Research Project

Improving Health Campaign Microplanning with Predictive Analytics Planning Tool

Findings from Rwanda


A study to determine if a web-enabled platform that automates many planning tasks (predictive analytics tool) improves the efficiency and effectiveness of health campaigns. 


The study location was 30 districts in Rwanda.

Neglected Tropical Diseases (NTDs) and Vitamin A

  • Predictive analytics provides the ability to analyze the performance of a health campaign in relation to social determinants such as poverty.
  • Predictive analytics can help assess health past campaign performance and trends that can be used for future programming.
  • See more.

Key Messages

The consortium of Connecti3, University of Rwanda and Orbital Media developed a web-enabled platform that automates many planning tasks (e.g., forecasting), incorporates numerous public and privately available datasets, and uses algorithms that can improve the efficiency and effectiveness of the health campaign microplans. This study aimed to determine if an automated, predictive analytics tool can increase efficiency by reducing planning time, ensuring better targeting of services to those with the greatest need, and optimizing resource deployment plans.

The tool’s efficiency was evaluated by looking at how much planning time was saved during a  nationwide integrated health campaign that offered vitamin A supplements, intestinal deworming, schistosomiasis treatment, screening and treatment for malnutrition, family planning, vaccination catch-up services, and promotion of hygiene and sanitation. 

The team evaluated the targeting of services by examining the relationship between coverage rates and poverty distribution, and computed resource optimization on whether the operational funds were dispersed to attain the highest coverage rates. The study also evaluated how well the tool was built, how well it addressed the customers’ planning difficulties, and how keen the end users were to use this new technology for the next health campaign planning.

General findings:

  • Predictive analytics can help save time.
  • Predictive analytics can help better target services to those in need.
  • Predictive analytics can help resource deployment.

“Too many people involved, leads to not everyone being trained properly and trouble getting everyone together, not to mention it is very expensive to bring all those people together." - KII participant



Rwanda uses paper-based tools for planning health campaigns. Manual systems take time, demand intensive departmental coordination, and make it challenging to determine whether the intended campaign goals were achieved. It also does not allow the use of predictive analytics, artificial intelligence (AI), or big data to improve health service targeting, or optimize resource deployment.

The consortium of Connecti3, University of Rwanda and Orbital Media developed a web-enabled platform that has automated many of the planning tasks. This technology now performs tasks like forecasting and calculating the required quantities of medical supplies automatically. A dashboard keeps tabs on the districts’ operational readiness for the health campaign, the status of district budget allocations, and national fundraising status and once the results are entered into the tool, the program generates a report template with all the charts and data annexes.

Image above: Study locations across 30 districts within Rwanda.

Study Location

  • The Ministry of Health conducts nationwide (across 30 districts), an integrated health campaign, known as Maternal and Child (MCH) Weeks two times a year.
  • The May 2022 MCH Week offered Vitamin A, intestinal deworming schistosomiasis treatment, screening and treatment of malnutrition, family planning, routine vaccination catch-up, promotion of hygiene and sanitation, education on snake and dog bites.
  • The MCH Week plans to reach 6,573,029 people, (a total of 48% of the total pop of 13,626,507. 
  • The total cost of this campaign is estimated at RWA1,143,227,670, about US$1,111,008.

Study Questions

  1. Can a predictive analytics tool improve health campaign efficiency and effectiveness?
  2. How does the tool compare to the current system?
  3. Are health campaign planners willing to use the tool? 

Image above: Screenshot of the predictive analytics planning tool.

See the research brief to view the methods of this study.


Over the course of 1.5 months, the Ministry of Health (MOH) staff held a total of six planning meetings, involving over 201 people to discuss the status of the microplans from March 15, 2022, to April 29, 2022. approximately 40 days were spent in these meetings. The predictive analytics tool was an improvement over the manual planning method in these key areas.


Key informant interviews confirmed that long, arduous planning time is a pain point for campaign managers. Comparatively, the predictive analytics microplanning tool took 10 days less than the manual system of planning and involved a total of 151 people instead of 201. Multiple people from the same department who were there at the meeting to simply know what was going on did not need to attend in person because they could monitor the status of the health campaign microplans on the online dashboard. This translates into a calculated savings of about 10 days of senior staff time per campaign (most staff have an average tenure of 12 years) by using the predictive analytics platform.

Analysis of coverage and poverty

Beyond coverage rates, the manual method of planning and reporting did not calculate any additional key performance indicators that could inform future health campaign programming. The predictive analytics tool expanded this knowledge, finding that the district poverty rates , and health campaign coverage performance were shown to be inversely correlated, with poorer districts obtaining higher coverage rates for vitamin A and deworming compared to affluent districts. The poorest district received the least amount of money (US$2.09) per person and achieved the highest coverage rates of 95% for Vitamin A and 97% for deworming.

User feedback

The direct observation of the tool used by 13 MOH health campaign staff showed that they were able to navigate the tool easily and found it intuitive. An online survey after the direct observation was completed by 10 out of 13 people and provided the following information:

Lessons learned

While the tool could reduce the staff workload and bring more visibility into the planning process, many staff members were reluctant to take on another digital tool, especially one that is built by a private sector company. See the research brief for more details on these lessons and potential solutions.

“Help Us - Digitalize the health campaign activities especially the reporting to help us save time." – FGD participant, Gisenyi District

The use of this team’s predictive analytics digitization platform shows promise. Based on the information gathered through interviews and focus groups, the predictive analytics tool addresses the challenges of planning health campaigns, especially the ability to have planning visibility across departments on a single platform.

Promising practices include:

  • Predictive analytics tools can provide the ability to analyze and better understand the performance of a health campaign in relation to social determinants such as poverty.
  • Predictive analytics provides an easier method to assess health past campaign performance and trends that can be used for future programming.
  • Predictive analytics can save planning time and provide better transparency across departments.

See the research brief for detail on lessons learned and policy implications.

Photo Credit: University of Rwanda

“The major strength for me is the diversity of people involved in the preparation.” – KII participant