BT

Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ

Topics

Choose your language

InfoQ Homepage Articles The Promise of Healthcare Analytics

The Promise of Healthcare Analytics

Bookmarks

This article first appeared in IEEE IT Professional magazine. IEEE IT Professional offers solid, peer-reviewed information about today's strategic technology issues. To meet the challenges of running reliable, flexible enterprises, IT managers and technical leads rely on IT Pro for state-of-the-art solutions.

 

“ In the next ten years, data sciences and software will do more for medicine than all of the biological sciences together.” That’s quite an assertion by venture capitalist Vinod Khosla. 1 The venture capitalists of the world and Silicon Valley entrepreneurs might all be excited about the promise of wearable technologies with biophysical interfaces, physiological sensors, and embedded diagnostic tools to measure “the quantified self,” which will then empower individuals to take control of their healthcare, improve the effectiveness of treatments, and “replace 80% of what doctors do.” 2

However, it’s one thing to have information, and it’s another to act on it. Patients are notorious for not following through on “doctor’s orders,” and many of the diseases caused by lifestyle choices are due to them avoiding what they should be doing (quitting smoking, exercising regularly, eating healthy foods, and so on). Following through on treatment regimens requires the missing discipline that contributed to the condition that the treatment is meant to remediate.

What Works in Treatment Adherence

Behavioral changes and adherence to medical treatment are better addressed through patientprovider communication than they are through an application. “Patient non-adherence is a physician-patient communication challenge—not a health information technology challenge,” according to Stephen Wilkins, a consumer health behavior researcher.4 He goes on to say that when doctors prescribe a new medication, they typically spend less than a minute explaining it to the patient: why it needs to be taken, how to take it, when to stop, and so forth.

“In healthcare, the actionability and effectiveness of data science hinge on communication between providers and patients, and on patients’ ability to act on those insights. There are a few methods of provider-to-patient communication and actionability” that are effective at getting patients to change behaviors, according to Kyle Samani, a veteran of the electronic medical records industry. 5

These mechanisms for encouraging change represent the virtual embodiment of things that help people make the day-to-day decisions that lead to long-term changes in unhealthy behaviors and habits—communication, education, reminders, and, in some cases, nagging spouses. Point-ofcare information from physician to patient is delivered in a very small window of time. This information needs to be reinforced with ongoing education from justin-time sources such as portals, mobile tips, and daily motivators.

A most interesting idea is for mobile sensors to detect behaviors in real time and offer alternatives, interventions, or, in an extreme example, an alert to a health coach or spouse to intervene and discourage a poor choice. (Google Glass identifying a cigarette about to be lit. “Leave me alone, Glass.”) The key is understanding context and offering the correct messaging and mechanisms when the patient is receptive and in need of the information, encouragement, motivation, or admonition. Context is determined and enabled by text analytics by correlating unstructured content consumption with measured outcomes through electronic health records, observational outcomes, and sensor data. Integrating and making sense of these diverse information sources, though challenging, will be an important innovation in medical informatics and healthcare analytics, and will become the critical missing element between healthcare intervention and outcome.

The Role of Analytics

Data analytics won’t reduce the need for physicians but will alter how they spend their time. In the future, doctors will spend less time in diagnosis because this function can be better accom- plished by computers that can absorb and analyze vast amounts of up-to-date medical information. Doctors and other medical professionals can then spend more time on patient education and providing the compassionate care they might have originally envisioned.

All of this data, whether it is obtained from sensors, communication, and reminders to patients, or from detailed notes taken by doctors at the point of service, provide a rich and growing repository of information that can be mined for a variety of purposes. Beyond diagnoses, analytics can be used for a vast array of purposes in healthcare, depending on the data that’s being analyzed, the hypotheses being developed, the framework for analysis, and the domain expert’s perspective. For example, the Controlled Risk Insurance Corporation (CRICO), a captive medical liability insurance company owned by and serving the Harvard medical community, analyzes malpractice insurance claims to identify high-risk patient populations, conditions, treatments, procedures, physicians, and contributing factors to improve the safety of medical procedures and treatments. 6

Another source of rich historical data across large patient populations is that captured by healthcare payers—the insurance companies and public health agencies that mine millions of claims each year for trends in service delivery, quality, efficacy, abuse, waste, fraud, and errors. The data holds great promise for outlier detection in healthcare services. The US Center for Medicare and Medicaid has just enacted policies to release claims data that until recently wasn’t publicly available. 7 (The American Medical Association received an injunction in 1979 that prevented the public from knowing how much taxpayer money individual doctors received from the Medicare program, which effectively closed this data off from analysis. 8 ) This data will allow for greater scrutiny of costs and provide visibility into unusual claim patterns that could be indicative of fraud. In 2011, more than US$4 billion in fraudulent healthcare payments was recovered, but that amount is a small fraction of the estimated total for fraudulent payments. Therefore, the incentive is high for such analyses. 9

In addition, data analysis could identify the most effective treatments for specific subpopulations at a more granular level than has been possible before.

Evidence-based medicine is a broad concept that applies descriptive, statistical, and analytical approaches to evaluating the efficacy of treatment through review of experimental (structured clinical trials) as well as analysis of unstructured observational data (typically, electronic health records). To a layperson, all healthcare might seem to be “evidence-based.” Isn’t medical science based on evidence?

The challenge is that though medicine is founded in science, the practice of medicine is considered an art based on science.10 Evidence-based medicine combines research approaches with clinical observations of treatments and outcomes. It also combines a variety of approaches for developing, disseminating, and implementing practices that are clinically appropriate and cost effective. 11

Analytics and big data approaches for dealing with large amounts of structured and unstructured heterogeneous data will help support evidence-based medicine by providing another analytical tool in the researcher’s and clinician’s toolkit. Other efforts hold great promise in helping to identify low-frequency adverse drug events through the analysis of observational medical data.

Personalized medicine is an important, emerging area of healthcare. The ability to personalize medical treatment is based on the data-intensive fields of pharmacogenomics, nutrigenomics, and pharmacoproteomics, 12 all of which use the understanding of the molecular behavior of bioactive molecules to develop advanced medical treatments.

Biological systems are variable, dynamic, complex chemical systems in which slight variations in an individual’s genomic makeup have significant implications with regard to a therapy’s effectiveness. Analysis of the data could identify the most effective treatments for specific subpopulations at a more granular level than has been possible before. Decoding the mechanisms of action of compounds and biologicals in this environment depends on researchers’ ability to model interactions with thousands of potential variables and millions of possible data points.

Tailoring treatment to individual needs based on genetic makeup will require highly sophisticated analyses—personalized medicine is mind bogglingly data-and analytics-intensive.

The healthcare system is undergoing tremendous change, and analytics plays a central role in improving outcomes and quality of life while helping to control costs. The next several years will see many new mechanisms and tools making significant contributions in all of these areas.

References

  1. F. Lardinois, “Vinod Khosla: In the Next 10 Years, Data Science Will Do More for Medicine than All Biological Sciences Combined,” Tech Crunch, 11 Sept. 2013.
  2. V. Khosla, “Technology Will Replace 80% of What Doctors Do,” Fortune, 4 Dec. 2012.
  3. A. Atreja, N. Bellam, and S.R. Levy, “Strategies to Enhance Patient Adherence: Making it Simple,” Medscape General Medicine, vol. 7, no. 1, 2005.
  4. S. Wilkins, “Patient Non-Adherence (Like Engagement) Is a PhysicianPatient Communication Challenge—Not a Health Information Technology Challenge,” Center for Advancing Health, 23 July 2013.
  5. K. Samani, “Unlocking the Power of Data Science In Healthcare,” EMR & HIPAA, 29 Jan. 2014.
  6. Malpractice Risks of Routine Medical Procedures,” CRICO press release, 17 Dec. 2013.
  7. Medicare Provider Utilization and Payment Data: Physician and Other Supplier,” Centers for Medicare and Medicaid Services, last updated Apr. 2014.
  8. Wall Street Journal Sues to Open Up Secret Medicare Database,” Dow Jones press release, 25 Jan. 2011.
  9. Efficient Use of Big Data Could Reduce Instances of Healthcare Fraud,” Govplace, 2012.
  10. S.C. Panda, “Medicine: Science or Art?” Mens Sana Monographs, vol. 4, no. 1, 2006, pp. 127–138.
  11. J. Belsey, What Is Evidence-Based Medicine? May 2009.
  12. V. Ozdemir et al., “Personalized Medicine Beyond Genomics: New Technologies, Global Health Diplomacy and Anticipatory Governance,” Current Pharmacogenomics and Personalized Medicine, vol. 7, no. 4, 2009, pp. 225–230.

About the Author

Seth Earley is CEO of Earley & Associates. He’s an expert in knowledge processes and customer experience management strategies. His interests include customer experience design, knowledge management, content management systems and strategy, and taxonomy development. Contact him at seth@earley.com.

 

This article first appeared in IEEE IT Professional magazine. IEEE IT Professional offers solid, peer-reviewed information about today's strategic technology issues. To meet the challenges of running reliable, flexible enterprises, IT managers and technical leads rely on IT Pro for state-of-the-art solutions.

Rate this Article

Adoption
Style

Hello stranger!

You need to Register an InfoQ account or or login to post comments. But there's so much more behind being registered.

Get the most out of the InfoQ experience.

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

Community comments

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

BT