Readmissions of medical patients: an external validation of two existing prediction scores

Assessing racial bias in healthcare predictive models: Practical lessons from an empirical evaluation of 30-day hospital readmission models

OBJECTIVE

Despite increased availability of methodologies to identify algorithmic bias, the operationalization of bias evaluation for healthcare predictive models is still limited. Therefore, this study proposes a process for bias evaluation through an empirical assessment of common hospital readmission models. The process includes selecting bias measures, interpretation, determining disparity impact and potential mitigations.

METHODS

This retrospective analysis evaluated racial bias of four common models predicting 30-day unplanned readmission (i.e., LACE Index, HOSPITAL Score, and the CMS readmission measure applied as is and retrained). The models were assessed using 2.4 million adult inpatient discharges in Maryland from 2016 to 2019. Fairness metrics that are model-agnostic, easy to compute, and interpretable were implemented and apprised to select the most appropriate bias measures. The impact of changing model's risk thresholds on these measures was further assessed to guide the selection of optimal thresholds to control and mitigate bias.

RESULTS

Four bias measures were selected for the predictive task: zero-one-loss difference, false negative rate (FNR) parity, false positive rate (FPR) parity, and generalized entropy index. Based on these measures, the HOSPITAL score and the retrained CMS measure demonstrated the lowest racial bias. White patients showed a higher FNR while Black patients resulted in a higher FPR and zero-one-loss. As the models' risk threshold changed, trade-offs between models' fairness and overall performance were observed, and the assessment showed all models' default thresholds were reasonable for balancing accuracy and bias.

CONCLUSIONS

This study proposes an Applied Framework to Assess Fairness of Predictive Models (AFAFPM) and demonstrates the process using 30-day hospital readmission model as the example. It suggests the feasibility of applying algorithmic bias assessment to determine optimized risk thresholds so that predictive models can be used more equitably and accurately. It is evident that a combination of qualitative and quantitative methods and a multidisciplinary team are necessary to identify, understand and respond to algorithm bias in real-world healthcare settings. Users should also apply multiple bias measures to ensure a more comprehensive, tailored, and balanced view. The results of bias measures, however, must be interpreted with caution and consider the larger operational, clinical, and policy context.

The Early HOSPITAL Score to Predict 30-Day Readmission Soon After Hospitalization: a Prospective Multicenter Study

BACKGROUND

The simplified HOSPITAL score is an easy-to-use prediction model to identify patients at high risk of 30-day readmission before hospital discharge. An earlier stratification of this risk would allow more preparation time for transitional care interventions.

OBJECTIVE

To assess whether the simplified HOSPITAL score would perform similarly by using hemoglobin and sodium level at the time of admission instead of discharge.

DESIGN

Prospective national multicentric cohort study.

PARTICIPANTS

In total, 934 consecutively discharged medical inpatients from internal general services.

MAIN MEASURES

We measured the composite of the first unplanned readmission or death within 30 days after discharge of index admission and compared the performance of the simplified score with lab at discharge (simplified HOSPITAL score) and lab at admission (early HOSPITAL score) according to their discriminatory power (Area Under the Receiver Operating characteristic Curve (AUROC)) and the Net Reclassification Improvement (NRI).

KEY RESULTS

During the study period, a total of 3239 patients were screened and 934 included. In total, 122 (13.2%) of them had a 30-day unplanned readmission or death. The simplified and the early versions of the HOSPITAL score both showed very good accuracy (Brier score 0.11, 95%CI 0.10-0.13). Their AUROC were 0.66 (95%CI 0.60-0.71), and 0.66 (95%CI 0.61-0.71), respectively, without a statistical difference (p value 0.79). Compared with the model at discharge, the model with lab at admission showed improvement in classification based on the continuous NRI (0.28; 95%CI 0.08 to 0.48; p value 0.004).

CONCLUSION

The early HOSPITAL score performs, at least similarly, in identifying patients at high risk for 30-day unplanned readmission and allows a readmission risk stratification early during the hospital stay. Therefore, this new version offers a timely preparation of transition care interventions to the patients who may benefit the most.

Characteristics of Patients with Cancer Readmitted Within 30 Days to an Acute Palliative Care Unit

OBJECTIVE(S)

For patients with cancer, the emergence of acute palliative care units (APCU) may hold promise in curtailing hospital readmissions. The study aims to describe the characteristics of patients readmitted to an APCU.

METHODS

This retrospective study examined patients with cancer readmitted within 30 days to an APCU. Readmissions were further classified as either potentially preventable or non-preventable.

RESULTS

Out of 734 discharges from July 1, 2014 to July 1, 2015, 69 (9%) readmissions were identified and analyzed. For index admissions, median length of stay was five days, and one (1%) was discharged home with hospice care. For readmissions, median time from index admission to readmission was nine days, median length of stay was six days, three (4%) patients died, and 20 (30%) went home with hospice. Ten (14.5%) readmissions were deemed potentially preventable (95% CI 7.2-25.0%). Race/ethnicity-White/Black/Hispanic/Others-was 60%, 10%, 20% and 10%, respectively, among potentially preventable readmissions and 76%, 22%, 2% and 0%, respectively, among potentially non-preventable readmissions (P = .012). Potentially preventable readmissions were more likely to have venous thromboembolism (40% vs. 12%, P = .046) and more reasons for readmission (median 2 vs. 1, P = .019).

CONCLUSIONS

Among patients with cancer readmitted to an APCU, one out of seven was potentially preventable and a far larger proportion was discharged with hospice care compared to the index admission. Recognition of disease course, meaningful goals of care discussions and timely transition to hospice care may reduce rehospitalization in this population.

Prediction of 30-Day Readmission in Hospitalized Older Adults Using Comprehensive Geriatric Assessment and LACE Index and HOSPITAL Score

(1) Background: Elders have higher rates of rehospitalization, especially those with functional decline. We aimed to investigate potential predictors of 30-day readmission risk by comprehensive geriatric assessment (CGA) in hospitalized patients aged 65 years or older and to examine the predictive ability of the LACE index and HOSPITAL score in older patients with a combination of malnutrition and physical dysfunction. (2) Methods: We included patients admitted to a geriatric ward in a tertiary hospital from July 2012 to August 2018. CGA components including cognitive, functional, nutritional, and social parameters were assessed at admission and recorded, as well as clinical information. The association factors with 30-day hospital readmission were analyzed by multivariate logistic regression analysis. The predictive ability of the LACE and HOSPITAL score was assessed using receiver operator characteristic curve analysis. (3) Results: During the study period, 1509 patients admitted to a ward were recorded. Of these patients, 233 (15.4%) were readmitted within 30 days. Those who were readmitted presented with higher comorbidity numbers and poorer performance of CGA, including gait ability, activities of daily living (ADL), and nutritional status. Multivariate regression analysis showed that male gender and moderately impaired gait ability were independently correlated with 30-day hospital readmissions, while other components such as functional impairment (as ADL) and nutritional status were not associated with 30-day rehospitalization. The receiver operating characteristics for the LACE index and HOSPITAL score showed that both predicting scores performed poorly at predicting 30-day hospital readmission (C-statistic = 0.59) and did not perform better in any of the subgroups. (4) Conclusions: Our study showed that only some components of CGA, mobile disability, and gender were independently associated with increased risk of readmission. However, the LACE index and HOSPITAL score had a poor discriminating ability for predicting 30-day hospitalization in all and subgroup patients. Further identifiers are required to better estimate the 30-day readmission rates in this patient population.

A bias evaluation checklist for predictive models and its pilot application for 30-day hospital readmission models

OBJECTIVE

Health care providers increasingly rely upon predictive algorithms when making important treatment decisions, however, evidence indicates that these tools can lead to inequitable outcomes across racial and socio-economic groups. In this study, we introduce a bias evaluation checklist that allows model developers and health care providers a means to systematically appraise a model's potential to introduce bias.

MATERIALS AND METHODS

Our methods include developing a bias evaluation checklist, a scoping literature review to identify 30-day hospital readmission prediction models, and assessing the selected models using the checklist.

RESULTS

We selected 4 models for evaluation: LACE, HOSPITAL, Johns Hopkins ACG, and HATRIX. Our assessment identified critical ways in which these algorithms can perpetuate health care inequalities. We found that LACE and HOSPITAL have the greatest potential for introducing bias, Johns Hopkins ACG has the most areas of uncertainty, and HATRIX has the fewest causes for concern.

DISCUSSION

Our approach gives model developers and health care providers a practical and systematic method for evaluating bias in predictive models. Traditional bias identification methods do not elucidate sources of bias and are thus insufficient for mitigation efforts. With our checklist, bias can be addressed and eliminated before a model is fully developed or deployed.

CONCLUSION

The potential for algorithms to perpetuate biased outcomes is not isolated to readmission prediction models; rather, we believe our results have implications for predictive models across health care. We offer a systematic method for evaluating potential bias with sufficient flexibility to be utilized across models and applications.

Comparison of LACE and HOSPITAL Readmission Risk Scores for CMS Target and Nontarget Conditions

This study evaluated the utility and performance of the LACE index and HOSPITAL score with consideration of the type of diagnoses and assessed the accuracy of these models for predicting readmission risks in patient cohorts from 2 large academic medical centers. Admissions to 2 hospitals from 2011 to 2015, derived from the Vizient Clinical Data Base and regional health information exchange, were included in this study (291 886 encounters). Models were assessed using Bayesian information criterion and area under the receiver operating characteristic curve. They were compared in CMS diagnosis-based cohorts and in 2 non-CMS cancer diagnosis-based cohorts. Overall, both models for readmission risk performed well, with LACE performing slightly better (area under the receiver operating characteristic curve 0.73 versus 0.69; P ≤ 0.001). HOSPITAL consistently outperformed LACE among 4 CMS target diagnoses, lung cancer, and colon cancer. Both LACE and HOSPITAL predict readmission risks well in the overall population, but performance varies by salient, diagnosis-based risk factors.

HOSPITAL Score and LACE Index to Predict Mortality in Multimorbid Older Patients

Background

Estimating life expectancy of older adults informs whether to pursue future investigation and therapy. Several models to predict mortality have been developed but often require data not immediately available during routine clinical care. The HOSPITAL score and the LACE index were previously validated to predict 30-day readmissions but may also help to assess mortality risk. We assessed their performance to predict 1-year and 30-day mortality in hospitalized older multimorbid patients with polypharmacy.

Methods

We calculated the HOSPITAL score and LACE index in patients from the OPERAM (OPtimising thERapy to prevent Avoidable hospital admissions in the Multimorbid elderly) trial (patients aged ≥ 70 years with multimorbidity and polypharmacy, admitted to hospital across four European countries in 2016–2018). Our primary and secondary outcomes were 1-year and 30-day mortality. We assessed the overall accuracy (scaled Brier score, the lower the better), calibration (predicted/observed proportions), and discrimination (C-statistic) of the models.

Results

Within 1 year, 375/1879 (20.0%) patients had died, including 94 deaths within 30 days. The overall accuracy was good and similar for both models (scaled Brier score 0.01–0.08). The C-statistics were identical for both models (0.69 for 1-year mortality, p = 0.81; 0.66 for 30-day mortality, p = 0.94). Calibration showed well-matching predicted/observed proportions.

Conclusion

The HOSPITAL score and LACE index showed similar performance to predict 1-year and 30-day mortality in older multimorbid patients with polypharmacy. Their overall accuracy was good, their discrimination low to moderate, and the calibration good. These simple tools may help predict older multimorbid patients’ mortality after hospitalization, which may inform post-hospitalization intensity of care.

Prediction value of the LACE index to identify older adults at high risk for all-cause mortality in South Korea: a nationwide population-based study

Background

As a tool to predict early hospital readmission, little is known about the association between LACE index and all-cause mortality in older adults. We aimed to validate the LACE index to predict all-cause mortality in older adults and also analyzed the LACE index outcome of all-cause mortality depending on the disease and age of the participants.

Methods

We used the National Health Insurance Service (NHIS) cohort, a nationwide claims database of Koreans. We enrolled 7491 patients who were hospitalized at least once between 2003 and 2004, aged ≥65 years as of the year of discharge, and subsequently followed-up until 2015. We estimated the LACE index using the NHI database. The Cox proportional hazards model was used to estimate the hazard ratio (HR) for all-cause mortality. Furthermore, we investigated all-cause mortality according to age and underlying disease when the LACE index was ≥10 and < 10, respectively.

Results

In populations over 65 years of age, patients with LACE index ≥10 had significantly higher risks of all-cause mortality than in those with LACE index < 10. (HR, 1.44; 95% confidence interval, 1.35–1.54). For those patients aged 65–74 years, the HR of all-cause mortality was found to be higher in patients with LACE index≥10 than in those with LACE index < 10 in almost all the diseases except CRF and mental illnesses. And those patients aged ≥75 years, the HR of all- cause mortality was found to be higher in patients with LACE index ≥10 than in those with LACE index < 10 in the diseases of pneumonia and MACE.

Conclusion

This is the first study to validate the predictive power of the LACE index to identify older adults at high risk for all-cause mortality using nationwide cohort data. Our findings have policy implications for selecting or managing patients who need post-discharge management.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12877-022-02848-4.

HOSPITAL Score, LACE Index and LACE+ Index as predictors of 30-day readmission in patients with heart failure

This study aimed to evaluate the accuracy of the HOSPITAL Score (Haemoglobin level at discharge, Oncology at discharge, Sodium level at discharge, Procedure during hospitalization, Index admission, number of hospital admissions, Length of stay) LACE index (Length of stay, Acute/emergent admission, Charlson comorbidy index score, Emerency department visits in previous 6 months) and LACE+ index in predicting 30-day readmission in patients with diastolic dysfunction. Heart failure remains one of the most common hospital readmissions in adults, leading to significant morbidity and mortality. Different models have been used to predict 30-day hospital readmissions. All adult medical patients discharged from the SIU School of Medicine Hospitalist service from 12 June 2016 to 12 June 2018 with an International Classification of Disease, 10th Revision, Clinical Modification diagnosis of diastolic heart failure were studied retrospectively to evaluate the performance of the HOSPITAL Score, LACE index and LACE+ index readmission risk prediction tools in this patient population. Of the 730 patient discharges with a diagnosis of heart failure with preserved ejection fraction (HFpEF), 692 discharges met the inclusion criteria. Of these discharges, 189 (27%) were readmitted to the same hospital within 30 days. A receiver operating characteristic evaluation showed C-statistic values to be 0.595 (95% CI 0.549 to 0.641) for the HOSPITAL Score, 0.551 (95% CI 0.503 to 0.598) for the LACE index and 0.568 (95% CI 0.522 to 0.615) for the LACE+ index, indicating poor specificity in predicting 30-day readmission. The result of this study demonstrates that the HOSPITAL Score, LACE index and LACE+ index are not effective predictors of 30-day readmission for patients with HFpEF. Further analysis and development of new prediction models are needed to better estimate the 30-day readmission rates in this patient population.

Validation of the LACE Index (Length of Stay, Acuity of Admission, Comorbidities, Emergency Department Use) in the Adult Neurosurgical Patient Population

BACKGROUND

The LACE index (Length of stay, Acuity of admission, Comorbidities, Emergency department use) quantifies the risk of mortality or unplanned readmission within 30 d after hospital discharge. The index was validated originally in a large, general population and, subsequently, in several specialties, not including neurosurgery.

OBJECTIVE

To determine if the LACE index accurately predicts mortality and unplanned readmission of neurosurgery patients within 30 d of discharge.

METHODS

We performed a retrospective, cohort study of consecutive neurosurgical procedures between January 1 and September 29, 2017 at our institution. The LACE index and other clinical data were abstracted. Data analysis included univariate and multivariate logistic regressions.

RESULTS

Of the 1,054 procedures on 974 patients, 52.7% were performed on females. Mean age was 54.2 ± 15.4 yr. At time of discharge, the LACE index was low (1-4) in 58.3% of patients, moderate (5-9) in 32.4%, and high (10-19) in 9.3%. Rates of readmission and mortality within 30 d were 7.0, 11.4, and 14.3% in the low-, moderate-, and high-risk groups, respectively. Moderate-risk (odds ratio [OR] 1.62, 95% CI 1.02-2.56, P = .04) and high-risk LACE indexes (OR 2.20, 95% CI 1.15-4.19, P = .02) were associated with greater odds of readmission or mortality, adjusting for all variables. Additionally, longer operations (OR 1.11, 95% CI 1.02-1.21, P = .02) had greater odds of readmission. Specificity of the high-risk score to predict 30-d readmission or mortality was 91.2%.

CONCLUSION

A moderate- or high-risk LACE index can be applied to neurosurgical populations to predict 30-d readmission and mortality. Longer operations are potential predictors of readmission or mortality.

Readmission of older acutely admitted medical patients after short-term admissions in Denmark: a nationwide cohort study

Background

Knowledge of unplanned readmission rates and prognostic factors for readmission among older people after early discharge from emergency departments is sparse. The aims of this study were to examine the unplanned readmission rate among older patients after short-term admission, and to examine risk factors for readmission including demographic factors, comorbidity and admission diagnoses.

Methods

This cohort study included all medical patients aged ≥65 years acutely admitted to Danish hospitals between 1 January 2013 and 30 June 2014 and surviving a hospital stay of ≤24 h. Data on readmission within 30 days, comorbidity, demographic factors, discharge diagnoses and mortality were obtained from the Danish National Registry of Patients and the Danish Civil Registration System. We examined risk factors for readmission using a multivariable Cox regression to estimate adjusted hazard ratios (aHR) with 95% confidence intervals (CI) for readmission.

Results

A total of 93,306 patients with a median age of 75 years were acutely admitted and discharged within 24 h, and 18,958 (20.3%; 95% CI 20.1 - 20.6%) were readmitted with a median time to readmission of 8 days (IQR 3 - 16 days). The majority were readmitted with a new diagnosis. Male sex (aHR 1.15; 1.11 - 1.18) and a Charlson Comorbidity Index ≥3 (aHR 2.28; 2.20 - 2.37) were associated with an increased risk of readmission. Discharge diagnoses associated with increased risk of readmission were heart failure (aHR 1.26; 1.12 - 1.41), chronic obstructive pulmonary disease (aHR 1.33; 1.25 - 1.43), dehydration (aHR 1.28; 1.17 - 1.39), constipation (aHR 1.26; 1.14 - 1.39), anemia (aHR 1.45; 1.38 - 1.54), pneumonia (aHR 1.15; 1.06 - 1.25), urinary tract infection (aHR 1.15; 1.07 - 1.24), suspicion of malignancy (aHR 1.51; 1.37 - 1.66), fever (aHR 1.52; 1.33 - 1.73) and abdominal pain (aHR 1.12; 1.05 - 1.19).

Conclusions

One fifth of acutely admitted medical patients aged ≥65 were readmitted within 30 days after early discharge. Male gender, the burden of comorbidity and several primary discharge diagnoses were risk factors for readmission.

Assess the Performance and Cost-Effectiveness of LACE and HOSPITAL Re-Admission Prediction Models as a Risk Management Tool for Home Care Patients: An Evaluation Study of a Medical Center Affiliated Home Care Unit in Taiwan

The LACE index and HOSPITAL score models are the two most commonly used prediction models identifying patients at high risk of readmission with limited information for home care patients. This study compares the effectiveness of these two models in predicting 30-day readmission following acute hospitalization of such patients in Taiwan. A cohort of 57 home care patients were enrolled and followed-up for one year. We compared calibration, discrimination (area under the receiver operating curve, AUC), and net reclassification improvement (NRI) to identify patients at risk of 30-day readmission for both models. Moreover, the cost-effectiveness of the models was evaluated using microsimulation analysis. A total of 22 readmissions occurred after 87 acute hospitalizations during the study period (readmission rate = 25.2%). While the LACE score had poor discrimination (AUC = 0.598, 95% confidence interval (CI) = 0.488–0.702), the HOSPITAL score achieved helpful discrimination (AUC = 0.691, 95% CI = 0.582–0.785). Moreover, the HOSPITAL score had improved the risk prediction in 38.3% of the patients, compared with the LACE index (NRI = 0.383, 95% CI = 0.068–0.697, p = 0.017). Both prediction models effectively reduced readmission rates compared to an attending physician’s model (readmission rate reduction: LACE, 39.2%; HOSPITAL, 43.4%; physician, 10.1%; p < 0.001). The HOSPITAL score provides a better prediction of readmission and has potential as a risk management tool for home care patients.

Frailty and hospital outcomes within a low socioeconomic population

BACKGROUND

Clinical frailty scales (CFS) predict hospital-related outcomes. Frailty is more common in areas of higher socioeconomic disadvantage, but no studies exclusively report on the impact of CFS on hospital-related outcomes in areas of known socioeconomic disadvantage.

AIMS

To evaluate the association of the CFS with hospital-related outcomes.

DESIGN

Retrospective observational study in a community hospital within a disadvantaged area in Australia (Social Economic Index for Areas = 0.1%).

METHODS

The CFS was used in the emergency department (ED) for people aged ≥ 75 years. Frailty was defined as a score of ≥4. Associations between the CFS and mortality, admission rates, ED presentations and length of stay (LOS) were analysed using regression analyses.

RESULTS

Between 11 July 2017 and 31 March 2018, there were 5151 ED presentations involving 3258 patients aged ≥ 75 years. Frail persons were significantly more likely to be older, represent to the ED and have delirium compared with non-frail persons. CFS was independently associated with 28-day mortality, with odds of mortality increasing by 1.5 times per unit increase in CFS (95% CI: 1.3-1.7). Frail persons with CFS 4-6 were more likely to be admitted (OR: 1.2; 95% CI: 1.0-1.5), have higher geometric mean LOS (1.43; 95% CI 1.15-1.77 days) and higher rates of ED presentations (IRR: 1.12; 95% CI 1.04-1.21) compared with non-frail persons.

CONCLUSIONS

The CFS predicts community hospital-related outcomes in frail persons within a socioeconomic disadvantage area. Future intervention and allocation of resources could consider focusing on CFS 4-6 as a priority for frail persons within a community hospital setting.

Vital Sign Abnormalities on Discharge Do Not Predict 30-Day Readmission

OBJECTIVE

Hospital readmissions are common and expensive. Risk factors for hospital readmission may include vital sign abnormalities (VSA) at the time of discharge. The study aimed to validate VSA at the time of discharge as a useful predictor of hospital readmission within 30 days of discharge. VSA was compared to the validated HOSPITAL score and LACE index readmission risk prediction models.

DESIGN

All adult medical patients discharged from internal medicine hospitalist service were studied retrospectively. Variables such as age, gender, diagnoses, vital signs at discharge, 30-day hospital readmission, and components for the HOSPITAL score and LACE index were extracted from the electronic health record for analysis.

SETTINGS

A 507-bed university-affiliated tertiary care center.

PARTICIPANTS

During the 2-year study period, a cohort of 1,916 discharges for the hospitalist service were evaluated. The final analysis was based on the data from 1,781 hospital discharges that met the inclusion criteria.

RESULTS

VSA was found in 13% of the study population. Only one abnormal vital sign was present in a higher proportion readmitted to the hospital within 30 days of discharge. No discharges had three or more unstable vital signs. Receiver operating characteristic (ROC) comparisons of the HOSPITAL score (C statistic of 0.67, P < 0.001), LACE index (C statistic of 0.61, P < 0.001), and VSA (C statistic of 0.52, P = 0.318) indicated that VSA at time of discharge was not a useful predictor of hospital readmission within 30 days of discharge.

CONCLUSION

Our study indicated that VSA at the time of discharge is not a useful predictor of 30-day hospital readmission at a university-affiliated teaching hospital. The more complex and validated HOSPITAL score and LACE index were useful predictors of hospital readmission in this patient population.

Multicentre, prospective observational study of the correlation between the Glasgow Admission Prediction Score and adverse outcomes

OBJECTIVES

To assess whether the Glasgow Admission Prediction Score (GAPS) is correlated with hospital length of stay, 6-month hospital readmission and 6-month all-cause mortality. This study represents a 6-month follow-up of patients who were included in an external validation of the GAPS' ability to predict admission at the point of triage.

SETTING

Sampling was conducted between February and May 2016 at two separate emergency departments (EDs) in Sheffield and Glasgow.

PARTICIPANTS

Data were collected prospectively at triage for consecutive adult patients who presented to the ED within sampling times. Any patients who avoided formal triage were excluded from the study. In total, 1420 patients were recruited.

PRIMARY OUTCOMES

GAPS was calculated following triage and did not influence patient management. Length of hospital stay, hospital readmission and mortality against GAPS were modelled using survival analysis at 6 months.

RESULTS

Of the 1420 patients recruited, 39.6% of these patients were initially admitted to hospital. At 6 months, 30.6% of patients had been readmitted and 5.6% of patients had died. For those admitted at first presentation, the chance of being discharged fell by 4.3% (95% CI 3.2% to 5.3%) per GAPS point increase. Cox regression indicated a 9.2% (95% CI 7.3% to 11.1%) increase in the chance of 6-month hospital readmission per point increase in GAPS. An association between GAPS and 6-month mortality was demonstrated, with a hazard increase of 9.0% (95% CI 6.9% to 11.2%) for every point increase in GAPS.

CONCLUSION

A higher GAPS is associated with increased hospital length of stay, 6-month hospital readmission and 6-month all-cause mortality. While GAPS's primary application may be to predict admission and support clinical decision making, GAPS may provide valuable insight into inpatient resource allocation and bed planning.

Risk scores identifying elderly inpatients at risk of 30-day unplanned readmission and accident and emergency department visit: a systematic review

OBJECTIVES

The aim of this systematic review was to describe and analyse the performance statistics of validated risk scores identifying elderly inpatients at risk of early unplanned readmission.

DATA SOURCES

We identified potentially eligible studies by searching MEDLINE, EMBASE, COCHRANE and Web of Science. Our search was restricted to original studies, between 1966 and 2018.

ELIGIBILITY CRITERIA

Original studies, which internally or externally validated the clinical scores of hospital readmissions in elderly inpatients.

DATA EXTRACTION AND SYNTHESIS

A data extraction grid based on Strengthening the Reporting of Observational Studies in Epidemiology and Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis statements was developed and completed by two reviewers to collect general data. The same process was used to extract metrological data of the selected scores.

QUALITY ASSESSMENT OF THE INCLUDED STUDIES

Assessment of the quality and risk of bias in individual studies was performed by two reviewers, using the validated Effective Public Health Practice Project quality assessment tool.

PARTICIPANTS

Elderly inpatients discharged to home from hospital or returning home after an accident and emergency department visit.

RESULTS

A total of 12 studies and five different scores were included in the review. The five scores present area under the receiving operating characteristic curve between 0.445 and 0.69. Identification of Senior At Risk (ISAR) and Triage  Risk Screening Tool (TRST) scores were the more frequently validated scores with ISAR being more sensitive and TRST more specific.

CONCLUSIONS

The TRST and ISAR scores have been extensively studied and validated. The choice of the most suitable score relies on available patient data, patient characteristics and the foreseen clinical care intervention. In order to pair the intervention with the appropriate clinical score, further studies of external validation of clinical scores, identifying elderly patients at risk of early unplanned readmission, are needed.

PROSPERO REGISTRATION NUMBER

CRD42017054516.

The use of analytic hierarchy process for measuring the complexity of medical diagnosis

Diagnostic complexity is an important contextual factor affecting a variety of medical outcomes. Existing measurements of diagnosis complexity either rely on crude proxies or use fine-grained measures that employ indicators from proprietary data that are not readily available. Hence, the study of this important construct in fields such as medical informatics has been hampered by the difficulty of measuring diagnostic complexity. This article presents a novel approach for conceptualizing and operationalizing diagnostic task complexity as a multi-dimensional construct, which employs the readily available International Classification of Diseases codes from medical encounters in hospitals and uses Analytic Hierarchical Process methodology. We demonstrate the reliability of the proposed approach and show that despite using a relatively simple procedure, it is able to predict readmission rates just as well as (or even better) than some of the sophisticated measures that have been used in recent studies (namely, the LaCE score index).

An improved support vector machine-based diabetic readmission prediction

BACKGROUND AND OBJECTIVE

In healthcare systems, the cost of unplanned readmission accounts for a large proportion of total hospital payment. Hospital-specific readmission rate becomes a critical issue around the world. Quantification and early identification of unplanned readmission risks will improve the quality of care during hospitalization and reduce the occurrence of readmission. In clinical practice, medical workers generally use LACE score method to evaluate patient readmission risks, but this method usually performs poorly. With this in mind, this study presents a novel method combining support vector machine and genetic algorithm to build the risk prediction model, which simultaneously involves feature selection and the processing of imbalanced data. This model aims to provide decision support for clinicians during the discharge management of patients with diabetes.

METHOD

The experiments were conducted from a set of 8756 medical records with 50 different features about diabetic readmission. After preprocessing the data, an effective SMOTE-based method was proposed to solve the imbalance data problem. Further, in order to improve prediction performance, a hybrid feature selection mechanism was devised to select the important features. Subsequently, an improved support vector machine-based (SVM-based) method was developed and the genetic algorithm was used to tune the sensitive parameter of the algorithm. Finally, the five-fold cross-validation method was applied to compare the performance of proposed method with other methods (LACE score, logistic regression, naïve bayes, decision tree and feed forward neural networks).

RESULTS

Experimental results indicate that the proposed SVM-based method achieves an accuracy of 81.02%, a sensitivity of 82.89%, a specificity of 79.23%, and outperforms other popular algorithms in identifying diabetic patients who may be readmitted.

CONCLUSIONS

Our research can improve the performance of clinic decision support systems for diabetic readmission, by which the readmission possibility as well as the waste of medical resources can be reduced.

Validation of the health assets index in the Australian inpatient setting: a multicentre prospective cohort protocol study

INTRODUCTION

It is well known that frail older adults are at increased risk for mortality and functional decline on admission to hospital. Systematic review demonstrates that health assets are associated with improved outcomes for hospitalised older adults. The health assets index (HAI) has been developed to measure health assets in the hospital setting. A protocol has been developed to determine the predictive validity of the HAI for frail older adults.

METHODS AND ANALYSIS

The HAI was developed based on a systematic review and secondary analysis of the interRAI-Acute Care (interRAI-AC) dataset. A pilot study was undertaken to refine the tool.The validation study will be a multicentre prospective cohort. Participants will be adults aged 70 years and older with an unplanned admission to hospital. Frailty, illness severity and demographic data will also be recorded. The primary outcomes are mortality at 28 days postdischarge and functional decline at the time of discharge from hospital. The primary hypothesis is that a higher score on the HAI will mitigate the effects of frailty for hospitalised older adults. The secondary outcomes to be recorded are length of stay, readmission at 28 days and functional status at 28 days postdischarge. The correlation between HAI and frailty will be explored. A multivariate analysis will be undertaken to determine the relationship between the HAI and the outcomes of interest.

ETHICS AND DISSEMINATION

Ethical approval has been obtained from Austin Health Human High Risk Ethics Committee. The results will be disseminated in peer-reviewed journals and research conferences. This study will determine whether the HAI has predictive validity for mortality and functional decline for hospitalised, frail older adults.

Different Perspectives on Predictability and Preventability of Surgical Readmissions

BACKGROUND

Although unscheduled readmissions are increasingly being used as a quality indicator, only few readmission studies have focused on surgical patient populations.

METHODS

An observational study "CURIOS@" was performed at three centers in the Netherlands. Readmitted patients and treating doctors were surveyed to assess the discharge process during index admission and their opinion on predictability and preventability of the readmission. Risk factors associated with predictability and preventability as judged by patients and their doctor were identified. Cohen's kappa was calculated to measure pairwise agreement of considering readmission as predictable/preventable. PRISMA root cause categories were used to qualify the reasons for readmission.

RESULTS

In 237 unscheduled surgical readmissions, more patients assessed their readmissions to be likely preventable compared with their treating doctors (28.7% versus 6.8%; kappa, 0.071). This was also reflected in poor consensus about risk factors and root causes of these readmissions. When patients reported that they did not feel ready for discharge or requested their doctor to allow them to stay longer at discharge during index admission, they deemed their readmission more likely predictable and preventable. Doctors focused on measurable factors such as the clinical frailty scale and biomarkers during discharge process. Health-care worker failures were strongly associated with preventable readmissions.

CONCLUSIONS

There is no consensus between readmitted patients and treating doctors about predictability and preventability of readmissions, nor about associated risk factors and root causes. Patients should be more effectively involved in their discharge process, and the relevance of optimal communication between them should be emphasized to create a safe and efficient discharge process.

Frailty status at admission to hospital predicts multiple adverse outcomes

Aims

frailty is proposed as a summative measure of health status and marker of individual vulnerability. We aimed to investigate the discriminative capacity of a frailty index (FI) derived from interRAI Comprehensive Geriatric Assessment for Acute Care (AC) in relation to multiple adverse inpatient outcomes.

Methods

in this prospective cohort study, an FI was derived for 1,418 patients ≥70 years across 11 hospitals in Australia. The interRAI-AC was administered at admission and discharge by trained nurses, who also screened patients daily for geriatric syndromes.

Results

in adjusted logistic regression models an increase of 0.1 in FI was significantly associated with increased likelihood of length of stay >28 days (odds ratio [OR]: 1.29 [1.10-1.52]), new discharge to residential aged care (OR: 1.31 [1.10-1.57]), in-hospital falls (OR: 1.29 [1.10-1.50]), delirium (OR: 2.34 [2.08-2.63]), pressure ulcer incidence (OR: 1.51 [1.23-1.87]) and inpatient mortality (OR: 2.01 [1.66-2.42]). For each of these adverse outcomes, the cut-point at which optimal sensitivity and specificity occurred was for an FI > 0.40. Specificity was higher than sensitivity with positive predictive values of 7-52% and negative predictive values of 88-98%. FI-AC was not significantly associated with readmissions to hospital.

Conclusions

the interRAI-AC can be used to derive a single score that predicts multiple adverse outcomes in older inpatients. A score of ≤0.40 can well discriminate patients who are unlikely to die or experience a geriatric syndrome. Whether the FI-AC can result in management decisions that improve outcomes requires further study.

Patients' and providers' perceptions of the preventability of hospital readmission: a prospective, observational study in four European countries

OBJECTIVES

Because of fundamental differences in healthcare systems, US readmission data cannot be extrapolated to the European setting: To investigate the opinions of readmitted patients, their carers, nurses and physicians on predictability and preventability of readmissions and using majority consensus to determine contributing factors that could potentially foresee (preventable) readmissions.

DESIGN

Prospective observational study. Readmitted patients, their carers, and treating professionals were surveyed during readmission to assess the discharge process and the predictability and preventability of the readmission. Cohen's Kappa measured pairwise agreement of considering readmission as predictable/preventable by patients, carers and professionals. Subsequently, multivariable logistic regressionidentified factors associated with predictability/preventability.

SETTING

15 hospitals in four European countries PARTICIPANTS: 1398 medical patients readmitted unscheduled within 30 days MAIN OUTCOMES AND MEASURES: (1) Agreement between the interviewed groups on considering readmissions likely predictable or preventable;(2) Factors distinguishing predictable from non-predictable and preventable from non-preventable readmissions.

RESULTS

The majority deemed 27.8% readmissions potentially predictable and 14.4% potentially preventable. The consensus on predictability and preventability was poor, especially between patients and professionals (kappas ranged from 0.105 to 0.173). The interviewed selected different factors as potentially associated with predictability and preventability. When a patient reported that he was ready for discharge during index admission, the readmission was deemed less likely by the majority (predictability: OR 0.55; 95% CI 0.40 to 0.75; preventability: OR 0.35; 95% CI 0.24 to 0.49).

CONCLUSIONS

There is no consensus between readmitted patients, their carers and treating professionals about predictability and preventability of readmissions, nor associated risk factors. A readmitted patient reporting not feeling ready for discharge at index admission was strongly associated with preventability/predictability. Therefore, healthcare workers should question patients' readiness to go home timely before discharge.

The HOSPITAL score and LACE index as predictors of 30 day readmission in a retrospective study at a university-affiliated community hospital

Introduction

Hospital readmissions are common, expensive, and a key target of the Medicare Value Based Purchasing (VBP) program. Validated risk assessment tools such as the HOSPITAL score and LACE index have been developed to identify patients at high risk of hospital readmission so they can be targeted for interventions aimed at reducing the rate of readmission. This study aims to evaluate the utility of HOSPITAL score and LACE index for predicting hospital readmission within 30 days in a moderate-sized university affiliated hospital in the midwestern United States.

Materials and Methods

All adult medical patients who underwent one or more ICD-10 defined procedures discharged from the SIU-SOM Hospitalist service from Memorial Medical Center (MMC) from October 15, 2015 to March 16, 2016, were studied retrospectively to determine if the HOSPITAL score and LACE index were a significant predictors of hospital readmission within 30 days.

Results

During the study period, 463 discharges were recorded for the hospitalist service. The analysis includes data for the 432 discharges. Patients who died during the hospital stay, were transferred to another hospital, or left against medical advice were excluded. Of these patients, 35 (8%) were readmitted to the same hospital within 30 days. A receiver operating characteristic evaluation of the HOSPITAL score for this patient population shows a C statistic of 0.75 (95% CI [0.67–0.83]), indicating good discrimination for hospital readmission. The Brier score for the HOSPITAL score in this setting was 0.069, indicating good overall performance. The Hosmer–Lemeshow goodness of fit test shows a χ² value of 3.71 with a p value of 0.59. A receiver operating characteristic evaluation of the LACE index for this patient population shows a C statistic of 0.58 (95% CI [0.48–0.68]), indicating poor discrimination for hospital readmission. The Brier score for the LACE index in this setting was 0.082, indicating good overall performance. The Hosmer–Lemeshow goodness of fit test shows a χ² value of 4.97 with a p value of 0.66.

Discussion

This single center retrospective study indicates that the HOSPITAL score has superior discriminatory ability when compared to the LACE index as a predictor of hospital readmission within 30 days at a medium-sized university-affiliated teaching hospital.

Conclusions

The internationally validated HOSPITAL score may be superior to the LACE index in moderate-sized community hospitals to identify patients at high risk of hospital readmission within 30 days.

Assessing the Performance of a Modified LACE Index (LACE-rt) to Predict Unplanned Readmission After Discharge in a Community Teaching Hospital

BACKGROUND

The LACE index was designed to predict early death or unplanned readmission after discharge from hospital to the community. However, implementing the LACE tool in real time in a teaching hospital required practical unavoidable modifications.

OBJECTIVE

The purpose of this study was to validate the implementation of a modified LACE index (LACE-rt) and test its ability to predict readmission risk using data in a hospital setting.

METHODS

Data from the Canadian Institute for Health Information's Discharge Abstract Database (DAD), the National Ambulatory Care Reporting System (NACRS), and the hospital electronic medical record for one large community hospital in Toronto, Canada, were used in this study. A total of 3855 admissions from September 2013 to July 2014 were analyzed (N=3855) using descriptive statistics, regression analysis, and receiver operating characteristic analysis. Prospectively collected data from DAD and NACRS were linked to inpatient data.

RESULTS

The LACE-rt index was a fair test to predict readmission risk (C statistic=.632). A LACE-rt score of 10 is a good threshold to differentiate between patients with low and high readmission risk; the high-risk patients are 2.648 times more likely to be readmitted than those at low risk. The introduction of LACE-rt had no significant impact on readmission reduction.

CONCLUSIONS

The LACE-rt is a fair tool for identifying those at risk of readmission. A collaborative cross-sectoral effort that includes those in charge of providing community-based care is needed to reduce readmission rates. An eHealth solution could play a major role in streamlining this collaboration.

Predicting 30-Day Readmissions in an Asian Population: Building a Predictive Model by Incorporating Markers of Hospitalization Severity

Background

To reduce readmissions, it may be cost-effective to consider risk stratification, with targeting intervention programs to patients at high risk of readmissions. In this study, we aimed to derive and validate a prediction model including several novel markers of hospitalization severity, and compare the model with the LACE index (Length of stay, Acuity of admission, Charlson comorbidity index, Emergency department visits in past 6 months), an established risk stratification tool.

Method

This was a retrospective cohort study of all patients ≥ 21 years of age, who were admitted to a tertiary hospital in Singapore from January 1, 2013 through May 31, 2015. Data were extracted from the hospital’s electronic health records. The outcome was defined as unplanned readmissions within 30 days of discharge from the index hospitalization. Candidate predictive variables were broadly grouped into five categories: Patient demographics, social determinants of health, past healthcare utilization, medical comorbidities, and markers of hospitalization severity. Multivariable logistic regression was used to predict the outcome, and receiver operating characteristic analysis was performed to compare our model with the LACE index.

Results

74,102 cases were enrolled for analysis. Of these, 11,492 patient cases (15.5%) were readmitted within 30 days of discharge. A total of fifteen predictive variables were strongly associated with the risk of 30-day readmissions, including number of emergency department visits in the past 6 months, Charlson Comorbidity Index, markers of hospitalization severity such as ‘requiring inpatient dialysis during index admission, and ‘treatment with intravenous furosemide 40 milligrams or more’ during index admission. Our predictive model outperformed the LACE index by achieving larger area under the curve values: 0.78 (95% confidence interval [CI]: 0.77–0.79) versus 0.70 (95% CI: 0.69–0.71).

Conclusion

Several factors are important for the risk of 30-day readmissions, including proxy markers of hospitalization severity.

Roles of disease severity and post-discharge outpatient visits as predictors of hospital readmissions

Background

Risks prediction models of 30-day all-cause hospital readmissions are multi-factorial. Severity of illness (SOI) and risk of mortality (ROM) categorized by All Patient Refined Diagnosis Related Groups (APR-DRG) seem to predict hospital readmission but lack large sample validation. Effects of risk reduction interventions including providing post-discharge outpatient visits remain uncertain. We aim to determine the accuracy of using SOI and ROM to predict readmission and further investigate the role of outpatient visits in association with hospital readmission.

Methods

Hospital readmission data were reviewed retrospectively from September 2012 through June 2015. Patient demographics and clinical variables including insurance type, homeless status, substance abuse, psychiatric problems, length of stay, SOI, ROM, ICD-10 diagnoses and medications prescribed at discharge, and prescription ratio at discharge (number of medications prescribed divided by number of ICD-10 diagnoses) were analyzed using logistic regression. Relationships among SOI, type of hospital visits, time between hospital visits, and readmissions were also investigated.

Results

A total of 6011 readmissions occurred from 55,532 index admissions. The adjusted odds ratios of SOI and ROM predicting readmissions were 1.31 (SOI: 95 % CI 1.25–1.38) and 1.09 (ROM: 95 % CI 1.05–1.14) separately. Ninety percent (5381/6011) of patients were readmitted from the Emergency Department (ED) or Urgent Care Center (UCC). Average time interval from index discharge date to ED/UCC visit was 9 days in both the no readmission and readmission groups (p > 0.05). Similar hospital readmission rates were noted during the first 10 days from index discharge regardless of whether post-index discharge patient clinic visits occurred when time-to-event analysis was performed.

Conclusions

SOI and ROM significantly predict hospital readmission risk in general. Most readmissions occurred among patients presenting for ED/UCC visits after index discharge. Simply providing early post-discharge follow-up clinic visits does not seem to prevent hospital readmissions.

International Validity of the HOSPITAL Score to Predict 30-Day Potentially Avoidable Hospital Readmissions

IMPORTANCE

Identification of patients at a high risk of potentially avoidable readmission allows hospitals to efficiently direct additional care transitions services to the patients most likely to benefit.

OBJECTIVE

To externally validate the HOSPITAL score in an international multicenter study to assess its generalizability.

DESIGN, SETTING, AND PARTICIPANTS

International retrospective cohort study of 117 065 adult patients consecutively discharged alive from the medical department of 9 large hospitals across 4 different countries between January 2011 and December 2011. Patients transferred to another acute care facility were excluded.

EXPOSURES

The HOSPITAL score includes the following predictors at discharge: hemoglobin, discharge from an oncology service, sodium level, procedure during the index admission, index type of admission (urgent), number of admissions during the last 12 months, and length of stay.

MAIN OUTCOMES AND MEASURES

30-day potentially avoidable readmission to the index hospital using the SQLape algorithm.

RESULTS

Overall, 117 065 adults consecutively discharged alive from a medical department between January 2011 and December 2011 were studied. Of all medical discharges, 16 992 of 117 065 (14.5%) were followed by a 30-day readmission, and 11 307 (9.7%) were followed by a 30-day potentially avoidable readmission. The discriminatory power of the HOSPITAL score to predict potentially avoidable readmission was good, with a C statistic of 0.72 (95% CI, 0.72-0.72). As in the derivation study, patients were classified into 3 risk categories: low (n = 73 031 [62.4%]), intermediate (n = 27 612 [23.6%]), and high risk (n = 16 422 [14.0%]). The estimated proportions of potentially avoidable readmission for each risk category matched the observed proportion, resulting in an excellent calibration (Pearson χ2 test P = .89).

CONCLUSIONS AND RELEVANCE

The HOSPITAL score identified patients at high risk of 30-day potentially avoidable readmission with moderately high discrimination and excellent calibration when applied to a large international multicenter cohort of medical patients. This score has the potential to easily identify patients in need of more intensive transitional care interventions to prevent avoidable hospital readmissions.

Predicting 30-Day Readmissions: Performance of the LACE Index Compared with a Regression Model among General Medicine Patients in Singapore

The LACE index (length of stay, acuity of admission, Charlson comorbidity index, CCI, and number of emergency department visits in preceding 6 months) derived in Canada is simple and may have clinical utility in Singapore to predict readmission risk. We compared the performance of the LACE index with a derived model in identifying 30-day readmissions from a population of general medicine patients in Singapore. Additional variables include patient demographics, comorbidities, clinical and laboratory variables during the index admission, and prior healthcare utilization in the preceding year. 5,862 patients were analysed and 572 patients (9.8%) were readmitted in the 30 days following discharge. Age, CCI, count of surgical procedures during index admission, white cell count, serum albumin, and number of emergency department visits in previous 6 months were significantly associated with 30-day readmission risk. The final logistic regression model had fair discriminative ability c-statistic of 0.650 while the LACE index achieved c-statistic of 0.628 in predicting 30-day readmissions. Our derived model has the advantage of being available early in the admission to identify patients at high risk of readmission for interventions. Additional factors predicting readmission risk and machine learning techniques should be considered to improve model performance.