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

Jones, Stephen L.; Cheon, Ohbet; Manzano, Joanna-Grace; Park, Anne K.; Lin, Heather; Halm, Josiah; Baek, Juha; Graviss, Edward A.; Nguyen, Duc T.; Kash, Bita A.; Phillips, Robert A.

doi:10.1097/jmq.0000000000000035

Cited by 4 publications

(4 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Validated readmission prediction models are used in health care, such as the LACE index and HOSPITAL scores. 20 - 23 However, these models do not perform well in patients with cancer, as shown recently in a study by Jones et al, 48 where both LACE and HOSPITAL underperformed in patients with colon and lung cancers relative to the general population. This highlights the importance and the need for cancer-specific models.…”

Section: Discussionmentioning

confidence: 89%

Thirty-Day Unplanned Hospital Readmissions in Patients With Cancer and the Impact of Social Determinants of Health: A Machine Learning Approach

Stabellini

Nazha

Agrawal³

et al. 2023

JCO Clinical Cancer Informatics

View full text Add to dashboard Cite

PURPOSE Develop a cancer-specific machine learning (ML) model that accurately predicts 30-day unplanned readmissions in patients with solid tumors. METHODS The initial cohort included patients 18 years or older diagnosed with a solid tumor. Two distinct cohorts were generated: one with and one without detailed social determinants of health (SDOHs) data. For each cohort, data were temporally partitioned in 70% (training), 20% (validation), and 10% (testing). Tree-based ML models were developed and validated on each cohort. The metrics used to evaluate the model's performance were receiver operating characteristic curve (ROC), area under the ROC curve, precision, recall (R), accuracy, and area under the precision-recall curve. RESULTS We included 13,717 patients in this study in two cohorts (5,059 without SDOH data and 8,658 with SDOH data). Unplanned 30-day readmission occurred in 21.3% of the cases overall. The five main non-SDOH factors most highly associated with an unplanned 30-day readmission (R, 0.74; IQR, 0.58-0.76) were: number of previous unplanned readmissions; higher Charlson comorbidity score; nonelective index admission; discharge to anywhere other than home, hospice, or nursing facility; and higher anion gap during the admission. Neighborhood crime index, neighborhood median home values, annual income, neighborhood median household income, and wealth index were the main five SDOH factors important for predicting a high risk for an unplanned hospital readmission (R, 0.66; IQR, 0.56-0.72). The models were not directly comparable. CONCLUSION Key drivers of unplanned readmissions in patients with cancer are complex and involve both clinical factors and SDOH. We developed a cancer-specific ML model that with reasonable accuracy identified patients with cancer at high risk for an unplanned hospital readmission.

show abstract

Section: Discussionmentioning

confidence: 89%

Thirty-Day Unplanned Hospital Readmissions in Patients With Cancer and the Impact of Social Determinants of Health: A Machine Learning Approach

Stabellini

Nazha

Agrawal³

et al. 2023

JCO Clinical Cancer Informatics

View full text Add to dashboard Cite

show abstract

“…showed that the "LACE and HOSPITAL readmission risk-prediction algorithms performed well and equivalently in a heterogeneous and diverse population." 24,25 It is important to note that during the study period of 2014-2019, the observed average annual 30-day readmission rate in the study facility varied from 10.2% to 12.7% as compared with the observed national average of 18.4% in 2016, for example. 26 The low rates of readmission in the study facility likely reflect multiple interventions enacted to reduce readmissions, for example, increased use of intravenous therapies, increased on-site diagnostic interventions, and other systematic changes designed to increase the diagnosis and treatment of patients in place.…”

Section: Discussionmentioning

confidence: 93%

“…As shown in Figure 2, our study further indicates a greater risk of readmission within 30 days with rising LACE and HOSPITAL scores. This is supported by the 2021 study by Staples et al, which demonstrated that “mLACE+, LACE+ and LACE scores predicted hospital readmission with excellent calibration and adequate discrimination,” and the 2022 study by Jones et al, which showed that the “LACE and HOSPITAL readmission risk‐prediction algorithms performed well and equivalently in a heterogeneous and diverse population.” 24,25 It is important to note that during the study period of 2014–2019, the observed average annual 30‐day readmission rate in the study facility varied from 10.2% to 12.7% as compared with the observed national average of 18.4% in 2016, for example 26 . The low rates of readmission in the study facility likely reflect multiple interventions enacted to reduce readmissions, for example, increased use of intravenous therapies, increased on‐site diagnostic interventions, and other systematic changes designed to increase the diagnosis and treatment of patients in place 27 .…”

Section: Discussionmentioning

confidence: 98%

Increasing illness severity of skilled nursing facility patients over time: Implications for readmission penalties

Siskind,

Dhansew,

Burns

et al. 2023

J American Geriatrics Society

View full text Add to dashboard Cite

BackgroundCurrent financial penalties for rehospitalization of skilled nursing facilities (SNFs) patients are based in part on the studies by Ouslander et al., 2011, and Mor et al., 2010, demonstrating that many SNF hospitalizations were avoidable. With increasing age, complex illness severity, and use of SNFs for subacute rehabilitation, readmission metrics and financial penalties based on previous data may be due for reevaluation.MethodsRetrospective electronic medical record (EMR) review of 21,591 admissions and discharges between 2010 and 2019 inclusive. Data extracted included demographics, LACE, Charlson comorbidity index (CCI), and simplified HOSPITAL score parameters. The scores were calculated for the study years from the extracted data. Patients readmitted to the hospital within 30 days were identified.ResultsMean yearly score of all three indices rose steadily: LACE score 10.76–12.04 (0.43 estimated annual increase, 95% CI [0.39, 0.46]), CCI 4.26–5.05 (0.31 estimated annual increase, 95% CI [0.27, 0.34]), and simplified HOSPITAL score 3.46–4.03 (0.21 estimated annual increase, 95% CI [0.18, 0.24]). The estimated probability of readmission across observed CCI scores ranged from 15.4% to 15.9%, 95% CI bounds (10.8%, 22.7%). The estimated probability of readmission across observed LACE scores ranged from 4.7% to 36.3%, 95% CI bounds (3.4%, 54.7%). The estimated probability of readmission across observed HOSPITAL scores ranged from 5.8% to 54.1%, 95% CI bounds (6.2%, 66.0%).Conclusions and ImplicationsThe study confirms anecdotal experience that the illness acuity of patients admitted to SNFs increased progressively over time and was associated with an increased risk of 30‐day readmissions to the hospital. Our study suggests that the use of clinically validated readmission risk assessment tools instead of the Skilled Nursing Facility Value‐Based Purchasing Program (SNF VBP) current risk adjustors may be a more accurate reflection of the current illness severity of a facility's patient population at the time of payment adjustment.

show abstract

“…Ademais, são identificados estudos voltados exclusivamente para a análise da readmissão de coorte de pacientes com câncer, como por exemplo o câncer de cólon e pulmão (JONES et al, 2022) e de pacientes submetidos a procedimentos cirúrgicos decorrentes do câncer, como pacientes com tumor cerebral (CAPLAN et al, 2019;WINTER et al, 2020), câncer de bexiga (SALUK et al, 2018) e câncer colorretal (ALYABSI et al, 2022).…”

Section: Discussionunclassified

Modelo preditivo do risco de readmissão hospitalar no contexto brasileiro

Dias¹

View full text Add to dashboard Cite

Hospital readmission has been used as a health care outcome indicator. Considering that readmissions are considered avoidable, expose patients to unnecessary risks and represent excessive costs, it is relevant to identify patients with high risk of readmission. In this perspective, several predictive models have been developed, among which the LACE index (L -length of stay; A -acuity of the admission; C -Charlson comorbidity index score; e E -emergency department use) stands out as an easily applicable tool, structured in data frequently available in health services and with clinically relevant results. Objective: to analyze the predictive ability of hospital readmission by a score of identification of readmission risk in the Brazilian context. Methods: a retrospective analytical observational study was carried out in a general hospital with important insertion in the health care network. All hospital admissions whose discharges occurred in the period from January to December 2019 were considered eligible. The following variables were analyzed: age, age group, gender, main and secondary diagnosis, origin, length of stay, character of hospitalization, Charlson Comorbidity Index, emergency department visits, and LACE index. Analysis performed according to lines of care, conditions of interest and age groups.Demographic data and characterization of hospitalization were analyzed according to the outcome (readmitted or not readmitted), using the nonparametric Mann-Whitney test and the chi-square test. Analysis of the predictive capacity of the score was performed using the discrimination measure with calculation of the area under the curve of receiver operating characteristics. Results: 22,687 patients were included, of whom 13.44% were readmitted within 30 days. The median LACE index was 5 (IQR: 3.00; 8.00) and was higher for readmitted patients 6.00 (IQR: 4.00; 9.00), p<0.001. 12.86% of patients were classified as high risk for readmission and the index discrimination was 0.561 (CI: 0.549; 0.572). In the analysis of the LACE index across lines of care, conditions of interest and age groups, difference was observed between readmitted and non-readmitted for cardiovascular diseases (p=0.002), cancer (p=0.003), under 5 years (p<0.001), children and adolescents (p<0.001), adults (p<0.001), elderly (p<0.001). The discrimination by analysis group was 0.524 (p=0.209) for Primary Care Sensitive Conditions, 0.542 (p=0.374) for Diabetes Mellitus, 0.416 (p=0.244) for Systemic Arterial Hypertension, 0.548 (p=0.004) for Cardiovascular Diseases, 0, 644 (p=0.096) for Stroke, 0.525 (p=0.546) for Chronic Respiratory Diseases, 0.539 (p=0.003) for Cancer, 0.591 (p<0.001) for children under five, 0.593 (p<0.001) for children and adolescents, 0.543 (p<0.001) for adults, and 0.539 (p<0.001) for the elderly. Conclusions: the differences observed between the groups of readmitted and non-readmitted patients indicate that the LACE index helps to identify the risk of readmission, although its low predictive capacity prevents its isolated use. ...

show abstract

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

Cited by 4 publications

References 20 publications

Thirty-Day Unplanned Hospital Readmissions in Patients With Cancer and the Impact of Social Determinants of Health: A Machine Learning Approach

Thirty-Day Unplanned Hospital Readmissions in Patients With Cancer and the Impact of Social Determinants of Health: A Machine Learning Approach

Increasing illness severity of skilled nursing facility patients over time: Implications for readmission penalties

Modelo preditivo do risco de readmissão hospitalar no contexto brasileiro

Contact Info

Product

Resources

About