Deployment of community health workers across rural sub-Saharan Africa: financial considerations and operational assumptions
Gordon C McCord a, Anne Liu b & Prabhjot Singh c
a. University of California, San Diego, United States of America (USA).
b. Earth Institute, Columbia University, New York, USA.
c. School of International and Public Affairs, Columbia University, 475 Riverside Drive, Suite 825, New York 10115, USA..
Correspondence to Prabhjot Singh (e-mail: email@example.com).
(Submitted: 08 July 2012 – Revised version received: 28 October 2012 – Accepted: 26 December 2012 – Published online: 13 February 2013.)
Bulletin of the World Health Organization 2012;91:244-253B. doi: http://dx.doi.org/10.2471/BLT.12.109660
The evidence is overwhelming that community-based interventions are an effective platform for extending health care delivery and improving health outcomes. Such evidence indicates that a well-implemented community-based health programme can: (i) reduce infant and child mortality and morbidity; (ii) improve health-care-seeking behaviour (e.g. increase rates of institutional delivery and immunization); and (iii) provide low-cost interventions for common maternal and paediatric health problems while improving the continuum of care.1–3 Such community-level programmes can be particularly effective for addressing the most common causes of paediatric mortality and morbidity, such as pneumonia, diarrhoea, undernutrition, malaria, human immunodeficiency virus (HIV) infection and acquired immunodeficiency syndrome (AIDS) and measles.4–11 These community-based health programmes are often successfully executed through community health workers (CHWs). These are lay people who live in the communities they serve and who function as a critical link between those communities and the primary-health-care system.12 The well-documented success of CHW programmes over the last few decades has increasingly pushed investment in CHW subsystems to national and international policy platforms as part of coordinated efforts to improve health-care systems.13
Although several reviews have documented the positive impact of CHWs on health outcomes, CHW programme costs have barely been examined, probably because cost data are much less widely available than data on programme outcomes. In one revealing review of 53 studies on CHW programmes in the United States of America, only six studies referenced costs and the authors considered such data to be insufficient to draw any conclusions.14 When included, costs are generally associated with interventions rather than with comprehensive CHW programmes; furthermore, estimates do not account for economies of scale or year-on-year efficiencies.15 To date, the official costs of national CHW programmes in pioneering countries such as Ethiopia, Malawi or Rwanda have not been estimated, partly because tracking unit costs is difficult and because methods for isolating the CHW subsystem from an integrated primary-health-care system have been elusive.
This paper provides cost guidance for one adaptable configuration of a CHW “subsystem”: a provider system housed within a larger primary-health-care system that includes clinics and referral hospitals. Costing is done by function (e.g. diagnosing and treating malaria) and by local epidemiologic characteristics (e.g. each country’s prevalence of HIV infection), so that components and assumptions can be easily modified. National scale-up of CHW programmes and of primary-health-care systems more broadly is likely to reduce the incidence of many of the diseases discussed in this paper. This model allows costs to be easily recalculated as incidence rates change. New functions, such as the care of patients with chronic conditions, could be added and costed once a vetted CHW protocol for these functions has emerged.
Model community health worker system
We designed the model CHW system presented in this paper using the health system building blocks of the World Health Organization (WHO) (Table 1), empirical evidence and lessons learnt from several global CHW programmes, and data from the Millennium Villages Project. In this costing model we focus on full-time, paid public sector CHWs who are formally recognized as an integral part of the rural primary-health-care system and who perform their work primarily in the community and through household visits.
The CHWs in this subsystem visit households to provide community case management for diarrhoea, malaria and malnutrition, management of pregnancy and health promotion. This subsystem is supported by robust supervision, information-driven management and a provider network within the health system. This model is also based on the assumption that CHWs receive short training and strong supervision and that the health system is strong. In the particular model costed, we also assumed that an mHealth platform was used to collect data and provide decision support, given the increasing popularity of mHealth programmes in primary care. Detailed inputs, processes and outputs from this framework can be found in the One Million Community Health Workers report.16
This costing exercise also illustrates one possible additional configuration in which a subset of “generalist” CHWs is further trained to work closely with skilled birth attendants to enhance maternal care. Although this CHW subset relies on a skilled clinical team member for supporting pregnancy and childbirth, for the sake of discussion we herein refer to these CHWs as “childbirth specialists”.
Cost inputs for household services
We estimated the cost of fully deploying CHW subsystems across the rural areas of sub-Saharan Africa and develop a costing framework for countries to refine and adapt in their own national planning. Fig. 1, Fig. 2 and Fig. 3 illustrate the algorithm for costing, including junctures (represented by green boxes) requiring data on population, disease incidence and supply and commodity costs. Unless otherwise indicated, unit costs were provided by the Millennium Villages Project.
Fig. 1. Costing algorithm for generalist community health workers (CHWs)
Fig. 2. Costing algorithm for childbirth specialist community health workers (CHWs)
Fig. 3. Costing algorithm for community health worker (CHW) referral system costs
Costs per community health worker
- Backpack and mobile phone: Each CHW is given a mobile smartphone costing US$ 150, a solar charger costing US$ 40 and a data plan costing US$ 480 per year for supervisory support, communication and data collection. To reduce costs, national CHW programmes can establish closed user groups – a nationwide network that allows free calls and short message service between identified health staff. Each CHW is also given a backpack costing US$ 10 to carry supplies.
- Training: CHWs receive one year of training: 3 months in a classroom and 9 months while in service, plus supervision and feedback. They also receive a yearly refresher course lasting 20 hours. In the Millennium Villages Project, training 50 CHWs for 80 hours costs US$ 300. This includes transport, meals, supplies and trainer’s honorarium, as necessary. We used this figure to compute an hourly cost. Following the experience of Pakistan’s national Lady Health Worker programme, we assumed a 5% student attrition rate during training.17
- Salaries: In the costing model, each CHW receives a monthly salary of US$ 80. This is the average CHW salary across Millennium Village Project settings in 10 countries. This figure will fluctuate depending on national rural wage levels.
- Management: Each CHW Manager is assumed to manage 30 CHWs and receives an annual salary of US$ 9600. This 10-fold difference between CHW and CHW manager salaries represents the average difference across Millenium Village sites. Note that CHW managers could be physicians in primary-health-care facilities with supplemental training in management, which would parlay some of this cost into existing primary-health-care system budgets.
- Overhead: We added 15% to total programme costs in order to account for the overhead costs at the local, national and international levels of a global effort to reach full coverage, or around 1 million CHWs across low-income sub-Saharan Africa by 2015. This includes components of the operational design not listed as segregated costs, such as community engagement and information management.
Generalist CHWs are expected to generate various costs. The cost modelling we have conducted is based on country-specific disease prevalence rates (and subnational endemicity data in the case of malaria). If the prevalence of any disease drops, the resulting commodity expenditures will decrease accordingly.
- Monitoring for undernutrition: CHWs will use mid-upper arm circumference (MUAC) strips to monitor children between the ages of 1 and 5 years for undernutrition at least once every 90 days. We estimate a requirement of one MUAC strip per child per year, at a cost per strip of US$ 0.05.18
- Treating diarrhoea: CHWs will treat episodes of diarrhoea with oral rehydration salts (ORS) and zinc tablets. They should teach caretakers how to administer ORS and leave two packets in the household, along with 20-mg tablets, to be taken once daily for 14 days (half a tablet in the case of children younger than 6 months).19 We calculate the average cost per episode to be around US$ 0.42.18 Following published data,20 we assume that children under 5 years of age experience 5 diarrhoeal episodes annually, and that people older than 5 years’ experience 1.28. This leads to an estimated 1.1 billion diarrhoeal episodes per year in rural Africa. Notably, diarrhoeal disease is projected to be the most expensive health condition covered by CHWs: US$ 0.79 out of US$ 2.88 in supply and commodity costs per person served by the CHW subsystem. A reduction in diarrhoeal disease incidence would have important cost implications for CHW programmes.
- Testing for and treating malaria: In the absence of microscopy, WHO protocol dictates that all people with fever in malarious areas should undergo a rapid diagnostic test (RDT) before being treated with antimalarials. We used the United Nations Population Division’s World Population Prospects’ demographic breakdown for 2010 to divide the population in malarious zones into groups aged 0–4, 5–9, 10–14 and 14+ years. We assume that the 0–4 age group will experience two fever episodes per year; the 5–9 and 10–14 age groups, one fever episode per year, and the 14+ age group, 0.5 fever episodes per year. These figures are the average annual number of fever episodes from all causes, across all malaria endemicity levels. One RDT kit costing US $0.50 is needed for each of these fever episodes.18,21 We also calculate the costs of treating malaria with artemisinin-based combination therapy (ACT). We assume that RDT will reveal that around 30% of the fevers are due to malaria and require ACT. The total number of fever episodes requiring treatment for malaria by 2015 is estimated at 150 million, which is within the range of other published estimates. Each case can be treated with either artemether–lumefantrine or artesunate–amodiaquine, so we use the average cost. The costs per treatment course are as follows: for artemether–lumefantrine, US$ 0.45 for the 0–4 age group, US$ 0.90 for the 5–9 age group, US$ 1.35 for the 10–14 age group and US$ 1.80 for the 14+ age group; for artesunate–amodiaquine, US$ 0.23 for infants, US$ 0.45 for children aged 1–6 years, US$ 0.80 for children aged 7–13 years and US$ 1.48 for people older than 13. We used the dosages recommended by the Roll Back Malaria Partnership and average drug prices obtained from Sanofi-Aventis (Paris, France), Ipca (Mumbai, India) and Cipla (Mumbai), with adjustment to match the age groups in our analysis. The total projected expenditure is US$ 153 million in RDTs and US$ 87 million in ACTs, with national per capita expenditure varying in accordance with each country’s endemicity levels.
- Deworming: Using country-level data on the prevalence of ascariasis, trichuriasis and hookworm,22 we applied the following protocol: if the country-level prevalence of any of the three infestations is over 50%, all individuals receive three doses of albendazole (costing US$ 0.02 per dose) per year.18 If the prevalence is above 20% and below 50%, then everyone receives one dose of albendazole. For schistosomiasis, if the prevalence is above 50%, everyone older than 4 years is treated with praziquantel once a year.23 If the prevalence is above 10%, every eligible person is treated once every two years. If the prevalence is below 10%, only school-age children are treated twice during their school-aged years (once on entry and once on exit). Praziquantel costs US$ 0.22 a tablet,18 and since the number of tablets in the treatment varies from 1 to 5 depending on body weight, we assume an average of 2.5 tablets per treatment.
- Pneumonia: WHO reported 131.3 million pneumonia episodes in Africa in 2004.24 From this we calculate the number of episodes per person and, on the assumption that incidence in a given country remains constant, we estimate the number of pneumonia episodes in the population covered by CHWs each year. We include timers for each CHW to assess children’s respiratory rate; respiratory timers from the United Nations Children’s Fund cost US$ 3.50.18 Since they are good for 10 000 uses, we assume one timer per CHW per year. The average cost of antibiotics is US$ 0.27 per case.24
- Screening for tuberculosis: CHWs should collect sputum samples from suspected tuberculosis cases. We estimate the number of screenings using each country’s incidence of tuberculosis and assume a positivity rate of 10%. The cost of screening one person covers the collection of three sputum samples in containers (one on the first day and two on the second day). The containers are labelled with a marker and put in a reusable sealable plastic bag. The cost of each container is US$ 0.054. Each CHW is given two markers for the year and three reusable sealable plastic bags every 3 months. Finally, CHWs use two new pairs of surgical gloves for each screening, since the screening is conducted over two days. In the Bonsaaso Millennium Village in Ghana, markers are priced at US$ 10 for a pack of 8, plastic bags are priced at US$ 8 for a pack of 25 and surgical gloves are priced at US$ 0.60 a pair.
The childbirth specialist CHW would generate the same human resource expenditures (training, salary, management) as the generalist CHW, but different commodity costs:
- Screening pregnant women for HIV infection: All pregnant women will undergo at least one HIV test. Following Joint United Nations Programme on HIV/AIDS cost estimates for the prevention of mother-to-child transmission of HIV, we estimate a cost of US$ 3.90 for each HIV serological test. If a pregnant woman tests positive for HIV, the CHW will also test her husband and children (assumed to be 2, on average). HIV-positive individuals will be referred to the national AIDS programme for counselling and treatment.
Ambulances are included in the costing model as a necessary link between the CHWs and the primary-health-care system.
- Ambulances: One ambulance is assumed to cover a population of 50 000 people in a rural area. Based on experience in Millennium Villages, an ambulance costs an average of US$ 36 000. The ambulance driver receives an annual salary of US$ 4782. Ambulance fuel and maintenance costs average US$ 4825 per year across Millennium Villages.
In total, it will take an average of US$ 3750 annually to train, equip and support each CHW between 2012 and 2015. Maintenance of the CHW programme after 2015 will cost US$ 3150 per CHW. These costs do not include two potential CHW services: family planning and HIV screening for the general population. Since public health specialists have not reached consensus on whether CHWs should be tasked with these services, we have separated them from the costing framework above and consider them below:
- Family planning: The average cost of family planning in Africa, including contraceptives, has been estimated at around US$ 26.90 annually per woman of reproductive age (15–49 years).25 If generalist CHWs offer family planning services to women in this age group, the total cost comes to an average of US$ 2.4 billion per year, or US$ 6.36 per person serviced by the CHW programme.
- HIV screening in the general population: Studies have begun documenting the benefits of active case finding and outreach by CHWs.26 Health service provision is moving towards increased decentralization and at some point CHWs may be tasked with testing the general population for HIV. (Note that in the costing presented earlier, pregnant women and their families are already tested through childbirth specialist CHWs). Although the appropriate frequency of testing remains to be determined, we assume one annual test per person to get an idea of costs. Extending HIV testing once a year to all HIV-negative people older than 14 years costs an average of US$ 791 million per year, or US$ 2.12 per rural inhabitant.
Estimated costs of scale-up by 2015
The model ramps up CHW coverage evenly from 2012 to 2015: in 2012 one quarter of the rural population receives services, but by 2015 the entire rural population will be serviced. Note that in the costing model rural population and population growth are based on United Nations Population Division data, which come from national statistics offices with varying definitions of urban and rural areas. Thus, our model is based on each country’s definition of “rural” and “urban”. The number of CHWs per inhabitant will be different for each country, especially in light of differences in rural population density. For this reason, Fig. 4 shows programme costs by major category as a function of the number of rural residents per CHW, which varies from 500 to 1000 people.
Fig. 4. Average annual CHW costs as a function of population served
Table 2 (available at: http://www.who.int/bulletin/volumes/91/4/12-109660) shows the country-by-country breakdown of the number of CHWs required by 2015 and the average yearly costs for supplies, training, salaries, management and overhead for the combined workforce of generalist CHWs and childbirth specialist CHWs. At a rate of 1 generalist for every 650 rural Africans and of 1 childbirth specialist for every 3500 Africans, the average annual cost of CHW system ramp-up across Africa is around US$ 2.6 thousand million (billion), or US$ 6.86 per person covered by CHW services, or US$ 2.72 per inhabitant. The cost of training, equipping and supporting each CHW amounts to US$ 3750 per year.
Table 2. Average annual expenditure on the community health worker (CHW) programme (with 1 CHW for every 650 rural inhabitants)
Steady-state system costs listed in the last three columns of Table 2 include the training of new CHWs to keep up with rural population growth and refresher courses for all CHWs in the system. Steady-state estimates include capital costs such as phones, backpacks and ambulances; these are amortized over 3, 3 and 7 years, respectively.
Fig. 5 shows how expenditure is spatially distributed across rural Africa. To obtain the distribution we used the costs per rural inhabitant given earlier and 2.5 arc-minute GIS data on population and urban extents in 2010.27 We eliminated the urban population and calculated the average annual cost per grid cell. The map allows international and national policy-makers to see how CHW programme expenditures would be distributed within and across sub-Saharan African countries.
Fig. 5. Spatial distribution across sub-Saharan Africa of the annual cost of a community health worker programme
According to studies published by the Commission on Macroeconomics and Health, the United Nations Millennium Project and the International Task Force on Innovative Financing for Health, in a low-income country a primary-health-care system should cost from US$ 50 to US$ 55 per capita per year in 2011 prices.28–30 According to this model, the CHW subsystem configured in this paper would cost approximately 5% of the total cost of a primary-health-care system.
Although this framework provides an estimate of CHW-related costs across sub-Saharan Africa, country-specific financing models reflecting the local costs of training and deploying the CHWs should emerge, along with country-specific strategies to suit different CHW programme characteristics. Nigeria is currently planning to deploy a CHW subsystem nationally.
The effectiveness of a CHW subsystem depends on how well the rest of the health system functions. Although we excluded hospitals, clinics and non-CHW health staff from our costing model, they provide vital logistical support to CHWs by making available to them adequate supplies of medicines and possibilities for patient referral. When national CHW initiatives have been taken to scale without supporting infrastructure, retrospective analysis reveals weaknesses in management, supply chain, financing and other components of the supporting system.31
Depending on existing policies or disease prevalence, national CHW strategies may be designed to train CHWs to provide supplemental services beyond those described in this model. As countries engage in the financial and operational planning of their CHW programmes, model inputs will differ in accordance with the local context, as exemplified in Table 3. The adaptation of the model to country-specific needs will become increasingly important as ministries of health prepare their community-based health workforces for the management of the increasing burden of non-communicable diseases. However, the management of patients with these diseases by CHWs requires further study and operational detailing in low-resource settings, as well as a dedicated supplemental source of financing.
Table 3. Factors to be modified when adapting operational design of model community health worker (CHW) costing system to local context
In summary, we recommend that countries wishing to develop a CHW strategy perform a similar costing exercise to design programme budgets. According to previous studies, investment in a well-organized and comprehensive CHW subsystem that is embedded in a primary-health-care system can reduce maternal and child morbidity and mortality. Our assessment suggests that the costs of the core elements of such a system are a fraction of the costs of primary-health-care services overall. This costing exercise sets the stage for a costing framework that can be used to determine the cost–benefit and cost–effectiveness of CHW programmes.
- Lewin S, Munabi-Babigumira S, Glenton C, Daniels K, Bosch-Capblanch X, van Wyk BE, et al., et al. Lay health workers in primary and community health care for maternal and child health and the management of infectious diseases. Cochrane Database Syst Rev 2010CD004015 pmid: 20238326.
- Lehmann U, Sanders D. Community health workers: what do we know about them? The state of the evidence on programmes, activities, costs and impacts on health outcomes of using community health workers. In: Evidence and information for policy. Geneva: Department of Human Resources for Health, World Health Organization; 2007.
- Perry H, Freeman P, Gupta S, Rassekh BM. How effective is community-based primary care in improving the health of children? Washington: Community-Based Primary Health Care Working Group, American Public Health Association; 2009.
- Bhutta ZA, Darmstadt GL, Hasan BS, Haws RA. Community-based interventions for improving perinatal and neonatal health outcomes in developing countries: a review of the evidence. Pediatrics 2005; 115: 519-617 pmid: 15866863.
- Black RE, Morris SS, Bryce J. Where and why are 10 million children dying every year? Lancet 2003; 361: 2226-34 http://dx.doi.org/10.1016/S0140-6736(03)13779-8 pmid: 12842379.
- Claeson M, Gillespie D, Mshinda H, Troedsson H, Victora CG, Bellagio Study Group on Child Survival. Knowledge into action for child survival. Lancet 2003; 362: 323-7 http://dx.doi.org/10.1016/S0140-6736(03)13977-3 pmid: 12892965.
- Victora CG, Wagstaff A, Schellenberg JA, Gwatkin D, Claeson M, Habicht J-P. Applying an equity lens to child health and mortality: more of the same is not enough. Lancet 2003; 362: 233-41 http://dx.doi.org/10.1016/S0140-6736(03)13917-7 pmid: 12885488.
- Darmstadt GL, Bhutta ZA, Cousens S, Adam T, Walker N, de Bernis L, Lancet Neonatal Survival Steering Team. Evidence-based, cost-effective interventions: how many newborn babies can we save? Lancet 2005; 365: 977-88 http://dx.doi.org/10.1016/S0140-6736(05)71088-6 pmid: 15767001.
- Knippenberg R, Lawn JE, Darmstadt GL, Begkoyian G, Fogstad H, Walelign N, et al., Lancet Neonatal Survival Steering Team, et al. Systematic scaling up of neonatal care in countries. Lancet 2005; 365: 1087-98 pmid: 15781104.
- Martines J, Paul VK, Bhutta ZA, Koblinsky M, Soucat A, Walker N, et al., Lancet Neonatal Survival Steering Team, et al. Neonatal survival: a call for action. Lancet 2005; 365: 1189-97 http://dx.doi.org/10.1016/S0140-6736(05)71882-1 pmid: 15794974.
- Bhutta ZA, Ahmed T, Black RE, Cousens S, Dewey K, Giugliani E, et al., Maternal and Child Undernutrition Study Group, et al. What works? Interventions for maternal and child undernutrition and survival. Lancet 2008; 371: 417-40 http://dx.doi.org/10.1016/S0140-6736(07)61693-6 pmid: 18206226.
- Haines A, Sanders D, Lehmann U, Rowe AK, Lawn JE, Jan S, et al., et al. Achieving child survival goals: potential contribution of community health workers. Lancet 2007; 369: 2121-31 http://dx.doi.org/10.1016/S0140-6736(07)60325-0 pmid: 17586307.
- Hongoro C, McPake B. How to bridge the gap in human resources for health. Lancet 2004; 364: 1451-6 http://dx.doi.org/10.1016/S0140-6736(04)17229-2 pmid: 15488222.
- Viswanathan M, Kraschnewski JL, Nishikawa B, Morgan LC, Honeycutt AA, Thieda P, et al., et al. Outcomes and costs of community health worker interventions: a systematic review. Med Care 2010; 48: 792-808 http://dx.doi.org/10.1097/MLR.0b013e3181e35b51 pmid: 20706166.
- Bhutta ZA, Lassi ZS, Pariyo G, Huicho L. Global experiences of community health workers for delivering health related millennium development goals: a systematic review, country case studies, and recommendations for integration into national health systems. Geneva: WHO Global Health Workforce Alliance; 2010.
- Singh Pa. One million community health workers: technical task force report. New York: Earth Institute, Columbia University; 2011.
- Hunt S. Lady health worker programme: third party evaluation of performance. Oxford: Oxford Policy Management; 2009.
- United Nations Children’s Fund [Internet]. Supply catalogue. New York: UNICEF; 2012. Available from: https://supply.unicef.org/unicef_b2c/app/displayApp/(layout=7.0-12_1_66_67_115&carea=%24ROOT)/.do?rf=y [accessed 29 January 2013].
- United Nations Agency for International Development & United Nations Children’s Fund & World Health Organization. Diarrhoea treatment guidelines: including new recommendations for the use of ORS and zinc supplementation for clinic-based healthcare workers. Arlington: Management of Social Transformations Project; 2005.
- Bern C. Diarrhoeal diseases. In: Global epidemiology of infectious diseases. Geneva: World Health Organization; 2004.
- Teklehaimanot A, McCord GC, Sachs JD. Scaling up malaria control in Africa: an economic and epidemiological assessment. Am J Trop Med Hyg 2007; 77: 138-44 pmid: 18165486.
- de Silva NR, Brooker S, Hotez PJ, Montresor A, Engels D, Savioli L. Soil-transmitted helminth infections: updating the global picture. Trends Parasitol 2003; 19: 547-51 http://dx.doi.org/10.1016/j.pt.2003.10.002 pmid: 14642761.
- World Health Organization [Internet]. Neglected tropical diseases: PCT databank, schistosomiasis. Geneva: WHO; 2005. Available from: http://www.who.int/neglected_diseases/preventive_chemotherapy/sch/en/index.html [accessed 29 January 2013].
- Pneumonia: the forgotten killer of children. New York & Geneva: United Nations Children’s Fund & World Health Organization; 2006.
- Singh S, Darroch JE. Adding it up: costs and benefits of contraceptive services – estimates for 2012. Guttmacher Institute & United Nations Population Fund; 2012. Available from: http://www.guttmacher.org/pubs/AIU-2012-estimates.pdf [accessed 29 January 2013].
- Mukherjee JS, Eustache FE. Community health workers as a cornerstone for integrating HIV and primary healthcare. AIDS Care 2007; 19: S73-82 http://dx.doi.org/10.1080/09540120601114485 pmid: 17364390.
- Center for International Earth Science Information Network/Columbia University, International Food Policy Research Institute, The World Bank & Centro Internacional de Agricultura Tropical. Global Rural-Urban Mapping Project (GRUMP): urban/rural extents. Palisades: NASA Socioeconomic Data and Applications Center; 2004. Available from: http://sedac.ciesin.columbia.edu/data/set/grump-v1-urban-extents [accessed 29 January 2013].
- Taskforce on Innovative International Financing for Health Systems. More money for health, and more health for the money: final report. Geneva: International Health Partnership; 2009.
- Millenium Project [Internet]. Investing in development: a practical plan to achieve the Millennium Development Goals. London: United Nations Development Programme; 2005: London.
- Macroeconomics and health: investing in health for economic development. Geneva: World Health Organization; 2001.
- Liu A, Sullivan S, Khan M, Sachs S, Singh P. Community health workers in global health: scale and scalability. Mt Sinai J Med 2011; 78: 419-35 http://dx.doi.org/10.1002/msj.20260 pmid: 21598268.
- Crigler L, Hill K, Furth R, Bjerregaard D. Community Health Worker Assessment and Improvement Matrix (CHW AIM): a toolkit for improving CHW programs and services. Washington: United States Agency for International Development; 2011.