Bulletin of the World Health Organization

Strategies for delivering insecticide-treated nets at scale for malaria control: a systematic review

Barbara A Willey a, Lucy Smith Paintain b, Lindsay Mangham a, Josip Car c & Joanna Armstrong Schellenberg b

a. Faculty of Public Health and Policy, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, England.
b. Faculty of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, England.
c. School of Public Health, Imperial College London, London, England.

Correspondence to Barbara Willey (e-mail: barbara.willey@lshtm.ac.uk).

(Submitted: 24 August 2011 – Revised version received: 31 January 2012 – Accepted: 02 February 2012 – Published online: 06 July 2012.)

Bulletin of the World Health Organization 2012;90:672-684E. doi: 10.2471/BLT.11.094771

Introduction

Malaria continues to represent a major public health problem in areas of endemicity, with an estimated 225 million cases worldwide in 2009.1 The 2015 goals of the World Health Organization’s (WHO’s) Roll Back Malaria Partnership are to reduce global malaria cases by 75% from 2000 levels and to reduce malaria deaths to near zero through universal coverage by effective prevention and treatment interventions.1 Among other preventive interventions, WHO recommends the use of insecticide-treated nets (ITNs), particularly long-lasting insecticidal nets, which have been shown to be cost-effective,24 to reduce malaria episodes among children < 5 years of age (hereafter, “children under 5”) by approximately 50% and all-cause mortality by 17%.5,6 Universal coverage with ITNs is defined as use by > 80% of individuals in populations at risk.6 WHO recommends supplying ITNs without charge or with a high subsidy and using a combination of periodic mass campaigns and routine delivery channels to deliver ITNs at scale.6 Other strategies include supporting the existing commercial sector and distributing vouchers exchangeable for partially subsidized ITNs through retailers.7

In response to the Roll Back Malaria Partnership’s targets for universal coverage, considerable efforts have been made recently to scale up ITN delivery. However, there is still low coverage in many countries and a need to understand the lessons learnt from experiences of scaling up ITN delivery. We therefore conducted a systematic review to synthesize recent evidence on the delivery of ITNs (including long-lasting insecticidal nets) at scale in malaria-endemic areas by documenting and characterizing the strategies for delivering ITNs at scale (at the district level or higher); summarizing ITN ownership among households and ITN use among children under 5, stratified by measures of equity when possible; summarizing the reported cost or cost-effectiveness of different strategies; and synthesizing information on reported factors influencing delivery of ITNs at scale.

Methods

A systematic review was conducted to identify studies that reported on the delivery of ITNs (including long-lasting insecticidal nets) at scale. The findings reported here form part of a larger systematic review on the scale-up of WHO-recommended malaria control interventions.8 We used a definition of “scaling up” that characterized this activity as the expansion of a health intervention beyond the initial geographical area or population group covered.9,10 We considered “at scale” to be ITN delivery in at least one district or the equivalent lowest level of health service administration in a given country.

Search strategy

Medline (Ovid), EMBASE, CAB Abstracts, Global Health and Africa Wide databases were searched using subject heading classification terms and free-text words. The following categories were combined using the AND Boolean logic operator: malaria terms, ITN and long-lasting insecticidal net terms and scaling-up terms (Box 1, available at: http://www.who.int/bulletin/volumes/90/9/11-094771). Filters were used to limit the search to humans and to publication dates from January 2000 to December 2010. Relevant papers from the grey literature were identified by searching Eldis and WHOLIS databases and Roll Back Malaria, Malaria Consortium, Africa Malaria Network Trust, and The Global Fund to Fight AIDS, Tuberculosis and Malaria web sites. Citation data for identified papers were exported to EndNote (Thomson Reuters, Carlsbad, USA), where duplicates were removed.

Box 1. Ovid Medline search

1. (malaria* or severe malaria or plasmodium or Plasmodium falciparum or Plasmodium vivax).ot,tw,ab,fs,kw,ti,hw,nm.

2. Malaria/ or exp Malaria, Falciparum/ or Malaria, Cerebral/ or Malaria, Vivax/

3. Plasmodium ovale/ or Plasmodium falciparum/ or Plasmodium/ or Plasmodium malariae/ or Plasmodium vivax/

4. exp Anopheles/

5. 1 or 2 or 3 or 4

6. Mosquito Control/

7. Insect Vectors/

8. “Bedding and Linens”/

9. Mosquito Nets/

10. Insecticide-Treated Bednets/

11. exp Insecticides/

12. exp Pyrethrins/

13. DDT/

14. Housing/

15. Larva/

16. exp Anopheles/

17. exp Chemoprevention/

18. Sulfadoxine/

19. Pyrimethamine/

20. pregnancy complications, infectious/ or pregnancy complications, parasitic/

21. Infant/

22. exp Anti-malarials/

23. Diagnosis/

24. exp Microscopy/

25. exp Laboratories/

26. Diagnostic Tests, Routine/

27. Point-of-Care Systems/

28. exp Therapeutics/

29. exp Drug Therapy/

30. Artemisinins/

31. Amodiaquine/

32. Mefloquine/

33. exp Chloroquine/

34. Primaquine/

35. Insect Repellents/

36. Community Health Aides/

37. 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36

38. (LLIN* or long-last* net or (long-lasting adj5 net)).ot,tw,ab,fs,kw,ti,hw,nm.

39. (ITN* or insecticide-treat* net or insecticidal-treat* net or insecticide-net or insecticidal-net or bed-net or bednet or treated-net or mosquito-net).ot,tw,ab,fs,kw,ti,hw,nm.

40. (IRS or indoor-residual spray* or indoor-spray*).ot,tw,ab,fs,kw,ti,hw,nm.

41. (larvicid* or larval control or larvi* fish or environment* management or environment* control* or drain* or house-screen* or (mosquito-proof* adj5 house) or repellent* or insecticide-treat* veil or insecticide-treat* hammock or insecticide-treat* blanket or insecticide-treat* cloth*).ot,tw,ab,fs,kw,ti,hw,nm.

42. (IPT or IPTp or IPTi or IPTc or intermittent preventive treatment*).ot,tw,ab,fs,kw,ti,hw,nm.

43. (diagnosis or RDT* or rapid diagnos* test* or rapid test* or microscop* or laborator*).ot,tw,ab,fs,kw,ti,hw,nm.

44. (treatment or antimalaria* or artemisinin-combination treat* or artemisinin-combination therap* or artemether lumefantrine or artesunate or amodiaquine or mefloquine or chloroquine or primaquine).ot,tw,ab,fs,kw,ti,hw,nm.

45. (malaria control or malaria intervention* or vector control* or vector management).ot,tw,ab,fs,kw,ti,hw,nm.

46. (community health worker* or village health worker* or (home manag* adj5 malaria)).ot,tw,ab,fs,kw,ti,hw,nm.

47. 38 or 39 or 40 or 41 or 42 or 43 or 44 or 45 or 46

48. 37 or 47

49. (scale-up or scaling-up or at-scale or go* to-scale or large-scale or roll-out or universal coverage).ot,tw,ab,fs,kw,ti,hw,nm.

50. 5 and 48 and 49

51. limit 50 to (humans and yr = ”2000 -Current”)

Eligibility criteria

Screening was a two-stage process. First, two authors (BW and LSP) independently screened titles and abstracts to determine which papers should undergo full-text assessment for eligibility. Retained papers underwent full-text review (performed independently by BW and LSP) to determine whether they described studies that satisfied the following criteria: subjects resided in areas where Plasmodium falciparum and Plasmodium vivax are endemic; ITN delivery at scale was evaluated; ITN ownership among households, receipt by pregnant women and/or use among children under 5 was evaluated; and an individual or cluster-randomized controlled design, a nonrandomized design, a quasi-experimental design, a before-and-after design, an interrupted time series design or a cross-sectional design without temporal or geographical controls was used.1113 Papers meeting these criteria were termed “index papers”. In addition to documenting and characterizing the strategies for delivering ITNs at scale and summarizing ITN ownership among households and ITN use among children under 5, this review also aimed to summarize the reported cost or cost-effectiveness of different strategies and to synthesize information on reported factors influencing delivery of ITNs at scale. As such, we also included papers that described qualitative studies, case studies, process evaluations and cost-effectiveness studies that were linked to an index paper.

The reference lists from eligible papers were hand-searched for additional relevant citations. All data relevant to the review were extracted from final included papers into an Access database (Microsoft, Redmond, United States of America).

Analysis

The first objective was to document and characterize the strategies for delivering ITNs at scale and was guided by a framework adapted from Kilian et al.14 Strategies were characterized by target population, implementation scale, implementer type, user cost and implementation duration (Fig. 1).

Fig. 1. Characteristics of strategies for delivering insecticide-treated nets at scale
Fig. 1. Characteristics of strategies for delivering insecticide-treated nets at scale
Source: Adapted from Kilian et al.14

The effectiveness of ITN delivery strategies was not compared using meta-analysis because study designs were too variable.15 Rather, narrative synthesis with a Best Evidence Synthesis approach was used to summarize findings and compare results across the different delivery strategies.16,17

The extent to which ITN ownership or use changed over time and whether such changes were attributable to the delivery strategy were assessed according to study quality. The quality of studies with a randomized or nonrandomized control group and of those using an interrupted time-series design was assessed using the Cochrane risk of bias checklist15 and Grading of Recommendations Assessment, Development and Evaluation (GRADE) criteria.18

All reported costs were adjusted for inflation by two authors (LSP and LM) and are presented as 2010 United States dollars (US$) using the consumer price indices available from the International Monetary Fund.19 When possible, costs are reported separately as financial (i.e. monetary) costs or economic costs (including opportunity costs and costs of donated goods and services).

Content analysis and narrative synthesis were used to identify important influences on delivering ITNs at scale and themes were assessed across the different ITN delivery strategies.16,17

Results

Fig. 2 details the literature search and screening process, performed according to guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Group.20 We included 32 papers that described 20 studies from 12 African nations (Burkina Faso, Eritrea, Ghana, Kenya, Madagascar, Malawi, Niger, Nigeria, Togo, Uganda, the United Republic of Tanzania and Zambia) and one partially autonomous region (Zanzibar). Six studies were implemented on a national level, two on a regional scale and 12 at the district level (of which three took place in only one district). Fourteen studies delivered ITNs only to children under 5 and/or pregnant women (Table 1 and Table 2, both available at: http://www.who.int/bulletin/volumes/90/9/11-094771).

Fig. 2. Flow of selection process for inclusion of studies of strategies for scaling up delivery of insecticide-treated nets (ITNs) for malaria control in areas with endemicity for Plasmodium falciparum and Plasmodium vivax malaria
Fig. 2. Flow of selection process for inclusion of studies of strategies for scaling up delivery of insecticide-treated nets (ITNs) for malaria control in areas with endemicity for <em>Plasmodium falciparum</em> and <em>Plasmodium vivax</em> malaria
ACT, artemisinin combination treatment; IPTp, intermittent preventive treatment in pregnant women; IVC, integrated vector control; LLIN, long-lasting insecticidal net.

Strategies for delivering ITNs at scale

Fig. 3 summarizes the characteristics of the strategies used to deliver ITNs at scale using the categories presented by Kilian et al.14 Routine health services, retailers and community-based agents were used to deliver ITNs on a continuous basis. Time-limited strategies either integrated the distribution of ITNs with a public health campaign or delivered ITNs through a stand-alone campaign. Most continuous strategies partially subsidized the delivery of ITNs, whereas all time-limited strategies fully subsidized delivery of ITNs. Most strategies that used routine health services targeted pregnant women or children under 5. All strategies involving time-limited integrated campaigns and stand-alone campaigns targeted children under 5, whereas strategies using retailers and community-based delivery provided ITNs to the general population. Seven studies used a combination of strategies.

Fig. 3. Equity ratios and prevalence of household ownership of insecticide-treated nets (ITNs) and use among children aged < 5 years in areas with endemicity for Plasmodium falciparum and Plasmodium vivax malaria, by delivery strategya,3,2125,2730,3235,3951
Fig. 3. Equity ratios and prevalence of household ownership of insecticide-treated nets (ITNs) and use among children aged < 5 years in areas with endemicity for <em>Plasmodium falciparum</em> and <em>Plasmodium vivax</em> malaria, by delivery strategy<sup>a,</sup><sup>3</sup><sup>,</sup><sup>21</sup><sup>–</sup><sup>25</sup><sup>,</sup><sup>27</sup><sup>–</sup><sup>30</sup><sup>,</sup><sup>32</sup><sup>–</sup><sup>35</sup><sup>,</sup><sup>39</sup><sup>–</sup><sup>51</sup>
a Studies may appear in more than one category if multiple strategies were used to deliver ITNs at scale or if strategies changed over time.b Value of 1 indicated equitable distribution, marked on figure by a red dashed line. Ratios > 1 suggest that the poorest quintiles were favoured.c Campaign integrated with measles vaccination campaign.d Campaign integrated with polio vaccination campaign.e Campaign integrated with mass drug administration for lymphatic filariasis (LF) campaign.

Studies with high ITN ownership or use

Eighteen studies reported ITN ownership among households and/or ITN use among children under 5 (Table 2). ITN ownership among households ranged from 1.3% to 94% and ITN use among children under 5, which is typically lower than the prevalence of household ITN ownership, ranged from 12% to 94%. Ten studies reported a high prevalence of ITN ownership or use during at least one survey conducted after initiation of the ITN delivery strategy. Six reported ownership by > 60% of households,25,4244,47,48,50,51 two reported ownership by > 80% of households29,30,49 and two reported use by ≥ 87% of children under 5.27,41

Of the six studies reporting ownership by > 60% of households, four used an uncontrolled cross-sectional survey design, surveying 300–3000 households 1–3 years after delivery began.25,44,50,51 The other two used a before-and-after design in which approximately 2500 households were surveyed before and one year after ITN delivery during campaigns integrated with measles vaccination.42,43,47,48 During the 1–2-year period between baseline and endline surveys, ITN ownership among households increased from 24.5% to 79% in one study and from < 1% to 55–70% in the other.

The two studies reporting ITN ownership by > 80% of households were uncontrolled cross-sectional surveys. A total of 475 households in Ghana29,30 and 2074 households in Zambia49 were surveyed five months and six months, respectively, after ITN delivery campaigns. ITN ownership in Ghana was 90%, whereas ownership in Zambia was 88% in rural areas and 82% in urban areas. In Ghana, a follow-up survey conducted 38 months after the initial survey revealed that ownership among households had decreased by 18%, to 74%.

Both studies reporting a high prevalence of ITN use among children under 5 also had an uncontrolled cross-sectional design. A total of 378 households in the Adjumani district of Uganda were surveyed 5–7 months after distribution of partially subsidized ITNs to pregnant women through antenatal care clinics27 and 264 households in the North A district of Zanzibar were surveyed 5 months after ITN delivery during a stand-alone ITN campaign.41 Responses revealed use by 94% of children under 5 in households surveyed in the Adjumani district and by 87% of children under 5 in the North A district.

All 10 studies that reported a high prevalence of ITN ownership or use provided fully subsidized ITNs through at least one component of their delivery strategy (Fig. 3 and Table 2). Seven studies provided fully subsidized ITNs through a stand-alone campaign only (in one41) or through an integrated campaign only (in six4244,4751). One study considered the continuous delivery of free ITNs through antenatal clinics.25 Two studies evaluated combined strategies.27,29,30 In one, ITNs were delivered to pregnant women through antenatal clinics on a continuous basis by use of a partially subsidized voucher system and to children under 5 through a campaign integrated with measles vaccination, at full subsidy.29,30 In the other, ITNs were delivered under a full subsidy to pregnant women through antenatal clinics on a continuous basis and for free to children under 5 during a stand-alone campaign on a time-limited basis.27

Equity of ITN ownership and use

Thirteen studies reported coverage stratified according to socioeconomic status as a measure of equity (Table 2). One study evaluated equity on the basis of urban and rural residence and twelve studies evaluated it on the basis of a household asset index. Of the latter studies, three reported a concentration index and nine reported an equity ratio. A concentration index ranges from −1 to 1, with a value of 0 indicating equitable distribution and values > 0 indicating inequitable distribution benefiting the least poor group. An equity ratio measures the equity of distribution in the poorest quintile relative to that in the least poor quintile, with a value of 1 indicating equitable distribution and values between 0 and 1 indicating inequitable distribution benefiting the least poor group.

The study that evaluated equity in terms of urban and rural residence was based on data from a national survey performed after partially subsidized delivery of ITNs to pregnant women and children under five at health centres.28 The survey found greater use among children under 5 in urban areas, compared with those in rural areas (51% versus 17%).

Three studies presented the concentration index of ITN ownership among households or ITN use among children under 5. The concentration index in each revealed higher ITN ownership or use among the least poor groups. One study had a quasi-experimental design and evaluated continuous delivery of partially subsidized ITNs through health care facilities.21 The other two used a cross-sectional design to assess the fully subsidized delivery of ITNs during a stand-alone campaign27 or during a campaign integrated with measles vaccination.46

Nine studies presented the equity ratio, or sufficient data for its calculation, of ITN ownership among households or ITN use among children under 5 (Fig. 3). The highest ownership was reported in the poorest quintile in four campaigns that integrated the delivery of free ITNs with measles vaccination. Two of the four used a cross-sectional design to evaluate strategies at either the national or district levels.44,49 The other two used a before-and-after design and also reviewed delivery at the district or national levels.42,43,47,48 The change in equity index was available only for one of the before-and-after studies and involved a decrease from 1.2 to 1.1.42,43

ITN use was similar across quintiles in two studies, both of which used an uncontrolled cross-sectional survey design of delivery at the district level. The strategy evaluated in one delivered ITNs during a stand-alone campaign.41 The other investigated a combined strategy involving delivery of fully subsidized ITNs to children under 5 through a campaign integrated with measles vaccination and partially subsidized ITNs to pregnant women through antenatal clinics.29,30

In five studies, ITN ownership or use was higher in the least poor quintile. Three studies evaluated the delivery of free ITNs to children under 5 through a campaign integrated with polio or measles vaccination in Niger, in Lindi region of the United Republic of Tanzania, and four rural districts of Zambia (Chilubi, Kaputa, Mambwe and Nyimba).45,49,50 In the fourth study, the delivery of partially subsidized ITNs to pregnant women via antenatal care clinics in the United Republic of Tanzania was examined.3235 The fifth study reviewed a stand-alone ITN campaign involving distribution of fully subsidized nets to children under 5 in the Micheweni district of Zanzibar.41

Study quality

Table 2 shows the variety of study designs used to assess ITN delivery strategies. Of the 18 studies reporting data on ITN ownership among households and ITN use among children under 5, the study design in two (a cluster-randomized controlled trial2224 and a quasi-experimental study without randomization21) involved comparison areas, and the study design in four involved a temporal comparison. Two of the studies with a temporal comparison evaluated time-limited delivery of fully subsidized ITNs42,43,47,48 and two analysed continuous delivery of partially subsidized ITNs.3,36,3840 As such, the interpretation of ITN ownership among households and ITN use among children under 5 between survey years varies by study design and delivery strategy.

Only the cluster-randomized controlled trial directly compared different delivery strategies.2224 One strategy involved subsidized sale, promoted by social marketing, of ITNs to the general population plus free distribution of long-lasting insecticidal nets to pregnant women at antenatal care clinics. The other strategy involved only subsidized sale, promoted by social marketing, of ITNs to the general population through retailers. Ownership of ITNs was 35% in the dual-intervention arm and 23% in the retail-only arm (P < 0.001). Although the risk of bias was low in this study, the quality of the evidence was downgraded from high to moderate on the basis of the GRADE criteria because it was unclear whether analyses adjusted for the clustered design and because no relative measure of effect was provided.

One study described the delivery of partially subsidized ITNs at the district level through sales by health facility staff.21 ITN ownership was 14% in three intervention districts, compared with 1.3% in two comparison districts (P < 0.001). The risk of bias in this study was moderate principally because of the lack of randomization. The quality of evidence was very low on the basis of the GRADE criteria because there were important differences between intervention and comparison areas at baseline (e.g. socioeconomic status) that were not adjusted for in the analysis and because no relative measure of effect was provided.

In nonrandomized studies, identification of the channel through which the ITN is delivered (i.e. antenatal clinics or retail shops) may help determine whether the change in coverage achieved can be allocated to the delivery strategy.12 Studies in three countries did not stratify ITN ownership by delivery channel.3,3640,47,48 However, elsewhere, a decline in the proportions of unsubsidized ITNs sourced from retailers and partially subsidized ITNs sourced from maternal and child health clinics was seen among children under 5.42,43 Both decreases occurred after initiation of an integrated campaign in 2006 to distribute fully subsidized ITNs, with the campaign contributing almost half of the ITNs used by children under 5 surveyed during 2006–2007.

Costs

Ten studies reported on the cost or cost-effectiveness of ITNs (Table 3). Of these, seven described only cost per ITN delivered or cost per treated-net–year. The remaining three were cost-effectiveness studies that also presented cost per death or per disability-adjusted life year averted. All except one of the economic evaluation studies conducted sensitivity analyses around the major cost and outcome parameters.

Four studies investigated the cost of delivering free ITNs through antenatal care clinics, with three at the district level and one at the national level. In the district-level studies, financial costs ranged from US$ 8.20 to US$ 10.54 per ITN delivered22,26,27 and economic costs ranged from US$ 5.47 to US$ 5.89 per ITN delivered.22,27 The study at the national scale reported an economic cost of US$ 10.77 per ITN delivered.25,35

Of the four studies that evaluated the delivery cost of partially subsidized ITNs, three investigated delivery through the retail sector and one investigated voucher use. Studies of retail-based delivery reported financial costs of US$ 5.47 and US$ 11.16 per ITN delivered in Burkina Faso and Malawi, respectively, and of US$ 12.57 and US$ 18.72 per treated-net–year in the United Republic of Tanzania and Malawi, respectively.3,22,36 The studies in Burkina Faso and the United Republic of Tanzania were at the district level and the study in Malawi was at the national level; the length of protection afforded by ITNs in calculations of cost per ITN delivered was assumed to be 12 months in Burkina Faso and 6 months in Malawi. The fourth study investigated the Tanzanian National Voucher Scheme and found economic costs of US$ 10.77 per ITN delivered and US$ 6.02 per treated-net–year, with the latter calculation assuming 12-month protection from a treated net.35

The four studies that evaluated fully subsidized campaigns found financial costs per ITN delivered of US$ 3.71 to US$ 11.79 for those integrated with vaccination campaigns30,45,49 and US$ 9.48 for a stand-alone campaign.27 The stand-alone campaign considered in one of the studies had an economic cost per ITN delivered of US$ 4.76.27

Three studies presented some measure of health impact. The economic cost per child death averted was US$ 1242 for a national voucher scheme35 and US$ 2924 for a retail sector programme involving partially subsidized delivery.36 The economic cost per disability-adjusted life year averted was similar, at US$ 100 and US$ 107.25,36

Cost or cost-effectiveness estimates were most sensitive to the assumed ITN lifespan (i.e. physical viability and duration of insecticide protection) and the proportion of ITNs actually used (leakage). The main cost associated with ITN delivery programmes was the ITNs themselves, most often followed by staff and transport.

Factors influencing ITN delivery

Information on factors influencing delivery of ITNs at scale was available for 12 of 20 studies (Table 4). Important perceived influences on the delivery of ITNs at scale, from the perspective of actors involved, were categorized into those at the user level, the implementer or health system level and the policy level.52

Facilitators at the implementation level included provision of training and appropriate supervision and support. At the policy level, facilitators included involvement of relevant stakeholders during planning and implementation and cooperation across ministries, departments and sectors (e.g. health and retail). Several barriers were identified, including costs to users for partially subsidized strategies, variation in implementation due to insufficient supplies of ITNs and vouchers and to poor communication and adherence to distribution procedures, and, at the policy level, financial resources to sustain current and future distribution strategies.

Discussion

Strategies frequently used to deliver ITNs at scale reported in the published and grey literature include continuous delivery of partially subsidized ITNs through the health sector and retail outlets, continuous delivery of free ITNs though antenatal care clinics and time-limited delivery of free ITNs, either alongside other public health goods (usually vaccines) during integrated campaigns or through stand-alone ITN campaigns. Few experiences with continuous delivery by community-based agents were recorded. The majority of strategies delivered to a targeted population of children under 5 or pregnant women. Seven studies from six countries described multiple concurrent or sequential delivery strategies, particularly continuous strategies in combination with a time-limited campaign.

These studies showed wide variability in ITN ownership among households and ITN use among children under 5. Although findings of high ownership or use were largely drawn from uncontrolled studies, strategies reviewed in the majority of studies included at least one component that delivered ITNs at a full subsidy. The majority of equity evidence was from uncontrolled studies: in general, strategies that used time-limited delivery of fully subsidized ITNs were equitable or pro-poor, in contrast to strategies that used continuous delivery of partially subsidized ITNs. No equity evidence from fully subsidized continuous strategies was available.

Comparisons of costs and cost-effectiveness across these strategies are challenging because of variations in the methods of economic analysis used and in the scale of delivery, as emphasized previously.53 Nonetheless, the cost of delivering ITNs across the strategies was reasonably comparable. The main cost was the ITNs themselves, a cost frequently supported by donor funding, and all of the cost-effectiveness estimates were most sensitive to ITN lifespan and proportion of ITNs actually used.

This review aimed to synthesize details on the context of, barriers to and facilitators of strategies to deliver ITNs at scale, some of which were implemented under near-programmatic conditions. Important factors influencing the delivery of ITNs at scale were similar across delivery strategies. Barriers involving cost were common at the user level, whereas barriers involving stock-outs and poor logistics for ITN procurement and transport were common at the implementer level. Training and supervision of staff was often highlighted as a facilitator at the implementer level and cooperation across departments or ministries and stakeholder involvement were highlighted at the policy level.

The evaluation of large-scale health programmes has been highlighted as a “top priority in global health”54 and researchers have emphasized that the use of randomized designs for such evaluation may be inappropriate because of low external validity.11,55 Therefore, to characterize the full breadth of ITN delivery strategies and to synthesize evidence that corresponded to the conditions under which large-scale ITN delivery may occur in practice, we included a variety of study designs.56 However, this made interpretation of findings challenging, particularly because a before-and-after study of a campaign conducted at a single time point is qualitatively different from annual surveys conducted during a continuous distribution strategy.

The Medical Research Council recommends that the evaluation of complex interventions include information on the context and implementation of interventions. Our experience in conducting this review suggests that future synthesis of evidence involving large-scale delivery of complex public health interventions would benefit from improved consistency of reporting of the implementation process by included studies.57,58 Recommendations for reporting are available from the Transparent Reporting of Evaluations with Nonrandomized Designs (TREND) statement.59

It is simplistic to interpret the findings of this review as providing a single recommendation to policy-makers on which ITN delivery strategy to adopt. Rather, the review highlights that choosing among alternatives depends on contextual factors, such as the epidemiologic characteristics of malaria, attributes of health systems and contextual constraints. Moreover, the review demonstrates how a framework for characterizing delivery strategies can prove useful in synthesizing evidence, which may help policy-makers formulate implementation strategies to deliver ITNs to populations in their local settings.


Acknowledgements

We thank Mark Petticrew and Neil Spicer (Faculty of Public Health and Policy, London School of Hygiene and Tropical Medicine) and Rifat Atun (Imperial College London, formerly with The Global Fund to Fight AIDS, Tuberculosis and Malaria) for helpful comments.

Funding:

This review was supported by the Alliance for Health Services and Policy Research, World Health Organization, which commissioned this work as a background paper for the First Global Symposium on Health Systems Research.

Competing interests:

None declared.

References

Share