Big ideas from Dominic Mutai

The following is an interview I did with my colleague Dominic Mutai
for the PATH M&E newsletter. Dominic is a disruptive thinker in Kisumu, Kenya.

Q: Before you joined PATH you developed a mobile app to predict a patient's risk of tuberculosis, and you were funded by a BMGF Grand Challenges grant. Tell us about your app and where you see these types of predictive technologies going in the future.

We developed a TB screening mobile app that predicts a person's probability of having TB.  The app's algorithm used the WHO 4-symptom rule for investigating TB suspects; (1) cough (with or without coughing blood), (2) hotness of body or sweating at night even when it is cold, (3) noticeable weight loss and, (4) night sweats to establish the probability of the person having TB. Based on the score by the app, the person was either ruled out, asked additional questions, or recommended for further evaluation.

Photo courtesy D. Mutai

mHealth technologies are turning phones into miniature labs and instant healthcare delivery systems. Mobile apps are improving care coordination, raising the level of patient engagement, and managing chronic diseases, but the technologies could be made even more effective with better use of predictive modelling. Similar predictive mHealth apps can suffice for predicting the risk lifestyle diseases like heart disease/ hypertension/stroke, or even predicting possibility of occurrence of maternal complications. Mobile apps that will put control of the user's health in their palms and give them a certain level of confidence will be very useful. Such apps will be the first point of care providing accurate information to the users, managing doubts and urging them to take control of their health


Q: You work on APHIAplus and manage a lot of data. What types of systems, tools, or processes would you like to see adopted by projects across PATH to make your work easier. In other words, what is your dream for data management in large-scale health projects?

Yes, APHIAplus is a big project that generates a lot of data. There is a constant demand by the program for better use of data in program management and decision making. This puts pressure on the M&E team. Thus we need to be armed with the best tools that are easy to use and deliver on demand analytics to the users at their convenience. Big projects like APHIAplus should have well documented systems and processes for managing vast data. First things first will be development of Data Management Plans (DMP) or an analysis plan will guides the entire data management processes- what data is to be collected, who to collect it, how to collect it, where will it be stored and how the data is be analyzed and used. Then, systems are developed that collect and store the data. Web applications and mobile apps are lately being more accepted for their ease of use and availability than the traditional desktop systems. A central and important factor in data management is where the data is stored. Good data management practices need a central database or even a data warehouse where all the project's data is stored and managed centrally than using desktop databases and spreadsheets where every person has their copy. What I would call proliferation of spreadsheets is every M&E person's nightmare because each person has their own copy of data which at many times won't be the same data. It's hard to back up and maintain all these disparate data sources. Thus it would be prudent to invest in network based relational database system fit for small enterprises; like SQL Server. After that users are then given their right analysis tools which they are comfortable with. Excel is still (and will be) a tool that many users are comfortable with thus having means where users can connect Excel to the central database works well. SQL Server being a Microsoft product integrates well with Excel and other Microsoft products. It provides Analysis Services where users can query with Excel and dice and slice the data the way they want. Bundled with SQL server are Reporting Services and Integration Services. Integration Services perhaps to be is a magical tool in data cleaning and ETL- Extracting, Transforming and Loading data from different sources while Reporting Services provides an easy to use report portal. I see that our PATH report portal makes use of it. Then there is Tableau. Tableau is just a tool beyond tools. I met Tableau when I joined PATH, and in it, I met a wonderful visualization software close to none.


Q: Since you're a Grand Challenges winner, you must have many creative ideas. Do you have any innovative ideas for how the data you manage could be better used?

Perhaps the best thing would be where the data is sourced and how to improve that process for all involved stakeholders. Our project is anchored on the Ministry of Health systems (DHIS-2). Innovation is needed to ease documentation burden on the health care worker and assist the Ministry manage entire data flow process and data use for real time disease surveillance and tracking. There is a big documentation burden on the Ministry of Health's health care workers. Paper based registers are used and there is a register for every program area - Malaria, HIV/AIDS, TB, Immunization, nutrition. Then there are monthly reports that are expected of the health care worker to manually tally and aggregate the data from the registers to the reporting tools. This is usually a big problem as the wrong data is usually tallied and entered. Another problem is the problem of patients getting lost to follow up, defaulting or transferring out. This is a common problem to all HIV care is implementing partners and the Ministry.

My idea would be a cheap point of care Electronic Health or Medical Record (EMR) that works anywhere in Africa. We have had the introduction of OpenMRS EMR but its impact has not been very successful since it requires desktop computers and electricity. Its adoption as a point of care EMR is also limited because of the typing and its ease of use. Security has also been a problem and we have had a number of computers stolen. OpenMRS also works only for the HIV care and treatment and TB areas. An alternative would be a holistic Electronic Health Record (HER) that is easy to use, using touch screens and accepts handwritten clinical notes. It should also integrate with other systems, notably DHIS 2 and at the end of the month, the health care worker just transmits the data straight to DHIS. It should give the health care workers dashboards, alerts and clinical decision support where they can track patients and diseases burden on the real time. It should provide a means of patient accessing their data and when a patient moves, the EHR and patient data can be transferred by a click of a button to the next health provider.

A Tablet based EHR would be a good idea since it will work in the remotest places in Africa without electricity (they can be charged by simple solar chargers), they are easy to use – with touch screens and they provide hand writing support; health care workers can touch and write their clinical notes instead of clicking and typing. Data will be secured using encryption while the tablets physical security is easy to implement. Patient scheduling, interactive visualizations and dashboards will be built into the system to aid in decision making. Manual report generation will be replaced by electronic report generation by a click of a button and the right data submitted straight to DHIS eliminating another point of data entry at the district level thus reducing data entry errors. We as a program, the ministry and all stakeholders would have the accurate data, instantly while the documentation burden will be relieved off the health care workers to that they can concentrate on seeing patients. Tablet prices have been dropping and we currently have tablets going for less than 100 dollars.  Hand in hand with this innovation will be creation of policies that safeguard patient's sensitive data. Many countries don't have  the US HIPAA equivalent acts and laws that safeguard patients data confidentiality, integrity and availability.  Thus a point of care Tablet based EHR would a welcome innovation.​


Short Bio​

Dominic Mutai is a Data Manager with PATH's APHIAplus western Kenya.  Prior to joining PATH, Dominic worked on vaccine and drug trials while working with KEMRI/CDC; a collaboration between Kenya Medical Research Institute and US Centers for Disease control and prevention as a Data Manager and previously, at Columbia University's International Centers for AIDS Care and treatment Programs (ICAP) as a M&E Officer. Dominic holds a Bs\Sc Degree in Computer Science and Engineering and is finalizing his MSc in IT Security and Audit. Dominic is passionate on developing tools and devices that make use of Information Technology to solve Africa's health problems. When he is not analyzing data or writing code, Dominic spends his time with his wife and two sons in their home in the lakeside town of Kisumu.​

Knowledge brokers: the bridge to somewhere

This post is based on a blog in the PATH M&E Community of Practice newsletter. 

Why do we need knowledge brokers?

We know that very little of the health data and evidence produced is systematically used to inform programs and policies. This observation has led to the creation of entire disciplines of “knowledge translation” and “evidence-based policy-making.” One of the key observations of these disciplines is that relationships matter. We can build all the databases, write all the policy briefs, or create all the data in the world, but it won’t be used unless we invest in the social element.

In her seminal article about the diffusion of innovations in healthcare organizations, Trisha Greenhalgh said, “knowledge depends for its circulation on interpersonal networks, and will only diffuse if these social features are taken into account and barriers overcome.”

Barriers exist. Even when networks exist, people are less likely to exchange information than to engage in other types of exchanges. And when we look at the direction of information exchange, people are much more likely to provide it than request it (although neither process happens very often). Look at the networks below – the same people, but three different types of relationships. On the left, the ties represent whether they reported interacting at all during policy development; in the middle the ties represent whether they provided research evidence to one another; on the left, ties represent whether they requested evidence from one another. 

Figure 1. Policy actors in Burkina Faso according to three types of relationship ties

Why are we reticent to ask for information? Wouldn't it vastly improve our work? Yes, I found in Burkina Faso that policy actors who exchanged information were more likely to use it to inform their decision-making. 

That’s where knowledge brokers come in. Knowledge brokers can be ad-hoc and informal, or a formal role in an organization. Evaluations of formalized knowledge brokering roles in high- and low-income countries have suggested that they are effective at building individual and organizational capacity to use data. So what can we do to identify, support, or be effective knowledge brokers?

1. Brokers are bridge builders

Knowledge brokers are strategically connected in their networks such that they are able to reach many otherwise unconnected actors. They build bridges between communities. They understand the context. They possess an intuitive mental map of their networks and know where to build the next bridge. We call this metric of strategic connectedness betweenness centrality, and the graph below shows the most strategically connected health systems researcher according to betweenness centrality. This person can theoretically reach the most other actors in this community to share new knowledge.

Figure 2. Co-authorship network of health systems researchers, nodes sized by betweenness centrality

2. Brokers come from the inside 

Knowledge brokers are perceived to be more credible and trustworthy if they are embedded in the organizations they target, like ministries of health or health services organizations. Detailed network mapping and qualitative interviews in Burkina Faso demonstrated that policy actors were more likely to adopt ideas from someone within the ministry than from development partners (even though the development partners had better network connectivity). This is further evidence to refute the "two communities" hypothesis prevalent in KT, which has also been challenged by others and stands in our way of designing interventions that recognize the role flexibility of actors in policy-making.  

Our team’s work on the Gavi Full Country Evaluation is also showing the importance of trust in the provision of technical assistance for vaccine decision-making. This is not the first time that approachability and patience has been identified as necessary traits of a knowledge broker.

3. Brokers are translators

The biggest skill of a knowledge broker is their ability to translate across various users’ and stakeholders’ languages, skills and perspectives. We should all keep this in mind when discussing our work. You don’t need fancy network maps to connect with a colleague and talk about research evidence, project data, or new knowledge you have. And don't forget to ask for information -- brokering goes both ways. Go forth and broker!

Could playing games improve immunization rates?

I have been following Michael Simmons’ interesting writing on networks and relationships. His most recent article in Forbes presents research from Brian Uzzi and others about how to build stronger ties in collaboration networks. This is an area of interest for myself and colleagues at PATH and IHME working on an evaluation of Gavi support in four countries. As part of this, we are studying the structure of immunization policy partnerships – networks of policy-makers, partners, and other stakeholders involved in planning and managing national immunization programmes -- in order to understand whether some are more effective than others. Effective at doing what? Effective at doing all the activities necessary to operate a successful immunization programme, eventually leading to more children immunized and fewer preventable childhood deaths. Our colleagues at IDRC Uganda presented some preliminary results at HSR 2014 that suggested that trust was perceived by people in the network to be essential for carrying out their activities, and yet it was not distributed equally throughout the immunization partnership network they studied.

One question that comes up often is: “Fine, we know trust and good relationships are important, but can we artificially intervene on them? Don’t they just happen organically?”

Photo from Outlook Business

Trust and relationship building do happen organically, but the process of building them can be accelerated. We know this from social psychology, from network science, from management science, and Michael's many articles provide a great overview. What is compelling about Michael’s new article is that he points to activities that involve a heightened physical state as being more effective at building stronger connections. He reports Prof. Uzzi’s findings that MBA students built their deepest connections on the sporting pitch, not in the classroom, and suggests collaborators take a walk together or play catch when discussing ideas and solving problems. 

So what does this mean for immunization partnerships? Should we be walking on treadmills while preparing immunization plans? Maybe (and I do know someone in Seattle whose office recently installed treadmills). Perhaps Gavi should team up with Right to Play and introduce sports days as part of regional immunization meetings or national planning retreats. Or perhaps Bill Gates should sponsor bridge tournaments as part of the Foundation's support to Gavi. 

What sacred crocodiles and meat festivals can teach us about evidence-informed health policy

Last week my colleagues and I published a paper in Implementation Science: “Exchanging and Using Research Evidence in Health Policy Networks: a statistical network analysis.” We set out to understand which network factors might increase the likelihood that actors exchange research evidence with each other. Here's why this is important:

• When people exchange evidence, they are more likely to use evidence,
• The use of research evidence can help improve the effectiveness and equity of health policies, and;
• Evidence exchange is a social process and thus should be studied using methods that enable the measurement and modeling of the social dimension.

Using exponential random graph models, we found that network structure, more than individual attributes, explained why people provided or requested evidence in their policy networks, and whether they used it to inform their own decision-making.

To help explain these findings, I provide some anecdotes from my life in Burkina Faso.

Point 1: Leverage existing relationships

People in Burkina Faso are entrepreneurial. My good friend Ai is young but has held any number of roles as research assistant, language instructor, translator, guest house owner, CEO, etc. How does Ai succeed in these roles? In addition to her intelligence, she understands the power of her networks and the importance of trustworthiness.

People's networks are large, and effective. It would take 20 minutes to drive 100 meters with Mme Salimata Ki, my mentor in the Ministry, as she knew everyone we drove past, and stopped to either have a conversation with them or give them food or money if they needed it. My access to respondents was thanks to her, and the woman could move mountains. I would love to map her network. Actually, it reminds me of a study underway by JP Onnela at Harvard where they hypothesize that the structure of participants at the Kumbh Mela replicates the global network structure in India. Salimata is the Kumbh Mela of Burkina Faso. 

The point is: Burkinabe’s are used to playing many roles in their networks. We found that policy actors were more likely to exchange evidence with each other if they had other types of relationships together. I called this ‘layering’ in the paper but it is also called multiplexity in network terms.

What can we do? If you have evidence to disseminate, encourage people to share it with their friends, colleagues, officemates, family, tailor, butcher, etc. Journal websites do this by allowing users to share on social media, but we know that face-to-face ties – strong ties – are more likely to lead to useful knowledge exchange because they help make information 'sticky.' So instead of the Facebook icon, let’s make a ‘talk’ icon and put it on evidence outputs. Of course journal clubs are perfect for this. 

Point 2: Don’t eat too much street meat

There can be too much of a good thing, even in the world of grilled meats and evidence exchange. On the meat front, I attended the annual Festivale de grillades (grilled meat festival) in Ouagadougou with a close friend from the U.S. There was an entry fee, so we were incentivized to eat as much as possible. We egged each other on to eat and eat, to the point that we began to perspire and felt dizzy, and this lasted for at least 24 hours. I believe clinicians refer to this as "meat sweats." I doubt we would have succumbed to the meat if we had been with a more diverse group of friends. Salimata would not have let me eat so much street meat!

In our study of evidence exchange networks, one policy issue – community integrated management of childhood illness -- showed much higher frequency of exchange (and use) of evidence than the other issues. This network also showed a propensity to form cliques, aka triangles. Triangles represent a social process that we observe in networks at a greater frequency that can be explained by chance alone: people are more likely to form a relationship if they have a friend in common. When I see a network with lots of triangles, I think, "empirical signature of cohesive and congruent policy communities!" (or, female adolescent friendship network). Indeed, analyses from my other chapters show that despite all this exchange and ‘use’ in this network, evidence wasn’t used instrumentally to solve tough problems about child health. It was used symbolically to justify pre-determined policy positions, mainly of development partners and donors. In other words, people came together to eat meat, it was tasty (i.e., the Lancet Child Survival series), they ate a lot, but eventually it became difficult to discern whether the meat still tasted good or  they continued to eat because everyone else was.

In other words, we need to support evidence exchange and use, but we also need to make sure that people have the capacity to find and interpret the evidence and that network structure does not encourage group-think over innovation.

What can we do? Try to avoid attending meat festivals with your best friend. Instead, go with a diverse group. Invite some vegetarians! You will be more likely to try new things, and your stomach will thank you. The National AIDS Council (CNLS-IST) did this in the early 2000s by requiring representation of civil society and PLHIV, ensuring the diversity of the Council and entrenching the capacity and access of civil society organizations in the HIV policy domain. This has had a lasting positive effect on the HIV policy network structure and its effectiveness in Burkina Faso; we found the HIV actors were more likely to use evidence when it was provided to them, and this network ultimately produced and used evidence to instrumentally achieve policy change. 

Point 3: Use person-net, not Internet

If I had a nickel for every person (from North America and Europe) who said, “Oh! You’re studying Facebook in Burkina Faso,” I’d have about 85 cents (which is significant). Social media and online social networking is exciting, but not very useful in Burkina Faso circa 2011.

Poor internet infrastructure poses one of the biggest challenges to the fidelity of many of the knowledge transfer, policy translation, and implementation science interventions that are shipped off to West Africa and elsewhere. When I arrived in Burkina in 2011 I was kindly given an office in the Ministry of Health. I had no internet where I was living, there was really no public internet (although I single-handedly supported the tonic water industry in hotel lobbies), and the connection at the Ministry was really slow. It was virtually impossible to read emails, much less download a research article to inform policy with. So I looked around my office and realized that if I bypassed the wifi router and plugged the Ethernet cord directly from the wall into my computer, the internet performed much faster. Did it cross my mind that this router might be used by other offices? Briefly, but no one said anything. And then ONE MONTH LATER someone knocked on my door and asked whether they could look at the router. I had turned off the internet to the entire Department. 

Other common reasons for internet outages included:

• Deep-sea underground digging in Guinea or Cote d’Ivoire had cut the internet cable. THE internet cable. For the continent.
• The popularity of copper jewelry meant that local internet cables were often dug up to make bracelets.

Needless to say, downloading research articles is not high on your list of things to do when you have a brief window of connectivity. I found that this meant that certain actors (i.e., development partners) had better access to evidence than others, leading to the inequitable distribution of normative power.

What can we do? We should be much more creative about the media we use to disseminate research evidence. Knowledge brokers, policy dialogues, printed policy briefs are all possible examples. Salimata puts research evidence on a CD to share with colleagues. Evidence at your fingertips. 

Final point: My network made me do it!

I shook hands with a crocodile in Burkina Faso. My personal attributes – risk preference, IQ, education, past experience – would not have predicted such a thing. But my husband encouraged me, and Salimata facilitated it, and so I went to a village of sacred crocodiles and sat on one. We know about social influence and heuristics in decision-making – both came into play in my case -- but what is even more compelling, and also more abstract, is the role of social structure on our decisions. That is, we are influenced not only by what our friends are doing, but how are friends are connected to each other and to others. And this is the main point of our paper. Controlling for individual attributes and shared traits, whether or not two people exchanged evidence was best explained by their network's structure.

What can we do? As noted above, we can design networks that are more conducive to appropriate evidence exchange and use. I'm beginning to work towards this with NetworkRx, a very beta simulation tool to see the effect of network structure on policy outcomes. And stay away from crocodiles.

Hope you find the paper delicious.

Who is the Kevin Bacon of Health Systems Research?

Earlier, I wrote about network mapping at the Third Global Symposium on Health Systems Research. We mapped participants' social ties, collaboration ties, and information seeking ties as reported by them on an online survey. 

Today I will show you the collaboration network generated through this process, as well as the 'real' network of co-authorship relationships generated through data mining. 

We asked: "With whom have you collaborated on health systems or policy research in the past two years." Here's what we got: 

In this graph, 91% of all nodes belong to a single, connected component. And, any two people in this network are connected by four or fewer people. We knew that the HSR community was… well-connected… (insular also comes to mind), but really? 

In network science this is referred to as the 'small-world phenomenon.' Ever heard of 'six degrees of separation'? This refers to the fact that any two people in the United States are connected by six or fewer acquaintances (as confirmed by the famous postcard experiment of Stanley Milgram). Or that everyone in the movie world is connected to Kevin Bacon* by four or fewer connections (you might have also heard six). 

In other words, Kevin Bacon is connected to the most other actors in the fewest steps. Why does this matter? Imagine that you had an important message to deliver to everyone in your network, and that the message travels from person to person starting with one person. Who would you give it to for fastest diffusion? Test your intuition here by trying to slow the spread of an infectious disease in a network. 

Bottom line, you want to find the person who can access different parts of the network in the fewest steps. They sit on paths between the most number of people. They have the highest betweenness centrality. 

But, our network data were not complete. Only about 70 people completed the survey, so we have plenty of missing edges and attributes for those who didn't complete the survey. 

This begs two questions: 

1. Can we mine citation data to get a fuller picture?

2. Who is the Kevin Bacon of HSR?

To answer the first, yes, we can use publicly available citation data. I searched ISI/Web of Science for all the names of Health Systems Global members (which was our proxy for who attended the conference), and downloaded the names of everyone they have published with since 2012, mirroring the survey question. Using Sci2, I constructed a co-authorship network with our respondents and their co-authors as nodes, and co-authorship relationships as edges/ties. 

This is the network we get: 

Complete co-authorship network of survey respondents

This network has 5576 nodes, representing 1433 publications since 2012. 75% of these authors are connected in one large component, with a relatively dense core but sparser periphery structure. We also see many more components -- a total of 191 unconnected smaller networks. No longer is everyone as connected. Now, like in the real world, everyone is connected by 6 degrees of separation, compared with 4 above. 

Who has the shortest paths to everyone else in the network? Who is our Kevin Bacon?

Co-authorship network of survey respondents, nodes sized by (scaled) betweenness centrality, 

six highest betweenness scored labeled 

John Lavis has the highest betweenness centrality. John** was on the greatest number of paths between other actors. He is in a position to best connect otherwise unconnected parts of the network, and to reach all other nodes most efficiently, all else being equal.

When it comes to ideas, John Lavis is a broker (which is appropriate, as he studies knowledge brokers). He can:

• Access new ideas from diverse parts of the network

• Disseminate new ideas to diverse parts of the network

• Act as a translator between groups

What makes John different than the rest of us? Since 2012, he published 30 articles with 78 different people. That's an average of 2.6 new co-authors per paper! Imagine how many people they are connected to. And the fact that you might not recognize all the names is a good thing for innovation. 

Does he intentionally try to broker the world of HSR? Maybe (we should ask him), but he also has many factors in his favor: he's an expert, so diverse authors seek him out and he's based in Canada where institutions are smaller and thus one cannot rely on within-institutional collaborations alone. 

Did you happen to notice that of the six highest scores in the HSR network above, four are Canadian by birth? Natural peacekeepers?

Back to the big picture. What does this mean for the field of Health Systems Research?

Connectivity is a good thing -- it's how we share and interpret ideas and knowledge, access resources and opportunities, and function as a society -- but there are two concerns with connectivity.

1. More connectivity isn't always better

2. The distribution of connections matter

First, there is a danger of networks becoming overly connected. Multiple, redundant connections with the same groups of people (represented by high network density or many triangles in a network) leads to 'group think' and may stymie innovation. Densely connected network cores begin to operate like an echo chamber.

Second, we need to think of how the connections are distributed across network nodes. At the node level, centrality (i.e., number of connections) equals social capital. We learned that betweenness centrality is one particularly powerful measure of capital, brokerage, and potential influence. The HSR co-authorship network had extremely large variation in people's betweenness scores, meaning that there is a lot of "power" in this network, and it is not evenly distributed.

So high network density is bad? 

No. We need to separate connectedness into two components: density (i.e., the proportion of actor pairs who are connected) and centralization (i.e., the distribution of connections across the network). Density in and of itself isn't bad; sometimes all those repeated connections are necessary. It is the centralization of those connections around one or a few nodes that is not ideal for innovation. We used to think that density and decentralization were at odds but we now know they can co-exist.

Reza Yousefi-Nooraie found that academic medical research teams were more productive if their networks were dense, but also decentralized and open to external networks. A recent study in Nature, corroborating other studies of knowledge-intensive organizations, found that dense networks aid knowledge transfer when the knowledge is complex. The same study also found that team performance was positively associated with the strength of expressive ties (i.e., friendships).

However, I found in Burkina Faso that while dense networks aided the dissemination and spread of research evidence (i.e., complex information), the same density protected the policy status quo and prevented the new ideas from being adopted.

So what does success look like for a HSR network? An ideal academic co-authorship network would be productive (i.e., publish a large quantity of research) but it should also be innovative (i.e., willing to adopt new ideas) and relevant (i.e., research priorities are identified systematically and democratically and research findings are used). What would that look like as a network?

• Dense, with strong ties, including layered ties based on friendship and trust

• Decentralized, with brokers at the margins who can access external networks and new ideas

• Diverse, with varied representation of people who can translate across communities

How do we move towards such a network?

 

As individuals, we can continue to develop strong relationships with co-authors that lead to intellectual productivity. The HSR network, particularly the core, could be more decentralized. We could all be better at developing new collaborations with people outside of our existing networks. Finally, but related to the previous point, we can do better at building diverse research teams, as measured by geographic location, sex, age, discipline, and job role. At the network level, there are interventions that could be applied, but I can't think of examples where co-authorship networks are actively governed. What about research collaboration networks? Is this the role of Health Systems Global (or donors)?

* Note that newer data show that Sean Connery is on more shorter paths than Kevin Bacon.

** John was my PhD supervisor, but this in no way informed the network analysis.

Burkina Faso: Community Health Revolution 2.0?

Ouagadougou is burning, and it looks like the end of Blaise Campaoré’s 27 year rule. Forget Ebola on its border, Burkina Faso is experiencing ‘Revolution 2.0,’ a nod to Thomas Sankara’s revolution in 1983 and the bold social programs that followed. 

While Sankara’s rule was brief (he was killed in a coup led by his best friend, Blaise Campaoré – yes, that Blaise), he created a legacy of community health care that exists to this day. Sankara introduced a policy of «un village, un PSP» [one village, one primary health care post]. This was in the early 80s, post Alma-Ata, when populist rulers were beginning to stand up in Africa and elsewhere and recognize the importance of primary health care, on one hand, and community empowerment, on the other. Sankara’s revolution was supported by the people – as seems to be the case today – and people need health care. A small health hut was built in each village and the community nominated a lay health worker to work there. She (normally, but sometimes he) received small supplies to treat wounds, provide paracetemol for fever, and perhaps refer the sick to health facilities further afield. Their work was voluntary.

In the years since Sankara’s death, the government paid little attention to sustaining or expanding community health care programs. In some cases the community health workers continued their work without external support. In other cases, NGOs intervened and used the CHWs to deliver their health programs. As a country hit hard by malaria, CHWs soon became essential human resources for preventing and treating malaria in communities, and were rolled into the country’s Home Management of Malaria program in the late 1990s. 

More often than not, the fate of CHWs followed the trends set by global policy elites. The Bamako Initiative, decentralization, and user fees? Sure, let the CHWs charge patients a fee. A shift towards facility-based care? Sure, tell CHWs to refer all sick kids to health centers for IMCI; make birth attendants illegal. The rise of public private partnerships? Yes, CHWs can dispense chloroquine for Roll Back Malaria and the Global Fund. What, user-fees aren’t working? IMCI isn’t working? Chloroquine is no longer effective?

When I interviewed health policy-makers, health bureaucrats, politicians and civil society actors in 2011, it was common for respondents to become nostalgic at the mention of community health care. “Un village, un PSP,” they would whisper. Many had started off as junior civil servants or medical officers in rural districts, responsible for implementation of the policy. No one forgot the slogan. Ideas live, even if their creators do not. 

Despite the roller-coaster of policies involving CHWs, the fact is, they remain in communities and they remain a viable policy instrument. Recent efforts to retrain them (or recruit younger, more educated workers) for integrated community case management have been partly successful, but I found a consensus amongst actors that greater emphasis and resources needed to be directed towards the implementation of the country’s revamped community health policy. 

The community health policy and related community health programs are more than nostalgia. The effective implementation of such program are necessary to address the persistent high rates of under-five mortality, with major causes of malaria, diarrhea and pneumonia. Local and international NGOs have had success with community-based treatment of severe acute malnutrition – the leading risk factor for child morbidity and mortality – but resources and political will are needed. 

This week’s Revolution marks a crossroads. There will be a new leader, and they could throw their political will behind community health strategies as did the President of Niger to tremendous success. Let’s hope so.

And because this is a network blog, a word about networks and social movements in Burkina. One respondent told me that policy change required a «masse critique» [critical mass]. The idea of networks resonated with everyone I spoke to, who acknowledged the role of relationships in every day life. Civil society actors, particularly, wanted to know how to use their networks to achieve social change. Today #lwili is trending on Twitter, faster than CNN can find Burkina on a map.

Here is the network of policy actors involved in the development of the community integrated management of childhood illness (C-IMCI) policy in Burkina in 2011. Note that the most influential actor, as measured by a combination of betweenness and degree centrality and interview and documentary data, was not the most central, not the highest ranking, but a mid-level technical bureaucrat in the Ministry of Health. Anyone can start a revolution.  

Mapping Networks from HSR 2014

At the recent Third Global Symposium for Health Systems Research, Jeff Knezovich and I asked participants to complete an online network survey. Our aim was to map the networks (social, collaboration, information seeking) of conference participants. We had some technical glitches with the online tool, slow internet access, and the apathy towards completing a survey that is commonly observed. The experience confirmed:

1. Network mapping, especially in internet-limited settings, should be done off-line;
2. Network mapping (probably anywhere) should be done face-to-face. Otherwise respondents are unlikely to respond;
3. One should always pilot their data collection tools!

The idea of mapping the network seemed to resonate, but in total we had only 71 responses (give or take – cleaning the data outputted from the app was more grueling than a hike up Table Mountain!). Nevertheless, let’s see what these data look like. I will be running the analysis in Rstudio, using the statnet suite of packages. Feel free to download the .csv files and work along with me.

Step 1: Make sure the packages are installed

Install the latest version of ergm and sna: install.packages('ergm'); install.packages('sna')

library(ergm)
library(sna)

I also changed the color palette because I don’t like missing node attributes to be colored as black. It just doesn’t look nice. As a note, these colors are also good for color-blindness, and for printing in grey-scale.

col.list <-c("white", "darkblue", "cornflowerblue", "darkorange1", "darkred")
palette(col.list)

Step 2: Import data and convert to network

ONASurveys.com exported the data as an edge-list, and I had to do some extensive work to delete duplicate names, ensure IDs matched, etc. But you can use the cleaned files. There is one for each network, as well as an attributes file that is used for all the networks.
Save the files somewhere and set that folder as your working directory in R.

setwd("~/Dropbox/Dissertation_jan4/Conferences/Cape Town 2014")
attr <- read.csv(file="attr.csv", header=T, stringsAsFactors=FALSE)
social <-read.csv(file="social.csv", header=TRUE, stringsAsFactors=TRUE)
spre <- network(social, matrix.type="edgelist", directed=F)
smat <- as.matrix(spre)
snet <- network(smat, matrix.type="adjacency", directed=F, vertex.attr=attr)

Note that I had to coerce the edgelist into a network object, then into a matrix (to match up with the attributes), then back into a new adjacency-type network object with attributes attached.

summary(snet)

The summary() command shows us the network vertices, edges and density. Note that this is the entire network of all 1515 participants, 90% who did not complete the survey. Let’s plot that network to see what it looks like (and then we’ll get rid of the isolates).

Step 3: Plot the network

plot.network(snet, edge.col="darkgrey", vertex.border="black")

You can see a cluster of activity in the center, with edges in grey. But otherwise this is not a helpful graph. Let’s delete isolates and check out the summary stats.

sno_iso <- delete.vertices(snet, which(degree(snet)<1))

Run the summary(sno_iso) command to see ALL the details, or simply:

centralization(sno_iso, degree, mode="graph")

## [1] 0.1035052

network.density(sno_iso)

## [1] 0.007237999

Still a pretty sparse network! (Although these are incomplete data, so we can’t say much about the actual network density or centralization). The exact question was: “who did you, or do you plan to have lunch or dinner with during the conference.” One can imagine that there are likely clusters of friends/colleagues who are likely to socialize, and fewer connections between these clusters. And what drives the propensity to form social ties? Let’s look at a few graphs before we test hypotheses in ergm models.
Do people socialize with others from their region?

s2coord <- plot.network(sno_iso, edge.col="darkgrey", vertex.border="black")

plot.network(sno_iso,  coord=s2coord, vertex.col="region", edge.col="darkgrey", vertex.border="black")
legend("bottomleft", legend=c("Africa", "Americas", "Europe", "South-East Asia", "Unknown"), pch=21,
       cex=1, pt.bg=c("darkblue", "cornflowerblue", "darkorange1", "darkred", "white"))

Again, it is somewhat difficult to tell with the missing attribute data, but it doesn’t seem as though there is clustering by region. (On a side note, if these data were very important, I could look up all the alters’ regions. For smaller networks this would certainly be worth it).
Is sociality based on similar organization?

Hmm… first of all, we see that most of our respondents are from research organizations. Second, they seem to be more central in the network. Are they more likely than chance to eat lunch with other researchers? We will find out soon. But first, let’s examine by age.

Finally, which nodes are in the most strategic position to broker other nodes? This is measured by betweenness centrality, and can be applied to understand who the brokers are, and how to most efficiently disseminate ideas or information. We will calculate the betweenness centrality scores for all nodes, and then size our graphed nodes according to their betweenness.

s2between<-betweenness(sno_iso, g=1, gmode="graph", cmode="undirected")

plot.network(sno_iso, coord=s2coord, vertex.col="region", vertex.cex=s2between/1500, edge.col="darkgrey", vertex.border="black")
legend("bottomleft", legend=c("Africa", "Americas", "Europe", "South-East Asia", "Unknown"), pch=21,
       cex=1, pt.bg=c("darkblue", "cornflowerblue", "darkorange1", "darkred", "white"))

The most strategically located brokers are from South-East Asia.

Step 4: Construct ergm models to test hypotheses about why conference participants socialize with each other

Ok, now let’s examine these in ergm models. Exponential random graph models (ergm) are a class of logistic regression model that allow us to test hypotheses related to dyads, i.e., network ties/edges. See the statnet website for a list of ergm resources. I highly recommend “Birds of a Feather or Friend of a Friend” by Goodreau, Kitts and Morris (2009) for both a master class in ergm modeling as well as a wonderful application to adolescent friendship networks.
Unlike traditional statistical models, where the covarariates are some function of the units of analysis, ergm models allow us to alo reprensent covariates that are functions of the network itself. I usually build my ergms in two waves: 1. A set of attribute-only models where covariates are tested separately and then added to the final model stepwise if they improve model fit; 2. A set of structural-only models (following the same process as above).
Starting with the attributes, let’s test each in a model with an edges term, which is like an intercept in a traditional regression model.

smodel.02 <- ergm(sno_iso ~ edges+nodematch("region"))
summary(smodel.02)

## 
## ==========================
## Summary of model fit
## ==========================
## 
## Formula:   sno_iso ~ edges + nodematch("region")
## 
## Iterations:  20 
## 
## Monte Carlo MLE Results:
##                  Estimate Std. Error MCMC % p-value    
## edges            -3.66743    0.05216     NA  <1e-04 ***
## nodematch.region -3.02832    0.14465     NA  <1e-04 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
##      Null Deviance: 82741  on 59685  degrees of freedom
##  Residual Deviance:  4375  on 59683  degrees of freedom
##  
## AIC: 4379    BIC: 4397    (Smaller is better.)

The nodematch.region coefficient reports the change in log odds of a tie existing between any two given nodes if they share the same region (as compared to not sharing a region). The exponent of -3.02832 is 0.0484. Two actors are less likely to socialize if they are from the same region. BUT! Big caveat: so much of the alter data is missing, that the majority of ties are between known regions and unknown regions (i.e., different regions).
Let’s look at some models with structural covariates. What are these magical structural covariates? They are underlying social processes which have been documented empirically to occur more than chance alone. Today we will examine transitivity, or triangle formation, which describes the propensity for people to form relationships with ‘friends of friends.’ Transitivity has many implications. Think of triangles, literally, as cliques. Cliques might be fun for lunch, but they are not conducive to exposure to new ideas, innovation, behavior or policy change, etc. In the first model I test whether social ties are more likely to exist if they close a triangle. We expect that a person is more likely to socialize with their friends’ friends.
*A note about missing edge data: While our structural models will not be affected by missing attribute data, they will be affected by missing edge data. We only know the edges of respondents, not the edges of the alters

smodel.06 <- ergm(sno_iso ~ edges+gwesp)

summary(smodel.06)

## 
## ==========================
## Summary of model fit
## ==========================
## 
## Formula:   sno_iso ~ edges + gwesp
## 
## Iterations:  20 
## 
## Monte Carlo MLE Results:
##             Estimate Std. Error MCMC % p-value    
## edges       -5.19997    0.05653      0  <1e-04 ***
## gwesp        0.99650    0.17815      0  <1e-04 ***
## gwesp.alpha  1.18096    0.10367      0  <1e-04 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
##      Null Deviance: 82741  on 59685  degrees of freedom
##  Residual Deviance:  5011  on 59682  degrees of freedom
##  
## AIC: 5017    BIC: 5044    (Smaller is better.)

The gwesp term measures the change in log odds of a tie forming between two nodes given that this tie will close a triangle. Yes, even with missing edges, ties are more likely to exist if they close a triangle between three nodes. This is the result we expected for sociality, but let’s check to see whether this happens with collaboration ties. We would expect that people are more likely to collaborate with their collaborators’ collaborators.

cmodel.06 <- ergm(cno_iso ~ edges+gwesp)

summary(cmodel.06)

## 
## ==========================
## Summary of model fit
## ==========================
## 
## Formula:   cno_iso ~ edges + gwesp
## 
## Iterations:  20 
## 
## Monte Carlo MLE Results:
##             Estimate Std. Error MCMC %  p-value    
## edges       -4.96450    0.06448      0  < 1e-04 ***
## gwesp        0.93190    0.25254      0 0.000225 ***
## gwesp.alpha  1.27732    0.12876      0  < 1e-04 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
##      Null Deviance: 50343  on 36315  degrees of freedom
##  Residual Deviance:  3729  on 36312  degrees of freedom
##  
## AIC: 3735    BIC: 3761    (Smaller is better.)

Indeed, people are much more likley to collaborate with their collaborators (based on our incomplete data).
There are a few ways to deal with missing edge data:

• We could have asked the respondents to report their alters’ edges (this is typical in ego-network sampling, but its accuracy depends on the relationship being measured)
• We can remove the nodes we didn’t interview (and thus their edges). This will leave us with a network of complete edges, but not a complete network. I.e., the network we will be left with is not a real network. But neither is the missing edges network…
• We could try to impute edges based on attribute data. Wait! This is what we’d do if we didn’t know that edges are predicted not just on attributes, but also on network structure! Network dependencies make it difficult to impute edges. Man, these networks! Next time

Finally, we could mine existing data (i.e., citation data, Twitter data) to construct relevant networks. Maybe next week…