Dealing with highly skewed hospital length of stay distributions: The use of Gamma mixture models to study delivery hospitalizations

Williford, Eva M; Haley, Valerie B.; McNutt, Louise‐Anne; Lazariu, Victoria

doi:10.1371/journal.pone.0231825

Cited by 17 publications

(13 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We note that the gamma distribution accounts for the skewness of the density function of the LOC random variable (Fig. 3), as suggested by several authors [30, 31]. The results obtained for the GLM models are presented in Tables 4 and 5.…”

Section: Methodssupporting

confidence: 55%

Identifying the Main Predictors of Length of Care in Social Care in Portugal

Lopes

Guerreiro

Esquível

et al. 2021

Port J Public Health

View full text Add to dashboard Cite

In this paper, we aim to identify the main predictors at admission and estimate patients’ length of care (LOC), within the framework of the Portuguese National Network for Long-Term Integrated Care, considering two care settings: (1) home and community-based services (HCBS) and (2) nursing home (NH) units comprising Short, Medium, or Long Stay Care. This study relied on a database of 20,984 Portuguese individuals who were admitted to the official long-term care (LTC) system and discharged during 2015. A generalised linear model (GLM) with gamma distribution was adjusted to HCBS and NH populations. Two sets of explanatory variables were used to model the random variable, LOC, namely, patient characteristics (age, gender, family/neighbour support, dependency levels at admission for locomotion, cognitive status, and activities of daily living [ADL]) and external factors (referral entity, number of beds/treatment places per 1,000 inhabitants ≥65 years of age), maturity and occupancy rate of the institution, and care setting. The features found to most influence the reduction of LOC are: male gender, having family/neighbour support, being referred by hospitals to NH (or by primary care to HCBS), and being admitted to units with a lower occupancy rate and with fewer months in operation. Regarding the dependency levels, as the number of ADL considered “dependent” increases, LOC also increases. As for the cognitive status, despite the opposite trend, it was only statistically significant for NH. Furthermore, two additional models were applied by including “death,” although this feature is not observable upon admission. By creating a model that allows for an estimate of the expected LOC for a new individual entering the Portuguese LTC system, policy-makers are able to estimate future costs and optimise resources.

show abstract

Section: Methodssupporting

confidence: 55%

Identifying the Main Predictors of Length of Care in Social Care in Portugal

Lopes

Guerreiro

Esquível

et al. 2021

Port J Public Health

View full text Add to dashboard Cite

show abstract

“…Our statistical analysis is summarized in the GitHub repository 19 . A log-linked Gamma model was used to evaluate the association between each variable and the length of the hospital stay 20 . Since repeated observations were obtained from a single patient, we used the generalized estimating equation method with an exchangeable correlation structure and robust standard error estimation.…”

Section: Methodsmentioning

confidence: 99%

Effectiveness of empirical anti-pseudomonal antibiotics for recurrent COPD exacerbation: a multicenter retrospective cohort study

Shiroshita

Miyakoshi

Tsutsumi³

et al. 2021

Sci Rep

View full text Add to dashboard Cite

Although frequent chronic obstructive pulmonary disease (COPD) exacerbation has been associated with the isolation of Pseudomonas aeruginosa (PA) in sputum cultures, it remains unknown whether the empirical use of anti-pseudomonal antibiotics can improve outcomes in patients with frequent COPD exacerbations. This multicenter retrospective cohort study evaluated whether the empirical use of anti-pseudomonal antibiotics improves the length of the hospital stay in patients with recurrent COPD exacerbation (≥ 2 admissions from April 1, 2008 to July 31, 2020). For statistical analysis, a log-linked Gamma model was used. Parameters were estimated using a generalized estimating equation model with an exchangeable correlation structure accounting for repeated observations from a single patient. Covariates included age, body mass index, home oxygen therapy use, respiratory rate, heart rate, oxygen use on admission, mental status, systemic steroid use, activities of daily living, and the number of recurrences. Hospital-specific effects were specified as fixed effects. In total, 344 patients and 965 observations of recurrent COPD exacerbations were selected. Anti-pseudomonal antibiotics were used in 173 patients (18%). The estimated change in the length of the hospital stay between anti-pseudomonal and non-anti-pseudomonal antibiotics groups was 0.039 days [95% confidence interval; − 0.083, 0.162]. Anti-pseudomonal antibiotics could not shorten the length of the hospital stay.

show abstract

“…A log-linked Gamma model was used to evaluate the association between each variable and the length of the hospital stay. 19 Since repeated observations were obtained from a single patient, we used the generalized estimating equation method with an exchangeable correlation structure and robust standard error estimation. Hospital-speci c effects were implemented as xed effects.…”

Section: Discussionmentioning

confidence: 99%

Effectiveness of empirical anti-pseudomonal antibiotics in patients with recurrent COPD exacerbation: A multicenter retrospective cohort study

Shiroshita¹,

Miyakoshi²,

Tsutsumi³

et al. 2021

Preprint

View full text Add to dashboard Cite

Background: Although frequent chronic obstructive pulmonary disease (COPD) exacerbation has been associated with the isolation of Pseudomonas aeruginosa (PA) in sputum cultures, it remains unknown whether the empirical use of anti-pseudomonal antibiotics can improve outcomes in patients with frequent COPD exacerbations. Research Question: Does empirical use of anti-pseudomonal antibiotics improve the length of the hospital stay in patients with recurrent COPD exacerbation?Study Design and Methods: This multicenter retrospective cohort study was conducted in seven hospitals in Japan. We included patients aged ≥40 years who were admitted to the hospital with COPD exacerbation more than twice during the study period (April 1, 2008 to July 31, 2020). The primary outcome was the length of the hospital stay. For statistical analysis, a log-linked Gamma model was used. Parameters were estimated using a generalized estimating equation model with an exchangeable correlation structure accounting for repeated observations from a single patient. Covariates included age, body mass index, home oxygen therapy use, respiratory rate, heart rate, oxygen use on admission, mental status, systemic steroid use, activities of daily living, and number of recurrences. Hospital-specific effects were specified as fixed effects.Results: Among 1573 patients with COPD exacerbation, 344 patients and 965 observations of recurrent COPD exacerbations were selected. Anti-pseudomonal antibiotics were used in 173 patients (18%). The estimated change in the length of the hospital stay between the anti-pseudomonal and non-anti-pseudomonal antibiotics groups was 0.044 days [95% confidence interval; -0.077, 0.166]. Interpretation: Use of empirical anti-pseudomonal antibiotics based only on recurrence may not be necessary. Further large-scale studies are needed to more precisely evaluate the effectiveness of empirical anti-pseudomonal antibiotics.

show abstract

Dealing with highly skewed hospital length of stay distributions: The use of Gamma mixture models to study delivery hospitalizations

Cited by 17 publications

References 35 publications

Identifying the Main Predictors of Length of Care in Social Care in Portugal

Identifying the Main Predictors of Length of Care in Social Care in Portugal

Effectiveness of empirical anti-pseudomonal antibiotics for recurrent COPD exacerbation: a multicenter retrospective cohort study

Effectiveness of empirical anti-pseudomonal antibiotics in patients with recurrent COPD exacerbation: A multicenter retrospective cohort study

Contact Info

Product

Resources

About