Diagnosis made during or shortly after emergency hospital care (here termed emergency diagnosis) is well recognised in oncology as being associated with advanced disease and poorer outcomes. The reasons for emergency diagnoses are complex—ranging from rapid disease progression to patient- and system-level delays—and these drivers are unlikely to be unique to cancer. There is limited evidence documenting how often emergency diagnosis occurs and how it relates to prognosis across a range of non-neoplastic conditions. This study examined the frequency of emergency diagnosis and its association with clinical outcomes across 13 exemplar non-neoplastic conditions using linked routine health records in England.
Researchers analysed linked data from primary care, secondary care, and mortality records for 1,701,154 patients recorded in Clinical Practice Research Datalink (CPRD) and linked Hospital Episode Statistics (HES) and Office for National Statistics (ONS) datasets. The study window spanned 1999–2019. The 13 exemplar conditions examined were axial spondyloarthritis, coeliac disease, coronary/ischaemic heart disease, chronic obstructive pulmonary disease (COPD), inflammatory bowel disease (IBD), Lyme disease, multiple sclerosis (MS), Parkinson’s disease, polycystic ovary syndrome, rheumatoid arthritis, schizophrenia, subacute bacterial endocarditis, and tuberculosis.
The primary exposures were whether the recorded diagnosis occurred at or soon after an emergency hospital contact (emergency diagnosis) versus not. Outcomes compared between groups were 1-year mortality and time spent in hospital in the year after diagnosis. Analyses adjusted for age and year of diagnosis, socioeconomic deprivation, comorbidity burden, and the healthcare setting in which the diagnosis was made. Results were examined across two CPRD subsets (Aurum and GOLD) and supplementary analyses of five cancer sites were performed to assess concordance with prior literature.
Emergency diagnosis occurred frequently across the examined conditions. At least 20% of patients experienced emergency diagnosis for 9 of the 13 conditions. In particular, more than 30% of patients with Parkinson’s disease and 35% of patients with COPD were diagnosed following an emergency presentation.
Patients diagnosed as emergencies had notably worse outcomes in the year after diagnosis. For 9 of the 13 conditions there was a substantial absolute increase (at least a +10% difference) in 1-year mortality among those with emergency diagnoses compared with those without. After multivariable adjustment, emergency diagnosis was typically associated with markedly higher odds of death within one year. Examples reported in the source include adjusted odds ratios for women with coeliac disease of 7.53 (95% CI 5.64–10.1) and for men with IBD of 5.99 (95% CI 5.30–6.78).
Time spent in hospital during the year following diagnosis was also substantially greater in emergency-diagnosed patients. The study reports adjusted rate ratios such as 8.76 (95% CI 6.52–11.8) for men with coeliac disease and 5.59 (95% CI 2.57–12.2) for women with Lyme disease.
Results were consistent when analyses were stratified by CPRD data subset (Aurum and GOLD). Supplementary analyses applying the same methods to five cancer sites (brain, colon, lung, pancreas, and ovary) produced findings concordant with existing literature, supporting the internal validity of the analytic approach.
Emergency diagnosis is a common feature across a diverse set of non-neoplastic conditions and is associated with worse short-term clinical outcomes, including higher 1-year mortality and increased hospital utilization. The magnitude of association was large for some conditions that typically have a favorable prognosis when diagnosed non-emergently (for example, coeliac disease and IBD), indicating that emergency presentation may identify a subgroup at particularly high risk.
The observed associations persisted after adjustment for measurable confounders (age, year, deprivation, comorbidity burden, and care setting), suggesting that emergency diagnosis status is an important marker of prognosis across multiple conditions. However, causality cannot be inferred from these observational data: emergency diagnosis may reflect underlying disease severity, delayed detection, or other unmeasured factors.
Strengths of the study include the large, population-based sample (≈1.7 million patients), the use of linked primary care, hospital, and mortality data, and consistency of findings across CPRD subsets and supplementary cancer-site analyses.
The main limitation is the reliance on routine health-record coding: the study assumes accurate recording of diagnoses and dates. The operational definition of emergency diagnosis may capture some patients whose emergency episode was unrelated to the subsequently recorded condition, and residual confounding from unmeasured factors is possible. Further detail on exact case definitions, code lists, and temporal windows used for assigning emergency diagnosis are reported in the full manuscript and appendices; those specifics were not reproduced in this summary.
These findings indicate that emergency diagnosis is not confined to cancer and is associated with worse outcomes across many non-neoplastic conditions. The authors recommend targeted research to understand the drivers of emergency diagnosis both generally and for individual diseases. They suggest investment to reduce emergency diagnoses, especially for conditions where diagnosis in community settings would be expected (for example, rheumatoid arthritis and COPD) and for conditions where emergency diagnosis correlates with particularly poor outcomes (such as coeliac disease, IBD, and MS).
It remains to be established whether preventing emergency diagnoses or improving post-emergency management would change outcomes; interventional studies or system-level evaluations would be required to assess the effect of such strategies.
In this large, linked-records study in England, emergency diagnosis was common across a range of 13 non-neoplastic conditions and was associated with substantially worse 1-year mortality and increased hospital use in many conditions. The authors call for further research and health-system action to understand and reduce emergency diagnoses where feasible.