Physical pain is a significant problem for many older adults. It has been estimated that between 25% and 86% of community-dwelling elderly individuals suffer from significant pain problems (Brattberg, Thorshmd, & Wikman, 1989; Chrischilles et al., 1992; Mobily, Herr, Clark, & Wallace, 1994; Thomas & Roy, 1988). In nursing homes, the prevalence of pain is also high, with as many as 83% of nursing home residents reportedly having pain (Ferrei!, Ferrell, & Osterweil, 1990; Ferrei!, Ferrell, Sc Rivera, 1995; Parmelee, Smith, & Katz, 1993, Sengstaken & King, 1993).
The high prevalence of pain in advanced age is primarily related to chronic health disorders, particularly painful musculoskeletal conditions such as arthritis and osteoporosîs (Wallace, 1994). In addition, there is an increased incidence of cancer, the need for surgical procedures, pressure ulcers, and cardiovascular disease in this age group (Ferrell, 1991; Parmelee, 1994). Cancer, in particular, is associated with significant pain for one third of patients with active disease and for two thirds of those with advanced disease (Ferrell & Ferrell, 1990).
Complications of unrelieved pain are widespread and varied. Physical, functional, and mental health conditions are associated with or exacerbated by pain (Won et al., 1999). Depression, withdrawal, sleep disturbances, impaired mobility, decreased activity involvement, and increased health care utilization have been associated with physical pain (Ferrell, 1995; Ferrell & Ferrell, 1990; Herr Oc Mobily, 1991; Kahana, Kahana, Namazi, Kercher, & Strange, 1997; Parmeíee, Katz, & Lawton, 1991; Lavsky-Shulan et al., 1985). Other geriatric conditions that can be worsened by pain include falls, deconditioning, malnutrition, gait disturbances, and slow rehabilitation (Ferrell, Ferrell, & Rivera, 1995). Thus, pain has major implications for the quality of life and quality of care for elderly individuals (Ferrell & Ferrell, 1990).
Despite the prevalence and consequences of pain, evidence suggests pain is underdetected and poorly managed among older adults. According to Cleeland (1987), there are a number of factors contributing to this situation, including individual-based and caregiver-based factors. Individualbased factors in older adults that may impair accurate pain assessment include (American Geriatrics Society, 1998; Forrest, 1995; Fulmer, Mion, & Bottrell, 1996; Wells, Kaas, & Feldt, 1997):
* A belief that pain is a normal part of aging.
* Concern of being labeled a hypochondriac.
* Fear of the prognostic consequences of pain.
* Fear of narcotic addiction.
* Worry about health care costs.
* A belief that pain is of low priority when multiple health conditions exist.
Other factors, such as hearing and speech disabilities, may prevent elderly individuals from communicating pain to caregivers (Brockopp, Warden, Colclough, & Brockopp, 1996; Feldt, Warne, & Ryden, 1998; Forrest, 1995). In addition, cognitive impairment is an important factor in diminishing elderly individuals' reporting of pain (Parmelee, Smith, & Katz, 1993; Weiner, Peterson, Logue, & Keefe, 1998). In fact, Weiner et al. (1998) found that cognitive impairment was the strongest predictor of pain self -report in nursing home residents. Lower Mini-Mental Status Exam (MMSE) scores were significantly associated with lower reported pain intensity.
Pain detection and management are also influenced by caregiver-based factors. Caregivers have been found to share the belief that pain is a pan of the normal aging process and to fear using narcotics because of potential addiction or adverse side effects (Wells, Kaas, St Feldt, 1997). Similarly, elderly individuals' cognitive status influences health care providers' assessment and treatment of pain (Kaasalainen et al., 1998). Recently, Horgas and Tsai (1998) reponed that cognitively impaired nursing home residents were prescribed and administered significantly less analgesic medication than cognitively intact older adults.
This finding may reflect impaired elderly individuals' inability to perceive, recall, or report the presence of pain to health care providers. It may also reflect caregivers* inability to detect pain among patients who are less able to verbally report its presence, as well as caregivers' lack of belief in the validity of pain reports in this subgroup of elderly individuals. In addition, there is evidence to suggest that care providers' judgments of pain may be influenced by gender expectations (Miaskowski, 1997; Vallerand, 1995; Won, et al, 1999).
In the nursing home setting, 80% to 90% of all direct caregiving is performed by nursing assistants (NAs) (Smyer, Brannon, & Cohn, 1992). Nursing assistants usually have a high school education or less, and are the least trained caregivers in the facility. The passage of the Omnibus Budget Reconciliation Act in 1987, however, mandated that NAs receive standardized training in basic skills and pass a certification examination. This training includes little or no information about pain assessment (Wells, Kaas, & Feldt, 1997). This is unfortunate because NAs have more direct contact with the residents than other health care providers, and thus have more opportunities to assess pain. This lack of training, coupled with the heavy caregiving demands and the need to complete daily routine care, may contribute to inaccurate pain assessment among NAs (Feldt, Warne, & Ryden, 1998). Thus, licensed nursing personnel, those largely responsible for pain treatment, may be receiving inaccurate or incomplete répons of residents' pain from the NAs.
Several studies have been conducted to examine the accuracy with which health care providers assess pain in various settings. Most of these studies have investigated the congruence of self -ratings and caregiver ratings of cancer, postoperative, or burn pain. For instance, Yeager, Miaskowski, Dibble & Wallhagen (1995) compared pain reports between cancer patients and their family caregivers and found that family members rated the patient's pain as significantly more intense than did the patient. Grossman et al. (1991) compared self -report and health care provider ratings of pain in a sample of 103 hospitalized patients with cancer. Significant correlations between pain intensity ratings by patients and their nurses, house officers, and oncology fellows were noted for the entire sample.
However, when only the subset of patients with intensity scores of 4 or greater on a visual analogue scale were considered, no statistically significant correlations were found between any of the paired raters. These findings demonstrate discrepancies between health care providers' and patients' perceptions of pain may be more pronounced in those patients with significant pain. Further, Camp (1988) reported that cancer patients* and nurses' descriptions of pain agreed in only 14% of cases.
With regard to postoperative pain, ratings among patients and nurses were found to be congruent in only 35% of cases, with nurses underestimating pain in 45% and overestimating pain in 20% of cases (Zalon, 1993). Similar findings were reported in a study of burn patients. Choiniere, Melzack, Girard, Rondeau, and Paquin (1990) compared the pain ratings of 42 hospitalized burn patients and 42 nurses. This study found that nurses' and patients' ratings of pain intensity were significantly, but moderately, correlated (r = .47), but that pain was underestimated in 43% and overestimated in 27% of cases. Interestingly, these authors found that the amount of nursing experience was significantly associated with under- and overestimation of pain: Less experienced nurses were more likely to overestimate pain intensity and more experienced nurses were more likely to underestimate pain intensity among burn patients.
Teske, Daut, and Cleeland (1983) reported that congruence between nurses' and patients' reports of pain were higher in the case of acute pain than chronic pain. Finally, HallLord, Larsson, and Steen (1998) compared pain ratings among patients, nurses, and assistant nurses in a sample of elderly patients in an intensive care unit in Sweden. Significant discrepancies were found among all three groups. Nurses were more likely to overestimate pain and assistant nurses (equivalent to NAs in the United States) were more likely to underestimate pain relative to the patients' experience.
Until recently, very few studies have examined the congruence of pain reports in geriatric patients or in long-term care settings. In a sample of 200 elderly participants attending an adult day care program, Werner, Cohen-Mansfield, Watson, and Pasis (1998) compared pain ratings among elderly individuals, family caregivers, and day care staff. They concluded that elderly individuals and staff caregivers agreed on the presence of pain in 67.7% of cases overall, but that agreement differed according to the elderly individual's cognitive status. Agreement was lower among seniors with MMSE scores lower than 24 (67.1% agreement) versus those with MMSE scores 24 or greater (68.6% agreement).
In the nursing home setting, Kaasalainen et al. (1998) studied the relationship between nurses* ratings of residents* pain and the amount of analgesics administered. The results indicated there was no significant correlation between nurses1 perceptions of pain and their actions with regard to pain medications. In another study, Feldt, Warne, and Ryden (1998) examined agreement of pain reports in three nursing homes by asking NAs and family members to rate the intensity of residents' pain. These authors concluded that NAs and family members were congruent in approximately 47% of cases when rating an individual's pain. These studies, however, did not directly investigate whether caregivers' perceptions of pain were congruent with the residents* perceptions of their own pain.
Individuals perceptions of their health and functioning are a key feature of comprehensive gerontological assessment, providing important information for making accurate diagnoses and realistic treatment goals (Stone, Wyman, & Salisbury, 1999). This is particularly true in the case of pain, which is a highly subjective experience (Melzack & Casey, 1968). In fact, pain is defined as "whatever the person says it is, existing whenever the experiencing person says it does" (Pasero, Paice, & McCaffery, 1999, p. 17). As such, residents' selfreports of pain should serve as the reference standard by which other caregivers' ratings are compared.
Thus, the purpose of this study is to examine the relationship between NAs* and nursing home residents' perceptions of pain. Specifically, the following questions are addressed:
* What are the prevalence, location, and intensity of self-rated and NA-rated pain?
* What is the congruence between residents' and NAs' ratings of pain prevalence, location, and intensity?
* What are the resident-based correlates of congruent or incongruent pain ratings?
* What caregiver-based factors are associated with congruent or incongruent pain ratings?
The participants in this study were 45 resident-NA dyads from one private, for-profit nursing home located in a large metropolitan area in the Midwest. Residents in the sample had a mean age of 82.5 years (range = 58 to 98 years, SD = 9.7). The majority of participants (n = 35, 75.6%) were women and widowed (« = 26, 65%). Approximately one third (32%) of residents had a medical diagnosis of Alzheimer's disease or dementia.
Sixteen NAs participated in the study, with each NA providing data on more than one resident in the facility. All of the NAs were women and were, on average, 36.5 years of age (range = 23 to 53 years, SD = 8.1 years). The NAs had been employed in the facility for a mean of 3.6 years (range = 4 months to 1 7.8 years, SD = 4.4 years) and had worked as NAs for an average of 8.2 years (range = 2 to 17.8 years, SD = 4.9 years). With regard to educational level, approximately one third of NAs reported they had a high school education (n - 6, 37.5%) and the remaining two thirds reported they had some college or were college graduates (n = 10, 62.5%).
All residents in the nursing home were invited to participate in this study through flyers, letters, and personal contact. Of the 145 residents residing in the facility, 45 were both willing and able to participate in an individual interview. Residents were interviewed in thenprivate rooms by trained research staff. Informed consent was obtained from the residents.
All NAs employed in the facility on the day or evening shift were recruited to participate in the study via individual letters and follow-up personal contact by the researchers. Of the 58 NAs who were eligible to participate, 16 (27.6%) agreed to complete the pain-rating questionnaire. Nursing assistants were asked to complete one rating form for residents for whom they had provided direct care during the previous week. Each NA completed the questionnaire for between one and five residents, but each resident was rated by only one NA. Because the focus of this study was on resident-NA dyads, the final sample consisted of 45 pairs with complete data on both NA-ratings and self -ratings of pain.
Residents' self-rated pain. In an interview format, residents were asked about the presence, location, and severity of pain in 12 body sites. Specifically, residents were asked the following question: "During the last week, did you have any pain?" If the participants responded yes, they were probed for whether or not they experienced pain in specific body sites, as well as the intensity of the pain in each location. Intensity was assessed using a verbal descriptor scale where O was no pain, 5 was moderate pain, and 10 was the worst pain possible. The 12 locations assessed included head, neck, stomach, back, shoulder, elbow, wrist, hand, hip, knee, ankle, and feet. A sum score was computed to represent the number of painful sites reported. In addition, a mean pain intensity score was computed across all 12 body sites.
Nursing assistant ratings of residents' pain. On the same day the residents were interviewed, NAs were asked to provide data about the residents under their care. Nursing assistants responded to the question, "During the last week, did this resident, (resident's name), have any pain?" If the NAs responded yes, they were also probed to identify the location and intensity of the resident's pain in the same 12 body sites described above. Pain intensity was rated on a verbal descriptor scale where O was no pain, 5 was moderate pain, and 10 was the worst pain possible. A sum score was computed to represent the number of painful sites reported and a mean pain intensity score was computed across all 12 body sites.
RESIDENT RATINGS OF PAIN PREVALENCE AND INTENSITY IN SUBSET OF RESIDENTS REPORTING PAIN (n = 22)
Resident data. The Minimum Data Set (MDS) Version 2 was used to obtain demographic and diagnostic data related to the residents in this study. This federally mandated, standardized instrument provides data on all nursing home residents and was intended for clinical, quality assurance, and research purposes (Hawes, et al., 1995). Specially trained staff, usually RNs who are unit managers, completes the MDS assessment process. Research personnel obtained data about residents' age, sex, marital status, and medical diagnoses from this chart data. Minimum Data Set diagnoses of Alzheimery disease (code = Ilq) or dementia (code = Uu) were coded as cognitive impairment. MDS diagnoses of arthritis (code = III), hip fracture (code = Um), osteoporosis (code = Ilo), pathological bone fracture (code = Up) or cancer (code = Ilpp) were coded as painful diagnoses (Feldt et al., 1 998; Horgas 6t Tsai, 1998). These specific diagnoses were summed to create an indicator of the total number of painful diagnoses present.
Nursing assistant data. A questionnaire was administered to the NAs to obtain data related to their age, education, and gender. In addition, open-ended questions were used to obtain data about caregivers' work history as NAs and how long they had been employed in the study facility.
Geriatric Depression Scale. The Geriatric Depression Scale (GDS) - Short Form (Brink, Yesavage, Lum, Heersema, Adey, Oc Rose, 1982) was used to assess self-reported depression among the nursing home residents. The GDS consists of 15 yes/no ítems and is considered the most reliable measure of depression among elderly individuals with dementia (Parmelee, Lawton, & Katz, 1989). Higher scores indicate higher levels of depressive symptomology, with a score of 5 representing the cut-off for depression.
NA RATINGS OF PAIN PREVALENCE AND INTENSITY IN SUBSET OF RESIDENTS PERCEIVED TO H AVE PAIN(n = 16)
Philadelphia Geriatric Center (PGC) Well-Being Scale. The modified PGC Morale Scale Lawton, 1975} was used to assess subjective well-being among the nursing home residents. This scale consists of 15 items (McCulloch, 1991) and asks participants to indicate the extent of their agreement with the statements on a 5-point scale (1 = strongly agree, 5 = strongly disagree). Higher scores indicate more positive levels of well-being.
Residents' self-rated pain prevalence, location, and intensity. The results indicated that 48.9% (n = 22) of the residents reported they experienced pain within the week prior to interview. Of those participants reporting pain (n = 22), a mean of 3.4 painful body locations was reported (range = 1 to 10, 50 = 2.8). The most frequently reported pain sites were the back (50%) and knees (40.9%). In those residents with pain, the mean pain intensity rating across the 12 body locations was 1.5 on a 1 to 10 scale (range = O to 5.3, SD = 1 .4), indicating mild pain. Participants' rated back pain as the most intense (mean pain rating = 2.9, 50 = 3.5, range = O to 10), followed by stomach pain (mean pain rating = 2.0, SD = 3.2, range = O to 10). See Table 1 for a summary of the prevalence of pain in the 12 locations and the average pain intensity ratings for those residents reporting pain.
CONGRUENCE BETWEEN NA AND RESIDENT RATINGS OF PAIN PREVALENCE
RESIDENT-BASED DEMOGRAPHIC, DIAGNOSTIC, AND AFFECTIVE CORRELATES OF PAIN CONGRUENCE AND INCONGRUENCE
Nursing assistants' ratings of pain prevalence, location, and intensity. Nursing assistants reported 35.6% (n = 16) of residents experienced pain within the last week. In those residents that were perceived to have pain (n = 16), NAs reported an average of 6.3 painful body locations (range = O to 12, 5D = 4.0). Nursing assistants identified pain most frequently in residents' back (75% of residents), feet (68.8%), shoulders (62.5%) and knees (62.5%). Averaged across all 12 locations, NAs described residents* pain intensity as 1.9 on a O-to-10 scale (range = 0.25 to 10, SD = 2.3). Back pain was rated as the most intense (mean pain intensity rating = 3.4, SD = 3.5, range = O to 10), followed by foot pain (mean pain intensity rating = 2.8, SD = 3.5, range = O to 10). Table 2 summarizes NA reports of the prevalence of pain in the 12 body locations and the mean pain intensity ratings for each site for those residents perceived as having pain.
Congruence between resident and NA ratings of pain. There was no significant association between resident self-rating of pain and NA ratings of resident pain (?2=3.1, off = 1, p = .08). The kappa statistic, which assesses the percent agreement corrected for chance, showed low congruence between the different raters [kappa = -0.25]. As shown in Table 3, when asked whether pain was experienced during the past week, resident and NA responses were congruent in 17 of 45 cases (37.7% of cases). In 12 cases, residents and NAs agreed no pain was experienced and in 5 cases (11.1%), they agreed the resident experienced pain. In contrast, resident and NA ratings were incongruent in 62.2% of the cases. Pain was underdetected in 17 cases (37.8%); that is, residents stated they experienced pain, but NAs stated that the residents did not. In addition, pain was overreported in 1 1 cases (24.4%) because residents reported no pain, but NAs stated residents did report pain. Table 3 summarizes the proportion of dyadic ratings that were congruent, overreported, and underdetected.
In the 12 body locations assessed, there was no significant association between residents' selfrating and NAs' ratings of resident pain. The number of pain sites reponed by NAs and residents was not statistically significant (r = -0.16, /> = .29). A series of 12 chi-square analyses were conducted for each of the 12 locations and none reached statistical significance. Kappa statistics ranged from -.19 for stomach pain to 0.004 for arm pain, indicating very low agreement.
With regard to pain intensity, no significant correlations were found between the two raters' perceptions of pain. Correlation coefficients were conducted for overall mean level of pain (r = -Q.17,p - .28) and for pain intensity in each of the 12 sites (r ranged from -0.04 for elbow pain to -0.13 for stomach pain). None reached statistical significance, suggesting very low agreement between residents' and NAs* perceptions of pain intensity.
Relationships between residents' demographic, diagnostic, and affect characteristics and pain congruence, underdetection, and overreporting. The question remains about what differentiates residents for whom pain is accurately detected from those for whom pain ratings did not correspond (e.g., either overreported or underdetected). For the purpose of these analyses, agreement that pain was or was not present was considered congruent. These congruent responses were compared to both the overreported and underdetected responses. A number of resident-based characteristics were considered, including resident demographic characteristics (age, sex), diagnoses (dementia and painful conditions), and affective factors (e.g., depression, wellbeing). Table 4 summarizes the results of these analyses.
No significant differences were found for any of the demographic factors or diagnoses. That is, whether pain was reported accurately, was underdetected, or was overreported by the NAs was not related to residents' age, sex, diagnosed cognitive impairment, or the number of diagnosed painful conditions.
In contrast, the self-rated affective dimensions (e.g., well-being, depression) did differ significantly in the three congruence outcome groups. Level of well-being was significantly lower among residents for whom pain was underdetected than in those residents who either had congruent pain perceptions or whose pain was overreported (F [2, 42] = 5.70, p = .01). Well-being was highest in residents who denied pain, but who were perceived as having pain by their caregivers.
Self-rated depression was also significantly associated with the congruence outcomes (F [2, 40] = 3.03, p = .05). Residents whose pain was underdetected had significantly higher levels of depression than did those whose pain was overreported. Level of depression was lowest in the overreported group; that is, among residents who denied pain but in whom caregivers perceived pain.
Relationships between NA characteristics and pain congruence, underdetection, and overreporting. The nature of the sample (e.g., 16 NAs rated 45 residents) precluded direct statistical analysis of the relationship between NA characteristics and the three congruence outcomes because of the non-independence of observations. Thus, to approximate this analysis, the percentage of congruent and non-congruent pain reports per NA were analyzed. Three scores were computed: percentage of ratings that were congruent, percentage that were overdetected, and percentage that were underdetected. For instance, one NA rated pain among four residents. One rating was congruent with the resident's self-report (25%), 2 ratings were overreported (50%), and 1 rating was underdetected (25%).
These computed percentage scores were correlated with NA age, education, and work experience. The results indicated no significant relationship between the age of the NA; the duration of employment in the facility; the duration of experience as an NA; or educational level and the percentage of congruent, overdetected, or underreported pain reports.
The results of this study illustrate the complexities of assessing pain in nursing home residents. These difficulties are manifest in the lack of agreement between elderly nursing home residents and their paid caregivers in relation to the presence of pain. Nursing assistants, the primary caregivers in the nursing home setting, failed to report pain in approximately two thirds of the residents who stated they experienced it. Further, there was very little agreement between residents and NAs related to location or intensity of pain.
The lack of congruence in pain reports found in this study is interesting to note. Based on prior research in different settings and with different caregiver/care recipient dyads, the lack of agreement on the presence of pain and the proportion of pain that is underdetected is not surprising. The overreporting of pain, however, raises some questions. In approximately 25% of participants, NAs reported the resident had experienced pain when the resident denied it. It is difficult to speculate what might be contributing to this phenomenon. If one adopts the McCaffery and Beebe (1989) definition, residents* report of pain holds the highest weight. This finding suggests an overattribution of pain by the NA. Nursing assistants may be observing behaviors they are ascribing to pain (e.g., disruptive behaviors [Feldt et al., 1998], crying) when these behaviors actually reflect another emotion or phenomenon such as sadness or fear.
In those residents for whom pain was reported, NAs perceived a slightly higher intensity of pain and more painful sites than reported by the residents themselves. This trend, however, could not be tested statistically because of the different sample sizes for the subsamples reporting pain (n = 22) or rated as having pain by the NAs (« = 16). This trend, if empirically validated, would appear to be consistent with other work addressing this question (Werner et al., 1998; Yeager et al., 1995). Yeager et al. (1995) reported that family caregivers perceived cancer pain as more intense than did the patients. Similarly, Werner et al. (1998) found that both NAs and family caregivers of older adult day care participants rated pain intensity higher than did the elderly individuals themselves. These findings may reflect the close nature of the relationships between caregivers and residents, and the extent to which seeing someone suffer is disturbing. This finding may, in fact, be an endorsement of the nature of the caregiving relationship between NAs and the residents for whom they provide care, and may reflect the compassion and caring one would hope would be present in this caregiving context. Nonetheless, it suggests pain assessment by these care providers is inaccurate. This may lead to inappropriate pain treatment by the licensed personnel.
The question remains: What is contributing to the discrepant pain reports found in this study? Few of the resident-based factors assessed in this study were related to the congruent or incongruent reporting of pain. In contrast to previous studies, age and sex were unrelated to the accuracy of pain reports (Miaskowski, 1 997; Vallerand, 1995). Further, diagnosed cognitive impairment and the number of diagnosed painful conditions were unrelated to whether pain was congruently assessed, overreported, or underdetected. This is a somewhat puzzling finding given the preponderance of evidence suggesting cognitiveìy impaired older adults are less likely to report pain (Parmelee, 1997), to be treated for pain (Horgas Sc Tsai, 1998), and to be consistent in their pain reports (Weiner et al., 1998). As such, one would expect a higher proportion of participants in the incongruent pain groups to be diagnosed with cognitive impairment. The failure to detect this relationship may reflect unreliability of the medical diagnoses (Spore, Horgas, Smyer, & Marks, 1992) and a lack of data related to the severity of cognitive impairment. Weiner et al. (1998) found that level of cognitive impairment was an important factor, such that higher MMSE scores were related to reports of higher pain intensity and more stability in pain reports.
In the present study, cognitive tests were not administered to the study participants; thus, information related to the level of impairment is not available. Anecdotally, however, participants in this study were able to respond to interview questions about their pain, and were able to provide informed consent. The ability to provide consent was ascertained from facility administrators familiar with the resident's status, and the ability of residents to briefly summarize the intent of the study after it was described by the researchers (Raway, 1 994). Future research, however, is warranted that directly assesses cognition with psychometrically validated neuropsychological tools to investigate the cognitive status and pain relationship more thoroughly and its influence on caregiver assessments of pain.
Residents' affective states, in particular depression and perceived well-being, did differ significantly across the congruence groups. The relationship between pain and depression has consistently been reported (Parmelee, Katz, & Lawton, 1991; Williamson & Schulz, 1992). In the present study, residents with underdetected pain had significantly higher levels of depression and lower levels of well-being. These findings may illustrate the consequences of unacknowledged pain for residents* mental health. Given the cross-sectional nature of this study, however, it is not possible to determine whether these affective states are a cause or a consequence of underdetected pain.
In contrast, residents who denied pain, but were perceived as having pain by the NAs had the highest levels of well-being and the lowest levels of depression. This may reflect methodological issues about assessing pain and, in particular, the possible confounding of pain history and current pain. It is possible that NAs were influenced by residents* prior pain complaints, injury, or illness and believed pain persisted after it had ended. According to the behavioral model of pain (Fordyce, 1976), NAs may have focused additional attention, assistance, or support toward these residents, resulting in more positive affect. This finding does not fully support the behavioral model because it hypothesizes that individuals would be more likely to report pain to solicit these responses from others. Additional research is needed to more fully explore this subgroup of residents for whom pain is overreported by caregivers.
There are several limitations to this study that must be acknowledged. First, the sample was small and selected from one nursing home. Thus, the extent to which these findings can be generalized to a larger or more diverse population of nursing home residents is not clear. Second, the assessment of cognitive status relied on charted, medical diagnoses instead of independent neuropsychological assessments. Thus, cognitive status may have not been adequately measured. Third, the data were collected such that a single NA rated the pain of multiple residents, resulting in nonindependence of observations for multi-level data. Thus, the types of statistical analyses that could be performed to analyze the influence of NA characteristics on pain congruence ratings was limited and may not adequately address this important question. Future research would avoid this methodological limitation by having one NA rate a single resident.
Despite these limitations, the results of this study highlight an important issue: the incongruence between self-reported pain and NAs' reports of pain. Because NAs provide most of the direct caregiving in nursing home settings, these findings may have serious implications for the adequacy and accuracy of pain management in this setting. If NAs do not accurately detect pain in the residents for whom they provide direct care, they are unable to convey appropriate and correct information to the licensed personnel who administer analgesics or non-pharmacological pain treatments. Thus, a large proportion of nursing home residents may be at risk for inappropriate treatment of pain - either under- or overtreatment.
These findings highlight the need to include NAs in in-service education for assessing pain, increasing their knowledge and skill in accurate assessment strategies. Further, these results illustrate the broader need for increased attention to pain management in nursing homes. This is especially true in light of the importance older adults place on pain management, rating it as one of the top priorities for the nursing profession (Hudson & Sexton, 1996). As it stands, many elderly individuals may not only be disappointed in the care that they receive in this domain, but may be needlessly suffering.
- American Geriatrics Society (1998). The management of chronic pain in older persons. Journal of the American Geriatrics Society, 46,635-651.
- Brattberg, G., Thorslund, M-, & Wikman, A. (1989). The prevalence of pain in a general population: The results of a postal survey in a county of Sweden. Pain, 37(2), 215-222.
- Brink, T.L., Yesavage, J.A., Lum, O., Heersema, P.H., Adey, M-, & Rose, T.L. (1982). Screening tests for geriatric depression. Clinical Gerontologist, 1, 37-43.
- Brockopp, D., Warden, S., Colclough, G., & Brockopp, G. (1996). Elderly people's knowledge of and attitudes to pain management. British Journal of Nursing, 5(9), 556-558, 560-562.
- Camp, L.D. (1988). A comparison of nurses' recorded assessments of pain with perceptions of pain as described by cancer patients. Cancer Nursing, 11(4), 237-243.
- Choiniere, M., Melzack, R., Girard, N., Rondeau, J., & Paquin, M. (1990). Comparisons between patients' and nurses' assessment of pain and medication efficacy in severe burn injuries. Pain, 40(2), 143-152.
- Chrischilles, E.A., Foley, D.J., Wallace, R.B., Lenke, J.H., Hanlon, J.T., Glynn, R.J., Ostfeld, A.M., & Guralnik, J.M. (1992). Use of medications by persons 65 and over: Data from the established population for epidemiologie studies of the elderly. Journal of Gerontology, 47(5), M137-M144.
- Cleeland, C. (1987). Barriers to the management of cancer pain. Oncology, I (2 Supplement), 19-26.
- Feldt, K.S., Warne, M.A., & Ryden, M.B. (1998). Examining pain in aggressive cognitively impaired older adults. Journal of Gerontological Nursing, 24, 14-22.
- Ferrell, B.A. (1991). Pain management in elderly people. Journal of the American Geriatric Society, 39, 64-73.
- Ferrell, B.A. (1995). Pain evaluation and management in the nursing home. Annals of Internal Medicine, 123(9), 681-687.
- Ferrell, B.R., & Ferrell, B.A. (1990). More research needed on geriatric pain management. Provider, 7, 31-32.
- Ferrell, S.A., Fendi, B.R., & Osterweil, D. (1990). Pain in the nursing home. Journal of the American Geriatric Society, 38, 409-414.
- Ferrei!, B.A., Ferrell, B.R., Sí Rivera, L. (1995). Pain in cognitively impaired nursing home patients. Journal of Pain and Symptom Management, 70(8), 591-598.
- Fordyce, W.E. (1976). Behavioral methods for chronic pain and illness. St. Louis, MO: Mosby.
- Forrest, J. (1995). Assessment of acute and chronic pain in older adults. Journal of Gerontological Nursing, 21, 15-20.
- Fulmer, T., Mion, L.C, St Bottreü, M.M. (1996). Pain management protocol: NICHE Faculty. Geriatric Nursing, 17(5), 222-227.
- Grossman, S.A., Sheidler, V.R., Swedeen, K., Mucenski, J., & Piantadosi, S. (1991). Correlation of patient and caregiver ratings of cancer pain. Journal of Pain and Symptom Management, 6(2), 53-57.
- Hall-Lord, M.L., Larsson, G., & Steen, B. (1998). Pain and distress among elderly intensive care unit patients: Comparison of patients' experiences and nurses' assessment. Heart ana Lung, 27(2), 123-132.
- Hawes, C, Morris, J.N., Phillips, C.D., Mor, V., Fries, B.E., & Nonemaker, D. (1995). Reliability estimates for the Minimum Data Set for nursing home resident assessment and care screening (MDS). The Gerontologist, 35(2), 172-178.
- Herr, K. ?., & Mobily, P.R. ( 1 99 1 ). Complexities of pain assessment in the elderly: Clinical considerations. Journal of Gerontological Nursing, 17, 12-19.
- Horgas, A.L., & Tsai, P.F. (1998). Analgesic drug prescription and use in cognitively impaired nursing home residents. Nursing Research, 47(4), 235-242.
- Hudson, K.A., & Sexton, D.L. (1996). Perceptions about nursing care: Comparing elders' and nurses' priorities. Journal of Gerontological Nursing, 22, 41-46.
- Kaasalainen, S., Middleton, J., Knezacek, S., Hartley, T., Stewart, N., Ife, C., & Robinson, L. (1998). Pain and cognitive status in the institutionalized elderly: Perceptions and interventions. Journal of Gerontological Nursing, 24(8), 24-31.
- Kahana, B., Kahana, E., Namazi, K., Kercher, K., & Strange, K. (1997). The role of pain in the cascade from chronic illness to social disability and psychological distress in late life. In D.I. Mostofsky & J. Lomranz (Eds.), Handbook of pain and aging (pp. 183-206). New York: Plenum Press.
- Lavsky-Shulan, M-, Wallace, R.B., Kohout, F.G., Lemke, J.H., Morris, M.C., & Smith, LM. (1985). Prevalence and functional correlates of low back pain in die elderly: The Iowa 65+ Rural Health Survey. Journal of the American Geriatrics Society, 33, 23-28.
- Lawton, M.P. (1975). The Philadelphia Geriatric Center Morale Scale: A revision. Journal of Gerontology, 30(1), 85-89.
- McCaffery, M., & Beebe, A. (1989), Pain in the elderly. In M. McCaffery & A. Beebe (Eds.), Paat: Clinical manual for nursing. St. Louis, MO: Mosby.
- McCulloch, BJ. (1991). A longitudinal investigation of the factor structure of subjective well-being: The case of the Philadelphia Geriatric Center Morale Scale. Journal of Gerontology, 46(5), P251-P258.
- Melzack, R., & Casey, K.L. (1968). Sensory, motivational, and central control determinants of pain: A new conceptual model. In D.R. Kenshalo (Ed.), The skin senses (pp. 423-443). Springfield, IL: Charles C. Thomas Press.
- Miaskowski, C. (1997). Women and pain. Critical Care Nursing Clinics of North America, 9(4), 453-458.
- Mobily, P.R., Herr, K.A., Clark, M.K., & Wallace, R.B. (1994). An epidemiologie analysis of pain in the elderly: The Iowa 65+ Rural Health Study. Journal of Aging and Health, ¿(ISSUE?), 139-154.
- Parmelee, P.A. (1994). Assessment of pain in the elderly. In M.P. Lawton & J.A. Teresi (Eds.), Annual review of gerontology and geriatrics: Vol. 14. Focus on assessment techniques (pp. 281-301). New York: Springer.
- Parmelee, P.A. (1997). Paul and psychological function in late life. In D.I. Mostofsky &, J. Lomranz (Eds.), Handbook of pain and aging (pp. 207-226). New York: Plenum Press.
- Parmelee, P.A., Katz, I.R., & Lawton, M.P. (1991). The relation of pain to depression among institutionalized aged. Journal of Gerontology, 46(\), PI5-P21.
- Parmelee, P.A., Lawton, M.P., & Katz, LR. (1989). Psychometric properties of the geriatric depression scale among the institutionalized aged. Psychological assessment. Psychological Assessment, 1(4), 331338.
- Parmelee, P.A., Smith, B., & Katz, IJL (1993). Pain complaints and cognitive status among elderly institution residents. Journal of American Geriatrics Society, 41, 517-522.
- Pasero, C-, Pake, J.A., & McCaffery, M. (1999). Basic mechanisms underlying the causes and effects of pain. In M. McCaffery & C. Pasero (Eds.), Pain Clinical Manual (2nd éd.) (pp. 15-34). St. Louis, MO: Mosby.
- Raway, B. (1994). Pain behaviors and confusion in elderly patients with hip fracture (Doctoral dissertation, The Catholic University of America, 1994). Dissertation Abstracts International, 5J(02-B), UMI # 9418593.
- Sengstaken, E.A. & King, S.A. (1 993). The problems of pain and its detection among geriatric nursing home residents. Journal of the American Geriatrics Society, 41, 541-544.
- Smyer, M., Brannon, D., & Conn, M. (1992). Improving nursing home care through training and job redesign. The Gerontologist, 32, 327-333.
- Spore, D.L., Horgas, A.L., Smyer, M.A., & Marks, L.N. (1992). The relationship of antipsychotic drug use, behavior, and diagnoses among nursing home residents. Journal of Aging and Health, 4(4), 514535.
- Stone, J.T., Wyman, J.F., 8c Salisbury, S.A. (1999). Clinical gerontological nursing: A guide to advanced practice (2nd ed.). Philadelphia: W.B. Saunders.
- Teske, K., Daut, R.L., & Cleeland, C.S. (1983). Relationships between nurses' observations and patients* self-reports oí pain. Pain, /6(3), 289-296.
- Thomas, M.R., & Roy, R. (1988). Age and pain: A comparative study of the "younger and older" elderly. Pain Management, 8, 174-179.
- Vallerand, A.H. (1995). Gender differences in pain. Image: The Journal of Nursing SchoUrship, 27(3), 235-237.
- Wallace, M. (1994). Assessment and management of pain in the elderly. MEDSURG Nursing, 3(4), 293-298.
- Wells, N., Kaas, M., & Feldt, K. (1997). Managing pain in the institutionalized elderly. In D.I. Mostofsky & J. Lomranz (Eds.), Handbook of pain and aging (pp. 129-154). New York: Plenum Press.
- Weiner, D.K., Peterson, B.L., Logue, P., & Keefe, FJ. (1998). Predictors of pain selfreport in nursing home residents. Aging Clinical and Experimental Research, 10(5), 411-420.
- Werner, P., Cohen-Mansfield, J., Watson, V, & Pasis, S. (1998). Pain in participants of adult day care centers: Assessment by different raters. Journal of Pain and Symptom Management, /5(1), 8-17.
- Williamson, G.M., Sc Schulz, R. (1992). Pain, activity restriction, and symptoms of depression among community-residing elderly adults. Journal of Gerontology, 47(6), 367-372.
- Won, A., Lapane, K., Gambassi, G., Bernabé!, R-, Mor, V., & Lipsitz, L.A. (1999). Correlates and management of no amalignan t pain in the nursing home. Journal of the American Geriatrics Society, 47, 936-942.
- Yeager, K.A., Miaskowski, C., Dibble, S.L., & Wallhagen, M. (1995). Differences in pain knowledge and perception of the pain experience between outpatients with cancer and their family caregivers. Oncology Nurse Forum, 22(8), 1235-1241.
- Zalon, M.L. (1993). Nurses* assessment of postoperative patients* pain. Pain, 54(5), 329-334.
RESIDENT RATINGS OF PAIN PREVALENCE AND INTENSITY IN SUBSET OF RESIDENTS REPORTING PAIN (n = 22)
NA RATINGS OF PAIN PREVALENCE AND INTENSITY IN SUBSET OF RESIDENTS PERCEIVED TO H AVE PAIN(n = 16)
CONGRUENCE BETWEEN NA AND RESIDENT RATINGS OF PAIN PREVALENCE
RESIDENT-BASED DEMOGRAPHIC, DIAGNOSTIC, AND AFFECTIVE CORRELATES OF PAIN CONGRUENCE AND INCONGRUENCE