A new study has validated QResearch-linked electronic health records as a nationally representative resource for cancer research in England.
Published in ESMO Real World Data & Digital Oncology and led by Professor Simon Lord, the study found that cancer cases in the linked dataset closely reflect national cancer statistics in terms of age, sex, cancer stage, and treatment characteristics, positioning it as a rich resource for population-level oncology research.
QResearch is a large primary care database containing longitudinal health records for more than 13 million registered patients from over 1500 GP practices (qresearch.org). Records extend back to 1990 and include detailed information on demographics, diagnoses, comorbidities, prescribed medications, laboratory results, and lifestyle factors. QResearch records have been linked to national hospital, cancer registry, and mortality datasets, creating one of the largest linked health data resources for oncology research in England.
Linked electronic health records are valuable for detailed analysis of real-world cancer care. Combining datasets enables researchers to interrogate lifestyle factors, medication exposure, and comorbidities alongside cancer diagnosis, treatment, and outcomes in ways that would not otherwise be possible. QResearch-linked data have already supported the development of risk prediction models, such as those estimating breast cancer-specific mortality.
Until now, the representativeness of the QResearch population against national cancer statistics has not been assessed. In this study, the research team conducted a retrospective validation of QResearch-linked oncology data between 2013-2020.
QResearch primary care records were linked to Hospital Episode Statistics (HES), the National Cancer Registration Dataset (NCRD), Systemic Anti-Cancer Therapy (SACT) data and death registrations, with patients included if a cancer record was identified in any of these sources.
Across seven major cancer sites, the QResearch-linked dataset captured 14.6% of national cancer registrations (n = 196,565/1,348,243). This proportion increased over the study period, consistent with increasing participation in QResearch. Importantly, the characteristics of patients within the linked dataset closely matched national cancer statistics. Distributions of age, sex and cancer stage were similar to national data, while the proportions of patients receiving systemic anti-cancer therapy and radiotherapy also aligned closely with national estimates.
The researchers also investigated whether the cancer population remained representative when NCRD records were unavailable.
Professor Simon Lord said:
“The National Cancer Registration Dataset has near complete data on NHS patients with malignant tumours in England. However, registry releases are subject to time lags, so it was important to assess whether cancer cases identified through primary care, hospital and death registration records remain representative in the absence of registry data. This is particularly important for longitudinal population-based studies that extend into more recent periods.”
The study found that more than 93% of cancer cases could be identified through QResearch, hospital, or death registration data without reliance on the NCRD.
Cases recorded only in the NCRD were more commonly older adults and those with earlier-stage disease. They also included a higher proportion of melanoma and kidney cancers and a lower proportion of prostate cancers. Excluding NCRD-only cases had negligible impact on the overall demographic structure or distributions of cancer site, stage and treatment.
Dr Sanat Kulkarni, NIHR Academic Clinical Fellow at the University of Oxford and study author commented:
“Our study demonstrates that the QResearch-linked dataset offers a nationally representative and comprehensive platform for population oncology research, even when NCRD data are unavailable. Integrating detailed primary care information with national cancer and treatment data creates important opportunities to understand cancer care and outcomes in the context of patients’ wider health.”
The study highlights the importance of multi-source data for creating a more complete picture of cancer care, with different records capturing different aspects of a patient’s journey. In particular, this linked dataset can support research into treatment selection, adherence, toxicity and outcomes, including the effects of multimorbidity, polypharmacy and lifestyle factors that are difficult to study using cancer records alone.
By demonstrating that this linked population closely reflects national cancer statistics, the study provides important validation for its use in future population-based studies.
This research was funded by World Cancer Research Fund (WCRF), as part of the World Cancer Research Fund International grant programme.'QResearch primary care records linked to national hospital and cancer registry data: a validation study’ was published in ESMO Real World Data & Digital Oncology in August 2026.
