30-day readmission rate высокий — predict at-risk patients

Ситуация: Hospital: 30-day readmission rate 18%. Medicare penalty kicks in > 15%. Need at-risk identification.

Hospital data. Discharge cohort 2K/мес. Readmission = повторная госпитализация в 30 дней. Penalty $X/year.

Доступные данные

• admissions: patient_id, admitted_at, discharged_at, primary_diagnosis
• patients: id, age, comorbidities
• readmissions: patient_id, admitted_at, prev_discharge_id

Задачи

• Decomposition: by diagnosis, age, comorbidities.
• Predictive model: at-risk score.
• Intervention strategy.
• Cost-benefit.

Все кейсы для подготовки → ← Все кейсы