Healthy Pregnant Women

In Kenya, 350 out of 100.000 women don’t survive pregnancy. In the Netherlands, this number is only 7. While certainly any number is too high, this remarkable difference suggests that there is something we can do to improve the situation. PharmAccess helps pregnant women to deliver their babies safely. We helped PharmAccess make the next step in achieving this mission by the analysis presented here. The study yielded five main findings:​

  1. We found patterns in which complications occur during pregnancies and when.

  2. A mere 4% of women in the program make up to 20% of the program's costs.

  3. Clinic visitation behaviors showed preferred visitation hours.

  4. We found similarities in demographics and behaviors between different groups of women.

  5. We observed differences in the amount of times women return to the various clinics, providing a valuable proxy for quality of care at the clinic.

Note that, to ensure privacy, all visualizations presented here are based on dummy data and not actual results.


PharmAccess wants to do healthcare better. With a focus on sub-Saharan Africa, they work on improving the whole system. This means mobilizing private and public resources to get more money into the system, measuring and improving quality of care, and reaching even the most excluded individuals through simple technologies like the mobile phone. This approach to healthcare is designed to ensure that everyone has access to better care when they need it, not just when they can afford it. 

For more information on PharmAccess' process, see the below video.

The project

One of the projects PharmAccess undertakes is the MomCare program in which insurance is provided to pregnant women. The brilliance in the approach of PharmAccess is the fact that both clinics and insurance beneficiaries affirm visits and their outcomes in an online environment. This creates transparency in money flows and providing information on health progression.

The analysis

To enable PharmAccess to make the next step in their mission, we analysed their data, visualizing it interactively and intuitively. A large part of PharmAccess their work is to allocate they funding wisely. Their money should flow into the projects and interventions that help vulnerable people in the most effective ways. In those places where theoretical research about effects is lacking or inapplicable, real world data has to fill in the gaps.

This analysis therefore focused on picking low hanging fruits that indicate how to focus internal efforts. Here we provide some of the main findings (fully anonymised). For each of the main findings, results were visualized interactively. We also provided suggestions to PharmAccess on the implications that the findings have.


  • Patterns are found as to which anomalies occurring during pregnancies and when. This gives doctors additional insights into when medicine should be in stock, what complications they should be looking out for, and what specialist they should have available based on women their individual journeys.

  • A mere 4% of the women make up 20% of the program's costs. Drilling into the complications these women were diagnosed with gave us insight into what complications are driving the programs' costs.

  • We were able to cluster women in different groups based on individual demographics. These clusters were subsequently found to display distinct behaviors. Using these clusters, we can make estimates of the approximate behaviors of women that sign up to the program. This allows to catch undesirable behavior early on.

  • Pregnant women were found to visit clinics at similar times. This allows doctors to adjust their opening hours and the presence of certain specialists in response to demand.

  • We observed differences in how often women come back for visits for the various clinics. This information gives a rough but valuable proxy for the quality of care provided by the clinics, allowing PharmAccess to monitor quality better.

  • Possibly best of all result, we find in such exploratory analysis projects that internal data analysis raises additional questions that help define how to better service the 'customer'. In this case, how to improve mother and baby health in pregnant women.

Customer Testimonies

"I really like how this looks! It gives a really clear overview on what is happening on the ground." 

- Mark van der Graaf, Director Digital Health

"This is what I want to see! It would be much nicer to send our clinics reports like these, rather than having to manually analyse excel files and type up reports."

- Nicole Spieker, Director East Africa

"This is great. We could use this to monitor in real time whether clinics are open."

- Mark van der Graaf, Director Digital Health

"I could continue looking at this visualisation for two hours, and I would still be learning new things."

- Andreas Chrysanthou, Business Analyst

Follow us:

© 2020 by Soda science