30-Day Readmissions among Seriously Ill Older Adults

Introduction

Inpatient palliative care (IPC) consultation teams are a predominant, fast growing service in the United States, with a 138% growth rate from 2000 to 2009. Nearly two-thirds (63%) of U.S. hospitals with 50 beds or more report providing palliative care services.^1,2 These services are fueled by data demonstrating high rates of unmitigated pain and symptom distress, hospital deaths, family burden, and general dissatisfaction with end-of-life hospital care. ³ Additionally, as the evidence builds with reports of improved symptom management, lower inpatient death rates, improved patient and family satisfaction, and lower costs, this growth is expected to continue.^4–9

The importance of considering the gaps in the palliative care consultation model for patients and families is being heightened by emerging models of payment and delivery stimulated by the Patient Protection and Affordable Care Act (ACA) and an increasing recognition that episodically oriented approaches to care don't serve chronic illness well. IPC services are provided during one discreet hospital episode and most lack mechanisms for continuity following hospital discharge. Underscoring a potential mismatch between a hospital-oriented model and what patients and families need, the majority of care in the last year of life takes place at home, provided largely by family members who are often untrained and ill prepared to address the myriad problems encountered by seriously ill patients.^10–12 This lack of preparation and training may result in increased use of medical services; a recent study found that 77% of patients who went to the emergency department in the last 30 days of life were subsequently admitted to the hospital, and among those admitted, 68% died in the hospital. ¹³ With most people preferring to die at home,^14–17 these hospital admissions and readmissions may lower quality of life for patients.^17,18

Medical care in the last year of life is costly, accounting for 27% to 30% of annual Medicare expenditures.^19–20 These high costs are driven by greater numbers and intensity of inpatient admissions,^21,22 accounting for at least half of expenditures in the last year of life. ²³ In fact, one study suggests that acute care in the last 30 days of life accounts for more than three-fourths of all costs of care in the last year of life. ²³ With new ACA provisions of reducing Medicare reimbursement to hospitals with high readmissions rates coupled with high costs of care in the last year of life and potentially reduced quality of life for patients, understanding factors associated with hospital readmissions among seriously ill inpatients is critical.

This study identifies factors associated with readmissions in the 30 days following hospital discharge among seriously ill recipients of an IPC consult. Using the Anderson Behavioral Model of Health Service Use, which includes predisposing characteristics, enabling resources, need, and use of health services (e.g., hospice, home health care, nursing facilities) as a conceptual model guiding our study, we hypothesized that ethnicity, age, pain, anxiety, and discharge disposition would be related to30-day hospital readmissions.

Methods

We conducted a retrospective cohort study to identify factors associated with hospital readmissions in the 30 days following discharge among seriously ill patients receiving an IPC consult. Institutional review board approval was received from Kaiser Permanente Medical Center and the University of Southern California.

The sample consisted of seriously ill patients aged 65 and over who received an IPC consult between October 2007 and June 2009, and subsequently survived to hospital discharge. The study site was an urban non-profit health maintenance organization (HMO) medical center located in Los Angeles County.

Results

From October 2007 through June 2009, 484 patients aged 65 and over received IPC consults. Among these, 45 (9.3%) died during hospitalization, and 31 (6.4%) had missing medical service or disposition data, leaving 408 in the analytic sample. The mean age was 80.1 years (standard deviation [SD]=8.2) and about half (48.5%) were female. The sample was diverse: 37.5% were white, 22.5% Latino, 20.3% black, 7.8% were of other ethnic background, and 11.8% had missing data. Cancer was the most common primary diagnoses (34.3%) followed by congestive heart failure (CHF) (16.4%), dementia (11.8%), coronary artery disease (11.8%), and chronic obstructive pulmonary disease (COPD). More than half (58.8%) were discharged to hospice, 14.7% to home-based palliative care, 14.2% to a nursing facility, 8.6% to home with no care, and 3.7% to home with home health (Table 1). Nearly all (99.2%) admissions to hospice or home-based palliative care were new referrals.

Overall, 10.0% of those discharged from the hospital were readmitted within 30 days. Comparison of sample characteristics among those that were and were not readmitted revealed that 30-day readmitted patients were more likely to have CHF or end-stage renal disease and less likely to have an advance directive. Additionally, 30-day readmitted patients were more likely to have no emergency contact or a distant relative as an emergency contact (Table 1). They also were more likely to be discharged without care at home or to a nursing facility (Table 2). Among the 41 patients readmitted within 30 days, more than half (51%) were readmitted within 5 days and 78% were readmitted within 10 days of hospital discharge.

Discussion

This study found that, overall, 10% of seriously ill patients receiving an IPC consultation were readmitted in the 30 days following hospital discharge, and having hospice or home-based palliative care post-discharge was associated with significantly lower odds of hospital readmission. Moreover, the lowered rates were similar between the home-based palliative care group and the hospice group with 4.6% of those discharged to hospice and 8.3% of those discharged to home-based palliative care being readmitted, much lower compared with the 25.7% rate among those discharged home with no in-home care. A randomized controlled trial of a home-based palliative care model for seriously ill patients found that costs of care can be reduced by 38%, largely by reducing emergency room visits and hospitalizations and increasing in-home multidisciplinary care. ²⁶ Additionally, a recent study found enrollment in hospice care was highly associated with lower rates of emergency department use in the last month of life. ¹³

Disease-specific case management studies also suggest that longitudinal monitoring and support may reduce avoidable hospitalization among high-risk patients. Meta-analyses of 18 studies of comprehensive discharge planning plus post-discharge support and 29 studies of multidisciplinary heart failure management teams found these models to be effective in reducing hospital readmissions, improving quality of life, and increasing survival rates^.31,32 These findings underscore the potential benefit of longitudinal follow-up, with a systematic review characterizing the similarities between successful case management programs and palliative care services. ³³

Our current study found the rate of readmission for our IPC patient sample was lower than the overall medical center 15% readmission rate among older adults, even when considering the sickness of the population under care. A previous randomized controlled trial of IPC found no difference in hospital readmissions between usual care and those receiving an IPC consult. ⁶ However, readmissions among seriously ill patients may entail significantly higher costs of care compared with the general population of older adults. A study of IPC patient costs found average costs per hospital episode ranged from $16,737 (nonterminal admission) to $30,494 (terminal admission). ³⁴ Exacerbating these implications, the Center for Medicare and Medicaid Services will lower reimbursement rates for those hospitals with high levels of readmissions, beginning in 2013. Given the potential for higher odds of readmissions among those discharged to home with no in-home care, elevated costs of these readmissions, and the significant evidence attesting to the ability to save money through the provision of home-based palliative care, broadening access to these longitudinal palliative care programs may offset higher costs accrued from hospital readmissions and may even provide opportunity for additional cost savings. For example, costs of monthly palliative care services may average about $2,000, clearly offsetting hospital episodes that exceed $16,000 on average. ²⁶

Although having an advance directive was not significantly associated with hospital readmission, others have found that end-of-life discussions are positively associated with having an advance directive and receiving hospice care. ²³ A study of health care costs in the last week of life found that those who had not have an end-of-life discussion with their physician had 55% higher costs of medical care than those who had had discussions. ²³ The addition of discharge disposition in the model reduced the significance level of this variable, understandable given that hospice services specifically include advance care planning. Other studies may not have had the comprehensive model that we used to examine this effect.

Our analysis has several limitations. As a study of a managed care medical center our findings may conservatively estimate the magnitude of this problem. The site for our study has cultivated particular excellence in home-based palliative care, which is generally unavailable elsewhere. However, this site provided an opportunity to examine the potential effects of services to promote earlier access to palliative care on hospital readmission rates. We are unable to determine if any of the patients discharged to a nursing facility received hospice, although the standard within the HMO does not include a full hospice team in nursing facilities. Additionally, as an administrative data analysis, we lack reasons for hospital readmission that might offer insight into the appropriateness of subsequent care. Disease progression and a new medical condition have been described as reasons for readmission among this population. ³⁵ However, to the extent we could measure them, including the presence of pain, anxiety, and serious underlying illness including co-morbidity, our population had similar markers of advanced, debilitating illness.

This study, one of the first to examine hospital readmissions among seriously ill patients, found receipt of palliative care to be positively associated with lower odds of readmissions and provides preliminary evidence for the need for improved longitudinal access to palliative care for some seriously ill patients following IPC consult and hospital discharge. Longitudinal palliative care models provide a natural platform to support patients and caregivers if disease progresses and they experience increased complications. Further research is needed to test the effectiveness of such longitudinal models in reducing readmissions among seriously ill patients.