Key Learnings from Building a Life Saving Distribution Model

Ruhani Walia
16 min readJan 31, 2021

--

We often confuse words.

Affect vs. Effect

Than vs. Then

Complement vs. Compliment

These mistakes simply result in a red notification on Grammarly or a blue underline on MS Word, nudging you to correct it. Unfortunately for others, the confusion perpetuates long-standing issues of accessibility and fairness.

You might wonder, how can two words possibly do that much damage simply by sounding alike?

Let’s talk about Equality vs. Equity.

It is undeniable that some of us are born with a head start. That head start can mean being born into wealth, in a wealthy country, or in a place with greater opportunity. That’s what this first graphic represents.

Each of us is born into different socioeconomic backgrounds. We each have a different outlook on life. Sidenote: if you’ve seen the classic model for equality vs. equity, you’ll notice these graphics differ. In these graphics, everyone is the same height. Therefore, the “fault” of inequity is not placed on people’s innate qualities but on environmental startings points, in this case, the hills.

Equality has become synonymous with “a level playing field” and giving everyone the same resources. It is being “fair” across the board.

In this graphic, each person has been given the same amount of support. It is equal. Notice how the disparity has not been solved.

Equity should become known as “more for those who need it” and differences in resource allotment by need.

This graphic results in the same view for all because each has received what they need. In some ways, equity results in more equal access to opportunity.

What, then, is inequity?

It is a lack of fairness and the existence of resolvable barriers

To reference the first graphic, it is turning a blind eye to the obvious and undeniable advantages that some have over others.

Inequity can be frustrating. It feels like seeing a solution but watching the problem unfold before you anyways.

The frustration of inequity is why we’re building a framework to reduce maternal mortality.

Mothers are needlessly dying while giving life.

Maternal mortality is a tragedy that affects far too many women around the world, particularly in Africa and South Asia. Northern Nigeria is afflicted by some of the highest maternal mortality ratios (MMRs) in the world. For example, Jigawa State, in North-Western Nigeria, has an MMR of 1,012 deaths per 100,000 live births.

If Jigawa were to try and reach the Sustainable Development Goal (SDG) 3.1 goal set by the United Nations, the MMR would need to be 93% lower!

That’s why, about a year ago, my team and I started our intervention, Ending Maternal Mortality (EMM). EMM aims to supply women in rural areas with misoprostol, a drug that reduces the onset of postpartum hemorrhaging, to reduce the rate of maternal mortality deaths by 20–31.5%.

Postpartum hemorrhaging (I know it sounds like a mouth full) is simply excessive bleeding that occurs once a woman has given birth. Unfortunately, it’s the world’s leading cause of maternal mortality.

Fortunately, misoprostol can reduce 85–90% of PPH cases when administered properly.

Our first pilot to increase access to misoprostol is planned to take place in Jigawa State, Nigeria. With support from the Ministry of Health, EMM hopes to further this project to other states and countries. Maternal mortality is a global issue — yet we seldom hear about it in developed countries, where instead, binge-eating is more likely to kill you (I’m not even joking).

On the other hand, African women are 175 times more likely to die due to childbirth than women in developed countries.

Strong maternal healthcare systems are a benefit to more than just mothers; strong maternal healthcare is a step towards a stronger healthcare system for families, and ultimately, resilient communities for all. The tools necessary to reduce maternal mortality exist; they are simply not accessible or affordable to those who need them the most.

See where the frustration comes from?

This is an issue of inequity stemming from poor distribution of and accessibility to solutions that can save lives

The focus of this article is going to be on key learnings when designing a model for rural, remote, and low-income areas.

Building the model for EMM has taught me three key mistakes that must be avoided:

  1. A Weak Monitoring and Evaluation System
  2. Failing to Carefully Consider Cultural Barriers
  3. Incorrect Selection and Attribution of Stakeholders

1. Monitoring and Evaluation

I know this sounds like a buzzword, but I promise it’s meaningful. From designing the EMM pilot-study, pilot model and post-pilot analysis, there are a few key learnings I’ve gathered.

M&E is essentially the creation of feedback loops for a project. Each part, monitoring and evaluation, work in tandem but mean very different things:

Monitoring: the continuous process of ensuring activities under the program are executed as they should be and providing regular feedback should it deviate

Evaluation: collecting and analyzing information to answer questions about the programme’s progress. (This is done through process, outcome and impact evaluation — more on this soon).

Let’s take a look at each in a little more detail.

Monitoring

Monitoring, according to the WHO, is achieved by regularly collecting information about all activities carried out by the program. This means receiving updates from stakeholders, reviewing data collection, and identifying whether the intervention is running into any problems that need to be immediately solved.

Evaluation

On the other hand, there are several types of evaluation. A few are listed below in chronological order of use:

  1. Formative Evaluation during conceptualization: generate data on the need for the intervention, develop baseline indicators and program priorities
  2. Process Evaluation during implementation: measuring the effectiveness of program procedures, data generated identifies inefficiencies and areas of potential improvement
  3. Outcome Evaluation during implementation: data on intervention outcomes and whether the outcomes being identified are due to the program/to what extent (basically, how valuable the program actually is)
  4. Economic Evaluation during implementation: measures benefits of the program vs. its costs, measures the efficiency of a program (similar to an audit)
  5. Impact Evaluation post project: studies intervention from start to finish, identifies metrics to quantity whether it was successful or not
  6. Summative Evaluation post project: how well did the intervention reach its targets, used to ask for project expansion or continuation (if positive)
  7. Goals-Based Evaluation post project: evaluation of SMART targets

In addition to all the varying types of evaluation, the data being evaluated is a collection of indicators. There are several types of indicators. Further details on which is the most useful/impactful can be found under best practices:

  1. Input Indicators: measures of what is needed to start the intervention (e.g., funding, stakeholders, infrastructure)
  2. Process/Output Indicators: are activity driven and measure changes in the steps leading up to the desired outcome process (e.g., the number of contraceptives distributed, the number of children reached etc.)
  3. Outcome Indicators: results of activities, change-driven
  4. Impact Indicators: some interventions consider their most long-term outcome indicators impact indicators (e.g., % of deaths from Malaria after 3 months vs. at the start of the intervention)

The Logic Model

The Logic Model presented by the WHO is a great visual tool to understand the chronological order of indicators during evaluation:

After the pre-pilot research and baseline indicators are complete, reporting requirements must be created and met:

  1. Disaggregate baseline data (separate into groups and analyze)
  2. Identify sub-groups missing the program’s services to promote equitable access (the need in the community)
  3. Use (periodically) 3rd party assessments/audits to demonstrate objective results should they need to be reported (considered more legitimate than if the program conducts an audit itself)
  4. Clearly define roles and responsibilities for all stakeholders involved in M&E data collection and analysis
  5. Disseminate M&E results post-pilot study and pilot through reports or other mediums

So far, I’ve outlined some of the deeper parts of evaluation that have been helpful for me in program development.

Beyond the nitty-gritty, here’s my synthesized list of evaluation best practices:

  1. Do not solely rely on impact indicators. It is difficult to determine the specific environmental factors that lead to an impact indicator (e.g., a lower maternal mortality rate) and to associate it with your intervention is less accurate than if you were to instead use an outcome indicator. However, impact indicators are great to measure the system as a whole
  2. Set SMART Indicators (WHO Framework): indicators that are Specific Measurable Achievable Realistic and Time-bound (e.g., population with real income below the equivalent of $1 a day calculated as total household income divided by household members)
  3. Determine who the evaluation belongs to. Decide on ownership ahead of time to avoid conflict. Does it belong to those who paid for it (subjects)? Does it belong to on-ground workers who collected the indicator data? Does it belong to all participants? Does it belong to your organization? To your partner’s organization (if applicable)?
  4. Standardize M&E forms to avoid discrepancies in M&E data entry interpretation. By creating evaluation guidelines and standard processes, chances of varying definitions, estimation methods or explanations are reduced.
  5. Adjust indicators for local variants: output, process, and product indicators, all need to be adjusted to be accurate by local standards (e.g., a target for the number of facility births will be very different between villages if one is close to a facility than another (in distance))
  6. Determine endline results after determining the baseline. To create accurate and manageable goals, pre-pilot research must be conducted (to see an example of EMM’s pre-pilot research, click here). This can take the form of a needs-assessment. A needs assessment is a systematic analysis used to identify gaps in required infrastructure or resources. It will result in data about current infrastructure availability.

2. Cultural Barriers

Misoprostol is not the only thing that must be properly distributed; education on safe birthing practices and the perils of maternal mortality must be emphasized.

The issue itself seems to lack the gravity it deserves to spur action.

The precursor to changing accessibility is addressing cultural and stigmatic opposers to change. Even if the solution exists and is accessible, the ultimate barrier to change is how communities receive it. If it’s frowned upon, the solution, no matter how impactful and positive, is not going to get past the warehouse where it’s stored if it even makes it that far.

While this is something that our model definitely addresses, EMM has committed to developing cultural views on safe birthing practices by up-skilling health practitioners with a curriculum on misoprostol use. If you’re interested in learning more about EMM’s training curriculum, take a look at this article written by Isabella for an overview.

Within the model, there are three steps to determine what adaptations must be made per with cultural beliefs:

1. Understanding the community’s hierarchy and religious aptitudes.

Learning whether or not the population is religious will be a guiding force to the model adjustments you need to pursue.

For example, if your intervention is distributing contraceptives and the community is Christian with strong views on abstinence from sex until marriage, an unmarried woman may take offence to your offering contraceptives.

Learning the religion of a community will often highlight what the hierarchy of influence is. More importantly, it will help you uncover whether politics and religion both play a role in who influences the most people. For example, should you seek to gain support from the community mayor or the community pastor? If you require approval from both, whom must you ask first to follow the chain of command? Is there a government-endorsed faith?

The chain of command also applies to understanding government. What type of political system is used? What are the different levels of regulation (e.g., local, provincial, federal)? This information is necessary to ensure that your intervention does not infringe on any policies. For example, if your intervention is agricultural, are pesticides permitted?

2. Gaining contextual validation through primary research.

While reading research papers and pilot studies published by other interventions and organizations is good research, basing model design decisions on secondary research is risky. The best way to validate what you have learned is to ask a local of the community in which your intervention will take place.

It’s become part of my weekly routine to talk to Nigerians. EMM has assembled various advisory boards for each section of the framework, so an advisory board for model development, one for the training curriculum etc. Expert advice is always important, and if possible, experts with the same cultural context as your intervention location.

Checking that cultural assumptions are true is also necessary. Members of our team have interviewed mothers, community members, and those who left Nigeria for work elsewhere to verify whether our understanding of the cultural and religious aptitudes was correct.

Learning that you’ve made an incorrect assumption is one of the most valuable points of feedback you can gain before beginning anything on-ground.

A best practice here is to have an agenda and general questions prepared for each meeting.

3. Developing an understanding of attitudes towards technology and change.

This one is only applicable if your intervention is planning on using technology. EMM plans to use mobile phones and an application to store and collect data. A thorough analysis of technological barriers is important — and this includes looking at digital (il)literacy levels. Good questions to ask include:

  • What apps are most popular?
  • What age groups use phones?
  • What is the internet coverage like?
  • Is purchasing data expensive?
  • What networks/network providers exist?

Christina, who works on pictorials and anthropology research at EMM, introduced the idea of a Localization Guide. This research looks at how previous interventions altered their marketing, educational, and training materials based on geographic location and literacy level. If you’re interested in learning more about our localization guide, you can contact Christina at wchristina2512@gmail.com.

In addition to these steps, the EMM model majorly considers the history of non-traditional medical practices. These traditional practices in Nigeria are very closely related to spiritual and religious beliefs, so the introduction of a Western solution requires careful study of existing health norms.

The Heuristic Model on Religion Environment Effects on the Framing of Health Norms may be a helpful framework to apply to interventions in the healthcare space to understand what competing methodologies exist.

Normative Rules: religious followers gain from participating in normal faith activities. Coercive Rules: rules that must be followed to maintain religious identity. Mimetic Rules: rules that followers model off of others of the same faith

This framework outlines how religion affects the formation of health norms. A behavioural economic theorem known as the framing effect is used within this framework to explore how religion, as a force, shapes what becomes acceptable human behaviour.

The study of the heuristic model described a religious identity as something that obliges its follower into faith-aligned practices which guide them in their health choices.

People are more likely to participate in a public health program that aligns with their priority health norms. This means that the promotion of these initiatives should address faith and community-oriented standards for credibility.

3. Stakeholders

Common mistakes:

  1. Involving too many stakeholders and complicating the model
  2. Poorly allocating duties amongst (stakeholders/potentially overburdening a stakeholder/underusing a stakeholder). To resolve, can use the IAP2 Spectrum (referenced below)
  3. Not aligning the incentives of a stakeholder to their role in the model
  4. Overlooking community champions (enlisting advocates and community member support)

To avoid involving too many stakeholders and more accurately allocate duties among them, the IAP2 Spectrum can be used. This is a stakeholder engagement framework that is used to designate the appropriate level of participation to stakeholders.

The spectrum includes five levels of engagement that can be applied to stakeholders by objective, outcomes, timeframes, resources, and level of interest in the intervention. The levels include:

  1. Inform
  2. Consult
  3. Involve
  4. Collaborate
  5. Empower

Below is an example of an IAP2 Spectrum.

This example uses the framework to determine public engagement levels for stakeholders. The framework can be adjusted to be used in a variety of contexts such as level data involvement.

To avoid overpopulating sections of the model, a list of duties should be written out. Often, any overlap or repetition in tasks across stakeholders can be identified this way. This process requires the use of the MECE Framework.

The MECE Framework ensures that the options listed are all mutually exclusive and collectively exhaustive. This approach eliminates duality.

Lastly, to avoid overlooking community champions…create a process to identify and engage them! Enlisting help from existing on-ground organizations who are willing to help spread the word of your intervention ahead of time or distribute pamphlets will improve the likelihood that the solution will be well-received.

4. EMM Pilot Model

Having described various key learnings under the topic of M&E, Culture, and Stakeholders in model development, below is an overview of EMM’s current pilot model as of January 2021. This model has been through several iterations and will likely undergo many more before we pilot on-ground.

Summary: The proposed misoprostol distribution system begins at our supplier’s warehouse, DKT International. The misoprostol then moves to a distribution centre. Our model outlines three distribution centres, including Primary Healthcare Centres (PHCs), Health Clinics, and Health Posts.

At this point, there are two options. (1) A Mother may visit any of the distribution centres to receive a prescription, or a prescription and the misoprostol.

The second route continues: (2) Community Health Extension Workers (CHEWs) register and educate women on the intervention and provide them with a registration number and prescription. A Traditional Birth Attendant (TBA) will visit a PHC, health clinic or health post to receive a stock of misoprostol.

Here another two options occur. (1) A TBA gives a mother the misoprostol at 7 months of pregnancy then visits during the birth to help her deliver. (2) The TBA holds onto the misoprostol until birth, then administers it to the woman.

Finally, in the case of a birthing complication, a TBA may need to refer the woman to a healthcare facility to save her life.

Note: All TBAs attend an EMM training and must pass a certification exam to be part of this model. One of the things they’re trained on is when to refer women to a higher facility

More Detailed Overview

Stage 1 Stock

To distribute misoprostol to 10,000 women in 1 year or approximately 2,500 interventions in 3 months, we need to start with a pilot in 3 local government areas.

A local government area is an administrative section of a country that a local or municipal government is in charge of. By starting in three local government areas, we are more likely to transfer resources easily and maintain trained staff (due to brain drain and staff rotation.)

  1. Our supplier, DKT international, ships misoprostol to the registered PHC(s), Health Clinics and/or Health Posts which all have their own distributor ID.* They may also ship to a central warehouse where the EMM on-ground team will deal with more granular distribution.
  2. Using a digital form on a mobile application, the focal person at the distribution centre inputs the amount of stock received, center ID* and the date.

*Input from the MoH will help us select (a) distribution centre(s).

*We will generate and assign the distribution centre ID numbers pre-pilot.

Stage 2.1: Mother Registration and Misoprostol Supply

The first option is that a mother may visit any of the PHCs, Health Clinics or Posts to receive education, a registration number and prescription. This can also take place near community areas like markets or places of worship. In some cases, (depending on guidance from the MoH), the mother may also receive the misoprostol right away to store for use until after birth. In this case, the cycle is complete.

Stage 2.2: Door-to-Door Registration and Misoprostol Education

The second option, which will happen regardless of if the first is considered or not, includes a CHEW registering and prescribing women door to door. Married women who are not pregnant may also be registered for future enrollment in the EMM program. Given the high fertility rate, they are likely to become pregnant if married. A prescription copy is sent to a centralized database for any EMM/FAYOHI affiliated health facility to access.

Stage 3: TBA Distribution

This stage connects the misoprostol to the TBAs. Note that each TBA must pass the EMM training exam, thus receiving a certification ID.

TBAs travel to a nearby Primary Healthcare Centre, Health Clinic or Health Post to receive misoprostol. From conversations we’ve had with owners of PHCs and our advisory board, we know that the frequency of PHCs within LGAs is quite low, making it difficult and distant for TBAs. This means it is more likely for a TBA to visit a clinic, post, or dispensary.

Stage 4.1: Misoprostol to Mother and Delivery

This option involves a TBA delivering the misoprostol to a woman when she is 7 months pregnant. The reason a mother may be given this option is so that she does not have to rely on the timely arrival of a TBA on the day of her delivery to ensure she has a safe birth. If this option is selected, the mother holds on to the misoprostol until her delivery. At that time, a TBA will arrive to help deliver the child and administer the misoprostol.

Stage 4.2: Misoprostol Storage and Delivery

This option involves a TBA holding on to the misoprostol until it is time for a mother to give birth. After helping with the delivery, the TBA will administer the misoprostol.

Stage 5: Post-Delivery

Once a TBA has helped a mother give birth, the TBA will phone a data collector (using an automatic line/toll number.) The data collector may be a CHEW, however, if their capacity is full, we would consider leveraging other parties such as community volunteers or our on-ground partner’s data collectors. These collectors will be incentivized via a stipend.

An accurate number of how many data collectors we will need per TBA will be determined once we receive further data from the Ministry of Health about the number of births that occur in the pilot LGA. As a temporary hypothesis, we believe one data collector will be able to service 10 TBAs. This would mean a total of 10 data collectors for the 3-month pilot (assuming we train 100 TBAs).

The data collector will have a week upon receiving the call notification to visit the mother and record:

  • the prescription number connected to the mother who has given birth
  • if the mother and child survived the birth
  • the date and time of the birth
  • the mother and TBA’s net promoter score

For the sake of brevity and simplicity, this overview does not include many details on the data collection involved in each step. Further information about the data collection can be found here.

If you are interested in learning more about the EMM model, you can request a copy of our Model Writeup upon publication by emailing ruhani.walia@gmail.com.

Contact

Our goal is to build an intervention that local communities can sustain. We have a strong emphasis on building and testing the infrastructure to create a foundation conducive to reducing maternal mortality. Focusing on building a robust and scalable model that localized entrepreneurs will maintain ensures we can help the most people possible.

In the next five years, we want misoprostol to be readily available across Nigeria and at least two other countries. This project aims to help pioneer an international policy and strategy to reduce maternal mortality through a community-first distribution model that leverages female entrepreneurs and community health workers.

If you’re interested in learning more about EMM, the model, the training curriculum, anthropology research or have feedback, please email ruhani.walia@gmail.com or reach out via Twitter @Ruhani_Walia.

You can also visit us at emm.health or our Instagram @emm.health.

--

--