GNS Healthcare Blog

Population Health to Personalized Medicine: Why Shooting for “Average” Misses the Mark

  

Current cancer treatments provide benefits to only one out of every four patients. Drugs for Alzheimer’s are ineffective for 70 percent of patients. Medicines for arthritis (50%), diabetes (43%), and asthma (40%) have no benefits for large portions of afflicted individuals. In general, according to a report from the Personalized Medicine Coalition (PMC), many FDA-approved drugs are ineffective on average for nearly half the targeted patient populations.

These sobering statistics are hard to believe in an era of advanced technology, but not so surprising when you get to the core issue. Medicine is currently practiced using evidence-based treatment guidelines developed for the “hypothetical average patient”. The problem is none of us are average.

Population Health efforts generate huge amounts of healthcare data and providers are striving to figure out the best way to leverage the information. Unfortunately, using this data for broad-stroke solutions fails to get to the critical causes of disease at the individual patient level.

Providing the right treatment to the right patient at the right time sounds simple enough in theory. In practice, however, matching an individual patient’s condition with the most effective treatment or intervention is one of the most challenging aspects of healthcare. Tailoring the treatment to the individual is the key to achieving personalized medicine since what works for one patient may be ineffective for others.

 

One-size-fits-all approach doesn’t work

It’s become clear that a top-down, standard-of-care approach – giving the same drugs to anyone with a given disease or condition – will only be effective some of the time. Solving this patient/treatment matching problem requires determining which specific individuals will benefit from a given treatment- whether it’s a drug, medical device, procedure or care management program. The best way to drive beneficial care outcomes requires taking a holistic approach to the individual patient to determine what works best for each.

The costs of an ineffective, one-size-fits all treatment approach are significant. Take the example involving metastatic colorectal cancer. According to a report from the PMC, using genetic tement would save $604 million in annual health care costs.[1]

With that kind of savings from one type of treatment for one type of disease, and considering total annual healthcare spend in the United States now tops $3 trillion, it’s not a stretch to estimate that billions of dollars are being wasted on ineffective treatments.

Realizing the goals of Triple Aim – enhancing the patient experience, improving care outcomes and reducing the cost of healthcare – may mean moving from a population health mindset to a personalized medicine approach.

 

Getting to What-if?

With so much at stake both in terms of healthcare spend and improved patient outcomes, how do we take the next step to better match patients with the targeted treatments that are right for them?

What if we could provide each patient with a personally optimized treatment plan? Or predict which diabetes medication will help a specific patient best manage his or her disease? What if we could use genomic data to identify patients who would benefit from a stem cell transplant? And what if we could deliver this type of care at similar or lower cost than we are able to do today?

Incorporating these “what if” questions into the healthcare decision making process is the essence of personalized medicine. When we answer these questions -  something biopharmaceutical companies and health plans are trying to do today - we as an industry will achieve a new, higher level of precision in patient care, one that delivers far better outcomes at far less cost.

Advances in artificial intelligence (AI), like the powerful causal machine learning technology, are making great strides in helping us make the move from the “hypothetical average” to optimized intervention plans for individual patients. No one wants to be a “guinea pig” for a series of “maybe-this-will-work” treatments while his or her disease continues to progress. Patients want to be treated based on their own biology and other personal data factors that will lead to specific treatments that will benefit them. 

Healthcare has been attempting to transition from a fee-for-service model to a value-based system for a few years now.  The current environment of skyrocketing costs, the rise of big healthcare data and breakthroughs in machine learning presents the perfect opportunity to bring precision matching of interventions to individuals.

Population Health initiatives have made great contributions to improved care outcomes, but it’s time to take the next step to personalized medicine. Real value is going to come from leveraging artificial intelligence to determine what works for whom. That’s the needed shift that will get us closer to the goal of true value-based change.

 

 

[1]The Personalized Medicine Report, 2017 – Opportunities, Challenges, and the Future, Personalized Medicine Coalition

Subscribe to the GNS Blog

Recent Posts:

Twitter