High Symptom Burden and Low Functional Status in The Setting of Multimorbidity

Abstract

Objective

To enhance understanding of the relationship between multimorbidity, symptom burden, and functional status among patients with life-limiting illness.

Design

Secondary analysis of baseline data from a randomized clinical trial conducted in the Palliative Care Research Cooperative Group (PCRC). Group differences were tested with a t-test; multivariate regression analysis was used to determine effect of multiple variables on functional status and symptom burden.

Participants

Adults (N=381) recruited from 15 Palliative Care Research Cooperation sites who participated in the parent statin-discontinuation clinical trial were included in the analysis. Inclusion criteria for the trial was: diagnosis of a life-limiting illness, statin use for ≥ 3 months, life expectancy >1 month, and declining functional status.

Measurements

Cancer diagnosis: solid organ and hematologic malignancies; Multimorbidity: >2 on Charlson Comorbidity Index (CCI); High symptom burden: a) number of symptoms with a severity score >4, and b) any score > 7 on Edmonton Symptom Assessment Scale (ESAS); Functional status: Australia-modified Karnofsky Performance Scale (AKPS).

Results

Fifty-one percent had a primary diagnosis of cancer, mean age 74.1 years (SD 11.6). Participants had multiple comorbid illnesses (CCI=4.9, SD=2.8), multiple symptoms (ESAS=27.2, SD=15.9) and reduced functional status (AKPS=53, SD=13). In both univariate and multivariate analyses, multimorbidity was associated with greater symptom burden (4.2 vs. 3.1 moderate or severe symptoms (t=−3.2, p=.002); 12% vs. 6% with severe symptoms (t=−3.7, p<0.001), while cancer diagnosis was not. In both univariate and multivariate analyses, higher symptom burden was associated with lower functional status (F = 11.6, p<0.001) while multimorbidity was not.

Conclusion

Symptoms cannot be attributed solely to a diagnosis of cancer. The association between symptom burden and functional status underscores the importance of clinical attention to symptoms in patients with multimorbidity.

INTRODUCTION

Multimorbidity, while difficult to define¹ typically refers to the presence of two or more chronic conditions². Multimorbidity is experienced by 70% of adults over the age of 75³. Nearly 55% of Medicare recipients with stroke and heart failure have five or more additional chronic conditions⁴. Increases in the prevalence of multimorbidity is anticipated due to rapid growth in the aging population, medical advances, and improved longevity^3,5.

Multimorbidity is negatively associated with health status and quality of life⁶, and is related to increased health care utilization, costs^7,8, and mortality rates^5,9. Interventions targeting the assessment, treatment, and symptom management of patients with multimorbidity^10,11 are lacking, especially for those with advanced, life-limiting illness. Additionally, multimorbidity is often excluded from disease specific research¹². While there are some data describing the impact of multimorbidity on health outcomes, little is known about the impact of multimorbidity in the setting of life-limiting illness^13,14. This analysis investigates the relationship between multimorbidity, symptom burden, and functional status in those with life-limiting illness. Better understanding the relationship between symptom burden, functional status and multimorbidity has the potential to inform improvements in care for older adults with life-limiting illness. We sought to understand: a) the relationship between multimorbidity and symptom burden, controlling for functional status, primary diagnosis of cancer, age and gender; and b) the relationship between multimorbidity and functional status controlling for symptom burden, primary diagnosis of cancer, age, and gender.

METHODS

This was a secondary analysis of baseline data from a multi-site randomized clinical trial of statin discontinuation conducted in the Palliative Care Research Cooperative Group (PCRC). The results of the parent trial are published elsewhere¹⁵. All participants of the parent trial were included in the analysis. Eligibility criteria included having been on statin for ≥3 months for primary or secondary prevention, a progressive life-limiting illness which was determined by the treating physician stating that they would not be surprised if the patient died in the next year, life expectancy >1 month also determined by the treating physician, and declining functional status defined as a reduction in Australia-modified Karnofsky Performance Status (AKPS) score to <80% in the previous 3 months. Cancer diagnosis was defined as a primary diagnosis of solid organ or hematologic malignancies. Multimorbidity was defined as a score >2 on the Charlson Comorbidity Index (CCI), which weights a combination of chronic conditions documented from the medical record based on a scale of 1–6, with higher scores indicating increased multimorbidity¹⁶. Symptom burden was measured by the Edmonton Symptom Assessment Scale (ESAS) which captures ratings of symptom severity for 9 symptoms (pain, tiredness, nausea, depression, anxiety, drowsiness, appetite, shortness of breath, and wellbeing) on scale of 0–10 for a total score calculated between 0–90 with higher scores indicating greater symptom burden¹⁷. An individual symptom score > 4 was considered to be ‘moderate’, and > 7 was considered to be ‘severe’. Moderate symptom burden was defined as the number of symptoms with a severity score >4, symptoms severity was defined as the proportion of symptoms with a severity score >7. Functional status, reported by the patient, was measured by the Australia-modified Karnofsky Performance Scale (AKPS) with scores ranging from 100 (normal, no complaints, no evidence of disease) to 0 (dead)¹⁸. Group differences were tested with a t-test; multivariate regression analysis was used to determine effect of multiple variables on functional status and on symptom burden (Supplementary Figure S1. Conceptual Model for Multivariate Analyses).

RESULTS

Patients (N=381) had multiple comorbid illnesses, numerous symptoms and, consistent with study entry eligibility criteria, reduced functional status (Table 1). The most common non-cancer comorbid illnesses were chronic obstructive pulmonary disease (32%), congestive heart failure (26%) and cardiovascular disease (22%).

Table 1.

Patient Characteristics (N=381)

Characteristic
Female, n (%)	171 (44.9)
Age, mean±SD	74.1±12
Primary diagnosis of cancer, n (%)	197 (51.7)
Non-cancer primary diagnosis, n (%)
Chronic obstructive pulmonary disease	47 (12.3)
Congestive heart failure	32 (8.4)
Dementia	28 (7.3)
Renal disease	16 (4.2)
Other	61 (16.1)
Most common non-cancer comorbid illness, n (%)
Chronic obstructive pulmonary disease	120 (31.5)
Congestive heart failure	99 (26.0)
Cardiovascular disease	83 (21.8)
Multimorbidity (CCI), mean±SD	4.9±2.8
Symptom burden (ESAS), mean±SD	27.2±15.9
Functional status (AKPS), mean±SD	53±13.0

Charlson Comorbidity Index (CCI), Edmonton Symptom Assessment Scale (ESAS), Australia-modified Karnofsky Performance Scale (AKPS)

Univariate results are summarized in Table 2. Presence of multimorbidity was associated with the number of symptoms with ESAS score >4 (P = .002), and the proportion with severe symptoms, ESAS >7, P < .001. However, multimorbidity was not associated with functional status. Those with a primary diagnosis that was not cancer had worse functional status (AKPS 48.5±12.9) compared to those with a primary diagnosis of cancer (AKPS 58.1±11.4), P < .001. There were no univariate associations between a primary diagnosis of cancer and symptom burden.

Table 2.

Univariate Relationships between Multimorbidity and Cancer Diagnosis and Symptom Burden and Functional Status

			T-Test	P-Value
	Multimorbidity Present (CCI >2)
	Yes	No
Number of Symptoms > 4 (ESAS)	4.2±2.8	3.1±2.5	−3.2	.002
Proportion of Severe Symptoms (ESAS >7)	.12±.16	.06±.10	−3.7	<.001
Functional Status (AKPS)	53.5±13.0	53.3±13.2	−0.1	.89
	Primary Diagnosis
	Cancer	Non-Cancer
Number of Symptoms > 4 (ESAS)	4.06±2.7	3.8±2.8	−0.8	.42
Proportion of Severe Symptoms (ESAS > 7)	.10±.15	.11±.15	0.05	.96
Functional Status (AKPS)	58.1±11.4	48.5±12.9	−7.8	<.001

Charlson Comorbidity Index (CCI), Edmonton Symptom Assessment Scale (ESAS), Australia-modified Karnofsky Performance Scale (AKPS)

Multivariate analyses (Table 3) found that multimorbidity was associated with greater symptom burden (t = 2.44, P = .02), but not with functional status. Cancer diagnosis was not independently associated with symptom burden, while decreased functional status was associated with greater symptom burden (t = −3.40, P = .001). Non-cancer primary diagnosis and greater symptom burden were associated with worse functional status (t =4.03, P < .001; t =−3.40, P =.001).

Table 3.

Multivariate Results^a

	Estimate	SE	T Value	P-Value
Outcome: Symptom Burdenb
Multimorbidity	.92	.38	2.44	.02
Cancer Primary Diagnosis	.16	.35	.45	.65
Functional Status (AKPS)	−.05	.01	−3.40	.001
Outcome: Functional Statusc
Multimorbidity	−1.79	1.61	−1.11	.27
Cancer Primary Diagnosis	5.89	1.46	4.03	<.001
Number of Moderate or Severe Symptomsb	−.85	.25	−3.40	.001

^aadjusted for age and sex

^bSymptom burden defined as number of symptoms >4 Edmonton Symptom Assessment Scale (ESAS)

^cAustralia-modified Karnofsky Performance Scale (AKPS)

DISCUSSION

We found that multimorbidity was associated with symptom burden, but not with functional status. We also identified that in this population, a cancer diagnosis was associated with better functional status, while greater symptom burden was associated with worse functional status. Those with multimorbidity experienced higher symptom burden; while participants with a non-cancer primary diagnosis and those with greater symptom burden had lower functional status.

While multimorbidity was associated with higher symptom burden cancer was not. Severe and persistent symptoms have previously been linked to multimorbidity among community dwelling older adults^19,20, and increased symptom burden has also been associated with multimorbidity among cancer patients^21,22. These findings in concert with that of previous studies suggest that in many instances, symptom burden is driven by noncancer diagnoses, whether the patient has cancer or not. In contrast to previous research indicating an association between multimorbidity, impairments in function²³ and health related quality of life²⁴, our data did not show a relationship between multimorbidity with functional status, while controlling for symptom burden. However, multimorbidity has been associated with symptoms that are directly related to restrictions in activities of daily living, particularly in the last 5 months of life²⁵. Therefore, symptoms may be the pathway by which multimorbidity influences functional status.

Symptom burden, rather than multimorbidity, was associated with lower functional status. Previous studies highlighting the relationship between symptom severity, quality of life and functional status corroborate this result^26–28. Symptom burden may be a mediating variable between multimorbidity and functional status. However, it is also possible that specific symptoms, such as depression and fatigue, drive this relationship, rather than overall moderate to severe symptom burden. For example, both depression and fatigue have previously been shown to exacerbate symptom burden and functional limitations among patients with life-limiting illness^29–31. Therefore, our results may be detecting the relationship between depression, fatigue, and functional status. The potential association between specific symptoms, symptom burden and functional status underscores the importance of clinical attention to symptoms in patients with multimorbidity.

Although cancer was not independently related to symptom burden, patients with a primary diagnosis of cancer had better function compared to those with a non-cancer diagnosis. Our findings suggest that symptoms experienced by patients with life-limiting illness cannot be attributed solely to a diagnosis of cancer. Previous studies suggest that patients with life-limiting illness experience high symptom burden, particularly pain, regardless of diagnosis³². Multimorbidity is also common among cancer patients, and clinicians have difficultly addressing these patients’ symptoms²². Providers in cancer care may not consider that symptoms might be emanating from conditions other than cancer. As such, cancer patients with multimorbidity have been found to take more medications, experience increased illness burden, and are at greater risk of under treatment³³.

Given the association between multimorbidity and symptom burden, and the relationship of symptom burden with functional status, these findings suggest careful clinical attention should be made to symptoms, regardless of diagnosis, particularly in the setting of multimorbidity. However, symptom management is more complex in the context of multimorbidity due to limitations in pharmacologic options, multi-provider coordination, difficulty applying standardized guidelines, and complicated self-care regimens^34,35. Our findings support the growing evidence recommending a shift from disease-specific treatment to addressing symptoms in the context of multimorbidity. Nonpharmacologic behavioral treatments with the potential to improve more than one kind of symptom may be particularly valuable in the context of multimorbidity³⁶.

There are limitations to this analysis. This was a secondary analysis of data collected from a clinical trial. Because participants agreed to participate in the trial and were all using statins at the time of study entry, the findings may not represent the diagnosis distribution, symptom burden or functional status of all older adults with life-limiting illness. Although our models were adjusted for age, gender, symptom burden, and functional status, it is likely that we were unable to control for every confounding variable among the tested relationships. There was also a difference in patient survival by diagnosis; patients with a non-cancer primary diagnosis lived longer than those with cancer. Multimorbidity was assessed as a presence or absence of multiple comorbid conditions, but not severity of the conditions. An indicator of illness severity in addition to multimorbidity might have demonstrated a stronger relationship between multimorbidity and function. Inclusions of diagnosis not specified in the CCI may also have strengthened our measurement of multimorbidity. In addition, functional status may not have been fully captured by the AKPS. The AKPS only looks at generic assistance and is subject to interpretation of what that assistance constitutes. Other functional measures that provide more granularity of assistance might have identified other relationships³⁷.

In conclusion, our data suggest that multimorbidity is associated with symptom burden. Our work also suggests that symptom burden is associated with decreased functional status. Future research should focus on better understanding the relationship between symptom burden and multimorbidity. Improved measures to include both diagnosis and illness severity for capturing multimorbidity and effective approaches to address symptom burden in older adults with life-limiting illness and multimorbidity are needed.

Supplementary Material

Supp FigS1

Supplementary Figure S1. Conceptual Model for Multivariate Analyses