Emergency Presentation (EP), where cancer is diagnosed following unplanned hospitalization, is a significant concern, particularly for colorectal cancer. While improvements have been made, a substantial proportion (around 22%) of colorectal cancer patients in England are diagnosed through EP, leading to poorer survival and patient experience. EP reflects delays in diagnosis stemming from both patient-related factors (symptom awareness, cancer awareness) and healthcare system issues (access, service delivery). Studies consistently show inequalities in EP, with older, more deprived, female, and non-white patients, as well as those with comorbidities, facing a higher risk. This increased use of emergency services in deprived areas can't be fully explained by higher comorbidity rates, pointing to systemic healthcare barriers. This study hypothesizes a link between pre-diagnosis hospital admission patterns (both elective and emergency) and EP for colon cancer. By comparing the most and least deprived patients, the researchers aim to understand how pre-diagnosis hospital use contributes to socio-economic inequalities in EP. The primary aims are to assess whether more deprived patients and those diagnosed through EP have higher rates of HEAs in the two years before diagnosis and to identify the conditions most frequently leading to these HEAs.

Existing research indicates that emergency presentation (EP) in colorectal cancer diagnosis is a global issue, linked to poorer outcomes. Inequalities in EP are evident, with socio-economically deprived patients experiencing higher rates of emergency diagnosis. This disparity is partially attributed to factors such as lower symptom awareness, increased comorbidities, and potentially, limited access to healthcare. Previous studies have explored the association between specific comorbidities and EP, but the influence of admission type for these conditions across different socio-demographic groups remains unclear. This research aims to bridge this knowledge gap by examining the pre-diagnostic hospital admission patterns in relation to socio-economic status and their association with EP.

The study utilized data from the English Cancer Registry and Hospital Episode Statistics (HES) for all patients diagnosed with colon cancer in England in 2013. Patient characteristics, including deprivation quintile (based on income domain of the English Index of Multiple Deprivation), sex, and age at diagnosis, were obtained from the Cancer Registry. HES data provided information on hospital admissions (HAs) in the two years preceding diagnosis, including admission codes (ICD-10) and admission type (emergency or non-emergency). A total of 15,263 patients with at least one HA were included after excluding those without HES records. The ICD-10 codes were grouped into 42 aggregate condition groups. The analysis involved two main measures: monthly rate of HAs per patient and monthly proportion of patients with at least one emergency admission. A multi-step purposeful variable selection approach was adapted to mixed-effects logistic regression to identify conditions predicting HEAs. This involved stepwise variable selection based on p-values and checking for confounding. Complete separation issues in some variables were addressed. The final model included age and a selected set of condition groups, stratifying by sex and deprivation. Marginal effects and average changes in the probability of HEA were calculated.

The analysis revealed that colon cancer patients diagnosed through EP had a significantly higher number of HEAs than those diagnosed via non-emergency routes, particularly in the seven months preceding diagnosis. The rate of HEAs was consistently higher among the most deprived patients, with the difference increasing in the months leading up to diagnosis. However, the proportion of patients with at least one HEA showed less pronounced deprivation-related differences. Analysis of conditions predictive of HEA showed that some conditions with potentially related symptoms (abdominal/pelvic pain, appendicitis, digestive disorders) increased the probability of HEA. Other conditions unrelated to colon cancer (infectious diseases, injury, poisoning, cardiovascular and respiratory issues, mental health disorders) also strongly predicted HEA. Notably, the baseline probability of HEA was considerably higher in the most deprived male patients than in the least deprived, while the difference was less pronounced in females. Although individual condition effects were similar across deprivation groups, the higher baseline probability in the most deprived males meant conditions explained less of the HEA probability in that group. Fewer conditions predicted HEA in the most deprived males compared to the least deprived, primarily concerning cardiovascular, respiratory, and mental health issues. Differences between the most and least deprived were also noted for specific conditions such as anaemia and gynaecological conditions.

The study's findings highlight that while the overall rate of hospital admissions was similar across deprivation levels, emergency admissions prior to colon cancer diagnosis showed significant socio-economic disparities. The increased HEA rate in the most deprived is not attributable to the specific conditions involved but potentially reflects repeated use of emergency services for non-specific symptoms. This suggests delays in the diagnostic pathway, potentially due to system-level factors rather than solely patient-related factors like symptom awareness or comorbidity burden. The higher risk of EP in women, despite generally higher symptom awareness, might be explained by less specific symptoms that receive benign initial diagnoses. Differences in the association of conditions like anaemia with HEA across sex and deprivation groups suggest potentially delayed recognition or management in certain populations. The prevalence of both acute and chronic conditions associated with HEA emphasizes the need for proactive care management to prevent emergency admissions, particularly among vulnerable groups.

This study demonstrates that socio-economic inequalities in colon cancer diagnosis via emergency presentation are associated with higher rates of pre-diagnostic hospital emergency admissions among the most deprived, especially in the months leading up to diagnosis. This is likely due to system-level barriers, rather than solely patient-related factors. Future research should focus on investigating these systemic barriers and developing interventions to improve access to timely and appropriate healthcare for all populations, regardless of socio-economic status.

While the study used high-quality, population-based data, some limitations exist. The exclusion of patients without HES records (12%) might introduce bias. The purposeful variable selection method, while advantageous for risk factor modeling, tests variables one at a time rather than jointly in later steps. A sensitivity analysis using a penalized algorithm showed some discrepancies with the primary method, potentially reflecting issues with highly correlated covariates. The study focused on colon cancer in 2013; generalizability to other cancers and time periods needs further investigation.