Adoption of Patient Generated Health Data in Oncology: A Report from the NCCN EHR Oncology Advisory Group

Abstract

Introduction:

Collecting, monitoring, and responding to patient-generated health data (PGHD) are associated with improved quality of life and patient satisfaction, and possibly with improved patient survival in oncology. However, the current state of adoption, types of PGHD collected, and degree of integration into electronic health records (EHRs) is unknown. The PRO Workgroup sought to perform an assessment and provide recommendations for cancer centers, researchers, and EHR vendors to advance the collection and use of PGHD in Oncology.

Methods:

The NCCN EHR Oncology Advisory Group formed a Patient Reported Outcomes (PRO) Workgroup to evaluate these issues via a survey of Member Institutions. Questions were designed to assess the current state of PGHD collection including how, what, and where PGHD is collected. Additionally, detailed questions about governance and data integration into EHRs were asked.

Results:

Twenty-three of 28 Member Institutions responded to the survey. The collection and use of PGHD is widespread among NCCN Members Institutions (96%). The vast majority of centers (90%) embed at least some PGHD into the EHR although challenges remain as evidenced by 88% of respondents reporting the use of instruments not integrated. Forty-seven percent of respondents are leveraging PGHD for process automation and adherence to best evidence. Content type and integration touchpoints vary among the members, as well as governance maturity.

Conclusion:

The reported variability regarding PGHD suggests that it may not yet have reached its full potential for Oncology care delivery. As the adoption of PGHD in Oncology continues to expand, opportunities exist to enhance their utility. Among the recommendations for cancer centers is establishing a governance process that includes patients. Researchers should consider determining which PGHD instruments confer the highest value. It is recommended that EHR vendors collaborate with cancer centers to develop solutions for the collection, interpretation, visualization, and use of PGHD.

Introduction:

Collecting, monitoring, and responding to patient-generated health data (PGHD) are associated with improved patient satisfaction and quality of life in oncology.^1–7 They are also potentially associated with improved patient survival.^8,9 Many electronic health record (EHR) and patient engagement vendors have begun to enable remote monitoring and electronic PGHD collection. Healthcare organizations can now collect vast amounts of PGHD electronically with the potential to inform predictive models. This includes direct data entry by patients and their caregivers. An additional potential benefit of this evolution includes improvements in data quality to support care team coordination and clinical research.^10–13 For example, activity data from devices have the potential to inform performance status pre-operatively, or for clinical trials eligibility screening and monitoring. The science around using PGHD has also matured, with prior work done to identify best practices for statistical analyses in controlled research studies.¹⁴

Questions remain, however, regarding operationalizing PGHD management in the clinic, such as 1) the optimal amount and type of PGHD to collect, 2) standardizing and visualizing actionable PGHD¹⁵, 3) integrating PGHD within EHRs to support clinician workflow, 4) triaging appropriate alarm notifications to the care team, 5) mitigating patient “survey fatigue”, and 6) identifying which PGHD instruments confer realized benefits to patients being monitored. We report here on the current state of adoption and implementation of PGHD collection and monitoring across leading cancer centers in the U.S. We discuss current gaps and recommendations for cancer centers, researchers, and EHR vendors.

Methods:

The National Comprehensive Cancer Network (NCCN^®) formed a Patient Reported Outcomes (PRO) Workgroup in 2019 to assess the current state of PGHD collection at NCCN Member Institutions and provide recommendations for integrating PGHD into EHRs. The Workgroup is comprised of representatives from 11 of the 31 NCCN Member Institutions, who have expertise in EHR optimization. The PRO Workgroup operates under the auspices of the NCCN EHR Oncology Advisory Group which serves as a forum for Group Members to share challenges and innovative practices regarding the optimization of EHR systems.

The PRO Workgroup initiated its charge by developing a conceptual model for patient reported outcomes (Figure 1). For the purposes of this report we refer to the broader concept of patient-generated health data (PGHD),^13,16–18 a subset of which are PROs. PGHD consists of five structured features (collection method; content; setting; purpose; integration touchpoints) under the broader agency of governance. Governance of PGHD emerged as a necessary component of a mature health system’s approach to operationalizing the collection and monitoring of PGHD. We identified the following four key outputs from collecting and monitoring PGHD: improved patient outcomes; efficient hospital operations; increased EHR data quality and integrity; and enhanced patient safety. The PRO Workgroup utilized the PGHD conceptual model to guide the development of the surveys used in this report.

Figure 1:

Patient Generated Health Data Model

Our initial survey assessed the current state of PGHD collection at NCCN Member Institutions. It was piloted by five NCCN Member Institutions to ensure content accuracy and question clarity. In May 2019 the survey was distributed to all 28 (at the time) EHR Oncology Advisory Group members, using an electronic survey tool (Survey Monkey). Each institution designated a single respondent responsible for completing the survey and data were collected over a 4-week period.

Upon review of the initial survey responses, the PRO Workgroup requested a follow-up survey to gain a greater understanding of “point-of-care” collection, EHR integration of validated survey instruments, barriers to EHR integration, and governance structures. This second survey was administered in August 2019 and response data were collected over a 4-week period through Survey Monkey.

Data from the surveys were presented to the NCCN EHR Oncology Advisory Group on October 21, 2019 and approved for publication.

Results:

The full results are available in the online appendix.

Survey 1 - Patient Generated Health Data

Twenty-three of twenty-eight NCCN Member Institutions responded to the survey (82% response rate). Some survey respondents did not answer all questions and, therefore, the total number of responses varied slightly per question.

Collection:

Of the respondents, 96% reported that their institutions collect PGHD, 95% use more than one method of collection, and 57% utilize three or more collection methods. The type and frequency distribution of PGHD collected is shown in Figure 2. The most common purposes for PGHD collection included active symptoms or events monitoring (86%) and screening (81%). The most widely used instrument to collect PGHD at NCCN Member Institutions was the Patient Healthcare Questionnaire (PHQ-2 or PHQ-9) (85%). Eighty-one percent of respondents indicated PGHD is collected within the research setting and 57% in the standard of care setting. Sixty-seven percent indicated PGHD is collected at point-of-care and 62% reported the data is collected remotely.

Figure 2:

Type of PGHD Collected at NCCN Member Institutions (n=21)

Integration:

Ninety percent of responding centers embed at least some PGHD into the EHR, while only four responding centers indicated that all PGHD is included in the EHR. Seventy-one percent reported that the center’s clinical decision support system alerts the care team about concerning responses or scores above or below a threshold. Sixty-seven percent reported that their care teams receive referrals or activities assigned to them based upon PGHD responses.

Governance:

Responses regarding the degree to which centers have a governance process are shown in Figure 3. In each of the three governance areas, only 33% to 43% reported having a fully implemented governance process (Figure 3). Forty-eight percent of responding centers reported that patients participate in the design of PGHD collection methods and processes.

Figure 3:

The degree to which centers have a governance body or process (n=21)

Survey 2 - Patient Generated Health Data Follow-up

This survey was distributed to the twenty-three Member Institutions that completed Survey 1 and 17 responded. Some survey respondents did not answer all questions and, therefore, the total number of responses varied slightly per question.

Collection:

Respondents indicated that point-of-care collection of PGHD occurs primarily in the outpatient waiting room (94%), but also in the outpatient exam room (41%), in the outpatient infusion room (18%), and one center collects PGHD in the inpatient room (6%) at point-of-care. Centers varied in the percent of PGHD that is collected at point-of-care versus remotely/at home with 75% of responding centers collecting 60% or more of PGHD at the point-of-care, and three indicating that 100% of PGHD is collected at point-of-care.

Integration:

Most centers indicated that PGHD are integrated into their EHR via different methods across their center and several centers used multiple methods. Eighty-eight percent of centers reported that some PGHD electronically integrates into the EHR as discrete data, however, 29% of those centers also have some PGHD pushed into the EHR as consolidated information in one document. Additionally, 53% of responding centers still scan paper documents containing PGHD into the EHR (without discrete or summary integration). While almost half (47%) indicated PGHD triggers automatic actions in the EHR based on thresholds, 88% of centers also utilize validated PGHD instruments that are not integrated into the EHR.

Governance:

Responding centers described PGHD governance structures that vary widely and feature unique staffing models. Examples include the following: departmental governance led by administrative and physician leaders; oncology stakeholder governance group lead by physicians, directors, disease line administrators, registered nurses and information technologists; oversight/advisory committee staffed by executive operations and research leadership with additional multi-disciplinary membership. One center reported efforts to decentralize a steering group while another reported efforts to centralize for an entire university health system.

Discussion:

NCCN has conducted the first assessment of the state of the science regarding PGHD collection and management across its Member Institutions. This study extends prior work by Zhang et al¹⁹ in 2015 with additional insights, both for uptake and integration. Our study focused on a different population of Oncology centers (NCCN, n=28) where Zhang et al focused on Quality Oncology Practice Initiative (QOPI) groups (11 of 28 NCCN Member Institutions are QOPI-certified). Like Zhang et al, we confirmed that cancer centers are utilizing different types of PGHD (not just symptoms or outcomes data), thus forming the basis of our recommendation for a more inclusive conceptual model for PGHD that is not limited to PROs alone (Figure 1). We also observed a higher percentage of cancer centers collecting PGHD (96% vs. 69%). This may reflect temporal trends since 2015, or more rapid uptake among comprehensive cancer centers specifically, or other factors not measured in this analysis. We observed a different distribution of survey types, including all the types reported by Zhang et al, plus social history, family history, social determinants of health, medications, preferences and values, implanted devices, and images. Our study demonstrated a higher rate of collection electronically, and at home in between visits. Our extended model of PGHD (Figure 1) could be of value to various standards organizations and researchers who are working to draw insights from PGHD, or for comparative effectiveness studies.

Our major finding is that adoption of the collection of PGHD is widespread among survey respondents (96%). NCCN Member Institutions currently collect a wide variety of PGHD content types, and it varies by institution (Figure 2). Ninety percent of respondents reported inclusion of at least some of their data into the EHR, though methods to store and visualize PGHD in the EHR varied from discrete data to scanned documents, and varied within each respondent’s center based on what PGHD was being collected. Challenges remain as evidenced by 88% of respondents reporting the use of numerous instruments that are not integrated into the EHR and only four of the responding centers having all PGHD included in the EHR. Given that the evidence for PGHD collection is potentially associated with improved survival from Basch et al^8,9 only if also monitoring and responding to these data, how one ingests and visualizes PGHD may be important. Subjective comments suggest factors such as copyright regulations, competing priorities, and technical challenges impede centers’ ability to integrate all PGHD into their commercial EHR.

A significant percentage of institutions (47%) are leveraging this data for process automation and adherence to best evidence, including generating automated EHR referral orders from the data (such as for psychosocial referrals for positive PHQ-2 or PHQ-9 scores, or smoking cessation). This has the potential to add the patient as an active participant in the care team’s efforts to adhere to best practice, and to reduce clinical burden through automated triage and referrals (though we recognize this could also increase unwanted referrals from the patients’ perspectives).

There is wide variability in the maturity and composition of PGHD governance models as is shown in Figure 3. Even at comprehensive cancer centers that have widely adopted PGHD, the majority of respondents self-reported that they have not yet fully implemented a governance structure to determine what to collect, how to incorporate PGHD into clinical workflows, and identify secondary use cases. Less than half responding centers include patients in the governance process.

The reported variability in data management, EHR integration, and governance suggests that PGHD may not yet have reached its full potential for advanced Oncology care delivery. In narrative comments, centers described oncology-specific challenges they have encountered, such as intra-departmental agreement, standardization, governance, personnel support, communication with and identification of eligible patients, volume and timing of surveys, integration into the clinic workflow, display format in EHR, interpretation of results, and ensuring clinical follow-up of PGHD.

These findings represent opportunities for the development of new knowledge regarding required competencies that organizations should employ as they continue to develop and expand their collection and utilization of PGHD. Recommendations for how cancer centers, researchers, and EHR vendors can optimize the collection, integration, and visualization of PGHD are summarized in Table 1.

Table 1.

PGHD recommendations for cancer centers, researchers and EHR vendors

Domain	Patient Generated Health Data (PGHD) Recommendations
Cancer Centers	1. Establish a PGHD governance process20, with a focus on clinical operational value
	2. Include patients in the governance process.
	3. Advocate for enhanced data standards to support comparative Oncology research, including adding key PGHD survey concepts to emerging pan-cancer standards within OMOP/OHDSI21 and mCODE (http://standardhealthrecord.org/guides/mcode/)
Researchers	1. Determine which PGHD instruments confer the highest value for quality of life, patient satisfaction, avoiding survey fatigue, and survival22 (Note: We anticipate the results of three ongoing NCI IMPACT Moonshot initiative studies evaluating impact of ePROs on symptom management and hospitalization will also provide insight into optimal implementation strategies.)
	2. Conduct feasibility assessments for workflow and systems integration to identify best practices. (Note: The Centers for Medicare and Medicaid Services (CMS) has launched a program in this regard which will reveal helpful insights in Oncology.23,24 Additional information is forthcoming from the eSyM (aka SIMPRO) project.25,26)
	3. Identify solutions to avoid worsening the digital divide.27
	4. Design optimal PGHD visualizations at the point-of-decision for care teams, including which clinicians are best suited to review PGHD, are open cognitive support questions for Informatics experts.
	5. Determine the incremental impact of engaging patients’ caregivers
	6. Conduct user-centered design research for optimal patient and care-giver experience when interacting with such digital tools and how they understand their role in their care.
EHR Vendors	1. Develop solutions in EHRs and portals for the easy collection, interpretation, visualization, and use of PGHD, especially for the use of already validated instruments, with the ability to move the data into notes, flowsheets, and decision rules.
	2. Employ design thinking in the engineering of visualization tools of PGHD, given its potentially varied types and potentially longitudinal collection.
	3. Collaborate directly with cancer centers where PGHD adoption is well established as is demonstrated in this study, including advocating for enhanced PGHD data standards.

A limitation of our study was not fully assessing the current state of how PGHD is used in clinical care. Although we did ask centers if the EHR triggers alerts and referrals in response to the collected PGHD, further study is needed to understand other ways PGHD is utilized in clinical care, including whether it is used to facilitate shared decision making between providers and patients. We also did not obtain the technical or operational specifics regarding how those centers that have had the most success with PGHD integration have accomplished this, meriting further analysis. Additional limitations include not assessing where PGHD responses are being recorded within the EHR (as part of the billable provider note, a separate document, or some combination of both). Furthermore, investigation is needed to understand how centers fund and support their PGHD collection and monitoring activities.

Conclusion:

The collection and use of PGHD is widespread among NCCN Member Institutions. Content type and integration touchpoints vary among the members (Figure 2). Governance maturity varies among NCCN Member Institutions (Figure 3). As the adoption of PGHD in Oncology continues to expand, opportunities exist to enhance the collection and monitoring process, standardize PGHD instruments, integrate PGHD into vendor designed EHR workflows, develop governance structures, and maximize the state of science around realizing value from PGHD.