Evidation White Paper Filling Gaps in RWD.pdf
Filling gaps in real-world data (RWD)
New approaches to incorporating meaningful data from everyday life to drive greater value across the product development and commercialization lifecycle
!
Table of Contents
→ Introduction
→ How to build a more complete picture of health and the lived experience
→ How to provide more value to healthcare stakeholders and decision-makers
→ How to generate more robust data sets by engaging the right people at the right time
→ How to increase ROI across the lifecycle and organization
→ Gain higher-quality data while sharing meaningful insights
Introduction
Electronic health records (EHRs) and claims often fail to fully show how treatments work and diseases progress in everyday life, as well as real-world factors impacting people’s health and health-related decisions. To fill this gap, life science companies are increasingly incorporating data collected directly from individuals in their day-to-day lives into their real-world evidence (RWE) strategies.
Person-provided ePROs, surveys, and continuous objective data (e.g., heart rate, sleep, mobility) from devices and wearables enhance real-world data sets with a more comprehensive and granular view of the following:
- Disease progression, symptoms, and treatment impact in day-to-day life
- Subsegments by disease, treatment, and lifestyle characteristics
- Quality of life (QoL) and activities of daily living
- Behavioral, attitudinal, and socioeconomic factors influencing outcomes
- Barriers to diagnosis, treatment initiation, and adherence
- The value of therapy from a health economics and outcomes perspective
Data from these sources can also create more representative data sets by engaging underserved and underrepresented populations who may not or cannot regularly access traditional healthcare services.
How to build a more complete picture of health and the lived experience
The FDA and payers are increasingly demanding more representative and person-centric data to generate RWE that describes the whole person and their lived experience with their health or disease/condition. This more complete picture provides benefits in terms of better health outcomes as well as informing decision-making across the product development and commercialization lifecycle.
How different data sets work together to describe the whole person within the context of their daily life
| EHR and EMR | Claims | ePROs, surveys, objective digital measures from wearables and devices (e.g., sleep, mobility, heart rate) | |
|---|---|---|---|
| Health profile | ○ | × | √ |
| Treatments | × | ○ | √ |
| Condition progression and quality of life | ○ | × | √ |
| Healthcare visits and costs | ○ | √ | ○ |
| Demographics, lifestyle, and social determinants of health | ○ | ○ | √ |
| Individual perspectives | × | × | √ |
| Frequency | Episodic | Episodic | Continuous |
Digital data from wearables and other sensors are playing an increasingly important role in life sciences as well as healthcare. According to the Biomarkers, EndpointS and other Tools (BEST) glossary developed by the FDA and NIH Biomarker Working Group, a biomarker is a “defined characteristic that is measured as an indicator of normal biologic processes, pathogenic processes, or biological responses to an exposure or intervention.” This was further extended to digital biomarkers by the FDA, as “a characteristic or set of characteristics collected from digital health technologies that is measured as an indicator…” This definition captures the ability to use information from one or more digital health technologies simultaneously to derive one or more biomarkers — and to provide the context needed to enrich a single data point.
In clinical research, RWD from sensors and ePROs have been used to deliver interventions for participants with respiratory diseases, neurological disorders, cancer, mental health issues, and more. They have also been used to measure outcomes in clinical research for Parkinson’s disease, physical activity programs, diabetes, and more.
Digitally derived data facilitate the understanding of the true disease burden, unmet needs, and treatment impact outside of traditional care settings by:
- Adding a level of objectivity that can be missing from retrospective, self-reported data.
- Minimizing the white coat syndrome that can be present in clinic-based measurements.
- More closely reflecting meaningful functional abilities or responses to treatment in real life.
- Maximizing the ability to reach a more inclusive, representative group of people outside of the clinic walls.
How digital measures can augment the information found in traditional data sources
| TYPE OF INFORMATION | TRADITIONAL DATA SOURCE(S)* | DIGITAL MEASURES FROM WEARABLES AND OTHER DEVICES |
|---|---|---|
| Mobility | Clinician observation, 6-minute walk test (6MWT), Retrospective patient or caregiver report, Current medication(s) | Activity sensors (gait speed, daily activity levels) |
| Pain/stiffness | Retrospective patient report, Written daily diaries, Current medication(s) | Activity sensors (number of afternoon naps, time of the first step) |
| Heart health | Episodic, clinician-recorded heart rate and blood pressure, ECG conducted in clinic, Current medication(s) | Continuous heart rate sensor, At-home wireless blood pressure monitor |
| Diabetes management | Written record of blood glucose levels, HbA1c measured at the lab, Written food and activity diaries, Current medication(s) | Wireless blood glucose monitor, Wireless data collected from insulin pumps, Electronic diaries with dropdown menus |
| Neurodegenerative movement disorders | Visual assessment and scales for tremor and bradykinesia, “On/off” diaries, Current medication(s) | Accelerometers (tremor and bradykinesia), Activity sensors (patterns of activity in relation to medication), Medication trackers |
| Sleep | Retrospective patient report, Written daily sleep diaries, Current medication(s) | Activity sensors (sleep duration and sleep disturbances), Pulse oximetry (breathing patterns during sleep) |
| Fatigue | Retrospective patient report, Current medication(s) | Activity sensors (number of afternoon naps, time of the first step) |
CASE STUDY
How ePROs and surveys filled gaps in understanding the lived experience with obstructive sleep apnea
A program for excessive daytime sleepiness (EDS) related to obstructive sleep apnea, a partnership between Evidation and Jazz Pharmaceuticals
Challenge
EDS is common for people with obstructive sleep apnea (OSA) and can persist despite positive airway pressure (PAP) therapy.
Data collected
An app-based survey was conducted with 2289 participants to collect:
- Epworth Sleepiness Scale (ESS)
- PAP use
- Satisfaction with care
Insights
- EDS was common in this real-world population with OSA, even among highly adherent PAP users.
- PAP adherence was associated with higher patient satisfaction.
- EDS was associated with lower patient satisfaction.
How to provide more value to healthcare stakeholders and decision-makers
Life science companies have an opportunity to use person-generated RWD to enhance the evidence and value they provide to other healthcare stakeholders, including healthcare providers, government organizations, non-profit organizations, regulatory agencies, payers, and the individuals participating in research.
By identifying the types of data that are useful to each of these groups, companies can share the stakeholder-specific value around the medical product of interest. Data relevancy as well as the format, amount, and timing of data sharing are important to minimize the effort to use the data. Engaging with these stakeholder groups early in the study design process will help identify which data points are of interest and how those could be captured.
The value person-generated RWD provides for each stakeholder group
| DATA NEEDS | ROLE OF PERSON-GENERATED RWD | |
|---|---|---|
| Individuals | Monitor one's own health, Facilitate conversations with health care providers | Education about conditions and risk factors for illness, Awareness of symptom relevance, Monitoring of outcomes during treatment |
| Healthcare providers | Use limited resources (e.g., time) more effectively, Monitoring of patient outcomes, Communicate more effectively with patients, Understand social determinants affecting health-related decisions | Disease management, Treatment decisions, Remote monitoring |
| Government and non-profit organizations | Monitor public health surveillance trends (i.e. Influenza, COVID, upper respiratory infections), Understand factors driving health care access | Observational cohorts, Disease surveillance |
Data collection
Collected permissioned data from Evidation’s digital health measurement and engagement platform using:
- Wearables and devices: activity, sleep, blood pressure, and symptoms
- Survey questions about health and experience over time
Outputs
- Personalized Heart Health report delivered to the patients
- Information presented in a way that could be used with their health care team
Participant engagement
Used the collected data to provide health event-driven alerts based on ACC guidance for symptom changes.
Outcomes
- 50% of the first 15,000 participants downloaded their personalized insights report.
- 87% of individuals who shared their report with their care team found it helpful or extremely helpful.
- Uncovered key findings around:
- Disease burden
- Adherence barriers (including behavioral and socioeconomic factors)
- Overall patient experience
How to generate more robust data sets by engaging the right people at the right time
Once the types of data have been chosen, the right people to provide those data need to be identified — at the right point in their health journey. Media ads and physician referrals to research can be hit-and-miss, attracting people who might not be eligible and missing entire segments of the population who do not receive (or potentially do not trust) the presented information.
Digital platforms house a rich data set of characteristics about their members, independent of their engagement with the traditional healthcare system and enabling efficient, precise identification of people eligible for participation. In addition, digital platforms provide low-burden methods of data collection, when and where it's convenient. When combined with personalized support, these data collection methods can serve as a powerful motivator to encourage further participation.
Adhering to privacy principles establishes trusted data exchange, instilling confidence for participants that their data are being used with the right intentions and for healthcare stakeholders that the collected data accurately and reliably reflect the individual’s situation.
How direct-to-patient research minimizes the data-sharing burden and enriches the experience
| TRADITIONAL DATA SOURCES | DIRECT-TO-PATIENT DIGITAL RESEARCH |
|---|---|
| Episodic data collection offers little opportunity to offer support for new or emerging symptoms or changing health. | Continuous tracking of health and symptoms can identify subtle changes and opportunities to provide education about actions to prevent more serious illness or side effects. |
| Relying on data from EHRs or claims records can delay opportunities for intervention. | Detection of health concerns in real time can trigger suggestions for timely interventions. |
| Manual person-generated data collection methods such as person-reported outcomes (PROs) are often burdensome. | Gamified rewards and challenges provide motivation to continue sharing data, and automated reminders can prompt completion of assessments. |
| Manual person-generated data collection methods such as PROs are often not easily personalizable to the person's situation. | Digital surveys can be tailored to the individual's characteristics, situation, and health status, ensuring only relevant questions are presented. |
| Entry into clinical trials or other supportive programs typically requires awareness of them by the health care team. | Opportunities to participate in related studies or access related care can be presented based on the person's characteristics and changes in health situation. |
CASE STUDY
How the FluSmart program drives long-term engagement through personalized insights and data-driven nudges to take action for influenza-like illness
Challenge
- Many individuals experiencing flu-like symptoms do not receive treatment.
- There is low awareness of what actions to take or which care options are available.
Overview
- Started in 2017, with 189,000+ enrolled in 2023.
- Longitudinal, permissioned data from wearables, smartphones, and surveys: symptoms, activity, sleep, heart rate.
- Machine learning flu algorithm developed in partnership with BARDA – detects meaningful changes in sleep and activity correlated with onset of ILI.
- Personalized flu insights and event-driven nudges to proactively seek care or clinical research opportunities.
Engagement
- 90% completion rate across content types.
- 55% response rate to predictive alerts of flu-like symptoms.
- 60% of survey completions within first 48 hours.
Results
- 20,000 instances of flu-like symptoms identified (October 2022 to January 2023).
- 1,100+ individuals referred to clinical trials.
Other outcomes
- Up to 2.7x greater precision in identifying individuals with flu-like symptoms with our predictive algorithm than general targeting.
- Key learnings on flu burden, experience, and barriers to care not visible in traditional data sets.
- Iterative improvements based on passive data collection and collaborations with government and biopharma organizations.
- Ability to implement this for emerging ILI outbreaks and to support development of vaccines and therapies for ILI such as respiratory syncytial virus (RSV).
How to increase ROI across the lifecycle and organization
The power of RWD to inform the product development lifecycle will only be realized if the data are appropriate, acceptable, and meaningful for all stakeholders. Therefore, the decision to include digital data and which data points to collect in research should not be arbitrary.
- Adopt a systematic, best practice approach to identify which RWD to collect.
- Lead with a person-centered perspective about what works within their life and constraints, rather than with a technology-first approach.
- Engage all stakeholders early to determine a “meaningful” measure of health (and therefore, endpoints that matter): one that detects a change in what matters to people in their everyday lives.
Achieving meaningfulness is important and requires contextualizing the data individuals provide. For example, for someone with chronic pain, any of the following could be considered important outcomes: less sleep disturbance, less fatigue, improved physical function, less pain at specific times of the day (e.g., first thing in the morning), and more. Compared with sleep diaries or in-clinic function assessments, activity sensors and other wearables provide objective, continuous measures of sleep and activity levels. Regularly completed ePRO assessments complement the objective data with the individual’s subjective view of the intervention’s impact.
When selected with intention, RWD points add specific value throughout development and commercialization
RWD helps measure pre-intervention function to better detect changes as the intervention progresses, facilitates the detection of patterns in a person’s life that affect the outcomes, and allows the individual to provide real-time feedback regarding whether a change in measured outcome had a tangible effect on their QoL.
How to address common life science data challenges
| CHALLENGE | SUCCESS STORIES |
|---|---|
| Understanding the daily burden of disease and the factors affecting patient outcomes | Sanofi: Activity trackers provided objective measures of headache burden in daily life; individuals slept more, had reduced physical activity, and had lower maximum heart rate on days with headache. As headache-specific impact on QoL increased, activity and maximum heart rate decreased and sleep increased. Headache days with higher self-rated health were associated with less napping, higher step count and maximum heart rate, correlating with increased activity. |
| Quickly finding the right participants for the right trials | BARDA: An algorithm used demographic, socioeconomic, and behavioral data from 128,629 people as well as predicted local COVID-19 prevalence data to generate risk scores for COVID-19 infection. These risk scores were used to identify individuals to follow up regarding COVID-19 infection. The follow-up of these high-risk individuals achieved a 4-to 7-fold greater incidence of COVID-19 infection than similar real-world study cohorts. |
| Moving individuals from awareness to action, to reduce the risk of illness that could exacerbate their condition | Sanofi: In a joint project regarding type 2 diabetes, the platform identified behavioral markers that could guide positive lifestyle change, such as influenza vaccination, to prevent complications in this high-risk group. |
| Identifying high-risk individuals who would benefit from earlier care and treatment | COPD: An algorithm used retrospective claims and wearable data to determine when someone with COPD was experiencing worsening symptoms that could require hospitalization. Monitoring passive data could identify and alert individuals to seek care before hospitalization is required. Asthma: The platform identified distinct differences in symptom control that were not captured by traditional RWE, providing opportunities to intervene when poorly controlled asthma is detected. |
| Continuously and reliably collecting surrogate endpoint data | Digital methods of remotely collecting data such as blood pressure and continuous glucose monitoring (CGM) for surrogate endpoints could provide more objective data than measurements collected during clinic visits (e.g., to avoid the white coat effect on blood pressure and overcome limitations with A1c). |
Gain higher-quality data while sharing meaningful insights
There is great potential to drive value for all healthcare stakeholders and increase ROI across the product lifecycle by incorporating person-generated RWD within the overall data collection strategy. An intentional approach can help identify low-burden methods that fill data gaps and provide meaning for the participants providing their data. When assessing the current RWE approach, consider the following questions:
→ Could person-generated data provide additional, useful information about symptoms, QoL, or treatment outcomes?
→ Are there lower burden methods of collecting endpoint data from participants?
→ Which endpoints provide greater meaning to participants, and can those endpoints be collected digitally?
→ What information has been previously requested by regulatory agencies or payers that could be augmented by person-generated data?
→ Could digital biomarkers help identify eligible participants, faster?
To learn how Evidation engages individuals in order to characterize and quantify disease progression, treatment experience, and quality of life, connect with a member of our commercial team.