Home
HOME ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Services
Right arrow Download to citation manager
PubMed
Right arrow PubMed Citation
The Gerontologist 45:68-73 (2005)
© 2005 The Gerontological Society of America

Characteristics Associated With Pain in Long-Term Care Residents With Dementia

Christianna S. Williams, PhD1,2,, Sheryl Zimmerman, PhD1,3, Philip D. Sloane, MD, MPH1,4 and Peter S. Reed, PhD, MPH5

Correspondence: Address correspondence to Christianna S. Williams, PhD, Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill, 725 Martin Luther King Jr. Blvd., Chapel Hill, NC 27514. E-mail: Christianna_Williams{at}unc.edu


   Abstract
TOP
 Abstract
Methods
 Results
 Discussion
 References
 
This article describes the prevalence, assessment, and treatment of pain, as well as characteristics associated with pain, in 331 residents with dementia in 45 assisted living facilities and nursing homes. Overall, 21% of residents had pain, and pain was more commonly reported in for-profit facilities, and for those receiving professional assessment and treatment.

Key Words: Discomfort • Cognitive impairment • Nursing homes • Assisted living • Residential care

Pain is common in residents of long-term care facilities, with prevalence estimates in nursing homes ranging from 49% to 83% (Fox, Raina, & Jadad, 1999). Pain leads to negative consequences, including impaired mobility, depression, and social withdrawal (Ferrell, 1991; Ferrell, Ferrell, & Osterweil, 1990; Parmelee, Katz, & Lawton, 1991); for those with dementia, pain may exacerbate cognitive decline and agitated behavior (Buffum, Miaskowski, Sands, & Brod, 2001; Feldt, Warne, & Ryden, 1998). Pain is likely no less common in those with dementia than in the cognitively intact (Farrell, Katz, & Helme, 1996), although it is less frequently reported (Cohen-Mansfield & Lipson, 2002; Parmelee, Smith, & Katz, 1993; Sengstaken & King, 1993) and treated (Horgas & Tsai, 1998; Morrison & Siu, 2000). Reputed causes of underdetection include inability of residents with dementia to communicate their discomfort verbally (Weissman & Matson, 1999); lack of staff training (Ferrell, 1995; Weissman & Matson) and formal assessment (Weiner, 2002); and staff attitudinal barriers (Mrozek & Werner, 2001; Weiner & Rudy, 2002). Though challenging, particularly in noncommunicative demented residents, standardized assessment is recommended (Ferrell, 1995; Weiner). Once pain is identified, treatment should be administered, using both pharmacologic and nonpharmacologic approaches (AGS Panel on Persistent Pain in Older Persons, 2002; Ferrell, 1995).

While the process of pain care for cognitively impaired nursing home residents has received some research attention, less work has focused on pain in demented individuals in residential care/assisted living (RC/AL). Furthermore, existing research provides little understanding of factors that are associated with staff reporting of resident pain in both types of long-term-care settings. Therefore, this study has two objectives: (a) describe the prevalence and relevant components of pain management in nursing homes and RC/AL; and (b) identify resident and facility characteristics associated with staff reports of pain in individuals with dementia.


   Methods
TOP
 Abstract
 Methods
Results
 Discussion
 References
 
Sample
As part of the Dementia Care project of the Collaborative Studies of Long-Term Care, (Zimmerman et al., 2005, this issue), residents aged 65 and older with diagnosed dementia were randomly selected from each of 10 nursing homes and 35 RC/AL facilities in four states. The participation rate was 73% among eligible residents; a total of 421 residents were enrolled. These analyses are based on the 331 residents for whom the care supervisor provided pain data. The care supervisors in this study were predominantly nurses (77% were registered nurses [RNs] or licensed practical nurses [LPNs]), and most (75%) had been in their current position for at least one year.

Measures
Pain was measured using the Philadelphia Geriatric Center–Pain Intensity Scale (PGC–PIS; Parmelee et al., 1991), and the 0–10 pain numeric rating scale (NRS; Jensen & Karoly, 1992). For these analyses, residents scoring 2 or more on the PGC–PIS (range 1 = no pain to 5 = extreme pain) are considered to have pain. This threshold corresponds to an average across the six items of "a little" pain or worse; all residents meeting this definition had a response of "moderate" pain or worse on at least one item. Both instruments were administered to the supervisor overseeing each resident's care (for all participants) and to those residents who scored 10 or greater on the Mini-Mental State Examination (MMSE; Folstein, Folstein, & McHugh, 1975). The PGC–PIS had excellent internal consistency reliability for both supervisor and resident samples (Cronbach's {alpha} = 0.89, 0.88, respectively). Both also correlated well with the NRS (Spearman r = 0.69, supervisor; r = 0.68, resident).

In accordance with the conceptual framework of structure, process, and outcome (Donabedian, 1980) as operationalized for this study, supervisors were administered a series of questions regarding pain assessment, treatment, and training. They reported perceived presence of pain in the past week ("a moderate amount" or more) and formal pain assessment in the past year (written, standardized assessment and/or by physician or nurse). Treatment included ongoing professional treatment, other treatment (including informal), and treatment success among residents with pain in the past year (with "quite" or "extremely" well constituting "perceived success"). Staff also reported perceived training adequacy (with "quite" or "extremely" well trained in identifying and helping with resident physical pain and discomfort classified as feeling adequately trained). Facility administrators reported on facility demographic characteristics and estimated the proportion of supervisory and direct care staff who received formal training in detecting and treating pain.

Resident cognition was assessed using the MMSE and the Minimum Data Set Cognition Scale (MDS-COGS; Hartmaier, Sloane, Guess, & Koch, 1994). Functional status was measured with the MDS-ADL (Morris, Fries, & Morris, 1999), and comorbidity as the number of 10 supervisor-reported conditions (detailed in Table 1). Residents were classified as having arthritis based on supervisor report of resident arthritis, rheumatism, degenerative joint disease, lupus erythematosis, or scleroderma.


View this table:
[in this window]
[in a new window]
  		Table 1. Prevalence of Pain and Relevant Components of Care in Dementia Care Study Sample, by Setting.

 
Current medication use was ascertained by asking the supervisor to list all prescription and nonprescription medications used four or more days in the prior week. Residents taking one or more analgesics, antipyretics, nonsteroidal anti-inflammatory drugs (NSAIDs), and/or narcotics were identified as receiving pain medication. Aspirin was excluded because most aspirin use in long-term care is for cardiovascular disease prevention and is at a dosage that is subtherapeutic for pain.

Analysis
Descriptive statistics were computed separately for RC/AL facilities and nursing homes, with statistical comparison of settings for these characteristics computed using Generalized Estimating Equations (GEE; Diggle, Heagerty, Liang, & Zeger, 2002), applied to linear or logistic (for continuous and binary characteristics, respectively) models and adjusting for participant clustering by facility. Descriptive statistics were similarly computed for those for whom the supervisor reported pain and those without pain. Odds ratios and 95% confidence intervals were estimated using separate logistic regression models for each characteristic, controlling for clustering using GEE empirical standard error estimates and exchangeable correlation. Adjusted odds ratios were estimated controlling for resident age, gender, non-White race, cognitive impairment, comorbid conditions, and activities of daily living (ADL) dependencies. To ensure that factors associated with pain did not differ between nursing homes and assisted living facilities, a setting-by-characteristic interaction term was tested in each multivariable logistic model.


   Results
TOP
 Abstract
 Methods
 Results
Discussion
 References
 
Among the 331 residents for whom the supervisor reported on pain, the average age was 84.4 years, 82% were female, and 11% non-White. Ninety-two (27.8%) residents were sufficiently cognitively intact (MMSE ≥ 10) to report their own pain.

Table 1 describes the pain status of the sample. None of the measures of pain prevalence differ significantly between nursing home and RC/AL residents. Based on supervisor report, 20% to 23% had pain; among residents who self-reported, the prevalence was higher (25% to 39%). For residents with both supervisor and self-report, there was 62% agreement ({kappa} = 0.10) for the presence of pain (PGC–PIS ≥ 2).

More than one third of residents (34.7%) had no formal pain assessment. Formal assessment, all types of pain treatment, and staff training in assessment and treatment were all reported more often by nursing homes than RC/AL facilities. More than three fourths of supervisors in both settings reported success in treating resident pain and more than 90% reported feeling adequately trained to identify and help with resident pain.

When associations between resident characteristics and pain were evaluated (Table 2), depression and arthritis were identified as significant, independent correlates of pain. No association was noted between resident cognitive status and supervisor pain report.


View this table:
[in this window]
[in a new window]
  		Table 2. Characteristics Associated With Substantial Pain, Unadjusted and Adjusted.

 
When facility factors were related to pain status, residents in for-profit facilities were found to be more likely to have supervisors reporting resident pain, an effect undiminished by covariate adjustment. To explore this finding, we examined whether other aspects of pain care differed between profit and nonprofit facilities. In these post hoc analyses, we found that residents in for-profit facilities were less likely to undergo professional pain assessment (54% vs 81%, p =.008). Also, fewer residents in for-profit homes were receiving pain medication, both overall (34% vs 61%, p =.015) and among those with pain (53% vs 91%, p =.041). Consequently, when the association between for-profit status and pain was adjusted for professional assessment and medication use, along with the other covariates, the relationship strengthened (AOR = 4.3, 95% CI 1.8–10.1).

Other statistically significant associations noted between facility factors and reported pain included professional assessment and all forms of pain treatment. Still, one quarter of those with pain had not been professionally assessed, 60% had not received a standardized assessment, and 19% of residents with pain were receiving neither pain medication nor ongoing professional treatment. Treatment was less likely to be perceived as successful in those in whom current pain was reported (AOR = 0.4, 95% CI 0.2–0.8). No interactions between resident or facility characteristics and setting (nursing home vs RC/AL facility) were statistically significant (all p >.05).


   Discussion
TOP
 Abstract
 Methods
 Results
 Discussion
References
 
We found few differences in pain prevalence and components of pain management between nursing homes and RC/AL facilities, the notable exception being that more nursing home residents received pain treatment, especially medications. Roughly 40% of residents with pain in both settings were receiving no pain medication and nearly one third (31.3%) were receiving no treatment for pain. Given the Joint Commission on Accreditation of Healthcare Organizations' mandate for regular pain assessment in nursing homes (Phillips, 2000), it is notable that a substantial minority of residents received no formal pain assessment and that assessment was only slightly more common in nursing homes than RC/AL facilities.

The prevalence of pain reported here, whether by resident or supervisor report (20% to 39%) appears lower than the frequently cited prevalence of up to 80% in previous nursing home studies. However, this disparity is due at least in part to our use of a threshold of 2.0 to denote pain, rather than a broader definition of "any" pain. Applying the latter criterion to our sample, the prevalence is 62% and 76% for supervisor and resident report, respectively. Further, our supervisor and resident mean PGC–PIS scores (1.53 and 1.80, respectively) are remarkably similar to those for markedly impaired (1.47) and cognitively intact (1.80) groups previously reported (Parmelee et al., 1993). Finally, analgesic use—commonly used as an indirect measure of pain—was similar in our study (34% to 59%) to the prevalence reported in a recent review (Fox et al., 1999).

We found pain more common in for-profit facilities, independent of resident demographics and health status. This could be because more residents in for-profit facilities have pain or simply because their pain is more frequently recognized. The fact that both professional assessment and pain medication use are less common in for-profit facilities makes the latter explanation less plausible, however, and suggests a need for more assessment and care in these facilities.

We used proxy report as the primary pain measure. For the 27% of residents who administered a self-reported pain measure, agreement with the supervisor was modest, albeit similar to (Werner, Cohen-Mansfield, Watson, & Pasis, 1998) or better than (Horgas & Dunn, 2001; Weiner, Peterson, Ladd, McConnell, & Keefe, 1999) that reported by others. While a potential study limitation, proxy report allowed the standardized assessment of pain using a reliable, valid pain instrument (Parmelee, 1994; Parmelee et al., 1991; Parmelee et al., 1993) for the entire sample of cognitively impaired long-term-care residents (mean MMSE = 8.1). In fact, a recent study found that correlation between caregiver and care-recipient pain report was stronger for the PGC–PIS than for other pain measures examined (Krulewitch et al., 2000). The great majority (84%) of supervisors had known the residents for whom they reported pain for at least a year, and 88% reported that they knew the residents "pretty well" or "very well." Further, proxy report is salient because recognition of pain by staff is an essential first step in its effective management (Kamel, Phlavan, Malekgoudarzi, Gogel, & Morley, 2001), and we wished to identify correlates of this important endpoint. While it is not surprising that pain was noted in those for whom current treatment is reported, the finding that treatment is less likely to be deemed successful in those with pain suggests that staff are cognizant of the residents for whom pain management remains incomplete. This finding may be considered in conjunction with the very high proportion of staff who report feeling adequately trained to assess and treat pain in the residents under their care, perhaps implying that some staff are unaware of other techniques to treat pain or feel that additional intervention is either not warranted or unavailable.


   Footnotes
 
This research was supported by grants from the National Alzheimer's Association, Program and Community Services Division and Medical and Scientific Division (Grant No. IIRG-00-2222). The authors express appreciation for the cooperation of the staff, residents, and families participating in the Collaborative Studies of Long-term Care. Gratitude also is extended to Malaz Boustani, MD, MPH, for his helpful comments on early drafts and to Ms. Jane Darter for expert data management. Back

1 Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill. Back

2 Department of Epidemiology, School of Public Health, University of North Carolina at Chapel Hill. Back

3 Department of Social Work, University of North Carolina at Chapel Hill. Back

4 Department of Family Medicine, University of North Carolina at Chapel Hill. Back

5 Alzheimer's Association, National Office, Chicago, Illinois. Back

Decision Editor: Richard Schulz, PhD

Received for publication June 22, 2004. Accepted for publication September 24, 2004.


   References
TOP
 Abstract
 Methods
 Results
 Discussion
 References
 





This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Services
Right arrow Download to citation manager
PubMed
Right arrow PubMed Citation

HOME ARCHIVE SEARCH TABLE OF CONTENTS