360-Degree Data Capture to Revolutionize Patient Care

Discover how self-reported lifestyle data gathered from wearables, social media and other sources can help identify potential behaviors and risks that predict disease progression or even early disease development.

Today, we are trying to understand the patient journey from a very narrow lens of data collected during health care interventions, which limits our ability to fully explain diverse outcomes. Self-reported lifestyle data gathered from wearables, social media and other sources can help identify potential behaviors and risks that predict disease progression or even early disease development. This can lead to timely and customized interventions that can not only provide better prognosis but can even prevent illnesses, which will reduce the burden of diseases on society.

The next holy grail in real-world data capture

Contemporary medicine is largely focused on decision-making based on data collection about sick patients when they are receiving treatment, whether as part of clinical trial data collection, the result of emergency room visits or through regular physician interactions.

Human disease pathways and outcomes, however, are defined by variables outside of clinical trials, hospitals and physicians’ offices. At present, there is no one easy way to capture all that relevant information. Most of the information that is currently captured comes through the clinical trials or health care system, through means including electronic medical records and claims data.

Accessing all the other data generated as people go about their daily lives is the “holy grail” for achieving more effective drug development, as well as more effective prevention and treatment of disease. Utilization of those data could lead to significant reductions in morbidity and mortality, as well as economic savings. Disease could be identified earlier, and behaviors and attitudes could be corrected sooner, which could reduce the severity of disease development.

The possible categories of informative data that could be collected are considerable. At this point, it is difficult to know which data points will be valuable or how much detail is needed. The goal is to use an approach similar to stepwise regression, starting with unconstrained data capture and honing in on the data points that are significant predictors of outcomes.

Obvious parameters for data collected

There are some obvious parameters that are known to be linked to general good health: exercise, movement and activity data, caloric intake and nutrition, sleep schedules, oxygen levels, and heart rate. These types of real-time data are already being collected by smart watches and other wearable devices across populations and captured in a fairly clean way for use in the development of outcomes models.

In fact, most exercise apps are increasing in sophistication and are able to provide valuable information. The same is true of home monitoring devices, such as those for blood pressure. For these parameters, there is a clearly visible path forward for systematically collecting data that is useful.

Less typical data sources

Other, less obvious data sources carry more noise and require more careful analysis. Examples include people’s buying habits — their consumer behavior in terms of food and other consumables. The potential is there to access very rich data that can, if separated from the noise, be integrated with traditional health care data.

Another potential source of behavioral — and attitudinal — data is social media. These data can help in the fine-tuning of subpopulations or phenotypes. For instance, by systematically examining social media posts, it is possible to follow how attitudes about the COVID-19 vaccines are driving behaviors. We have seen a divide around attitudes toward modern science, which can result in delayed diagnosis and hence poor prognosis or mortality, although public confidence in physicians remains higher than in researchers.

There are challenges, of course, with social media data because some people present a picture of themselves that they want others to see rather than one that is entirely honest, and negativity supercedes positivity. Separating real and projected personas is thus important. In some cases, it may be essential to collect social media in an unobtrusive and regulatory compliant manner while being cognizant of inherent limitation.

Managing the bias of wearables and electronic data

While personal wearables and other similar technology serve as an attractive source of personal electronic data generated outside of the health care setting, they tend to be used by healthier, highly motivated, engaged individuals who are more disciplined in following healthier lifestyles. Thus, they are likely to have better prognoses and outcomes.

Since most patient data collection is in the context of illness and the provision of treatment, data on healthier individuals are elusive, but nonetheless very valuable. The ideal situation, however, will require a broader subset of the population gaining interest in capturing these types of data. Before that can happen, the enabling technologies and devices must become affordable enough to reach the masses and become truly user friendly.

Insurance companies and governments also have roles to play in encouraging the personal tracking of health data. Payers could incentivize their policyholders to use wearables by offering free devices and discounts on policy costs. Once people are more aware that certain behaviors are being tracked, they are generally motivated to improve their performance. If issued a monthly report, they will likely watch their results. This can motivate people to make better choices; if they know they are being measured, they will always try to perform well. Once people get engaged with their health care, outcomes will be better, with significant reduction in morbidity, mortality and economic costs.

Potential to drive population-tailored health care policies

Collected patient-reported data generated outside of the health care setting can benefit not only the individual patient but entire populations. Those data can be aggregated and used for epidemiological analyses and policymaking. Attitudes (e.g., vaccine hesitancy) and prevalent conditions (e.g., diabetes, gastrointestinal cancers) in a population can be identified, and health care policies can be customized to address them.

To achieve this goal, of course, the key stakeholders in any health care situation — patients, providers, payers and policymakers (the four Ps) — must be aligned. People also need to make health care an electoral issue and be willing to invest in health care systems. Public officials must then be held accountable for establishing sufficient health care systems, from primary care physicians to hospital infrastructure. This is an issue primarily seen outside of Organisation for Economic Co-operation and Development (OECD) countries, but the COVID-19 pandemic has opened eyes to health care across borders. Progress will occur with increased patient involvement and when people understand that their actions today — taking 10,000 steps per day, for instance — will impact their future quality of life and security.

Ethical data collection and management are essential

Success in data collection of patient-reported non-health care information can only happen if the people collecting and analyzing the data are ethical and transparent about their activities. Without the highest level of commitment on these fronts, individuals will never trust the process. It’s necessary to provide some sort of contract that outlines how the data will be collected and treated and what value will be provided in return, such as reports that can help individuals monitor and improve their own health and well-being and ultimately impact the health of society as a whole.

Clinical trial data remains imperative

While self-reported daily life data are essential to establishing a complete picture of individuals when they become patients, collection of this type of data will not in the foreseeable future displace the need for clinical trials and clinical trial data collection, including the double-blind randomized controlled trials used for registration of candidate drugs.

No clinical studies are going to be able to address all these different types of data and possible confounders. Patient-reported non-health care-related data will be used to enhance clinical trials and supplement post-marketing studies. While randomized controlled trials will remain the gold standard to establish safety and efficacy, collecting and assessing patient data in the real world can facilitate studies investigating safety and efficacy in specific patient populations (including those with excluded comorbidities) that could not be included in trials, as well as comparative effectiveness studies. Real-world data can also be leveraged to increase the diversity of recruited patients and improve the representativeness of patient populations. After a new drug or medical device has been launched on the market, real-world data can be generated and collected from patients receiving treatment in a real-world setting, with no limitation on patient numbers and no criteria that can exclude participation. Having gene sequencing and lifestyle data captured and analyzed can allow for very precise tailoring of medicines, shifting drug development further from the blockbuster era and into personalized precision therapies.