Psychometric Properties of Patient-Reported Quality Measures for Community-Based Serious Illness Care

Quality measures from the Serious Illness Survey for Home-based Programs have been endorsed by the Centers for Medicare & Medicaid Services's consensus-based entity for use in initiatives to promote quality improvement and accountability of home-based serious illness care. However, no patient-reported quality measures have been endorsed for assessment of serious illness care across a range of outpatient settings and providers. To address this gap and create quality measures for use by Medicare Advantage (MA) organizations or other entities responsible for care of seriously ill individuals, we adapted and field-tested the survey among MA enrollees.

We refined the previously validated survey by removing questions specific to home-based services, adapting question wording, and adding new items to inform quality improvement. Following cognitive interviews among 20 seriously ill individuals to ensure consistent interpretation of survey items, we finalized the survey and field-tested it among a sample of seriously ill enrollees from a state-wide MA plan. Using the 1412 survey responses, we assessed item performance, used factor analysis to construct composite quality measures, evaluated item-scale correlations, and examined validity by calculating the degree to which quality measures predicted respondents' overall ratings of care.

The overall survey response rate was 41.5%. Cronbach's alpha estimates for proposed composite measures assessing communication, care coordination, help for symptoms, planning for care, and understanding own health ranged from 0.673 to 0.864, indicating adequate internal consistency in assessing their underlying constructs. Together, the composites explained 42.0%-44.3% of the variance in respondents' overall ratings of their care. Communication and care coordination were the strongest predictors of overall ratings.

Patient-reported measures derived from the Serious Illness Survey for Community-based Care can be used to inform quality improvement, monitor care over time, and assess the effectiveness of new initiatives for seriously ill individuals receiving care across a range of community settings.