Big data revolution in healthcare sector

an article that appeared in the Mint on using Big data for healthcare 

What if an app on your phone could predict a malaria outbreak in your neighbourhood or your treadmill could diagnose your worsening cardiovascular health? 

If you think these ideas are too far-fetched, watch closely the big data revolution underway in India’s healthcare sector. The aim is to use the large, unstructured sets of data to discern patterns that can lead to the delivery of personalized treatment. 

 

KGB or Kooda, Gandagi, Badboo (garbage, dirt, bad smell) is an app developed by a team of researchers at the affordable healthcare division in the Public Health Foundation of India (PHFI). The app uses social networking platforms to address issues of sanitation and hygiene in urban areas. 

 

“We were just having fun with technology. One can click a picture of garbage or stagnant water, and post on any of the social networking platforms,” said Kanav Kahol, who is an expert in the use of information, mobile and sensor technology for public health. 

 

“There is nothing new about the technology. What is new is how it is now being put to use. People have been posting pictures on social networking sites. We are just channelling it to the right people to address issues of health and sanitation. The pictures on KGB will come to our central database and we forward it to relevant authorities. If there are more pictures from a single neighbourhood, we can tell what kind of vector-borne diseases can be expected in the coming weeks,” he said. This process is known as participatory epidemiology.

 

“There is so much to be done in this space. Gymnasiums have machines that already record speed and time of people working out on machines. This data could be used to develop a pattern of how much a person can exercise and any deviations can be used to predict worsening health conditions. Consistently deteriorating time period on a treadmill could be a warning of a heart check-up. We already have all the information. We just have to start looking at things in a new way,” said Kahol. 

 

At the Institute of Genomics and Integrative Biology, part of the Council of Scientific and Industrial Research (CSIR), Anurag Agrawal, a doctor and researcher, is even more enthused by the possibility of collecting copious amounts of data.

 

Agrawal is coordinating with hospitals as well as experimenting with standalone clinics as part of the CSIR project that Mint reported on 27 January. A clinic at Lakhimpur Kheri in Uttar Pradesh is one among several such centres in the country, the others being in Haryana and Hyderabad, which are equipped with equipment that analyses blood samples and heartbeat rhythms. The data is streamed into centralized servers in Delhi. Such data is gold, Agrawal said. “More than public health, it is about creating massive data sets that will give us a sense of what’s going to happen in terms of recognizing diseases in their early stages,” said Agrawal in the 27 January interview. 

 

Analysis of the data can throw light on hitherto unexplained phenomena, he said.

 

Drawing on his own work, Agrawal offers the example of impulse oscillometry (breath measurement) readings.

 

The impulse oscillometry reading of an asthmatic looks no different from that of a normal person between attacks. Yet, biopsies of an asthmatic’s lung indubitably reveal that there’s something distinctly different. “We asked ourselves if we can spot these differences beforehand,” said Agrawal. For that, Agrawal used computer algorithms to parse the oscillometry readings and found that at the level of mathematical functions, there were distinct signs of sickness among asthmatics even when undiagnosed for the illness.

Agrawal is now modifying these findings so as to be able to detect early signs of chronic obstructive pulmonary disease, globally ranked among the top three causes of death in smokers.

 

In a research paper on the electronic health centre concept published in the June edition of the journal PlOS Medicine, Agrawal along with other authors wrote that fungal skin infections, scabies, and allergic rashes accounted for a large number of patient visits on the basis of data from 3,677 such users.

 

Muscle or joint pain were reported at early ages, possibly, according to the authors, due to the demands of a strenuous agriculture-based lifestyle. Antacids, analgesics/antipyretics, and anti-histamines comprised the bulk of the medicines prescribed and relatively few antibiotic prescriptions were recorded. 

 

“That’s a good sign given a tendency to over-use antibiotics in India,” said Agrawal. “But on the other hand there were disproportionately fewer women between 6 to 18 years (relative to their proportion in that region’s population) who were visiting the clinics. Now that’s something we’d want to look deeply into.”

Others recognize the power of big data to make an impact in the area of health care.

 

Ashutosh Pande, an engineer who’s spent most of his professional life designing GPS applications entirely unconnected to the world of medicine, quit his job in 2010 to launch his own company, Arogya Mobile Health Pvt. Ltd. Arogya has just begun preliminary trials in Uttarakhand to collect and scan assorted health parameters in the hope that they will throw up patterns. As part of his plan, which essentially aims to marry the ubiquity of mobile phones with burgeoning concerns over lifestyle diseases, primary school-educated health workers will collect weight, temperature, blood pressure and electrocardiogram readings from roughly 50,000 villagers in 50 villages. The Bluetooth-enabled devices that they use will send the information to the cloud and a pre-programmed algorithm will instantly determine whether someone needs to go see a doctor. In return for a monthly fee of less than Rs.100 a month, Pande hopes to access a trove of data that can be parsed through analytics and then be used for predicting early onset heart disease, diabetes and cardiovascular disease. “These were once supposedly the diseases of the affluent but new data shows that such afflictions are increasingly found in rural areas,” he added. 

 

Although India’s prowess in information technology and the myriad health afflictions of its people should have made it a natural hotspot for big data applications in health, “progress has been painfully slow”, Agrawal said.

Getting even basic data was exceedingly difficult and few companies and institutions were investing enough to improve access, he said. While sectors such as banking, insurance and retail have already benefited from the big data revolution, the health sector is yet to use the unstructured data generated by hospitals and other institutions to generate predictive health interventions. 

 

Healthcare in India has always lagged behind in adopting radical ideas such as big data due to “risk of lives”, said Kapil Khandelwal, who runs a start up, EquNev Capital Pvt. Ltd, which specializes in the use of big data, referring to any radical change in standard procedure that could have an impact on mortality.

“The health sector will be the last to adopt big data, after it has been proven in other regulated industries such as consumer products, banks and financial services. We are seeing adoption in health financing, clinical trials and drug discovery, pharma and wellness segments of healthcare in India but hospitals are yet to use it to define tailor-made care for patients,” he said.