|Home | About | Journals | Submit | Contact Us | Français|
Homebound adults experience significant symptom burden.
(1) To examine demographic and clinical characteristics associated with high symptom burden in the homebound, and (2) to examine associations between symptom burden and time to hospitalization, nursing home placement, and death.
318 patients newly enrolled in the Mount Sinai Visiting Doctors Program, an urban home-based primary care program, were studied. Patient sociodemographic characteristics, symptom burden (measured via the Edmonton Symptom Assessment Scale, ESAS), and incidents of hospitalization, nursing home placement, and death were collected via medical chart review. Multivariate Cox proportional hazards models were used to analyze the effect of high symptom burden on time to first hospitalization, nursing home placement, and death.
43% of the study sample had severe symptom burden (i.e. ESAS score ≥6 on at least one symptom). Patients with severe symptom burden were younger (82.0 vs 85.5 yrs, p<0.01), had more comorbid conditions (3.2 vs 2.5 Charlson score, p<0.01), higher prevalence of depression (43.4% vs 12.0%, p<0.01), lower prevalence of dementia (34.3% vs 60.6%, p<0.01), and utilized fewer hours of home health services (86.2 vs 110.4 hrs/wk, p<0.01). Severe symptom burden was associated with a shorter time to first hospitalization (hazards ratio=1.51, 95% CI 1.06–2.15) in adjusted models, but had no association with time to nursing home placement or death.
The homebound with severe symptom burden represent a unique patient cohort who are at increased risk of hospitalization. Tailored symptom management via home-based primary and palliative care programs may prevent unnecessary healthcare utilization in this population.
There is growing recognition of the prevalence of symptoms associated with chronic diseases in the elderly. Common conditions such as dementia, congestive heart failure, chronic obstructive pulmonary disease (COPD), and cancers are often accompanied by substantial symptom burden, including symptoms such as depression, nausea, shortness of breath, fatigue, and pain (1–5). Currently an estimated two out of every three older adults live with multiple chronic medical conditions and as the US population ages and the number of chronically ill older adults with multimorbidity increases, the burden of resulting symptoms will continue to rise (3, 6, 7). High symptom burden, independent of associated underlying diseases, has been linked to poorer quality of life, increased mental and physical disability, greater use of health care services, and increased risk for morbidity and mortality (2, 8–13).
The estimated 2 million elderly homebound living in the US have a disproportionate chronic disease burden (14). The homebound often have complex and interrelated medical and social comorbidities that contribute both to substantial functional and cognitive limitations, greater use of health care services, and higher mortality rates compared to the non-homebound elderly population (11, 14–19). Unfortunately, the homebound have also historically lacked adequate primary medical care due to disabilities limiting access to routine care, lower income, or inability to find appropriately trained and experienced providers to manage their complex health care needs (14, 20–24).
A growing body of literature suggests that this homebound population suffers from substantial symptom burden related to their chronic conditions (11, 20, 25, 26). In a recent study of a nationally representative sample of Medicare beneficiaries, the homebound reported more than twice as many symptoms as their non-homebound counterparts (25). While symptom burden has been documented as an independent predictor of hospitalization (2, 12), nursing home placement (13), and mortality (2) in community-dwelling older adults, the impact of symptom burden on these healthcare outcomes in the homebound has not been studied. As the number of elderly individuals in the US continues to grow, the homebound population will grow as well (27). A better understanding of symptom burden and its consequences in the homebound population is essential in order to improve patient quality of life, improve healthcare service delivery, and decrease unnecessary healthcare utilization among the homebound.
Therefore we 1) examined characteristics associated with high symptom burden in the chronically ill homebound and 2) examined associations between high symptom burden and time to hospitalizations, nursing home placement, and death.
Mount Sinai Visiting Doctors (MSVD) is the largest academic home-based primary care program in the United States. Described elsewhere in detail (28, 29), the MSVD team of physicians, social workers, nurse practitioners, and nurses collaborate with community nursing agencies to provide multidisciplinary primary care to more than 1,000 homebound individuals in Manhattan annually (30, 31). Any patient over the age of 18 living in Manhattan who meets the Medicare homebound definition (able to leave home only with great difficulty and for absences that are infrequent or of short duration) is eligible to enroll in MSVD. In addition to routine and urgent home visits, the MVSD team provides extensive telephonic support and disease management between visits (30, 31). High priority is placed on quality of life, comfort, and minimizing unnecessary emergency department visits and hospitalizations (28).
This observational cohort study included all patients newly enrolled in the MSVD program between September 2008 and February 2010. Patients were followed over time from enrollment in the MSVD program until discharge from the program (e.g., due to death, nursing home placement, changed providers) or until the censoring date of July 5, 2014. The study was approved by the Mount Sinai IRB.
Baseline characteristics for all patients were collected by the primary care physician as a routine part of the comprehensive initial home visit for all newly enrolled patients, as described in Wajnberg 2012. Measures included: 1) sociodemographic variables such as age, gender, race, insurance type, primary language, and living situation, 2) functional status as measured by ability to perform activities of daily living (ADLs: physical ambulation, feeding, dressing, grooming, bathing, toileting, and incontinence) and instrumental activities of daily living (IADLs: telephone use, shopping, food preparation, housekeeping, house repairs, laundry, transportation, taking medicine, and financial management) (32), 3) home health and nursing service utilization, 4) comorbidity as measured using the Charlson Comorbidity Index (CCI) (33), and 5) symptom burden as measured by the Edmonton Symptom Assessment Scale (ESAS) (34). The ESAS consists of visual analogue scales (VAS) scored from 0 (indicating no symptoms) to 10 (the worst possible symptom severity) for 10 symptoms, including pain, tiredness, nausea, depression, anxiety, drowsiness, appetite, well-being, and shortness of breath, and other (to be filled out by the provider) (34). The ESAS was selected as it is a standardized, simple (with only 10 VAS scales) and short (requiring approximately 5 minutes to complete) symptom assessment tool that has been validated for both patient and caregiver report in a number of care settings, including those with older people with multimorbidities (34–37); whether assistance was required to complete the ESAS was noted. Based on previous research, patients were categorized as having a severe symptom burden if they had a score greater than 6 on any one or more ESAS symptoms (38).
Study outcomes included incidence and timing of: 1) hospitalizations and 30-day readmissions at Mount Sinai Hospital, 2) nursing home placement, and 3) death. Location of death was also determined.
Baseline characteristics and study outcomes were abstracted via review of patients’ electronic medical records.
Study variables were descriptively characterized, with means and standard deviations reported for continuous variables and frequency and percentages reported for categorical variables.
Severe and non-severe symptom burden patients were compared with respect to baseline measures using Pearson Chi-square test or independent sample t-tests. Outcomes related to hospitalization, nursing home placement, and death were descriptively characterized and incidence and rates of hospitalization were calculated. Severe and non-severe symptom burden patients’ hospitalization, nursing home, and death rates were compared using Pearson Chi-square test or independent sample t-tests.
Given symptom burden was assessed at patient admission and length of follow-up was highly variable (range = 0 to 68 months), we used adjusted Cox proportional hazards models with 95% confidence intervals to investigate the association between severe symptom burden and time to (1) hospitalization, (2) nursing home placement, and (3) death. By accounting for differential follow-up times by modeling time to event rather than any event occurrence, the Cox proportional hazards regression provides a more clinically relevant estimate of risk. Unadjusted Kaplan Meier survival curves were plotted to model the relationship between symptom burden and time to hospitalization, nursing home placement, and death.
A p value <0.05 was considered statistically significant and all analyses were conducted using SAS version 9.3 (SAS Institute, Inc., Cary, NC).
A total of 318 homebound patients were followed for up to 6 years after enrollment, with an average follow-up time of 2.2 years. The typical MSVD patient was elderly, female, non-White, English speaking, had significant functional impairment, multiple chronic diseases, and received home health and nursing services (Table 1).
Forty-five percent of the study sample was classified as having severe symptom burden. The most severe symptoms reported were tiredness (average ESAS score of 5.8 out of 10, 49% reported), lack of well-being (5.3, 47%), anxiety (5.3, 27%), and depression (5.3, 33%). Patients classified as having severe symptom burden had a statistically significant (p < 0.01) higher prevalence and higher average ESAS scores for all symptoms except for anxiety.
Patients with and without severe symptom burden varied on a number of key characteristics at baseline (Table 1). Compared to non-severe symptom burden patients, patients with severe symptom burden were younger (82.0 vs 85.5 years, p < 0.01), more independent in IADLs (36% vs 48% with IADL scores of 0 or 1, p < 0.05) and utilized fewer hours of home health services per week (86.2 vs 110.4 hours, p < 0.01). Patients with severe symptom burden also had more comorbid conditions (3.2 vs 2.5 Charlson score, p < 0.01), with a higher prevalence of depression (43% vs 12%, p < 0.01) and diabetes (43% vs 25%, p < 0.01) and a lower prevalence of dementia (34% vs 61%, p < 0.01). Patients with severe symptom burden were more likely to self-complete the ESAS (61% vs 49%, p = 0.03).
A total of 161 patients (50.6%) were hospitalized at least once during the study, with an average rate of 1.2 hospitalizations per year (Table 2). There were no significant differences in number, rate, or average time to hospitalizations between symptom burden groups. Of the patients who were hospitalized, 39 (24%) were readmitted to the hospital at least once within 30 days. A significantly lower percentage of patients with severe symptom burden were readmitted within 30 days of a hospitalization (18% vs 31%, p < 0.05). Approximately one-eighth (12.6%) of all patients were placed in a nursing home and over half (57.2%) died during the study period. Of patients who died, 65% died at home. There were no significant differences between symptom burden groups in nursing home placement and death.
Results from the unadjusted Cox regression model showed an increased risk for time to first hospitalization for patients with severe symptom burden (hazards ratio, HR = 1.75, 95% confidence interval, CI = 1.28 – 2.40), as also evidenced by the Kaplan Meier curves in Figure 1. After controlling for age, depression, disease burden (CCI >5), and self-completion of the ESAS, severe symptom burden patients were still 50% more likely to be hospitalized compared to non-severe symptom burden patients (HR = 1.51, 95% CI = 1.06 – 2.15) (Table 3). Severe symptom burden was not a significant predictor of time to nursing home placement or death.
In this study, homebound patients with severe symptom burden tended to be younger with more comorbid conditions and less dementia, IADL impairment, and utilization of home health services. Among this homebound elderly patient population, severe symptom burden was an independent predictor of time to first hospitalization.
These findings confirm that in our homebound population, multimorbidity was associated with severe symptoms. While dementia is a common reason for referral to MSVD, those with dementia were less likely to have severe symptom burden. Those with dementia were also less likely to self-report symptoms on the ESAS as compared to patients without dementia (36.8% vs 78.3%) and it is possible that caregivers underreported the symptoms that dementia patients suffered from. However, the ESAS has been validated for caregiver responses for symptom burden assessment and prior studies also suggest that informal primary caregivers tend to overestimate symptom intensity (34, 36, 37, 39–44). Thus, rather than underreporting, we believe these findings reflect that those with dementia had less chronic diseases and therefore less severe symptom burden; we believe that multimorbidity drives severe symptom burden in our population.
It is also notable that patients with severe symptom burden used on average 20 fewer hours of home health services per week as compared to those without severe symptom burden. These observed differences may be due to the fact that patients with severe symptom burden had less dementia and less IADL impairment and therefore had less need for home health services. However, given there was significant ADL impairment in both patients with and without severe symptom burden, these differences could also be due to unmet health care needs with symptomatic patients underutilizing needed home health services. Further study is needed to explore how home health care services impact patient symptom burden.
Like previous studies of community dwelling elders in general (2, 12), we found that severe symptom burden was associated with shorter time to first hospitalization in the homebound. Time to first hospitalization best demonstrates the impact of severe symptom burden on hospitalization in this study due to varying periods of patient follow-up. The association between symptom burden and hospitalization may be in part due to the fact that those with severe symptom burden rely more on hospitals for acute symptom management and care, resulting in earlier and more frequent hospitalizations. For these patients, their symptom burden rather than the presence or absence of chronic disease itself may be driving their utilization. This demonstrates an opportunity for clinical intervention; while a clinician is unlikely to reverse a patient’s severe osteoarthritis, adequate treatment of pain may not only improve the patient’s quality of life but also decrease his/her risk of hospitalization. In addition, it suggests the need for development and incorporation of symptom management protocols into the care of this patient population. Further studies should explore how the use of validated symptom assessment instruments to monitor symptom burden can not only guide symptom management but also impact hospitalization rates.
Unlike previous studies, in our population severe symptom burden did not predict nursing home placement or death as it had in community-dwelling elders (2, 13). This may be because many patients enter the MVSD program specifically to avoid nursing home placement and this contributes to a low nursing home placement rate overall. Future nursing home placements may have also been missed given the average follow-up time of only 2.2 years. Similarly, a high baseline mortality rate at MSVD may have limited the potential predictive value of symptom burden severity. It is also important to note that the studies conducted in community dwelling elders used a composite score to measure symptom burden (2, 12, 13). We chose to define symptom burden as a score of 6 or greater in one or more symptoms on the ESAS as this definition was more clinically relevant in our practice; severe symptoms in any category not only identifies a discreet cause of patient suffering but also provides a potential opportunity to intervene.
Symptom burden for pain, anxiety, depression, and tiredness in homebound patients has been successfully managed and decreased with home based primary care (HBPC) programs (45). Fortunately, the number of home based primary care (HBPC) programs that are able to provide comprehensive treatment of chronic conditions and symptoms for the homebound is growing (23, 26, 29, 46). These programs have already been shown to improve the health of the homebound population while decreasing healthcare spending (47–52). With 45% of patients in the current cohort reporting severe symptom burden, the need for the incorporation of palliative care into HBPC programs is evident. Many models of HBPC are emphasizing the importance of palliative approaches for people with complex chronic illnesses. Evidence in non-homebound elders suggests that for older primary care patients, interventions such as in-home palliative care programs and “symptom treatment” protocols can lead to significant symptom improvement, higher satisfaction with services, and reductions in health care utilization (23, 48, 53–56). Further study should specifically assess if effective symptom management has a similar impact in the homebound.
Our study had several limitations. First, symptom burden data was only available for a single time point and our analysis could therefore not capture the dynamic changes in symptom burden that may occur for patients over time. Second, we utilized a definition of severe symptom burden that included 45% of the study sample and did not differentiate between specific symptoms. While we believe our definition of symptom burden reflects clinically useful indices of disease and health needs, we recognize that other definitions have been and could be used with potentially differing results. Third, we limited our chart review to hospitalizations at the Mount Sinai Hospital System, potentially underestimating the number of hospitalizations in this patient cohort. And lastly, we studied a sample of elderly homebound patients enrolled in an urban home-based primary and palliative care program; the results from this study may not be generalizable to the broader homebound patient population or to homebound elderly enrolled in other in-home care programs.
Despite these limitations, our research characterizes homebound elders with multiple medical conditions who are most likely to have severe symptom burden. This study also describes a positive association between severe symptoms and time to hospitalization in the elderly homebound population. Because of their multiple chronic medical issues, the symptomatic homebound need a more holistic approach to care that home based primary care and palliative care programs are uniquely positioned to provide. As this patient population grows, the need for innovative models of in-home symptom burden management will increase. We propose that symptom burden severity, irrespective of the specific symptom or underlying chronic disease, may be an important way to guide symptom management and treatment strategies with the potential benefit of reducing unnecessary hospitalizations.
The authors would like to acknowledge the Mount Sinai Visiting Doctors program for their support.
Nancy Yang received funding from the Mount Sinai Patricia S. Levinson Summer Fellowship for Community Oriented Research and Service and support from the Medical Student Training in Aging Research program to complete this research. Dr. Ornstein was supported by: National Institute on Aging K01AG047923 and The National Palliative Care Research Center. The authors declare no potential personal or financial conflicts of interest.
Nancy Yang, Department of Geriatrics and Palliative Medicine, Icahn School of Medicine at Mount Sinai. New York, NY, USA.
Katherine Ornstein, Department of Geriatrics and Palliative Medicine, Icahn School of Medicine at Mount Sinai. New York, NY, USA.
Jennifer M. Reckrey, Department of Geriatrics and Palliative Medicine, Icahn School of Medicine at Mount Sinai. New York, NY, USA.