Patient-generated health data (PGHD) contains entire datasets which, contrary to electronic health records (EHRs), are collected based on patient input. This data may be directly entered or provided by patients (e.g. via quality of life questionnaires) or may be generated indirectly (e.g. via wearable devices).
Patient-Generated Health Data in A Shifting Healthcare Model
In healthcare, the biggest focus has traditionally been mostly around EHR data, since this is the foundation for various transactions (e.g. insurance companies or healthcare systems needing to bill for specific medical interventions). While this is gradually changing and the EHR data becomes more central to patient care, the way healthcare is being administered is also shifting: the average patient is spending less time in the clinic and is receiving care related services in many different settings (at home, via the pharmacy, in the community, etc). Therefore, the importance of patient-generated health data is growing rapidly.
There are opportunities to leverage PGHD throughout the patient journey and on various levels. A high-level description is presented below, made more concrete through the example of a cancer patient:
Self-care: A patient, especially if diagnosed with a chronic condition, is constantly faced with several challenges, including physical side-effects, psychological burden, social barriers and more. PGHD in this context includes self-reported adverse events, evaluation of mental wellbeing via questionnaires, physical activity data via fitness trackers, social interactions via communications devices, etc. Based on these datasets, the patient can be presented with self-care material, provided with information on relevant support organizations, referred to the clinic if necessary, and so on. For example, a cancer patient who has returned from a radiotherapy session but is also taking chemotherapy may take a picture of a rash that has developed in the irradiated area, and report severe nausea.
Remote patient monitoring: Similar to the above self-care setting, data can be collected and reported real-time to healthcare professionals in order to ensure proper monitoring and timely care. While this is not always recommended for resource utilization reasons, autonomous systems can evaluate the severity of each case and engage with a doctor or nurse as necessary. A patient’s weight and eating habits, for example, can trigger alerts for possible intervention due to malnutrition or even high risk of cancer cachexia, which may even be an irreversible condition.
Correlations: Shifting away from the real-time nature of interconnection, an intelligent system can perform various correlations among possible known risk factors or trigger points for specific actions. For instance, persistent bone pain could raise concerns for cancer metastasis to the bones. With the advent of artificial intelligence and machine learning, in particular, these correlations may be performed across factors that have not been identified in the past – leading to potentially interesting discoveries.
Real World Evidence (RWE): The industry, payers, and providers are increasingly looking for evidence and specific outcomes that are reported beyond a clinical trial setting. For RWE, patient-generated health data constitute a critical asset which is already being incorporated in many healthcare decisions (primarily through the form of patient-reported outcomes). For example, two treatments with the same medical efficacy are no longer viewed equivalent if they impose the different burden of disease to patients.
The benefits derived from identifying, collecting, reacting to, and processing patient-generated health data span, multiple stakeholders – a small list follows.
As described above, the patient stands to gain substantially through a large-scale quantification of their health status. Improved self-care, prompt interaction with healthcare professionals, elaborate warning systems, and a better understanding of the patient journey, are only some of the direct benefits for patients.
Healthcare professionals can also benefit from properly structured PGHD. More specifically, having access to each patient’s longitudinal data can not only help with their ongoing healthcare duties but can assist with capturing trends and correlations that are not immediately obvious.
Healthcare systems are already experiencing the value of PGHD. For example, in the telehealth and virtual care settings, where PGHD can assist with ongoing monitoring, the derived savings in costs and resource utilization have increased acceptance in many regulatory bodies. Moreover, value-based care is no longer weighed based on medical outcomes alone, but also on the long-term quality of life of its recipients – another great application for patient-generated health data.
Researchers with access to (anonymized) PGHD are finding numerous ways to process the data, perform analyses, and even form research hypotheses. In some cases, the depth and breadth of the data can lead to interesting applications of Artificial Intelligence as well.
The life sciences industry can also benefit. For example, a common “trial endpoint” in cancer is “Progression-Free Survival” (PFS) and “Overall Survival’ (OS). While the industry keeps making improvements through heavily researched medications, PGHD can shed light into the long-term quality of life of patients – revealing opportunities to provide “beyond the pill” services that can deliver additional value.
Finally, insurance companies are among the recipients of gains from patient-generated health data. Making more rational choices and streamlining operations can only be made possible through long-term engagement with the patient; this can only scale through structured PGHD. And while the fear of a “Big Brother” will never go away when it comes to health-related data, responsible and ethical approaches can be coupled with proper alignment of incentives.
If “data is the new oil”, what does that make “Patient Generated Health Data”?
The opportunities and benefits described above have already illustrated the value of PGHD. The main concepts driving some of the relevant business models to follow:
– Patient-oriented: While this is not always a relevant approach, expecting the patient to pay for a service based on their own data may be an option (for remote monitoring by a registered nurse, for instance).
– Data-oriented: Both life science and insurance companies stand to gain substantially from PGHD (especially related to specific conditions). Clearly, patient privacy and anonymity must be preserved, not only as per the established regulations but also abiding by the patient consent agreed upon.
– Access-oriented: While the value of PHGD is undisputable, emerging business models are being structured around the resulting access to individual patients. For example, a researcher may be interested to interview patients diagnosed with a specific tumor, receiving a specific medication, and reporting an intriguing combination of side-effects.
– Marketplace: Such a highly valued asset is also open to a marketplace model, where buyers and sellers contract for specific data points. Therefore, patients may actually benefit (beyond the above avenues) by selling their own data, and many blockchain-based initiatives have been building momentum on this tangent.
The Elephant in the Room: Privacy, Anonymity, and Respect for Patient Data
While patients tend to be generally more open to healthy citizens when it comes to data sharing, their data must be treated with extreme care. The legal framework for collecting and processing data, especially in a healthcare setting, is definitely of great importance. In Europe, the evolution of the Data Protection Act to the GDPR, and in other parts of the world, less centralized but similarly important trends, is marking a drastic shift into patients controlling their own data and the related privacy policies.
As a result, any entity aiming to process data must absolutely have the patients at the center and exercise extreme diligence in the way data processing is done. Moreover, it should not be an afterthought, but be embedded in the architecture of the system, the service, and its interfaces.
Therefore, and without being prescriptive, patients should be made aware of how their data is being used, in what form it is being shared and with what entities. Moreover, the entity collecting and processing the data should be very clear as to the patients’ rights; the General Data Protection Regulation (GDPR) has set a good framework for this.
Each of us generates enormous amounts of data every single day, including healthcare-related data. Capturing, structuring, and processing this data is part of an ongoing shift towards more data-oriented healthcare systems and services. Placing the patient at the center of all these processes is fundamental for identifying the opportunities and devising mechanisms to leverage them.
Cancer may be a prime context of such data, but not the only one. After all, being patient-centric means to recognize that patients often suffer from multiple conditions (co-morbidities). Many stakeholders, including healthcare systems and the industry, are gradually evolving to incorporate PGHD. It is unclear whether this will lead to a seamless mesh (or a siloed mess) of data, but the potential is enormous.
About the Author
Thanos Kosmidis is the Cofounder and CEO of Care Across, a digital health company focusing on cancer. It runs platforms with dynamic algorithms that provide patients with 100% personalized services based on peer-reviewed publications. The company collaborates with the industry around real-world evidence, patient engagement platforms, and clinical trial recruitment. Follow him on Twitter at @ThanosKosmidis or @CareAcross.