- India does not count harm caused by missing medical history, so this page estimates it the way epidemiologists estimate dengue or malaria: start with everyone, then narrow down step by step (how many are treated, how many are harmed, how many of those harms come from missing history, how many die), using a published rate at each step[1][2][3].
- Central estimate: about 6,400 deaths a year in Indian hospitals from preventable medication errors caused by an incomplete history (very likely between 2,900 and 12,800), about 17 a day.
- In clinics and outpatient care: about 38 lakh medication incidents a year linked to missing history, of which about 1.6 lakh cause severe harm (range 60,000 to 3.4 lakh).
- By age: about 1,400 of the deaths are children and about 2,000 are people over 65, who are 7% of the population but nearly a third of the deaths, because they take more medicines and are hospitalised far more often for drug harm[4].
- Suffering that lasts: about 4,600 people a year are left permanently disabled by these errors in hospital alone; most others recover within six months[5].
- These are estimates with wide ranges, not counts. The method, every input and the code are below so anyone can check or change them.
Why an Estimate, and How Scientists Make One
No Indian register records "harm because the doctor did not know the history". Most errors are never reported, and a death is recorded under the disease, not the error. Epidemiologists face the same gap with dengue and malaria, where most cases are never notified. They do not wait for a perfect count. They build the number from parts that have been measured:
- 1Who is exposedPeople hospitalised or seen in clinics each year
- 2How often harm happensMeasured rates of medication harm per patient or visit
- 3How much is due to the causeThe share of errors traced to an incomplete history
- 4What it leads toSevere harm and death per event
Every input is uncertain, so each is given as a range (low, most likely, high) rather than one number. The model draws 100,000 random combinations from those ranges and reports the middle answer, plus the range that holds 9 in 10 answers. That range is the honest answer: the true number is very likely inside it. Disease-burden studies report uncertainty the same way.
Channel 1: Patients Admitted to Hospital
We start with everyone in India and narrow down, one step at a time, to the people who die because a doctor did not have their medicine history. Each step uses a rate measured in a published study.
Hospital deaths, step by step
- People in India146.39 crore
- About 3 in 100 of them are treated in hospital in a year2.9%, national health survey4.25 crore
- Of those, about 1 in 40 is harmed by a medicine error that could have been prevented2.5%, World Health Organization10.6 lakh
- About 1 in 7 of those errors happens because the doctor did not have the patient's history15%; studies find up to 27%1.59 lakh
- About 3 in 100 of those patients die3%; World Health Organization and a Mumbai hospital study4,776 deaths
This uses the most likely value at each step. Because every value is uncertain, we also repeated the calculation 1 lakh times with values drawn from each range. The middle answer was about 6,400 deaths, and 9 in 10 answers fell between 2,900 and 12,800.
| Input | Value used | Source and reasoning |
|---|---|---|
| Population, 2025 | 146.39 crorefixed | UNFPA, 2025[6] |
| Share of people treated as in-patients in a year | 2.9%range 2.6% to 3.4% | 2.6% rural, 3.4% urban, NSS 75th round[7] |
| In-patients with preventable medication-related harm | 2.5%range 1.5% to 4.1% | WHO: about 1 in 10 patients harmed, over half preventable, half of that medication-related, so 10% × 50% × 50% = 2.5%[1]. Mumbai: 6.89% of admissions caused by drug reactions, 59.6% avoidable, so 4.1%[8]. Low value is our assumption. |
| Share of those errors caused by an incomplete medication history | 15%range 5% to 27% | "Up to 27%" of hospital prescribing errors come from incomplete histories at admission[2]. Low and likely values are our assumptions, deliberately below that ceiling. |
| Deaths per harm event | 3%range 1.9% to 6.4% | WHO: 2.6 million deaths from 134 million adverse events in low- and middle-income hospitals = 1.9%[9]. Mumbai: 17 deaths among 265 patients with drug reactions = 6.4%[8]. |
Channel 2: Clinics and Outpatient Care
Most Indians are treated in clinics, not hospitals. Here the studies measure safety incidents per consultation and the share causing severe harm, but not deaths, so this channel reports severe harm only.
Most Indians are treated in clinics, not hospitals. The same narrowing works here, but studies of clinics measure severe harm rather than deaths.
Severe harm in clinics, step by step
- People in India146.39 crore
- In any two weeks, about 7 or 8 in 100 are ill7.5%, national health survey10.98 crore
- A year has about 24 two-week periods, so clinic visits in a yearcounting one visit per illness; the real number is higher267 crore
- About 1 visit in 40 has a safety problem2.5%, worldwide review of clinic records6.68 crore
- About 1 in 3 of those problems involves a medicine35%, our assumption2.34 crore
- About 1 in 7 of those happens because the history was missing15%; studies find up to 27%35 lakh
- About 1 in 25 of those causes severe harm4%, same worldwide review1.4 lakh a year
| Input | Value used | Source and reasoning |
|---|---|---|
| Share reporting an illness in the last 15 days | 7.5%range 7% to 9% | About 7% rural and 9% urban[7] |
| Safety incidents per consultation | 2.5%range 1.5% to 3.5% | Typically 2 to 3 per 100 consultations in record-review studies worldwide[3] |
| Share of incidents involving medicines | 35%range 25% to 50% | Our assumption: prescribing and diagnosis are the incidents most linked to harm[3], but the review gives no single share |
| Share due to missing history | 15%range 5% to 27% | Same range as the hospital channel[2] |
| Share of incidents causing severe harm | 4%range 1% to 8% | Typically 4% (range under 1% to 44%)[3] |
Results
| Outcome per year | Middle estimate | Very likely between |
|---|---|---|
| People treated as in-patients | 4.32 crore | 3.97 to 4.77 crore |
| In-patients harmed by a history-related, preventable medication error | 1,76,000 | 88,000 to 3,04,000 |
| Deaths among them | 6,370 | 2,890 to 12,840 |
| Outpatient consultations (one per illness) | 277 crore | 257 to 307 crore |
| History-related medication incidents in outpatient care | 38.3 lakh | 18.9 to 67.1 lakh |
| Of which severe harm | 1,57,800 | 60,000 to 3,43,000 |
Children, Adults and Older People
Harm is not spread evenly by age. To split it, each age group's share of the population is weighted by how much more likely its members are to be harmed by medicines compared with adults aged 15 to 64.
Older people take more medicines and are far more likely to be hospitalised because of one. So we weight each age group by its size and by how much more likely it is to be harmed.
Who the hospital deaths are, step by step
- Children: 24 in every 100 Indians, same risk as adultsrisk taken as equal to adults; could be up to 3 timesweight 24
- Adults aged 15 to 64: 69 in every 100 Indiansweight 69
- People over 65: 7 in every 100 Indians, but 6.8 times as likely to be hospitalised by a medicine7 × 6.8; US national dataweight 47.6
- Add the weights140.6
- Each group's share of the deaths is its weight divided by 140.634% over 65 · 49% adults · 17% children
| Group | Share of people | How much more likely to be harmed | Source and reasoning |
|---|---|---|---|
| Children, 0 to 14 | 24%[6] | Same as adults (could be half to 3 times) | Per medication order, children had three times the rate of potentially harmful errors, mostly wrong doses[10]; per child, fewer medicines are taken, so the likely value is 1 (our assumption). Drug reactions cause 2.9% of children's admissions[11] |
| Adults, 15 to 64 | 69% | The comparison group | The comparison group |
| Older people, 65 and over | 7%[6] | 6.8 times for hospital stays (range 4.3 to 9.2); 2.4 times for emergency visits (1.8 to 3.0) | In US national surveillance, people over 65 were 2.4 times as likely to need emergency care for drug harm and 6.8 times as likely to be hospitalised for it[4] |
| Per year: middle estimate, with the likely range below | Children | Adults 15 to 64 | Over 65 |
|---|---|---|---|
| Hospital patients harmed | 39,00016,800 to 80,200 | 79,90039,400 to 1,41,200 | 54,30026,400 to 97,900 |
| Hospital deaths | 1,420560 to 3,300 | 2,8901,290 to 5,940 | 1,960860 to 4,090 |
| Severe harm in clinics and outpatient care | 43,70015,200 to 1,06,700 | 89,80033,500 to 1,99,500 | 21,7008,100 to 49,000 |
| History-related medication incidents in outpatient care | 10.6 lakh4.6 to 21.5 lakh | 21.8 lakh10.6 to 39.3 lakh | 5.3 lakh2.5 to 9.7 lakh |
The age split is the least certain part of the model. The risk figures for older people come from the United States, and India has no national data on drug harm by age. The ranges are wide for that reason.
Lasting Harm: People Still Suffering After Treatment
Many people harmed by a medication error are treated and recover. Some are not. The best long-term evidence on what happens to hospital patients after an adverse event is the Harvard Medical Practice Study, which reviewed 30,121 records from 51 New York hospitals in 1984: 70.5% of adverse events caused disability lasting under six months, 2.6% caused permanent disability and 13.6% led to death[5].
What happens to the people harmed in hospital
- Hospital patients harmed each year by these errorsmiddle answer of the model1,76,000
- About 7 in 10 recover within six months70.5%, Harvard Medical Practice Study1.24 lakh
- About 1 in 40 are left permanently disabled2.6%, same study; range 2,300 to 7,900about 4,600
The Harvard study found that 13.6% of adverse events ended in death. Our model assumes far fewer (1.9% to 6.4%), so our death figure is on the cautious side.
In clinics, the 1.6 lakh severe harms a year are by definition the ones that cause serious or lasting damage[3]. Combined with the hospital figures, an incomplete medical history leaves about 1.6 lakh Indians a year with severe or permanent harm from a medicine error, on top of the deaths.
Is the Number Plausible?
A good estimate should sit sensibly inside larger, independently measured numbers.
- Against the WHO figure. WHO counts 2.6 million deaths a year from unsafe hospital care in low- and middle-income countries[9]. Our hospital estimate covers only one cause (missing history) and one type of harm (medicines).
- Against Indian hospital data. In one Mumbai emergency department, 0.83% of all admitted patients died from a drug reaction[8]; our model assumes far fewer deaths per admission because most Indian admissions are less severe than a tertiary emergency.
What This Estimate Does Not Do
- It is not a count. It is a structured estimate. The range is the honest answer; the middle figure is a summary.
- It leaves out harm we cannot quantify yet: wrong diagnoses because earlier test results were unavailable, emergencies where no one knows the blood group or allergies, Ayurvedic and homeopathic products the doctor never hears about, and families who never learn of an inherited disease.
- It counts no deaths in outpatient care, because no reliable death rate per outpatient incident exists. The true death toll is therefore higher than the hospital figure alone.
- Several inputs come from other countries. Indian rates are likely higher, because Indian clinics have less time and fewer records; we did not inflate them.
The model is a short Python program with every input labelled by its source. Read the model code; anyone can change an input and rerun it. Next: what changes when the history is available →
Sources
- World Health Organization. Patient safety, fact sheet, 11 September 2023. who.int
- Tam VC, Knowles SR, Cornish PL, et al. Frequency, type and clinical importance of medication history errors at admission to hospital: a systematic review. CMAJ, 2005;173(5):510. cmaj.ca
- Panesar SS, deSilva D, Carson-Stevens A, et al. How safe is primary care? A systematic review. BMJ Quality and Safety, 2016; NIHR Evidence summary. evidence.nihr.ac.uk
- Budnitz DS, Pollock DA, Weidenbach KN, et al. National surveillance of emergency department visits for outpatient adverse drug events. JAMA, 2006. read.qxmd.com
- Brennan TA, Leape LL, Laird NM, et al. Incidence of adverse events and negligence in hospitalized patients: results of the Harvard Medical Practice Study I. NEJM, 1991. read.qxmd.com
- UNFPA State of World Population 2025, as reported by ThePrint, 10 June 2025: India 1,463.9 million, 24% aged 0 to 14. theprint.in
- National Statistical Office. Key Indicators of Social Consumption in India: Health, NSS 75th round (July 2017 to June 2018), Report 586. mospi.gov.in
- Patel KJ, Kedia MS, Bajpai D, et al. Evaluation of the prevalence and economic burden of adverse drug reactions presenting to the medical emergency department of a tertiary referral centre. BMC Clinical Pharmacology, 2007 (KEM Hospital, Mumbai). link.springer.com
- World Health Organization. 10 facts on patient safety (134 million adverse events and 2.6 million deaths a year in hospitals in low- and middle-income countries). who.int
- Kaushal R, Bates DW, Landrigan C, et al. Medication errors and adverse drug events in pediatric inpatients. JAMA, 2001. scispace.com
- Smyth RMD, Gargon E, Kirkham J, et al. Adverse drug reactions in children: a systematic review. PLoS One, 2012. pmc.ncbi.nlm.nih.gov
- Kruk ME, et al. Lancet Global Health Commission on high-quality health systems, 2018, as reported by The Quint (IndiaSpend). thequint.com