Published in


The Future of Global Digital Health: Notes from GDHF 2017

Originally published December 6, 2017 on

Where are we now? Where do we want to go? How do we want to get there? These are the questions our co-founder and managing director Hannah Cooper proposed at the closing plenary at the Global Digital Health Forum this year. We’ve included her slides and speech below if you’re interested in learning more about some of the great ideas that she talked about and where we think the future is heading.

Thank you to the Global Digital Health Network for hosting the conference and to all that attended!


As hard as it is given the headline, I’d like you to pay attention to these two ads that popped up the other day when I was reading the paper.

It’s as if someone is reading my mind…. How do they know I’ve been thinking about changing credits cards, maybe finding one that gives me better deals when I travel, more rewards, access to more lounges?

Or that I’m thinking about buying Google Home or Amazon Echo? That it might be fun for my young kids to ask Alexa to play them their favorite songs.

How many of you have had this happen? Where you have an add pop-up, seemingly innocuously, suggesting that you buy the very item that you’ve been contemplating?

These are just some examples of how the private sector, and e-commerce in particular, is generating insights into customers and segmenting its customer base to offer products that meet their needs.


Let’s compare that experience with where we are now in global digital health.

We have a lot of data, electronic medical records, repositories of patient-level data, program data.


…but we’re not adequately using those data to generate insights

Instead we’re finding ourselves in situations like this:

Thousands of data elements, heavy reporting burden, emphasis on collection.


As we have seen this week, through the wide range of excellent presentations, this is really starting to change.

As we move forward, we will likely be able to do two things:

1 — Generate population level insights that can inform policy.

I wanted to highlight some work that we doing with the Digital Impact Alliance (DIAL) and the Government of Malawi, using anonymized location data from telco providers.

  • Triangulate telco data (pop density, footfalls, gatherings), UNICEF population & GIS data, with disease burden
  • Develop a dynamic model that helps MOH modify health posts placement
  • Decision support: better data use in routine budget execution & program planning.

2 — In the future, we’ll also likely be able to create individualized profiles of health-system users to help us more precisely target health service.

In a way, this is what the Washington Post did in the example of the ads I showed at the start of my presentation.

It is important to flag, however, to know that building these profiles comes with serious risks: Governments could use this data to monitor their populations, suppress dissent, or redirect resources away from disfavored groups.

As was raised in a number of sessions this week, we need to clearly commit to protecting privacy. Otherwise companies, organizations, and governments will understandably be reluctant to share their data, even for population-level analysis.

We need to establish trust by building privacy protection into our efforts from the beginning.

That being said, think about what we could do if we were able to better-match health services to need and preference.

We could see an increased health-seeking behavior and health outcomes. Like those ads that targeted my needs, that seemed to read my mind, in global health we’ll be able to do the same. It is exciting to think about what this means in terms of how we could, for example, target prevention & behavioral interventions.


What does data use mean — what is it like to experience — how do we know when we’ve achieved a “data use culture”?

As an M&E person — it is helpful to make sure we all have the same understanding of where we are going — in order to move towards a common goal.


My colleague, Tyler, and I talk a lot about what it’s like to experience a data use culture change. We’ve had the opportunity to work for a number of different organizations around the work, in different settings.

The one place that we keep coming back to is our experience working for PEPFAR (President’s Emergency Plan for AIDS relief).

In 2014 with the appointment of Amb. Birx, who was highly focused on ensuring the HIV response was centered around data — so that the program and its resources (which were flat-lining) could effectively match need. We worked with her to develop the guidance and implement this new data driven resource allocation approach.

It was exciting to bear witness to this extremely rapid turn around of a massive ship (or aeroplane) as it was still moving.

We often ask ourselves why it was particularly successful? How was it that change happened and this program — which spans over 50 countries, 7 government agencies, and has annual budget of $6 billion dollars — was able to become so data driven in such a short amount of time?

We think that it is because the guidance was directly tied to resources, there was structure approach, tools and standard methods, incentives were aligned. So an effective combination of a carrot and stick approach. Alain Labrique (carrot stick graph).

We’re not sure if this approach is able to be replicated elsewhere. There were a certain set of conditions that made it possible in such a short amount of time, and without a doubt there are a lot of things that could be improved in both the development of the guidance and its execution. That being said, it is interesting to think about what we can learn from such an approach.

It is important for us to think about, describe, document what we think a data use culture looks look. Document data use culture change when we experience it, and think through what we can learn from these experiences.


To follow-up on my PEPFAR example, PEPFAR is just swimming in data. Data that is being used for the most part internally to make decisions about the program. This also raises the issue of who data belongs to and who can access it?

So we currently have all of these awesome pockets of data — that unfortunately aren’t accessible except to those that created them.

What could be different in 5 years if we changed that?

Data owners, especially the large organizations and multilaterals — need to come together around a common set of data standards on how data is warehoused and made publicly available.


In addition, we’re not effectively providing feedback and data to people at lower levels. For example, mobile phone data — the landscape is changing so dramatically, how are we leveraging what people already have in their hands?

We have a plethora of apps focused on the collection of data, not on the access of results and decision support, i.e. how to understand the data that people have in their hands, and how to use it.

Currently so focused on collection, we are not effectively providing feedback and data to people lower levels — and this is preventing us from truly creating a data use culture.

[Luzi Orphan Care CBO example]


Despite millions of dollars in data-related training, three recent studies in Malawi showed that 45–55% of data handlers at point of care had never been trained in data management and use.

How is this possible? Could it be that we’re not providing a custom approach to training — we’re not tailoring training to peoples’ different skills sets, backgrounds, attitudes?

In addition, training comes with incentives- perhaps the wrong people are coming to the training? Instead of sending the janitor who also doubles as a data entry clerk, it is the facility in-charge who is attending the training? Or selection of who gets to go to the training is based on who’s turn it is, or the need to reward a hard worker.

Wouldn’t it be great if in 5 years we had a consolidated repository of training: who has been trained on what, what their backgrounds are; and from an investment perspective if we could figure out where the gaps are and which people are being missed both in terms of capacity and geographically?

How can we identify true agents of change? Could we have better mechanisms to find people who are particularly adept at using data, and use them to light the fire in the health facility or ministry?

Better mentorship — both in the country or outside the country — a paired mentoring model that would allow partner staff with mentors to help them think through and solve specific questions they are grappling with.

We should form better links with universities, and think more for example about pre-service curriculum. For example as my colleague, Director for Policy & Planning in Ministry of Health in Malawi mentioned yesterday in our panel discussion: currently, in the pre-service curriculum for health workers in Malawi, bio-stats is not for credit, it’s optional. That needs to change.


Currently, either we don’t align incentives and data use or there is a misalignment of incentives.

Here is a slide from a discrete choice experiment we conducted in Malawi — looking at incentives and data use. And you can see there is a big mismatch between current preferences and what staff are actually receiving.

For example, I would have assumed that a big incentive for staff would be presentation at a national conference in the capital. But as it turns out — most people preferred a certificate in data management. In some places, with some groups, a certificate means a lot — and it appears that Malawi is one of the places.

Bottom line — we can’t expect our preferences and incentives that motivate us — to motivate others — particularly around data analysis and use. Hence the need to study this further. Discrete choice experiment is a nice way to come up with some number behind a some time tricky to quantify area.


We need to better understand the power of the sematic web.

The semantic web “provides a common framework that allows data to be shared and reused across application, enterprise, and community boundaries” Better data documentation and standards that enhances harmony of data sources across the web.

Seems basic, but we can’t even begin fathom the power this unsexy and unglamorous work unlocks.

Data janitorial work — data hygiene as I call it — has the power to democratize big data.

We don’t all need to be data scientists to enjoy the insights better linked data will produce. We just need to line things up more effectively and let some amazing learning tools and bright minds step in.

Brett Hurt CEO and co-founder at said: “The NSA gets it. Palantir gets it. Facebook, Google, they get it.” But the rest of us aren’t there yet.

We can’t expect machines to really make our lives easier until we make it easier for machines to understand our data.

For this to work, the development community has to rally around data hygiene as a first principle and actually mean it.


To achieve any of these things — we have to get over our hesitance to share our work, warts and all.

This will push to us to where we really want to be in 5 years.

About Cooper/Smith

When implemented correctly, data collection and analysis ensures that programs succeed and achieve actionable results. In international development, that means concrete improvements for those who need it most.

At Cooper/Smith, we use hard data to increase the effectiveness and efficiency of development programs worldwide.

Write to us at or visit to learn more about our data-driven approach to health and development.




Cooper/Smith uses hard data to improve the outcomes of health and development programs worldwide.

Recommended from Medium

Parse date columns with Pandas read_csv()

The Importance of a High Quality Mattress.

Part 1: How to use your data to fight COVID-19, a roadmap for countries in Sub-Saharan Africa

The Things They Said in the 70's…

Cartoon drawing of a speeding car with an electric plug flying behind it, while there is a Welcome to NJ sign with a palm tree in the background, and a speed limit sign written for kilometers.

Pytorch: Examining the Titanic Sinking with Ridge Regression

Plot a Treemap in Python

COVID-19 Analysis on 3D Bioprinting Market Size, Growth Estimation, Key Players, Share Value…

Progress update on Data Analytics for Canadian Climate Services

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store


We use hard data to increase effectiveness and efficiency of health and development programs worldwide.

More from Medium

What dogs can teach us about noticing the things that most people overlook

Monte Carlo Hits the Circuit Breaker on Bad Data

What are the big picture contents of Indonesian most controversial laws during 2019–2022?

Computational Thinking: A Secret Ingredient in the Making of a Unicorn Data Analyst