Symptom Burden Among Cancer Survivors: Impact of Age and Comorbidity

Article Figures & Data

Download figure
Open in new tab
Figure 1.
Probability of ongoing symptoms stratified by cancer survivorship, age, and sex. The final model included cancer survivorship, age, sex, education, race/ethnicity, and income (see Table 2). Average prediction was based on this model stratified by cancer survivor status, age, and sex.

Table 1.

Sociodemographic Characteristics of Respondents With and Without a History of Cancer

Characteristics	Cancer Survivors (n = 1,904)	General Population Controls (n = 29,092)	P^*
Age (years)
<50	22.9	65.3	<.001
50 to 64	26.9	20.8
≥65	50.2	13.9
Sex
Male	41.2	48.4	<.001
Female	58.8	51.6
Education attainment
High school graduate or less	52.4	46.1	<.001
Some college or technical school	27.4	29.2
College or more advanced degree	20.2	24.7
Race/ethnicity
Hispanic	3.8	11.5	<.001
Non-Hispanic white	89.0	72.2
Non-Hispanic black	5.4	11.7
Other (non-Hispanic)	1.8	4.6
Marital status
Currently married/living with partner	64.2	63.8	0.74
Not currently married/partnered	35.8	36.3
Annual household income
<$20,000	24.6	19.1	<.001
≥$20,000	75.4	80.9
Comorbid conditions
None	39.2	76.4	<.001
One	39.1	18.1
Two	12.2	3.9
Three or more	9.5	1.6
Time since cancer diagnosis (years)
1	18.4
2 to 5	29.8
6 to 10	18.7
≥11	33.1

Data source: National Health Interview Survey 2002; weighted percentages were extrapolated from the US census (2000) for the adult civilian. All data presented as percentages.
* Categorical variables were compared with univariate logistic regression; P values are two-sided.

Table 2.

Symptom Burden Among Cancer Survivors Differs by Age

Characteristics	Pain			Psychological Distress			Insomnia
Characteristics	OR	95% CI	P	OR	95% CI	P	OR	95% CI	P
CS versus control (unadjusted)	2.44	2.16–2.74	<.001	1.98	1.76–2.22	<.001	2.09	1.83–2.38	<.001
Age (years)
<50	2.96	2.28–3.85	<.001	2.94	2.35–3.67	<.001	2.70	2.09–3.48	<.001
50 to 64	2.23	1.76–2.83	<.001	1.84	1.45–2.34	<.001	1.80	1.40–2.31	<.001
>64	1.36	1.14–1.63	.001	1.48	1.23–1.79	<.001	1.44	1.18–1.75	<.001
Sex (female vs male)	1.24	1.15–1.32	<.001	1.51	1.39–1.64	<.001	1.47	1.37–1.59	<.001
Education
High school or less^*	1			1			1
Some college	0.98	0.90–1.06	.56	0.86	0.79–0.94	.001	1.01	0.92–1.10	.877
College or more advanced	0.73	0.66–0.81	<.001	0.57	0.51–0.64	<.001	0.70	0.63–0.76	<.001
Race/ethnicity
Non-Hispanic white^*	1			1			1
Non-Hispanic black	0.84	0.75–0.94	.003	0.90	0.80–1.02	.102	0.78	0.70–0.88	<.001
Hispanic	0.55	0.48–0.63	<.001	0.85	0.75–0.96	.007	0.77	0.69–0.87	<.001
Other (non-Hispanic)	0.62	0.49–0.78	<.001	0.67	0.52–0.86	.002	0.57	0.45–0.72	<.001
Annual household income
(<$20,000 vs. ≥$20,000)	1.37	1.25–1.51	<.001	1.85	1.69–2.04	<.001	1.52	1.39–1.67	<.001

* Models included cancer survivor status and age as interaction terms and controlled for sex, education, race/ethnicity, and income.
CS, cancer survivors; OR, odds ratio; CI, confidence interval.

Table 3.

Comorbidity and Increased Symptom Burden Among Cancer Survivors

Comorbidities^* (n)	Pain		Psychological Distress		Insomnia
Comorbidities^* (n)	Control	CS	Control	CS	Control	CS
0	1 (1.00–1.00)	2.23 (1.73–2.86)	1 (1.00–1.00)	2.37 (1.86–3.00)	1 (1.00–1.00)	2.38 (1.86–3.03)
1	7.44 (6.67–8.29)	10.02 (8.08–12.43)	3.37 (3.08–3.70)	4.57 (3.71–5.61)	3.52 (3.21–3.86)	4.54 (3.64–5.66)
2	15.44 (13.33–17.87)	17.39 (12.95–23.35)	5.62 (4.90–6.43)	7.38 (5.46–9.99)	6.38 (5.57–7.30)	7.71 (5.60–10.60)
3+	27.34 (22.49–33.23)	35.56 (25.34–49.91)	11.26 (9.37–13.54)	15.12 (11.08–20.64)	9.92 (8.36–11.78)	10.25 (7.40–14.21)

Models included comorbidity and cancer survivor status as interaction terms and were controlled for age, sex, education, race/ethnicity, and income. Data presented as odds ratios (95% confidence intervals).
* Comorbidities were measured by the number of health conditions causing functional limitations.
CS, cancer survivors.

Table 4.

Symptom Burden by Cancer Type and Time Since Cancer Diagnosis, Adjusted for Age, Education, Race/Ethnicity, and Income

Analysis included only cancer survivors. Data presented as odds ratios (95% confidence intervals).
* Because several cancer types are sex specific, the models did not include sex as a covariate.
† P <.05.
‡ P <.01.