Tracking India’s Deadly Flu Outbreak in Real Time

Early one morning in late August, before the sun has inched over the horizon, thousands of people have already taken to the streets of the ancient city of Nashik in India. They walk alone or with families during a reprieve from the heavy rain. Some are barefoot. Women, clad in colorful saris, carry bags in their hands or on their heads. Young children grasp the hands of their grandparents. Everyone is heading in the same direction. They walk past shops and houses. Police dot the streets, some wearing surgical masks as a precaution against the latest H1N1 influenza outbreak.

34
Police wear surgical masks as a precaution against H1N1 influenza during the Kumbh Mela.

The streets narrow and snake downward. A man tells his aging and frail mother, who is walking with the aid of a stick, “we’re almost there.” Her eyes light up and a smile creeps onto her lips as she sees hundreds of people bathing in the Godavari River, its green-brown water gently flowing through the middle of Nashik and around the ghats, the wide steps that form the river’s banks.

The old woman and her son join the bathing crowd. They dip into the waters once, twice, thrice, and then take water in their cupped hands and let it trickle back into the river as an offering to the gods and their ancestors. This was the first royal bath day of the Kumbh Mela, a Hindu festival held this summer in Nashik. Millions of Hindus attend the Kumbh Mela, a mass pilgrimage considered to be the largest gathering of people at one place in the world. Crowds swell to a peak on royal bath days, held on auspicious dates when the power and magnetism of the holy waters are believed to be amplified.

Nearby, in the same waters, people are washing their clothes and brushing their teeth. Still others are filling water in plastic cans to carry home. The river is more than a sacred destination—it is a source of drinking water, a laundry, and more.

35
A pilgrim brushes his teeth while others bathe around him. Despite the vigilance of thousands of staff, several pilgrims still perform daily chores such as brushing their teeth and washing their clothes in the river.

The crowd of pilgrims descending on Nashik this year coincides with an outbreak of particularly virulent H1N1 influenza. So far Nashik has recorded 487 cases of H1N1, with 82 deaths—the highest in the district in the last five years. Across the country, the virus has affected thousands more, mutating and becoming even more dangerous than ever before.

Dozens of pop-up clinics tend to the sick and injured during the Kumbh Mela. They are scattered along the ghats and walkways, at the bus stands and at the railway station, and near the pilgrims’ tents. The white tarps stretched over poles are just big enough to accommodate two tables, a couple of chairs, a bed, a cupboard, and cartons of medicines. Out front, red boards marked “dispensary” hang from the tents. Beside one clinic, water is leaking from a plastic tank and pooling on the floor, an ideal breeding ground for mosquitos.

51
Thousands of pilgrims bathe in the Godavari River on an auspicious royal bathing day.

Despite the makeshift appearance of the clinics, many are on the cutting edge of disease surveillance. For the first time ever, the doctors who man these stations will have at their fingertips a digital tablet loaded with a sophisticated software that can track disease and quickly spot the early signs of a potential epidemic.

The introduction of the software, called EMcounter, is part of a larger joint effort of Harvard University, the state government of Maharashtra, UNICEF, the United Nations’ Children Fund, and other partners—and for many in the group, it’s the latest iteration in a collaboration that dates back nearly a decade. Now called the JanaSwasthya Project, which means “public health” in Hindi, the initiative aims to screen Kumbh Mela pilgrims for a range of health diseases and conditions, with EMcounter as its cornerstone. “We believe that the potential for mobile health to revolutionize healthcare access and delivery in India is enormous and untapped,” the project organizers write on their website.

It’s the first time such a technology has been used to track diseases in real time at the Kumbh Mela or any other mass gathering. In a country with vast discrepancies in access to health care, where reporting of diseases can take weeks, the project’s backers believe EMcounter has the potential to improve the speed of health reporting and, ultimately, save lives.

Documenting the Spread

Just before taking a dip in the Godavari, 58-year-old Gauri Devi Solanki enters a pop-up clinic. She is wearing a long red skirt and matching top, and her head is covered in a patterned scarf. Solanki’s eyes are watery and her nose runny. Kalyani Bunage, a 28-year-old doctor, greets her under a whizzing fan that hangs from the tented roof.

“What happened?” she asks. Solanki has traveled over 600 miles by bus with her husband and sister from Rajasthan, a state in Western India. The trio reached Nashik about five days ago. “I have had a cold, cough, and headache for the past two or three days,” says Solanki, who also has diabetes. Solanki says that she purchased some antibiotics over-the-counter when she arrived in Nashik, but they did not alleviate her symptoms. Dr. Bunage asks if she feels chills in the morning and night and a burning sensation while urinating. Solanki nods yes.

8
An old woman consults a doctor from a religious group at a clinic set up for the festival. At many of the government clinics, these doctors worked alongside government-appointed doctors.

Dr. Bunage taps on a new black tablet computer, recording Solanki’s age and gender in an electronic form that is part of EMcounter. She scrolls through a list of 40-odd presumptive diagnoses, checking the boxes next to cough and cold, headache, and urinary tract infection, and then taps “submit.”

It’s a quick and unremarkable process, but it represents a dramatic departure from the norm here in India, where doctors record illegible, incomplete, and sometimes incorrect entries on paper forms.

37
Doctors at one of the clinics study the EMcounter application on a tablet.

Solanki’s basic information and symptoms are whisked away via Wi-Fi and cellular networks to servers in the cloud. Over the course of the festival, which ran from July through September, doctors wielding EMcounter tablets capture similar information for about 35,000 other patients who visited clinics. A remote team of Harvard researchers and government officials monitors the data as it streams in, using the EMcounter tools on their end to analyze the data for trends like which disease is most prevalent and where. In the days surrounding Nashik’s first royal bathing day, most of the patients who visited these clinics had a cough and cold, fever, or joint pain. But about 23 people had a fever as well as an upper respiratory infection—telltale symptoms of H1N1.

Early Warning

EMcounter is the brainchild of Satchit Balsari, a 37-year-old faculty member at Harvard University’s FXB Center for Health and Human Rights who also leads the global emergency medicine program at Weill Cornell Medical College in New York. He grew up in Mumbai and attended medical school there, and he came to the U.S. to attend a Master’s program in public health at Harvard. On his first day of class, the attack on the World Trade Center on September 11 would forever change how Dr. Balsari viewed his career.

“That in some sense shaped my interest in disaster and humanitarian medicine,” Dr. Balsari tells me in the lobby of his hotel in Nashik. He’s wearing a blue kurta—a loose, collarless, knee-length shirt—and grey cargo pants that hang neatly on his lean frame. After completing his Master’s, Dr. Balsari enrolled in New York-Presbyterian Hospital’s emergency medicine program in 2004. He then returned to India briefly, eager to understand how the country’s fledgling emergency medicine specialty was developing. His experiences in the country’s casualty wards sparked the idea that would become EMcounter.

Around the same time, India launched its Integrated Disease Surveillance Program with the help of the World Bank. Under the program, the paper records of most village-level health visits are gathered district-wide and transcribed on computers a week later. It’s a disjointed system that can delay the response to infectious disease outbreaks. A recent analysis showed that, although the reporting of the number of outbreaks in India has increased three-fold from 2008 to 2013, there is “an urgent need of improving the quality of reporting and investigations.” Communicable diseases account for nearly half of India’s disease burden.

Dr. Balsari collaborated with a few friends to develop a simple computer program that would allow them to transfer paper records from the ER wards to computers on a daily basis. They called the interface EMcounter—”EM” for emergency medicine, and the rest because “it rhymed with encounter that doctors have with their patients,” Dr. Balsari says. In 2006, they piloted the EMcounter in the Sundaram Medical Foundation in Chennai. Since then, the team has been revising and improving the tool, eventually moving it to a tablet interface that was first tested in 2012 at Akobo County Hospital in South Sudan.

4
Dr. Satchit Balsari conducts the training session for doctors who will be using the EMcouter tablets during the Kumbh festival.

The same year, Harvard University’s South Asia Institute began a multi-disciplinary project, “Mapping the Kumbh Mela,” and reached out to experts in the U.S. and India to study various facets of the festival, including public health. Dr. Balsari immediately saw an opportunity to digitize disease surveillance using EMcounter.

The first Kumbh Mela Dr. Balsari’s team studied was in 2013 in the city of Allahabad in Northern India. There, at the end of each day, the group of young doctors and medical students from the U.S. and India used EMcounter to transcribe the patients’ health and demographic data from the paper charts maintained by the doctors. “That gave us interesting data, but it was not exactly real-time,” Dr. Balsari says. He approached the government of Maharashtra to pilot real-time disease surveillance in this year’s Kumbh Mela in Nashik. UNICEF also expressed interest in the project because of the large number of women and children attending as pilgrims. The project, which took the name JanaSwasthya, piloted EMcounter for three days around each of the three royal baths—one in August and two in September. Beginning in July, Dr. Balsari’s team began to train Indian doctors from Nashik and neighboring villages to use EMcounter.

“We liked the idea, especially because we were concerned about the H1N1 outbreak. EMcounter promised to catch early warning signals,” says Archana Patil, a senior official of health services for the government of Maharashtra.

Incomplete Records

North of the city of Nashik, inside a gated area in Ware village, 36-year-old physician Dilip Janardan Bhojane examines patients in the front room of his modest two-room house. Outside, barefoot kids walk on half-paved streets lined with lush trees. Makeshift tea stalls dot a village of paddy fields, huts, concrete houses, and open drains.

Dr. Bhojane’s room is actually a temporary outpatient wing of the Ware Primary Health Centre, which stands dilapidated just a few yards away. The faded yellow brick walls are peeling in several places, and part of the roof fell down a year ago.

Until a new building is constructed, Dr. Bhojane’s front room is now a patient’s first point of contact with the health center, known as the outpatient department, or OPD. The other, smaller room is where Dr. Bhojane sleeps. Most of his clothes hang on the walls, and a suitcase and the rest of his belongings are stacked in the room’s open closet. Water is seeping through the wall next to a small mirror. As is the case in villages across India, the government has provided this house to Dr. Bhojane to ensure that he stays close to the center. Dr. Bhojane’s wife and four-year-old daughter live in a different district because there are no good schools or parks in Ware. Dr. Bhojane sees 40 to 50 patients a day in this office, which serves a community of 24,000 people in all.

27
A doctor checks a sick patient's blood pressure during in one of the pop-up clinics.

As he sits behind a wooden desk waiting for the next patient, I ask Dr. Bhojane to show me the center’s patient records of the last week. The records include the prescription papers on which Dr. Bhojane scribbles the name and age of each patient along with the medicines they need. But he does not write down their symptoms and diagnosis. Similarly, he leaves the disease column in the OPD ledger largely blank.

Using these records, the health center’s pharmacist completes the presumptive diagnosis form, or P-form, which contains a list of about 20 possible diseases and ailments. The P-form is one of the key documents used to track public health in India. Every Saturday, pharmacists record how many patients were diagnosed with a specific disease in the past week. The form is then hand-delivered to the office of a data entry specialist, who types this information into the online portal of the government’s Integrated Disease Surveillance Program.

Looking at the incomplete and illegible prescription paper and OPD ledger, I ask Dr. Bhojane how the pharmacist will complete the P-form. He nods. “Hmmm.”

All the P-forms I checked had “zero” written under the columns of various presumptive diagnoses, even when more than 250 patients visited the OPD during each of those weeks. Without more information, there is no way to know what medical problem these patients had. If there had been a disease outbreak, there is no record of it.

It takes about eight days for the patient data to make its way from the Ware Primary Health Centre to district government officials. In the midst of a fast-moving outbreak, a week is a long time.

The Ware Primary Health Centre is one of four primary health centers I visited in the villages around Nashik. Invariably, each had discrepancies in reporting. The OPD registers, prescription papers, and P-forms were either incomplete or illegible.

46
Patients wait to see a doctor at one of the busiest clinics in Ram Kund, a neighborhood of Nashik filled with temples.

Yogita Sudhakar, a pharmacist at a nearby primary health center where about 2,000 patients visit in a month, explains that filling out so many forms can be burdensome. “Most of my time goes in just giving out medicines. There is no time to maintain proper daily records,” she says. As a result, a stack of 400 to 500 prescription papers stares at her at the end of each week when she sits down to fill the P-form. “I rarely have time to go through each paper,” she says, admitting that she often makes a guess based on her experience when completing the form. At other health centers, pharmacists were less careful about maintaining accurate records.

Dhananjay Pansare, a data entry officer, tells me that he calls on the doctors only if he finds or suspects glaring problems in the forms. A recent P-form showed 27 cases of acute diarrhea in a week, which Pansare suspected to be higher than usual and so he called the doctor, who told him to lower the number. “He told me to put just the number seven. So I entered seven in the surveillance portal.” In another form, Pansere found a number that looked like “28” scrawled under the column of “influenza-like” illness. “When I called the doctor, he asked me to make it zero. So I did,” he says.

Barriers Remain

With such glaring discrepancies in disease surveillance on the ground, it’s hard to spot an epidemic in a timely manner, leaving India vulnerable to outbreaks. But tools like EMcounter could help address those shortcomings. “Fast reporting and analysis helps us act fast,” says Dawal Shrihari Salve, Nashik’s disease surveillance officer.

Archana Patil of Maharashtra’s health department was satisfied with EMcounter’s performance during the first royal bath, where it recorded, in real-time, various diseases and their symptoms including the early warning signals of H1N1. During the three days when the tablets were piloted during my visit, ten suspected patients of H1N1 were admitted in the district hospital. Of them, three were found positive. Two died.

43
Women washing clothes in the Godavari River during the Kumbh Mela.

Salve, who received both paper and EMcounter records from all Kumbh clinics, says that the electronic tablet was helpful. “In non-tablet clinics, doctors had to stay up late in the night to do the analysis,” he says.

Satish Pawar, director health services for the Maharashtra government, says that EMcounter would be additionally useful in primary health care centers in rural areas. “But we can’t do it without the financial support from central government or from some private body to buy equipment and train a large number of healthcare workers—and most importantly to sustain it,” he says.

Others working in local government sense that the barriers to tracking infectious disease may be insurmountable without significant changes. “The government is least interested in disease surveillance,” says an officer in Nashik’s health department, who asked to speak anonymously out of fear that criticizing the government might cost him his job.

Uday Barve, an epidemiologist at the same department, points to another problem. Some diseases such as H1N1 must be reported within 48 hours, but many cases are not reported at all because private hospitals are not part of the government’s surveillance system. A recent survey reports that about 70% of people in rural areas visit private clinics.

Experts studying disease surveillance say that developing countries like India need to modernize their disease tracking systems to efficiently detect and transparently report outbreaks as well as catch unusual patterns like the re-emergence of a disease.

Vivek Singh, an assistant professor at Indian Institute of Public Health in Hyderabad and vice president of the U.S.-based nonprofit International Society for Disease Surveillance, tells me about a 2008 pilot of an SMS-based disease surveillance system by the government of Andhra Pradesh. It was sophisticated enough to generate automatic alerts when the cases of particular disease crossed a threshold. But the system came to an end because the country’s central government at that time was busy scaling up the paper-based disease surveillance program nationwide. “Central government wanted paper records, and state said send texts,” says Parvathalu Sabhapati, a pharmacist at a primary health center close to Hyderabad. “I preferred texts, but after sometime, nobody bothered whether or not I sent them. Officials shouted at me if I did not send paper records.”

41
A woman, like many others, drinks the river water as a part of her pilgrimage, even though it is unfit for human consumption.

“Unless the political leadership at the center makes it a priority, no amount of pilots are going to help,” Singh says. In November, he will begin working with a team of experts from World Health Organization to evaluate India’s disease surveillance program.

Singh, however, is hopeful that things will change. The Digital India Program introduced by the new government aims to improve citizen access to government services, but he is quick to add that it shouldn’t rely on one kind of technology. “A place where tablets might not work, mobile texts can help,” he says.

India’s central government, however, has not even begun to think about implementing real-time surveillance, let alone which technology to use. “The government has no plan for real-time disease surveillance right now,” says a senior officer with India’s disease surveillance program in New Delhi, who spoke on the condition that he remain anonymous. “The fact is that we need real-time disease surveillance, but like that we need many other good things. Are we ready to do it—or is it a pressing priority now—is the question,” he says.

The officer, who is vaguely aware of EMcounter, feels that a wide rollout of a tablet based real-time disease surveillance system would be a massive undertaking financially. It would also require multiple training sessions for hundreds of thousands of health workers. “Also, we do not even have the basic infrastructure like electricity in many remote areas of the country—forget the internet,” he says.

But for those who worked on EMcounter, the financial investment in prevention is worth the cost. According to UNICEF, the entire budget of renting the tablets and capacity and infrastructure building was about $70,000. “It is very cost effective if you look at the number of people who visited Kumbh,” said Rajlakshmi Nair of UNICEF. It wouldn’t cost more than a dollar or two per person if you take into account the huge number of people that visit Kumbh and the outbreaks that such technologies can prevent, she says.

To Dr. Balsari, the costs are minimal relative to India’s health budget and easily offset by the savings that come from applying the information collected. A developing country, he says, does not have hundreds of millions of dollars to invest in fixed infrastructure such as land-lines or even computers. Using EMcounter on the tablet allows India to “leapfrog” health information systems, he says.

A Change

On the evening after the first royal bath day, a van stops in front of Dr. Bunage’s clinic. Two workers come out and collect the tablet and the paper records. Dr. Bunage zips up her green backpack and emerges from the clinic with the nurse and pharmacist. They switch off the fan, close the windows, and hang a lock on the door. The clinic won’t open again until September 12, a day before the second royal bath.

“It was nice to use the tablets,” she says. “The entries are more accurate. In regular practice, sometimes under stress, we write illegibly and wrong entries creep in which create trouble in analysis later,” says Dr. Bunage, who works in a government hospital in Nashik. Not all doctors share Dr. Bunage’s opinions. Some were reluctant to use the tablets and saw them as an additional task on top of the required paper records. Some entered their data retrospectively when they were free from seeing patients. Even Bunage had to briefly put the tablet aside when she was inundated with patients. “I found it hard to catch up because it is new to me,” says Dr. Bunage. “But I quickly got back to it as the numbers lessened.”

11
Thousands of devotees gather to take a dip at one of the many ghats in the city.

Dr. Balsari is in talks with the Maharashtra government to take EMcounter beyond the Kumbh to primary health centers like the one in Ware. The government is currently reviewing the data. “I would be disappointed if the government does not take it up,” Dr. Balsari says, though he acknowledges that policies don’t always change very rapidly.

The job is too big for one team of researchers to build up India’s health surveillance system, he says. “We can just show—look this works! And we can help facilitate this if the government wants,” he says leaning forward on the couch at the Nashik hotel. “What is satisfying to me is that most doctors who used EMcounter liked it. It is about incremental behavior change,” he says. “It is about pushing the needle.” And, at Kumbh, he felt, “something changed.”

This article is part of the “Next Outbreak” series, a collaboration between NOVA Next and The GroundTruth Project in association with WGBH Boston.