Program evaluation of an ORS and zinc scale-up program in 8 Nigerian states

Background In Nigeria, diarrhea is the second leading killer of children under five. Between 2012-2017, the Clinton Health Access Initiative, Inc. (CHAI) and the Government of Nigeria implemented a comprehensive program in eight states aimed at increasing the percentage of children under five with diarrhea who were treated with zinc and oral rehydration solution (ORS). The program addressed demand, supply, and policy barriers to ORS and zinc uptake through interventions in both public and private sectors. The interventions included: (1) policy revision and partner coordination; (2) market shaping to improve availability of affordable, high-quality ORS and zinc; (3) provider training and mentoring; and (4) caregiver demand generation. Methods We conducted cross–sectional household surveys in program states at baseline, midline, and endline and constructed logistic regression models with generalized estimating equations to assess changes in ORS and zinc treatment during the program period. Results In descriptive analysis, we found 38% (95% CI = 34%-42%) received ORS at baseline and 4% (95% CI = 3%-5%) received both ORS and zinc. At endline, we found 55% (95% CI = 51%-58%) received ORS and 30% (95% CI = 27%-33%) received both ORS and zinc. Adjusting for other covariates, the odds of diarrhea being treated with ORS were 1.88 (95% CI = 1.46, 2.43) times greater at endline. The odds of diarrhea being treated with ORS and zinc combined were 15.14 (95% CI = 9.82, 23.34) times greater at endline. When we include the interaction term to investigate whether the odds ratios between the endline and baseline survey were modified by source of care, we found statistically significant results among diarrhea episodes that sought care in the public and private sector. Among cases that sought care in the public sector, the predictive probability of treatment with ORS increased from 57% (95% CI = 50%-65%) to 83% (95% CI = 79%-87%). Among cases that sought care in the private sector, the predictive probability increased from 41% (95% CI = 34%-48%) to 58% (95% CI = 54%-63%). Conclusions Use of ORS and combined ORS and zinc for treatment of diarrhea significantly increased in program states during the program period.


Contents
. Detailed program activities and statistics 2 Appendix S1. Survey weighting 5  Worked with government to develop a national Essential Childhood Medicines Scale-Up Plan outlining a coordinated strategy for scaling up ORS and zinc use.
Coordinated mass media demand generation activities with BBC Media Action and the Sustaining Health Outcomes through the Private Sector (SHOPS) to reach additional states. Established National and State Essential Medicine Coordinating Mechanisms (N/SEMCM) to coordinate domestic and donor resources for implementing the Essential Childhood Medicines Scale-Up Plan.
Coordinated community inter-personal engagements and PPMV outreach with Pact, Society for Family Health (SFH), and SHOPS.
Supported government to revise the national standard treatment guidelines, national standing orders, and the essential medicines list to include zinc as well as broadly communicate policy allowing zinc to be given over-the-counter. Mapped partner activities and investments to identify opportunities for coordination. Market shaping and improving availability of affordable, highquality ORS and zinc Facilitated the entry of 4 new low-osmolarity ORS (L-ORS), 5 new zinc DT products, and 4 co-packaged products to the Nigerian market.
Technical support to suppliers to introduce an additional 3 copacks and 1 zinc DT product.
Supported local suppliers to reduce wholesale prices by connecting them to more affordable sources of supplies. This led to a 77% reduction in the wholesale cost of a treatment course (2 1L L-ORS sachets & 10 zinc tablets) from $1.56 to $0.34 Supported NAFDAC to push existing ORS suppliers to switch to the low-osmolarity formulation.
Implemented "wholesale activation" to prime the supply chain for ORS and zinc. The activity entailed placing brand-agnostic representatives at regional wholesale hubs to encourage purchase of ORS and zinc and distribute promotional and point-of-sale materials. Reached 5,655 PPMVs.
Conducted 3 additional rounds of wholesaler activation reaching an additional 15,841 PPMVs.
Signed incentive agreements with local suppliers to implement a rural salesforce. Local suppliers hired and trained 33 rural salesforce team members which sold 273,000 ORS sachets and 288,000 zinc strips.
Continued incentive agreements with local suppliers to implement a rural salesforce resulting in the distribution of 1.7M ORS sachets and 1.6M zinc blister packs to PPMVs and other retailers.
Provided guidance to state governments on ORS and zinc procurements, including technical assistance with quantification and advocacy to secure budget and funding.
Signed performance-based agreements with nine local distributors to use their existing supply chains to expand distribution of ORS and zinc. Performance based incentives were based on the share of rural PPMV shops carrying ORS and zinc based on independent assessments. Continued support to state governments to budget and procure  Adapted caregiver demand generation activities to leverage existing structures, such as female vanguard organizations, religious schools and educational institutions. Over 1,000 sessions discussing children's health and diarrhea were conducted with female vanguard associations and reached over 40,000 women. Over 700 sessions were conducted at Islamiyah schools reaching over 47,000 women. Trained 38,825 key community influencers as 'diarrhea champions' to educate caregivers on appropriate diarrhea management and disseminate flyers and other reference materials to share recommendations to their communities.
Expanded "health talks" to over 500 PHCs reaching over 2M caregivers.
Incorporated diarrhea management content into regular "health Launched radio campaign in 5 states (Bauchi, Kaduna, Kano, talks" at 96 secondary health facilities. "Health talks" targeted mothers visiting antenatal, immunization, and general outpatient clinics. Reached an estimated 448,835 women in the first half of the program. Katsina, and Rivers). Aired 3,400 radio spots promoting ORS and zinc.

Appendix S1. Survey weighting
Probability selection weights were calculated for each household. We estimated the probability of a household being selected for the survey using the following equation: where, Phi is the overall probability of a household in the ith EA of stratum h being selected, P1hi is the first-stage sampling probability of the ith EA in stratum h being selected, and P2hi is the second-stage probability of selecting a household in the ith EA from stratum h.
In the first-stage probability selection, let nh be the number of EAs selected in stratum h and Nh be the total number of EAs in stratum h from the sampling frame. The probability of selecting the ith EA in stratum h is calculated using the following equation: However, the census does not provide information on the number of households living in urban and rural areas. Therefore, we use the proportion of urban and rural EAs from the EA database to estimate the proportion of households living in urban and rural areas. Table 4 summarizes the calculations and estimates. We calculate the proportion of households living in each stratum to the population in the 8 program states. To calculate the post-stratification weights, we take the ratio between the estimated 2006 household population distribution in the 8 states and our weighted household sample distribution in the dataset.