Page 120 - ITU KALEIDOSCOPE, ATLANTA 2019

2019 ITU Kaleidoscope Academic Conference

prevention and research side, the stack would also receive Through data collection from a variety of sources, the
data from R&D repositories on health and nutrition of India knowledge management platform will be a treasure trove of
like the Indian Council of Medical Research (ICMR); big data. With the help of artificial intelligence tools, big data
national research institutions like the National Institute of analytics as well as information systems like GIS, this data
Nutrition (NIN); National Centre for Disease Control can be analyzed to extract important insights needed for
(NCDC) implementing the disease surveillance programs of answering relevant questions like reasons for a disease
the government; radio and diagnostic systems, laboratories; outbreak, areas involved, etc. as well as mapping and
and various government departments that have linkages with predicting outbreaks, triggering response mechanisms and
human health including academia. To enable forecasting for taking preventive action. However, there may be a need to
prevention of diseases, effective linkages established democratize this data in a way so as to make it available for
between food control agencies and the public health systems use of machine learning (ML) and AI.
including epidemiologists and microbiologists can provide
information on food-borne diseases, which may be linked to Decision support algorithms employing quantitative data
food monitoring data and lead to appropriate risk-based superimposed on qualitative understanding of local contexts
policies. This information includes annual incidence trends, would help to undertake risk assessments of public health
identification of susceptible population groups, domains. Predictive modeling would help to improve
identification of hazards, identification and tracking of estimates and thereby allow quantification of health risks and
causes of diseases and the development of early warning also find applications for assessing prevention strategies in
systems for disease outbreaks. Therefore, IDSP, HMIS, risk management. The processed data from the stack can be
AINPPR and similar other data emerging from various made available to various stakeholders through open
sources will have to be collected and analyzed centrally by application programming interfaces (APIs).
the knowledge management system (KMS). Once this data
mapping and feeding mechanism is strategized, implemented STEP 4: Use of NHS information for evidence-based
and executed over an extended period of time, the NHS shall decision making, forecasting, planning and research by
act as a centralized health record repository for all citizens. different stakeholders

Once the sources of data are identified by mapping the The insights generated based on the analysis of data can
patient journey, the next step would be to focus on the data provide not only straightforward information that is useful to
formats/databases, and then connecting them. The need for the health functionaries directly but also enable cross-
uniform standards to make multiple EMR systems functional collaboration between various stakeholders
compatible and the information interoperable is paramount (Figure 1, Block 3).
as it will tie up isolated pools of data. A consortium can be
setup consisting of representatives from various consenting For example, information on the immunization status in a
data-sharing stakeholders to identify and list the various particular area can help the health officer to plan resource
current formats being used, come up with short-term allocation of both staff and material for those areas that are
interoperability solutions and envisage long-term data lagging in immunization coverage. On the other hand, cases
sharing standards on common agreed formats. Effective of nicotine toxicity in tobacco harvesters or cases of silicosis
change management would play a pivotal role in aiding the from mining may require collaboration with research
stakeholders to adopt the new agreed formats to process and institutions that can provide technological solutions like
share the data being collected at their end. The costs involved suitable nylon gloves for tobacco farmers or well-designed
in the change can be managed in a way that is offset by the masks for the miners.
overall commercial gains incurred due to the implementation
of the NHS. In terms of channel usage, high speed STEP 5: Social learning: awareness, sensitization and
communication technology is proposed to facilitate data training
collection, analysis and reduce reaction time as well as
enable effective sharing. This digitized data will then be The implementation of a project with an all-encompassing
stored in a central place like a cloud. It will be accessed vision would be meaningful only if stakeholders’ capabilities
remotely by all stakeholders. Also, standards of data security are augmented at all levels ranging from the top till the
need to be strengthened with the use of blockchain ‘bottom of the pyramid’. Political leaders and policy makers
technology so as to protect this data from cyber threats. at the highest level must be encouraged to stay aligned to the
successful culmination of the ‘Health for All’ goal.
This is the most critical step towards building the KMS as it Awareness is equally critical amongst patients whose public
strives to bring together “stove piped” data and needs health data and the related socioeconomic indicators are the
substantial investment of resources not only in terms of funds mainstay of the system. In addition, health data may also be
but also manpower. Here, buy-in from the decision-making crowdsourced from citizens, therefore the citizens need to be
authorities is important as it will drive the project. sensitized about the ‘principle of consent’ with regard to
their health data and personal health records (PHR). Equally
STEP 3: Applying data analytics relevant is capacity building drive for every constituent. As
an example, the capabilities of the grass-root level public
health worker, who is expected to input the information at

– 100 –