Investigation of the causal relationship between patient portal utilization and patient's self-care self-efficacy and satisfaction in care among patients with cancer.

Objective: The objective of this study was to examine the causal relationship between the usage of patient portals and patients' self-care self-efficacy and satisfaction in care outcomes in the context of cancer care.

Methods: The National Institute's HINTS 5 Cycle 1-4 (2017-2020) data were used to perform a secondary data analysis. Patients who reported being ever diagnosed with cancer were included in the study population. Their portal usage frequency was considered as an intervention. Patient's self-care self-efficacy and satisfaction in care were the primary outcomes considered and they were measured by survey respondents' self-reported information. A set of conditional independence tests based on the causal diagram was developed to examine the causal relationship between patient portal usage and the targeted outcomes.

Results: A total of 2579 were identified as patients with cancer or cancer survivors. We identified patient portals' impact on strengthening patients' ability to take care of their own health (P = .02, for the test rejecting which is necessary for the expected causal relationship, ie, the portal usage impacts the target outcome; P = .06, for the test rejecting which is necessary for the reverse causal relationship), and we identified heterogenous causal relationships between frequent patient portal usage and patients' perceived quality of care (P = .04 and P = .001, for the tests rejecting both suggests heterogeneous causal relationships). We could not conclusively determine the causal relationship between patient portal usage and patients' confidence in getting advice or information about health or cancer care related topics (P > .05 for both tests, suggesting inconclusive causal directions).

Conclusions: The results advocate patient portals and promote the need to provide better support and education to patients. The proposed statistical method exploits the potential of national survey data for causal inference studies.