Maximize the value of patient data and engage patients with an analytics strategy
Treatment costs for people with one or more chronic conditions account for a staggering 86% of U.S. healthcare dollars, including $176 billion to care for the nation’s 29 million diabetics.1 Among patients treated for heart failure, 25% are re-hospitalized within 30 days and 50% within six months – creating an additional $17 billion in direct costs per year. New value-based reimbursement models that prompt providers to improve patient outcomes and deliver more cost-effective care are forcing clinicians to seek ways to improve the health of people with chronic conditions and minimize the need for expensive hospital care.
Big Data Challenges
The widespread adoption of mobile devices and wearable sensors as well as the growing use of remote monitoring have created new “big data” challenges, including challenges associated with the growing field of personalized medicine. It’s also driving a need for better analytics. The emergence of these new technologies holds great promise for improving the lives of chronically ill individuals and for reducing care costs. At the same time, the enormity and complexity of data can be overwhelming for clinical caregivers who want to leverage information to glean actionable insights that positively impact the care process.
Consider, for example, the traditional sources of data for analytics solutions: historical insurance claims, patient electronic health records and public health databases. While these sources are helpful, the data are retrospective, are not typically customized to the individual, and fail to incorporate real-time information. Data collected from remote monitoring systems creates a unique and personal profile but lacks the context of the patient history.
While the combination of retrospective and real-time data is ideal, it is not feasible for providers to actually track, much less correlate and interpret such a huge volume of data. In fact, unless an individual employs some sort of advanced machine learning technology, it’s virtually impossible to synthesize contextual data with all the data generated by monitoring systems.
To understand the health trajectories of people with chronic conditions, caregivers must incorporate an analytics strategy that eliminates data gaps, synthesizes information from multiple sources, and encourages people to engage in their own care.
Gaps in the Strategic Use of Analytics
The effective use of analytics has the potential to improve the quality of care and reduce costs. However, analytics initiatives often fall short of their potential because of gaps in strategy.
Data vs. Knowledge
Data is just data unless caregivers have the ability to interpret the data and take timely action. With remote monitoring, for example, providers need more than just data points; they also need to correlate findings with other patient data in order to understand their significance and communicate it effectively to the individual.
Engaging the Consumer
When it comes to optimizing outcomes, the consumer is arguably the most influential participant. The effectiveness of science-based medicine is thus dependent and limited by the consumer’s understanding of and engagement in his/her prescribed care plan.
For example, use of sensor technology outside of traditional healthcare facilities has made it easier to monitor an individual’s health. However, if a consumer is not engaged and adhering to his/her care plan, remote monitoring systems are of minimal benefit. In order to ensure the active participation of consumers, we must transfer the knowledge gleaned from collected data in a way that is meaningful to the individual.
Precision Medicine Through Machine Learning
Analytics strategies that leverage machine learning result in richer and more precise insights, especially when data is generated from multiple sources. With machine learning, data from an individual can be analyzed based on a population of one. Real-time data from sensor technology can be incorporated on an ongoing basis to assess everything from vitals, self-observed symptoms, activity levels, mood, etc.
This results in a higher degree of accuracy when predicting adverse events. Caregivers are thus able to customize their interactions with the individual to ensure productive measures are taken to stabilize an individual’s health.
Less Intrusive, More Engagement
The incorporation of machine learning into analytics also reduces the volume of required data. For example, people with diabetes are expected to test their blood sugar, inject insulin, and record their activities multiple times a day. They are forced to become their own data managers and make sense of any fluctuations in test results.
With the introduction of machine learning, the algorithms are far more advanced, requiring fewer data points and resulting in more accurate predictive values. For a person with diabetes, this reduces the testing and recording requirements and simplifies his/her life. The individual is thus more likely to remain engaged and committed to following the prescribed care plan.
Thanks to advances in technology, providers have access to more data than ever from across the full healthcare continuum. To maximize the value of this information, machine learning technologies can now simplify the management of data and facilitate more engaged consumers. When consumers are engaged, patients and caregivers collaborate to make more informed decisions, which ultimately drives better outcomes.
References
- Gerteis J, Izrael D, Deitz D, LeRoy L, Ricciardi R, Miller T, Basu J. Multiple Chronic Conditions Chartbook. AHRQ Publications No, Q14-0038. Rockville, MD: Agency for Healthcare Research and Quality; 2014. Accessed January 14, 2016.
