Data Integration Challenges in Healthcare (Part 2)
In Part 1, I discussed how (1) Digitization and (2) Data Accuracy are the foundational pillars to solving healthcare’s biggest data integration challenges.
Teams on the ground need to feel involved in the process, from ideation, to vendor selection, and implementation.?
Data Accuracy is then possible when you have the right software vendors that your teams both like and use.??
The question is becomes;?
How do you ensure that the data capture process is being followed?
That brings me to the third pillar, Incentives
Incentives
The late, great Charlie Munger once said:?
“Show me the incentive and I will show you the outcome”
Incentives are things that you use to get someone to do what you need them to do. They can be rewards, punishments, or both.?
Usually the right incentive structures have both the carrot and the stick inbuilt.?
Incentive building is its own discipline so I won’t do it any justice here, however two concepts that I refer back to frequently when designing incentives are economics terms. Revealed and Stated preferences.?
A Revealed Preference was first introduced by an economist in the 1930s. Simply put, it is deriving someone’s true preferences by monitoring what people do (Reveal) instead of what they say.?
A Stated Preference is a market research term developed in the 1970, whereby people would be asked directly about their preferences.?
The key to building a fit-for-purpose incentive structure is in both asking people what they think, then monitoring how they behave.?
When what people say and how they behave is consistent, you’re closer to the “truth”.
In a data collection context, you have to understand why people aren’t using the tools that you provide, you then make the correction in the tool / process, and see if the problem persists.?
If it does, then you introduce some rewards for the behaviors that you want to see, and punishments for those you want to discourage.?
Measure again and alternate between changing the tools, incentives, and processes until you reach a compliance level that you’re happy with.?
The final pillar is to adopt the right interoperability standards.?
Interoperability
If you are integrating health data, especially in Africa, you will find almost every possible format, architecture, and type of data being collected.
That makes the art of integrating this data incredibly difficult.?
Thankfully the Fast Healthcare Interoperability Resources (FHIR) was developed by HL7 to simplify the process of data exchange by providing a consistent and easy-to-implement framework.?
FHIR allows for healthcare software to “speak with each other” through exchanging data.?
At Kapsule we learned that when you have loads of different software systems, it is most efficient to get them all to talk to one central system, rather than talk to each other.?
In tech we call this central database a “Data Lake”.?
Once you have everyone feeding into the Data Lake, you only need to do analytics on that Data Lake to understand what is happening everywhere else?
(Thankfully HL7 made this tasks infinitely easier through creating the Clinical Document Architecture)?
When you have a very fragmented data market and you have different companies, universities and organizations all with their own software solutions, like is the case in most of Africa, then you begin to appreciate how incredible these standards are.?
If you’ve:
Then controlled for 80% of the digital integration problems that will arise. You’ll also be better placed to handle the 20% of challenges that still arise.