Ascensión Doñate-Martínez

Application of screening tools to detect risk of hospital readmission in elderly patients in Valencian Healthcare System (VHS) (Spain)

The Sustainable Social and Healthcare Model (SSHM) is aimed to establish new care pathways in primary care systems contributing to the decrease of health services use and costs and improve the integration and efficiency of social and health care for elderly people with long-term care (LTC) needs. One of these strategies is the segmentation of population in risk groups through standardized tools. This paper is a retrospective study aimed to determine the viability of the implementation of the screening tools Probability of Repeated Admission - Pra - and The Community Assessment Risk Screen - CARS - to detect patients at risk of hospital readmission in a sample of 500 elderly people (65+) from the VHS in Spain. Patients were recruited from three Health Departments. Data from selected tools and predictive variables were collected through the healthcare database from the VHS. The most important results indicate that both instruments predict with high efficacy the proportion of patients not readmitted (negative predictive value between 91% and 92%). Moreover, the tools performed with a moderate efficiency being the Pra less sensitive (54%) and more specific (81%) than CARS (with a sensitivity and specificity of 64%). Results from this study suggest that the application of instruments as Pra and CARS are of interest to the Valencian Health Administration as they can be a good strategy to improve the management of elderly patients at risk with comorbidities and guiding clinical decision.

Aplicación de The Community Assessment Risk Screen en centros de atención primaria del Sistema Sanitario Valenciano

Resumen Objetivo Aplicar la herramienta The Community Assessment Risk Screen (CARS) para detectar pacientes mayores con riesgo de reingreso hospitalario y estudiar la viabilidad de su inclusión en los sistemas de información sanitaria. Diseño Estudio de cohortes retrospectivo. Emplazamiento Departamentos de salud 6, 10 y 11 de la Comunidad Valenciana. Participantes Pacientes de 65 años o más atendidos en diciembre de 2008 en 6 centros de salud. La muestra fue de 500 pacientes (error muestral = ± 4,37%, fracción de muestreo = 1/307). Mediciones Instrumento CARS formado por 3 ítems: diagnósticos (enfermedades cardiacas, diabetes, infarto de miocardio, ictus, EPOC, cáncer), número de fármacos prescritos e ingresos hospitalarios o visitas a urgencias en los 6 meses previos. Los datos procedían de SIA-Abucasis, GAIA y CMBD, y fueron contrastados con profesionales de atención primaria. La variable de resultado fue el ingreso durante 2009. Resultados Los niveles de riesgo del CARS están relacionados con el futuro reingreso (p < 0,001). El valor de la sensibilidad y la especificidad es de 0,64, el instrumento identifica mejor a los pacientes con baja probabilidad de ser hospitalizados en el futuro (valor predictivo negativo = 0,91; eficacia diagnóstica = 0,67), pero tiene un valor predictivo positivo del 0,24. Conclusiones El CARS original no identifica adecuadamente a la población con alto riesgo de reingreso hospitalario. No obstante, si fuese revisado y mejora su valor predictivo positivo, podría ser incorporado en los sistemas informáticos de atención primaria, siendo útil en el cribado y la segmentación inicial de la población de pacientes crónicos con riesgo de rehospitalización.OBJECTIVE Application of The Community Assessment Risk Screen (CARS) tool for detection of chronic elderly patients at risk of hospital readmission and the viability study for its inclusion in health information systems. DESIGN Retrospective cohort study. LOCATION Health Departments 6, 10, and 11 from the Valencia Community. PARTICIPANTS Patients of 65 and over seen in 6 Primary Care centres in December 2008. The sample consisted of 500 patients (sampling error=±4.37%, sampling fraction=1/307). VARIABLES The CARS tools includes 3items: Diagnostics (heart diseases, diabetes, myocardial infarction, stroke, COPD, cancer), number of prescribed drugs and hospital admissions or emergency room visits in the previous 6months. The data came from SIA-Abucasis, GAIA and MDS, and were compared by Primary Care professionals. The end-point was hospital admission in 2009. RESULTS CARS risk levels are related to future readmission (P<.001). The value of sensitivity and specificity is 0.64; the tool accurately identifies patients with low probability of being hospitalized in the future (negative predictive value=0.91, diagnostic efficacy=0.67), but has a positive predictive value of 0.24. CONCLUSIONS CARS does not properly identify the population at high risk of hospital readmission. However, if it could be revised and the positive predictive value improved, it could be incorporated into the Primary Care computer systems and be useful in the initial screening and grouping of chronic patients at risk of hospital readmission.

Application of the Community Assessment Risk Screen in Primary Care centres of the Valencia Health System

Resumen Objetivo Aplicar la herramienta The Community Assessment Risk Screen (CARS) para detectar pacientes mayores con riesgo de reingreso hospitalario y estudiar la viabilidad de su inclusión en los sistemas de información sanitaria. Diseño Estudio de cohortes retrospectivo. Emplazamiento Departamentos de salud 6, 10 y 11 de la Comunidad Valenciana. Participantes Pacientes de 65 años o más atendidos en diciembre de 2008 en 6 centros de salud. La muestra fue de 500 pacientes (error muestral = ± 4,37%, fracción de muestreo = 1/307). Mediciones Instrumento CARS formado por 3 ítems: diagnósticos (enfermedades cardiacas, diabetes, infarto de miocardio, ictus, EPOC, cáncer), número de fármacos prescritos e ingresos hospitalarios o visitas a urgencias en los 6 meses previos. Los datos procedían de SIA-Abucasis, GAIA y CMBD, y fueron contrastados con profesionales de atención primaria. La variable de resultado fue el ingreso durante 2009. Resultados Los niveles de riesgo del CARS están relacionados con el futuro reingreso (p < 0,001). El valor de la sensibilidad y la especificidad es de 0,64, el instrumento identifica mejor a los pacientes con baja probabilidad de ser hospitalizados en el futuro (valor predictivo negativo = 0,91; eficacia diagnóstica = 0,67), pero tiene un valor predictivo positivo del 0,24. Conclusiones El CARS original no identifica adecuadamente a la población con alto riesgo de reingreso hospitalario. No obstante, si fuese revisado y mejora su valor predictivo positivo, podría ser incorporado en los sistemas informáticos de atención primaria, siendo útil en el cribado y la segmentación inicial de la población de pacientes crónicos con riesgo de rehospitalización.OBJECTIVE Application of The Community Assessment Risk Screen (CARS) tool for detection of chronic elderly patients at risk of hospital readmission and the viability study for its inclusion in health information systems. DESIGN Retrospective cohort study. LOCATION Health Departments 6, 10, and 11 from the Valencia Community. PARTICIPANTS Patients of 65 and over seen in 6 Primary Care centres in December 2008. The sample consisted of 500 patients (sampling error=±4.37%, sampling fraction=1/307). VARIABLES The CARS tools includes 3items: Diagnostics (heart diseases, diabetes, myocardial infarction, stroke, COPD, cancer), number of prescribed drugs and hospital admissions or emergency room visits in the previous 6months. The data came from SIA-Abucasis, GAIA and MDS, and were compared by Primary Care professionals. The end-point was hospital admission in 2009. RESULTS CARS risk levels are related to future readmission (P<.001). The value of sensitivity and specificity is 0.64; the tool accurately identifies patients with low probability of being hospitalized in the future (negative predictive value=0.91, diagnostic efficacy=0.67), but has a positive predictive value of 0.24. CONCLUSIONS CARS does not properly identify the population at high risk of hospital readmission. However, if it could be revised and the positive predictive value improved, it could be incorporated into the Primary Care computer systems and be useful in the initial screening and grouping of chronic patients at risk of hospital readmission.