant women cost US$ 4966.25 more (95 % CI US$ 3703.53; 6376.83) and led to a small excess of 28.36 DALYs (95 % CI −75.78; 134.18), with LBW contributing 81.3 % of this difference. The incremental cost-effectiveness ratio was −175.12 (95 % CI −1166.29; 1267.71) US$/DALY averted. Simulations show that cost-effectiveness of ISTp-AL increases as the efficacy of IPTp-SP decreases, though the specific threshold at which ISTp-AL becomes cost-effective depends on assumptions about the contribution of bed nets to malaria control, bed net coverage and the willingness-to-pay threshold used.

Conclusions

At SP efficacy levels currently observed in the trial settings it would not be cost-effective to switch from IPTp-SP to ISTp-AL, mainly due to the substantially higher costs of ISTp-AL and limited difference in outcomes. The modelling results indicate thresholds below which IPT-SP efficacy must fall for ISTp-AL to become a cost-effective option for the prevention of malaria in pregnancy.

Background

Malaria in pregnancy (MiP) is associated with poor health outcomes in the mother and child, primarily during the first and second pregnancies. The most notable adverse health outcomes in moderate or high transmission settings include maternal anaemia, perinatal mortality and low birth weight (LBW) [1].

The approach to MiP prevention currently recommended by WHO consists of long-lasting insecticide treated bed nets (LLIN) and provision of sulfadoxine-pyrimethamine (SP) as intermittent preventive treatment (IPTp-SP) at each scheduled antenatal care (ANC) visit from the beginning of the second trimester until delivery [2, 3]. By clearing existing infections caused by drug-sensitive parasites as well as preventing incident infections, IPTp-SP reduces the risk of maternal anaemia, LBW and neonatal mortality [4]. Currently, 39 countries in malaria endemic sub-Saharan Africa have an IPTp policy [5]. However, analysis of national survey data from 27 countries from 2009 to 2011 estimated that despite high ANC coverage (≥2 visits, 75.1 %), only 21.5 % of the total births at risk of malaria were born to mothers who received IPTp-SP [6].

Over the last decade, the emergence and spread of high-level parasite resistance to SP in eastern and southern Africa has led to growing concerns about the effectiveness of IPTp-SP [710], although the loss of efficacy may be lower in pregnant women than in children under five years of age [11]. SP resistance occurs through point mutations in the genes encoding the target enzymes of SP, dihydropteroate synthase (dhps) and dihydrofolate reductase (dhfr) with prevalence of different mutant alleles varying throughout Africa. In areas with >90 % prevalence of the quintuple mutation, named after its its molecular maker “K540E”, there was continued beneficial impact of IPTp-SP on birthweight and maternal haemoglobin levels. However, IPTp-SP effectiveness becomes threatened in areas where prevalence of the sextuple Pfdhps-A581G mutation exceeds 10 %, such as northern Tanzania, western Kenya and southern Uganda [9, 10, 1215]. For such areas there is a criticial need to evaluate alternative drugs to replace SP or alternative strategies using a diagnostic based test-and-treat intervention to replace IPTp. Trials of mefloquine and azithromycin with chloroquine showed that these alternatives were not sufficiently well tolerated or efficacious in pregnant women to be used for IPTp [16, 17]. IPTp with dihydroartemisinin/piperaquine has recently shown more encouraging results in Kenya and Uganda [13, 18], but there are concerns about using artemisinin combinations for widespread prophylaxis. Intermittent screening of women at each ANC visit with a rapid diagnostic test for malaria (RDT) and treatment of those with a positive result with an artemisin-based combination therapy (ACT) (ISTp) is a potential alternative approach to IPTp-SP in areas where the latter is no longer effective. A pilot study of this approach conducted in Ghana showed that ISTp was non-inferior to IPTp-SP in preventing LBW and maternal anaemia [19]. To confirm this result a non-inferiority trial which compared ISTp with artemether-lumefantrine (AL) versus IPTp-SP was conducted in four countries in West Africa (Burkina Faso, Ghana, Mali and The Gambia) [20]. The trial found ISTp-AL to be non-inferior to IPTp-SP with respect to all the primary outcomes, which were risk of LBW, mean maternal haemoglobin concentration prior to delivery and the prevalence of placental malaria [20]. However, the incidence of clinical attacks of malaria was significantly higher in women who received ISTp-AL than in those who received IPTp-SP.

Policymakers need to know the costs and cost-effectiveness of ISTp-AL and IPTp-SP as well as their clinical impact in order to inform any decisions about a change of strategy in the management of MiP. It was anticipated that the cost of ISTp-AL would be higher than IPTp-SP because (i) every woman receives an RDT (which itself is slightly more expensive than SP), (ii) it requires a more expensive drug to be administered to women screened as positive for malaria parasites, and iii) the procedure is more time-consuming. ISTp-ALis likely to be less effective where IPTp SP works well because RDTs have a low sensitivity [13] and women derive prophylactic benefit from IPTp-SP even if they are uninfected at the time of drug administration. For these reasons, the objective of this analysis was to estimate the incremental cost effectiveness of ISTp-AL versus IPTp-SP and then to simulate the effects on cost effectiveness of changes in SP efficacy due to spread of resistance. The cost effectiveness model incorporated the primary and key secondary endpoints published by Tagbor et al. [20]. In addition a cost-consequences analysis (CCA) was conducted in order to: (1) identify and consider effects that cannot easily be translated into health outcomes and (2) to report outcomes and costs separately, instead of aggregating them into a composite measure such as DALYs or QALYs and incremental cost-effectiveness ratios (ICERs) [21].

Methods

Trial setting, population, procedures and analysis

The trial (ClinicalTrials.gov reference: NCT01084213) enrolled a total of 5354 women in their first or second pregnancy at five sites in Burkina Faso, Ghana, Mali and The Gambia, with seasonal moderate or high malaria transmission and a prevalence of the K540E haplotype <1 % [14, 20]. All participants received an LLIN at enrolment, and then received either IPTp-SP or ISTp-AL at ANC contacts during their pregnancies. The primary objective of the trial was to establish whether ISTp-AL was non-inferior to, or in other words not worse by a clinically important amount than IPTp-SP in terms of preventing the following primary outcomes: (1) LBW (non-inferiority odds ratio (OR) margin <1.263), (2) maternal haemoglobin (Hb) concentration prior to delivery (non-inferiority margin <0.2 g/dl Hb concentration reduction) and (3) the prevalence of placental malaria (non-inferiority OR margin <1.286) [22]. More details on the trial population, procedures and data analysis are provided in Tagbor et al. [20].

Outcomes

For the analysis described below, model outcomes were selected from the outcome categories affecting the mother or the offspring based on (1) clinical and economic importance, and (2) availability of disability weights to calculate disability-adjusted-life-years (DALYs). Based on these criteria, LBW, moderate to severe anaemia prior to delivery (both primary trial outcomes for which ISTp-AL was shown to be non-inferior) and clinical malaria (defined as an unscheduled clinic visit due to illness with a positive blood smear,- a secondary trial outcome, for which ISTp-AL was shown to be inferior to IPTp-SP) were included as model outcomes in the cost-effectiveness analysis as well as in the subsequent resistance modelling. The total DALYs in each arm were calculated by summing the DALYs from the three outcomes, as disability weights of concurring events (i.e. malaria infection and anaemia) are unknown. In addition, adverse events measured in the trial as well as intangible outcomes, such as the value of treating according to a test result rather than presumptively, were included in a cost-consequences analysis.

Costs

The incremental fixed and variable cost to the health provider of delivering the interventions including and excluding the costs arising from the consequences of MiP (i.e. clinical malaria disease, moderate/severe anaemia and LBW) was calculated. Household costs were excluded since both interventions would be part of routine ANC visits, provided free of charge in most settings in sub-Saharan Africa and, therefore, not expected to differ between the interventions.

Economic costs were calculated and expressed in constant 2012 US$, using the local consumer price index [23] and average 2012 exchange rates [24]. Cost data collection comprised (1) health facility costing studies in the Ségou region, Mali (N = 4) and in the Upper East Region, Ghana (N = 4) and (2) two observational studies of the time nurses took to administer ISTp and IPTp in Ghana (NISTp = 18; NIPTp = 20) and Malawi (NISTp = 30; NIPTp = 18), conducted by the Economics Working Group of the Malaria in Pregnancy Consortium. All cost data collection was approved by the Ethics Committee of the London School of Hygiene and Tropical Medicine and by the relevant ethics committees in the countries where cost data were collected. Informed consent was obtained from each participant being observed. For more details on cost data collection methods see Additional file 1: Appendix S1.

Analysis and model

The cost-effectiveness analysis was conducted from the health provider perspective. To account for the effects of a potential increase in SP resistance, the cost effectiveness of ISTp-AL versus IPTp-SP for decreasing efficacy of SP was simulated. In addition, the effects of reducing the costs of RDTs and AL by one half were analysed. Separate, but structurally identical decision tree models were developed for each outcome (see Fig. 1 for LBW). A lifetime horizon was adopted to show the lifelong (discounted) mortality effects of the consequences of MiP.

Fig. 1
figure 1

The decision tree. The decision tree model illustrates the example for LBW. All numbers represent the results published in Tagbor et al. [20]. The same structure was used for moderate/severe anaemia and clinical malaria. IPTp-SP intermittent preventive treatment with sulfadoxine-pyrimethamine; ISTp-AL intermittent screening and if positive followed by treatment with arthemether-lumefantrine; LBW low birth weight

DALYs were estimated using disability weights from the Global Burden of Disease Study (GBD) 2010 and 2004 [25, 26], applying local life expectancies, no age weighting, and discounting at 3 %. For more details on the calculation of DALYs, see Additional file 1: Appendix S1 [25, 26].

The incremental cost-effectiveness ratio (ICER) was calculated for a hypothetical cohort of 1000 women by dividing the incremental cost of the intervention by the incremental DALYs averted [(CostsISTp-AL-CostsIPTp-SP)—Cost from health consequences from MiP/(DALYIPTp-SPp-DALYISTp-AL)].

The CCA separated costs and consequences into four categories, calculated per 1000 women where applicable: (i) costs, (ii) measurable outcomes contributing to DALYs, (iii) measurable outcomes that do not contribute to DALYs, and (iv) non measurable outcomes, such as the value of not giving medicines to all pregnant women. The costs were calculated including and excluding the costs associated with the health consequences, and presented with 95 % confidence intervals based on percentiles.

To illustrate the uncertainty of all estimates simultaneously we conducted a probabilistic sensitivity analysis (PSA) with 10,000 iterations, producing a point estimate and 95 % confidence interval based on percentiles for the differences in costs and effects, and an average ICER. Cost-effectiveness guidelines were used to assign an appropriate distribution to each parameter [27] (Table 1). The PSA results were plotted on the cost effectiveness plane and assessed against three frequently applied policymaker willingness-to-pay (WTP) thresholds, referred to as low (US$ 39.72), middle (US$ 238.33) and high (US$ 861.33), respectively. The low and middle threshold are the historic WHO thresholds of US$ 25 and US$ 150 adjusted for inflation [28], while the high threshold is the unweighted mean GDP per capita calculated across the four countries [29]. No deterministic sensitivity analysis was conducted, as only unrealistically large changes to a single parameter could lead to a conclusion-changing base case ICER.

Table 1 Input variables for the cost consequence analysis and the base case and probabilistic cost-effectiveness analysis

The impact of different levels of SP resistance on the ICER was subsequently modelled, by changing the efficacy of IPTp-SP while holding the efficacy of ISTp-AL constant as measured in the trial. The measured risk of the model outcomes (LBW, severe/moderate anaemia and clinical malaria) in the IPTp-SP arm of the trial was assumed to be the maximum possible efficacy and labelled as 100 %. All trial participants were assumed to have slept under an LLIN. Details of how SP resistance affects IPTp-SP efficacy are not clearly understood, therefore we assumed a linear relationship between decreasing IPTp-SP efficacy and outcomes. Data from the Cochrane review by Radeva-Petrova et al. [30] (comparing IPTp-SP with placebo or no intervention) were used to calculate the extrapolated risk of the model outcomes in the IPTp-SP group by percentage point change in efficacy, from 100 % (maximum) to 0 % (minimum). As the Cochrane review authors were unable to stratify IPTp efficacy by bed net use, the effect of bed nets on these modelled outcomes was explored in three scenarios, by assuming: (1) bed nets have no effect and the full extrapolated risk was attributable to IPTp-SP (used as the starting point for simulation only), (2) bed nets bear 1/2 and (3) 2/3 of the burden and, therefore, only 1/2, or 1/3, respectively, of the extrapolated risk of the model outcomes was attributable to IPTp-SP. Subsequently, the incremental costs, incremental DALYs and ICERs comparing ISTp-AL with IPTp-SP at each percentage point of decreasing SP efficacy for each of the three bed net scenarios were calculated.

Finally, these hypothetical results were used to calculate the threshold level of SP efficacy at which ISTp-AL would become cost effective, stratified by WTP and assumption about the contribution of bed nets. For more details on how the resistance modelling was conducted, please refer to Additional file 1: Appendix S1. All resistance simulations included the costs arising from health consequences of MiP and at bed net coverage levels estimated in the most recent Demographic and Health Surveys in the four trial countries (unweighted average of 47.7 % across the four countries).

Analysis of international procurement databases indicated an overall downwards trend in the unit cost of RDTs and AL from 2010 to 2014. Therefore, the impact of reducing these commodity costs by 50 % (with the standard error held constant) on the ICER and in the resistance analysis was explored.

To be able to show simultaneously how the cost effectiveness changes by level of SP efficacy, assumption of the contribution of bed nets and bed net coverage, we used a net monetary benefit (NMB) function instead of the ICER, because (a) the simulated results cannot easily be depicted when the denominator of a ratio approaches 0 and (b) because NMB includes the WTP threshold in the formula. The formula for NMB is as follows: NMB = (∆DALY averted x WTP)-∆Costs. When using NMB, an intervention becomes cost effective when the NMB is ≥0.

Stata (version 12, StataCorp, College Station, Texas) was used to analyse the observational data whilst the international procurement data, the health facility costs and the decision tree model were analysed in Excel (Microsoft Office 2013), the latter also using Visual Basic for Applications.

Results

Consequences

Per 1000 women there were 151 (95 % CI 137;167) vs 156 (95 % CI 141;172) LBW babies, 17 (95 % CI 12;25) vs 23 (95 % CI 17;32) cases of severe or moderate anaemia and 74 (95 % CI 61;92) vs 115 (95 % CI 96;144) episodes of clinical malaria at unscheduled visits in the IPTp-SP and ISTp-AL groups, respectively (Table 3). The total burden was estimated as 755.1 and 782.9 DALYs per 1000 women in the IPTp-SP and ISTp-AL arms, respectively (Table 3).

Costs

Table 2 shows the itemized cost results. The total cost (including health worker time and cost of supplies) per administration of IPTp-SP was US$ 0.79 (95 % CI 0.53; 1.12). If screening was negative, the cost per administration of ISTp-AL was US$ 1.45 (95 % CI 1.12; 1.85) and US$ 4.06 (95 % CI 3.29, 4.97) if screening was positive (see Table 2).

Table 2 Itemized cost results

The cost per outpatient visit for malaria treatment was US$ 11.78 (95 % CI 9.18; 14.68), per adult inpatient day US$ 35.27 (95 % CI 19.99; 55.07) and per paediatric inpatient day US$ 63.51 (95 % CI 36.30; 99.27), the latter two excluding medical supplies. The final cost of health consequences of MiP was estimated to be US$ 40.93 (95 % CI 19.52; 71.19) for the short-term consequences of LBW, US$ 10.15 (95 % CI 6.76; 15.17) for a moderate/severe anaemia case and US$ 13.71 (95 % CI 6.12; 24.65) for a clinical malaria case during pregnancy.

Table 3 shows the costs of administering the interventions to 1000 pregnant women, which amounted to US$ 1631.84 (95 % CI 1100.11; 2316.97) for IPTp-SP and US$ 5778.77 (95 % CI 4701.65; 7039.15) for ISTp-AL, when excluding the costs arising from health consequences of MiP. After including costs arising from health consequences it rose to US$ 9006.54 (95 % CI 5610.30; 13,680.56) for IPTp-SP and US$ 13,972.79 (95 % CI 10,199.65; 18,983.86) for ISTp-AL (Table 3). In summary, if 24.7 % of women screened in the ISTp-AL arm have a positive RDT test result, as measured in the trial, and if the costs of the consequences of MiP are included, ISTp-AL costs on average around US$5 more per woman per pregnancy than IPTp-SP. If 10 % (as in The Gambia) or 50 % (as in Ghana) of the RDT screenings were positive the costs per woman would amount to US$ 3.88 and US$ 6.73, respectively.

Table 3 Cost consequence analysis

Cost effectiveness analysis

Compared with IPTp-SP, delivering ISTp-AL to 1000 pregnant women led to an excess of 27.8 DALYs, of which 81.3 % were attributable to LBW, 5.9 % to severe/moderate anaemia and 12.8 % to clinical malaria. With an incremental cost of US$ 4929.0 per 1000 women, this produces an ICER of US$ −177.1/DALY averted (Table 4).

Table 4 Cost effectiveness analysis

The negative ICER is not driven by cost savings, but rather by a slightly lower (not statistically significant) point estimate of efficacy in the ISTp-AL group. The results of the PSA including the costs of the health consequences of MiP are shown in Table 4 and Fig. 2a. Per 1000 women the difference in cost amounted to US$ 4966.25 (95 % CI US$ 3703.53; 6376.83) more for ISTp-AL with an excess of 28.36 (95 % CI −75.78; 134.18) DALYs. The average ICER from the PSA was −175.12 (95 % CI −1166.29; 1267.71) US$/DALY averted. The results excluding the costs of health consequences of MiP are plotted in Fig. 2b yielding a difference in costs of US$ 4146.93 (95 % CI 3420.66; 4978.48) per 1000 women resulting in an ICER of −146.29 (95 % CI −1039.11; 1123.15) US$/DALY averted. The less widely dispersed simulation points on the CE plane reflect less uncertainty around the cost estimate when costs of the health consequences of MiP are excluded.

Fig. 2
figure 2

Cost effectiveness planes: the graphs display the results of two different Monte Carlo simulations with 10’000 iterations each using the value ranges and distributions specified in Table 1. The different coloured lines illustrate the median and three willingness-to-pay (WTP) thresholds. WTP threshold 1 of US$ 861.33 is the GDP/capita averaged over the four countries. WTP threshold 2 and 3 represent the original thresholds defined by WHO in 1993 as highly attractive and attractive, both inflated to US$ 2012. a shows the results including the costs occurring from the consequences of malaria during pregnancy and b excluding these costs

Cost-consequences analysis

Table 3 shows the results of the cost-consequences analysis. Consequences that could not be included in the calculation of DALYs included placental malaria, a primary outcome of the trial, and side effects measured in the trial, all expressed as cases per 1000 women. All side effects which occurred at statistically significantly different rates in the ISTp-AL and IPTp-SP arms are shown. Of the measured side effects there were fewer cases of dizziness, sleeplessness, weakness, nausea and vomiting reported in the ISTp-AL arm. In addition to these measurable differences, there were also other non-measurable outcomes that should be considered. These include benefits, such as the intrinsic value of giving fewer drugs to pregnant women and only administering drugs to women who are infected; the value to women of knowing their infection status; and the value to society of identifying malaria infection, particularly for surveillance purposes. There are also potential deleterious effects, such as failure to treat women with a false negative test.

Modelling of cost effectiveness for decreasing efficacy of SP and decreased commodity costs

Figures 3 and 4 shows the simulation results for a hypothetical reduction of SP efficacy to 50 and 0 % respectively, of the current level. Figures 3a and 4a replicate Fig. 2a, while Figs. 3b, c and 4b, c show the simulation results assuming that bed nets bear 1/2 or 2/3 of the extrapolated outcome burden when SP efficacy is reduced. Simulation results assuming no contribution of bed nets can be found in the Additional file 1: Appendix S1. By reducing SP efficacy levels from 100 % to 50 % or 0 %, more simulation points lie in the North Eastern (top right) quadrant of the cost effectiveness plane and to the right of the WTP threshold (the criterion for being cost-effective with respect to this threshold).

Fig. 3
figure 3

Simulation results: cost effectiveness plane at SP efficacy level of 50 % with a bed net coverage of 47.7 % [37] and with costs from consequences included. a shows the trial results replicated from Fig. 2a for comparison purpose, b, c show the simulation results for the assumption that bed nets bear 1/2 and 2/3 respectively, of the predicted burden of LBW, severe/moderate anaemia and clinical malaria

Fig. 4
figure 4

Simulation results: cost effectiveness plane at SP efficacy level of 0 % with a bed net coverage of 47.7 % [37] and with costs from consequences included. a shows the trial results replicated from Fig. 2a for comparison purpose, b, c show the simulation results for the assumption that bed nets bear 1/2 and 2/3 respectively, of the predicted burden of LBW, severe/moderate anaemia and clinical malaria

Table 5 (in column 2 and 3) indicates the efficacy levels at which ISTp-AL becomes cost effective stratified by WTP threshold and the assumptions concerning burden prevented by bed net use. For example at the middle WTP threshold ISTp-AL becomes cost-effective when the efficacy of IPTp-SP is between (i) 63–64 % or (ii) 59–60 % of current SP efficacy levels, assuming that bed nets bear (i) 1/2 or (i) 2/3 of the extrapolated outcome burden. For the highest and lowest WTP threshold this range lies between 69–73 % and 12–20 %, respectively.

Table 5 Threshold modelling for decreasing SP efficacy: Efficacy levels of IPTp-SP at which ISTp-AL becomes cost effective stratified by (1) two assumptions made on burden prevented by bed net use at a bed net coverage of 47.7 % and (2) willingness to pay threshold

Additional results summarizing the effects of three key variables on cost effectiveness: (i) the level of IPTp-SP efficacy, ranging from 0 to 90 % of current efficacy, (ii) bed net coverage and (iii) the WTP threshold are shown in the Additional file 1: Appendix S1.

Finally, reducing the costs of ACT and RDT by 50 % leads to an ICER of US$ −107.7/DALY averted (∆Cost = US$ 2998.1, ∆ DALY = −27.8). Table 5 (in column 4 and 5) shows the results of the threshold analysis when the costs of the ACT and RDT were halved. The effect of reducing commodity costs leads to a higher threshold of SP efficacy at which ISTp-AL becomes cost effective, in other words, with lower commodity costs, SP needs to lose less of its efficacy for ISTp-AL to become cost-effective.

Discussion

At current levels of cost and efficacy, it is not cost-effective to switch from IPT-SP to IST-AL in areas where Plasmodium falciparum remains sensitive to SP. This is because IST-AL is more costly than IPTp-SP and no more effective at reducing the adverse effects of malaria during pregnancy. However, it is possible to use a cost-effectiveness model to simulate how cost-effectiveness would change with the loss of efficacy of IPTp-SP resulting from a potential development of SP resistance, and with a reduction in commodity costs. The threshold at which IST-AL becomes cost effective compared with IPTp has been assumed to vary depending on the contribution of bed nets to reducing the burden of MiP; another key variable is decision maker WTP to avert a DALY. The analysis presented here shows that for the highest WTP threshold ($861 per DALY averted), ISTp-AL becomes cost-effective at levels of IPTp efficacy ranging from 69 to 79 % of current levels. At a threshold of $238/DALY averted, a greater loss of IPTp-SP efficacy is needed for ISTp-AL to become cost-effective, and for the lowest threshold ($40/DALY averted), IPTp-SP efficacy must fall to 12–36 % of current levels before IST-AL is more cost-effective. These thresholds are higher when commodity costs are halved, i.e. this level is found between 71 and 80 % for the highest and 35–54 % for the lowest WTP threshold.

The cost-consequences analysis represents an effort to introduce into the decision-making metrics the effects of important consequences that cannot be readily incorporated into the calculation of DALYs. For example, as malaria incidence diminishes, the justification for presumptive treatment will need to be re-examined, however safe the drug used for preventive treatment. Valuations of this benefit could be elicited from women and communities. However, integrating this benefit into outcome measures may require moving to a cost-benefit or impact evaluation framework. Similarly, the value of information on infection, either to women individually or as a contribution to surveillance in a context of malaria elimination, could be examined empirically. On the other hand there is also the possible negative impact of women who falsely test negative, who could potentially be deprived of essential treatment. However, for the purpose of this analysis the authors have elected to enumerate, rather than to quantify, these benefits or disadvantages.

This analysis is subject to a number of important limitations. First, the Cochrane review [30] includes results from more recent trials, including some conducted in areas where some level of SP resistance is expected, biasing downwards the efficacy of IPT-SP prior to the simulated decrease due to resistance. Second, the presence of bed nets introduces a series of complications related to distinguishing the incremental effect of IPTp-SP when bed nets are already addressing some share of the burden of MiP. The Cochrane review efficacy estimates were not stratified by bed net use, therefore assumptions about the incremental effectiveness of IPTp-SP in addition to bed nets were needed. All trial participants were given a bed net but actual adherence to bed nets was not measured and, therefore, it was not possible to stratify efficacy by bed net use, although it is expected to have been high under trial conditions.

The costs of ISTp-AL were affected by the proportion of women who test positive when screened with an RDT, which in turn is dependent on malaria transmission intensity. However, even at the lowest observed level of positive test results, ISTp-AL was not cost-effective as the cost of the test itself exceeds that of IPTp-SP.

Costs also depend on the number of times a woman receives each intervention during a pregnancy. At the time of the trial, the policy in three of the trial countries was for women to receive two doses of IPTp-SP, resulting as measured in the trial in 2.06 administrations per woman in the IPTp-SP and 2.76 in the ISTp-AL arm. Although it would be feasible to explore how the frequency of administration would affect costs, it is impossible to estimate how this would influence efficacy, therefore trial estimates were used. As an indication, increasing the number of administrations of IPTp-SP to 2.76 per woman reduced the difference in costs between the two interventions to US$ 4.4 per woman. Transmission level is also likely to influence cost-effectiveness. More evidence is needed from low transmission settings, which also partially overlap with sextuple mutant settings as to how this balance will work out.

This model has focused on simulating how cost effectiveness of ISTp-AL changes if SP efficacy decreases. It is currently not known how SP resistance levels translate into IPTp efficacy and ISTp has yet to be shown to be more effective in a high resistance setting. One trial conducted in Western Kenya, where the prevalence of the sextuple A581G mutant was measured at 5.8 %, found ISTp with dihydroartemisinin–piperaquine (DP) to be inferior to IPTp-SP [13]. It is thought that in settings with even higher prevalence of the A581G sextuple mutant, IPTp-SP has almost no efficacy in preventing clinical malaria, MiP attributable LBW and anaemia [14, 15], however the loss of SP efficacy in such settings remains to be studied. Naidoo and Roper suggested that most P. falciparum resistance patterns appear first in East Africa and usually spread to Central and West Africa within 15 years [31].

Decisions about a change in strategy to prevent MiP will be based on many factors in addition to effectiveness and cost-effectiveness. Feasibility and acceptability to mothers and health workers will be important, as will be the feasibility of policy change. In addition, total cost and affordability will be crucial.

Conclusion

At the current level of SP efficacy and transmission intensity found in this trial in West Africa, switching from IPTp-SP to ISTp-AL is not recommended as ISTp-AL is not more effective and costs considerably more per woman. However, as our modelling suggests, in settings with >10 % prevalence of the sextuple mutant, where IPTp has reduced or no efficacy, ISTp-AL has the potential to be a viable and cost-effective option. Any continued decrease in commodity costs would positively affect this transition.

Abbreviations

AL:

artemether-lumefantrine

ACT:

artemisin-based combination therapy

ANC:

antenatal care

CCA:

cost consequence analysis

CEA:

cost effectiveness analysis

DALY:

disability adjusted life year

Hb:

haemoglobin

HCW:

health care worker

ICER:

incremental cost effectiveness ratio

IPTp:

intermittent preventive treatment of malaria during pregnancy

ISTp:

intermittent screening for malaria during pregnancy and if positive followed by treatment

GBD:

global burden of disease

LBW:

low birth weight

LLIN:

long-lasting insecticide-treated net

MiP:

malaria during pregnancy

NMB:

net monetary benefit

PSA:

probabilistic sensitivity analysis

RDT:

rapid diagnostic test

SP:

sulfadoxine-pyrimethamine

WHO:

World Health Organization

WTP:

willingness to pay

References

  1. Desai M, ter Kuile FO, Nosten F, McGready R, Asamoa K, Brabin B, et al. Epidemiology and burden of malaria in pregnancy. Lancet Infect Dis. 2007;7:93–104.

    Article  PubMed  Google Scholar 

  2. Intermittent preventive treatment in pregnancy (IPTp): key documents and guidelines. http://www.who.int/malaria/areas/preventive_therapies/pregnancy/en/index.html. Accessed 15 Feb 2016.

  3. WHO. A strategic framework for malaria prevention and control during pregnancy in the African region. AFR/MAL/04/01. Brazzaville: World Health Organization: Regional Office for Africa; 2004.

    Google Scholar 

  4. ter Kuile FO, van Eijk AM, Filler SJ. Effect of sulfadoxine-pyrimethamine resistance on the efficacy of intermittent preventive therapy for malaria control during pregnancy: a systematic review. JAMA. 2007;297:2603–16.

    Article  PubMed  Google Scholar 

  5. van Eijk AM, Hill J, Alegana VA, Kirui V, Gething PW, ter Kuile FO, et al. Coverage of malaria protection in pregnant women in sub-Saharan Africa: a synthesis and analysis of national survey data. Lancet Infect Dis. 2011;11:190–207.

    Article  PubMed  PubMed Central  Google Scholar 

  6. van Eijk AM, Hill J, Larsen DA, Webster J, Steketee RW, Eisele TP, et al. Coverage of intermittent preventive treatment and insecticide-treated nets for the control of malaria during pregnancy in sub-Saharan Africa: a synthesis and meta-analysis of national survey data, 2009-11. Lancet Infect Dis. 2013;13:1029–42.

    Article  PubMed  Google Scholar 

  7. Coulibaly SO, Kayentao K, Taylor S, Guirou EA, Khairallah C, Guindo N, et al. Parasite clearance following treatment with sulphadoxine-pyrimethamine for intermittent preventive treatment in Burkina-Faso and Mali: 42-day in vivo follow-up study. Malar J. 2014;13:41.

    Article  PubMed  PubMed Central  Google Scholar 

  8. Gesase S, Gosling RD, Hashim R, Ord R, Naidoo I, Madebe R, et al. High resistance of Plasmodium falciparum to sulphadoxine/pyrimethamine in northern Tanzania and the emergence of dhps resistance mutation at Codon 581. PLoS One. 2009;4:e4569.

    Article  PubMed  PubMed Central  Google Scholar 

  9. Harrington WE, Mutabingwa TK, Muehlenbachs A, Sorensen B, Bolla MC, Fried M, et al. Competitive facilitation of drug-resistant Plasmodium falciparum malaria parasites in pregnant women who receive preventive treatment. Proc Natl Acad Sci USA. 2009;106:9027–32.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Naidoo I, Roper C. Drug resistance maps to guide intermittent preventive treatment of malaria in African infants. Parasitology. 2011;138:1469–79.

    Article  PubMed  PubMed Central  Google Scholar 

  11. Kalanda GC, Hill J, Verhoeff FH, Brabin BJ. Comparative efficacy of chloroquine and sulphadoxine–pyrimethamine in pregnant women and children: a meta-analysis. Trop Med Int Health. 2006;11:569–77.

    Article  CAS  PubMed  Google Scholar 

  12. Kayentao K, Garner P, van Eijk AM, Naidoo I, Roper C, Mulokozi A, et al. Intermittent preventive therapy for malaria during pregnancy using 2 vs 3 or more doses of sulfadoxine-pyrimethamine and risk of low birth weight in Africa: systematic review and meta-analysis. JAMA. 2013;309:594–604.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Desai M, Gutman J, L’Lanziva A, Otieno K, Juma E, Kariuki S, et al. Intermittent screening and treatment or intermittent preventive treatment with dihydroartemisinin-piperaquine versus intermittent preventive treatment with sulfadoxine-pyrimethamine for the control of malaria during pregnancy in western Kenya: an open-label, three-group, randomised controlled superiority trial. Lancet. 2015;386:2507–19.

    Article  CAS  PubMed  Google Scholar 

  14. Naidoo I, Roper C. Mapping ‘partially resistant’, ‘fully resistant’, and ‘super resistant’ malaria. Trends Parasitol. 2013;29:505–15.

    Article  PubMed  Google Scholar 

  15. Minja DT, Schmiegelow C, Mmbando B, Bostrom S, Oesterholt M, Magistrado P, et al. Plasmodium falciparum mutant haplotype infection during pregnancy associated with reduced birthweight, Tanzania. Emerg Infect Dis. 2013;19:9.

    Article  Google Scholar 

  16. Gonzalez R, Mombo-Ngoma G, Ouedraogo S, Kakolwa MA, Abdulla S, Accrombessi M, et al. Intermittent preventive treatment of malaria in pregnancy with mefloquine in HIV-negative women: a multicentre randomized controlled trial. PLoS Med. 2014;11:e1001733.

    Article  PubMed  PubMed Central  Google Scholar 

  17. Pfizer, London School of Hygiene and Tropical Medicine, Medicines for Malaria Venture. Evaluate azithromycin plus chloroquine and sulfadoxine plus pyrimethamine combinations for intermittent preventive treatment of falciparum malaria infection in pregnant women in Africa. 2010–2013. Clinical Trials NCT01103063 (Unpublished).

  18. Kakuru A, Jagannathan P, Muhindo MK, Natureeba P, Awori P, Nakalembe M, et al. Dihydroartemisinin-piperaquine for the prevention of malaria in pregnancy. N Engl J Med. 2016;374:928–39.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Tagbor H, Bruce J, Agbo M, Greenwood B, Chandramohan D. Intermittent screening and treatment versus intermittent preventive treatment of malaria in pregnancy: a randomised controlled non-inferiority trial. PLoS One. 2010;5:e14425.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. Tagbor H, Cairns M, Bojang K, Coulibaly SO, Kayentao K, Williams J, et al. A non-inferiority, individually randomized trial of intermittent screening and treatment versus intermittent preventive treatment in the control of malaria in pregnancy. PLoS One. 2015;10:e0132247.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Mauskopf JA, Paul JE, Grant DM, Stergachis A. The role of cost-consequence analysis in healthcare decision-making. Pharmacoeconomics. 1998;13:277–88.

    Article  CAS  PubMed  Google Scholar 

  22. Piaggio G, Elbourne DR, Altman DG, Pocock SJ, Evans SJ, Consort Group. Reporting of noninferiority and equivalence randomized trials: an extension of the CONSORT statement. JAMA. 2006;295:1152–60.

    Article  CAS  PubMed  Google Scholar 

  23. World economic outlook database. Consumer price index by country for 2012. http://www.imf.org/external/pubs/ft/weo/2012/02/weodata/index.aspx. Accessed 10 Oct 2015.

  24. Historical currency converter. http://www.oanda.com/currency/historical-rates/. Accessed 10 Oct 2015.

  25. Salomon JA, Vos T, Hogan DR, Gagnon M, Naghavi M, Mokdad A, et al. Common values in assessing health outcomes from disease and injury: disability weights measurement study for the Global Burden of Disease Study 2010. Lancet. 2012;380:2129–43.

    Article  PubMed  Google Scholar 

  26. Global burden of disease. 2004 update: disability weights for diseases and conditions. http://www.who.int/healthinfo/global_burden_disease/GBD2004_DisabilityWeights.pdf. Accessed 10 Oct 2015.

  27. Briggs A, Sculpher M, Claxton K. Decision modelling for health economic evaluation. Oxford: Oxford University Press; 2006.

    Google Scholar 

  28. WHO. Investing in health research and development: report of the ad hoc committee on health research relating to future intervention options. Geneva: World Health Organization; 1996.

    Google Scholar 

  29. WHO. Choosing interventions that are cost effective (WHO-CHOICE): cost effectiveness thresholds. http://www.who.int/choice/costs/CER_thresholds/en/. Accessed 10 Oct 2015.

  30. Radeva-Petrova D, Kayentao K, ter Kuile FO, Sinclair D, Garner P. Drugs for preventing malaria in pregnant women in endemic areas: any drug regimen versus placebo or no treatment. Cochrane Database Syst Rev. 2014;10:CD000169.

    PubMed  Google Scholar 

  31. Naidoo I, Roper C. Following the path of most resistance: dhps K540E dispersal in African Plasmodium falciparum. Trends Parasitol. 2010;26:447–56.

    Article  CAS  PubMed  Google Scholar 

  32. Wang H, Dwyer-Lindgren L, Lofgren KT, Rajaratnam JK, Marcus JR, Levin-Rector A, et al. Age-specific and sex-specific mortality in 187 countries, 1970–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380:2071–94.

    Article  PubMed  Google Scholar 

  33. Price RN, Simpson JA, Nosten F, Luxemburger C, Hkirjaroen L, ter Kuile F, et al. Factors contributing to anemia after uncomplicated falciparum malaria. Am J Trop Med Hyg. 2001;65:614–22.

    CAS  PubMed  PubMed Central  Google Scholar 

  34. Marchant T, Willey B, Katz J, Clarke S, Kariuki S, ter Kuile F, et al. Neonatal mortality risk associated with preterm birth in East Africa, adjusted by weight for gestational age: individual participant level meta-analysis. PLoS Med. 2012;9:e1001292.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Sicuri E, Bardaji A, Nhampossa T, Maixenchs M, Nhacolo A, Nhalungo D, et al. Cost-effectiveness of intermittent preventive treatment of malaria in pregnancy in southern Mozambique. PLoS One. 2010;5:e13407.

    Article  PubMed  PubMed Central  Google Scholar 

  36. Brabin BJ, Hakimi M, Pelletier D. An analysis of anemia and pregnancy-related maternal mortality. J Nutr. 2001;131:604S-14 (discussion 614S-15).

    Google Scholar 

  37. Measure DHS Statcompiler. http://www.statcompiler.com/. Accessed 28 Nov 2015.

Download references

Authors’ contributions

KH was responsible for the economic study supervision and oversaw the design and conduct of the economic study. SF, KH and ES developed the model concept and design. SF and ES had full access to all data in the model and take full responsibility for integrity of the data and the model as well as accuracy of all analysis. SF, ES, DH, JA, KB, DC, SC, SID, KK, FtK, PM, HT, JW, AW, MC, BG were involved in the data collection or acquisition of data from other sources. SF, KH, ES participated in the data analysis and interpretation. SF and KH drafted the manuscript. ES, DH, JA, KB, DC, SC, SID, KK, FtK, PM, HT, JW, AW, MC, BG and KH revised the manuscript critically and contributed to redrafting. FtK and KH obtained the funding for this study. All authors read and approved the final manuscript.

Acknowledgements

We are very grateful to the Economic Evaluation group at the LSHTM for valuable discussions about the model and methods; and to our fieldworkers in Ghana, Mali and Malawi.

Competing interests

The authors declare that they have no competing interests.

Availability of data and material

We are happy to make our spreadsheet available, but are still looking into different options.

Ethics approval and consent to participate

All cost data collection was approved by the Ethics Committee of the London School of Hygiene and Tropical Medicine (Reference Number 5590 and 6543) and by the relevant ethics committees in the countries where cost data were collected, which was Ghana, Mali and Malawi. Informed consent was obtained from each participant being observed.

Funding

This work was funded by the Malaria in Pregnancy Consortium, which is funded through a grant by the Bill and Melinda Gates Foundation to the LSTM. MC was supported by a fellowship jointly funded by the UK Medical Research Council (MRC) and the UK Department for International Development (DFID) under the MRC/DFID Concordat agreement, and part of the EDCTP2 programme supported by the European Union. The funding institution had no role in the model design and development, data collection, analysis and interpretation; or preparation, review or approval of the manuscript.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Silke Fernandes.

Additional file

12936_2016_1539_MOESM1_ESM.docx

Additional file 1: Appendix S1. The additional file contains further detail on the methods used in the analysis and additional results.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/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://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Fernandes, S., Sicuri, E., Halimatou, D. et al. Cost effectiveness of intermittent screening followed by treatment versus intermittent preventive treatment during pregnancy in West Africa: analysis and modelling of results from a non-inferiority trial. Malar J 15, 493 (2016). https://0-doi-org.brum.beds.ac.uk/10.1186/s12936-016-1539-4

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://0-doi-org.brum.beds.ac.uk/10.1186/s12936-016-1539-4

Keywords