A few years ago, I worked with a team to implement a metrics-based population health program for our Care Managers. We built a program that collected advanced clinical data on various segments of the patient population to better manage their care. We built reports that classified patients using a risk score that calculated the likelihood that they would experience a preventable hospitalization. We started with Diabetes patients, then moved on to patients with cardiac issues, obesity, and other conditions.
So what is population health?
Population Health, or Population Health Management is the management of clinical outcomes of defined population segments, and the technology used in that effort.
It’s no secret that Healthcare in the US has under-performed against our modern peers in spite of spending vastly more than other developed countries. Part of the problem has been that providers get paid for services delivered as opposed to the outcomes of patients who they care for. So when Congress and President Obama were working on Health Reform in 2010, one of the main components was the creation of a new entity called an Accountable Care Organization (ACO). This voluntary arrangement allows Healthcare providers to enter into agreements with large populations of patients, usually through employers to provide services that are partially paid for by the outcome of the patients, not by the quantity of services delivered. It’s called an Accountable Care Organization because providers really are to some degree accountable for how healthy patients become.
To pull this off, Healthcare organizations must collect and analyze clinical data on huge numbers of patients, and provide reporting tools to help Care Managers and Physicians to know how well their patient population is doing.
Population Health Metrics
A key component in a population health program is the use of metrics, or measures to gather clinical data on many patients. That data is used to assign point values to various metrics to create risk scores. One risk score example is the LACE Score.
The LACE score is a set of data gathered on patients to help determine if they are at risk for a hospital readmission. The idea is to catch problems that can be treated by a primary care physician before the problems advance to a critical stage. LACE stands for:
Length of Stay
Acuity of admissions
Emergency Department admissions in the past six months
Every patient who is being tracked will be given a score on each of the above components. Let’s look at Bob, who has Type 2 Diabetes. He had a 4 day hospital stay in the past month, giving him a score in the “L” component. If that hospital stay was serious, then he gets a high Acuity score in the “A” component. He also has high blood pressure, which is a co-morbidity adding to his health issues in the “C” component. Then if he has had any ER visits in the past six months, there are more points assigned. In the end, he gets a LACE composite score, which predicts his overall health and risk for readmission.
So, what then do we do about it? A clinic which is enrolled in an Accountable Care Organization will have an RN Care Manager who works with a defined population of patients. Here is where Healthcare IT comes in. The Care Manager needs advanced data collection tools to evaluate the patient population. Today’s electronic health record (EHR) systems have reports and measurement tools that enable Care Managers to instantly view all of the patients under their care, then sort and group them by disease class, age, and risk scores such as the LACE.
Another widely used score is the Framingham Cardiovascular Score, which is used to estimate cardiovascular risk.
Intensive Outpatient Care Program
Once the Care Manager has identified the patients who are at risk, or otherwise need attention, they then contact the patients to schedule visits with a Physician and/or Care Manager. Remember, the goal here is prevention. Once the patient is under the clinic’s care, the Care Manager helps to set goals for the patient, such as lowering cholesterol or quitting smoking. Patients who are at a very high risk are flagged as IOCP, which stands for Intensive Outpatient Care Program. As time passes and the care plan unfolds, the success or failure of these goals is tracked in the EHR. In an Accountable Care Organization, some of the payment to Providers is based on these measures.
This is certainly a bold step that comes with some risk. What happens if large numbers of patients simply don’t succeed in improving their health? I guess my response would be that since health outcomes in the US are so far behind other developed nations, we’ve got to try something different from what we have been doing, and this seems to be a reasonable attempt.
Where Does The Data Come From?
Many Healthcare organizations manage millions or tens of millions of patients, so they need ways to extract data from lots of complicated angles. Let’s say we want to identify all of our patients who are either Diabetic or are at risk of developing the disease. The easy part of course is to start with patients who have been diagnosed. We need to go a lot deeper though to affect those who are at risk. Population health analysts will use technical tools to build metrics that answer queries like, “show me all patients who…”
- Are diagnosed with Diabetes and haven’t had a recent appointment
- Have not filled their prescriptions
- Have a high or borderline A1C (a lab test to show blood sugar trending over time)
- Have not had an eye or foot exam, which is needed to prevent complications
The data will be extracted from the organization’s EHR system and compiled into a data repository. We can also pull data from insurance companies to find Diabetes claims records on these patients from other organizations. Once the data has been compiled, the analysts will work with clinicians to develop risk scores, like the LACE mentioned above, or customized scores.
When the data has been deemed to be accurate, clinicians can then reach out to the patients to see if they can be brought into the fold to better manage their conditions. Patients with a high risk score would be contacted by phone, but those with a medium risk score might be contacted by letter. EHR systems can auto-generate letters on large numbers of patients.
Over time, the data is re-analyzed to see how well the effort is paying off.
Population Health Success Stories
As population health software has matured, many organizations have had success with improved outcomes, reduced readmissions, and lower costs. Here are just a few examples:
Baystate Health of Springfield, MA used a population health system to manage patients with chronic conditions and behavioral health issues. They were able to lower readmissions by up to 60%, and decrease repeat ER visits by 60%.
Catholic Health Initiatives of Englewood, CO used a population health to achieve a 21 percent decrease in pneumonia mortality, a 27 percent decrease in catheter-associated UITs, a 34 percent decrease in surgical site infections following colon surgery, and 45 percent reduction in surgical site infections after hysterectomies.
Bassett Medical Center of Cooperstown, NY used the IBM Watson platform to reach out to non-compliant patients with chronic conditions. The medical group was able to book more than 43,000 appointments from the effort.
Source: Becker’s Hospital Review
Other organizations have had success in using population health metrics to reach the homeless and otherwise under-served segments.
Population Health Vendors
The field of population health solutions has gotten fairly crowded in a short time. We have population health modules built into the architecture of major EHR vendors such as Allscripts, Cerner, Epic, eClinicalWorks, McKesson, and NextGen. There are also solutions from Forward Health Group, Wellcentive, IBM, and others who don’t have their own EHRs.