How Machine Learning Can Improve Patient Communication

Communicating effectively with patients is often literally a matter of life and death, so being able to do so well is crucial.  New research from Texas A&M highlights how machine learning can help clinicians communicate effectively with cancer patients.

The researchers argue that many of the 1.7 million new cases of cancer per year in the United States could be avoided if more regular screening tests were performed and detection made earlier.  They cite reduced mortality rates of 28% for lung cancer, 24% for breast cancer and 37% for liver cancer when regular screening is performed.

Screening is typically bolstered by effective marketing interventions that encourage people to enroll.  These can consist of emails, letters, community outreach, and various other approaches, yet those in greatest need seldom get the screening they so urgently require.

Effective communication

A group of at-risk patients were assigned to one of three groups: usual care, outreach alone, or outreach with patient navigation.  The usual care group received the baseline preventative care recommendations given at the discretion of the physician during a usual care visit.

The outreach alone and outreach with patient navigation groups provided a variety of direct marketing efforts, containing mailings, phone calls and customized educational material from trained patient navigators.

By using machine learning, the team were able to understand how the effectiveness of outreach programs differed markedly across both time and patients.  For instance, outreach tended to be most effective when patients were female, in better health, and visited their doctor more regularly.  When outreach was combined with patient navigation, it tended to work best with older patients who lived in higher-income neighborhoods.

The researchers believe that a fully optimized outreach program could achieve a better return of up to 96%, which would produce not only significant health benefits, but also financial gains running into millions of dollars too.

Of course, the challenge is to implement these findings and provide personalized outreach to specific patients, which is easier said than done, especially for healthcare organizations that aren’t renowned for their high-technology approaches to work.

Facebooktwitterredditpinterestlinkedinmail