by Andrea Fletcher
Recently while visiting the Dimagi office in Boston we got into a heated discussion about the difference between a geek and a nerd (I know – nerds debating whether or not they are nerds definitively make them nerds). I responded by pulling up the following Venn Diagram:
Nerd vs. Geek. vs. Dweeb vs. Dork
Hence forth we can all agree on the proper definition of when someone is a nerd, or just geeking out.
EUREKA! This argument lead to a moment of inspiration — in order to better understand a situation you need a visual representation of exactly what is going on.
As soon as I pulled up the Venn diagram our discussion was stimulated and we were able to gather a consensus on issues — we need to do more of this with data out in the field. We need a visual representation for presenting information.
I am a self-described “data junkie.” While I like the math behind data, anyone can slap a bunch of numbers together and say that they came up with something interesting. Rather, what I like about data is what you DO with that information once it’s analyzed. I like how data can point you in a direction and show you how to prioritize different options.
I like when data tells the a beautiful visual story. I like “pretty graphs.” I think graphs are beautiful – I am actually quite obsessed with them (this probably puts me in the geek/nerd category, but definitely not in the dweeb section).
Let me share with you some of my all-time favorite “pretty graphs” to further illustrate how graphs/ pretty visuals can make a big difference in an understanding of what is happening.
The first is a graph from worldmapper.com, a website that takes maps and uses health data to draw proportional maps of the world.
Childhood Diarrhea Prevalence Worldwide presented in a map/graph (What is better than a graph? Cartography and a graph all in one!!)
“Pretty graph” number two comes from gapmider.com and is titled “Africa is not a country.”
“There are huge differences between the countries in Africa. It’s misleading to treat them as one and the same. Life expectancy in Cape Verde is for example 22 years longer than in Swaziland, both are African countries.” –Gapminder.com
If you haven’t played on gapminder.com it is worth it to go on and watch a few demonstrations (Hans Rosling is a fantastic speaker) but this shows how a bubble graph can be used to demonstrate really powerful and interesting relationships in simple data. It tells a story of life expectancy versus income per person (GDP) and shows how the world doesn’t exist in the first world/third world paradigm that we often use, but rather a continuum.
This third graph is close to my heart. Having worked with HIV/AIDs orphans prior to coming to Dimagi, this graph is a big part of how I ended up back in Africa. It depicts the rise of the orphan crisis that is occurring as the result of HIV/AIDS, and although it is a sad story, it is one that the world must hear. This graph is over a decade old and we still haven’t fully grasped the scale of the situation.
Every time I look at this I can’t help but as the question – “Now what?”
My final graph is funny and ironic and there for awesome:
My favorite pie is definitely blackberry.
NGOs should use data to drive decision-making
Ok, so hopefully I have now convinced you that graphs are cool, pretty, and tell amazing stories of what is going on in global health. So what? How does this make a difference in the life of a community healthcare worker, or a patient in need of care?
“Pretty graphs” are a way of interpreting data – and data can relay powerful information. The key to all of this is that we use the data, and therefore the graphs, to drive our decision-making. For some reason this is often a tough sell in global health. Funds are low and good data can be expensive. That doesn’t mean that it can’t be done or that it isn’t important, it just means that we need to find ways to make it easier.
In our personal lives we use data to drive decisions all the time. Before you choose to buy a house you check out the crime rates in the area, the the local schools, and property values. Other sectors use data-driven decision making regularly as well. Teachers use test grades to determine which students need more help and whether or not they understand the material. Modern medicine is inherently data-driven, using lab results and tests to determine what the diagnosis of a patient is and whether or not their treatment is working.
Yet for some reason this often is not transferred to the setting of global health. Global health often relies on what I like to call “fuzzy data”—that is unreliable and often inaccurate data, and the lag time can be incredibly slow. It can literally take years for data to be gathered and disseminated. So when we look at something like disease rates within a given country there is oftten little consensus in the data between organizations such as the WHO, NGOs, and the local government.
I have heard many times that the people on the ground don’t have the capacity to understand the information, or that they don’t find it helpful. This isn’t necesarrily the truth. A major problem is in how we present the information. In order to figure out how to allocate resources and improve the health of their communities clinics and healthcare workers need data presented in a format that they can understand—they need “pretty graphs.” They need a story so they can then ask meaningful questions. Questions like “Why is that happening?” or “How do we fix this?” can lead to a dialogue that changes the way they make decisions.
There are barriers as to why NGOs don’t use data to drive decision-making. First off, it takes a specific skill set to be able to turn raw data and build it into a “pretty graph.” Sure excel can do a lot but there is a serious and difficult barrier to overcome in having a workforce with the technical capabilities to handle large amounts of data and interpret them in a meaningful way.
Secondly, funders often require data to be collected for monitoring and evaluation purposes (which is a step in the right direction don’t get me wrong), but their interests often take priority over providing feedback to those on the ground. The indicators that are turned into “pretty graphs” might not be the same indicators that are truely important to a healthcare worker in the field.
There are also ethical dilemmas in collecting data, mainly deciding if the cost outweighs the benefit. Collecting and analyzing data can be extremely expensive and it may not be cost effective or an efficient use of resources.
This is where something like mobile platforms can make a huge impact.
What mobile platforms need to make this happen
Mobile platforms (such as Dimagi’s CommCare) provide real time data collected directly at the source. The next step that needs to happen is the data needs to be presented to healthcare workers to help drive their decision-making.
The old way of data collection goes something like this: person on the ground collects the data and then they never see it again until it comes back as an aggregate number that doesn’t really tell them how to solve the problem. The story of community ends up lost in the “fuzzy data” world. Even if they are using a platform like commcare in order to manage their patients they need the feedback of what is going on in real time. They need “pretty graphs.” A visual representation of how they can do a better job and what the data is telling them in order to make decisions.
If a picture is worth a thousand words, then a graph tells you what those words are. Whether it’s a bar, pie, line, bubble, or map, graphs can articulate a community’s story in a way that is both useful and interesting. We need to do a better job of starting that conversation.
Imagine being able to show CHWs a pie chart of the percentage of patients they missed following up, or a graph of how they have helped decrease their community’s infant mortality rate. Problems can be fixed within hours rather than years. Information can be gathered and processed quickly and issues can be resolved before they spiral out of control. Mobile health platforms using “pretty graphs” has the potential to make huge strides in healthcare management.
How do we put this in the palm of their hand—literally.
Maybe I am a bit biased because I am a self-described “data junkie.” Maybe my nerdy (or geeky?) love of graphs has skewed my perception on whether or not this is necessary. Or maybe, just maybe, there is a story to be told, and a way to improve the way that we manage healthcare and drive decision-making that involves “pretty graphs.”