Pain Assessment in Elderly Patients with Severe Dementia

Transcription

1 48 Journal of Pain and Symptom Management Vol. 25 No. 1 January 2003 Original Article Pain Assessment in Elderly Patients with Severe Dementia Paolo L. Manfredi, MD, Brenda Breuer, MPH, PhD, Diane E. Meier, MD and Leslie Libow, MD Pain and Palliative Care Service (P.L.M.), Department of Neurology, Memorial Sloan-Kettering Cancer Center; Department of Neurology (B.B.), New York University; Department of Geriatrics and Adult Medicine (D.E.M., L.L.), Mount Sinai School of Medicine; The Jewish Home and Hospital (P.L.M., B.B., L.L.); and The Hospital for Joint Diseases (B.B.), New York, New York, USA Abstract The purpose of this study was to assess the reliability and validity of facial expressions as pain indicators in patients with severe dementia. Based on interviews with patients who could report pain, we defined characteristics of decubitus ulcers associated with reports of pain during dressing changes. We then evaluated 9 patients who had ulcers with these characteristics but were unable to communicate verbally because of severe dementia. We videotaped their facial expressions before and during their decubitus ulcer dressing change. We showed the videotape segments, in random order, to 8 medical students and 10 nurses. The 18 viewers were asked to infer the presence or absence of pain based on their observations of the patients facial expressions and vocalizations. The dressing change of decubitus ulcers extending beyond the subcutaneous tissue, covering an area of at least 9 cm 2, and with a moist surface, was always reported as painful by study patients able to report (95% confidence interval of %). The intraclass correlation coefficient for the answers of the 18 viewers evaluating each videotape segment for the presence of pain was Sensitivity, specificity, and positive and negative predictive values of viewers ratings of facial expressions and vocalizations as a measure of the presence of pain were: 0.70, 0.83, 0.90, and The intraclass correlation coefficient for the answers rating pain intensity was only 0.10, indicating only slight agreement beyond chance. Assuming dressing changes of ulcers reported as painful by communicative patients are also painful in non-verbal severely demented patients, clinician observations of facial expressions and vocalizations are accurate means for assessing the presence of pain, but not its intensity, in patients unable to communicate verbally because of advanced dementia. J Pain Symptom Manage 2003;25: U.S. Cancer Pain Relief Committee. Published by Elsevier. All rights reserved. Key Words Pain, dementia, assessment, decubitus, ulcer, geriatrics Address reprint requests to: Paolo L. Manfredi, MD, Pain and Palliative Care Service, Department of Neurology, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021, USA. Accepted for publication: February 25, Introduction In nursing homes, the prevalence of pain and dementia are high, ranging from 45% to 84% for pain 1,2 and from 40% to 78% for dementia. 3,4 Although the patient s report is the most accurate method for measuring pain, 5 pa U.S. Cancer Pain Relief Committee /03/$ see front matter Published by Elsevier. All rights reserved. PII S (02)

2 Vol. 25 No. 1 January 2003 Pain Assessment in Severe Dementia 49 tients with very advanced dementia cannot convey the experience of pain verbally. These patients are therefore at risk for undetected and untreated pain. 6,7 The reliability and validity of facial expressions and vocalizations as markers of pain in non-verbal, severely demented patients remain poorly defined. This study was conducted in two parts. The objective of the first part of the study was to identify a clinical condition consistently described as painful by patients who were able to verbally communicate the experience of pain. This condition also had to be common in patients with advanced dementia. We hypothesized that the dressing change of deep decubitus ulcers fulfilled these criteria. The objective of the second part of the study was to assess the reliability and validity of facial expressions as pain indicators in patients with severe dementia undergoing a painful procedure. The reports of communicative patients from the first part of the study were the basis for the definition of the procedure as painful. Methods This study was conducted in a nursing home that has 1427 long-term beds and 202 subacute care beds. The nursing home is affiliated with a university hospital. The study was approved by the Institutional Review Board (IRB) of the nursing home and by the IRB of the affiliated university hospital. Informed consent was obtained from patients for the first part of the study and from the patients surrogate for the second part of the study, as these patients were all severely demented and unable to consent. Over 9 months, we identified all patients with decubitus ulcers who, in the opinion of the nursing staff responsible for their care, were able to reliably answer questions about pain. We used a standardized questionnaire (Table 1) to confirm the ability of these patients to reliably answer questions about pain. Although this questionnaire has not been validated as a method for identifying patients able to report pain accurately, it is simple and face valid. We then assessed their pain while undergoing a dressing change for the care of their decubitus ulcer by asking standardized questions: Please tell me if you feel pain at any time during the dressing change. Whenever the patient would indicate the presence of pain the evaluator would ask: Is the pain a little or a lot? The Table 1 Questionnaire to Screen for Eligibility a 1. Do you have pain now? 2. Are you free of pain now? 3. Are you hurting now? 4. Is there any part of your body that hurts now? a Patients whose answers to questions 2, 3, and 4 were consistent with the answer for question 1 were eligible for the first part of the study. Patients whose answers to questions 2, 3, and 4 were not consistent with the answer for question 1 were eligible for the second part of the study. analysis of these interviews identified ulcer characteristics associated with pain during dressing changes. To confirm that these ulcer characteristics were indeed consistently associated with pain, we identified 8 additional consecutive patients with similar decubitus ulcers, able to report their pain, based on the Table 1 questionnaire, and we interviewed them with the same questions to assess the presence and intensity of pain during their dressing change. For the second part of the study, we identified 9 patients with a diagnosis of dementia and decubitus ulcers fulfilling the same characteristics. The diagnosis of dementia was based on the diagnoses recorded in the patient chart and was confirmed with a Mini Mental State Examination (MMSE). 8 We excluded patients able to answer the Table 1 questionnaire, patients with diabetes or a history of spinal cord pathology, and patients who would not wake up with verbal and tactile stimuli. We videotaped the facial expressions of these 9 patients, before and during the dressing change of the decubitus ulcer. For each patient, we then randomized the sequence of the two videotape segments, before and during the dressing change. We showed the videotapes, in the randomized sequence, to 8 fourth-year medical students and 10 nurses. At the end of each videotape segment, the viewers answered independently the questions outlined in Table 2. The viewers were blind to each other s ratings. We used the following precautions to ensure that the viewers would base their answers only on the patients facial expressions and vocalizations: the videotapes showed only the patient s face and, therefore, the viewers were blind to the clinical context; before recording the videotape at rest we positioned the patients for the dressing change so that each patient was in the same position during the two videotape segments; during the video segment of the pa-

3 50 Manfredi et al. Vol. 25 No. 1 January 2003 Table 2 Videotape Questionnaire for Medical Students and Nurses 1. Did the patient experience pain during ANY PART of this episode? Definitely not Probably not Undecided Probably yes Definitely yes 2. If you answered probably yes or definitely yes, please rate the pain experienced: Mild Moderate Severe Cannot rate degree of pain tient at rest we stimulated the patient verbally and with gentle touch to ensure a state of wakefulness; we edited the videotapes to ensure that the duration and the background sounds were the same for each patient. Validity analyses were based on the assumption, derived from the findings of the first part of the study, that patients were in pain in the segments videotaped during the dressing change and not in pain in the segments taped before the dressing change. Intraclass correlation coefficients for the ratings of the 18 viewers evaluating each videotape segment for the presence of pain were computed using the SAS macro program INTRACC, which is based on the method of Shrout and Fleiss. 9 Results The analysis of interviews and clinical data of 31 patients able to reliably answer questions about pain showed that ulcers with all of the three following characteristics were painful during dressing changes: 1) extension beyond the subcutaneous tissue, 2) surface of at least 9 cm 2, and 3) moist surface. The analysis of interviews of 8 additional, consecutive patients confirmed the painful nature of the dressing change in this context. All 8 patients stated that they experienced pain during the dressing change of the ulcer and no pain at rest, before the dressing change. For 6 of the 8 patients the pain was rated as a lot and for 2 as a little. Based on this 8/8 report of pain, the 95% confidence interval that patients with similar ulcers would experience pain during the dressing change is %. Table 3 lists the demographic data and analgesic use for these 8 patients. Table 4 lists the same data for the 9 nonverbal, severely demented patients who participated in the videotaping part of the study. Among the 9 patients unable to answer questions about pain, the MMSE was 0 for 8 patients and 3 for 1 patient. When selecting these 9 non-verbal patients, 23 patients were excluded because of diabetes, 4 because of possible spinal cord injury, and 1 because of coma. Table 5 summarizes the results for the first of the two questions shown in Table 2, administered to the 8 fourth-year medical students and the 10 nurses. The intraclass correlation coefficients were 0.65 for the answers of medical students, 0.64 for the answers of the nurses, and 0.64 for the entire group. These values indicate substantial agreement beyond chance (10). The sensitivity, specificity, positive predictive value, and negative predictive values of viewers ratings of facial expressions as a measure of the presence of pain were: 0.70, 0.83, 0.90, 0.81 for all 18 viewers, respectively; 0.67, 0.89, 0.92, 0.79 for the 10 nurses, respectively; and 0.74, 0.75, 0.88, and 0.82 for the 8 medical students, respectively. For the only patient who was receiving an opioid medication (transdermal fentanyl patch, Patient Table 3 Demographic Data and Analgesics for the 8 Patients Able to Answer Questions About Pain Age, median (range) 85.9 ( ) Race (6 W, 2 B) Sex (4 females, 4 males) Analgesics Administered White Female acetaminophen White Female morphine White Male acetaminophen White Female acetaminophen White Male acetaminophen Black Female none Black Male none White Male none

4 Vol. 25 No. 1 January 2003 Pain Assessment in Severe Dementia 51 Patient Table 4 Demograpic Data and Analgesics for the 9 Demented Patients Unable to Answer Questions About Pain Age, median (range) 83.8 ( ) Race (5W, 2B, 2H) Sex (7 females, 2 males) Analgesics Administered White Female acetaminophen Hispanic Female acetaminophen Black Female acetaminophen Black Male acetaminophen Hispanic Male acetaminophen White Female acetaminophen White Female fentanyl patch White Female none White Female acetaminophen 50 g/hour), 14 of 18 raters chose no pain while viewing the videotape segment of the dressing change ( patient in pain ) This represented the highest rate of a no pain answer for a segment videotaped during a dressing change. For all 18 raters, the intraclass correlation coefficient for the second question shown in Table 2 rating pain intensity was 0.10, indicating only slight agreement beyond chance. 10 Discussion Pain is an unpleasant sensory and emotional experience, induced by sensory stimuli and interpreted and modulated by individual emotions, memories, and expectations. Pain is a subjective feeling, with no known biological markers: proof of its presence and measurement of its intensity rely entirely on the patient s selfreport. Although cognitive impairment that is not severe enough to interfere with interviewing has not been found to mask pain complaints, 2 severely demented patients, who cannot communicate verbally, are at risk for under-detection and under-treatment of pain. 6,7 Decubitus ulcers have been shown to cause pain, 11 an observation confirmed in the present study by the patients able to reliably report pain. In these patients, the dressing change of deep, moist ulcers of at least 9 cm 2 always caused pain, thereby reliably identifying this procedure as almost always painful. Eight of the 9 patients with advanced dementia, videotaped during the dressing changes of similar decubitus ulcers, consistently demonstrated facial expressions that were identified by observers as indicative of pain. The poorest recognition of a patient in pain, where 14 of 18 viewers did not infer the presence of pain while viewing a videotape segment during the dressing change, occurred in the only patient receiving opioids. Although it is possible that, due to the opioid analgesic effect, this patient did not experience pain as severe as the other patients, observations from a single patient are difficult to interpret. These results suggest that facial expressions, even when viewed in isolation from the clinical context are valid and reliable indicators of the presence of pain. The use of interviews of cognitively intact patients to extrapolate their experience of pain to patients with advanced dementia has been previously employed by other investigators. 12 In our study, facial expressions of demented patients suggestive of pain correlated with a procedure presumed to be painful. These results reinforce that it is reasonable to assume that procedures Table 5 Reliability, Sensitivity, Specificity and Positive and Negative Predictive Value of Observer Evaluations of Facial Expressions as Indicators of Pain Parameter Medical Students (n 8) Nurses (n 10) Combined group (n 18) Reliability (lower end of 95% CI) 0.65 (0.50) 0.64 (0.51) 0.64 (0.52) Sensitivity (SD) 0.74 (0.14) 0.67 (0.13) 0.70 (0.14) Specificity (SD) 0.75 (0.26) 0.89 (0.09) 0.83 (0.20) Positive predictive value (SD) 0.88 (0.10) 0.92 (0.074) 0.90 (0.088) Negative predictive value (SD) 0.82 (0.11) 0.79 (0.063) 0.81 (0.088)

5 52 Manfredi et al. Vol. 25 No. 1 January 2003 that are painful in cognitively intact patients will also cause pain in patients unable to communicate because of dementia, although the observations made for patients with pain during dressing changes may not be applicable to patients with other type of pains. While raters were reliably able to identify the presence of pain, the reliability for rating pain intensity (second question on Table 2) was low, underscoring the need for more studies to assess pain in patients with severe dementia. Due to the exclusion of patients in a comatose state our results may not apply to patients who are not alert. In summary, clinicians relying on the close observation of facial expressions of patients with severe dementia to infer the presence or absence of pain can expect to be correct between 80% and 90% of the time. When facial expressions are integrated within the clinical context of a painful procedure or disease process, the ability of the clinician to detect the presence of pain is likely to be even higher. Acknowledgments This research was supported by a grant to the Jewish Home & Hospital from the New York State Department of Health, Bureau of Long- Term Care Services. References 1. Stein WM, Ferrell BA. Pain in the nursing home. Clin Geriatr Med 1996;12(3): Parmelee PA. Pain in cognitively impaired older persons. Clin Geriatr Med 1996;12(3): Rovner BW, Kafonek S, Filipp L, et al. Prevalence of mental illness in a community nursing home. American J Psy 1986;143(11): Magaziner J, Zimmerman SI, German PS, et al. Ascertaining dementia by expert panel in epidemiologic studies of nursing home residents. Ann Epidemiology 1996;6(5): Merskey H, Bogduk N, eds. Classification of chronic pain. Seattle: International Association for the Study of Pain Press, Feldt KS, Ryden MB, Miles S. Treatment of pain in cognitively impaired compared with cognitively intact older patients with hip-fracture. J Am Geriatr Soc 1998;46: Morrison RS, Siu AL. A comparison of pain and its treatment in advanced dementia and cognitively intact patients with hip fractures. J Pain Symptom Manage 2000;19: Folstein F, Folstein SE, McHugh PR. Mini-Mental State : a practical method for grading the cognitive state of patients for the clinician. J Psychiatric Research 1975;12: Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull 1979;86: Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33: Dallam L, Smyth C, Jackson BS, et al. Pressure ulcer pain: assessment and quantification. J Wound Ostomy Continence Nurs 1995;22: Morrison RS, Ahronheim JC, Morrison GR, et al. Pain and discomfort associated with common hospital procedures and experiences. J Pain Symptom Manage 1998;15(2):