Professional Documents
Culture Documents
2014 - Alex Morris - HPP - Price Subsidies For Private Sector ACTs - Systematic Review
2014 - Alex Morris - HPP - Price Subsidies For Private Sector ACTs - Systematic Review
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0/), which permits
unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Health Policy and Planning 2014;1–9
Published by Oxford University Press in association with The London School of Hygiene and Tropical Medicine doi:10.1093/heapol/czu013
REVIEW
Methods
and findings A systematic literature review was conducted to identify reports of experimental
or programmatic ACT subsidies to assess the impact of subsidies on consumer
use. Relationships between price, use and potential confounding factors were
examined using logistic and repeated measures binomial regression models, and
approximate magnitudes of associations were assessed with linear regression. In
total, 40 studies, 14 peer-reviewed and 26 non-peer-reviewed, were eligible for
inclusion in the analysis. The reviewed studies found a substantial increase in
private sector ACT use following the introduction of a subsidy. Overall, each $1
decrease in price was linked to a 24 percentage point increase in the fraction of
suspected malaria cases purchasing ACTs (R2 ¼ 0.302). No significant differences
were evident in this relationship when comparing the poorest and richest groups,
rural vs urban populations or children vs adults.
Conclusions These findings suggest that ACT price reductions can increase their use for
suspected malaria, even within poorer, more remote populations that may be
most at risk of malaria mortality. Whether a subsidy is appropriate will depend
upon local context, including treatment-seeking behaviours and malaria preva-
lence. This review provides an initial foundation for policymakers to make
evidence-based decisions regarding ACT price reductions to increase use of
potentially life-saving drugs.
Keywords Subsidy, artemisinin combination therapies, malaria, systematic review
1
2 HEALTH POLICY AND PLANNING
KEY MESSAGES
Subsidies increased private sector ACT usage by children under the age of five in all of the studies identified, from
7 percentage points in Nigeria to 49 percentage points in Uganda (both at national scale).
On average, a 1 dollar reduction in the price of ACTs was associated with a 24 percentage point increase in ACT usage
among those of any age, and a 32 percentage point increase among children under five.
The observed increases in ACT use in the private sector following subsidization appear equitable; subsidies substantially
increased ACT use even among the poorest individuals studied, and no significant differences were apparent in ACT use
increases between those living in rural vs urban areas.
All ACT subsidies described in the reviewed literature were implementation of Global Fund grants and other interventions
categorized as either corresponding to (1) the AMFm, (2) an (WHO 2011). However, no known private sector interventions
experimental subsidy or (3) a non-AMFm, non-experimental at national scale were implemented during this period other
programmatic subsidy, typically implemented by the govern- than the subsidies under consideration.
ment or an non-governmental organizations. These categories Where possible, trends were examined within specific sub-
were selected to examine the effect of the AMFm specifically, to groups of interest, including children under five, those of lowest
observe rigorously measured impact of subsidization in research socioeconomic status (SES) and those living in rural areas. SES
studies that compared changes against controls and to review was measured by comparing the highest and lowest wealth
evidence on the impact of prior programmatic subsidies. Any index quintiles, generated from principal components analysis of
information provided on anti-malarial use or purchasing assets by ACTwatch using demographic and health survey
behaviour, patient characteristics and ACT stocking and pricing methodology. Rurality was considered as a dichotomous rural/
in retail outlets or health facilities was recorded. The ACT urban variable defined separately by each study. Differences in
stocking and pricing data were used to look for associations use and increases in use during subsidies between high- and
between these measures and use rates. low-SES groups and those in urban and rural locations were
ACT use has been defined as both (1) the proportion of compared statistically using logistic regression models.
patients with fevers (or suspected malaria illness) who receive In a second set of analyses, all studies reporting not only use
an ACT (Kangwana et al. 2011) and (2) the proportion of anti- but also stocking of ACTs and sales prices were compiled to
malarial recipients who receive an ACT (Talisuna et al. 2012). specifically examine the linkage between commonly measured
through database
screening (n=370)
PubMed (n=134)
sources (n=36)
EMBASE (n=215)
Cochrane (n=21)
40 records included in
Included
A B
through private but accredited drug dispensing outlets. Under Relationships between ACT use, stocking and price
the subsidy, use of ACTs in the private sector increased from 0% Figure 5 depicts all reviewed reports that presented data on
of anti-malarial drugs purchased for children under five to 10% both the fraction of private sector outlets found to be stocking
almost a year after subsidy implementation (Simba and Kakoko ACTs and the price of the ACTs in those locations at proximal
2012). In Kenya, a subsidy was launched in 2007 to make ACTs time periods. Accounting for correlation between outlets
available to Community and Family Wellness shops in three measured in the same survey, the price of ACTs for any age
rural districts. In 2006, household surveys found that 15% of group was negatively but not significantly associated with ACT
patients receiving anti-malarials in either the private shops or stocking (OR ¼ 0.870, Z ¼ 1.66, P ¼ 0.098). However, each $1
public health facilities received ACTs. After 15 months of the increase in the price of a paediatric ACT was associated with an
intervention, this fraction increased to 42%, with 9% of those OR for stocking of 0.223 (Z ¼ 2.62, P ¼ 0.009). Ignoring
receiving ACTs obtaining them from the subsidized shops correlation, linear regression output suggests that this $1
(Sabot et al. 2009a,b). increase in price is associated with a 3.49% decrease in the
fraction of outlets stocking any ACT (R2 ¼ 0.051), and a 23.5%
decrease in the fraction of outlets stocking a paediatric ACT
The AMFm specifically (R2 ¼ 0.532).
The ACTwatch Group collected comparable data on ACT use in In Figure 6, all reviewed studies are depicted that recorded
Madagascar, Nigeria and Uganda, with surveys conducted data on both ACT stocking in the private sector and their use in
1–2 years before implementation of the subsidy and again the same regions at proximal time periods. A repeated measures
after almost 2 years of subsidization in Nigeria and Madagascar regression model suggests that, on average, each 10% increase
and 1 year in Uganda (Figure 3) (ACTwatch Group and in stocking is associated with an OR for ACT use of 1.27
ABMS/Benin et al. 2012). In addition, reports were identified (Z ¼ 3.82, P < 0.001) among all ages. For children, the OR is
reporting ACT use in Ghana, Tanzania, Uganda and Madagascar 1.46 (Z ¼ 3.63, P < 0.001). Ignoring correlation between meas-
following implementation of the AMFm (Quakyi et al. 2012; urements from the same surveys, this 10% increase in stocking
P. Yadav et al., unpublished data). All reports described is associated with a 5% increase in the fraction using ACTs
increases in ACT use after implementation of the subsidy. At (R2 ¼ 0.363) among all ages, and an 8% increase in the fraction
the time of analysis, no reports were available on the subsidy’s of children using ACTs (R2 ¼ 0.462).
effect in Zanzibar, Kenya or Niger. Finally, a negative association was found between mean ACT
The increases in use observed by ACTwatch household price and use in the same regions (Figure 7). The overall trend
surveys of children under five in Madagascar, Nigeria and was strongly influenced by a single outlier from Madagascar,
Uganda varied within SES groups and within urban and rural where price was >5 dollars. Excluding this point from analysis
strata between baseline and endline survey rounds (Figure 4). and accounting for correlation, on average a $1 increase in ACT
These classifications reflect categorizations made by the price was associated with an OR for use of 0.30 (Z ¼ 3.02,
ACTwatch Group in their reports on each country (ACTwatch P < 0.001) among all ages, and 0.153 (Z ¼ 6.29, P < 0.001)
Group and ABMS/Benin et al. 2012). among children. Ignoring correlation and again excluding the
6 HEALTH POLICY AND PLANNING
A B
Figure 4 Changes in ACT use in AMFm countries by rurality and SES. (A) Increases in use of ACTs in the private sector by children under five in
the lowest and highest SES groups reported in ACTwatch household surveys before and after implementation of the AMFm in Madagascar, Nigeria
and Uganda. (B) Increases in use of ACTs by children under five in urban and rural areas as reported in ACTwatch household surveys before and
after implementation of the AMFm in Madagascar, Nigeria and Uganda.
outlier, a $1 decrease in price was associated with a 24 percent- experimental studies, programmatic subsidies and the AMFm
age point increase in ACT use among those of any age all consistently demonstrate a negative association between
(R2 ¼ 0.302). Restricting only to surveys measuring price for ACT use and price. In most non-experimental examples, the
paediatric drugs and use by children under five, a $1 decrease lack of controls for comparison means it is impossible to assess
in ACT price was associated with a 32 percentage point increase the attributable impact precisely, but in the absence of other
in ACT use (R2 ¼ 0.594). major interventions aimed at increasing use among private
sector treatment-seekers, it is likely that subsidy implementa-
tion plays an important role.
The influence of subsidies on ACT use appears consistent
Discussion within subgroups such as children under five and more remote
The results of this review suggest that reducing the retail price and poorer populations. Contrary to some stated concerns
of ACTs can successfully increase ACT use, although evidence (Oxfam 2012), this finding suggests that ACT price reductions
from rigorous studies remains scant. Available results from may have substantial impact on populations that have been
SUBSIDIES INCREASE ACT USE: A SYSTEMATIC REVIEW 7
Figure 5 ACT prices and stocking in the private sector. The dashed
treat outside the formal health sector and the prevalence of ACTwatch Group, Population Services International. 2011. Household
malaria infection in this population (Cohen et al. 2012a,b). Survey Report Kingdom of Cambodia. Nairobi, Kenya: Population
This review has important limitations. Each of the types of Services Internationalhttp://www.actwatch.info/countries/evidence.
studies reviewed has specific weaknesses: the experimental php?id_country¼29, accessed 22 October 2012.
studies were in limited geographies and controlled environ- Arnold F, Ye Y, Ren R et al. 2012. Independent Evaluation of the
ments that may not be fully representative of programmatic Affordable Medicines Facility—malaria (AMFm). Multi-country
Independent Evaluation Report: Final Report. Calverton, MD: ICF
conditions, while the programmatic findings tended not to have
International.
controls nor proximal baselines, making it impossible to
Barat LM, Palmer N, Basu S et al. 2004. Do malaria control interventions
understand the counterfactual scenario of how ACT usage
reach the poor? A view through the equity lens. American Journal of
might have changed in the absence of a subsidy. Several of the
Tropical Medicine and Hygiene 71: 174–8.
AMFm evaluations included in the analysis were conducted
Boerma JT, Black RE, Sommerfelt AE, Rutstein SO, Bicego GT. 1991.
shortly after the implementation of the subsidy, making it
Accuracy and completeness of mothers’ recall of diarrhoea occur-
difficult to ascertain the total magnitude of the effect of rence in pre-school children in Demographic and Health Surveys.
the intervention. In some cases, such as Uganda, the imple- International Journal of Epidemiology 20: 1073–80.
mentation of the subsidy was significantly delayed, and the Cohen J, Dupas P, Shaner S. 2012a. Price Subsidies, Diagnostic Tests, and
evaluation period covered just 11 months. Finally, treatment Targeting of Malaria Treatment: Evidence from a Randomized Controlled
outcomes are the product of a complex series of factors Trial. [Manuscript in progress]. http://www.povertyactionlab.org/pub-
including social norms, supply chains and interactions between lication/price-subsidies-diagnostic-tests-and-targeting-malaria-
FUNDING Liberati A, Altman D, Tetzlaff J et al. 2009. The PRISMA statement for
reporting systematic reviews and meta-analyses of studies that
This work was supported by the Bill and Melinda Gates evaluate healthcare interventions: explanation and elaboration.
Foundation. BMJ 339: b2700.
Moon S, Perez Casas C, Kindermans JM, de Smet M, von Schoen-
Angerer T. 2009. Focusing on quality patient care in the new global
Acknowledgements subsidy for malaria medicines. PLoS Medicine 6: e1000106.
The authors would like to thank ACTwatch Group (Population Oxfam. 2012. Salt, Sugar, and Malaria Pills: How the Affordable Medicine
Facility—Malaria Endangers Public Health. Oxfam Briefing Paper. http://
Services International), in particular Tanya Shewchuk and
www.oxfam.org/sites/www.oxfam.org/files/bp163-affordable-medi
Stephen Poyer, as well as Catherine Goodman, Kara Hanson
cine-facility-malaria-241012-en.pdf, accessed 24 October 2012.
and Prashant Yadav for reviewing earlier versions of this paper.
Population Services International. 2006. Cambodia Malaria TRaC Survey
among Populations Living in Malaria Endemic Areas. Washington, DC:
Population Services International Research Division.
References Quakyi I, Winch P, Adjei A, Sullivan D. 2012. Summary Report: Improving
ACTwatch Group and ABMS/Benin, ASF/DRC, PSI/Madagascar et al. Correct Diagnosis and Treatment of Malaria in Rural Ghana. Baltimore,
2012. Household Survey Report Benin, Democratic Republic of the Congo, MD: Johns Hopkins School of Public Health.
Madagascar, Nigeria, Uganda and Zambia. Roll Back Malaria Partnership. 2011. GMAP—the Global Malaria Action
ACTwatch Group, Population Services International. 2009. Household Plan. Refined/Updated GMAP Objectives, Targets, Milestones and Priorities
Survey Report Kingdom of Cambodia. http://www.actwatch.info/ Beyond 2011. http://www.rbm.who.int/gmap/gmap2011update.pdf,
countries/evidence.php?id_country¼29. accessed 22 October 2012. accessed 24 October 2012.
SUBSIDIES INCREASE ACT USE: A SYSTEMATIC REVIEW 9
Sabot O, Yeung S, Pagnoni F et al. 2009a. Distribution of artemisinin- Talisuna A, Daumerie PG, Balyeku A et al. 2012. Closing the access
based combination therapies through private sector channels: barrier for effective anti-malarials in the private sector in rural
lessons from four country case studies. Resources for the Future Uganda: consortium for ACT private sector subsidy (CAPSS) pilot
Discussion Paper: Prepared for the Consultative Forum on the study. Malaria Journal 11: 356.
AMFm—the Affordable Medicines Facility—malaria, September White NJ. 2008. Qinghaosu (Artemisinin): the price of success. Science
27–28, 2008, Washington, D.C. http://www.rff.org/RFF/Documents/ 320: 330–4.
RFF-DP-08-43_FINAL.pdf, accessed 7 September 2012. World Health Organization. 2010. Guidelines for Treatment of Malaria—
Sabot OJ, Mwita A, Cohen JM et al. 2009b. Piloting the global subsidy: Second edition. http://www.who.int/malaria/publications/atoz/
the impact of subsidized artemisinin-based combination therapies 9789241547925/en/index.html, accessed 15 October 2012.
distributed through private drug shops in rural Tanzania. PLoS One 4: World Health Organization. 2011. World Malaria Report 2011. http://www.who.
e6857. int/malaria/world_malaria_report_2011/en/, accessed 7 September
Schäferhoff M, Yamey G. 2011. Estimating Benchmarks of Success in the 2012.
Affordable Medicines Facility—Malaria Phase 1. San Francisco: World Health Organization. 2013. Financing Malaria Control—Allocating
Evidence-to-Policy Initiative. http://www.theglobalfund.org/docu Limited Resources. MPAC Presentation. http://www.who.int/malaria/
ments/amfm/E2PI_EstimatingBenchmarksInAMFm_Report_en/, mpac/resource_allocation_mpac_presentation_march_2013.pdf,
accessed 7 September 2012. accessed 20 May 2013.
Simba D, Kakoko D. 2012. Access to subsidized artemether-lumefantrine Yeung S, Patouillard E, Allen H, Socheat D. 2011. Socially-marketed
from the private sector among febrile children in rural setting in rapid diagnostic tests and ACT in the private sector: ten years of
Kilosa, Tanzania. Tanzania Journal of Health Research 14: 2. experience in Cambodia. Malaria Journal 10: 243.