The Autopsy of Bad Data
The complaint remained identical across three decades. A researcher would appear at Dr. Naresh Tyagi’s door—sometimes an anxious postgraduate student, sometimes a senior faculty member clutching a sheaf of handwritten data—and explain that they had conducted a study and now required his help with the analysis.
Tyagi would adjust his glasses and look at the numbers. He would ask when the study had been designed. Inevitably, he would learn that the design had been finalized months or even years earlier, entirely without his involvement.
“One of the problems I encountered all my life was that people didn’t ask me to help them design the study. They would do the study, come along with the data, and then ask for my help.”
He said this not with bitterness, but with the mild, enduring exasperation of a man who had spent his entire life explaining a fundamental truth of science: a statistician consulted after an experiment is like a doctor called after the patient has died. He could sometimes salvage the analysis; he could perform a brilliant statistical autopsy. But he could never salvage a flawed design. Yet, knowing this, he kept coming to work anyway, chasing the numbers in the hope of catching them early.
Saharanpur to Sevagram
Born on January 6, 1951, in Saharanpur, Uttar Pradesh, Tyagi lost his mother when he was still a child. After the fifth grade, he left home, drifting from hostel to hostel, staying ahead of his studies through the kind of fierce, quiet self-sufficiency that early loss either shatters or cements. He completed his matriculation in Muzaffarnagar, earned a BSc from Vardhaman College in Bijnor, and finished an MSc in Statistics at DAV College in 1974.
He moved to Delhi for a difficult year, surviving a job at a private college that routinely failed to pay its staff. He abandoned a half-started MA in Economics only when a Government of India fellowship came through for the International Institute for Population Sciences (IIPS) in Mumbai. Population studies—the sweeping mathematics of human existence—became the specialization that would eventually lead him to Sevagram.
Between 1977 and 1980, he worked as a Junior Scientific Officer at K.E.M. Hospital in Mumbai, mapping the statistical effects of environmental pollution on human health. A WHO fellowship then took him to Indonesia and Thailand. Upon submitting his findings, the universe presented him with a rare, binary choice. On a single morning in 1980, two envelopes arrived simultaneously: one offering a position in the Indian Police Service, the other offering a post at a rural medical college in a Gandhian village.
He looked at the two divergent paths. He chose Sevagram.
Birth Weight and the BASIC Interpreter
He joined MGIMS on September 10, 1980, as a Lecturer in Statistics and Demography. It was a deeply unusual position: a non-medical statistician embedded inside a medical institution. His value was entirely dependent on the quality of the questions the clinicians around him were willing to ask—and on how early in their research they thought to ask them.
His closest intellectual partnership was forged with Dr. B.D. Bhatia, an Associate Professor of Paediatrics. Crucially, Dr. Bhatia did not bring Tyagi dead data; he brought him questions before the data existed. Together, they designed and published a groundbreaking series of papers on the relationship between low birth weight and maternal measurements. This work informed Tyagi’s own MPhil thesis and his 1990 PhD in Statistics. The central question—why some babies are born underweight while others are not—was investigated through rigorous multivariate analysis across Wardha district. It was exactly how Tyagi believed all clinical research should function: a doctor and a mathematician, arriving at the truth together.
In 1982, when MGIMS acquired a computer from the ICMR, Tyagi sat down and wrote original programs in BASIC and FORTRAN to handle salary calculations and student admissions. Before the machine arrived, he had run the entire biostatistics operation on fiercely systematic paper-based methods, designed to impose rigid consistency on data that would otherwise arrive in chaotic, incompatible forms from different wards. His wife, Sudesh, joined MGIMS as a computer operator, working alongside him in the Medical Records Department to collect, enter, and clean the hospital’s data. The institute’s records from that era carry the unmistakable, reliable hallmark of two people who took numbers seriously enough to treat them with reverence.
The Bombay Stock Exchange and the Wardha Road
Numbers governed his professional life, but they also dictated his personal rhythms. Every morning, Tyagi walked the Sevagram-Wardha road carrying a walking stick. It was not a social stroll; it was the kind of solitary, kinetic movement that a highly analytical mind requires to process complex probability. Those who passed him on those mornings knew better than to interrupt unless they had an equation worth solving.
In the evenings, the rhythm changed. Chess was his particular arena. The campus provided a rotating cast of formidable opponents—Dr. S.C. Varshney, Dr. R.M. Raizada, Dr. U.S. Jajoo, Dr. S.P. Kalantri, and Dr. G.V.S. Murthy. The chess games were fiercely strategic; the conversation around them was wonderfully light.
He was known for speaking his mind with mathematical precision. In an institutional setting, where diplomacy often dilutes the truth, his directness was legendary. He refused to soften assessments he knew to be accurate, and he never inflated praise for sloppy research. In a discipline where the integrity of a study depends entirely on an honest critique at the design stage, his bluntness was professionally essential, even if occasionally unwelcome.
His other great passion, entirely removed from medicine, was the Bombay Stock Exchange. Beginning in 1976 with the purchase of Tata Motors shares, Tyagi applied the exact same analytical rigor to the market that he brought to epidemiology: tracking trends, calculating moving averages, and making his own assessments rather than following the herd. Whether this obsession made him wealthy is unrecorded. That it engaged his brilliant, number-hungry mind is absolute fact.
The Department He Founded Twice
He retired from MGIMS in January 2011, packing up thirty years of data. But men who chase numbers rarely stop counting. He immediately joined KLE University in Belagavi, Karnataka, as Professor and Head of Epidemiology and Biostatistics. In a sense, he was simply founding the same department he had built at MGIMS in 1982 all over again. Over the next decade, he developed curricula, expanded a PhD program, and was made a Fellow of the Indian Society for Medical Statistics.
His legacy at MGIMS, however, remained profoundly personal. His son, Vishal, completed his MS in Otolaryngology at MGIMS and now practices in Mumbai. His daughter, Swati, completed her MBBS at MGIMS and practices Obstetrics and Gynaecology in Haryana. Both were trained by the very colleagues their father had sparred with over chessboards and datasets for three decades.
What the Numbers Finally Said
Biostatistics is the subject medical students dread and researchers routinely abuse. They dread it because its logic is alien—the cold language of probability rather than the warm art of diagnosis. They abuse it by consulting its practitioner too late, asking a statistician to bless a badly designed experiment.
Dr. Tyagi spent thirty years at MGIMS fighting to be present at the beginning of the story rather than the end. He succeeded often enough to accumulate a monumental body of work: the birth weight studies, a hundred published papers, the original BASIC programs, and a generation of students who finally understood that a statistician is not a calculator, but a collaborator.
The core problem he described—that people came to him too late—was never fully solved. But the people who did sit across from his desk learned what it looked like to do science correctly. And that is exactly what a master teacher leaves behind: not a perfectly solved equation, but a much clearer understanding of the math.