“As patients near the end of life, an ACO will want to make sure that they receive the appropriate treatments to be comfortable, maintain a good quality of life and have their preferences about dying honored.  Is there any way a medical director can use data to monitor this across a whole population?”

The answer is definitively Yes!  Physician leaders can use data to identify patients nearing the end of life, ensure that their wishes are properly documented, and provide the care and comfort those patients need.  An ACO’s clinical and technical leaders can work together to implement the following plan:

Step one: Algorithm identifies patients of concern.

Claims data combined with EHR data provide a rich source of information to figure out via an algorithm which patients in an ACO are more likely to die within the following 12 months.  (The breadth of claims data helps the ACO see all of the services a patient is receiving in and out of network, while the depth of rich clinical data from the EHR provides a clear picture of the patient’s needs.)

Data items especially useful in constructing the algorithm include age, recent hospitalizations, and significant co-morbidities such as heart failure, cancer or neurologic function.  The Levine score, a well-known older claims-based method for predicting mortality, also uses the element of discharge to nursing home.

Critically, clinical and technical leaders must collaborate to build the right algorithm and data foundation for end of life care.

Step two: Physicians review patient list.

Algorithms can be extremely useful, but still require human review and insight.  Once the algorithm has generated the list of potential patients, the names of these patients should be presented to their own primary care doctors with a simple question – “Would you be surprised if this patient died in the next 6-12 months?”

Step three: Physicians discuss and document end of life wishes.

If the physician answers that she would not be surprised if a patient were to die in the following 6-12 months, the next question for the doctor is, “Have you had a discussion about your patient’s end of life preferences, including issues of hospice and DNR status?”

If the answer is no, the doctor is encouraged to schedule the patient to have that important conversation, and document the answers in the correct field in the EHR.  This is a critical workflow process point, as the ACO needs to be able to harvest the plans for these very ill patients from a stable field, and not from random text in a progress note.

This conversation can be uncomfortable, so the clinical leader also needs to support providers with a training program – mostly for the PCPs, but also for relevant specialists such as oncologists and cardiologists who frequently work with patients near the end of life.  This training program can be delivered via a 2-hour live meeting or via a mandatory webinar, and it should help educate doctors on the how to have a conversation about patient choices around the dying process and the basics of hospice care.

Step four: Care management services assist with end of life challenges.

The next step involves offering Care Management services to these ill patients as directed by the primary physician.  Usually nurses with some extra training in end of life care can take on some of the care of these patients before they might transition to hospice.

A successful program like this requires ACO leadership to make excellent end of life care a priority, and it depends on effective collaboration between clinical and IT leaders.

Presently in the United States, many patients are not offered the opportunity to declare their wishes for how the end of their life should proceed.  Unfortunately, the default option under the law requires mandatory resuscitation regardless of the chances of survival.  This often results in an undignified and inhumane death.

This sad scenario is part of the larger issue of older Americans often receiving unneeded care in the last months or year of life.  We recently shared data with the New York Times for the article Skin Cancers Rise, Along with Questionable Treatments documenting the large number of potentially unnecessary dermatological procedures performed on older patients in the last year of life.  The good news is with thoughtful medical leadership and powerful IT, many of these overutilization problems near the end of life can be effectively circumvented.  I have witnessed a successful, data-fueled end of life care program significantly increase the percent of patients dying with hospice and decrease the percent of patients dying in the hospital.

In summary, a motivated Medical Director can put together a team within the ACO that will result in a greatly improved record of giving patients near the end of life choices and having the resources to honor those choices.

The Final Year: Where and How We Die

For all our efforts to preserve and maintain health and quality of life, an equally important but often overlooked conversation surrounds how our lives come to an end. Approximately 30% of Medicare costs are attributed to the 5% of beneficiaries who die each year, with 78% of those costs stemming from life-sustaining acute care during the final thirty days of life. Compounding the considerable financial cost is the emotional cost of difficult choices made during this time, sometimes without full knowledge or awareness of options and likelihoods.

Arcadian Nick Stepro explores these issues in his data visualization The Final Year, which has been featured on NPR, U.S. News and World Report, and other publications.