Medical records can be a practical tool in hospitals’ quest to lower patients’ risk of 30-day readmission, says the lead researcher of one study aimed at identifying patients most in need of transitional care interventions and another designed to define a limited number interventions based on a patient's identified risks and needs.
In the first study, Angie Hochhalter, Ph.D., and her associates found that longer length of stay, major or extreme illness severity, and unscheduled admission were each significantly associated with a higher risk for 30-day hospital readmission. The assessed records for 6,287 Medicare beneficiaries hospitalized for at least 1 day in 2008 in Texas, noting that the 13.5% 30-day readmission rate was better than Medicare average.
"It’s nice that we have some things that come out as significant, but it is not helpful if we are looking to do personalized transitions," said Dr. Hochhalter, a member of the internal medicine faculty at Scott & White Healthcare and Texas A&M Health Science Center, both in Temple, Texas. So, she said, to determine how people’s behavior might play a role, "we used medication adherence in study No. 2."
In that investigation, the team evaluated data for 2,816 older patients at Scott & White Memorial Hospital. Patients were 65 years and older (median age 77 years) and had a 9.2% 30-day readmission rate at 30 days. They identified two significant factors: readmission was less likely for patients discharged to home vs. a nursing home (hazard ratio, 0.58) or who were more adherent to medications (HR, 0.065).
Empirical evidence is needed to address the problem of readmissions among older hospitalized patients, Dr. Hochhalter said. A high amount of patient heterogeneity is a challenge to designing effective, patient-directed transitional care interventions, she said.
"It’s a Cinderella problem – one size fits few. There are evidence-based interventions we can use, but to pick one and use [it] as a hospital system is probably not going to reach everyone" said Dr. Hochhalter at the most-recent annual meeting of the Gerontological Society of America.
"Offering many sizes is impractical," Dr. Hochhalter said. So her aim was to identify a limited number of transitional care interventions, based on identified risks and needs. The three main drivers of this personalized approach would be predetermined risk, clinical judgment, and family choice.
Scott & White has an integrated electronic record system that includes health plan billing, claims, clinical data, information on use of services, and more, she said. The first study data included members of a Medicare cost contract offered by Scott & White Health Plan. The 6,287 beneficiaries were hospitalized for at least 1 day in 2008, were discharged, and had a subsequent contact with Scott & White within 1 year. The 180-day readmission rate approached 35%.
"The DRG [diagnosis-related groups] severity score gives you an idea of how serious their condition was," Dr. Hochhalter said. Not surprisingly, the 30-day readmission rate increased with severity of illness. "This is important when we think about who we want to target with interventions."
The specific increased risks of 30-day readmission were longer length-of-stay (HR, 1.01), major or extreme severity of illness (HR, 1.61), and unscheduled admission (HR, 1.59).
The results of the second study suggest a person’s behavior, specifically medication adherence, can lower readmission rates. The investigators used medication possession ratio (MPR) from July 2006 to June 2008 as an indirect measure of adherence. "It is more a measure of behavior in the long term and [also reflects their] engagement in their own health," Dr. Hochhalter said. Data included patients hospitalized for at least 1 day at Scott & White in 2007 and then discharged. The 180-day readmission rate in this study was nearly 31%, she added.
They assessed the MPRs for overall drug use, as well as medications used to combat diabetes, hyperlipidemia, hypertension, and depression. "Adherence rates to disease-specific medications were high. We consider anything over 80% high," Dr. Hochhalter said. "But, in terms of overall adherence, it was moderate."
Both discharge to home (vs. a nursing home transition) and higher adherence to medication were significantly associated with lower readmission rates, Dr. Hochhalter said. Regarding adherence, she added, "It is good to show this long-term behavior [is beneficial]."
Unexpectedly, seeing a physician within 30 days of discharge was associated with a higher likelihood of readmission. "It is interesting, because it’s in the wrong direction," Dr. Hochhalter said. "We’ve struggled with this—is it an overall utilization pattern or sicker people? We don’t know the explanation, but it’s unusual to find that."