https://orcid.org/0000-0002-0548-2905
Correction
30 Apr 2025: Celdir MG, Wehby G, Prakash S, Tanaka T (2025) Correction: July effect in hospitalized cirrhosis patients: A US nationwide study using difference-in-differences analysis. PLOS ONE 20(4): e0323476. https://doi.org/10.1371/journal.pone.0323476 View correction
Figures
Abstract
Background
The July effect in US teaching hospitals has been studied with conflicting results. We aimed to evaluate the effect of physician turnover in July on the clinical outcomes of patients hospitalized with cirrhosis.
Methods
We utilized the Nationwide Inpatient Sample database (2016–2019) to identify patients hospitalized with cirrhosis and liver-related complications (variceal bleeding, hepatorenal syndrome, acute-on-chronic liver failure). We used difference-in-differences analysis to compare teaching and non-teaching hospital differences in mortality and length of stay (LOS) in May and July, and trends in outcomes in other months before and after July.
Results
We included 78,371 hospitalizations in teaching and 23,518 in non-teaching hospitals in May and July. Teaching hospital admissions had overall higher complication rates and mortality compared to non-teaching hospitals. We did not find a difference in mortality between teaching and non-teaching hospitals in all cirrhotic patients (adjusted odds ratio 1.01, 95%CI [0.88–1.16]) or in those with severe complications (0.87, [0.72–1.06]). There was greater LOS in July vs. May in teaching hospitals relative to non-teaching hospitals for all patients with cirrhosis (adjusted rate ratio 1.03, 95%CI [1.02–1.05]) and for those with severe complications (1.19, [1.17–1.21]). The months after July were associated with longer LOS in teaching hospitals, with the effect gradually diminishing over the subsequent months.
Citation: Celdir MG, Wehby G, Prakash S, Tanaka T (2025) July effect in hospitalized cirrhosis patients: A US nationwide study using difference-in-differences analysis. PLoS ONE 20(1): e0316445. https://doi.org/10.1371/journal.pone.0316445
Editor: Usama Waqar, Emory University, UNITED STATES OF AMERICA
Received: July 24, 2024; Accepted: December 11, 2024; Published: January 13, 2025
Copyright: © 2025 Celdir et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data that support the findings of this study are openly available in the Healthcare Cost and Utilization Project website (https://hcup-us.ahrq.gov/).
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
List of abbreviations: ACLF, acute-on-chronic liver failure; aRR, adjusted rate ratio; aOR, adjusted odds ratio; CirCom, Cirrhosis Comorbidity Score; HCUP, Healthcare Cost and Utilization Project; LOS, length of stay; NACSLED-ACLF, North American Consortium for the Study of End-Stage Liver Disease Acute on Chronic Liver Failure; NIS, National Inpatient Sample; OR, odds ratio; RR, rate ratio
Introduction
The anticipated departure of experienced trainee physicians and entry of a new cohort of trainees in the transition time in July has raised questions about potential disruptions in workflow due to the exit and entry of a sizeable proportion of staff within a short period and the consequent effects on productivity and patient outcomes. The influx of less experienced physicians in July may disrupt workflow, potentially leading to adverse outcomes when new residents and fellows begin training in US teaching hospitals. This potential “July effect” has been studied for surgical outcomes, in-hospital mortality, efficiency measured by length of stay (LOS) and hospital charges, and medication errors, with conflicting results.[1–3] Some studies have indicated small but significant increases in mortality during this period, especially for patients requiring more complex care, such as patients with high-risk myocardial infarction [4] and those undergoing complex cardiac surgeries [5], while other studies have not shown differences in mortality or morbidity outcomes [2, 6–8].
Cirrhosis is the eighth-leading cause of death in the US and imposes a significant healthcare burden with high and increasing rates and costs of hospitalization [9, 10]. Management of hospitalized patients with cirrhosis is complex; early recognition and treatment of complications are necessary, and hospitalizations often require involvement of multidisciplinary care teams and advanced discharge planning [11–13]. Severe in-hospital complications, including acute-on-chronic liver failure (ACLF) occur in 25% of hospitalized patients with decompensated cirrhosis and are associated with high in-hospital mortality [14, 15]. In decompensated hospitalized patients with severe complications and ACLF, identification and management of complications, and prognostication for liver transplantation or initiation of palliative care are complex and challenging. Furthermore, care pathways and protocols targeting hospitalized cirrhosis patients are not well-established [13, 16–18]. Due to trainee turnover and training and the adjustment time of new trainees starting in July, time-sensitive clinical decisions when caring for this complex patient population may be delayed at the beginning of the academic year, potentially leading to adverse clinical outcomes, including mortality, health complications, and longer LOS. However, the impact of trainee physician turnover on the clinical outcomes of hospitalized patients with cirrhosis occurring in July has not been sufficiently examined in the existing literature.
We aimed to evaluate the effect of physician turnover in July on the clinical outcomes of a representative sample of hospitalized patients with cirrhosis in the US using the National Inpatient Sample dataset, employing a quasi-experimental difference-in-differences design. We hypothesized that mortality and LOS may increase in teaching hospitals in July due to trainee physician turnover.
Material and methods
Data source and study sample
Our study used data from the National Inpatient Sample (NIS) of the Healthcare Cost and Utilization Project (HCUP) collected by the Agency for Healthcare Research and Quality (https://hcup-us.ahrq.gov/nisoverview.jsp) [19]. The NIS represents approximately 20% of all discharges from community hospitals in the US in a given year. We included years 2016 through 2019 because ICD-10 codes were first used in the fourth quartile of 2015 and because admissions and outcomes in 2020 were affected by the COVID-19 pandemic. In the NIS database, a hospital is categorized as a teaching hospital if it has ACGME-approved residency programs, is a member of the Council of Teaching Hospitals, or maintains a resident-to-bed ratio exceeding 0.25. Due to the low number of rural hospitals in the dataset and the absence of information regarding their teaching or non-teaching status (i.e., facilities are coded as urban-teaching, urban-non-teaching, or rural), we excluded rural hospitals from our analyses. We obtained information on the following additional hospital-level characteristics from the NIS database: hospital bed size (small, medium, large), nine census divisions, and ownership status (public, private not-for-profit, private investor-owned).
Identification of comorbidities and outcomes
We identified hospitalized cirrhosis patients using ICD-10 codes as listed in the S1 Table. The accuracy of these ICD-10 diagnostic codes in detecting cirrhosis admissions has been demonstrated in previous work [20]. In sub analyses, we focused on patients at higher risk of inpatient mortality who have severe complications during their hospital stay, including acute-on-chronic liver failure, hepatorenal syndrome, or variceal bleeding. Acute-on-chronic liver failure (ACLF) was identified based on the North American Consortium for the Study of End-Stage Liver Disease Acute on Chronic Liver Failure (NACSLED-ACLF) score, using two or more organ failures in patients with chronic liver failure [21]. NACSLED-ACLF defined organ failures are respiratory failure (need for mechanical ventilation), renal failure (acute kidney injury or need for dialysis), cardiovascular failure (shock, central venous pressure procedure, arterial line procedure), and neurological failure (brain death or hepatic coma) [22]. The ICD-10 codes that we used to identify the presence of ACLF, hepatorenal syndrome, and variceal bleeding are provided in S1 Table. We also identified relevant diagnoses to calculate Cirrhosis Comorbidity Score (CirCom) [23]. CirCom is a validated comorbidity scoring system for patients with cirrhosis developed for observational studies based on the following comorbidities: chronic obstructive pulmonary disease, acute myocardial infarction, peripheral arterial disease, epilepsy, substance abuse, heart failure, cancer, and chronic kidney disease [23–25].
Statistical analyses
We first provided a description of the patients, encompassing their demographic and clinical characteristics. Continuous variables were compared using the Student’s t test, while categorical variables were analyzed using the chi-square test.
We then estimated a difference-in-differences model that compared the change in in-hospital mortality and LOS of hospitalized patients with cirrhosis during May vs. July and between teaching and non-teaching hospitals. Specifically, the difference in outcomes for May vs. July in teaching hospitals was compared with the difference in outcomes for the month of May and the month of July in non-teaching hospitals. This model assumes that any seasonal variation in outcomes between these two months that is not related to resident turnover (in teaching hospitals) is otherwise similar between teaching and non-teaching hospitals. We used May as the pre-transition month rather than June because some residency programs start training within June, or residents may leave early at the end of their training. We estimate the difference-in-differences regression using logistic regression models for mortality. Because LOS was not normally distributed and positively skewed, we used Poisson regression where LOS was seen as a count of number of days that the patient remained in the hospital [26, 27]. The models included an interaction term, (July)×(Teaching), where July is a binary indicator for outcomes in July relative to May, and Teaching is a binary indicator for hospital teaching versus non-teaching status, as well as these indicators as covariates on their own. We first estimated a basic difference-in-differences specification without additional covariates. We also estimated a specification that adjusts for hospital ownership, census division, bed size, and patient’s sex, age, race, and Cirrhosis Comorbidity Score (CirCom). We estimated the models for the total sample of patients with cirrhosis and separately for those with severe complications as defined above. We report the estimates as odds ratios (OR) for in-hospital mortality and incidence rate ratios (RR) for length of stay, which are the exponentiated coefficients of the difference-in-differences interaction term. The RR for length of stay specifically compares the counted days of hospitalization between groups. For example, an RR of 2 in one group, relative to another, implies that the counted days of hospitalization (i.e., LOS) are twice as long in the former group.
To better understand seasonal variation in the months leading to July between teaching and non-teaching hospitals, as well as variation in outcomes post-July, we estimated an extended version of the difference-in-differences model that includes all other calendar months. This second model included separate 0/1 indicators for each month (except May, which was the reference month) and interactions with teaching status similar to the first model [28]. The exponentiated coefficients of the difference-in-differences interaction terms were plotted for each month (with May as the reference month). The interaction term in the form of (Month)*(Teaching) with an indicator variable for each month and hospital teaching status from the fitted adjusted logistic regression models for mortality and Poisson models for length of stay is presented in the Figs 1–4. The exponentiated coefficients were plotted according to the month. Error bars indicate 95% confidence intervals in the figures.
We used R version 4.3.0 for statistical analyses and considered p values <0.05 statistically significant.
Ethical statements
This cross-sectional nationwide study conforms to the ethical guidelines 1975 Declaration of Helsinki and its later amendments or comparable ethical standards. The Human Investigation Committee (IRB) of the University of Iowa was not required to approve or review this study as the National Inpatient Sample provides publicly available and de-identified data.
Results
Sample description
The analytical sample included 78,371 hospitalizations in teaching hospitals in May (N = 39283) and July (N = 39088) and 23,518 in non-teaching hospitals in May (N = 11837) and July (N = 11681). Patient characteristics are summarized in Table 1. Patients were slightly younger, on average, in teaching hospitals compared with non-teaching hospitals (58.4 vs. 59.2 years in teaching vs. non-teaching, respectively, p < 0.01). The proportion of women was also slightly lower in teaching hospitals (38.2% in teaching vs. 39.2% in non-teaching, p<0.01). Inpatient mortality for cirrhosis admission combining the two months was 5.7% in teaching vs. 5.1% in non-teaching hospitals. The distribution of CirCom scores was similar between teaching and non-teaching hospitals. Rates of inpatient cirrhosis complications were higher in teaching hospitals.
Difference-in-differences estimates, May versus July
Table 2 summarizes the results of difference-in-differences estimates for models for mortality and LOS mentioned in this section. Unadjusted inpatient mortality in the total sample was not different in teaching hospitals between May and July (4.9% versus 4.7%) or in non–teaching hospitals between May and July (5.6% versus 5.5%). The unadjusted difference-in-differences estimate comparing July and May mortality differences between teaching and non-teaching hospitals was not statistically significant (OR = 1.01 95% CI 0.89–1.16). Adjusting for hospital and patient characteristics resulted in a null difference-in-differences estimate (OR 1.01, 95% CI 0.89–1.15). Similarly, among patients with severe complications, the difference-in-differences comparisons showed no significant differences (OR = 0.87; 95% CI of 0.72–1.06 in the adjusted model).
LOS (i.e., the counted days of hospitalization) in non-teaching hospitals was 3% shorter in July than in May (adjusted rate ratio 0.97, 95% CI 0.97–0.98). In contrast, it was similar in teaching hospitals between May and July (adjusted rate ratio 0.99, 95% CI 0.98–1.00). The difference-in-differences analysis of LOS revealed a more extended hospitalization in July versus May in teaching hospitals compared to non-teaching hospitals (unadjusted and adjusted rate ratio 1.03, 95% CI 1.02–1.05), interpreted as 3% increase in LOS in teaching hospitals attributed to the effect of July, compared to non-teaching hospitals. Similarly, among patients with severe complications, non-teaching hospitals had shorter LOS in July than in May (adjusted rate ratio 0.93, 95% CI 0.91–0.96), while teaching hospitals had similar LOS between July and May (adjusted rate ratio 1.01, 95% CI 0.99–1.02). The difference-in-differences estimates indicated a longer hospitalization in July versus May in teaching hospitals (adjusted rate ratio 1.19, 95% CI 1.17–1.21), interpreted as 19% increase in LOS in teaching hospitals attributed to the effect of July.
Difference-in-differences estimates, month-by-month
In both teaching and non-teaching hospitals, mortality in January, February, November, and December was higher than in the rest of the calendar year. Teaching hospitals had higher unadjusted mortality rates than non-teaching hospitals throughout the year (S2 Table). Median LOS was 4 days in the overall patient population (IQR: 2–7, S3 Table). Consistent with the analyses of patients hospitalized only in May and July, there was no significant trend of mortality differences based on teaching status and before or after July. However, July and the following months were associated with longer LOS in teaching hospitals, and this difference gradually tapered over the subsequent months. The month-by-month (relative to May) difference-in-differences estimates for mortality and LOS are plotted in Figs 1–4.
Discussion
In a national inpatient sample of patients hospitalized with cirrhosis, we found no evidence of an effect of physician turnover in July on mortality in teaching hospitals using a difference-in-differences analysis comparing teaching and non-teaching hospitals in May vs. July. We also found no evidence of an effect when focusing on patients with severe complications and no differences between teaching and non-teaching hospitals in seasonal trends in mortality in other months leading to or following July. Our findings revealed a modest yet statistically significant increase in LOS at teaching hospitals in July, compared to non-teaching hospitals, determined using a difference-in-differences design. Moreover, this effect was more pronounced in patients with severe complications, where there was a 19% increase in LOS attributed to the ’July effect’ in teaching hospitals as opposed to non-teaching hospitals.
To our knowledge, this study is the first to assess the effect of trainee physician turnover in July in teaching hospitals on the health outcomes of hospitalized patients with cirrhosis. Contrary to our results, which revealed no observable impact of July on mortality rates among hospitalized patients with cirrhosis, a systematic review concluded that there is an association between July turnover and mortality, with an effect size ranging from an overall relative risk increase of 4.3 to 12%. However, there is significant heterogeneity in the existing literature [2]. The elevation in mortality risk was more consistently observed in studies on patient groups with elevated risk, such as complex cardiac surgeries [5]. Among medical subspecialties, studies were performed primarily for cardiac outcomes and did not show mortality differences [4, 8, 29]; except for a small but significant increase for higher-risk cardiac patients [4].
LOS is commonly used as a proxy for efficient clinical management and resource utilization in healthcare settings [30]. Our results are consistent with several prior studies that showed increased LOS during the trainee turnover period [1, 5, 31, 32]. However, most of these studies did not consider seasonal variation using a difference-in-differences framework. Noteworthy of our findings include the difference-in-differences estimate suggesting an increase in LOS in teaching hospital. However, this increase might be arguably attributable to a small decline in LOS in non-teaching hospitals between May and July, while LOS remained steady in teaching hospitals. Therefore, this finding on LOS should be cautiously interpreted. The difference-in-differences model and this finding hinge on the assumption that the May-July difference would have been similar between teaching and non-teaching hospitals without the trainee turnover, which cannot be directly tested. Therefore, we have also performed a sensitivity analysis on trends of LOS before and after July. The month-by-month analysis leading to July shows no evidence of pre-July trends that would explain the estimated increase in July but there was a consistent trend of increased LOS in month-by-month RRs in the months following July in teaching hospitals compared to non-teaching hospitals. Additionally, considering the unclear reasons for non-teaching hospitals having shorter LOS in July compared to May, it seems reasonable to hypothesize that common confounding factors, which potentially reduce the LOS in July, offset the “July-effect” due to trainee turnover in teaching hospitals, which could have led to a bias towards a null effect. The lower level of familiarity of new trainees with the teams, care coordination of cirrhosis patients with complex discharge planning needs [13, 16, 17], and delays in initiating palliative care [16] may have resulted in longer hospital stays in July and the following months. Establishing care pathway protocols may mitigate these factors, as it has been shown to improve outcomes in chronic diseases such as heart failure [33].
Our study also offers new evidence on seasonal variation in two important health outcomes for patients with cirrhosis. Specifically, our data suggest increased mortality and LOS during winter months, which is consistent with studies demonstrating seasonal variations in mortality and hospitalization outcomes in general medicine and other specialties [34, 35].
Our study has two notable strengths: the inclusion of data from a national inpatient sample and the use of difference-in-differences analysis with non-teaching hospitals as a control group to account for seasonal variation. Nevertheless, our study has several limitations. First, although we have adjusted for multiple potential patient and hospital level confounders, there may still be confounding between teaching and non-teaching hospitals over time that bias the difference-in-differences estimates. Second, there are potential misclassification errors. For example, we used the Cirrhosis Comorbidity Score to account for patients’ pre-existing morbidity; however, this scoring has been only validated in the outpatient setting and may not apply to our inpatient sample [23, 25, 36]. Also, we have relied on administrative diagnostic codes, and there may be differences in coding practices between teaching and non-teaching hospitals. Inherent to administrative databases, NIS ICD-10 codes may not always be accurate for cirrhosis and related complications [20]. Additionally, the NIS dataset does not provide laboratory or medication data. Third, severe complications defined in our dataset did not differentiate the cirrhosis complications present at admission from those developed during hospitalization. Due to the limitations of registry-based data, we did not have laboratory and further clinical data to stratify disease severity at the time of admission or further adjust for severity in the models. Fourth, each year, the NIS sample may include different hospitals, which could differentially alter the sample composition over time between teaching and non-teaching status. We adjusted for several patient and hospital-level covariates, which accounts for any sample composition changes in these specific characteristics and found similar results overall. Finally, our NIS dataset entries are based on admissions, not individuals; we were not able to consider multiple admissions for a patient and the difference in readmissions between teaching and non-teaching hospitals in our analyses.
As a potential future direction, examining additional outcomes, including readmission rates and post-discharge mortality, would be valuable. However, these outcomes cannot be studied using this specific dataset. The readmission rate for patients with cirrhosis-related complications has been reported to be approximately 25% [37]. Previous studies report that non-teaching hospitals have higher readmission rates, [37] and premature discharge may be associated with higher readmissions [38]. Furthermore, the trend of increasing 30-day post-discharge mortality despite a decrease in inpatient mortality in recent years [10], signifies a shift of mortality risks to after discharge. This underscores the importance of careful discharge planning to balance efforts to reduce LOS with the prevention of premature discharge [39].
Conclusion
We did not find any differences in in-hospital mortality in July in teaching hospitals for patients with cirrhosis, but there was an increase in LOS in July in teaching relative to non-teaching hospitals, especially in cirrhosis patients with severe complications. The findings support the need for appropriate supervision of new trainees, developing quality measures to minimize operational disruptions, and establishing care protocols for hospitalized cirrhosis patients, which may improve clinical outcomes in this high-risk patient population.
Supporting information
S1 Table. ICD-10 and CPT codes to identify the study population and severe cirrhosis complications.
https://doi.org/10.1371/journal.pone.0316445.s001
(DOCX)
S2 Table. Inpatient mortality among patients admitted to teaching and non-teaching hospitals, according to month.
https://doi.org/10.1371/journal.pone.0316445.s002
(DOCX)
S3 Table. LOS among patients admitted to teaching and non-teaching hospitals, according to month.
https://doi.org/10.1371/journal.pone.0316445.s003
(DOCX)
Acknowledgments
Presentation. A portion of the material presented in this article was also featured in a presentation at the AASLD Liver Meeting in November 2023.
References
- 1. Huckman R, Barro J. Cohort Turnover and Productivity: The July Phenomenon in Teaching Hospitals. National Bureau of Economic Research. 2005.
- 2. Young JQ, Ranji SR, Wachter RM, Lee CM, Niehaus B, Auerbach AD. "July effect": impact of the academic year-end changeover on patient outcomes: a systematic review. Ann Intern Med. 2011;155(5):309–15. Epub 2011/07/13. pmid:21747093.
- 3. Phillips DP, Barker GE. A July spike in fatal medication errors: a possible effect of new medical residents. J Gen Intern Med. 2010;25(8):774–9. Epub 2010/06/01. pmid:20512532; PubMed Central PMCID: PMC2896592.
- 4. Jena AB, Sun EC, Romley JA. Mortality Among High-Risk Patients With Acute Myocardial Infarction Admitted to US Teaching-Intensive Hospitals in July. Circulation. 2013;128(25):2754–63. pmid:24152859
- 5. Shuhaiber JH, Goldsmith K, Nashef SAM. Impact of Cardiothoracic Resident Turnover on Mortality After Cardiac Surgery: A Dynamic Human Factor. The Annals of Thoracic Surgery. 2008;86(1):123–31. pmid:18573410
- 6. “July Effect”: Impact of the Academic Year-End Changeover on Patient Outcomes. Annals of Internal Medicine. 2011;155(5):309–15. pmid:21747093.
- 7. Myers L, Mikhael B, Currier P, Berg K, Jena A, Donnino M, et al. The association between physician turnover (the “July Effect”) and survival after in-hospital cardiac arrest. Resuscitation. 2017;114:133–40. pmid:28285032
- 8. Patlolla SH, Truesdell AG, Basir MB, Rab ST, Singh M, Belford PM, et al. No “July Effect” in the management and outcomes of acute myocardial infarction: An 18‐year United States national study. Catheterization and Cardiovascular Interventions. 2023;101(2):264–73. pmid:36617382
- 9. Zou B, Yeo YH, Jeong D, Park H, Sheen E, Lee DH, et al. A Nationwide Study of Inpatient Admissions, Mortality, and Costs for Patients with Cirrhosis from 2005 to 2015 in the USA. Dig Dis Sci. 2020;65(5):1520–8. Epub 2019/10/11. pmid:31598919.
- 10. Kim D, Cholankeril G, Li AA, Kim W, Tighe SP, Hameed B, et al. Trends in hospitalizations for chronic liver disease‐related liver failure in the United States, 2005‐2014. Liver International. 2019;39(9):1661–71. pmid:31081997
- 11. Perumalswami PV, Schiano TD. The Management of Hospitalized Patients with Cirrhosis: The Mount Sinai Experience and a Guide for Hospitalists. Digestive Diseases and Sciences. 2011;56(5):1266–81. pmid:21416246
- 12. Muir AJ. Understanding the Complexities of Cirrhosis. Clinical Therapeutics. 2015;37(8):1822–36. pmid:26188836
- 13. Mellinger JL, Volk ML. Multidisciplinary Management of Patients With Cirrhosis: A Need for Care Coordination. Clinical Gastroenterology and Hepatology. 2013;11(3):217–23. pmid:23142204
- 14. Hernaez R, Kramer JR, Liu Y, Tansel A, Natarajan Y, Hussain KB, et al. Prevalence and short-term mortality of acute-on-chronic liver failure: A national cohort study from the USA. Journal of Hepatology. 2019;70(4):639–47. pmid:30590100
- 15. Das V, Boelle P-Y, Galbois A, Guidet B, Maury E, Carbonell N, et al. Cirrhotic patients in the medical intensive care unit: Early prognosis and long-term survival*. Critical Care Medicine. 2010;38(11). pmid:20802324
- 16. Hernaez R, Patel A, Jackson LK, Braun UK, Walling AM, Rosen HR. Considerations for Prognosis, Goals of Care, and Specialty Palliative Care for Hospitalized Patients With Acute‐on‐Chronic Liver Failure. Hepatology. 2020;72(3). pmid:32416642
- 17. Nadim MK, Durand F, Kellum JA, Levitsky J, O’Leary JG, Karvellas CJ, et al. Management of the critically ill patient with cirrhosis: A multidisciplinary perspective. Journal of Hepatology. 2016;64(3):717–35. pmid:26519602
- 18. Kanwal F. Decreasing Mortality in Patients Hospitalized With Cirrhosis. Gastroenterology. 2015;148(5):897–900. pmid:25805421
- 19. Geissler EK, Schnitzbauer AA, Zulke C, Lamby PE, Proneth A, Duvoux C, et al. Sirolimus Use in Liver Transplant Recipients With Hepatocellular Carcinoma: A Randomized, Multicenter, Open-Label Phase 3 Trial. Transplantation. 2016;100(1):116–25. Epub 2015/11/12. pmid:26555945; PubMed Central PMCID: PMC4683033.
- 20. Hayward KL, Johnson AL, McKillen BJ, Burke NT, Bansal V, Horsfall LU, et al. ICD-10-AM codes for cirrhosis and related complications: key performance considerations for population and healthcare studies. BMJ Open Gastroenterol. 2020;7(1). Epub 2020/09/19. pmid:32943463; PubMed Central PMCID: PMC7500192.
- 21. Rosenblatt R, Shen N, Tafesh Z, Cohen‐Mekelburg S, Crawford CV, Kumar S, et al. The North American Consortium for the Study of End‐Stage Liver Disease–Acute‐on‐Chronic Liver Failure Score Accurately Predicts Survival: An External Validation Using a National Cohort. Liver Transplantation. 2020;26(2):187–95. pmid:31785079
- 22. O’Leary JG, Reddy KR, Garcia‐Tsao G, Biggins SW, Wong F, Fallon MB, et al. NACSELD acute‐on‐chronic liver failure (NACSELD‐ACLF) score predicts 30‐day survival in hospitalized patients with cirrhosis. Hepatology. 2018;67(6):2367–74. pmid:29315693
- 23. Jepsen P, Vilstrup H, Lash TL. Development and Validation of a Comorbidity Scoring System for Patients With Cirrhosis. Gastroenterology. 2014;146(1):147–56. pmid:24055278
- 24. Jepsen P. Comorbidity in cirrhosis. World Journal of Gastroenterology. 2014;20(23):7223. pmid:24966593
- 25. Crooks C, West J, Jepsen P. A validation study of the CirCom comorbidity score in an English cirrhosis population using the Clinical Practice Research Datalink. Clinical Epidemiology. 2018;Volume 10:107–20. pmid:29391833
- 26. Austin PC, Rothwell DM, Tu JV. A Comparison of Statistical Modeling Strategies for Analyzing Length of Stay after CABG Surgery. Health Services and Outcomes Research Methodology. 2002;3(2):107–33.
- 27. Xiao J, Lee AH, Vemuri SR. Mixture distribution analysis of length of hospital stay for efficient funding. Socio-Economic Planning Sciences. 1999;33(1):39–59. https://doi.org/10.1016/S0038-0121(98)00006-8.
- 28. Fernandez GA, Vatcheva KP. A comparison of statistical methods for modeling count data with an application to hospital length of stay. BMC Medical Research Methodology. 2022;22(1). pmid:35927612
- 29. Garcia S, Canoniero M, Young L. The effect of July admission in the process of care of patients with acute cardiovascular conditions. South Med J. 2009;102(6):602–7. Epub 2009/05/13. pmid:19434039.
- 30.
OECD Length of Stay 2023 [cited 2023 10/30/2023]. [OECD Length of Stay Data]. Available from: https://data.oecd.org/healthcare/length-of-hospital-stay.htm#indicator-chart
- 31. Rich EC. The Relationship of House Staff Experience to the Cost and Quality of Inpatient Care. JAMA: The Journal of the American Medical Association. 1990;263(7):953. pmid:2299762
- 32. Jen MH, Bottle A, Majeed A, Bell D, Aylin P. Early In-Hospital Mortality following Trainee Doctors’ First Day at Work. PLoS ONE. 2009;4(9):e7103. pmid:19774078
- 33. Kittleson MM, Panjrath GS, Amancherla K, Davis LL, Deswal A, Dixon DL, et al. 2023 ACC Expert Consensus Decision Pathway on Management of Heart Failure With Preserved Ejection Fraction. Journal of the American College of Cardiology. 2023;81(18):1835–78. pmid:37137593
- 34. Gemmell I, McLoone P, Boddy F, Dickinson GJ, Watt G. Seasonal variation in mortality in Scotland. International Journal of Epidemiology. 2000;29(2):274–9. pmid:10817125
- 35. Enquselassie F, Dobson AJ, Alexander HM, Steele PL. Seasons, temperature and coronary disease. Int J Epidemiol. 1993;22(4):632–6. Epub 1993/08/01. pmid:8225736.
- 36. Younossi ZM, Henry L, Stepanova M. A New Comorbidity Model for Predicting Mortality in Patients With Cirrhosis: Does It Work? Gastroenterology. 2014;146(1):19–24. pmid:24287302
- 37. Shaheen AA, Nguyen HH, Congly SE, Kaplan GG, Swain MG. Nationwide estimates and risk factors of hospital readmission in patients with cirrhosis in the United States. Liver International. 2019;39(5):878–84. pmid:30688401
- 38. Graham KL, Auerbach AD, Schnipper JL, Flanders SA, Kim CS, Robinson EJ, et al. Preventability of Early Versus Late Hospital Readmissions in a National Cohort of General Medicine Patients. Annals of Internal Medicine. 2018;168(11):766–74. pmid:29710243
- 39. Volk ML, Tocco RS, Bazick J, Rakoski MO, Lok AS. Hospital Readmissions Among Patients With Decompensated Cirrhosis. American Journal of Gastroenterology. 2012;107(2):247–52. pmid:21931378