A study by the Lancet Global Health medical journal has found that more than 75,000 people were killed in the first 16 months of the two-year war in Gaza, greatly exceeding the estimate of some 50,000 given by the local authorities at the time.
The research also found that the proportion of women, children and elderly people among those killed as supplied by the Gazan Ministry of Health was accurate. In addition, they found that while non-violent deaths in the period were substantial they were lower than projected.
Researchers noted that “high-quality mortality estimates are crucial for understanding the human cost of conflict” and said that the reports from the Gaza Ministry of Health on the death toll for the Gaza Strip following Oct 7, 2023, had attracted criticism – meaning independent estimates of both violent and non-violent deaths were needed.
Israeli authorities had repeatedly questioned the death tolls during the conflict, arguing they had been “exaggerated” by the Hamas-run Health Ministry.
The Lancet researchers conducted a population-representative household survey, the Gaza Mortality Survey, between Dec 30, 2024, and Jan 5, 2025, surveying 2000 households across 200 primary sampling units.
They said the survey produced three main findings: First, violent deaths “substantially exceeded” even the official Gaza Ministry of Health figures, “with an estimated 3·4% of the Gaza Strip’s population killed up to early January, 2025.”
The finding that 75,200 people are estimated to have died in the period would make the number of deaths in that period of the Gaza war 50% higher than previous estimations.
Second, women, children, and older people comprised 56·2% of violent deaths, which the Lancet researchers said was consistent with the Ministry’s reporting. Third, although non-violent excess deaths were substantial (ie, 8540), they were far lower than some projections and did not exceed violent deaths, they said.
| Estimate | SE (95% CI) | |
|---|---|---|
| Violent deaths | ||
| Children younger than 18 years | 22 800 | 3090 (16 700–28 800) |
| Women aged 18–64 years | 16 600 | 2220 (12 200–20 900) |
| People older than 64 years | 2870 | 943 (1020–4720) |
| Men aged 18–64 years | 32 900 | 2700 (27 600–38 200) |
| Women, children, and older people | 42 200 | 4630 (33 100–51 300) |
| Total violent deaths | 75 200 | 5920 (63 600–86 800) |
| Proportion of violent deaths among women, children, and older people (%) | 56·2% | 2·94 (50·4–61·9) |
| Non-violent deaths | ||
| Total non-violent deaths | 16 300 | 2040 (12 300–20 200) |
| Excess non-violent deaths | 8540 | 2040 (4540–12 500) |
“The study showed that rigorous household mortality surveys are feasible even in extremely challenging conflict environments and challenges assumptions about inevitably high ratios of indirect to direct deaths in modern conflicts,” the researchers said.
The questionnaire used by researchers first established a roster of household members as of Oct 6, 2023. For each person, respondents reported current vital status as one of the following: alive and resident, left Gaza Strip, moved within Gaza Strip, or dead, missing, or imprisoned. For deceased people, respondents classified deaths as violent (killed in or because of fighting), non-violent (disease, pre-existing conditions, accidents, or other causes not directly related to combat) or unknown.
The questionnaire also recorded births after Oct 6, 2023. The Lancet Global Health study covers the period of the war considered the most intense and lethal period, but does not include the period where food shortages became most acute.
The war in Gaza began after a violent Hamas incursion in October 2023 killed about 1,200 people in Israel, mostly civilians, while 250 were taken hostage.
The Lancet Global Health researchers said their findings “contradict claims that the MoH has inflated the death toll from the war in the Gaza Strip.”
“Instead, the MoH appears to provide conservative, reliable figures while working under extraordinary constraints. Our estimates are also incompatible with projections that indirect deaths are at least four times higher than violent deaths or that 62 413 people died of starvation. Although subsequent conditions worsened, our data to early January, 2025, show that violent deaths outnumbered non-violent deaths,” they said.
“Our excess mortality estimate might appear inconsistent with reports of very high mortality among specific vulnerable groups, such as the finding that approximately 40% of patients on dialysis died or left the Gaza Strip. However, such groups represent small fractions of the pre-conflict population of the Gaza Strip, which was approximately 2·2 million people. Even if there were 1000–2000 patients on dialysis and a 40% mortality rate, this would contribute 400–800 excess deaths, which is consistent with our population-level estimate of 8540. Similar patterns probably exist for other medically vulnerable subgroups, each with elevated mortality but little effect on overall excess deaths due to small population sizes,” they said.
“A high ratio of indirect to direct deaths is not inevitable in warfare,” the researchers, who has examined other conflicts, said.
“The Kosovo Memory Book classified only 281 of 13 517 deaths in the 1998–99 war as indirect. In Yemen’s 2015–19 civil war, although overall mortality increased by 17·8% above baseline, direct combat accounted for 67·2% of excess deaths. In Darfur (2003–08), crude mortality rates varied substantially by period, ranging from 0·49 to 5·86 per 10 000 per day depending on the intensity of conflict. In Tigray (2020–23), conflict-attributable child mortality rates increased during active conflict but showed considerable variation by age group and time period,” they said.
“These examples show that patterns vary considerably across conflicts even in challenging humanitarian conditions, with the ratio of indirect to direct deaths depending on conflict characteristics, humanitarian access, and baseline health conditions,” they concluded.
“This study’s main strength was data collection using rigorous survey methodology in an active conflict zone, which was previously claimed to be unfeasible,” they added.
Amongst the limitations of the research noted were: an inability to sample households with zero remaining live members or no members who were 16 years or older; a failure to reach the population remaining in Northern Gaza, Gaza City, or Rafah (though this gap might have lead to underestimation, they said); and respondents from Deir al-Balah and Khan Younis who were displaced to shelters fell outside the sampling frame.
In addition, “Governorate-specific violent death rates should be interpreted cautiously because many displaced people died outside their Governorate of origin, and we did not collect information on death location,” they said.
Addressing the possibility that respondents might have over-reported deaths, the researchers said teams communicated no material benefit from participation, and the mean household size of the sample (4.9), was below the population figure (5·5) and argued “against inflation with non-household relatives”.
They warned that “the vital status of missing people (12 200 estimated) remains unclear, with implications for overall mortality and demographic composition.”