A statistician who learnt from the person holding the pen

The first data set I was trusted with was a household survey in 2000, in villages north of Kochi. I had completed a master's degree in statistics at Ponnezhuthu Institute of Applied Statistics that May, after a bachelor's degree at Mazhavilkara College. At Ithalmara Field Research, a small survey company, I checked questionnaires as they came back from field investigators. One investigator's forms were too tidy: ages ending in zero or five, and a working tap in every house. He had not invented anything deliberately. He had rounded what respondents were unsure of and filled gaps with what seemed likely. I learnt then that data quality is mostly decided by the person holding the pen, and that correcting a table later never recovers what was not asked at the time.

In March 2004 I moved to Varnodhi Claims Administration in Bengaluru, a licensed third-party administrator processing health-insurance claims for insurers and corporate group policies. I joined as a senior analyst and later led a six-person claims analytics team. My job was to study what was paid, reduced and rejected. The commonest reason a hospital lost money on a claim was neither fraud nor a policy exclusion. It was a missing document, an unexplained extra day or a diagnosis that did not account for the procedure billed. Those files gave me an unusual education in how hospitals write, because I read their discharge summaries as a stranger would.

I crossed to the provider side in September 2008 as manager of management information at Sevanthira Memorial Hospital, a 550-bed charitable trust hospital in Bengaluru. I reported to the hospital administrator and led nine people. The hospital had excellent doctors and poor records. Bed occupancy was calculated three different ways, and the medical-records department was treated as a storeroom. I spent much of my time with ward sisters and records clerks, learning why a figure was entered late or in the wrong field. Most of the reasons were sensible. Fixing them meant changing forms and routines, not issuing reminders.

Chembakanad Healthcare recruited me in January 2014 to lead medical records and information systems across its five hospitals. The company was growing into northern Kerala and then Karnataka, and each site brought its own registration habits. As vice-president for health information and analytics from July 2019, I took on clinical coding, group reporting and the data workstream of a new hospital information system. The pandemic years tested that work. Daily bed and oxygen returns to district health authorities had to be accurate by evening, from wards working under great strain. We cut the ward return to six fields and named one person in each hospital to confirm it.

The hardest decision of that period came in 2022, before the new system went live. I insisted that 61,000 uncertain patient matches be reviewed by medical-records staff rather than merged automatically. It delayed the unified record by fourteen weeks and meant our Mysuru hospital opened on the old system. Some colleagues still think the delay was too long. I think a wrongly merged chart, carrying another person's allergy or blood group, is a clinical risk that a project timetable should not decide.

I became chief data officer in April 2023. The appointment created a single office for coding, records, engineering, analytics and data governance, with an annual budget of ₹19 crore. I present to the board's audit committee on data quality and claims integrity, and I report to its risk management committee on privacy incidents and on our preparation for the Digital Personal Data Protection Act. Our privacy officer reports to the general counsel rather than to me. I asked for that arrangement so that the person who reviews our use of patient data is not the person who runs it.

I am a statistician, not a clinician. I do not decide what is clinically correct in a note, and I do not set treatment protocols. My work is to make sure that what doctors, nurses and registration staff record can be relied upon by the next person who reads it, whether that is a doctor on another shift, an insurer's medical officer or a patient opening their own report. I lead mainly by asking department heads to own their data elements and by giving them evidence of where their records fail.

I now want a different kind of responsibility. I passed the IICA online proficiency self-assessment in 2025 and would like to serve on a board where data about people is central to the business. I expect to learn at least as much as I contribute.

Open to Board appointments

Languages / English · Malayalam · Kannada · Hindi