Complexity: Can health data and systems be black-boxed in service models?

Given the high level of complexity of contemporary ICT and information systems, the different service models described here are seen as a way for businesses to outsource the handling of complexity to service providers. Development of the cloud and big data may be seen as a transition of increasing sophistication through the various service models: Infrastucture, Platform, Software, and Analytics.

Based on the inherent openness and fuzziness of public health data and systems, some of their various components cannot go through such a transition of commodification and black-boxing. Efforts to achieve such commodification are important parts of the currently ongoing ‘big data’ and outsourcing efforts. An important question will be how such efforts to streamline everything as

‘easy to go’ boxes may eventually backfire and trigger the reopening of the black boxes and debunk beliefs of steady technological progress. This may then cause setbacks and the ‘opening of the black boxes, or calling the bluff, if the plan has been to be the technology optimist. Such examples abound. Reverse salient is an old military research term, used by Thomas Hughes (1983) on setbacks in infrastructure development when it has been too fast or too poorly grounded. In military terms, reverse salient is when one part of the frontline becomes weak and the enemy threatens to break through, and the whole frontline has to step back and consolidate.

The example of the DHIS 2 implementation in a big African country where the metadata structure had been developed ad hoc, is a good illustration of a reverse salient. While the system was in the process of being rolled out after a pilot period, performance of the system deteriorated as more provinces and users were added. Poor metadata design caused the server to crash when many users logged in. This revelation came only one and a half years into the process, representing a breakdown and the opening of the black box in the Thomas Hughes sense. Poorly designed systems will ultimately break down. The health information exchanges and use areas supported and encompassed by the HIS in a typical country-level HMIS implementation are highly context-sensitive and complex, with many linkages to different organizational structures and ‘moving’ parts. We have seen that the DHIS 2 platform is gaining tremendous popularity, but this in itself can push it into getting black-boxed. Huge potential of breakdowns will be created if designs are not sensitive to the health system complexities and are not allowed the time and space to adapt, evolve, and embed themselves in new contexts.

On another note, the history of DHIS versions 1 and 2 until now may be analysed in the light of reverse salient, where performance of certain components of the dominant—and always very techno-optimist—paradigm fails. Until now, DHIS 2 has had the role of the tortoise in Aesop’s fable about the tortoise and the hare, while larger technology projects have taken the role of the hare. DHIS has always been slowly walking forward regardless of the current fashion. In this new situation of increasing international donor acceptance of DHIS 2 and the rapid expansion and increased complexity in each country, DHIS 2 may risk taking over the role of the hare!

Concluding remarks on black-boxing services and big data

In the discussion in section 7.4.4 we have concluded that HIS and big health data cannot be commodified, black-boxed, and outsourced to service providers. This highlights the importance of the veracity dimension of big data.

However, the possible messiness and poor quality of health data are consequences of this complexity. Complexity is the root cause for why health data and HIS cannot be black-boxed, and need to be included in our understanding of big data. International hosting will typically provide a far better technocommercial option than local hosting. Governments, however, want to store their data within the country’s borders, not in the unspecified cloud. For the development of HIS, it is important to aim at leveraging all the advantages of cloud-based computing, while ensuring that the risks are well mitigated and managed. This is a non-trivial task given the institutional work required.

< Prev   CONTENTS   Source   Next >