Saturday, 3 August 2013

THE IT SOCIETY: In databases we trust!

Today I wish to explore the extent to which our IT-enabled society could go in making a seemingly straightforward situation more complicated and how we can make our technological expertise more grounded in reality.
Our modern society is what one would call a complex system rather than a linear system. In linear systems, outcomes between different variables are predictable with a clockwork regularity that can be described neatly with mathematical and physical principles. Examples include classical mechanics such as a swinging pendulum. Complex systems on the other hand comprise of thousands if not millions of variables which interact in a non-straightforward manners with such unpredictable outcomes that simplistic mathematical theorems that you learned in class simply fall apart. Unless you develop new ways in describing these non-linear systems (like chaos theory which I will not bother talking about here), you haven't a hope in deducing outcomes, let alone understanding the system in the first place. Classical examples include the weather. Forecasting the next day’s weather only became easier with the advent of sophisticated computers and, more importantly for today’s topic, better meteorological databases.

For many scientific disciplines, you need lots of up-to-date data. Research and survey projects are a dime a dozen nowadays but such data is useless if you don't have an efficient way to store, clean (removing input errors) and retrieve it easily. We usually rely on IT to do most of these things nowadays and we are so used to hearing from manufacturers about the benefits of these machines that we might be forgiven to believe that there aren't any catches at all with regard to IT.

Two months ago on the 15th of June I gave a review talk at my university on the subject of health information technology or HIT. I described its usefulness and the need for its implementation in the African health arena. My talk focused heavily on the history of one particular attempt in my home country of Tanzania. Many countries require all health institutions to record outpatient and inpatient details to allow abnormalities in outcomes to be tracked on a long-term basis. Typical systems employed include a centralised health database system, most of which are digital in nature. In Tanzania the database (referred to as the Health Management Information System or HMIS or MTUHA in the local language) was mostly paper-based, and still is at the district, level for a good number of years since its inception in 1993. Digitisation of the HMIS records only began at the turn of the new millennium and instead of improving data quality and efficiency in collection, IT augmentation has seemingly lowered the already poor quality of the data and has even made it more difficult for the users i.e. the health professionals to use. Researchers complain that far from being a source of life-changing knowledge, harvesting HMIS data is like looking for crops in a field of weeds. Useless!

What’s most disturbing however is the lacklustre enthusiasm that came out of the HMIS managing team based in the country’s ministry of health in the form of the following quote in a 2002 report:
“Contract out the review of the current HMIS software in the light of alternative packages available, with a view to recommending the best option for the national [health] system” (MoH, 2002:16).
In less than 2 years after digitisation was complete at the upper levels of governance, changing the system’s software was being contemplated by the team overseeing HMIS operations. More than a million U.S. dollars were used to build a broken system!

So what went wrong? The explanation I offered in the talk was that the stakeholders involved in the design and execution of the new system failed to come together in the open. Apparently the donors funding the program decided that it was better to give money directly to software vendors rather than risk losing it to the depths of government bureaucracy. But I insist that the more important reason is that these stakeholders failed to design the system to serve the health system effectively. IT is useless if the customer does not know how to use it, hence Samsung’s catchy phrase ‘designed for humans’ used in advertisements of the Galaxy series of smartphones. Although steps are now being taken to ensure that the next generation of Tanzanian health professionals are IT literate by introducing definite IT studies in the university curriculum, I think it would be better if the whole country were to prioritise IT in its national development goals just like in its neighbour Kenya which has recently seen a surge in investments from big names like IBM, Google and Hewlett Packard.

What we see here is a case of fundamental system design failure. The stakeholders forgot about the complexities of the African health system, hence the seemingly inexplicable nature of the digital HMIS’ outcome. Health workers are either too busy or too ignorant to work the gadgets. The solution is elegant but not necessarily simple; build a strong culture of IT, make people more aware of it then design the system to fit the people and not the other way around. This bottom-up approach will help us prevent building loaded but useless systems that will inhibit rather than allow us to harvest life-saving health knowledge.

You can also read my original paper here where you will also find a list of good references.