Identifying patients remains a challenge. Research has shown that clinical judgment (i.e., the surprise question) is not accurate and becomes even less reliable with greater familiarity between patient and provider. Existing claims-based algorithms do not reliably predict palliative care needs. Rules-based algorithms designed specifically for mortality prediction fall short.
We were invited to share some of our lessons learned with patient identification. We shared that lists should identify those most likely to benefit (not just predicting mortality). This involves predicting impending preventable events, such as inpatient admissions and emergency department visits. Use of more than just claims data is crucial. Efforts should be made to reduce false positives and eliminate those that clinical teams already know about. The list must be seamlessly integrated into clinicians’ workflow and actually lead to enrollment in the program. Otherwise it’s just another thing getting in clinicians’ way.
Engaging clinicians and training them in assessment and serious illness conversations, among other needed skills. There aren’t and won’t be enough specialty palliative care clinicians to meet demand. Just like we have come to embrace patient safety being everyone’s responsibility in health care, it is imperative that all clinicians see themselves as having a role in delivering basic pain and symptom management, as well as effective communication with patients with serious illness. It is important that they have the training and comfort level to do so effectively. This includes everyone from nurses, social workers, and physicians in primary care clinics to care managers at health plans, to clinicians in specialty offices (cardiology, oncology, pulmonology, etc.). This clinical training is available through CAPC's online curriculum, with free continuing education credits and ABIM MOC points for physicians.