A Multistate Model Predicting Mortality, Length of Stay, and Readmission for Surgical Patients.
Ontology highlight
ABSTRACT:
SUBMITTER: Clark DE
PROVIDER: S-EPMC4874834 | biostudies-other | 2016 Jun
REPOSITORIES: biostudies-other
Publications
A Multistate Model Predicting Mortality, Length of Stay, and Readmission for Surgical Patients.
Clark David E DE Ostrander Kaitlin R KR Cushing Brad M BM
Health services research 20151020 3
<h4>Objective</h4>Simultaneously evaluate postoperative mortality, length of stay (LOS), and readmission.<h4>Data source</h4>National Surgical Quality Improvement Program (NSQIP).<h4>Design</h4>Retrospective cohort.<h4>Methods</h4>Data from elective general surgical patients were obtained from the 2012 NSQIP Participant Use File. For each postoperative day, each patient's state was classified as index hospitalization, discharged home, discharged to long-term care (LTC), readmitted, or dead. Tran ...[more]
Similar Datasets
Project description:OBJECTIVE:Due to the defects in skin barrier function and immune response, burn patients who survive the acute phase of a burn injury are at a high risk of nosocomial infection (NI). The aim of this study is to evaluate the impacts of NI on length of stay (LOS) and hospital mortality in burn patients using a multistate model. DESIGN AND SETTING:A retrospective observational study was conducted in burn unit and intensive care unit in the First Affiliated Hospital of Wenzhou Medical University, Wenzhou, China. PARTICIPANTS:Data were obtained from 1143 records of patients admitted with burn between 1 January 2013 and 31 December 2016. METHODS:Risk factors for NIs were determined by binary logistic regression. The extended Cox model with time-varying covariates was used to determine the impact of NIs on hospital mortality, and cumulative incidence functions were calculated. Multiple linear regression analysis was applied to detect the variables associated with LOS. Using a multistate model, the extra LOS due to NI were determined. RESULTS:15.8% of total burn patients suffered from NIs and incidence density of NIs was 9.6 per 1000 patient-days. NIs significantly increased the rate of death (HR 4.266, 95% CI 2.218 to 8.208, p=0.000). The cumulative probability of death for patients with NI was greater that for those without NI. The extra LOS due to NIs was 17.68 days (95% CI 11.31 to 24.05). CONCLUSIONS:Using appropriate statistical methods, the present study further illustrated that NIs were associated with the increased cumulative incidence of burn death and increased LOS in burn patients.
| S-EPMC6278804 | biostudies-other
Project description:BackgroundEvidence on the association between newborn length of hospital stay (LOS) and risk of readmission is conflicting. We compared methods for modelling this relationship, by gestational age, using population-level hospital data on births in England between 2005-14.MethodsThe association between LOS and unplanned readmission within 30 days of postnatal discharge was explored using four approaches: (i) modelling hospital-level LOS and readmission rates; (ii) comparing trends over time in LOS and readmission; (iii) modelling individual LOS and adjusted risk of readmission; and (iv) instrumental variable analyses (hospital-level mean LOS and number of births on the same day).ResultsOf 4 667 827 babies, 5.2% were readmitted within 30 days. Aggregated data showed hospitals with longer mean LOS were not associated with lower readmission rates for vaginal (adjusted risk ratio (aRR) 0.87, 95% confidence interval (CI) 0.66, 1.13), or caesarean (aRR 0.89, 95% CI 0.72, 1.12) births. LOS fell by an average 2.0% per year for vaginal births and 3.4% for caesarean births, while readmission rates increased by 4.4 and 5.1% per year respectively. Approaches (iii) and (iv) indicated that longer LOS was associated with a reduced risk of readmission, but only for late preterm, vaginal births (34-36 completed weeks' gestation).ConclusionsLonger newborn LOS may benefit late preterm babies, possibly due to increased medical or psychosocial support for those at greater risk of potentially preventable readmissions after birth. Research based on observational data to evaluate relationships between LOS and readmission should use methods to reduce the impact of unmeasured confounding.
| S-EPMC5518288 | biostudies-literature
Project description:PurposeOur objective is to identify the predictive factors and predict hospital length of stay (LOS) in dengue patients, for efficient utilization of hospital resources.MethodsWe collected 1360 medical patient records of confirmed dengue infection from 2012 to 2017 at Max group of hospitals in India. We applied two different data mining algorithms, logistic regression (LR) with elastic-net, and random forest to extract predictive factors and predict the LOS. We used an area under the curve (AUC), sensitivity, and specificity to evaluate the performance of the classifiers.ResultsThe classifiers performed well, with logistic regression (LR) with elastic-net providing an AUC score of 0.75 and random forest providing a score of 0.72. Out of 1148 patients, 364 (32%) patients had prolonged length of stay (LOS) (> 5 days) and overall hospitalization mean was 4.03 ± 2.44 days (median ± IQR). The highest number of dengue cases belonged to the age group of 10-20 years (21.1%) with a male predominance. Moreover, the study showed that blood transfusion, emergency admission, assisted ventilation, low haemoglobin, high total leucocyte count (TLC), low or high haematocrit, and low lymphocytes have a significant correlation with prolonged LOS.ConclusionOur findings demonstrated that the logistic regression with elastic-net was the best fit with an AUC of 0.75 and there is a significant association between LOS greater than five days and identified patient-specific variables. This method can identify the patients at highest risks and help focus time and resources.
| S-EPMC8363490 | biostudies-literature
Project description:ObjectiveTo develop a predictive model for real-time predictions of length of stay, mortality, and readmission for hospitalized patients using electronic health records (EHRs).Materials and methodsA Bayesian Network model was built to estimate the probability of a hospitalized patient being "at home," in the hospital, or dead for each of the next 7 days. The network utilizes patient-specific administrative and laboratory data and is updated each time a new pathology test result becomes available. Electronic health records from 32 634 patients admitted to a Sydney metropolitan hospital via the emergency department from July 2008 through December 2011 were used. The model was tested on 2011 data and trained on the data of earlier years.ResultsThe model achieved an average daily accuracy of 80% and area under the receiving operating characteristic curve (AUROC) of 0.82. The model's predictive ability was highest within 24 hours from prediction (AUROC = 0.83) and decreased slightly with time. Death was the most predictable outcome with a daily average accuracy of 93% and AUROC of 0.84.DiscussionWe developed the first non-disease-specific model that simultaneously predicts remaining days of hospitalization, death, and readmission as part of the same outcome. By providing a future daily probability for each outcome class, we enable the visualization of future patient trajectories. Among these, it is possible to identify trajectories indicating expected discharge, expected continuing hospitalization, expected death, and possible readmission.ConclusionsBayesian Networks can model EHRs to provide real-time forecasts for patient outcomes, which provide richer information than traditional independent point predictions of length of stay, death, or readmission, and can thus better support decision making.
| S-EPMC7839923 | biostudies-literature
Project description:Aims/hypothesisHospital admissions for hypoglycaemia represent a significant burden on individuals with diabetes and have a substantial economic impact on healthcare systems. To date, no prognostic models have been developed to predict outcomes following admission for hypoglycaemia. We aimed to develop and validate prediction models to estimate risk of inpatient death, 24 h discharge and one month readmission in people admitted to hospital for hypoglycaemia.MethodsWe used the Hospital Episode Statistics database, which includes data on all hospital admission to National Health Service hospital trusts in England, to extract admissions for hypoglycaemia between 2010 and 2014. We developed, internally and temporally validated, and compared two prognostic risk models for each outcome. The first model included age, sex, ethnicity, region, social deprivation and Charlson score ('base' model). In the second model, we added to the 'base' model the 20 most common medical conditions and applied a stepwise backward selection of variables ('disease' model). We used C-index and calibration plots to assess model performance and developed a calculator to estimate probabilities of outcomes according to individual characteristics.ResultsIn derivation samples, 296 out of 11,136 admissions resulted in inpatient death, 1789/33,825 in one month readmission and 8396/33,803 in 24 h discharge. Corresponding values for validation samples were: 296/10,976, 1207/22,112 and 5363/22,107. The two models had similar discrimination. In derivation samples, C-indices for the base and disease models, respectively, were: 0.77 (95% CI 0.75, 0.80) and 0.78 (0.75, 0.80) for death, 0.57 (0.56, 0.59) and 0.57 (0.56, 0.58) for one month readmission, and 0.68 (0.67, 0.69) and 0.69 (0.68, 0.69) for 24 h discharge. Corresponding values in validation samples were: 0.74 (0.71, 0.76) and 0.74 (0.72, 0.77), 0.55 (0.54, 0.57) and 0.55 (0.53, 0.56), and 0.66 (0.65, 0.67) and 0.67 (0.66, 0.68). In both derivation and validation samples, calibration plots showed good agreement for the three outcomes. We developed a calculator of probabilities for inpatient death and 24 h discharge given the low performance of one month readmission models.Conclusions/interpretationThis simple and pragmatic tool to predict in-hospital death and 24 h discharge has the potential to reduce mortality and improve discharge in people admitted for hypoglycaemia.
| S-EPMC5423930 | biostudies-literature
Project description:ImportanceIn 2012, hospital reimbursement in Switzerland changed from a fee-for-service per diem system to a diagnosis-related group (SwissDRG) system. Whether this change in reimbursement is associated with harmful implications for quality of care and patient outcomes remains unclear.ObjectiveTo examine the association of the SwissDRG implementation with length of hospital stay (LOS), in-hospital mortality, and 30-day readmission rates in the overall adult inpatient population and stratified by 5 individual diagnoses.Design, setting, and participantsThis cohort study used administrative data from the Swiss Federal Statistical Office to investigate medical hospitalizations in Switzerland from January 1, 2009, through December 31, 2015. All hospitalizations for adult medical inpatients were included in the main analysis. Patients who presented with 1 of the 5 common medical diagnoses were included in the subanalyses: community-acquired pneumonia, exacerbation of chronic obstructive pulmonary disease, acute myocardial infarction, acute heart failure, and pulmonary embolism. An interrupted time series model was used to determine changes in time trends for risk-adjusted LOS, in-hospital mortality, and 30-day readmission after the implementation of SwissDRG in 2012. Analyses were performed from March 1, 2018, to June 30, 2018, and from November 1, 2018, to December 18, 2018.Main outcomes and measuresMonthly patient-level data for LOS, in-hospital mortality, and 30-day readmission rates.ResultsThe sample included a total of 2 426 722 hospitalized adult patients. Of this total, 1 018 404 patients (41.9%; 531 226 [52.2%] male, median [interquartile range (IQR)] age of 69 [55-80] years) composed the before-SwissDRG period; 1 408 318 patients (58.0%; 730 228 [51.9%] male, median [IQR] age of 70 [56-81] years) composed the after-SwissDRG period. The overall LOS gradually decreased from unadjusted mean (SD) 8.0 (12.7) days in 2009 to 7.2 (17.3) days in 2015. This reduction in LOS, however, was not substantially greater with the implementation of SwissDRG in 2012 (risk-adjusted slope, -0.0166 days; 95% CI, -0.0223 to -0.0110 days), with an adjusted difference in slopes of 0.0000 days (95% CI, -0.0072 to 0.0072 days). Risk-adjusted all-cause in-hospital mortality declined from 4.9% in 2009 to 4.6% in 2015, with a substantially greater decline after implementation of SwissDRG (difference between monthly slopes before and after implementation, -0.0115%; 95% CI, -0.0190% to -0.0039%). In the same period, risk-adjusted 30-day readmission rates increased from 14.4% in 2009 to 15.0% in 2015, with a greater increase after SwissDRG implementation (change in monthly slope, 0.0339%; 95% CI, 0.0254%-0.0423%). Patients with acute myocardial infarction were found to have a substantially greater increase after SwissDRG implementation in 30-day readmission rates (adjusted difference in slopes, 0.1144%; 95% CI, 0.0617%-0.1671%).Conclusions and relevanceAmong medical hospitalizations in Switzerland, SwissDRG implementation appeared to be associated with an increase in readmission rates and a decrease in in-hospital mortality but not with the gradual decrease in LOS observed in the historical control period.
| S-EPMC6484617 | biostudies-literature
Project description:Surgical patients frequently experience postoperative increases in creatinine levels. The authors hypothesized that even small increases in postoperative creatinine levels are associated with adverse outcomes.The authors examined the association of postoperative changes from preoperative baseline creatinine with all-cause in-hospital mortality and hospital length of stay (HLOS) in a retrospective analysis of surgical patients at a single tertiary care center between January 2006 and June 2012.The data of 39,369 surgical patients (noncardiac surgery n = 37,345; cardiac surgery n = 2,024) were analyzed. Acute kidney injury (AKI)-by definition of the Kidney Disease: Improving Global Outcome group-was associated with a five-fold higher mortality (odds ratio [OR], 4.8; 95% CI, 4.1 to 5.7; P < 0.001) and a longer HLOS of 5 days (P < 0.001) after adjusting for age, sex, comorbidities, congestive heart failure, preoperative hemoglobin, preoperative creatinine, exposure to radiocontrast agent, type of surgery, and surgical AKI risk factors. Importantly, even minor creatinine increases (?creatinine 25 to 49% above baseline but < 0.3 mg/dl) not meeting AKI criteria were associated with a two-fold increased risk of death (OR, 1.7; 95% CI, 1.3 to 2.4; P < 0.001) and 2 days longer HLOS (P < 0.001). This was more pronounced in noncardiac surgery patients. Patients with minor creatinine increases had a five-fold risk of death (OR, 5.4; 95% CI, 1.5 to 20.3; P < 0.05) and a 3-day longer HLOS (P < 0.01) when undergoing noncardiac surgery.Even minor postoperative increases in creatinine levels are associated with adverse outcomes. These results emphasize the importance to find effective therapeutic approaches to prevent or treat even mild forms of postoperative kidney dysfunction to improve surgical outcomes.
| S-EPMC4679549 | biostudies-literature
Project description:BackgroundThe population of persons seeking medical care is linguistically diverse in the United States. Language barriers can adversely affect a patient's ability to explain their symptoms. Among hospitalized patients, these barriers may lead to higher readmission rates and longer hospitalizations. Trained interpreters help overcome communication barriers; however, interpreter usage among patients is suboptimal.ObjectiveTo investigate differences among patients with limited English proficiency (LEP) in their length of stay (LOS) and 30-day readmission rate associated with their receiving professional interpretation at admission or discharge.DesignWe analyzed the rates of interpretation at admission and discharge of all LEP patients admitted to a tertiary care hospital over a three-year period. We calculated length of stay in days and as log of LOS. We also examined 30-day readmission. Using multivariable regression models, we explored differences among patients who received interpretation at admission, discharge, or both, controlling for patient characteristics, including age, illness severity, language, and gender.ParticipantsAll LEP patients admitted between May 1, 2004 and April 30, 2007.Main measuresLength of hospital stay as related to use of professional interpreters; readmission to the hospital within 30 days.Key resultsOf the 3071 patients included in the study, 39 % received language interpretation on both admission and discharge date. Patients who did not receive professional interpretation at admission or both admission/discharge had an increase in their LOS of between 0.75 and 1.47 days, compared to patients who had an interpreter on both day of admission and discharge (P<0.02). Patients receiving interpretation at admission and/or discharge were less likely than patients receiving no interpretation to be readmitted with 30 days.ConclusionsThe length of a hospital stay for LEP patients was significantly longer when professional interpreters were not used at admission or both admission/discharge.
| S-EPMC3445680 | biostudies-literature
Project description:Objective: How hospital length of stay after delivery for women with preeclampsia is associated with risk for readmission is unknown. The objective of this study was to evaluate risk for 60-day hypertension-related postpartum readmission based on length of stay after delivery.Methods: The 2014 Healthcare Cost and Utilization Project's (HCUP) Nationwide Readmissions Database was used to analyze risk for readmission for a hypertension-related diagnosis within 60 days from cesarean delivery hospitalization for women with preeclampsia who underwent cesarean delivery. Risk for readmission was evaluated based on postoperative length of stay as well as demographic, hospital, and other obstetric factors. Population weights were applied to create national estimates. Multivariable analyses were performed with adjusted risk ratios (aRR) and corresponding 95% confidence intervals as measures of effect. Mean and median hospital charges based upon postoperative length of stay were also evaluated. Time from delivery hospitalization to readmission was calculated.Results: In 2014, 65 401 women with preeclampsia underwent cesarean delivery. Of these, 1016 women (1.6%) were readmitted for a hypertension-related diagnosis. 921 of the 1016 readmissions occurred within 10 days of discharge (90.6%). In adjusted analyses, postoperative LOS 5-7 days and >7 days compared to LOS <3 days were associated with decreased risk of 60-day hypertension-related readmission (aRR 0.59 95% CI 0.45, 0.78; aRR 0.53 95% CI 0.29, 1.00, respectively). When the cohort was restricted to women with severe preeclampsia or eclampsia, LOS 5-7 days was associated with decreased risk of 60-day hypertension-related readmission in both unadjusted and adjusted analyses compared to LOS <3 days (risk ratios (RR) 0.34, 95% CI 0.18, 0.65; aRR 0.29, 95% 0.18, 0.46, respectively). Median delivery hospitalization charges were $26 512. Compared to LOS <3 days, mean and median charges increased significantly for patients with LOS 4, 5-7, and >7 days.Conclusion: Longer postoperative length of stay during cesarean delivery hospitalizations was associated with decreased risk for postpartum hypertension-related readmission. Most readmissions occurred soon after discharge. These findings support that post-delivery management may play a role in likelihood of women requiring subsequent readmission for complications related to preeclampsia after discharge.
| S-EPMC6380960 | biostudies-literature
Project description:BackgroundThe increasing number of patients treated with peritoneal dialysis (PD) and their consistently high rate of hospital admissions have placed a large burden on the health care system. Early clinical interventions and optimal management of patients at a high risk of prolonged length of stay (pLOS) may help improve the medical efficiency and prognosis of PD-treated patients. If timely clinical interventions are not provided, patients at a high risk of pLOS may face a poor prognosis and high medical expenses, which will also be a burden on hospitals. Therefore, physicians need an effective pLOS prediction model for PD-treated patients.ObjectiveThis study aimed to develop an optimal data-driven model for predicting the pLOS risk of PD-treated patients using basic admission data.MethodsPatient data collected using the Hospital Quality Monitoring System (HQMS) in China were used to develop pLOS prediction models. A stacking model was constructed with support vector machine, random forest (RF), and K-nearest neighbor algorithms as its base models and traditional logistic regression (LR) as its meta-model. The meta-model used the outputs of all 3 base models as input and generated the output of the stacking model. Another LR-based pLOS prediction model was built as the benchmark model. The prediction performance of the stacking model was compared with that of its base models and the benchmark model. Five-fold cross-validation was employed to develop and validate the models. Performance measures included the Brier score, area under the receiver operating characteristic curve (AUROC), estimated calibration index (ECI), accuracy, sensitivity, specificity, and geometric mean (Gm). In addition, a calibration plot was employed to visually demonstrate the calibration power of each model.ResultsThe final cohort extracted from the HQMS database consisted of 23,992 eligible PD-treated patients, among whom 30.3% had a pLOS (ie, longer than the average LOS, which was 16 days in our study). Among the models, the stacking model achieved the best calibration (ECI 8.691), balanced accuracy (Gm 0.690), accuracy (0.695), and specificity (0.701). Meanwhile, the stacking and RF models had the best overall performance (Brier score 0.174 for both) and discrimination (AUROC 0.757 for the stacking model and 0.756 for the RF model). Compared with the benchmark LR model, the stacking model was superior in all performance measures except sensitivity, but there was no significant difference in sensitivity between the 2 models. The 2-sided t tests revealed significant performance differences between the stacking and LR models in overall performance, discrimination, calibration, balanced accuracy, and accuracy.ConclusionsThis study is the first to develop data-driven pLOS prediction models for PD-treated patients using basic admission data from a national database. The results indicate the feasibility of utilizing a stacking-based pLOS prediction model for PD-treated patients. The pLOS prediction tools developed in this study have the potential to assist clinicians in identifying patients at a high risk of pLOS and to allocate resources optimally for PD-treated patients.
| S-EPMC8173398 | biostudies-literature