Identifying patients with advanced illness for needs assessment and care planning: SPICT in practice

Background Systematic identification of patients with advanced illnesses is central to UK end-of-life policies. Recognising a limited prognosis fosters decision-making about future care but probabilistic predictions such as the ‘surprise question’ tend to be overoptimistic. Even in cancer, complex mathematical models are only 60% accurate and open to confounders. The 2009 Supportive and Palliative Indicators Tool (SPICT) used clinical indicators derived from a review of prognostic tools to identify patients likely to have advanced illness and unmet supportive and palliative care needs for assessment. Aims To pilot the SPICT tool in the renal, liver, cardiac and respiratory wards of a large teaching hospital and assess its utility in an acute hospital setting. Methods Each ward team developed a customised checklist containing the SPICT clinical indicators. All unplanned admissions to each ward over 2 months were screened and followed up for 3 months to record re-admissions; hospital bed days; time, cause and place of death. The Charleson index of co-morbidities was calculated for comparison. Findings The indicators were evaluated and refined by successive teams. Multi-morbidity predicted a limited prognosis, re-admission risk and complex needs but should take account of disease severity. Most, but not all those who died, had a poor performance status (PS). Patients may spend more time bed-bound when hospitalised so usual status at home is important as an acute deterioration might be reversible. Repeated hospital admissions are well documented but increased community care needs are not always apparent initially. Conclusions The SPICT performs as well as the Charleson index in identifying patients with a limited prognosis and complex needs. The revised version contains fewer, clearly identifiable indicators that provide a rationale for reviewing care goals with patients and planning care; such as hospital admission, poorly controlled symptoms, increasing dependency and robust descriptors for common illnesses.