Pain is the most common symptom of many chrome diseases in older adults. Studies of the prevalence of pain in nursing home residents indicate between 66% and 83% of subjects interviewed indicate the presence of pain at least some of the time (Ferrell, Ferrell, & Osterweil, 1990; Ferrell, Ferrell, & Rivera, 1995; Parmelee, Smith, & Katz, 1993; Roy & Thomas, 1987; Sengstaken & King, 1993). The prevalence of pain complaints for people residing in board-and-care units was significantly higher (84%) than pain complaints for residents on skilled units (66%o) (Ferrell et al, 1990). Researchers noted the residents on the skilled units were more likely to be cognitively impaired but found no correlation between pain and mental status scores. Other researchers reported that as cognitive impairment increased, pain report decreased for hospitalized hip fracture patients, although more disabled, nonverbal, and markedly cognitively impaired patients were excluded from the study (Roberts & Eastwood, 1994). Parmelee, Smith, and Katz (1993) noted that pain complaints were less in people with greater cognitive impairment, yet there were no differences in the potential physical causes for pain in those subjects. Most of the above studies routinely excluded the more severely cognitively impaired subjects who were thought to be unreliable in their report of pain.
Self-report of pain has been the primary mode of assessing the location, intensity, and duration of pain. McCaffery and Beebe's (1989) definition that "pain is whatever the experiencing person says it is, existing whenever the person says it does" (p. 31) emphasizes the need to believe clients' statements about pain but creates problems in the assessment of pain in people who are nonverbal. Language skill changes in people with dementia that impair their ability to report pain include word-finding difficulties, paraphasias (using a word other than the intended word to explain something), anomias (inability to name objects or parts), and difficulties acting on what is said or heard. People with dementia may neglect to report pain because of an inability to adequately conceptualize, express, or explain the perceptions they have been experiencing (Parmelee et al., 1993). These cognitive changes limit the clients' ability to respond to questions on standardized pain assessment tools.
Few researchers have tested the ability of institutionalized cognitively impaired older adults to use standardized tools for pain report. One study of 217 cognitively impaired nursing home residents (mean MiniMental State Examination [MMSE] score = 12.1) on skilled units found 83% of subjects were able to complete at least one of five pain intensity tools tested (Ferrell et al., 1995). A study of postoperative hip fracture patients reported 73% of cognitively impaired subjects (mean MMSE score = 12.2) were able to complete the Verbal Descriptor Scale, a pain intensity tool containing seven brief pain severity descriptions (from "no pain" and "slight pain" to "extreme pain" and "pain as bad as it could be") (Feldt, Ryden, & Miles, 1998; Herr & Mobily, 1993).
Some research indicates cognitively impaired older adults who have retained their verbal skills may be as reliable as cognitively intact older adults when asked simple yes or no questions regarding pain (Feldt et al, 1998). However, the impaired group in the study was less accurate than the intact group when retest reliability for the Verbal Descriptor Scale was determined.
Nonverbal, cognitively impaired residents may communicate pain through behaviors that are not included in standardized pain assessment tools, such as vocalizations or crying out, facial expressions of grimacing or wincing, wrinkling of the forehead in response to movement, increased restlessness, rocking, rubbing or guarding of a body part, increased irritability, aggressive behaviors, or resistance to personal care which requires movement (Herr & Mobily, 1991; Marzinski, 1991; Parke, 1992; Ryden & Feldt, 1992; Sengstaken & King, 1993). Some attempts have been made to quantify pain and discomfort in nonverbal clients. A scale for measuring discomfort in noncommunicative clients with advanced Alzheimer'stype dementia has been developed (Hurley, Volker, Hanrahan, Houde, & Volicer, 1992). The Checklist of Nonverbal Pain Indicators (CNPI) is another tool tested for use with people with dementia (Feldt, 1996). This scale includes the behaviors most frequently cited by other researchers (e.g., verbal and nonverbal vocalization, grimacing, bracing, rubbing, restlessness).
The possibility that inadequate recognition and treatment of pain may precipitate aggressive behaviors has been suggested by several researchers (Boettcher, 1 983; CohenMansfield, Billig, Lipson, Rosenthal, & Pawlson, 1990). In one study, most aggressive behaviors in cognitively impaired nursing home residents were found to occur with physical contact or movement, suggesting pain may be triggered by movement, which in turn stimulates aggressive behaviors (Ryden, Bossenmaier, & McLachlan, 1991).
There are other informational sources that may aid in the assessment of pain in cognitively impaired older adults. Family members who have knowledge regarding clients' previous history of pain complaints and previous use of medications for pain related to chronic diseases could provide important insight to nursing staff. Caregivers who provide basic care activities may have observations of discomfort. Diagnoses listed in medical records are an important source of information in the assessment of pain for cognitively impaired older adults (Harkins, 1988; Witte, 1989). Some common pain-related diagnoses in the elderly are arthritis, osteoporosis, hip fractures, and cancer (Ferrell, 1993; Guccione, Meenan, & Anderson, 1989; Harkins, 1988; Melding, 1991; Witte, 1989).
The Model of Multifaceted Pain Assessment (Warne, 1994) (Figure) was developed as a guide for the assessment of pain in cognitively impaired older adults. This model identifies cognitive losses which contribute to reduction in selfreport of pain. The clients' inability to self-report or respond to standardized tools should trigger the health care providers' use of a multifaceted pain assessment, including:
* Family member or nursing assistant (NA) reports of possible client pain.
* Medical diagnoses known to commonly cause pain.
* Pain history demonstrated by the use of analgesics.
* Aggressive behavior patterns.
A multifaceted assessment of pain can develop a picture of possible pain without relying on client report. If the cognitively impaired individuals are able to relay information regarding pain, self-report should also be measured. Complaints of pain cannot be dismissed even if clients later forget they have reported pain.
The purpose of this study was to explore pain in cognitively impaired older adults and to determine the relationship between pain and aggressive behaviors. This study was based on the following assumptions:
* Chronic diseases commonly cause pain in older adults (McCaffery & Beebe, 1989).
* People with dementia can experience pain (Parke, 1992).
* People with severe dementia are often unable to self-report their pain.
* People with dementia who are suspected to have pain may benefit from pharmacological and nonpharmacological approaches to promote comfort and improve quality of life (Portenoy & Farkash, 1988).
This study was part of a larger project funded by the National Institute of Nursing Research to determine whether an intervention that provided dementia education for staff reduced aggressive behaviors of cognitively impaired residents. Baseline data collected from subjects at one site were used for this study to explore pain as a potential etiological factor for aggressive behaviors.
The specific research questions were:
* What proportion of cognitively impaired residents have pain reported by a family member and/or a NA?
* What proportion of cognitively impaired residents have medical diagnoses commonly known to cause pain?
* What proportion of residents with pain-causing medical diagnoses are treated with analgesics?
* Are aggressive behaviors more prevalent in cognitively impaired residents with medical diagnoses known to cause pain than in those without such diagnoses?
Setting and Sample
Three nonprofit metropolitan nursing homes were sites for the larger study of aggressive behaviors. Baseline data obtained from subjects at one of the nursing home sites were used for this descriptive study. The following eligibility criteria were used in the original study:
* Subjects must demonstrate one or more forms of aggression on a daily basis or several forms of aggression on a weekly basis.
* Subjects must have a score of 23 or less on the MMSE (Folstein, Folstein, & McHugh, 1975).
* Staff must expect subjects to have a long-term stay of at least 1 year.
The 38 subjects in this study had a mean age of 86.9. AU were White; 81.6% were women. The median length of stay was 53 months. Subjects had an average of 11 years of education. More than half of subjects were widowed; one third were currently married; and 18% had never been married. The mean case mix, a measure of dependency in activities of daily living (ADLs) and need for special nursing care, was 8 (H) on a scale of 1 to 11 (A to K). Higher case mix scores indicate greater dependency.
PERCENT OF SUBJECTS BY TYPE AND NUMBER OF PAIN-RELATED DIAGNOSES
All subjects had neuropsychiatrie diagnoses; 84% were diagnosed as having either dementia or Alzheimer's disease. The mean MMSE score was 6.4 (SD = 6.9), indicating severe cognitive impairment.
The MMSE, a 30-item widely used tool with established reliability and validity, was used to measure cognitive function (Folstein et al., 1975). Families of residents were asked, "Do you think R (residents name) experiences pain from some physical condition?" Nursing assistants were asked, "Do you think R experiences pain from some physical condition?" and "What tells you R has pain?" A Medical Records Data Form was designed to conduct a chart review on each subject to collect demographic information, functional status, diagnoses, and written documentation regarding pain identification and treatment. For purposes of this study, five diagnoses were identified as pain-related:
* History of cancer.
* History of fracture.
* Localized pain (defined as a history of headaches, knee pain, abdominal pain, or pain from contractures).
The Ryden Aggression Scale, a tool which identifies 26 aggressive behaviors, was used to document physically aggressive behavior (PAB), verbally aggressive behavior (VAB), and sexually aggressive behavior (SAB) of subjects (Ryden, 1988). The Ryden Aggression Scale, Form 1 (RASI), a retrospective measure, was used in subject selection. The Ryden Aggression Scale, Form 2 (RAS2), a concurrent measure of aggressive behaviors (Ryden, Bossenmaier, & McLachlan, 1991), was used to document current frequency of aggressive behaviors. In this study, the Cronbach's alpha for the RAS2 was .86.
ANALGESICS RECEIVED BY SUBJECTS WHO WERE REPORTED TO HAVE PAIN OR WHO HAD PAIN DIAGNOSES
AGGRESSION SCORES FOR SUBJECTS: COMPARISON OF THOSE WITH AND WITHOUT PAIN-RELATED DIAGNOSES
Staff recommended residents for the study based on consistency of aggressive behaviors, rating them with the RASI. Residents whom researchers determined met the aggression criterion were administered the MMSE to further evaluate their eligibility. Agreement to participate was obtained from residents' families.
Families and NAs were interviewed by researchers regarding their perception of pain in participating residents. The medical chart of each subject was reviewed to determine the presence of medical diagnoses commonly known to cause pain. Medication administration records were reviewed to determine the use of analgesics by each subject for 1 month prior to data collection. Aggressive behaviors were tallied for 15 days by NAs who had been instructed in the use of the RAS2.
According to families, 44.7% (n = 17) of subjects were identified as having pain. Families frequently stated the history of pain (e.g., "he had a herniated disk which causes back pain," "she has severe arthritis"). Nursing assistants reported pain in 65.8% (n = 25) of subjects, identified by verbal or behavioral cues (e.g., "she winces when we move her," "holds her head with her hands," "protects her arm and says 'ouch' when we get her up and dress her"). The families' and NAs' reports of pain were congruent in 47.3% of subjects.
One or more pain-causing diagnoses were found in 78.9 % of subjects (n = 30; mean = 1.59; SD = 1.2). Arthritis and history of hip fracture were the two pain-related diagnoses most frequently documented in subjects (Table 1).
Researchers determined whether residents reported to have pain had received analgesics in the previous month (Table 2). According to medical records, the analgesics most commonly prescribed for subjects in this study were acetaminophen and aspirin. Slightly more than one third of subjects believed to have pain by a family member received no medication for pain during the month prior to data collection. More than half of subjects (56%) who NAs identified as having pain had not received analgesics during the month prior to data collection.
Of the subjects who had at least one diagnosis known to cause pain, 60% had not received analgesics in the previous month. On closer examination, this low use of analgesics did not vary much by specific diagnosis. Fifty-nine percent (8 of 17) of subjects whose medical records indicated they had arthritis had received no pain medication in the previous month; 60% (3 of 5) of subjects with a history of cancer and 53.8% (7 of 13) of subjects with localized pain had received no pain medication in the previous month. Half of the 16 subjects with a history of hip fracture had received no pain medication in the previous month, and 43% (3 of 7) of subjects with a history of osteoporosis had received no pain medication in the previous month.
By examining subjects with each pain-related diagnosis separately, it was found the overall RAS2 scores were considerably higher for subjects with the pain-causing diagnosis than for subjects without that diagnosis (Table 3). The aggression scores indicate the mean number of observed aggressive behaviors per day. Residents with a history of cancer averaged more than 19 aggressive behaviors per day. The trend was for all aggression scores (overall and subscale scores) to be higher in subjects with pain-related diagnoses than in those with no pain diagnoses, with the exception of the VAB score in subjects with a history of hip fractures. Student t tests revealed people with arthritis had significantly higher overall mean aggression scores (t = -1.82; ? = .04) and mean PAB scores (t = -1.7; ? = .049). Initially, it appeared the overall aggression scores and the mean PAB score in people with cancer were significantly higher than in those without cancer; however, when using a test for unequal groups, the differences did not reach significance.
Subjects were then divided into two groups: those who had two or more of the pain-related diagnoses and those who had one or no painrelated diagnosis (Table 3). The subjects with two or more pain-related diagnoses had a significantly higher overall mean aggression score (t - - 1.93; p - .031) and a significantly higher mean PAB score (t = -2.43; P = -oi).
Pain in cognitively impaired nursing home residents was recognized and reported by some family members based on their knowledge of residents' previous history of pain and observations during current visits. Family advocates who observe signs of pain and report this to nurses appear to have some influence over treatment, which may account for the higher percent of subjects receiving pharmacological treatment for pain reported by family members than that reported by NAs.
An interesting finding was the higher percentage of reported pain by NAs than by families. Nursing assistants who provide physical care during ADLs may be more likely to observe signs of pain in residents, which is triggered by movement. Although NAs may recognize suspected pain through verbal statements or physical cues, pressure for them to complete ADLs may make them reluctant to interrupt their routines to report possible pain to nurses qualified to assess and treat pain. This may explain the lower percentage of pain treatment of subjects with pain reported by NAs. Nurses who assess residents for pain at a time when residents are more comfortably situated may not recognize the extent to which pain is a problem during physical activity. Memory loss in people with cognitive impairment may prevent them from remembering they had pain and what caused their pain.
Most subjects had one or more pain-related diagnoses; half had more than one. Yet, more than one third of subjects who had such diagnoses or who were reported to have pain by families or staff had received no pain medication in the previous month. The use of analgesics for pain may not have been appropriate for all subjects, and this study did not explore all possible treatments for pain subjects may have received. However, there clearly was inadequate pain treatment for some. Charts of five subjects enrolled in the aggression study (some at other sites) who had sustained hip fractures during the course of the study were reviewed. Of these five subjects, three received no pain medication on return to the nursing home (approximately 3 to 6 days, postoperatively), and two received two doses of 325 mg of acetaminophen with 30 mg of codeine within the first 36 hours at the nursing home but did not receive any analgesics during the following 2 weeks.
Aggressive behaviors were significantly more frequent in subjects with two or more pain-related diagnoses. Subjects with arthritis had significantly higher aggression scores than those without arthritis. Arthritis is the second most common diagnosis in nursing home residents (Guccione et al., 1989). Stiffness and decreased joint mobility may lead to pain associated with movement. Activities involving movement were a key trigger in stimulating aggressive responses in a study by Ryden, Bossenmaier, and McLachlan (1991). Aggressive behaviors may be a protective response by subjects who are not able to articulate their pain. They may also be indicators of ongoing chronic stress related to pain.
Almost 60% of residents with arthritis had received no pain medication in the previous month. The difference in the daily number of aggressive behaviors observed in those with arthritis versus those without arthritis has real clinical meaning (14 versus 8 aggressive behaviors) as well as statistical significance. This finding, plus the extent of aggression demonstrated in the small subgroup with a history of cancer (an average of more than 19 behaviors per day), supports the need for further study of the relationship between pain and aggressive behaviors in larger samples.
A multifaceted assessment of pain is critical if pain in cognitively impaired older adults is to be recognized and treated. Nurses who have knowledge of pain history, who elicit reports of pain from families and caregivers, who recognize painrelated medical diagnoses, and who appreciate pain may be a trigger for aggressive behaviors can more adequately detect pain in vulnerable residents. If pain is suspected, a trial of analgesic treatment may be warranted. Careful observation to determine if there is a change in behavior with treatment may validate pain was indeed a problem. Adequate pain treatment surely is a factor in the quality of life of any individual. If reduction in pain also reduces aggressive behaviors, the benefit extends to those affected by the behavior of aggressive residents - peers, families, and staff.
This is a secondary analysis of data from a previous study with a primary focus on aggressive behaviors in cognitively impaired institutionalized older adults. It included people with moderate to severe cognitive impairment who had consistent aggressive behaviors; therefore, the sample was not representative of those people with cognitive impairment who were not aggressive or who were community based. No pain assessment tool was included in the data collection process, therefore eliminating a specific pain assessment for each subject. Medical record forms at the facilities were inconsistent in documentation of pain. Administered analgesics were documented on medication records; however, nursing notes frequently lacked assessment or description of the pain for which the analgesic was given. The small sample size limits generalizability of findings to the broader population of cognitively impaired older adults.
IMPLICATIONS FOR PRACTICE
The findings of this study suggest the majority of cognitively impaired nursing home residents have one or more diagnoses for which pain is a frequent symptom. Though their pain is sometimes recognized by families or staff, the adequacy of treatment for some is questionable. Nursing staff need to consider all possible sources of information when assessing pain in cognitively impaired older adults. Nursing assistants who are made aware of signs of pain in this population are important reporters of pain-related behaviors. Pain appears to contribute to aggressive behaviors. Nurses need to consider the possibility of pain when exploring strategies for reducing aggressive behaviors. Studies of larger samples of cognitively impaired older adults are needed to provide knowledge needed to improve the assessment and treatment of pain in this population.
- Boettcher, E.G. (1983). Preventing violent behavior An integrated dieoretical model of nursing. Perspectives in Psychiatric Care, 21(2), 54-58.
- Cohen-Mansfield, J., Billig, N., Lipson, S., Rosenthal, A.S., & Pawlson, L.G. (1990). Medical correlates of agitation in nursing home residents. Gerontology, 36, 150-158.
- Feldt, K.S. (1996). Treatment of pain in cognitively impaired versus cognitively intact post hip fractured elders. Dissertation Abstracts International, 57-09B, 5574.
- Feldt, K.S., Ryden, M.B. & Miles, S. (1998). Treatment of pain in cognitively impaired compared with cognitively intact older patients with hip fractures. Journal of the American Geriatrics Society, 46, 10791085.
- Ferrell, B.A. (1993). Cancer and the older adult. Elderly Care, 5(3), 29-30.
- Ferrell, B.A., Ferrell, B.R., & Osterweil, D. (1990). Pain in the nursing home. Journal of the American Geriatncs Society, 38, 409-414.
- Ferrell, B.A., Ferrell, B.R., & Rivera, L. (1995). Pain in cognitively impaired nursing home patients. Journal of Pain and Symptom Management, 10(8), 591-598.
- Folstein, M.F., Folstein, S.E., & McHugh, PR. (1975). Mini-mental state: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189-198.
- Guccione, A.A., Meenan, R.F., & Anderson, J.J. (1989). Arthritis in nursing home residents. Arthritis and Rheumatism, 32(12), 1546-1553.
- Harkins, S.W. (1988). Pain in the elderly. In R. Dubner, G.F. Gebhart, & M.R. Bond (Eds.), Proceedings of the fifth world congress on pain (pp. 355-367). Amsterdam: Elsevier Science.
- Herr, K.A., & Mobily, RR. (1991). Pain assessment in the elderly. Journal of Gerontological Nursing, /7(4), 12-19.
- Herr, K.A., & Mobily, PR. (1993). Comparison of selected pain assessment tools for use with the elderly. Applied Nursing Research, 6, 39-46.
- Hurley, A.C., Volker, BJ., Hanrahan, P.A., Houde, S., & Volker, L. (1992). Assessment of discomfort in advanced Alzheimer's patients. Research in Nursing & Health, 15(5), 369-377.
- Marzinski, L.R. (1991). The tragedy of dementia: Clinically assessing pain in the confused, nonverbal elderly. Journal of Gerontological Nursing, 17(6), 25-28.
- McCaffery, M., & Beebe, A. (1989). Pain in the elderly. In M. McCaffery & A. Beebe (Eds.), Pain: Clinical manual for nursing. St. Louis: Mosby.
- Melding, RS. (1991). Is there such a thing as geriatric pain? Pain, 46, 119-121.
- Parke, B. (1992). Pain in the cognitively impaired elderly. The Canadian Nurse, 88(7), 17-20.
- Parmelee, P.A., Smith, B., & Katz, LR. (1993). Pain complaints and cognitive status among elderly institution residents. Journal of the American Geriatrics Society, 41, 517-522.
- Portenoy, R.K., & Farkash, A. (1988). Practical management of non-malignant pain in the elderly. Geriatrics, 43(5), 29-47.
- Roberts, H.C., & Eastwood, J. (1994). Pain and its control in patients with fractures of the femoral neck while awaiting surgery. Injury, 2$, 237-239.
- Roy, R., & Thomas, M.R. (1987). Elderly persons with and without pain: A comparative study. The Clinical Journal of Pain, 3, 102-106.
- Ryden, M.B. (1988). Aggressive behavior in persons with dementia who Uve in the community. Alzheimer's Disease and Related Disorders, 2, 342-355.
- Ryden, M.B., Bossenmaier, M., & McLachlan, C. (1991). Aggressive behavior in cognitively impaired nursing home residents. Research in Nursing & Health, 14, 87-95.
- Ryden, M.B., & Feldt, K.S. (1992). Goal directed care: Caring for aggressive nursing home residents with dementia. Journal of Gerontological Nursing, /S(Il), 35-42.
- Sengstaken, E. ?., & King, S.A. (1993). The problems of pain and its detection among geriatric nursing home residents. Journal of the American Geriatrics Society, 41, 541-44.
- Warne, M. (1994). The experience of pain in persons with cognitive impairments. Unpublished manuscript, University of Minnesota, Minneapolis.
- Witte, M. (1989). Pain control. Journal of Gerontological Nursing, 15(3), 32-37.
PERCENT OF SUBJECTS BY TYPE AND NUMBER OF PAIN-RELATED DIAGNOSES
ANALGESICS RECEIVED BY SUBJECTS WHO WERE REPORTED TO HAVE PAIN OR WHO HAD PAIN DIAGNOSES
AGGRESSION SCORES FOR SUBJECTS: COMPARISON OF THOSE WITH AND WITHOUT PAIN-RELATED DIAGNOSES