Accuracy and completeness of death recording in primary care electronic health records compared with national death registrations in England using OpenSAFELY-TPP
This study utilizing OpenSAFELY-TPP data reveals that while primary care electronic health records in England demonstrate high overall completeness and concordance with national death registrations, significant discrepancies persist for recent deaths and specific demographic groups, necessitating careful consideration of data recency and population characteristics in research.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
In the United Kingdom, the story of a person's life and death is recorded in two distinct places. One is the official, legal record kept by the government, a comprehensive ledger where every death is registered with a specific date and cause. This is the gold standard, the definitive account used to understand how many people are dying and why. The other record lives in the digital files of family doctors. These electronic health records are the daily logbooks of the National Health Service, tracking everything from a child's vaccination to an elderly patient's medication. While these files are primarily designed to help doctors care for the living, researchers have increasingly turned to them to study health trends, hoping to find patterns that the official government records might miss or that are too slow to reveal.
The question at the heart of this research is simple but vital: do these two records tell the same story? If a doctor's digital file says a patient died on a Tuesday, does the government's official register agree? And if there is a difference, who is missing from the picture? Researchers in England set out to answer this by comparing millions of entries from family doctors with the national death registry. They wanted to know if the digital files were reliable enough to be used for serious scientific study, or if gaps in the recording meant that certain groups of people were being overlooked in our understanding of mortality.
To find the answer, the team looked at data covering the years 2020 through 2025, a period that included the pandemic and its aftermath. They examined the records of over 1.4 million people who had died and were registered with a family doctor using a specific software system. By linking these individual files to the official government database, they could see exactly where the two sources agreed and where they diverged. The results showed that for the vast majority of people, the two records were in perfect harmony. In nearly 90 percent of cases, the date of death recorded by the family doctor was exactly the same as the date on the official government certificate. This high level of agreement suggests that, for most people, the digital files are a trustworthy reflection of reality.
However, the study also revealed that the picture is not perfectly clear for everyone, and the clarity depends heavily on when the death occurred and who the person was. For deaths that happened very recently, the official government records were often not yet available in the research database. Because the government process involves a legal registration period and verification steps, there is a natural delay. Consequently, in the months leading up to the study's end, a growing number of deaths appeared in the family doctors' files but had not yet shown up in the official government list. This gap was most noticeable in early 2026, where nearly 7 percent of recent deaths recorded by doctors had not yet been matched with an official government record.
The study also found that the completeness of these records varied across different groups of people. While the records were generally strong for older adults, they were less complete for younger people. Similarly, the records were more likely to be missing or delayed for people from certain ethnic backgrounds, including those identifying as Black, Asian, or from mixed heritage, compared to those identifying as White. Geographically, the records were less consistent in London and the West Midlands than in other parts of the country. These differences suggest that if researchers only look at one source of data without checking the other, they might accidentally miss a portion of the population, leading to an incomplete understanding of health trends.
Another significant finding concerned the timing of when a person is removed from a family doctor's list. In the past, it was common for a patient's registration to end long after they had died, or sometimes before. But the data showed a dramatic shift in recent years. By 2025, in 85 percent of cases, the family doctor's system was updated to remove the patient on the exact same day the official death was recorded. This rapid administrative update is a positive sign, indicating that the systems are becoming more synchronized. Yet, the study noted that for a small number of people, the doctor's system still did not record the death at all, even though the government knew about it.
The researchers concluded that while the digital records in family doctors' offices are highly reliable and match the official government data for the vast majority of cases, they are not perfect. The records are most reliable for older deaths and for the general population, but they require careful handling when studying recent events or specific demographic groups. The study recommends that scientists should primarily rely on the official government records for the most accurate dates, but they should also check the family doctors' files, especially when looking at very recent time periods or specific communities. By combining these two sources, researchers can ensure they are seeing the full picture, capturing the stories of those who might otherwise be missed in the data. This careful approach ensures that the lessons learned from these records are as accurate and inclusive as possible.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.