India recently reported outbreaks of diarrhoea in Indore and typhoid in Gandhinagar, among other instances. India’s Integrated Disease Surveillance Programme provides a weekly report of the outbreaks across India. As we have learned from Indore, reports of poor water quality were already out, but did not result in action. This is a failure of “intelligence”.We typically associate this word with security agencies, whose primary mandate is to gather intelligence and act on it to protect the nation. Using multiple sources of intelligence like human, electronic and advanced technologies, they take preemptive or reactive actions. Any explosion or terror attack is attributed to an intelligence failure. One of the primary reasons for such security failures is our inability to integrate intelligence from multiple sources due to a lack of sharing and unified analysis.AdvertisementLike security, there is a paradigm shift in the way we look at “public health intelligence”. Public health intelligence describes the process of monitoring health threats by gathering and monitoring information on events of public health importance with the specific aim of early detection for effective response.In public health, the relevant signals that we look for are disease outbreaks or clusters; potential pandemics, including from the animal world; outliers in performance of the health system, including coverage of health programmes; iatrogenic adverse events; antibiotic sensitivities; extreme weather warnings; and many more. Data needed to identify these signals would cover accidents, deaths and admissions with their causes, climate parameters, laboratory test results, national programme outputs, adverse events, pathogens in the environment, and so on. The data could come from social media, the internet of things, government portals, clinical records from public and private health facilities, non-health agencies, wastewater surveillance, etc. The types of data are also complex, like images, numbers and words, and need complex analytical approaches.Our current disease surveillance suffers from the problems of poor gathering of data and its conversion into meaningful intelligence. It continues to be fragmented, siloed, and limited and is mainly about reporting outbreaks. We continue to have a health management information system (HMIS) which is still not fully digitalised. Access to information in the HMIS is governed by privileges defined by vertical national programmes. Our surveillance system rarely leads to any preemptive action or a long-term course correction.AdvertisementThe Covid-19 pandemic taught us that we need to drastically up the game in health security. During the pandemic, we used multiple technologies to assist in the collection of data. These included web-based epidemic intelligence tools, machine learning and natural language processing, targeted public health messaging, social media and online searches, syndromic surveillance, wearable sensors, digital diagnostics and genomics, data visualisation tools, interactive geospatial maps, drones, computer vision, contact tracing apps, chatbots, smartphone apps, symptom-reporting apps, telemedicine, mobility pattern analysis, and security and privacy preserving technologies. These have come to stay now as standard processes.Has India reimagined its health security for the present times? The document, “Vision 2035: Public Health Surveillance in India,” a white paper prepared by NITI Aayog and the University of Manitoba in 2020, re-imagines public health surveillance to be a predictive, responsive, integrated, and tiered system of disease and health surveillance that includes prioritised, emerging, and re-emerging health conditions. It expects surveillance to be primarily based on de-identified (anonymised) individual-level patient information that emanates from health care facilities, laboratories, and other sources supported by an adequately resourced, effective administrative and technical structure. For a vision document, it is neither sufficiently visionary nor imaginative. It fails to recognise the rapid and massive changes happening globally and nationally in the way information is collected, processed and analysed. An exclusive focus on Electronic Health Records is inappropriate.Modern disease surveillance requires a complex architecture of an adaptable system with non-linear data flow, chaotic dynamics, emergent behaviour, and multi-scalar nested systems.The next-generation surveillance system will need to include the multiple data sources listed above. This requires data fusion, storage, data wrangling, data analysis and knowledge translation with a real-time dashboard, interactive data source, and analytic methods on a real-time basis. We need to make these inter-relational datasets amenable to collective synthesis.This can only be achieved by the use of artificial intelligence to augment traditional epidemiology in helping public health sort through large amounts of data. Machine learning can enhance predictive modelling, improve cluster analysis and social network analysis to increase the sensitivity of surveillance systems. Natural language processing can help extract meaning from unstructured data contained in clinical notes, social media and news.The key challenge in any intelligence system will always be the “human in the loop”. While we should invest in developing AI solutions to support public health surveillance, humans cannot be replaced. Our current approach of training Epidemic Intelligence Officers, while appropriate for the traditional approach, is inadequate for the future. We need to get more technology-trained people into the system.Given the surveillance needs, architecture, and oversight, I suggest setting up an autonomous, well-resourced national surveillance authority.Like a good security system goes beyond the police force, which neither has the time nor the capacity to handle security issues, we need to look beyond the routine health system if we want a good quality public health intelligence system. While the health system can continue to be its eyes (one among others) and hands, we can no longer afford for it to be its brain too!The writer is a Professor of Community Medicine at the All-India Institute of Medical Sciences, New Delhi. All views expressed are personal