Which Patient Populations Cause the Highest Readmissions?

In the third year of the Hospital Re­ad­mission Reduction Program (HRRP) (FY 2015), the proportion of hospitals receiving a penalty increased to 78 percent – up from 64 percent and 66 percent in 2013 and 2014 respectively. This increase is not surprising given that for 2015 penalties, CMS included additional diagnoses when assessing each hos­pital’s performance.

CMS estimated total hospital penalties under the HRRP to sum to $428 million in 2015, an increase over prior years ($290 million in 2013 and $227 million in 2014). This increase is due mostly to more hospitals receiving a penalty, given the expanded number of initial diagnoses included in calculating readmission rates, but a small part of the in­crease may also due to the increase in the maximum penalty.

“Very few studies have looked at the relationship between mild depression and readmissions.”

The Medicare Population And Readmissions

Most Medicare patients will stay in hospitals with either no penalty or penalties less than 1 percent

From a patient perspective, analysis finds that for 2015, most beneficiaries will stay in hospitals with low to no pen­alties. Specifically, it is estimated that 83 percent of beneficiary stays in 2015 will be in hospitals that scored well enough during their previous measurement period to receive either no penalty or penalties of less than 1 percent (Figure 1). Conversely, 17 percent of beneficiary stays will be in hospitals that will re­ceive pen­alties of 1 percent or higher. Less than a percent (0.5%) of Medicare admissions will occur in hospitals that re­ceived the maximum penalty of 3 percent.

Analysis of the variation in penalties by type of hospital suggests that Medi­care beneficiaries who go to certain types of hospitals – namely major teaching hospitals and hospitals with relatively greater shares of low-income beneficiaries – are more likely to stay in penalized hospitals (Table 1). To some degree there is overlap among these two types of hospitals as major teaching hospitals often serve as safety-net hospitals with higher proportions of low-income pa­tients. Further study by initial diagnoses could provide more insights, particularly for 2015 penalties which show some nar­rowing of performance among hospitals with greater proportions of low-income beneficiaries, compared to previous years. For example, among the top three quartiles, the percentage of hospitals receiving a penalty ranges by only 3 percentage points in 2015 (81% – 84%), but by 16 percentage points in 2013 (61% – 77%).

Main Findings:

• Both the share of hospitals receiving penalties for 30-day readmissions and total fines are higher in 2015, compared to previous years – due mostly to more medical conditions being measured, rather than increases in the penalty cap. The average financial penalty (Medicare payment reduction) for individual hospitals is less than 1 percent, as it has been in prior years.
• For 2015, 83 percent of Medicare patient admissions are projected to be in hospitals receiving either no readmission penalty or penalties of less than 1 percent.
• Across all three years of the HRRP, some types of hospitals are more likely than others to incur penalties, including major teaching hospitals and hospitals with relatively higher shares of low-income beneficiaries – two often overlapping characteristics.
• Beneficiary readmission rates started to fall in 2012, suggesting that hospital administrators and clinicians may have initiated strategies soon after the enactment of the HRRP and prior to the application of the fines – realizing that the penalties would be based on performance in preceding years. Other factors may also have played a role these declines.
• Researchers, hospitals, and policymakers are actively considering refinements to the HRRP and looking for ways to engage other providers and patients to reduce preventable patient readmissions to the hospital.

National Medicare Readmission Rates Started Falling Prior To HRRP Penalties

CMS has been posting individual hospital readmission rates on its Hospital Compare website, in addition to other measures of quality and patient satisfaction, since 2009. Designed for use by Medicare consumers as well as researchers, this website also provides comparisons of each hospital’s Medicare readmission performance to the national average by indicating whether the hospital is “better/worse/no different” than the U.S. National rate. In addition to readmissions following hospitalizations for se­lected diagnoses, the Hospital Com­pare website started reporting each hospital’s overall Medicare readmission rates.

While other studies have focused primarily on penalties imposed on specific hospitals, a database from “Hospi­tal­iza­tions among Patients in the Medicare Fee-for-Service Program,” New England Journal of Medicine provides additional analysis on the extent to which Medicare beneficiaries stay in hospitals with relatively high penalties and national trends in beneficiary readmission rates. Anal­ysis of this database shows that 2012 marks the first measurable declines in readmissions (Figure 2). Specifically, when the last six months of 2008 were replaced by the first six months of 2012, national re­admission rates fell across all three diagnosis categories.

Furthermore, these rates continued to decline in the subsequent measurement period. For each of the three diagnosis categories, therefore, hospital readmissions fell about 1-2 percentage points by mid-2013. These drops started after Congress enacted the HRRP, suggesting that hospitals may have initiated strategies to lower their readmission rates during the HRRP measurement periods preceding the start of the program’s penalties. Ad­ditionally, other clinician activities outside the hospitals may have contributed to the decline in readmissions. CMS has estimated reductions in overall hospital readmissions (not specific to any initial diagnoses) totaling 150,000 fewer Medi­care patient readmissions during the pe­riod between January 2012 and De­cem­ber 2013.

Super-Utilizer Population And Readmissions

The term “super-utilizer” describes individuals whose complex physical, behavioral, and social needs are not well met through the current fragmented healthcare system. As a result, these individuals often bounce from emergency de­part­ment to emergency department, from in­patient admission to readmission or in­stitutionalization – all costly, chaotic, and ineffective ways to provide care and im­prove patient outcomes.

Super-utilizers are the frequent fliers of the healthcare system, whose serious illnesses send them to the hospital multiple times every year and cost the system hundreds of thousands of dollars an­nually. Figuring out how best to address these patients’ needs and reduce their fi­nancial impact on the healthcare system is a subject of intense interest among policymakers. Now a new study has found that, in contrast to the notion that “once a super-utilizer, always a super-utilizer,” many patients who use healthcare services intensely do so for a relatively brief period of time.

Super-utilizers had an average all-cause 30-day readmission rate that was four to eight times higher than the readmission rate for other patients

The rate of all-cause 30-day readmissions was much higher among super-utilizers than among other patients, ranging from four times higher (among Medi­care patients) to almost nine times higher (among the privately insured). Across payers, the 30-day readmission rate among super-utilizers was in the 45-52 percent range compared with 5-12 percent among other patients.

Compared with other patients, Medicaid and privately insured super-utilizers had longer hospital stays and higher average hospital costs

Compared with other Medicaid patients, Medicaid super-utilizers had stays that were approximately 30 percent longer (6.1 vs. 4.5 days) and average hospital costs that were 30 percent higher ($11,800 vs. $9,000). Similarly, compared with other privately insured patients, privately insured super-utilizers had stays that were approximately 62 percent longer (5.9 vs. 3.6 days) and average hospital costs that were 43 percent higher ($14,600 vs. $10,200). There were no substantial differences in average length of stay or average hospital costs between Medicare super-utilizers and other Medicare patients.

Overall, super-utilizers accounted for between approximately 15 and 26 percent of aggregate hospital costs: 15.0 per­cent for Medicare patients aged 65+ years, 25.9 percent for Medicare patients aged 1-64.

A study by researchers at Denver Health, a medical center that serves many uninsured and underinsured pa­tients, ex­amined the characteristics and costs of patients who were hospitalized more than three times during a 12-month period in the two years from May 1, 2011, to April 30, 2013, or were hospitalized at least twice in 12 months and had a serious men­tal illness.

The results surprised researchers, says Tracy Johnson, director of healthcare reform initiatives at Denver Health and the study’s lead author, which was published in Health Affairs. At the end of the first year, only 28 percent of 1,682 patients who were originally identified as super-utilizers still met the criteria. At the end of two years, the figure had shrunk to 14 percent. Per person spending decreased in line with their de­creased use of healthcare services, from a baseline $113,522 per capita to $47,017 in year two.

Moreover, on an individual level, super-utilizers didn’t necessarily have characteristics of patients frequently as­sumed to fall into that group, Johnson says. “You’d think they’d all be people with multiple chronic conditions,” she says.

While a substantial 42 percent of high-cost patients with frequent hospital stays did have multiple chronic conditions, others did not. (Researchers grouped patients based on the category that best represented the reason for their hospitalizations. Some patients could have been placed in multiple categories.) Forty-one percent of super-utilizers’ hospitalizations were primarily related to serious mental health diagnoses. Smaller numbers of high-use patients were hospitalized because of orthopedic surgery, trauma, terminal cancer or for emergency inpatient dialysis.

Research and news reports often point out that super-utilizers are often un­insured or on Medicare and Medicaid and account for a large percentage of healthcare spending. Federal officials have suggested that their “large numbers of emergency department [ED] visits and hospital admissions might have been prevented by relatively inexpensive early interventions and primary care.” Many of the programs that have been developed to reduce super-utilizer healthcare use have focused on the needs of people with multiple chronic conditions, ensuring they have a medical home through which their care is coordinated, for example, or addressing their social services needs.

Behavioral Health Related To Readmissions

Mood disorders were among the top 10 principal diagnoses for super-utilizers aged 1-64 years who were covered by Medicare, Medicaid, or private insurance.

Mood disorders were among the top 10 principal diagnoses for patients aged 1-64 years across all payers, but not for Medicare patients aged 65 years and older. Specifically, mood disorders was the most common reason for hospitalization among Medicaid super-utilizers; it was also a top 10 condition for Medicare and privately insured super-utilizers aged 1-64 years.

Other mental health and substance use disorders also were among the most com­mon diagnoses for super-utilizers for some payers. Schizophrenia was the second most common condition for super-utilizers aged 1-64 years who were covered by Medicare or Medicaid. The share of hospital stays for this condition attributable to these super-utilizers was 31 per­cent among Medicare patients and 28 per­cent among Medicaid patients.

Alcohol-related disorders was a top 10 condition for Medicaid super-utilizers, who accounted for one-third of all hospital stays for this condition.

Researchers from Boston Medical Cen­ter (BMC) analyzed the data of pa­tients screened on admission for depression has identified what may be a major readmission risk factor that needs much more attention: depression. What the re­searchers found surprised them. While the prevalence of depression in the general adult population is under 7%, 16% of patients in their data set screened positive for mild depressive symptoms. An­other 24% tested positive for moderate or severe depression.

The authors also found a dose-re­sponse relationship between depression and readmission rates. Only 10% of pa­tients without depression were readmitted within 30 days. But that rate jumped to 14% for those found to have mild symp­toms, and it hit 19% – almost double – for patients with moderate to se­vere depression. “We were shocked at the sheer number,” says Ramon Cancino, MD, a BMC staff physician, chief medical officer of Mattapan Community Health Center in Mattapan, Mass., and the study’s lead author. As Dr. Cancino explains, the findings could point to an actual diagnosis of depression that calls for a psychiatric evaluation. But results could also be due to the fact that BMC is an academic and urban safety-net hospital. “Many of these pa­tients have other social determinants that contribute to poor health outcomes,” he points out. “Hav­ing depressive symptoms may be a direct result of those social determinants.”

Very few studies have looked at the relationship between mild depression and readmissions. Patients with mild symptoms may be readmitted due to a poor discharge process or to an outpatient fail­ure to diagnose or manage those symp­toms. But according to Dr. Cancino, even mild depressive symptoms can have major effects. The problem is that many physicians see mild symptoms as not very dangerous to a patient’s well-being or to the hospital system as a whole. “According to our paper, screening positive for even mild depressive symptoms is a marker for readmission and for utilization in general,” he says. “A simple screen for depressive symptoms might help physicians who are implementing readmission-reduction programs.”

Data Analytics And Social Determinants Can Tell Us More About Readmissions

Marilyn Szekendi, PhD, RN, director of quality research at University Health System Consortium (UHC) in Chicago, says learning the characteristics that lead to frequent admissions – defined as pa­tients who are admitted five or more times within one year – can help identify solutions for preventing repeated hospitalizations. UHC is an alliance of nonprofit academic medical centers and their affiliated hospitals.

“The good news here is that this is very doable,” Dr. Szekendi says. “Every hospital can run this analysis and actually create a list of who these patients are, along with their names and medical record numbers, so you can look at their diagnosis, you can look at other characteristics of the patient, and do a real-time assessment of who they are.”
For their report, Dr. Szekendi and col­leagues studied 28,291 patients admitted 180,185 times to academic medical centers in the U.S. from 2011 to 2012. While the cohort comprised just 1.6% of all pa­tients, it accounted for 8% of all admissions and 7% of direct costs.

Common factors linked with frequent readmissions included having significantly more comorbidities (an average of 7.1 versus 2.5), and 84% of their ad­missions are to medical services. In ad­di­tion, this patient population has higher rates of psychosis or substance abuse, the researchers note. Although frequently admitted patients are slightly more likely than other patients to be on Medicaid or to be uninsured (27.6% versus 21.6%), nearly three-quarters have private or Medicare coverage.

“We know that there are many other factors that we didn’t have data for, [such as] housing status, patients’ preexisting access to other kinds of medical care,” Dr. Szekendi says. “If we could do some further look at factors that define these patients, both nationally and individually, hospitals then would have some additional, really useful information about the patients that would further inform their improvement efforts. Going beyond the data … is the next step.”

---

Highlights: Readmission Stats

• The average all-cause 30-day readmission rate was four to eight times higher for super-utilizers than for other patients. Among patients aged 1-64 years, super-utilizers accounted for more than half of all 30-day readmissions.
• For all payers, patients with multiple chronic conditions accounted for a greater share of stays among super-utilizers than among other hospitalized patients. For example, among the privately insured, patients with two or more chronic conditions constituted 60.2 percent of all stays for super-utilizers compared with 36.4 percent for other patients.
• Super-utilizers were more likely to be admitted for medical conditions rather than surgical or other types of conditions. For example, among the privately insured, 65.1 percent of all stays for super-utilizers were admitted for medical conditions compared with 33.3 percent for other patients.
• Common chronic and acute conditions, such as congestive heart failure and septicemia, were among the 10 most common principal diagnoses for hospitalized super-utilizers across all payers.
• Mental health and substance use disorders were among the top 10 principal diagnoses for super-utilizers aged 1 to 64 years regardless of payer.

Resources:

1. Jencks, S. F. et al., “Hospitalizations among Patients in the Medicare Fee-for-Service Program,” New England Journal of Medicine Vol. 360, No. 14: 1418-1428, 2009.; Epstein, Arnold M. et al., “The Relationship between Hospital Admission Rates and Rehospitalizations,” New England Journal of Medicine Vol. 365, No. 24: 2287-2295, 2011.

2. The Hospitalist

3. Kaiser Foundation