Hospital trends in pancreatic cancer admissions: Insights from Florida, 2010-2020

In the United States (U.S.), pancreatic cancer (PC) ranks as the 12th most common cause of cancer, but it is the 4th most common cause of cancer-related deaths.^[1] This makes it the most challenging type of cancer because of its poor prognosis and high case fatality. Current estimates anticipate about 66,440 new cases and 51,750 deaths by 2024. With an aging population and changing demographics, the incidence of PC is expected to increase. While five-year survival rates have improved in the last 20 years, PC continues to be a deadly solid organ malignancy.^[2] The poor prognosis stems from several factors, which include late-stage presentation, aggressive tumor biology, and limited therapeutic responsiveness, which, combined, lead to increased hospitalizations, surgical interventions, and complication management.^[3] As one of the fastest-growing and aging states in the (U.S.), Florida exemplifies the national burden as it ranks the 2nd in PC deaths.^[4] This is also amplified when taking into account the multifaceted disparities due to social determinants of health disparities factors.^[5] Prior studies have indicated that various sociodemographic factors contribute to disparities in PC care and outcomes. A recent study found that patients with lower socioeconomic status often face delays in treatment, which can exacerbate health outcomes.^[6] A recent 2025 study found that Black/African American individuals face substantially elevated mortality risks, with adjusted odds ratios of 1.339 compared to White individuals.^[7] It is important to note that these disparities extend beyond mortality but also to care access and quality of care, such as length of stay (LOS).^[4]

In PC care, hospital characteristics have been shown to influence patient outcomes (in-hospital mortality and LOS). Factors such as teaching status, hospital size measured by the number of staffed beds, and nursing ratios are associated with patient outcomes, suggesting the need to understand these factors and optimize them.^[8,9,10,11] However, current gaps exist in the literature, and prior studies have inadequately examined relationships between patient demographics, hospital characteristics, and county-level social determinants of health in determining PC hospitalization outcomes in a temporal context.^[9] Furthermore, to our knowledge, no prior state-level study has applied a multilevel mixed-effects framework to model patient-, hospital-, and county-level predictors of both LOS and in-hospital mortality simultaneously in a PC-specific sample, leaving the relative contribution of each level undefined.

While overall survival for PC remains poor, understanding modifiable factors associated with LOS and in-hospital mortality remains clinically important. Shorter hospital stays with equivalent outcomes reflect more efficient care delivery and reduced patient burden, while in-hospital mortality represents potentially preventable deaths during acute care episodes that may be amenable to system-level interventions.^[12] Furthermore, there is no clear evidence on the temporal change in hospital outcome patterns.^[12] We hypothesized that hospital structural characteristics, particularly size and teaching status, would demonstrate volume-outcome relationships, while patient-level social determinants, including race/ethnicity and insurance status, would reveal persistent disparities in acute care outcomes. Therefore, this study addressed current literature gaps by investigating the temporal trends in PC hospital admissions, LOS, and in-hospital mortality in Florida from 2010 to 2020, examining patient, hospital, and county-level predictors of these outcomes.

2.1Data source

2.2Participants

2.3Measures

2.4Statistical analysis

2.5Ethical considerations

3.1Trends in admission diagnoses

Analysis of admission diagnoses from 2010 to 2020 revealed significant temporal trends and coding transitions. Abdominal pain is the leading diagnosis throughout the decade, accounting for almost 12% of admissions in 2016 before declining to 8.5% by 2020 (see Figure 1). PC, which remained the second most common diagnosis throughout the study period, showed a significant upward trend, increasing from 4.68% in 2010 to 7.09% in 2020 (p < .001). Chi-square analysis for trend confirmed this increase (χ² = 156.3, p < .001). Sepsis emerged dramatically after 2017, rising to become the third most common diagnosis by 2020 (approximately 6%). Fever and pancreatic head cancer remained relatively stable at 3%-4.5% of admissions. Statistical comparison, not shown, of diagnostic patterns before and after the ICD-10 transition showed no significant disruption in overall distribution (Mann-Whitney U = 45.3, p = .08), suggesting successful code mapping between systems.

Figure 1.

Top 5 diagnosis trends over time (2010-2020)

3.2Patient and hospital characteristics

3.3Predictors of length of stay

3.4Predictors of in-hospital mortality

The significant increase in pancreatic cancer as a primary admission diagnosis parallels national trends documented in previous studies. This trend parallels national data reported by Bhandari, Abdul ^[21] and Dahiya, Inamdar ^[9] suggesting systemic factors driving increased healthcare utilization for pancreatic cancer patients, consistent with national data. The magnitude of this increase cannot be attributed solely to population aging: the data demonstrate concurrent increases in comorbidity burden as measured by the Elixhauser Index, indicating more medically complex patients presenting for care. These trends directly address the study’s first objective and establish the epidemiological context in which hospital and patient-level predictors operate.

The emergence of sepsis as a leading diagnosis after 2017 merits particular attention from a clinical standpoint. This pattern may reflect enhanced sepsis recognition following implementation of Sepsis-3 criteria and systematic screening protocols, or may indicate genuine changes in disease presentation or complications.^{[22, 23]} From a gastroenterological perspective, this trend warrants investigation of potential relationships with interventional procedures, biliary obstruction management, or immunocompromised states associated with chemotherapy regimens.

Addressing the study’s temporal trend objective for in-hospital mortality, the data show stable rates over a decade despite a significant increase in patient complexity. Mean patient age and the Elixhauser Comorbidity Index both rose, indicating that hospitals were treating older and sicker patients. Maintaining stable mortality in this context is significant: it suggests that advances in perioperative care, critical care management, and complication response have offset the expected rise in mortality driven by patient acuity. This finding is clinically meaningful because it reframes stable mortality not as stagnation but as a quality achievement against a deteriorating case-mix baseline.^[24] Addressing the LOS objective directly, the data show a modest but significant reduction in LOS over the study period. This gain was achieved despite rising comorbidity burden and likely reflects improvements in care coordination, including the implementation of enhanced recovery after surgery (ERAS) protocols that have demonstrated effectiveness in major pancreatic surgeries.^[25] The significance of this finding extends beyond efficiency: shorter stays with stable mortality indicate improved care value, not simply cost compression. However, this analysis does not capture post-discharge outcomes, and the clinical implications of shorter stays must be weighed against potential readmission risk.

The observed racial disparities in mortality outcomes confirm the study’s hypothesis that patient-level social determinants reveal persistent disparities in acute care outcomes. Black patients faced 18% higher in-hospital mortality odds compared to Non-Hispanic White patients, a finding that persisted after adjusting for age, comorbidity, insurance status, and hospital characteristics. The significance of this finding lies in its magnitude and its resistance to risk adjustment: the disparity cannot be explained by the clinical factors measured, implicating structural barriers including differential access to subspecialty care, implicit bias in clinical decision-making, and unequal utilization of palliative resources. These findings align with established literature demonstrating reduced access to care, delayed diagnosis, and differential treatment patterns for racial minorities with pancreatic cancer.^[26,27,28]

The insurance-related findings address the hypothesis regarding patient-level predictors and extend it to the domain of financial access. Federal insurance coverage—primarily Medicare—is associated with 30% lower odds of in-hospital mortality (OR = 0.70, p < .001), and state insurance with 18% lower odds (OR = 0.82, p = .057). These associations carry specific significance for end-of-life care policy. The protective association between federal insurance (primarily Medicare) and in-hospital mortality warrants careful interpretation. For a disease with a poor overall prognosis, in-hospital mortality may reflect where death occurs rather than whether it can be prevented. Insured patients likely have greater access to hospice and palliative care services, enabling discharge to home or dedicated facilities for end-of-life care. A recent 2022 study showed that palliative care utilization has increased from 2.9% to 11.9% over time among insured patients.^[29] Uninsured patients may lack these options, resulting in death during hospitalization by default rather than care failure. Therefore, rather than suggesting that coverage expansion reduces PC mortality, these findings suggest that insurance facilitates appropriate transitions to community-based end-of-life care. A recent systematic review of 29 observational studies found evidence to support that health insurance interruption adversely affects cancer care.^[30] This interpretation aligns with the shorter LOS observed among insured patients, potentially reflecting more efficient discharge planning and post-acute care coordination. The substantially elevated mortality risk for patients with “other” insurance types (OR = 7.74) warrants additional investigation, as this category may include patients with specific coverage limitations or administrative barriers to care.^[31] The pronounced volume-outcome relationships observed in this analysis reinforce the importance of care centralization for complex pancreatic procedures.

Rural location is associated with substantially elevated in-hospital mortality (OR = 2.37, CI: 1.19–4.72, p = .014). Rural patients with pancreatic cancer face documented barriers, including delayed diagnosis, limited access to specialized oncology care, and lower rates of guideline-concordant treatment.^[32] The rural-by-bed-size interaction approaches but does not reach significance (OR = 0.73, p = .111), suggesting that larger rural hospitals may partially attenuate this mortality gap, though this warrants confirmation in multi-state analyses. These findings carry direct significance for healthcare delivery: they support selective referral of rural PC patients to regional teaching centers rather than closer community hospitals, and investment in rural oncology infrastructure where referral is not feasible. The association between market concentration and improved outcomes presents an intriguing finding that challenges^[33] conventional assumptions about healthcare competition.^[32] The 30% mortality reduction in more concentrated markets may reflect the development of centers of excellence, improved care coordination, or economies of scale in maintaining specialized capabilities. This finding warrants further investigation to understand optimal market structures for complex cancer care.

This analysis of county-level social determinants revealed the complex relationship between population health factors and individual clinical outcomes. The significant association between Social Economic Factors and mortality, while modest in magnitude, suggests that upstream interventions addressing community-level socioeconomic conditions may yield downstream benefits for cancer patients.^[34] This finding supports the growing recognition that healthcare quality cannot be divorced from broader social and economic conditions. The lack of association between most county-level health measures and acute care outcomes may reflect the temporal disconnect between long-term population health factors and immediate hospitalization outcomes. These factors likely exert greater influence on cancer prevention, early detection, and long-term survival rather than acute care mortality. The relatively homogeneous socioeconomic landscape within Florida may also limit the detectability of these relationships compared to more geographically diverse populations.^[35]

Study limitations

This study examined trends in PC hospital admissions, LOS, and in-hospital mortality, and the study findings have direct implications for patient care. First, the persistent racial disparities in mortality outcomes signal the need for targeted screening and early intervention programs in underserved communities. Reducing time to diagnosis and ensuring equitable access to specialized care could improve survival for Black patients and other at-risk populations. Second, the strong volume-outcome relationship supports continued regionalization of PC care. Patients may benefit from referral to high-volume centers for complex procedures, though healthcare systems must balance centralization with geographic access. Third, the protective effect of federal insurance suggests that coverage expansion could improve outcomes for uninsured patients. Policymakers should consider these findings when designing programs to reduce financial barriers to cancer care. Lastly, the stable mortality rates despite increasing patient complexity indicate that advances in perioperative and critical care management have been effective. Hospitals should continue implementing enhanced recovery protocols and evidence-based complication management. Addressing these modifiable factors through policy interventions, care coordination, and targeted outreach programs may lead to a reduction in preventable deaths and improve the quality of life for PC patients.

H.Y.H.: Conception, analyzing data, and design; drafting the article; final approval of the version to be published. Y.B.: Editing the article and revising it critically for important intellectual content; final approval of the version to be published.

No funding.

The authors declare they have no conflicts of interest.

Not applicable.

The Publication Ethics Committee of the Association for Health Sciences and Education. The journal’s policies adhere to the Core Practices established by the Committee on Publication Ethics (COPE).

Not commissioned; externally double-blind peer reviewed.

The data that support the findings of this study are not publicly available due to privacy or ethical restrictions.

No additional data are available.