Effect of Hospital Teaching Status on Outcomes of Patients With Acute Pancreatitis

Introduction Multiple studies have shown that outcomes of various diseases differ by the hospital teaching status. However, not much is known about the effects of hospital teaching status on outcomes of acute pancreatitis (AP). The aim of this study was to identify if there was an effect of hospital teaching status on the outcomes of AP. Methods The National Inpatient Sample (NIS) database was used to identify patients with a discharge diagnosis of AP from 2016 to 2019. Patients were classified according to whether they were admitted to teaching hospitals (TH) or non-teaching hospitals (NTH). Study outcomes were the length of stay (LOS), total hospitalization cost and charge, sepsis, shock, acute kidney injury, ICU admission, and mortality. Results A total of 1,689,334 patients were included in the study. Of these, 65.06% were in the TH group, while 34.94% were in the NTH group. Patients admitted to TH had a higher incidence of AKI (18.84% vs. 15.79%, p<0.001), shock (4.32% vs. 2.7%, p<0.001), sepsis (4.48% vs. 3.65%, p<0.001), and ICU admissions (4.78% vs. 2.90%, p<0.001) than NTH. Patients admitted to TH also had a higher length of stay (5.82 vs. 4.54 days, p<0.001) and higher hospitalization charges ($47,390 vs. $65,380, p<0.001). The mortality rate in TH was 2.25% compared to 1.5% in NTH (p<0.001). Conclusion Patients admitted to TH had higher mortality as compared to NTH. While the exact reason for this is unknown, it can be partially explained by a higher incidence of AKI, shock, and sepsis. Furthermore, ICU admissions were higher in TH, indicating higher resource utilization.


Introduction
The US healthcare system can be broadly divided into two major hospital systems, teaching hospitals (TH) and non-teaching hospitals (NTH). According to Health Care Utilization Project (HCUP), the largest collection of longitudinal hospital care data, a hospital is classified as a TH if it has one or more Accreditation Council for Graduate Medical Education (ACGME) approved residency programs or is a member of the Council of Teaching Hospitals (COTH) or has a ratio of full-time equivalent interns and residents to beds of 0.25 or higher [1]. TH have served as a pillar for advancing medicine by promoting research, educating trainees, and delivering care to the underserved population [2].
Prior studies have evaluated the effect of hospital teaching status on outcomes in various disease processes [9][10]. Based on our literature review, no studies have assessed the impact of the teaching status of the hospital on the outcomes of hospitalizations for AP. In this study, we investigated the effect of hospital teaching status on AP outcomes.

Data source
The National Inpatient Sample (NIS) database is maintained by the Agency for Healthcare Research and Quality (AHRQ). It is part of the healthcare utilization project (HCUP), which is a family of databases, software tools, and related products developed through a Federal-State-Industry partnership and sponsored by AHRQ. HCUP databases are derived from administrative data and contain encounter-level, clinical and non-clinical information, including all-listed diagnoses and procedures, discharge status, patient demographics, and charges for all patients, regardless of payer (e.g., Medicare, Medicaid, private insurance, uninsured), beginning in 1988. These databases enable research on a broad range of health policy issues, including cost and quality of health services, medical practice patterns, access to health care programs, and outcomes of treatments at the national, state, and local market levels. NIS is the largest inpatient database in the United States and contains data from 20% of all hospitalizations, representing approximately 8 million (unweighted) and 40 million (weighted) hospitalizations yearly. The NIS database contains information regarding clinical data and resource utilization in hospitalized patients while protecting patients' privacy. It includes one primary diagnosis, up to 40 secondary diagnoses, population baseline characteristics, patient comorbidities, and total charges.

Study population
We queried the NIS database from 2016 to 2019 using the International Classification of Diseases, Tenth Revision Clinical Modification (ICD-10 CM), for patients with a discharge diagnosis of AP. The ICD-10 codes used in the study are shown in Appendix 1. The hospital's teaching status was ascertained based on prespecified data by HCUP. Patients were classified according to admissions to TH or NTH. Patients who were under 18 or missing information on demographics were excluded from the analysis. This information is presented in Figure 1.

Study outcomes and variables
The categorical outcome measures were sepsis, shock, acute kidney injury (AKI), intensive care unit (ICU) admission, and inpatient mortality. We also measured the difference between continuous variables such as length of stay and total hospitalization charges. Hospital charges are defined as the amount charged by the hospital, before negotiating discounts with insurance companies.

Statistical analysis
Data for continuous variables are presented as population-weighted mean ± SE (standard error), while categorical variables are presented as a total number of patients with percentages. Univariate analysis was performed to assess differences between subjects admitted at TH and NTH. Continuous variables were compared using t-tests, and categorical variables were compared using chi-square tests. We also collected data on Charlson's Comorbidities. This is a well-validated index, which has been used in large administrative data to predict mortality and hospital resource utilization. This index has 17 comorbidities [11].
Hospital-level discharge weights provided by NIS were used to generate national estimates. Categorical variables were compared using the chi-square test, whereas an independent sample t-test was used for continuous variables. Univariate analysis was performed to study the effect of patient demographics, hospital characteristics, Charlson Comorbidities, and etiology of pancreatitis on categorical outcomes. A pvalue of 0.1 was considered a cut-off. A multivariate regression model was then built by including all variables found to be significant by univariate analysis to calculate the adjusted odds ratio. Logistic regression was used for categorical outcomes, and linear regression was used for continuous outcomes. A type I error of < 0.05 was considered statistically significant. Data analysis was performed using Stata (StataCorp. 2021. Stata Statistical Software: Release 17).

Demographics and hospital characteristics
A total of 1,689,334 patients were included in the study. Of these, 65.06% were in the TH group, while 34.94% were in the NTH group. Information on the total number of patients in the study population is presented in Table 1.   Table 2.   Table 3.

Length of Stay
The length of stay in TH was 5.82 days, while in NTH it was noted to be 4.54 days. Patients admitted to TH had a statistically significant higher length of stay compared to NTH (adjusted coefficient: 0.95, 95% CI-0.88-1.02, p<0.001).

Shock
The total incidence of shock in the study population was 3.76%. The incidence of shock in TH was 4.32% compared to 2.70% in NTH. There was a statistically significant increase in the likelihood of shock in TH (aOR:1.43, 95% CI-1.36-1.50, p<0.001).

Sepsis
The total incidence of sepsis in the study population was 4.19%. The incidence of sepsis in TH was 4.48% compared to 3.65% in NTH. A statistically significant increase was noted in the likelihood of shock at TH compared to NTH (aOR:1.21, 95% CI-1.14-1.28, p<0.001).

All-Cause Mortality
The total mortality in the study population was 1.99%. The incidence of mortality in TH was 2.25% as compared to 1.5% in NTH. A statistically significantly higher mortality was noted in TH compared to NTHs (aOR:1.37,95% CI-1.27-1.45 p<0.001). The differences in mortality and other categorical outcomes between TH and NTH groups are presented in Figure 2.

Discussion
In our retrospective study, 1.6 million patients were admitted with AP between 2016 and 2019. Of them, 65.06% were admitted to TH, while 34.94% were admitted to NTH. The study revealed that patients admitted to TH had a higher likelihood of sepsis, AKI, and shock. They also had more comorbidities compared to NTH. Similar to our study, Yiadom et al. revealed a higher median case acuity in patients admitted to TH compared to NTH [12]. THs can also serve as referral centers for a higher level of care due to the availability of subspecialists [13]. There is a possibility that some of the severe pancreatitis cases could have been transferred from NTH to TH.
Patients admitted to TH had significantly higher hospitalization charges than those admitted to NTH. This could be due to the possibility that patients admitted to TH required a higher level of care. There is also a possibility that more cholecystectomies and ERCP's were performed at TH as more patients with biliary pancreatitis received treatment at TH than NTH. A recent study by Rotundo et al. revealed that patients who required therapeutic ERCP had a higher length of stay and hospitalization charges in TH than NTH [14]. We believe that higher rates of ERCP in TH might have also led to these differences in hospitalization charges.
Rising admission rates for AP in the last few years have led to increased utilization of healthcare resources. Researchers examining the expenses incurred by patients found that while in-hospital charges were higher at TH, 30-day post-hospitalization charges were lower at TH [15]. This reduction can be credited to improved post-discharge planning and better care processes. Therefore, appropriately allocating limited resources is paramount in managing this rising demand.
We note the following limitation of this study. NIS lacks objective data such as laboratory tests, thus limiting the ability to calculate the severity of illness scores such as APACHE or BISAP scores. Secondly, information on treatment therapies such as the amount of intravenous fluids administered is not provided in NIS, which is an important confounder and can affect patient outcomes. In addition, NIS does not provide information on readmissions therefore, we can not track readmissions. As a result, it is difficult if it was an initial or recurrent episode. Due to the nature of the database, it is difficult to ascertain if the treatment was led by an expert or trainee, which can impact the outcomes. The major strength of this study is the large study population size from several hospitals across the country, which excludes selection bias based on the demographics. Our findings should be validated in a prospective cohort that captures more granular clinical data.

Additional Information Disclosures
Human subjects: All authors have confirmed that this study did not involve human participants or tissue. Animal subjects: All authors have confirmed that this study did not involve animal subjects or tissue.

Conflicts of interest:
In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.