The Creating Incentives and Continuity Leading to Efficiency staffing model: a quality improvement initiative in hospital medicine

Prediction of Chronic Disease-Related Inpatient Prolonged Length of Stay Using Machine Learning Algorithms

Objectives

The study aimed to develop and compare predictive models based on supervised machine learning algorithms for predicting the prolonged length of stay (LOS) of hospitalized patients diagnosed with five different chronic conditions.

Methods

An administrative claim dataset (2008-2012) of a regional network of nine hospitals in the Tampa Bay area, Florida, USA, was used to develop the prediction models. Features were extracted from the dataset using the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes. Five learning algorithms, namely, decision tree C5.0, linear support vector machine (LSVM), k-nearest neighbors, random forest, and multi-layered artificial neural networks, were used to build the model with semi-supervised anomaly detection and two feature selection methods. Issues with the unbalanced nature of the dataset were resolved using the Synthetic Minority Over-sampling Technique (SMOTE).

Results

LSVM with wrapper feature selection performed moderately well for all patient cohorts. Using SMOTE to counter data imbalances triggered a tradeoff between the model's sensitivity and specificity, which can be masked under a similar area under the curve. The proposed aggregate rank selection approach resulted in a balanced performing model compared to other criteria. Finally, factors such as comorbidity conditions, source of admission, and payer types were associated with the increased risk of a prolonged LOS.

Conclusions

Prolonged LOS is mostly associated with pre-intraoperative clinical and patient socioeconomic factors. Accurate patient identification with the risk of prolonged LOS using the selected model can provide hospitals a better tool for planning early discharge and resource allocation, thus reducing avoidable hospitalization costs.

Incorporating Continuity in a 7-On 7-Off Hospitalist Model and the Correlation With Patient Handoffs and Length of Stay

Little research in hospital medicine examines the effects of hospitalist continuity on patient outcomes. This study implemented a novel staffing model with approximately half of rounding teams starting their 7-day workweek on Monday and the others on Friday. Teams admitted their own patients on their first 4 days with additional nighttime admissions handed off to those teams. No admissions were given to teams on their last 3 days. Length of stay was significantly reduced from 6.34 days in 2015 to 5.7 days in 2016 (P < .002) with a significant decrease in handoffs. There was an increase in odds ratio of death (1.37, SE = .128) with each additional hospitalist involved in a patient's care while adjusting for year and number of patient diagnoses (P < .001). There was no statistical difference in charges, 30-day readmissions, or mortality between years.

Increased Mortality Rates During Resident Handoff Periods and the Effect of ACGME Duty Hour Regulations

BACKGROUND

Medical errors occur following handoff-related miscommunication. Data regarding the effect on patient-centered outcomes, specifically mortality, are lacking. Our objective was to investigate handoff-related mortality and the effect of duty-hour regulations.

METHODS

Retrospective cohort study of adult medical patients at a public, university-affiliated hospital from 2010 to 2012. Patients were divided into 2 cohorts: handoff group (discharged within 7 days following a change in resident physician team) vs control group (discharged the 3 weeks of each 4-week rotation before resident service change). The primary outcome was unadjusted and adjusted hospital mortality rate. As a secondary prespecified analysis, we examined the effect of 2011 Accreditation Council for Graduate Medical Education (ACGME) duty-hour changes.

RESULTS

Among 23,736 patients, unadjusted hospital mortality during the handoff group was higher than the control group (2.68% vs 2.08%, respectively; P = .007; odds ratio [OR] 1.30; 95% confidence interval [CI], 1.08-1.57). Following adjustment, this association remained statistically significant (adjusted OR 1.34; P = .003; 95% CI, 1.10-1.62). Similarly, pre-duty-hour unadjusted hospital mortality was higher in the handoff group vs control group (2.87% vs 2.01%, respectively; P = .006; OR 1.44; 95% CI, 1.11-1.86), which remained statistically significant following adjustment (adjusted OR 1.50; P = .002; 95% CI, 1.16-1.95). However, this association lost statistical significance following duty-hour revision with respect to both unadjusted (2.48% vs 2.15%, respectively; P = .30; OR 1.16; 95% CI, 0.88-1.53) and adjusted mortality (OR 1.18; P = .26; 95% CI, 0.89-1.56).

CONCLUSIONS

Resident transition in care was significantly associated with an increase in unadjusted and adjusted hospital mortality. Although improved by 2011 ACGME duty-hour amendments, a trend toward higher mortality remained following resident handoff.

The effect of hospitalist discontinuity on adverse events

BACKGROUND

Patient-physician continuity is difficult to achieve in hospital settings because of the need to provide care continuously. The impact of hospital physician discontinuity on patient safety is unknown.

OBJECTIVE

To determine the association between hospital physician continuity and the incidence of adverse events (AEs).

DESIGN

Retrospective observational study using multivariable models to adjust for patient characteristics.

PARTICIPANTS

Patients admitted to a nonteaching hospitalist service in a large academic hospital between March 1, 2009 and December 31, 2011.

MAIN MEASURE(S)

Two measures of continuity were used. The Number of Physicians Index (NPI) was the total number of unique hospitalists caring for a patient. The Usual Provider of Care (UPC) Index was the proportion of encounters with the most frequently encountered hospitalist. Outcome measures were AEs detected by automated queries of information systems and confirmed by 2 physician researchers.

KEY RESULTS

Our analysis included data from 474 hospitalizations. In unadjusted models, each 1-unit increase in the NPI (ie, less continuity) was significantly associated with the incidence of 1 or more AEs (odds ratio = 1.75; P < 0.001). However, UPC was not associated with incidence of AEs. Across all adjusted models, neither NPI nor UPC was significantly associated with the incidence of AEs. The direction of the effect of discontinuity on AEs was also inconsistent across models.

CONCLUSIONS

Hospitalist physician continuity does not appear to be associated with the incidence of AEs. Because hospital care is provided by teams of clinicians, future research should evaluate the impact of team complexity and dynamics on patient outcomes.

The impact of hospitalist discontinuity on hospital cost, readmissions, and patient satisfaction

BACKGROUND

Achieving patient-physician continuity is difficult in the inpatient setting, where care must be provided continuously. Little is known about the impact of hospital physician discontinuity on outcomes.

OBJECTIVE

To determine the association between hospital physician continuity and percentage change in median cost of hospitalization, 30-day readmission, and patient satisfaction with physician communication.

DESIGN

Retrospective observational study using various multivariable models to adjust for patient characteristics.

PARTICIPANTS

Patients admitted to a non-teaching hospitalist service in a large, academic, urban hospital between 6 July 2008 and 31 December 2011.

MAIN MEASURES

We used two measures of continuity: the Number of Physicians Index (NPI), and the Usual Provider of Continuity (UPC) index. The NPI is the total number of unique physicians caring for a patient, while the UPC is calculated as the largest number of patient encounters with a single physician, divided by the total number of encounters. Outcome measures were percentage change in median cost of hospitalization, 30-day readmissions, and top box responses to satisfaction with physician communication.

KEY RESULTS

Our analyses included data from 18,375 hospitalizations. Lower continuity was associated with modest increases in costs (range 0.9-12.6 % of median), with three of the four models used achieving statistical significance. Lower continuity was associated with lower odds of readmission (OR = 0.95-0.98 across models), although only one of the models achieved statistical significance. Satisfaction with physician communication was lower, with less continuity across all models, but results were not statistically significant.

CONCLUSIONS

Hospital physician discontinuity appears to be associated with modestly increased hospital costs. Hospital physicians may revise plans as they take over patient care responsibility from their colleagues.