Associations of back and leg pain with health status and functional capacity of older adults: Findings from the retirement community back pain study

Authors


Abstract

Objective

Low back pain (LBP) is the most frequently reported musculoskeletal problem in older adults, but its impact on health status is not well understood. Our objective was to determine whether LBP and concurrent leg pain are associated with health-related quality of life (HRQOL) and function in a cohort of older adults, and to examine care-seeking behaviors related to LBP.

Methods

This was a population-based, cross-sectional survey study of 522 community-dwelling men and women (67.4%) ages ≥62 living in 4 retirement communities in Maryland and northern Virginia. LBP status in the past year was categorized as no pain in the low back or leg, LBP only, and LBP with leg pain. HRQOL and function were measured with the Medical Outcomes Study Short Form 36 (SF-36).

Results

A total of 26.8% of the sample reported LBP only and 21.3% reported LBP plus leg pain. Participants with LBP and LBP plus leg pain had lower scores in all SF-36 domains, reflecting worse HRQOL (P < 0.0001). LBP and LBP plus leg pain were associated with 2-fold greater odds of falling and increased difficulty lifting grocery bags, walking several blocks, and bathing. LBP plus leg pain was associated with difficulty in social interactions (odds ratio 10.63, 95% confidence interval 3.57–31.60). Less than half sought care for LBP and those who did had poorer health status and greater pain burden.

Conclusion

LBP is common among older adults and strongly associated with reduced HRQOL and function. These findings argue strongly for both identifying cases of LBP by health care practitioners and pursuing effective treatments for LBP given the potential consequences.

INTRODUCTION

Low back pain (LBP) is the most frequently reported musculoskeletal problem and the third most frequently reported symptom in people over age 75 (1, 2). More than 17% of all visits to physicians for LBP involve individuals over age 65 (1, 3, 4). In terms of societal impact, Medicare data between 1991 and 2002 indicate a 132% increase in patients with LBP and a 387% increase in related charges for LBP (5). In older adults, LBP has been associated with several negative consequences, including decreased physical function (6–8) and increased disability (9), likelihood of falls (10), and mood disturbance (11). A significant proportion of elderly persons with LBP also report concurrent radicular leg pain (12), often due to lumbar spinal stenosis caused by degenerative narrowing of the spinal canal and/or intervertebral foramen leading to compression of the caudae equinae and/or lumbar nerve roots. Recent evidence suggests that the combination of LBP and leg pain among postmenopausal women is associated with greater deficits in health status than LBP alone (12). These findings need confirmation in a sample including older women and men.

Despite the high prevalence of LBP and its contribution to poor outcomes in older adults (6–9), surprisingly, little research has focused on LBP in people over age 65 (1, 13). LBP is not necessarily an inevitable consequence of aging and can potentially be managed with medication, exercise (14), cognitive–behavioral therapy (15), and electrical stimulation (16) to improve quality of life and reduce negative consequences. Given the expected dramatic increase in the older population over the next 25 years (17), prevalence of chronic pain and its associated societal costs will inevitably increase. To effectively treat older adults with back and leg pain, clinicians and researchers first need to gain a better understanding of the serious impact of back and leg pain on the overall health status of older adults.

The present study examined cross-sectional associations between back pain status (LBP alone or LBP with leg pain) and general health status, as well as functional capacity, in a cohort of older adults living in a continuing care retirement community (CCRC) setting. We hypothesized that LBP with leg pain would be associated with poorer health status and functional capacity than LBP alone or no pain. We also examined care-seeking behaviors related to back and leg pain to gain a sense of the level of assessment and treatment in a population with high access to health care.

PARTICIPANTS AND METHODS

Study design.

The Retirement Community Back Pain Study is a population-based survey study of adults ages 62 years or older living in 1 of 4 CCRCs in Maryland and northern Virginia that are privately owned and operated by 1 company. The CCRCs are designed similarly to the common condominium ownership model where residents buy into the community and pay a monthly maintenance fee. Pricing is geared toward homeowners from the middle income bracket (e.g., teachers, nurses, and police officers). Monthly maintenance fees are based on apartment size and include utilities, security, and 1 meal per day. Additional services (both personal and medical) are available on a fee-for-service basis. Eligibility for community entrance is based only on age (62 years) and financial ability to meet the entrance and monthly fee. The average age for community entrance is 78 years. Eligible participants comprised anyone age 62 years and older living independently in the CCRC. There were no exclusions. Study procedures were exempt from institutional review because the survey results could not be linked to respondents; however, the study was approved by the National Institute on Aging Intramural Program (via Clinical Investigator's Meeting) and the Research Advisory Board for the CCRCs.

Survey development.

A cover letter explaining the study and survey instrument was developed and tested on a small group of older adults to identify ambiguous or confusing questions and instructions. Adjustments were made as necessary.

The survey included specific questions regarding sociodemographic information, presence of chronic conditions, functional limitations, and pain, as well as the Medical Outcomes Study Short Form 36 (SF-36), a widely used and well-validated questionnaire developed to examine general health status (18, 19).

Recruitment and survey administration.

A random sample of 1,000 older adults living in the 4 CCRCs were surveyed. To generate the random sample, a staff member from one of the CCRCs entered all residents of the 4 communities into a database and assigned them a number. Within each community, 250 numbers were randomly generated using a computer program and the survey was sent to residents assigned to the randomly generated number. Only the staff member had access to the residents' number assignment key to ensure that investigators were masked to the respondents' identities. Surveys were delivered through the internal mail systems for each community to the selected sample. Potential respondents were given 4 weeks to complete and anonymously return the survey to a centralized drop box in each CCRC.

Pain measurement.

Participants were shown a homunculus depicting the low back region as extending from the bottom of the ribs down to the top of the legs. They were then asked, “In the past year, have you had any low back pain? If yes, please rate your usual back pain over the past year on a scale from 0 to 10.” The scale was shown with numbers from 0 to 10 on a line anchored with “no pain” and “extremely intense pain,” respectively. Next, participants were asked, “Have you had any radiating leg pain associated with your low back pain in the past year, excluding joint pain in the hip, knee or foot?” Participants responding “yes” were then asked to rate their leg pain in a similar manner as was done for LBP. Back pain status was categorized for each participant as no back or leg pain, LBP only, and LBP with leg pain.

Health-related quality of life.

Health-related quality of life (HRQOL) was measured using the SF-36, a generic self-administered questionnaire used to examine health in the following 8 domains: bodily pain, physical function, general health, vitality, mental health, social function, and role limitations due to social and physical problems. Scores on each domain were transformed into 0–100 scales, with higher scores representing better health status. Two subscale scores representing overall physical (Physical Component Summary [PCS]) and mental health (Mental Component Summary [MCS]) were also examined. These 2 scales use norm-based scoring with each scale scored to have the same mean ± SD of 50 ± 10 points (20, 21). Therefore, scores <50 represent below average health status compared with the rest of the population. Psychometric properties of this instrument are well established for use in multiple populations (19, 21).

Functional limitation.

In addition to analyzing the SF-36 items within their respective domains and subscales, individual items were examined to ascertain specific information on functional limitations. In the SF-36, participants were asked if they had difficulty performing any of the following tasks: lifting or carrying grocery bags, climbing a flight of stairs, bending/kneeling/stooping, walking several blocks, walking 1 block, and bathing. For this study, functional limitation was defined as reporting much difficulty with the specific task or inability to perform the task. Social interference was considered present if participants reported “quite a bit” or “extreme” interference due to physical problems. In addition to these specific items from the SF-36, participants were asked about their use of assistive devices for walking (yes/no) and whether they had fallen in the past year (yes/no).

Care-seeking behavior.

Participants were asked, “In the past year, have you sought any health care related to your low back pain?” If they answered yes, they were asked, “From whom did you seek care? Please choose all that apply: Physician, Chiropractor, Physical Therapist, Nurse, Massage Therapist, Acupuncturist, Other.” Participants were also asked, “In the past year, have you taken any prescription medications for your low back pain?”

Covariates.

Sociodemographic and health characteristics were examined as potential confounders of the relationship between pain and health/functional status. Covariates included age, sex, marital status, educational level, body mass index (BMI), depressive symptomatology, and chronic conditions. BMI was calculated as weight in kilograms divided by height in meters squared based on self-reported data. Depressive symptomatology was evaluated using the following question from the SF-36: “During the past week, have you felt downhearted and blue?” A response of “none of the time” was coded as negative for depressive symptomatology. For chronic conditions, participants were asked if a doctor had ever told them that they have any of the following conditions: osteoarthritis, osteoporosis, cancer, hip fracture, diabetes, stroke, heart disease, Parkinson's disease, or pulmonary disease. All of the aforementioned covariates were chosen due to their known associations with HRQOL. Additionally, sex, BMI, depressive symptoms, osteoarthritis, osteoporosis, and Parkinson's disease are all associated with back pain and, therefore, may be confounders.

Statistical analysis.

All analyses were performed using SAS software, version 9.1 (SAS Institute, Cary, NC). Participants were grouped based on back pain status, as described above. HRQOL variables from the SF-36 were examined across pain categories using analysis of variance. To better understand the magnitude of differences between pain-free participants and those with back and leg pain, standardized effect sizes were calculated for all domains of the SF-36, including the PCS and MCS subscales of the SF-36. Standardized effect size was calculated by subtracting the mean SF-36 value of the LBP plus leg pain group from the mean value of the no pain group and then dividing the difference by the pooled standard deviation of the 2 scores (22). Odds ratios (ORs) and 95% confidence intervals (95% CIs) were calculated using multinomial logistic regression to determine associations between back pain status and functional limitations, with the no pain group as the reference category. Respondents with LBP were also categorized by care-seeking behavior for their LBP in the past year. We examined participant characteristics across care-seeking categories using t-tests for continuous variables and chi-square tests for categorical variables. To examine associations independent of other factors, all analyses were adjusted for age, sex, marital status, educational level, BMI, depressive symptomatology, and each chronic condition.

RESULTS

Characteristics of the study sample are shown in Table 1. The average age of the participants was 81.3 years; 67.4% were female, 98.8% were white, and 49.4% were living with a spouse. Fifty-two percent of those surveyed responded and supplied data for this study. Although it is not possible to provide data regarding the nonresponders, the following is some basic demographic information on all residents in the 4 CCRCs that were sampled: the average age was 83.8 years, 67.1% were female, 99.4% were white, and 49.2% were living with a spouse. Of 522 participants, 251 (48.1%) reported LBP in the past year. Of the total sample, 26.8% reported LBP only and 21.3% reported LBP with leg pain in the past year. Participants with LBP plus leg pain were younger, had a smaller proportion of people in the ≥80 age group, had more falls, and had greater use of assistive devices for walking compared with others. They also had higher pain intensity and longer symptom duration.

Table 1. Participant characteristics according to back pain status*
CharacteristicWhole sample (n = 522)No pain (n = 271)LBP only (n = 140)LBP + leg pain (n = 111)
  • *

    Values are the number (percentage) unless otherwise indicated. LBP = low back pain.

  • P < 0.05 for comparison of LBP only versus no pain (chi-square tests for categorical variables and t-tests for continuous variables).

  • P < 0.05 for comparison of LBP + leg pain versus LBP only, and P < 0.05 for comparison of LBP + leg pain versus no pain (chi-square tests for categorical variables and t-tests for continuous variables).

  • §

    P < 0.05 for comparison of LBP + leg pain versus no pain (chi-square tests for categorical variables and t-tests for continuous variables).

  • P < 0.0001 for comparison of LBP + leg pain versus LBP only (chi-square tests for categorical variables and t-tests for continuous variables).

Categorical variables    
 Age category, years    
  60–6915 (2.9)5 (1.9)4 (2.9)6 (5.4)
  70–79177 (33.9)77 (28.4)54 (38.6)46 (41.4)
  ≥80330 (63.2)189 (69.7)82 (58.6)59 (53.2)
 Female352 (67.4)182 (67.0)98 (71.0)72 (65.5)
 White512 (98.8)266 (98.9)136 (97.8)110 (99.1)
 College graduate222 (42.5)111 (41.0)68 (48.2)43 (38.7)
 Living with spouse257 (49.4)128 (47.6)67 (47.9)62 (55.9)
 Fallen in past year138 (27.2)59 (22.6)43 (31.2)36 (33.3)§
 Uses a walking aid128 (24.8)58 (21.6)28 (20.0)42 (38.5)
 Depressive symptoms221 (42.9)100 (37.6)68 (48.9)53 (48.2)§
Continuous variables, mean ± SD    
 Age, years81.3 ± 5.782.0 ± 5.581.0 ± 5.579.8 ± 6.3§
 Weight, pounds154.9 ± 34.5152.4 ± 31.9154.2 ± 38.0161.7 ± 35.0§
 Average LBP intensity (range 0–10)4.1 ± 1.85.3 ± 1.9
 Consecutive weeks with LBP13.9 ± 21.527.3 ± 23.9

Tables 2 and 3 delineate unadjusted and adjusted differences, respectively, in HRQOL according to LBP status. Twenty participants provided incomplete data for the SF-36; therefore, all analyses examining HRQOL and functional limitations included data from 502 participants. There were no significant differences in baseline characteristics between participants with missing data and those without missing data (P > 0.05). There was a clear trend toward reduced health status with worsening of LBP status in all domains of the SF-36 (no pain → LBP only → LBP plus leg pain). Standardized effect sizes, which aid in understanding the magnitude of differences between groups, ranged from a moderate effect of 0.30 for mental health to a large effect of 1.01 for bodily pain in the adjusted model. As indicated by the summary scores in Table 2, participants with LBP plus leg pain had greater disruption in physical capacity than those without LBP (P < 0.0001). LBP was more highly associated with physical function than mental function.

Table 2. Unadjusted means and effect sizes for the association between back pain status and health-related quality of life as measured by the SF-36*
SF-36 domainsBack pain statusEffect sizeP
No pain (n = 259)LBP only (n = 135)LBP + leg pain (n = 108)LBP + leg pain vs. no pain
  • *

    Values are the mean (95% confidence interval) unless otherwise indicated. SF-36 = Short Form 36; LBP = low back pain; PCS = Physical Component Summary; MCS = Mental Component Summary.

  • Mean domain score for LBP + leg pain minus mean score for no pain group divided by pooled SD.

  • P values compare across LBP categories using analysis of variance.

Physical function68.6 (65.5–71.8)61.1 (56.8–65.4)47.9 (43.1–52.7)0.80< 0.0001
Role physical69.5 (64.7–74.4)56.9 (50.2–63.5)42.7 (35.2–50.2)0.72< 0.0001
Bodily pain76.3 (73.6–78.9)62.4 (58.7–66.0)48.3 (44.1–52.5)1.00< 0.0001
General health69.5 (67.2–71.8)66.8 (63.7–69.9)56.4 (52.9–59.9)0.69< 0.0001
Vitality62.4 (60.1–64.7)57.1 (53.9–60.3)45.3 (41.7–48.9)0.90< 0.0001
Social function87.9 (85.2–90.5)80.9 (77.3–84.7)71.7 (67.6–75.9)0.61< 0.0001
Role emotional81.9 (77.6–86.2)82.1 (76.3–87.9)68.8 (62.2–75.5)0.530.0028
Mental health83.2 (81.5–84.9)79.3 (76.9–81.7)75.7 (72.9–78.4)0.35< 0.0001
PCS50.0 (48.5–51.6)43.9 (41.8–45.9)36.7 (34.4–39.1)1.13< 0.0001
MCS52.2 (50.8–53.5)49.8 (47.9–51.6)44.3 (42.2–46.3)0.76< 0.0001
Table 3. Adjusted means and effect sizes for the association between back pain status and health-related quality of life as measured by the SF-36*
SF-36 domainsBack pain statusEffect sizeP
No pain (n = 259)LBP only (n = 135)LBP + leg pain (n = 108)LBP + leg pain vs. no pain
  • *

    Values are the mean (95% confidence interval) unless otherwise indicated. See Table 2 for definitions.

  • Mean domain score for LBP + leg pain minus mean score for no pain group divided by pooled SD.

  • P values compare across LBP categories using analysis of covariance. Analyses are adjusted for age, sex, marital status, educational level, body mass index, depressive symptom, and each chronic condition (osteoarthritis, osteoporosis, cancer, hip fracture, diabetes, stroke, heart disease, Parkinson's disease, and pulmonary disease).

Physical function68.3 (64.7–71.8)62.6 (57.9–67.4)48.2 (41.7–54.7)0.77< 0.0001
Role physical71.2 (65.9–76.6)58.6 (51.1–66.1)44.8 (34.7–54.9)0.71< 0.0001
Bodily pain76.6 (73.5–79.7)65.3 (61.1–69.4)48.3 (42.6–53.9)1.01< 0.0001
General health69.3 (66.7–71.9)68.5 (64.9–71.9)60.0 (55.1–64.9)0.49< 0.0001
Vitality62.5 (59.8–65.3)58.6 (54.9–62.2)45.9 (40.9–50.9)0.88< 0.0001
Social function89.0 (85.9–92.0)83.0 (78.9–87.1)73.9 (68.4–79.6)0.57< 0.0001
Role emotional85.2 (80.5–89.8)85.2 (78.9–91.5)67.2 (58.6–75.8)0.73< 0.0001
Mental health83.3 (81.5–85.1)80.5 (78.1–82.9)76.8 (73.5–80.1)0.30< 0.0001
PCS49.8 (48.1–51.5)44.9 (42.7–47.3)37.9 (34.9–41.0)1.01< 0.0001
MCS52.5 (51.1–53.9)50.9 (48.9–52.9)44.9 (42.3–47.5)0.73< 0.0001

Presence of LBP only, regardless of intensity, was associated with greater odds of difficulty with certain functional tasks, i.e., lifting, climbing stairs, and bathing (Table 4). However, participants with LBP plus leg pain had at least a 3-fold greater likelihood of having a lot of difficulty in all tasks examined as compared with those without LBP. For example, older adults with LBP plus leg pain had nearly 6-fold higher odds of having a lot of difficulty with lifting grocery bags. Compared with those without LBP, participants with LBP only or LBP plus leg pain had twice the odds of having fallen in the past year. Participants with LBP plus leg pain were more likely to use an assistive device for walking, report fair/poor self-rated health, and report social interference related to physical problems (OR 10.63, 95% CI 3.57–31.60).

Table 4. Odds ratios (95% confidence intervals) for associations between back pain status and task difficulty*
 No painLBP onlyLBP + leg pain
  • *

    Adjusted for age, sex, marital status, educational level, body mass index, depressive symptom and each chronic condition (osteoarthritis, osteoporosis, cancer, hip fracture, diabetes, stroke, heart disease, Parkinson's disease, and pulmonary disease). No pain is the reference group. LBP = low back pain.

Functional limitation   
 Difficulty lifting or carrying grocery bags1.002.07 (1.26–3.43)5.80 (3.17–10.62)
 Difficulty climbing a flight of stairs1.001.80 (1.07–3.01)4.12 (2.11–8.02)
 Difficulty bending, kneeling, or stooping1.001.56 (0.94–2.65)3.23 (1.64–6.37)
 Difficulty walking several blocks1.001.91 (1.16–3.17)4.60 (2.53–8.38)
 Difficulty walking 1 block1.001.13 (0.63–2.03)4.59 (2.52–8.38)
 Difficulty bathing and dressing1.002.49 (1.16–5.34)3.83 (1.74–8.44)
Health status   
 Fallen in past year1.001.95 (1.15–3.31)2.11 (1.17–3.82)
 Requires assistive device for walking1.000.78 (0.39–1.51)3.17 (1.67–6.01)
 Fair/poor self-rated health1.000.89 (0.44–1.84)2.29 (1.16–4.53)
 Social interference due to physical problems1.001.58 (0.41–6.09)10.63 (3.57–31.60)

Forty-five percent of participants with some form of LBP sought care, and all sought care from a physician; no other health care practitioners were seen for treatment. Of those with LBP only, 30% sought care compared with 64.5% of those with LBP plus leg pain. Only 37.7% reported use of prescription medications for LBP in the past year. As seen in Table 5, participants with LBP who sought care for their pain were more likely to use an assistive device for walking and to report osteoarthritis. Osteoarthritis was the only chronic condition to differ by care-seeking status. Participants with LBP who sought care also had lower health status as measured by the PCS and MCS, as well as higher pain severity and longer duration of LBP in the past year (Table 5).

Table 5. Characteristics of participants who sought care for LBP*
VariableDid not seek care (n = 138)Sought care (n = 113)P
  • *

    Values are the number (percentage) unless otherwise indicated. Analyses are adjusted for age, sex, marital status, educational level, body mass index, depressive symptom, and each chronic condition (osteoarthritis, osteoporosis, cancer, hip fracture, diabetes, stroke, heart disease, Parkinson's disease, and pulmonary disease). LBP = low back pain; PCS = Physical Component Summary; MCS = Mental Component Summary.

  • From chi-square tests for categorical variables and t-tests for continuous variables.

Categorical variables   
 Back pain status   
  LBP only99 (72.1)42 (37.5)< 0.0001
  LBP + leg pain39 (27.9)71 (62.5) 
 Age category, years   
  60–695 (3.7)5 (4.5)0.88
  70–7953 (38.9)46 (41.1) 
  ≥8078 (57.4)61 (54.5) 
 Female sex86 (64.2)83 (74.8)0.07
 College graduate61 (45.2)48 (42.9)0.71
 Living with spouse65 (47.8)62 (55.4)0.24
 Fallen in past year43 (32.1)35 (32.1)0.99
 Uses a walking aid29 (21.6)40 (35.7)0.01
 Osteoarthritis43 (31.1)79 (69.7)< 0.0001
 Depressive symptoms61 (44.1)62 (54.6)0.10
Continuous variables, mean ± SD   
 PCS44.34 ± 12.4137.33 ± 13.23< 0.001
 MCS50.05 ± 11.4144.05 ± 13.410.002
 Average LBP intensity (range 0–10)3.93 ± 1.665.32 ± 1.99< 0.0001
 Consecutive weeks with LBP10.63 ± 19.9426.35 ± 23.63< 0.0001

Given the lack of strong evidence for an association between care seeking and depressive symptoms using the single-item question from the SF-36, we performed a post hoc analysis using the MCS subscale of the SF-36. Walsh et al have previously demonstrated that an MCS score <35 can be used to screen for depressive symptoms in patients with spinal pain (23). Using the MCS score as a screening tool, we found that care seekers were more likely to demonstrate depressive symptoms than those who did not seek care (26% versus 12%; P = 0.0066).

DISCUSSION

LBP is a common but inadequately studied problem in older adults. Nearly half of the participants in the present sample reported some degree of LBP in the past year. Consistent with other studies of elderly persons (12), our findings demonstrate that LBP plus leg pain is associated with compromised HRQOL and multiple functional limitations. Our findings demonstrate that LBP is strongly linked to HRQOL in older adults, particularly when leg pain is present. The mean PCS score for those without LBP was slightly above the population norm of 50, while the mean PCS for those with LBP plus leg pain was nearly 1.5 SDs below the population norm for physical health status, a significant deficit. Vogt et al found a similar pattern between similarly defined groups in a cohort of postmenopausal women, but the magnitude of PCS difference was less (47.9 versus 42.1) (12). In contrast to the findings by Vogt et al, we also demonstrated lower MCS scores among participants with LBP plus leg pain. Similar to the HRQOL findings, LBP alone and LBP plus leg pain were associated with significant limitations in mobility performance and fall status. The strong association between LBP plus leg pain and social interference related to physical problems is quite concerning in this population because social interaction in late life is associated with fewer depressive symptoms and longer survival (24). Properly addressing LBP may be an effective way to prevent and ameliorate mobility limitations, as well as improve quality of life in older adults.

Remarkably, despite significant pain burden, less than half of those with LBP sought care for their symptoms. Whether or not the participants who sought care for their LBP were already actively seeing a physician for another reason could not be discerned. However, in an effort to gain a better understanding of why certain participants may have sought care, we examined differences in basic characteristics by care-seeking status. We expected that higher educational status might be associated with care seeking, but this was not the case. We also expected that women would be more likely to seek care, and there seemed to be some evidence of an association. Of all the chronic conditions examined, only self-report of osteoarthritis was associated with care seeking. We do not know if these participants were told that they had spinal osteoarthritis or osteoarthritis in some other joints. In addition, they may have been told they had osteoarthritis when they sought care for LBP.

As expected, participants who sought care had a greater pain burden as exemplified by higher pain intensity. The care seekers were also more likely to have a pain distribution that included both LBP and leg pain. Based on the other findings in this study, LBP plus leg pain is clearly associated with greater functional and quality of life deficits. Additionally, care seekers, on average, experienced a longer duration (∼15 weeks longer) of symptoms. The longer duration of pain may explain a substantial portion of the poor health status seen in those who sought care.

In terms of overall physical health, care seekers scored more than 0.5 SDs below those who did not seek care on the PCS subscale of the SF-36. This finding of lower physical health status is also supported by the increased likelihood of walking aid use among care seekers. As with physical health, lower scores on the MCS subscale of the SF-36 indicated worse overall mental health among care seekers. Although we did not see strong evidence for an association between care seeking and depressive symptoms using the single-item question from the SF-36, we did see a relationship in a post hoc analysis using the MCS subscale of the SF-36 as a screening tool for depressive symptoms. The overall worse physical and mental state of the care seekers seems to suggest that older adults are more likely to seek care after limitations become present; however, it is impossible to know for sure based on these cross-sectional findings.

Two mainstays in the conservative management of LBP are active rehabilitation and medication use. Unexpectedly, in this cohort, no respondents received physical therapy or chiropractic services and less than 40% took prescription medications for their LBP. We only asked about prescription medications for LBP, but it is certainly possible that participants were taking over-the-counter preparations based on physician advice or their own decision making. It would be interesting to know what advice or care was provided to those who sought care with a physician; however, these data were not available in this study. Why did so few older adults seek care for LBP and leg pain? A possible reason is that they were more concerned about other health issues they perceived as more important, i.e., cardiovascular health or diabetes; however, the proportion of other major chronic conditions was not higher in participants who did not seek care compared with those who did see care. Another possible reason is that respondents believed that pain is a normal part of aging that cannot be avoided. A better understanding of these issues is needed to improve management of LBP in the geriatric population. Particularly for older adults who are already at an increased risk of functional decline, it seems logical that conservative pain management be pursued and started as early as possible to potentially avoid the debilitating pain-disability cycle often seen in patients with chronic pain. Effective treatment of LBP in the general population is already difficult due to its heterogeneous and multifactorial nature, but evaluation and treatment become even more problematic in the geriatric population due to the significant amount of degenerative changes to the spine and its supporting muscular system (7, 8) seen with aging. Although interventions in the older population have been understudied, potential treatment options have begun to be evaluated in older adults, including percutaneous electrical stimulation (16), aerobic training (14), and cognitive–behavioral therapy (15).

A major advantage of this study is the use of a large community-dwelling sample of older men and women with specific data on both LBP and leg pain as well as measures of HRQOL and function. Nonetheless, several limitations must be considered in interpreting the findings. First, the cross-sectional nature of the data precludes identifying LBP and leg pain as the causes of reduced health status and function; however, it is certainly a plausible pathway. This pathway should be further examined longitudinally. Another limitation is the low response rate of 52% for this survey, which is not uncommon in studies of older adults, and the resultant potential for nonresponse bias. The sample is representative of the larger population in the 4 CCRCs examined with respect to demographic characteristics (age, sex, race, living arrangement). However, it is still possible that older adults with LBP might be more inclined to complete a survey about LBP; therefore, the proportion of people with LBP in this sample should not be assumed to represent prevalence of LBP in the CCRC setting. A further limitation is the racial and relative economic homogeneity of the sample, which limits the generalizability to older white adults living in a CCRC setting. Given that older adults living in these CCRCs come from the middle income bracket and that the CCRC is a best-case scenario for care-seeking behaviors in terms of available resources, it is very possible that the relationships seen here between pain and health status will likely be stronger in a more racially and economically diverse population of older adults; this is an area of research that is needed. An additional limitation is that, given the strong relationship between psychosocial functioning and LBP, it would have been beneficial to validate questionnaires specifically designed to assess anxiety and depressive symptoms in this sample to use as covariates in modeling the relationship between LBP and HRQOL. Although the SF-36 has been validated for use in multiple populations, some studies have demonstrated difficulty using this instrument in older adult populations (25). Finally, given the limitations of self-report measures in general, the results of the study would be stronger if, in addition to the self-report data, we also had performance-based measures of function to substantiate the relationship between pain and function. Further work that includes physical examination and performance testing is warranted to flesh out the relationship between LBP status and health status.

In conclusion, LBP is a common and debilitating condition in older adults; concomitant leg pain is also common and carries a much greater burden. The combination of LBP and leg pain is associated with a striking reduction in physical and mental quality of life, as well as functional capacity. Surprisingly, less than half of elderly persons living in the retirement community setting pursued treatment for their LBP and associated symptoms. This combination of high prevalence and low care-seeking behavior suggests that clinicians who see older patients should routinely ask targeted questions about back pain and make appropriate referrals as needed to potentially prevent functional decline in the already vulnerable older population. Additionally, both clinicians and researchers should expend efforts to pursue effective, active treatments for older adults with LBP and/or leg pain given the potential deleterious consequences.

AUTHOR CONTRIBUTIONS

Dr. Hicks had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Hicks, Gaines, Simonsick.

Acquisition of data. Hicks, Gaines.

Analysis and interpretation of data. Hicks, Shardell, Simonsick.

Manuscript preparation. Hicks, Gaines, Shardell, Simonsick.

Statistical analysis. Hicks, Shardell.

Acknowledgements

The authors thank the staff at the following Erickson Retirement Community facilities for their assistance with data collection: Charlestown, Oak Crest, Riderwood, and Greenspring.

Ancillary