The Association Between Noncancer Pain, Cognitive Impairment, and Functional Disability: An Analysis of the Canadian Study of Health and Aging

Abstract

Background.

Noncancer pain and cognitive impairment affect many older adults and each is associated with functional disability, but their combined impact has yet to be rigorously studied.

Methods.

This is a cross-sectional analysis of the Canadian Study of Health and Aging. Pain was collapsed from a 5-point to a dichotomous scale (no and very mild vs moderate and greater). Cognitive status was dichotomized from the Modified Mini-Mental State Examination (0–100) to no (>77) or mild-moderate (77–50) impairment. Five Instrumental Activities of Daily Living (IADL) and seven Activities of Daily Living (ADL) were self-rated as “accomplished without any help” (0), “with some help” (1), or “completely unable to do oneself” (2) and then summed to create a composite score of 0–10 and 0–14, respectively. Multivariate linear regression analysis was conducted to determine the associations between self-reported functional status with moderate or greater pain, cognitive impairment, and the interaction of the two.

Results.

A total of 5,143 (90.2%) participants were eligible, 1,813 (35.6%) reported pain at a moderate intensity or greater and 727 (14.3%) were cognitively impaired. The median IADL and ADL summary scores increased among the pain and cognition categories in the following order: no pain and cognitively intact (0.63 SD 1.24, 0.23 SD 0.80), pain and cognitively intact (1.18 SD 1.69, 0.57 SD 1.27), no pain and cognitively impaired (1.64 SD 2.22, 0.75 SD 1.57), and pain and cognitively impaired (2.27 SD 2.47, 1.35 SD 2.09), respectively. Multivariate linear regression found IADL summary scores were associated with pain, coefficient .17 (95% confidence interval [CI] 0.07–0.26), p < .01; cognitive impairment, coefficient .67 (95% CI 0.51–0.83), p < .01; and an interaction effect of pain with cognitive impairment, coefficient .24 (95% CI 0.01–0.49), p = .05. ADL summary scores were associated with pain coefficient .10 (95% CI 0.04–0.17), p < .01 and cognitive impairment, coefficient .29 (95% CI 0.19–0.39), p < .01, but had a nonsignificant interaction term, coefficient .12 (95% CI −0.03 to 0.29), p = .12.

Conclusions.

Noncancer pain and cognitive impairment are independently associated with IADL and ADL impairment and IADL impairment is even greater when both conditions are present.

MANY conditions associated with noncancer pain, such as osteoarthritis, spinal stenosis, and osteoporosis, are more prevalent with age and often affect multiple physical sites (1–3). Given this, it is not surprising that noncancer pain is reported by more than 40% of community-dwelling adults aged 75 or older (4–6). Of those reporting pain, a sizable proportion experience pain of at least moderate intensity. The impact of pain goes beyond physical discomfort and suffering as its presence is also associated with depression, decreased social interaction, and functional impairment (7,8).

The prevalence of cognitive impairment also increases with age so that by age 85 as many as 50% of older adults may be affected (9,10). The diagnosis of cognitive impairment is important as its course is typically characterized by gradual, yet progressive, memory loss and functional decline (11). The functional decline generally progresses in a stepwise fashion with the loss of more complex Instrumental Activities of Daily Living (IADL) such as driving, shopping, or cooking occurring earlier than basic Activities of Daily Living (ADL) such as bathing, walking, or dressing (12).

We conducted the current analyses to try to understand the relationships among noncancer pain, cognitive impairment, and functional disability. Our goal was to better characterize the contribution of self-reported pain and cognitive function to self-reported independence in IADL and ADL. We also wanted to determine whether noncancer pain and cognitive impairment interact, resulting together in more impaired physical function than can be explained by their independent contributions. A more in-depth understanding of these relationships is particularly relevant as functional status decrements are associated with poor outcomes including falls, institutionalization, and death (13,14). The clinical implication of such an interaction would be that effective pain management in those with cognitive impairment may help mitigate excessive functional decline.

METHODS

Participants

This study is a cross-sectional analysis of the Canadian Study of Health and Aging (CSHA). The CSHA was a national longitudinal study designed to provide information about the epidemiology of dementia. People aged 65 and older were recruited with representative samples drawn from 36 cities and their surrounding rural areas across Canada. Details of the study’s objectives, design, and methodologies have been described elsewhere (15,16). The interviews were conducted in participant homes by trained research personnel and included demographic characteristics along with health, disease, mood, physical function, and cognition measures.

This analysis includes participants who resided in the community and completed the questionnaire at the time of the second wave in 1996. Data from this wave incorporated a pain assessment measure that was not included in the other waves. Otherwise eligible participants were excluded from the analyses if the pain, cognition, or functional measure were missing, participants had greater than moderate cognitive impairment (as this affects the reliability of the pain and function questions) (8), or they reported having cancer in the past year (to focus only on noncancer pain). All participants provided written informed consent and the CSHA protocol was approved at each participating institution in Canada. The research plan for the secondary analyses was reviewed and approved by the University of Chicago's Institutional Review Board.

Measures

Pain and cognition were the independent variables in the current study. Noncancer pain was assessed using the 5-point Verbal Descriptor scale, which measures the presence and intensity of pain. The validity and reliability of this measure has been established in both cognitively intact and impaired populations (17–20). Participants were asked, “How much bodily pain have you had during the past 4 weeks?” The response categories offered were: 1 = none, 2 = very mild, 3 = moderate, 4 = severe, and 5=very severe. The responses were dichotomized as “none or very mild” and “moderate or greater” as the latter level is considered clinically relevant as it adversely affects health (21).

Cognitive status was assessed using the Modified Mini-Mental State Examination (3MS). The 3MS is similar to the Folstein Mini-Mental Exam, but adds four tasks (date and place of birth, animal naming, similarities, and a second delayed recall) and allows partial credit on some items (22). The 3MS has a range from 0 to 100 with lower scores indicating greater cognitive impairment. The 3MS has been validated in a variety of community-dwelling populations including the CSHA and a score of 77 or less indicates cognitive impairment with sufficient sensitivity and specificity to be used for screening purposes (23,24).

Functional status was ascertained using the Older Americans Resources and Services’ IADL and ADL (25). This self-report measure has previously been validated in cognitively intact persons as well as those with mild to moderate cognitive impairment (26). Each of five IADL (getting to distant places, using the telephone, going shopping, preparing own meals, and doing housework) and seven ADL (eating, dressing, personal care, walking, getting out of bed, taking a bath, and using the toilet) were included as part of the interview. Participants categorized their independence by indicating whether or not they could perform each activity without any help (0), with some help (1), or completely unable to do oneself (2). IADL and ADL self-reports were summed to develop a composite score for each, ranging from 0 to 10 for IADL and 0–14 for ADL, where higher scores indicate greater impairment. Functional impairment when used as a dichotomous variable was defined as any help needed to perform any of the five IADLs or any of the seven ADLs. In other words, impairment was defined as any summary score of 1 or greater for either measure.

The analyses also incorporated participant demographics, a depression screen (5-item mental health screening questionnaire) (27), self-reported medical comorbidities, and self-reported health status. Demographic variables included age, gender, race and ethnicity, education, and marital status. Mood was assessed using a 5-item self-report mental health screening questionnaire (27). Each of the five questions assess mental health (feelings of anxiety and depression) on a 6-point scale ranging from “none of the time” to “all of the time,” where higher scores indicate more symptoms of depression and total scores range from 5 to 30. Comorbidities were assessed by asking participants about conditions present in the past year from 15 body systems, and each was dichotomized as “yes” or “no.” The list included high blood pressure, heart or circulation problems, arthritis or rheumatism, Parkinson’s disease or other neurological problem, eye trouble, ear trouble, chest problems, trouble with stomach or digestive system, back problems, bladder control problems, problems controlling bowels, fractures, cancer, diabetes, and foot problems. Affirmative responses were summed to create an overall comorbidity index ranging from 0 to 15, with higher scores indicating greater morbidity. Self-reported health status was assessed using a 5-point scale ranging from poor to excellent health.

Statistical Analyses

A power analysis was conducted to determine the feasibility of using the CSHA data set to examine our hypothesis that noncancer pain and cognitive impairment interact to affect function in a nonlinear manner. In the CSHA, the prevalence of cognitive impairment was estimated to be 16%, so that among 5,000 participants, 800 persons would be cognitively impaired. Also, approximately 40% of participants, independent of cognitive status, had pain at a moderate intensity or higher, leaving 1,680 cognitively intact and 320 cognitively impaired participants with pain and 2,520 and 480 participants without pain, respectively. These sample sizes have an 80% power to detect an interaction effect between noncancer pain and cognition of 10%, assuming a Type I error rate of 5% and a two-tailed test.

Individual characteristics were analyzed by pain status (no or very mild pain versus moderate or greater pain) using chi-squared analysis for categorical variables. Mann–Whitney U tests (also referred to as Wilcoxon rank sum test) were used to analyze these differences for continuous variables due to their nonnormal distributions. Participants were subsequently divided into four groups based upon whether or not pain was reported at a moderate intensity or greater, and whether cognitive impairment was present. The four resultant groups were no pain and cognitively intact, pain and cognitively intact, no pain and cognitively impaired, and pain and cognitively impaired. Chi-square analysis was used to analyze differences among these aggregates.

Multivariate linear regression analysis was conducted to determine the associations between self-reported functional status with moderate or greater pain, cognitive impairment, and the interaction of the two. Analyses were performed separately for IADL and ADL summary scores as the dependent variable. Additional independent variables included factors significantly associated with the dependent variables in univariate analysis (p < .10), as well as factors that adjust for baseline differences in groups of interest. A correlation matrix was computed for all potential independent variables to assess collinearity and for variables that were highly correlated; inclusion in the final models was based on clinical relevance. Marital status was correlated with age (r = −.30) and male gender (r = .43) so it was left out of the multivariate models. Self-reported health was removed from the multivariate models because it was correlated with depression (r = .33) and comorbidity (r = .43). Statistical analyses were performed using STATA (version 10.0; StataCorp, College Town, TX). The funding source had no role in the study.

RESULTS

Of the 5,703 community-dwelling older adults who completed the 1996 wave of the CSHA, 560 were excluded: 478 were missing self-reported function (8.4%),191 had a 3MS score less than 50 (3.4%), and 116 had a cancer diagnosis in the past year (2.0%), leaving 5,143 (90.2%) of the original sample. Participant characteristics by reported level of pain is displayed in Table 1, in which 1,813 participants (35.6%) reported pain at a moderate intensity or greater. On average, persons with pain of this intensity were older, women, less educated, had higher depression scores and more comorbidities, and rated their health as poorer. No association was found between participants who reported this amount of pain and cognitive impairment, (p = .46).

Table 1.

Participant Characteristics by Severity of Noncancer Pain Self-Report, N = 5,086

Characteristic	Moderate or Greater Pain (N = 1813, 35.6%)	None or Mild Pain (N = 3273, 64.4%)	p Value
Age, mean (SD)	79.9 (6.0)	79.4 (6.1)	<.01
Female gender (%)	67.4	55.5	<.01
White race (%)	96.9	96.7	.98
Education (%)
8 years or less	31.9	28.9	.02
9–12 years	43.9	43.8
13 or more years	23.2	27.3
Marital status (%)
Single	6.5	6.7	.02
Married	45.7	49.6
Widowed/divorced/separated	47.8	43.7
Lives alone (%)	43.0	39.4	.01
Comorbidity, mean (SD)*	4.7 (2.3)	3.1 (2.0)	<.01
Health status, % very good or good	68.9	89.7	<.01
Depression score, mean (SD)†	11.0 (4.6)	8.8 (3.5)	<.01
Cognitively impaired‡ (%)	14.8	14.0	.46

Notes: * Comorbidity is the summary score of 15 possible chronic conditions reported by participants; higher scores indicate greater morbidity.

^†Depression score is the summed response to the mental health screening questionnaire; higher scores indicate greater depression.

^‡Cognitive status designated using the Modified Mini-Mental State Examination; scores of 78 or higher indicate cognitively intact and, scores between 50 and 77 cognitive impairment.

Among participants with no or mild pain, 37.1% reported IADL impairment, whereas 53.0% of participants with moderate pain or higher reported IADL impairment. At the same time, 39.7% of cognitively intact persons reported IADL impairment compared with 63.7% of cognitively impaired persons. Similarly, among participants with no or mild pain, 16.0% reported ADL impairment, whereas 29.6% of participants with moderate pain or higher reported ADL impairment. Finally, 17.2% of cognitively intact persons reported ADL impairment compared with 36.3% of cognitively impaired persons.

Figure 1 displays the proportion of participants with IADL and ADL impairment by pain and cognitive status. The percentage of participants reporting IADL and ADL impairment over the four pain and cognition groups increased in the following order: 1) no pain and cognitively intact, 2) pain and cognitively intact, 3) no pain and cognitively impaired, and 4) pain and cognitively impaired. Participants in groups 2 (pain and cognitively intact) and 3 (no pain and cognitively impaired) did not vary by IADL (p = .33) or ADL (p = .07) impairment and were subsequently combined into one group. Participants in groups 1 and 4 and the combined group differed in the proportion of participants with IADL p < .01 and ADL p < .01 impairment (Table 2).

Figure 1.

Functional impairment by pain and cognitive status*,^† *Noncancer pain self-report dichotomized as no or mild pain and pain at a moderate intensity or greater ^†Cognitive status designated using the Modified Mini-Mental State Examination; scores of 78 or higher indicate cognitively intact and, scores between 50 and 77 cognitive impairment.

Note: ADL = Activity of Daily Living; IADL = Instrumental Activity of Daily Living.

Table 2.

Relationship Between noncancer Pain and Cognitive Groups by IADL (Instrumental Activities of Daily Living) and ADL (Activities of Daily Living) impairment

	No Pain and Cognitively Intact	Pain or Cognitively Impaired*	Pain and Cognitively Impaired	p Value
IADL impairment (n)	2,814	2,005	268	<.01
Yes (%)	33.0	53.0	76.1
ADL impairment (n)	2,838	2,025	270	<.01
Yes (%)	12.3	27.0	47.4

Note: Of the four pain and cognition groups (no pain and cognitively intact, pain and cognitively intact, no pain and cognitively impaired, and pain and cognitively impaired), participants in pain and cognitively intact and no pain and cognitively impaired did not vary by IADL or ADL impairment and were subsequently combined into one group. ADL = Activity of Daily Living; IADL = Instrumental Activity of Daily Living.

The median IADL summary score was 1.05 (1.76 SD), range 0–10, and median ADL summary score was 0.5 (1.32 SD), range 0–14, both indicating high function. The median IADL and ADL summary scores increased among the pain and cognition categories in the following order: no pain and cognitively intact (0.63 SD 1.24, 0.23 SD 0.80), pain and cognitively intact (1.18 SD 1.69, 0.57 SD 1.27), no pain and cognitively impaired (1.64 SD 2.22, 0.75 SD 1.57), and pain and cognitively impaired (2.27 SD 2.47, 1.35 SD 2.09), respectively. Table 3 displays the linear regression analyses of summary IADL and ADL scores and pain (at a moderate intensity or greater) and cognitive status (intact or impaired) adjusted for demographics (age, gender, and education), depression, and comorbidity. IADL summary scores were associated with pain, coefficient .17 (95% confidence interval [CI] 0.07–0.26), p < .01; cognitive impairment, coefficient .67 (95% CI 0.51–0.83), p < .01; and an interaction effect, pain with cognitive impairment, coefficient .24 (95% CI 0.01–0.49), p = .05. ADL summary scores were associated with pain, coefficient .10 (95% CI 0.04–0.17), p < .01; cognitive impairment, coefficient .29 (95% CI 0.19–0.39), p < .01; the interaction term was not significant coefficient .12 (95% CI −0.03 to 0.29), p = .12.

Table 3.

Multivariate Linear Regression of Associations with IADL and ADL

	IADL Summary Score*		ADL Summary Score†
Explanatory variable	Coefficient (95% CI)	p Value	Coefficient (95% CI)	p Value
Age	.08 (0.07–0.09)	<.01	.03 (0.03–0.04)	<.01
Male	−.05 (−0.13 to 0.03)	.24	−.10 (−0.15 to −0.05)	<.01
Education	−.01 (−0.02 to 0.01)	.27	.01 (0.01–0.02)	<.01
Depression‡	.05 (0.03–0.06)	<.01	.01 (−0.01 to 0.01)	.06
Comorbidity§	.16 (0.14–0.18)	<.01	.08 (0.07–0.10)	<.01
Pain (moderate or greater)	.17 (0.07–0.26)	<.01	.10 (0.04–0.17)	<.01
Cognitive impairment‖	.67 (0.51–0.82)	<.01	.29 (0.19–0.39)	<.01
Pain × cognitive impairment	.24 (0.01–0.48)	.05	.13 (−0.03 to 0.28)	.12

Notes: ADL = Activity of Daily Living; CI = confidence interval; IADL = Instrumental Activity of Daily Living.

^*IADL summary score is self-reported function among five activities were each score ranges from 0 to 2 with higher scores indicating greater impairment giving a composite score of 0–10.

^†ADL summary score is self-reported function among seven activities were each score ranges from 0 to 2 with higher scores indicating greater impairment giving a composite score of 0–14.

^‡Depression score is the summed response to the mental health screening questionnaire, with higher scores indicating greater depression.

^§Comorbidity is the summary score of 15 possible chronic conditions reported by participants, with higher scores indicating greater morbidity.

^‖Cognitive status designated using the Modified Mini-Mental State Examination; scores of 78 or higher indicate cognitively intact and, scores between 50 and 77 cognitive impairment.

DISCUSSION

This is the first study to examine the relationship among noncancer pain, cognitive impairment, and functional disability. As hypothesized, pain and cognitive impairment were independently associated with functional disability, defined as IADL and ADL summary scores. A positive interaction effect between pain and cognition was associated with additional IADL loss. The presence of pain or cognitive impairment resulted in a similar proportion of participants reporting IADL or ADL impairment. IADL and ADL impairment increased across groups, with the no pain and cognitively intact and pain and cognitively impaired groups having the least and greatest functional impairment, respectively.

Other studies also suggest that medical conditions can interact with one another and increase the likelihood of a person having an adverse outcome in a nonlinear manner (28). One study reported that the presence of both arthritis and heart disease was associated with impaired mobility beyond their independent contributions. Compared with persons without arthritis or heart disease, the risk of disability was 4.3 times higher for persons with arthritis, 2.3 times higher for persons with heart disease, and 13.6 times higher for persons with both conditions (29). Moreover, coexistent health conditions can complicate the treatment of other conditions. For example, Thielke and colleagues reported that high pain levels among older adults were associated with worse response to depression treatment (30).

The current study is consistent with these and other previous studies and builds upon them by showing that pain and cognitive impairment may act synergistically in exacerbating functional loss. Although we do not have data on pain management in our sample, the literature consistently reports that pain is undertreated in older adults with and without cognitive impairment (31–33). Based upon our study and the published research, it is reasonable to suggest that improving pain management may shift the threshold of functional disability particularly for those with cognitive impairment. Even a small increase in treatment success may slow physical or cognitive decline and may keep older adults independent for longer periods of time and delay admission to nursing homes or other long-term care settings. On the other hand, studies are needed to ensure that more aggressive pain management is not associated with unanticipated functional loss. Finally, programs designed to improve function in persons with cognitive impairment that fail to detect and treat pain adequately may be much less likely to succeed.

Our study also suggests that optimizing health outcomes in frail elders may require comprehensive, interdisciplinary care simultaneously targeting more than one condition. The pilot study reported by Unutzer and colleagues designed to improve both pain and depression outcomes is consistent with this approach and their positive, albeit preliminary, results are promising (34). The interaction between pain and cognitive impairment as identified in our study will challenge geriatric and primary care providers to develop and test novel approaches to managing pain and minimizing disability in the growing numbers of older patients whose health care needs will be complicated by concurrent and progressive cognitive impairment.

The cross-sectional design does not allow for discernment of causality. Although the findings show a clear association among pain, cognitive status, and function, it is not clear from these findings which condition(s) can directly influence the other(s). In addition, the mechanisms by which pain and cognitive impairment interact to exacerbate functional impairment are unknown. Chronic pain in older adults is associated with decrements in neuropsychological performance and these decrements mediate the relationship between pain and disability (35). Furthermore, recent data indicate that chronic pain is associated with brain atrophy in older adults (36). It is possible, therefore, that the synergistic impact on function of pain and cognitive impairment occurs via alterations in brain morphology. Additional research is needed to elucidate the mechanism(s) of the interrelationship between pain, cognitive impairment, and physical function so that more effective treatments can be developed.

Strengths of the study include the large sample size that was drawn from a representative sample population. The study incorporated a validated pain measure rarely included in studies of aging and further underscores the importance of these findings. Also, the results provide a more complete picture of the complex relationship between noncancer pain and cognitive impairment with functional disability compared with what has been published to date.

Although our study has a number of strengths, its limitations should be noted. One weakness of the study is that the method of ascertaining cognitive impairment and function may underestimate the prevalence of these conditions and thus the study’s findings. Cognitive status was assessed using the 3MS that has acceptable sensitivity (87%) and specificity (89%) for detecting dementia in the community when the cutoff in this study of 77 or lower is used to indicate cognitive impairment (24,37). A proportion of individuals will be incorrectly designated as being cognitively impaired and would not be expected to experience as much functional disability as someone with an actual dementia. Also, the cutoff score used in this analysis infrequently categorizes participants as being cognitively intact when in fact they are impaired. Taken together, the study’s results may underestimate the true association of cognitive impairment with functional disability and potentially diminish the interaction effect with pain. Functional status was assessed using participant self-report of IADLs and ADLs that reflects clinical practice. However, cognitively impaired older adults typically overestimate functional ability; thus, as cognitive impairment worsens so does the discrepancy between reported function and actual performance abilities (26). As a result, self-report of functional abilities may differentially underestimate disability among participants with cognitive impairment, reducing the effect size of functional impairment in persons with cognitive impairment with and without pain. Despite these shortcomings, IADL loss displayed a multiplicative relationship with pain and cognitive impairment. Although the data were collected in 1996, we believe the study’s findings remain current and relevant as neither the assessment nor treatment of noncancer pain or cognitive impairment has significantly changed since the data were collected. The current study was a community-based population and our findings may not be generalizable to nursing home populations. Lastly, the nature of data collection required the exclusion of persons with advanced cognitive impairment so that the study findings may not be applicable to this population and warrants further study.

In conclusion, noncancer pain and cognitive impairment are independently associated with IADL and ADL impairment, and the association with IADL impairment is magnified when both conditions are present. This suggest that pain plays an underappreciated and significant role in the development of functional disability among older adults with cognitive impairment. Future efforts should investigate the impact of pain management programs in older adults with cognitive impairment including changes in pain intensity as well as variation in functional abilities.

FUNDING

All the data reported here were gathered using public funding from the National Health Research Development Program, which administrated a grant from the Seniors’ Independence Research program (6606-3954-MC(S)). This work was supported by a career development award from the National Palliative Care Research Center and the National Institute On Aging at the National Institutes of Health (grant K23AG029815) and from the Canadian Institutes of Health Research through an operating grant (grant MOP-62823). Additional support came from the Dalhousie Medical Research Foundation (career support to K.R. as Kathryn Allen Weldon Professor of Alzheimer Research) and the Fountain Innovation Fund of the QEII Research Foundation.