Development and validation of a core outcome measure for palliative care: the palliative care outcome scale. Palliative Care Core Audit Project Advisory Group.

Objectives: To develop an outcome measure for patients with advanced cancer and their families which would cover more than either physical symptoms or quality of life related questions. To validate the measure in various specialist and non-specialist palliative care settings throughout the UK.

Design: A systematic literature review of measures appropriate for use in palliative care settings was conducted. In conjunction with a multidisciplinary project advisory group, questions were chosen for inclusion into the scale based on whether they measured aspects of physical, psychological, or spiritual domains pertinent to palliative care, and whether similar items had shown to be valid as part of another measure. A staff completed version was developed to facilitate data collection on all patients throughout their care, and a patient completed version was designed to enable the patient to contribute to the assessment of their outcomes when possible. A full validation study was conducted to evaluate construct validity, internal consistency, responsiveness to change over time, and test-retest reliability. Assessments were timed.

Setting: Eight centres in England and Scotland providing palliative care, including inpatient care, outpatient care, day care, home care, and primary care.

Patients: A total of 450 patients entered care during the study period. Staff collected data routinely on patients in care long enough to be assessed (n = 337). Of these, 262 were eligible for patient participation; 148 (33%) went on to complete a questionnaire.

Main measures: The Palliative Care Outcome Scale (POS), the European Organisation for Research on Cancer Treatment, and the Support Team Assessment Schedule.

Results: The POS consists of two almost identical measures, one of which is completed by staff, the other by patients. Agreement between staff and patient ratings was found to be acceptable for eight out of 10 items at the first assessment. The measure demonstrated construct validity (Spearman rho = 0.43 to 0.80). Test/re-test reliability was acceptable for seven items. Internal consistency was good (Cronbach's alpha = 0.65 (patients), 0.70 (staff)). Change over time was shown, but did not reach statistical significance. The questionnaire did not take more than 10 minutes to complete by staff or patients.

Conclusion: The POS has acceptable validity and reliability. It can be used to assess prospectively palliative care for patients with advanced cancer.