Opinion: Focusing Care on Patients Who Need It Most—Easier Said Than Done?

Fred N. Pelzman, MD, MedPage Today , March 20, 2015

We all sort of know what a high-risk patient looks like, but it’s really hard to put that in writing with a set of hard and fast rules that lead to really great predictions. From Medpage Today.

They are coming for our Medicare patients.

Now that our accountable care organization has ramped up the hiring of care coordinators across our system, they have refocused their efforts on figuring out the complex issue of what makes a high-risk patient — that is, who needs care management?

As I have described previously, we all sort of know what a high-risk patient looks like, but it’s really hard to put that in writing with a set of hard and fast rules that lead to really great predictions.

The ACO understandably wants to put effort towards care coordination for the patients most in need of those services and most likely to benefit from them. In hard practical terms, that does mean saving money for the organization ultimately. Investment in efforts up front will hopefully help the patients get to a better state of health, all the while saving the organization some money.

Over the past year, as this project has developed, our office of managed care, along with multiple healthcare outcomes researchers, has worked on developing a model of predicting those patients most likely to be hospitalized and/or in need of intensive care management.

As you can imagine, lots of great minds, lots of effort, lots of math, goes into designing a system like this. It kind of reminds me of the old original Framingham equations for estimating 10-year mortality from coronary heart disease.

While we all know and love the easy-to-use calculators now available on Web-based apps or smart phones, how many of us actually remember the underlying math?

{For those of you interested:

1-exp(-exp[(ln(10) – u/s]), where u = 4.4284+m, and s = exp(-0.371-0.2825m)

For men, m = a – 1.6346 x ln(age) – 0.2082 x diabetes

For women, m = a – 6.5306 + 2.1059 x (ln[age/72])2 – 0.4055 x diabetes

For both men and women, a = 11.0938 – 0.8670 x ln(diastolic blood pressure) – 0.2789 x smoking – 0.7142 x ln(total cholesterol/HDL) – 0.7195 x EKG – LVH}

Complex predictive models can be incredibly powerful, incredibly useful, and also incredibly frustrating.

How often have we left the house dressed based on the weather report from the evening news, only to end up miserable and soaked on the street?

When our director of managed care e-mailed me my list of Medicare ACO patients, I was struck by how familiar most were to me (some just should not have been there, moved away and getting care elsewhere), how near and dear to my heart, how involved I was in their lives. And also looking at them I realized how much help I could use taking care of these complex patients. In many ways, this suggests that I and our healthcare system have surely failed them to some degree up till now.

Of the approximately 150 patients on my list, numerous stories of their lives, their interactions with the healthcare system, and my role in their health floated up off the page at me. An end-stage renal disease patient whose lifetime of poorly controlled diabetes and high blood pressure had led her to dialysis. Complex psychiatric illnesses for which patients have been unwilling to get care. Consequences of so many choices that patients had made that, with 20/20 hindsight, we all wished they had made otherwise.

Right off the bat, I noticed the names of five of my patients who had passed away. The data that led to this list was collected in 2013, and these patients have had difficult, complicated, unlucky, and subsequently tragic years.

The cynic in me made me want to tell the data collectors not to worry, these people were not going to cost them anything anymore.

But that seemed a little unfair. Still, refreshing your dataset to make sure you have the latest information before setting off makes sense, and helps us take it more seriously.

The team compiling the list had applied their statistical analysis to the patient’s demographics, medications list, problem list, lab results, number of hospitalizations and office visits, as well as costs, and highlighted 17 patients that they had elected to target as an initial sample.

Luckily, the team was open to hearing feedback from me and my partners about why some of these patients might not be a good candidate for intensive care management, and why some others on the list that they had missed were definitely where they could spend their hard earned money.

Some of these patients were just too highly functional, still gainfully employed, resistant to all efforts to change, likely to be insulted by someone trying to help.

Or was this just my own naivete, my stubborn feeling that if I could not make them better no one could?

Ultimately, I think more eyes on the problem is likely to only make things better.

Maybe my disorganized, homeless, schizophrenic patient will listen to the advice of someone other than me, and subsequently take his medicines and come to follow-up appointments.

Maybe there is another way to get my poorly-controlled diabetic to change her diet and start exercising, and I’m just not right person to help her find that.

And so on …

On my schedule today are two of their highlighted patients. I am not sure that either one of them wants to hear that they have been flagged as “high risk,” complicated, refractory, expensive.

I will try to gently nudge them towards intensive care management, to find the words to engage them in engagement.

And once the 17 are enrolled and on their way, we need to move on to the 150, and then on to everyone.

SOURCE: http://www.healthleadersmedia.com/print/QUA-314548/Opinion-Focusing-Care-on-Patients-Who-Need-It-MostmdashEasier-Said-Than-Done

Categories: Uncategorized

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s