Abstract

Objective. Explore the relationships between pain, depression, and functional disability in elderly persons.

Design. A cross-sectional, observational study of 228 independently living retirement community residents.

Methods. Self-report measures of pain (adaptation of McGill Pain Questionnaire), depression (Geriatric Depression Scale [GDS]) and physical functioning (Physical performance difficulties, activities of daily living [ADL], independent activities of daily living [IADL], and 3-meter walking speed) were employed.

Outcome Measures. Physical functioning variables were dichotomized. Individuals in the lowest quartiles of functional performance and of walking speed were contrasted to all others; for ADL and IADL, those needing some help were compared with those independent in activities.

Results. Pain and depression levels were strongly related to physical performance; depression levels were related to ADL and walking speed. In multivariate analyses, an interaction effect was observed where the effects of pain were a function of level of depression. Individuals reporting activity-limiting pain and slightly elevated depressive symptom levels, sub-threshold depression, or major depression were significantly more likely (AOR 7.8; 95% CI, 3.07–20.03) than non-depressed persons to be in the lowest quartile of self-reported physical performance.

Conclusions. While both pain and depression level affect physical performance, depressive symptoms rather than pain appear the more influential factor. When seeing elderly patients, identifying, evaluating, and treating both pain complaints and depressive symptoms and disorders may reduce functional impairment.

Chronic pain is a major obstacle to a satisfying old age [1]. With the number of persons 65 years of age and older in the United States currently about 35 million and projected to reach more than 64 million by 2030, chronic pain in the elderly is emerging as a significant public health problem of increasing magnitude. In their lifetime, an estimated 80% to 85% of persons age 65 and older will experience a significant health problem that will predispose them to pain [2]. In a recent study of community-dwelling senior citizens living in Western Ontario, Canada, Scudds and Robertson observed self-reported musculoskeletal pain during the past 2 weeks in 72.7% of their sample [3]. Fifteen percent indicated their pain was extremely severe. Higher prevalence rates of severe pain have been reported in the institutionalized elderly. The limited data available suggest that 33% to 61% report severe pain [4,5], 26 to 34% describe pain as occurring constantly or daily [6,7] and up to 47% report pain in multiple sites [5]. It is also estimated that 36% to 83% of the elderly population report some degree of pain that may interfere with life's activities and quality of life [8]. Beckar et al. [9] showed that, as compared with the pain free population, pain patients had significantly reduced physical, psychological and social functioning, and well being. Skevington et al. [10] reported a statistically significant association between pain and perception of general quality of life.

Clinical depression also negatively affects physical and social functioning and overall reported quality of life. It has been shown that depression is a major, independent contributor to disability [11] and that the presence of even minor depression is associated with significant functional disability [12]. There is also increasing evidence demonstrating a significant independent association between subthreshold depression and physical/social disability [13,14,15]. Depression also influences many aspects of the experience of pain including the perception of and reaction to pain. In their review of studies focused on the relationship between chronic pain and depression, Dworkin and Gitlin [16] reported between 1.5% and 57% of patients with chronic pain had major depression. It has also been documented that a majority of chronic pain patients, while not meeting criteria for major depression, have significant symptoms of depression [17].

When the relationships between chronic pain, depression, and functional performance are simultaneously examined, the literature is less clear. Won et al. [6] suggest that the associations between pain, depression, and functioning are interdependent and complex. Williamson et al. [18] reported that functional disability mediated the relationship between pain and depressed affect, while Hoizberg et al. [19] reported that depression was related to functional ability after controlling for the effects of pain level. In contrast, Haythornwiathe [20] reported that, in chronic pain patients, there were no group differences on measures of disability between depressed and nondepressed patients. However, the sample in the Haythornwaithe study was derived from a tertiary pain center population that was heterogeneous with respect to diagnosis and age. Thus, the study could not derive clinically coherent subgroups (e.g., with respect to diagnosis, age, and sex) of a large enough size to provide sufficient power to adequately investigate this question.

The purpose of this paper is to further explore the relationships between pain, depression, and functional disability in a sample of elderly persons. Particular emphasis is given to evaluating the presence of interaction effects between pain and depression and functional impairment. Given the relatively high prevalences of chronic pain, depressive symptoms and syndromes, and functional compromise in the elderly, a better understanding of the impact of depressive disorders on chronic pain and the relationship of these two factors to functioning is essential. Improved understanding would not only provide important insights regarding how to increase the quality of life for the elderly, but also clinical care practitioners would have insight regarding whether to emphasize treatment for the depressive disorder, the chronic pain, or both conditions simultaneously.

Methods

Sample Selection

Data for this analysis were derived from a 24-month longitudinal study funded by the National Institutes of Aging to investigate the personal and economic costs of depression in independently living elders. The study sample was drawn from the residents in the independent living sections of 25 Continuing Care Retirement Communities (CCRC) located in the greater Philadelphia area. Recruitment letters were sent to approximately 5,000 individuals over 60 years of age. One thousand, three hundred and forty-seven (1347) residents (27%) agreed to participate in an in-person screening interview.

Measures

Sociodemographic (e.g., age, gender, education, ethnic background) and physical health characteristics (e.g., self-rated health, self-reported medical conditions) were assessed. In addition, the 6-item self care Activities of Daily Living Scale (ADL), ranging from ‘0’, “inability to complete any activity without assistance”, to ‘6’, “ability to complete all activities without help”[21], the 30-item Geriatric Depression Scale (GDS) ranging from ‘0’, “no depressive symptoms”, to ‘30’, “depressive responses to all questions”[22], and the short form of the Blessed Dementia Scale (BDS) with a range of ‘0’, “all questions answered correctly”, to ‘15’, “no correct responses”[23], were administered to assess functional ability, depressive symptoms, and cognitive status respectively. To be eligible, fluency in English and a BDS less than 8, indicating the cognitive ability to complete the study, were required. All 220 eligible individuals with elevated depressive symptom levels (GDS ≥ 9) were invited to participate further in cross-sectional and longitudinal assessments of physical functioning, pain, and psychiatric status. One hundred seventeen of these (53%) agreed to participate. For each subject with a GDS greater than 8, a nondepressed individual (GDS < 9) of the same gender and within the same age stratum (60–74, 75–84,> 84) was recruited.

Psychiatric Assessment and Classification

After giving signed, informed consent to participate in additional baseline interviews and comparable 6-, 12-, 18-, and 24-month follow-up interviews, all study subjects completed the mood, anxiety, somatization, adjustment disorder, and psychotic screen modules of the Nonpatient Version of the Structured Clinical Interview for DSM-III (SCID) administered by a SCID-trained psychiatric clinical nurse specialist. The SCID has been shown to be a reliable tool for diagnosing these DSM-III R psychiatric conditions [24]. Using a diagnostic algorithm reported elsewhere [15], the SCID clinical data were combined with the screening and baseline Geriatric Depression Scale (GDS) scores to reclassify all study subjects, irrespective of their initial GDS score, into three depression groups: a) No Depression group (few symptoms [screening GDS score <6; n = 86]); b) Mild Depression group (some symptoms [screening GDS score 6–10; n = 53]); and c) High Depression group (sub-threshold or major depression symptoms on SCID and/or GDS scores at screening and baseline interviews> 10 [n = 89]).

The sample for this analysis consists of 228 individuals and is limited to data collected during the cross-sectional baseline assessments. Four individuals were dropped from the study after giving signed consent due to missing data on critical variables or to the presence of a SCID diagnosis other than major depression or dysthymia.

Pain and Physical Function Assessment

Following completion of the psychiatric assessment, a baseline interview that included both self-report and performance measures was completed. Pain was measured using 7 questions adapted from the McGill Pain Questionnaire (25). Two questions: (1) “How much have you been bothered by pain in the last two weeks?” (“Not at all” to “Extremely”) and (2) “How much has the pain interfered with your day-to-day activities?” (“Not at all” to “Extremely”), were used to classify individuals into three pain groups: 1) No Pain (n = 49); 2) Pain Without Activity Limitations (n = 93), and 3) Pain With Activity Limitations (n = 86). Baseline health status was determined by a question on global self-rated health and an adaptation of the medical condition checklist developed by Lawton, et al [21]. In addition to the screening activity of daily living, Instrumental Activities of Daily Living (IADL) (e.g., shopping, doing laundry, cooking) were assessed during the baseline interviews with an 8-item scale based on the work of Lawton, et al [21]. A self-report physical performance scale indicating the difficulty in performing 9 specific activities such as kneeling, stooping, pushing, grasping, was adapted from Nagi's work [26]; and a timed performance measure of an individual's normal speed to walk 3 meters was taken from the Structured Assessment of Independent Living Skills (SAILS), a scale of upper and lower limb function that has been validated in older samples [27].

Pain, depression variables, and combined pain/depression variables used in this study are listed in Table 1.

Table 1

Summary definitions of study main independent variables (Depression Level, Pain Level, Combined Depression and Pain)

Study Variables Measure Clinical Group 
Level of Depression for univariate and multivariate analyses   
     No Depression *GDS < 5 No evidence of depressive symptoms 
     Mild Depression GDS 6–10 Patient has some depressive symptoms 
     High Depression Repeated GDS> 6 or **SCID diagnosis Patient has clinical depression (eg, sub-threshold, minor, or major depression) 
Level of pain for univariate and multivariate analyses   
     No Pain Indicated: “Not at all” Bothered by pain in past 2 weeks  
     Pain without Activity Limitations Indicated: “A little”–“Extremely” Bothered by pain. In past 2 weeks & “Not at all” Limited in usual activities  
     Pain with Activity Limitations Indicated: “A little”—“Extremely” Bothered by pain in past 2 weeks & “A little”—“Extremely” Limited in usual activities  
Combined depression & pain variable for multivariate analyses   
     No Depression/All Levels Of Pain (No Pain, Pain Without Activity Limitations, Pain With Activity Limitations) GDS < 5 & All levels of pain No evidence of depressive symptoms but any level of pain may be present. Reference group for multivariate analyses 
     Mild Depression, High Depression/No Pain, Pain Without Activity Limitations GDS> 6; Subthreshold or major depression may be present; No pain or pain without activity limitations Mild or high depressive symptom levels are present. Pain may or may not be present but no activity limitations reported 
     Mild Depression, High Depression/ Pain With Activity Limitations GDS> 6; Subthreshold or major depression may be present Mild or high depressive symptoms are present. All individuals report pain with activity limitations 
Study Variables Measure Clinical Group 
Level of Depression for univariate and multivariate analyses   
     No Depression *GDS < 5 No evidence of depressive symptoms 
     Mild Depression GDS 6–10 Patient has some depressive symptoms 
     High Depression Repeated GDS> 6 or **SCID diagnosis Patient has clinical depression (eg, sub-threshold, minor, or major depression) 
Level of pain for univariate and multivariate analyses   
     No Pain Indicated: “Not at all” Bothered by pain in past 2 weeks  
     Pain without Activity Limitations Indicated: “A little”–“Extremely” Bothered by pain. In past 2 weeks & “Not at all” Limited in usual activities  
     Pain with Activity Limitations Indicated: “A little”—“Extremely” Bothered by pain in past 2 weeks & “A little”—“Extremely” Limited in usual activities  
Combined depression & pain variable for multivariate analyses   
     No Depression/All Levels Of Pain (No Pain, Pain Without Activity Limitations, Pain With Activity Limitations) GDS < 5 & All levels of pain No evidence of depressive symptoms but any level of pain may be present. Reference group for multivariate analyses 
     Mild Depression, High Depression/No Pain, Pain Without Activity Limitations GDS> 6; Subthreshold or major depression may be present; No pain or pain without activity limitations Mild or high depressive symptom levels are present. Pain may or may not be present but no activity limitations reported 
     Mild Depression, High Depression/ Pain With Activity Limitations GDS> 6; Subthreshold or major depression may be present Mild or high depressive symptoms are present. All individuals report pain with activity limitations 
*

30 item Geriatric Depression Scale. **Structured Clinical Interview for DSM-III.

Analyses

To evaluate the relationships between the depression symptom and pain level variables, all other independent variables (such as age, gender, and self-rated health), and the 4 measures of physical function, self-reported physical function, and timed walking speed were recoded into quartiles. The ADL and IADL variable categories were collapsed into 2 and 3 levels respectively. Univariate relationships among categorical variables were evaluated using chi square tests. Multivariate analyses using logistic regression were conducted to evaluate the relative importance of depression symptom level and pain level, while controlling for the other independent variables such as age, gender, and number of self-reported medical conditions. For these analyses, the physical function variables were dichotomized so that the most impaired category was contrasted with all other categories combined. The modest sample size precluded the use of logistic regression to meaningfully evaluate interactions between pain and depression while simultaneously controlling for other important independent variables. Because an interaction between pain and depression was suspected, alternative analytic approaches were undertaken. Specifically, four 3-way contingency tables were first constructed in which the depression level was the row variable, and pain level was the column variable. Within each cell, jointly defined by depression level and pain, the proportion of individuals in the most impaired category of each physical function variable level was displayed and the relevant chi square statistics were examined.

Based on the above analyses, a combined depression and pain variable was created with 3 categories defined as: 1) No Depression/All Pain Levels ; 2) Mild Depression or High Depression/No Pain or Pain Without Activity Limitations ; and 3) Mild Depression or High Depression/Pain With Activity Limitations. This variable was included in a multivariate logistic model to further elucidate the presence of an interaction between pain and depression, after controlling for age, gender, and number of reported medical conditions.

Results

Univariate Analyses

Sociodemographic and health-related characteristics of the study sample are displayed in Table 2. Consistent with the distributions of age, gender, and educational level in among CCRC residents, the sample was predominantly aged 75 to 84, female, and college or post-college educated. The distribution on self-rated health and reported medical conditions suggests the sample was in good general health. Data presented in Table 3 indicate generally good physical functioning for the majority of study individuals. For example, all but 19 were completely independent in self-care activities. On the self-report physical performance scale (range 0–27), the sample mean was 6.9 (SD 5.0).

Table 2

Selected sociodemographic and health characteristics of the CCRC residents studied

Characteristic 
Age:   
  60–74  30 13.2 
  75–84 141 61.8 
  85+  57  6.6 
Gender:   
  Male  44 19.3 
  Female 184 80.7 
Yrs school:   
  <12 Yrs   7  3.1 
  12 Yrs  31 13.5 
  13–16 Yrs 112 48.7 
  >16 Yrs  80 34.7 
Self-rated health   
  Excellent  28 12.3 
  Good 128 54.8 
  Fair  61 26.8 
  Poor/Bad  14  6.1 
# Conditions reported   
 0–2  64 28.1 
 3–4  73 32.0 
  5+  91 39.9 
Characteristic 
Age:   
  60–74  30 13.2 
  75–84 141 61.8 
  85+  57  6.6 
Gender:   
  Male  44 19.3 
  Female 184 80.7 
Yrs school:   
  <12 Yrs   7  3.1 
  12 Yrs  31 13.5 
  13–16 Yrs 112 48.7 
  >16 Yrs  80 34.7 
Self-rated health   
  Excellent  28 12.3 
  Good 128 54.8 
  Fair  61 26.8 
  Poor/Bad  14  6.1 
# Conditions reported   
 0–2  64 28.1 
 3–4  73 32.0 
  5+  91 39.9 
Table 3

Distribution of the CCRC residents studied according to self report measures of physical function: self-reported physical performance (FUNCTION), self care activities of daily living (ADL) and instrumental activities of daily living (IADL) and performance-based normal speed to walk three meters (WALK)

Function Measure    
Function: Mean 6. (SD 5.)   
  Range 0–27 Quartile Range  
  Best quartile: 0–2 50  22 
  Upper middle quartile: 3–6 66  29 
  Lower middle quartile: 7–10 53  23 
Lowest quartile: 11–23 59  26 
Walk (Normal speed 3 meter walk; Seconds/meter) Seconds/meter Mean 5.5 (SD 2.7)    
  Range 2.3–18.8 sec/meter Quartile range  
  Fastest quartile: 2.3–3.6 54  24.6 
  Upper middle quartile: 3.7–4.8 55  25.9 
  Lower middle quartile: 4.9–6.4 56  25.0 
  Slowest quartile: 6.5–18.8 57  24.6 
IADL (Instrumental ADL)   
  Independent all 8 activities: 91  39.9  
  Needs help 1–2 activities: 90  39.5  
  Needs help> 2 activities: 47  20.6  
ADL (Self Care ADL)   
  Independent all 6 activities 209  92  
  Needs help> 1 activities 19   
Function Measure    
Function: Mean 6. (SD 5.)   
  Range 0–27 Quartile Range  
  Best quartile: 0–2 50  22 
  Upper middle quartile: 3–6 66  29 
  Lower middle quartile: 7–10 53  23 
Lowest quartile: 11–23 59  26 
Walk (Normal speed 3 meter walk; Seconds/meter) Seconds/meter Mean 5.5 (SD 2.7)    
  Range 2.3–18.8 sec/meter Quartile range  
  Fastest quartile: 2.3–3.6 54  24.6 
  Upper middle quartile: 3.7–4.8 55  25.9 
  Lower middle quartile: 4.9–6.4 56  25.0 
  Slowest quartile: 6.5–18.8 57  24.6 
IADL (Instrumental ADL)   
  Independent all 8 activities: 91  39.9  
  Needs help 1–2 activities: 90  39.5  
  Needs help> 2 activities: 47  20.6  
ADL (Self Care ADL)   
  Independent all 6 activities 209  92  
  Needs help> 1 activities 19   

The distributions on level of depression and pain are shown in Table 4. Over-sampling individuals with an elevated screening GDS resulted in the high proportion (39%) of study members who met criteria for subthreshold or major depression or dysthymia. As per frequency, severity, and location of the pain, 79% were bothered by pain within the past 2 weeks, and about 50% of those reporting pain indicated that their pain resulted in limitations in their usual activities. Over 40% of the entire sample indicated they experienced pain in 4 or more locations. The most frequently reported pain sites were: 1) joints (64%), 2) arms and legs (49%), and 3) back (44%). The combined distribution of the pain and depression variables (shown in Table 5) reveals the presence of a strong, statistically significant association between the 2 measures. Almost 23% reported both Pain With Activity Limitations and High Depression while about 13% reported being free of both pain and depressive symptoms.

Table 4

Distribution of depressive symptom levels (DEPRESSION) and pain levels (PAIN)

Measure 
Depressive symptom group   
  No Depression group 86 37.7 
  Mild Depression group 53 23.2 
  High Depression group 89 39.0 
Self reported pain level   
  No pain 49 21.5 
  Pain Without Activity Limitations 93 40.8 
  Pain With Activity Limitations 86 37.7 
Measure 
Depressive symptom group   
  No Depression group 86 37.7 
  Mild Depression group 53 23.2 
  High Depression group 89 39.0 
Self reported pain level   
  No pain 49 21.5 
  Pain Without Activity Limitations 93 40.8 
  Pain With Activity Limitations 86 37.7 
Table 5

Combined distribution of depressive symptom level and pain level for elderly CCRC residents studied

 Level Pain
 
 
Depression Level No Pain Pain Without Activity Limitations Pain With Activity Limitations Total 
No Depression n = 28 (12.8%) n = 41 (18%) n = 17 (7.5%) n = 86 (37.3%) 
  Mild Depression n = 10 (4.4%) n = 26 (11.4%) n = 17 (7.5%) n = 53 (23.2%) 
  High Depression n = 11 (4.8%) n = 26 (11.4%) n = 52* (22.8%) n = 89 (39.0%) 
Total n = 29 (21.5%) n = 93 (40.8%) n = 86 (37.7%) n = 228 (100%) 
 Level Pain
 
 
Depression Level No Pain Pain Without Activity Limitations Pain With Activity Limitations Total 
No Depression n = 28 (12.8%) n = 41 (18%) n = 17 (7.5%) n = 86 (37.3%) 
  Mild Depression n = 10 (4.4%) n = 26 (11.4%) n = 17 (7.5%) n = 53 (23.2%) 
  High Depression n = 11 (4.8%) n = 26 (11.4%) n = 52* (22.8%) n = 89 (39.0%) 
Total n = 29 (21.5%) n = 93 (40.8%) n = 86 (37.7%) n = 228 (100%) 

X24 = 31.249 P value = <.0000

*

This table indicates the significant association (P < .0000) between pain and depression.

Univariate analyses of the relationships between the independent variables and the 4 physical function measures are summarized in Table 6. The associations between depression level and 3 of the 4 physical function variables are statistically significant and, as expected, impairment in physical function increased with level of depression. Conversely, pain level is only associated with self-reported physical performance level. The observed associations between age and walking speed and IADL were expected, but there were no comparable statistically significant associations between age and self-reported physical function or ADL. The associations between the physical function measures and self-rated health and number of medical conditions are in the expected direction and consistent with reports from other studies.

Table 6

Summary of the contingency table analyses to evaluate the univariate associations between Depression Level (DEPRESSION), Pain Level (PAIN) and other independent variables: age (AGE), sex (SEX), self-rated health (SRH) and number of self-reported medical conditions (CONDITIONS) and the study Physical Function Dependent Variable Measures: FUNCTION, WALK, IADL, ADL

 Physical Function Measures
 
Independent Variables FUNCTION WALK IADL ADL 
Depression More severe depression, Worse functioning (P <.0000)** More severe depression, Slower walking (P < .05) No association More severe depression, ADL help needed (P < .00) 
Pain More pain, Worse functioning (P < .0000) No association No association No association 
Age No association Increasing age, Slower walking (P < .000) Increasing age, IADL help needed (P < .000) No association 
Gender No association No association No association No association 
SRH Poorer SRH, Worse functioning (P < .0000) Poorer SRH, Slower walking (P < .00000) Poorer SRH, IADL help needed (P < .000) Poorer SRH, ADL help needed (P < .000) 
# Medical conditions More conditions, Worse functioning (P < .0000) More conditions, Slower walking (P < .0000) More conditions, IADL help needed (P < .000) No association 
 Physical Function Measures
 
Independent Variables FUNCTION WALK IADL ADL 
Depression More severe depression, Worse functioning (P <.0000)** More severe depression, Slower walking (P < .05) No association More severe depression, ADL help needed (P < .00) 
Pain More pain, Worse functioning (P < .0000) No association No association No association 
Age No association Increasing age, Slower walking (P < .000) Increasing age, IADL help needed (P < .000) No association 
Gender No association No association No association No association 
SRH Poorer SRH, Worse functioning (P < .0000) Poorer SRH, Slower walking (P < .00000) Poorer SRH, IADL help needed (P < .000) Poorer SRH, ADL help needed (P < .000) 
# Medical conditions More conditions, Worse functioning (P < .0000) More conditions, Slower walking (P < .0000) More conditions, IADL help needed (P < .000) No association 
**

All P values refer to the Pearson ChiSquare statistic computed during the contingency table analyses.

The associations of depression and pain level with age, gender, self-rated health, and number of medical conditions were evaluated. In analyses not shown here, neither pain level nor level of depression were related to age or gender. In contrast, there were strong, statistically significant associations between both pain and depression and self-rated health and number of medical conditions. As expected, poorer health measures were associated with higher levels of depression and pain (P<.05).

Multivariate Analyses

Multivariate analyses to evaluate if pain or depression had an independent association with each physical function variable were conducted using the logistical model. For these analyses, age, gender, self-rated health, and number of medical conditions were entered first into the logistic equation. After eliminating nonsignificant variables from the logistic equation, depression level and pain were entered alone and then together. The results of these analyses are shown in Table 7. Adjusted odds ratios are shown here with the upper and lower limits of the respective 95% confidence intervals. In computing the adjusted odds ratios, the “least impaired level” of each independent variable has been designated the “reference level”. For each independent variable, the displayed adjusted odds ratios indicate the increase in the odds of being in the most impaired physical function category for individuals in the designated category as compared to the reference level, while controlling for the effects of all other independent variables in the logistic equation. Table 7 shows that, after controlling for age, gender, and health status variables, the associations between pain and depression and reported physical performance remain statistically significant. For individuals who reported Pain With Activity Limitations, the odds of being in the most impaired quartile of physical performance was 4.15 times greater than those who reported No Pain. Likewise the odds of being in the lowest self-reported physical performance quartile were 2.82 greater among those in the High Depression group than for those in the No Depression group. For individuals in the High Depression group, there was almost an 8-fold increase in the odds of being in the ADL group who had difficulties with one or more self-care activities. Notably, level of depression was the only variable associated with ADL functioning. These findings are consistent with the univariate analyses. Controlling for age and number of reported medical conditions, however, the apparent univariate association between depression level and walking speed is eliminated. Age and number of self-reported medical conditions appear to have to the greatest impact on the physical function variables except ADL.

Table 7

Adjusted Odds Ratios (AOR) and 95% Confidence Limits (Lower [L], Upper [U]) for the independent variables with a statistically significant association with each of the four physical function variables: (FUNCTION), timed walking performance (WALK), Self Care Activities of Daily Living (ADL), and Instrumental Activities of Daily Living (IADL)

 function
 
WALK
 
ADL
 
IADL
 
 95% CI
 
95% CI
 
95% CI
 
95% CI
 
 AOR AOR AOR AOR 
Depression Group             
  No Depression — — — N.S. — — — N.S. 
  Mild Depression 2.59 (.92 7.34)    2.51 (.38 16.51)    
  High Depression 2.82 (1.02 7.35)    7.84 (1.71 35.16)    
Pain             
  No Pain — — — N.S. N.S. N.S. 
  Pain Without Activity Limits 1.11 (.33 3.7)          
  Pain With Activity Limits 4.15 (1.31 13.1)          
Conditions             
0–2 conditions — — — — — — N.S. — — — 
3–4 conditions 2.59 (.81 8.29) 2.26 (.83 5.98)    1.29 (.62 2.69) 
5+ conditions 5.26 (1.72 16.12) 3.73 (1.60 10.48)    2.33 (1.10 4.91) 
SRH             
Excellent N.S. N.S. N.S. — — — 
Good          2.97 (1.16 7.62) 
Fair, poor, bad          3.50 (1.23 9.86) 
Age             
60–74 — — — — — — N.S. — — — 
75–84 1.75 (.54 5.62) 2.38 (.75 9.97)    2.40 (1.01 5.69) 
85+ 4.23 (1.19 15.02) 7.87 (2.22 33.08)    6.35 (2.25 17.86) 
Gender             
Male — — — N.S. N.S. N.S. 
Female .18 (.05 .65)          
 function
 
WALK
 
ADL
 
IADL
 
 95% CI
 
95% CI
 
95% CI
 
95% CI
 
 AOR AOR AOR AOR 
Depression Group             
  No Depression — — — N.S. — — — N.S. 
  Mild Depression 2.59 (.92 7.34)    2.51 (.38 16.51)    
  High Depression 2.82 (1.02 7.35)    7.84 (1.71 35.16)    
Pain             
  No Pain — — — N.S. N.S. N.S. 
  Pain Without Activity Limits 1.11 (.33 3.7)          
  Pain With Activity Limits 4.15 (1.31 13.1)          
Conditions             
0–2 conditions — — — — — — N.S. — — — 
3–4 conditions 2.59 (.81 8.29) 2.26 (.83 5.98)    1.29 (.62 2.69) 
5+ conditions 5.26 (1.72 16.12) 3.73 (1.60 10.48)    2.33 (1.10 4.91) 
SRH             
Excellent N.S. N.S. N.S. — — — 
Good          2.97 (1.16 7.62) 
Fair, poor, bad          3.50 (1.23 9.86) 
Age             
60–74 — — — — — — N.S. — — — 
75–84 1.75 (.54 5.62) 2.38 (.75 9.97)    2.40 (1.01 5.69) 
85+ 4.23 (1.19 15.02) 7.87 (2.22 33.08)    6.35 (2.25 17.86) 
Gender             
Male — — — N.S. N.S. N.S. 
Female .18 (.05 .65)          

The Adjusted Odds Ratios (AOR) shown here indicate the increase in the odds of being in the “most impaired” physical function category for individuals in the designated category as compared to the reference category, “least impaired physical function”, after controlling for the effects of all other independent variables in the model.

The interaction of pain and depression:

As noted previously, an interaction between pain and depression was suspected. This possibility was first evaluated by examining the proportion of individuals in the most impaired physical functioning category for each group jointly defined by pain and depression. Table 8 illustrates the results of these analyses for self-report of physical performance. There is no consistent increase in the proportion of individuals in the most impaired physical performance quartile associated with increases in self-reported pain. At all pain levels, however, increase in depression level is associated with a higher probability of being in the lowest performance quartile. This effect is evident for those in the Mild Depression group as well as for those in the High Depression group. Strikingly, over 50% of the individuals in the Mild Depression group or the High Depression group and who report Pain With Activity Limitations are in the lowest self-reported physical performance quartile. These findings are consistent with an interaction between pain and depression levels in which the effect of pain is a function of one's level of depression, and in the presence of pain and depressive symptoms, the effects of both pain and depression are substantially greater than with either alone.

Table 8

Number and percentage of individuals in the most impaired quartile of self-reported physical performance group for individuals defined by both depressive symptoms and pain level

LEVEL PAIN
 
   
Depression Group No Pain Pain Without Activity Limitations Pain With Activity Limitations Total 
No Depression n=1 3.6% n=5 12.2% n=2 11.8% n=8 13.6% 
  Mild Depression n=2 20.0% n=4 15.4% n=9 52.9% n=15 25.4% 
  High Depression n=2 18.2% n=6 23.1% n=28 53.9% n=36 61.0% 
Total n=5 8.5% n=15 25.4% n=39 66.1% n=59 100% 
LEVEL PAIN
 
   
Depression Group No Pain Pain Without Activity Limitations Pain With Activity Limitations Total 
No Depression n=1 3.6% n=5 12.2% n=2 11.8% n=8 13.6% 
  Mild Depression n=2 20.0% n=4 15.4% n=9 52.9% n=15 25.4% 
  High Depression n=2 18.2% n=6 23.1% n=28 53.9% n=36 61.0% 
Total n=5 8.5% n=15 25.4% n=39 66.1% n=59 100% 

Comparable tables not shown here for the 3 remaining physical function variables are consistent with, but less convincing, than that for self-reported physical performance. It is difficult to confidently draw conclusions from these findings, however, because they do not take into account the previously observed important and potentially confounding effects of age, gender, and number of self-reported medical conditions. To address this issue, multivariate logistic regression analyses were conducted in which the most impaired level of function was modeled. Based on review of the contingency tables showing the proportion of individuals in the lowest quartile of each function variable (e.g., as seen in Table 8), the combined pain/depressive symptom variable was collapsed from 9 categories to 3 and entered as the independent variable of main interest in the multivariate analyses summarized in Table 9. These findings confirm that, even after controlling for health-related and sociodemographic factors that affect physical functioning, those in the Mild Depression or High Depression groups with Pain With Activity Limitations (30.3% of the sample studied), have over a 7-fold increased odds of being in the lowest quartile of physical performance. Moreover, such individuals were about 8 times more likely to have an ADL problem than those who were in the no depression group and reported they were free of pain. For ADL, the increase in the odds of being dependent in an activity is also greater for those who were in the Mild Depression group as well as the High Depression group. These findings confirm that both pain and depression level do affect some aspects of physical function with depressive symptoms having a greater impact than pain.

Table 9

Adjusted Odds Ratios (AOR) and 95% Confidence Limits (Lower (L), Upper (U)) for the Combined DEPRESSION/PAIN variable and other independent variables with a statistically significant association with one of the study physical function variables

 function
 
WALK
 
ADL
 
IADL
 
 95% CI
 
95% CI
 
95% CI
 
95% CI
 
Independent Variables AOR AOR AOR AOR 
DEPRESSION/PAIN combined             
  No Depression/all pain groups — — —    — — —    
Mild or High Depression/No Pain, Pain Without Activity Limitation 1.52 (.56 4.13) N.S.   3.76 (.73 19.14) N.S.   
Mild Or High Depression/Pain With Activity Limitation 7.80 (3.07 20.03)    7.96 (1.69 37.06)    
CONDITIONS             
0–2 conditions — — — — — —    — — — 
3–4 conditions 2.48 (.78 7.87) 2.17 (.81 5.88) N.S.   1.37 (.61 2.71) 
5+ conditions 5.49 (1.84 16.46) 3.71 (1.39 9.91)    2.60 (1.21 4.98) 
SRH             
Excellent          — — — 
Good N.S.   N.S.   N.S.   3.34 (1.83 7.91) 
Fair, poor, bad          4.10 (1.54 9.41) 
AGE             
60–74 — — — — — —    — — — 
75–84 1.75 (.54 8.14) 2.45 (.67 8.94) N.S.   2.64 (1.03 5.74) 
85+ 4.24 (1.19 15.14) 8.04 (2.08 31.08)    6.94 (2.37 18.41 
GENDER             
Male — — — N.S.   N.S.   N.S.   
Female .18 (.05 .66)          
 function
 
WALK
 
ADL
 
IADL
 
 95% CI
 
95% CI
 
95% CI
 
95% CI
 
Independent Variables AOR AOR AOR AOR 
DEPRESSION/PAIN combined             
  No Depression/all pain groups — — —    — — —    
Mild or High Depression/No Pain, Pain Without Activity Limitation 1.52 (.56 4.13) N.S.   3.76 (.73 19.14) N.S.   
Mild Or High Depression/Pain With Activity Limitation 7.80 (3.07 20.03)    7.96 (1.69 37.06)    
CONDITIONS             
0–2 conditions — — — — — —    — — — 
3–4 conditions 2.48 (.78 7.87) 2.17 (.81 5.88) N.S.   1.37 (.61 2.71) 
5+ conditions 5.49 (1.84 16.46) 3.71 (1.39 9.91)    2.60 (1.21 4.98) 
SRH             
Excellent          — — — 
Good N.S.   N.S.   N.S.   3.34 (1.83 7.91) 
Fair, poor, bad          4.10 (1.54 9.41) 
AGE             
60–74 — — — — — —    — — — 
75–84 1.75 (.54 8.14) 2.45 (.67 8.94) N.S.   2.64 (1.03 5.74) 
85+ 4.24 (1.19 15.14) 8.04 (2.08 31.08)    6.94 (2.37 18.41 
GENDER             
Male — — — N.S.   N.S.   N.S.   
Female .18 (.05 .66)          

The Adjusted Odds Ratios (AOR) shown here indicate the increase in the odds of being in the “most impaired” physical function category for individuals in the designated category as compared to the reference category, “least impaired physical funtion”, after controlling for the effects of all other independent variables in the model.

Discussion

The study data document the high prevalences of both pain and elevated depressive symptoms meeting criteria for subthreshold or more severe depression in this sample of cognitively intact, independently living elderly who reside in a retirement community. This is consistent with the findings of other studies of community-dwelling seniors [2]. Associations between pain and level of depression and measures of functional competence have also been reported in the literature [2,6]. This study has demonstrated independent associations between pain and depression and increasing impairment, but only for self-reported physical performance and activities of daily living (ADLs). When individuals were cross-classified by depression and pain, the effects of level of depression on self-reported physical performance were greater at each level of pain. The importance of pain, however, was limited to individuals who reported Pain With Activity Limitations and was only observed for those individuals in the Mild Depression or High Depression groups. For such individuals, the combined effects of pain and depression level were substantially greater than for either variable when considered alone. These findings are generally in agreement with other findings reported in an institutionalized sample of elderly [4]. The presence of an interaction between pain and depression is especially significant because of the size of the effect. Over 50% of individuals within the Mild and High Depression groups who reported Pain With Activity Limitations were in the most impaired quartile of self-reported physical performance. This percentage is over twice that for any other pain/depression group and is particularly important because 30% of the study sample was in the 2 high-risk pain/depression groups.

The associations of pain, depression, and activities:

The presence of a statistically significant association between depression level and Activities of Daily Living (ADL) is striking both because of the strength of the association (individuals in the Mild Depression or High Depression groups were 8 times more likely than those in the No Depression group to report some problem with ADL activities) and because other factors such as age, gender, number of medical conditions were unrelated to ADL level. Because of the cross-sectional design of this study, it is not possible to determine if the decrements in ADL competence contributed to depression onset or if individuals with high levels of depressive symptoms were more likely to report needing some help in ADL activities because of depression-associated psychomotor retardation, fatigue, or lack of motivation. Further longitudinal studies of this cohort will be conducted to elucidate the direction of this relationship.

The absence of statistically significant associations between pain and depression levels and Instrumental Activities of Daily Living (IADL) and Walking Speed was not entirely unexpected. IADL measures are somewhat problematic for residents of CCRC's since many household maintenance activities such as cleaning, laundry, and regular grocery shopping may be regular institutional services. As well, IADL measures are susceptible to sex role bias. Males are much more likely than females to report they “do not do” many household maintenance activities. To minimize potential biases due to differences in CCRC service provisions and gender roles, respondents who said they did not do an IADL activity were asked if they “could do” the activity. While it is anticipated that the above step minimized respondent bias, residual measurement error and the relative small sample size may have adversely affected the power to test hypotheses pertinent to IADL levels. With respect to the timed walking measure, WALK, the absence of associations with either pain or depression level may also reflect the small sample size. As well, it may be that walking 3 meters under test conditions may not be sensitive to the effects of differential motivation or compensatory behaviors that could reflect either depression level or pain.

Mechanisms

Although the associations between pain and depression and self-reported physical performance appear to be robust, the biological, psychological, or social mechanisms accounting for the associations are not clear. Depression-related lack of motivation and pain-related inability to participate in valued activities leading to depression and subsequent diminished functioning have been put forward as potential explanatory pathways. Dohrenwend et al. [28]—in a family study of depression in women with facial chronic pain and controls—suggested that it is the stress of living with pain, not the pre-morbid depression, or a family history of depression, that is associated with the onset of major depression in patients with pain. Kaplan [29] documented that anxiety, self-reported disability, and increased trend towards having symptoms of depression were more prevalent among back pain patients who did not exert themselves physically during functional capacity evaluations. Such lack of effort due to poor motivation because of depression in pain patients may explain some of the observed deficits in functioning. Loss of valued activities also has been shown to be a risk factor for development of depression whereas overall functional decline has not [30]. In the early stages of pain, such impairment in valued activities may precipitate onset of mild depressive symptoms, subthreshold and major depression, which, later in the individual's course, may contribute further to decline in physical performance. The cross-sectional study design and the absence of data on individual motivation or pain-related loss of ‘valued’ activities preclude comprehensive evaluation of these hypotheses with the present study data. The strong association between depression level and ADL functioning, after controlling for age, gender, self-rated health, and medical conditions, provides some indirect support for the importance of depression-related loss of motivation.

In addition to the cross-sectional study design and the potential problems with the validity of some measures as noted above, the homogeneous and socioeconomically advantaged sample limits our ability to generalize study findings to the large groups of elders who reside in private dwellings in the community, who are of African or Hispanic descent, and whose lifetime personal and health resources have been modest. On the other hand, the restricted, high socioeconomic status of the sample members and their residence in a structured, although independent, living arrangement has several important advantages for the present study. Potentially confounding effects of differences in health care and in life-long access to personal and financial resources are minimized. All study participants enjoyed excellent health care insurance and comparable, unrestricted access to health services. They had achieved high educational levels, the occupations of the head of household were high in social status, and their economic success had been sufficient to afford the initial and maintenance CCRC fees.

Implications For Health Care Planning

The homogeneity, advantaged socioeconomic status, and relatively good health status of the study participants and their unrestricted access to health care, however, make the high observed prevalences of pain (78%), high levels of depressive symptoms (39%), and limiting pain with high depressive symptoms (23%) very disturbing. The under-reporting of depressed feelings by older individuals and the lack of their detection by health care providers has been documented previously. The findings of this study underscore the importance of correcting these deficits so that early detection and aggressive treatment of depression, even at subthreshold levels, can be achieved.

While older individuals may be more stigmatized by depression than by chronic pain, under-reporting of pain is a problem for health care providers. For example, in a study of institutionalized older individuals with sufficient cognitive integrity for valid communication, the treating physicians failed to identify 34% of the patients with chronic pain [7]. As well, the apparent inadequate treatment of both the pain and depressive symptoms in this sample, where health service use is not limited due to lack of access or inability to pay, is alarming. This is particularly disturbing because depression and pain are treatable conditions in the elderly. Indeed, the study findings emphasize the value of aggressive treatment of depression, especially in the individual with pain, and aggressive treatment of pain itself. Specifically, based on the data shown in Table 8, it is apparent that treating the older person's depressive symptoms should be given highest priority since, in the absence of such symptoms, even activity-limiting pain is not associated with deficits in physical function. We are not suggesting, however, that treatment of pain problems should be considered secondary. It has been shown that, compared to younger patients, geriatric patients have significant and meaningful improvement with pain center treatments [31]. It has also been reported that multidisciplinary pain management centers yield positive outcomes in older individuals [32]. Given the reluctance of older individuals to report depressive symptoms, pain is likely to be the more common presenting complaint. This study emphasizes the importance of careful, persistent evaluation to rule out depression in patients with pain complaints, and of aggressive treatment of clinically important depression (Mild or High as defined here), when it has been identified.

It is recognized that comprehensive evaluation to diagnose depression among pain patients, and consistent initiation of effective treatments for both the depression and the pain problem may require further development of comprehensive pain treatment strategies based on bio-psycho-social models involving multidisciplinary approaches [33,34]. The problems facing primary care clinicians in obtaining access for one's elderly patients to comprehensive pain treatment services must be addressed [35,36].

Outcome studies assessing the effect of such approaches on the physical functioning and quality of life of elderly patients with chronic pain are also needed. In addition, patient and physician education regarding the high prevalence of comorbid pain and depression may be required before the patient and physician can work together to address both problems. The findings from this study suggest that giving high priority to identification and treatment of pain and depression will result in substantial benefits to the older individuals who experience pain and to the physicians who are committed to their care.

References

1
Gallagher
RM
Verma
S
Mossey
J
.
The epidemiology of chronic pain in the elderly: A public health problem.
Geriatrics
 
2000
;
55
:
40
48
.
[PubMed]
2
Ross
MM
Crook
J
.
Elderly recipients of home nursing services: pain, disability, and functional competence.
J Adv Nurs
 
1998
;
27
:
1117
26
.
[PubMed]
3
Scudds
RJ
Robertson
JM
.
Pain factors associated with physical disability in a sample of community-dwelling senior citizens.
J Gerontol A Biol Sci Med Sci
 
2000
;
55
:
M393
9
.
[PubMed]
4
Heuft
G
Kruse
A
Lohmann
R
Senf
W
.
Psychosomatic aspects of painperception by the elderly: Results of the Eldermen Study [in German].
Z Gerontol Geriatr
 
1995
;
28
:
349
57
.
[PubMed]
5
Brattberg
G
Parker
MG
Thorslund
M
.
The prevalence of pain among the oldest old in Sweden.
Pain
 
1998
;
67
:
29
34
.
6
Won
A
Lapane
K
Gambassi
G
Bernabie
R
Mor
V
Lipstiz
LA
.
Correlates and management of nonmalignant pain in the nursing home. SAGE study group. Systematic Assessment of Geriatric drug use via Epidemiology.
J Am Geriatr Soc
 
1999
;
47
:
936
42
.
[PubMed]
7
Sengstaken
EA
King
SA
.
The problems of pain and its detection among geriatric nursing home residents.
J Am Geriatr Soc
 
1993
;
41
:
541
4
.
[PubMed]
8
Herr
KA
Mobily
PR
.
Chronic pain in the elderly.
In:
Smithson
E
Tripp-Reiner
T
, eds.
Advances in Gerontological Nursing-Chronic illness and the older adult
 .
New York
:
Springer
;
1997
:
82
111
.
9
Becker
N
Bondigaard
TA
Olsen
AL
Sjorgren
P
Bech
P
Erickson
J
.
Pain epidemiology and health related quality of life in chronic nonmalignant pain patients referred to Danish multidisciplinary pain center.
Pain
 
1997
;
73
:
393
400
.
[PubMed]
DOI:
10
Skevinton
SM
.
Investigating the relationship between pain and discomfort and quality of life using the WHOQOL.
Pain
 
1998
;
76
:
395
406
.
[PubMed]
DOI:
11
Parker
JC
Wright
GE
.
The implications of depression for pain and disability in rheumatoid arthritis.
Arthritis Care Res
 
1995
;
8
:
279
83
.
[PubMed]
12
Jaffe
A
Froom
J
Galambos
N
.
Minor depression and functional impairment.
Arch Fam Med
 
1994
;
3
:
1081
6
.
[PubMed]
13
Johnson
J
Weissman
MM
Klerman
GL
.
Service utilization and social morbidity associated with depressive symptoms in the community.
JAMA
 
1992
;
267
:
1478
83
.
[PubMed]
14
Broadhead
WE
Blazer
DG
George
LK
Tse
CK
.
Depression, disability days, and days lost from work in a prospective epidemiologic survey.
JAMA
 
1990
;
264
:
2524
8
.
[PubMed]
15
Mossey
JM
Knott
JA
Higgins
M
Talerico
K
.
Effectiveness of a psychosocial intervention, Interpersonal Counseling, for subdysthymic depression in medically ill elderly.
J Gerontol A Biol Sci Med Sci
 
1996
;
51
:
M172
8
.
[PubMed]
16
Dworkin
RH
Gitlin
MJ
.
Clinical aspects of depression in chronic pain patients.
Clin J Pain
 
1991
;
7
:
79
94
.
[PubMed]
17
Romano
JM
Tuner
JA
.
Chronic pain and depression: Does the evidence support a relationship?
Psychol Bull
 .
1985
;
97
:
18
34
.
[PubMed]
DOI:
18
Williamson
GM
Schulz
R
.
Pain, activity restriction, and symptoms of depression among community residing elderly adults.
J Gerontol
 
1992
;
47
:
367
72
.
19
Hoizberg
AD
Robinson
ME
Geisser
ME
Gremillion
HA
.
The effects of depression and chronic pain on the psychosocial and physical functioning.
Clin J Pain
 
1996
;
12
:
118
25
.
[PubMed]
20
Haythornthwaite
JA
Sieber
WJ
Kerns
RD
.
Depression and chronic pain experience.
Pain
 
1991
;
69
:
177
84
.
21
Lawton
MP
Moss
M
Fulcomes
M
Kleban
MH
.
A research and service oriented multi-level assessment instrument.
J Gerontol
 
1982
;
37
:
91
9
.
[PubMed]
22
Yesavage
JA
Brink
TL
Rose
TS
et al
Development and validation of a geriatric depression scale: a preliminary report.
J Psychiatr Res
 
1983
;
17
:
37
49
.
23
Blessed
G
Tomlinson
BE
Roth
M
.
The associations between quantitative measures of dementia and of senile changes in the cerebral gray matter of elderly subjects.
Br J Psychiatry
 
1968
;
114
:
797
811
.
[PubMed]
24
Williams
JBW
Gibbon
M
First
M
et al
The structured clinical interview for DSM-III-R (SCID): II. Multisite test-retest reliability.
Arch Gen Psychiatry
 
1992
;
49
:
630
6
.
[PubMed]
25
Parmelee
PA
Katz
IR
Lawton
MP
.
The relation of pain to depression among institutionalized aged.
J Gerontol
 
1991
;
46
:
15
21
.
26
Nagi
SZ
.
An epidemiology of disability among adults in the U.S.
Milbank Q
 
1976
;
54
:
439
68
.
27
Mahurin
RK
DeBittignies
BH
Pirozzola
FP
.
Structured assessment of independent living skills: Preliminary report of a performance measure of functional abilities in dementia.
J Gerontol
 
1991
;
46
:
58
66
.
28
Dohrenwend
B
Marbach
J
Raphael
K
Gallagher
RM
.
Why is depression co-morbid with chronic facial pain? A family study test of alternative hypotheses.
Pain
 
1999
;
83
:
183
92
.
[PubMed]
29
Kaplan
GM
Wurtele
SK
Gillis
D
.
Maximal effort during functional capacity evaluations: an examination of psychological factors.
Arch Phys Med Rehabil
 
1996
;
77
:
161
4
.
[PubMed]
30
Katz
PP
Yelin
EH
.
The development of depressive symptoms among women with rheumatoid arthritis. The role of function.
Arthritis Rheum
 
1995
;
38
:
49
56
.
[PubMed]
31
Cutler
RB
Fishbain
DA
Osonoff
R
Rosomoff
HL
.
Outcomes in the treatment of pain in geriatric and younger age groups.
Arch Phys Med Rehabil
 
1994
;
75
:
457
-
64
.
[PubMed]
32
Helme
RD
Katz
B
Gibson
SJ
et al
Multidisciplinary pain clinics for older people. Do they have a role?
Clin Geriatr Med
 
1996
;
12
:
563
82
.
[PubMed]
33
Gallagher
RM
.
Integrating medical and behavioral treatment in chronic pain management.
Med Clin N Am
 
1999
;
83
:
823
49
.
[PubMed]
34
Gallagher
RM
Verma
S
.
Managing pain and co-morbid depression: A public health challenge.
Sem Clin Neuropsychiatry
 
1999
;
4
:
203
20
.
35
Gallagher
RM
.
Pain medicine and primary care: A community solution to pain as a public health problem.
Med Clin N Am
 
1999
;
83
:
555
85
.
[PubMed]
36
Gallagher
RM
.
The pain medicine and primary care community rehabilitation model: Monitored care for pain disorders in multiple settings.
Clin J Pain
 
1999
;
15
:
1
3
.
[PubMed]