The Predictive Factors on Extended Hospital Length of Stay in Patients with AMI: Laboratory and Administrative Data

Abstract

The length of hospital stay (LOS) is an important measure of efficiency in the use of hospital resources. Acute Myocardial Infarction (AMI), as one of the diseases with higher mortality and LOS variability in the OECD countries, has been studied with predominant use of administrative data, particularly on mortality risk adjustment, failing investigation in the resource planning and specifically in LOS. This paper presents results of a predictive model for extended LOS (LOSE - above 75th percentile of LOS) using both administrative and clinical data, namely laboratory data, in order to develop a decision support system. Laboratory and administrative data of a Portuguese hospital were included, using logistic regression to develop this predictive model. A model with three laboratory data and seven administrative data variables (six comorbidities and age ≥ 69 years), with excellent discriminative ability and a good calibration, was obtained. The model validation shows also good results. Comorbidities were relevant predictors, mainly diabetes with complications, showing the highest odds of LOSE (OR = 37,83; p = 0,001). AMI patients with comorbidities (diabetes with complications, cerebrovascular disease, shock, respiratory infections, pulmonary oedema), with pO2 above level, aged 69 years or older, with cardiac dysrhythmia, neutrophils above level, pO2 below level, and prothrombin time above level, showed increased risk of extended LOS. Our findings are consistent with studies that refer these variables as predictors of increased risk.

Hospital volume and outcomes for inpatients with acute myocardial infarction in Shanxi, China: A cross‐sectional study

“…LOS for inpatients with more serious conditions was longer than other patients, and numerous studies have drawn similar conclusions. 29 These results differ from those presented in previous studies, which reported that more resources in high-volume hospitals might result in lower costs. Our study found that a lower volume was related to lower costs.…”

“…One possible explanation for the difference in LOS for patients between medium‐ and high‐volume groups is that patients with more serious conditions are more likely to be treated in higher volume hospitals with a higher profile. LOS for inpatients with more serious conditions was longer than other patients, and numerous studies have drawn similar conclusions . These results differ from those presented in previous studies, which reported that more resources in high‐volume hospitals might result in lower costs.…”

Decision-making support systems on extended hospital length of stay: Validation and recalibration of a model for patients with AMI

“…A Recall (58%) and Specificity (73%), with a cut-off of 0.26, and a discriminatory capacity (area under the ROC curve [AUC] of 0.702) were observed for the population of patients with a primary diagnosis of AMI, who were discharged alive and for the variables defined in the previous model (VM). According to the data presented in Table 3, there is a decrease in the predictive capacity of the VM compared to the results obtained in the model by Magalhães et al (13). The VM considered LOSE ≥ 7 days.…”

Thirty‐day readmission, length of stay and self‐management behaviour among patients with acute coronary syndrome and type 2 diabetes mellitus: A scoping review

“…Seventy-four articles were retained for full-text reviews and additional 54 articles were excluded as they were (a) not A summary of the information about study characteristics and key findings is presented in Table 1. Most of the studies were performed in the USA (Deedwania et al, 2017;Dunlay et al, 2012;Oanh Kieu et al, 2018;Tisminetzky et al, 2014;Tisminetzky et al, 2015) and Australia (Wu, Chang, Courtney, & Kostner，2012b;Wu, Chang, & McDowell, 2009;Wu & Chang, 2008;Wu, Chang, Courtney, & Ramis, 2012a;Wu et al, 2017) (five studies in each); three studies were performed in Taiwan (Carral et al, 2003;Chiang et al, 2013;Hung et al, 2013); and one study each, was conducted in Canada (Southern et al, 2014), Italy (Valent, Tonutti, & Grimaldi, 2017), Spain (Rodriguez-Padial et al, 2017), Portugal (Magalhães, Lopes, Gomes, & Seixo, 2016), India (Kamalesh, Subramanian, Ariana, Sawada, & Tierney, 2005), the UK (Loudon et al, 2016), and France (Colin, Lafuma, & Gueron, 2007). The majority of articles used a FIGURE 1 PRISMA flow diagram of the study selection process…”

“…The remaining four articles examined the predictors of a longer LOS. Magalhães et al (2016) developed a model to predict extended LOS, which was defined as above the 75th percentile of LOS, and the results indicated that diabetes with complications were robust predictors, that showed the highest odds of extended LOS. In addition, Loudon et al (2016) and Valent et al (2017) found that diabetes is an independent predictor of increased LOS.…”