Risk adjustment, costs of care

Summaries from ACP Hospitalist Weekly.

Risk adjustment for more patient, population characteristics could reduce readmission penalties

Adjusting for clinical and social patient characteristics that CMS does not currently account for as part of the Hospital Readmissions Reduction Program may alleviate financial penalties for hospitals that disproportionately serve poorer and sicker patients, a recent study found.

Researchers used Medicare claims for hospital admissions in 2013 and 2014 and linked U.S. Census data to assess the impact on 30-day all-cause readmissions of accounting for clinical and social patient characteristics not currently used for risk adjustment in the program. The study sample included almost 1.2 million index admissions to 2,215 hospitals among 1,003,664 unique Medicare beneficiaries (41.5% men; mean age, 79.9 years).

The new characteristics for adjustment included hospitals' cumulative chronic disease burden, patients' dual enrollment in Medicare and Medicaid and disability status, the number of long-term nursing home residents, and multiple area-level socioeconomic variables. The researchers compared hospital readmission rates with and without adjustment for these additional characteristics, using only the average within-hospital associations between patient characteristics and readmissions as the basis for adjustment. Results were published online on Sept. 17, 2018, and in the November 2018 JAMA Internal Medicine.

Compared to current adjustments, adjustment for the additional characteristics decreased overall variation in hospital readmission rates by 9.6%. For the 10% of hospitals most affected by the new adjustments, readmission rates changed upward or downward by 0.37 to 0.72 percentage point. The study authors estimated that the new adjustments could reduce penalties (in relative terms) by 52%, 46%, and 41% for hospitals with the largest 1%, 5%, and 10% of penalty reductions, respectively.

In addition, the new adjustments reduced the mean difference in readmission rates between hospitals in the top and bottom quintiles of high-risk patients by 0.53 percentage point (95% CI, 0.50 to 0.55 percentage point; P<0.001), or 54% of the difference estimated with only CMS adjustments. Both clinical and social characteristics contributed to the reductions, which were much greater for conditions targeted by the Hospital Readmissions Reduction Program, although adjustments for social characteristics led to greater changes in rates of readmission and mortality than in rates of readmission alone.

The authors noted limitations of the study, such as the limited number of patient characteristics they were able to assess and their inability to evaluate causes of within-hospital associations between social risk factors and readmission.

Across all payment models, risk adjustment can induce clinicians “to either attend to the needs of vulnerable populations or shy away from them,” an accompanying commentary noted. “We must develop a consistent and equitable approach to risk adjustment that takes into consideration evidence on the costs of care associated with both clinical and social conditions as well as how organizations will respond when addressing the differential costs of managing the care of vulnerable populations.”

High costs among dual-eligible beneficiaries due mostly to long-term care

High health care costs in patients eligible for both Medicare and Medicaid appear to be related to long-term care rather than to preventable hospitalizations, according to a recent study.

Researchers performed an observational study using Medicare and Medicaid data to examine persistently high health care costs in patients eligible for both programs. Payments for 2008, 2009, and 2010 were examined, and patients were categorized as high cost in a particular year if their spending was in the top 10%. The authors also calculated differences in spending for patients who met the high-cost threshold in all three study years, those who met the threshold in 2008 but not in 2009 or 2010, and those who never met the threshold. The study results were published in the Oct. 16, 2018, Annals of Internal Medicine.

The study included 1,928,340 patients eligible for both Medicaid and Medicare in 2008, 2009, and 2010. In 2008, 192,835 patients (10%) met the threshold for high cost, and of these, 54.8% continued to fall into the high-cost category in the remaining two years of the study. Overall, 5.5% of patients were persistently high cost, 4.5% were transiently high cost, and 84.2% were non-high cost. Patients in the persistently high-cost group were younger than those in the transiently high-cost group and non-high-cost group and had fewer comorbid conditions and more intellectual impairment than those in the transiently high-cost group.

Spending in the persistently high-cost group was $161,224 per year versus $86,333 per year in the transiently high-cost group and $22,352 in the non-high-cost group. Among persistently high-cost patients, 68.8% of spending was on long-term care and less than 1% was related to potentially preventable hospitalizations related to ambulatory care-sensitive conditions, defined as those that can potentially be avoided if the patient receives appropriate, coordinated outpatient care. Persistently high-cost patients accounted for 34.0% of total spending over the three years of the study despite comprising only 5.5% of the dual-eligible population.

The authors noted that their study was based on administrative claims, that their identification of potentially preventable hospitalizations may not have been accurate in all cases, and that their data were relatively old, among other limitations. However, they concluded that high costs appeared to be persistent in more than half of patients eligible for both Medicare and Medicaid and that most of these costs were related to long-term care. “Strategies to control costs in dual-eligible beneficiaries may be more effective if they focus on reducing spending in long-term care rather than on reducing potentially preventable hospitalizations,” the authors wrote.