Using Electronic Patient-Reported Outcome Monitoring to Navigate Patients to Supportive Care Services

Electronic patient-reported outcome (ePRO) symptom monitoring is rapidly expanding in oncology clinical care. Studies of ePRO symptom monitoring have demonstrated benefits for health systems and patients, including a reduction in hospitalizations, emergency visits, and cost of care.¹ Patients with advanced cancer who receive ePRO symptom monitoring have improved quality of life, better symptom control, and in some cases improved overall survival.^2–5 The process of completing ePROs enables patients to engage in self-reflection and empowers patients to communicate their needs with the clinical team. Recognizing these health and patient-centered benefits, the Centers for Medicare and Medicaid Innovation added gradual implementation of ePRO monitoring to the new Enhancing Oncology Model.⁶ Large multisite trials and consortiums are studying how to support PRO infrastructure and implement ePRO symptom monitoring in community and academic practices and in diverse patient populations nationally.^3,7–9

Basch et al¹⁰ have called for recognition of ePRO symptom monitoring as a digital therapeutic intervention that improves cancer outcomes. If ePROs are recognized as a therapeutic, this would allow for insurance reimbursement to support the costs of implementing and sustaining an ePRO program. As we think about implementation, what would an optimized digital therapeutics program look like? In a future state, patients could be monitored with PRO items that are tailored to their personal disease, cancer experience, and needs using adaptive assessments that customize which items patient receive. Prediction algorithms could run in the background to optimize alert generation and identify patients who are at risk for adverse outcomes. Importantly, ePROs could be used to screen and navigate patients to supportive services such as a palliative care, psychosocial care, and/or social work based on patients’ reported needs. Finally, the loop could be closed with data on symptom reports, subsequent clinical interventions, and patient outcomes fed back into a learning health system for quality improvement and measurement.

As we work toward the ideal state, it is critically important to collect clinically meaningful PROs and determine thresholds for action that positively affect patient outcomes. Thresholds for action could be a meaningful change in a PRO item or longitudinal trend in scores that predicts an adverse event, such as an emergency department (ED) visit or hospitalization. Thresholds for action could also be a PRO response that requires intervention, such as a severe pain score or a positive screen for financial toxicity. For routine ePRO symptom monitoring during cancer treatment, nurse outreach is triggered by severe thresholds or worsening trends in PRO items that have been linked to ED visits and hospitalizations.^11,12 Less is known about monitoring other PRO domains beyond symptoms.

This important question is addressed in the accompanying article, where Sikorskii et al¹³ take a significant step toward understanding how ePRO reports in clinically meaningful domains could be used to predict risk of an adverse event. The authors aim to establish thresholds in the patient-reported outcomes measurement information system (PROMIS) scores for pain interference, physical function, fatigue, and depression that are associated with unplanned health care utilization. They investigate thresholds that are associated with subsequent ED or urgent care use among patients with cancer 14 and 30 days after a PRO assessment. They find that pain interference and physical function are associated with subsequent ED or urgent care visits and that specific score cutoffs for these domains have a higher association with these types of visits. These results suggest that thresholds for PROMIS scores in pain interference and physical function, if validated prospectively, could be used as potential triggers for interventions that prevent unplanned ED visits or urgent care use.

Although the authors do not find significant associations with PROMIS depression or fatigue scores and unplanned health care utilization, we should not conclude that monitoring for these symptoms is less impactful. Depression and fatigue are clinically important symptoms for patients and families, and ones that we manage frequently in oncology practice. Meaningful score changes may vary by population and context, and depression and fatigue scores would likely be significant with a different patient-based or clinical anchor.¹⁴ For example, severe responses to depression and fatigue PRO items could indicate a threshold for a psychosocial or palliative care intervention that will improve a patient’s quality of life on cancer treatment.¹⁵ The negative finding for these PRO domains in the accompanying article highlights gaps in knowledge about how to use clinically significant PROs to predict risk or navigate patients to ancillary services.

When building a digital therapeutics program, PRO domains and items should be prioritized depending on the clinical action and the anchor. In the article by Sikorskii et al,¹³ the anchor is unplanned health care utilization within a short interval after a PRO report and the action could be an intervention to prevent ED or urgent care visits. What other anchors should we consider when selecting PRO domains for symptom monitoring? Anchors that demonstrate cost-savings, health care utilization, and other value-added benefits are often picked because they are relatively easier to quantify and demonstrate impact to payers. However, we should also strive to understand how ePRO monitoring interventions can positively affect other novel outcomes that are important to patients and clinicians, such as timely palliative care referral, equitable access to care, excellent care coordination, and clinical efficiency, among others. Clinician satisfaction with using ePRO data to guide clinical care should also be equally weighted with value-based outcomes.

With these other anchors in mind, how can we link PROs to action? As a use-case example, consider the action of using ePRO monitoring to navigate patients to supportive care services. PRO monitoring holds enormous potential as a tool to screen patients for unmet needs and connect patients with appropriate services during treatment by capturing multidimensional supportive needs that may be both underreported by patients and uniquely captured with direct patient report. For example, our patients wonder if their treatment is working, how their loved one is holding up, if they can afford the next scan, and if they can find meaning in life after a cancer diagnosis.¹⁶ These worries, fears, and hopes for the future fall into the domains of communication, caregiver burden, financial toxicity, coping, and existential distress. PRO measures exist for patients to report in these domains, but have not been routinely used in oncology symptom monitoring programs or linked to supportive care services.

The framework presented in Figure 1 presents an approach to improving health care outcomes by using ePROs to navigate patients to supportive care services. Imagine a scenario in which patients are screened across multiple care domains at standardized intervals and then navigated via structured care pathways to supportive services based on thresholds or triggers for severe reports. Routine ePRO symptom monitoring can be viewed as a primary palliative care intervention that improves symptom management by the oncology team.¹⁷ Extending this concept, ePRO symptom monitoring should be linked with specialty supportive care services to identify and treat patients with complex, refractory symptoms and unmet care needs. In this scenario, routine ePRO symptom monitoring could be expanded to screen patients for physical symptoms and additional needs that supportive care services address, such as coping, social support, and existential distress, among others. ePRO reports could also be used to triage patients to other ancillary services, such as social work, psychosocial care, and financial navigation. Using ePROs to navigate patients to a network of ancillary services on the basis of specific unmet needs would scale our ability to support patients during their cancer treatment since many of these services are resource-constrained.

FIG 1.

Framework for using ePRO monitoring to navigate patients to supportive care services. ePRO, electronic patient-reported outcome.

The use of ePRO tools to capture supportive care domains also has the potential to overcome our personal biases and enhance equity. For example, we may not ask a young patient with metastatic cancer about prognostic understanding because it makes us feel uncomfortable. Or we may not ask a patient with low health literacy about how they are coping because we spend more time explaining the medical treatment. Within an ePRO screening system, all patients would be screened for multidimensional care needs using standardized ePRO assessments with associated decision support for referral, potentially reducing referral bias and increasing screening efficiency. This may be even more important for vulnerable patients who face barriers to accessing high-quality cancer care, such as those who are uninsured, have low income, low health literacy, face the effects of systemic racism, or lack social support.

Excellent symptom management is a shared mission between oncology and supportive care services.¹⁸ Yet, ePRO symptom monitoring is frequently not linked to supportive care services, and this gap is a missed opportunity to deliver timely supportive care to patients during treatment. Let us invite our palliative care colleagues to the table in the development of a digital therapeutics program that facilitates screening for supportive care needs and is anchored on outcomes important to patients and clinicians.