This resource reviews strategies to identify patients who are likely to require high-cost care.

• The high-cost population is very heterogeneous, and segmentation strategies can inform population subgroups and match appropriate care management strategies accordingly.
• This population can be segmented into three subgroups: patients with advanced illness, patients with persistent high spending, and patients with episodic high spending. The first two segments are good candidates for care management; however, it is difficult to predict the needs of the third.
• Common signs of patients with advanced illness include heart failure, cancer, coronary disease, increasingly frequent hospitalizations, and advanced age.
• Most patients with persistent high spending have chronic illnesses, and many have functional limitations as well as serious mental health and substance use diagnoses.
• There are two basic ways to identify enrollees for care management: using quantitative data (e.g., claims data) and using qualitative data (e.g., physician referral, patient-reported data). Evidence suggests that a hybrid approach is the most reliable.