Individual Research Article Critique Presentation Resource: The research study that you selected in Week Two Develop a 10- to 15-minute presentation in which you address the following points (7 pts): •Strengths and weaknesses of the study •Theoretical and methodological limitations •Evidence of researcher bias •Ethical and legal considerations related to the protection of human subjects •Relationship between theory, practice, and research •Nurse’s role in implementing and disseminating research •How the study provides evidence for evidence-based practice •

Identify the following for the research study selected (choose 1 or 2 NOT BOTH): 8 pts. • 1. Quantitative Research Article Critique (Follow the example pp. 433–442 of the text): • a. Phase 1: Comprehension b. Phase 2: Comparison c. Phase 3: Analysis d. Phase 4: Evaluation • 2. Qualitative Research Article Critique (Follow the example pp. 455–461 of the text): • a. 1. Problem (problem statement; purpose; research questions; literature review; frame of reference; research tradition) b. 2. Methodology (sampling & sample; data collection; protection of human subjects c. 3. Data (management; analysis . 4. Results (findings; discussion; logic; evaluation summary • Format the presentation as one of the following (5 pts): •Poster presentation in class •Microsoft® PowerPoint® presentation including detailed speaker’s notes •Video of yourself giving the presentation uploaded to an Internet video sharing site such as www. youtube. com --Submit the link to your facilitator, include a written reference page in APA format •Another format approved by your facilitator Pain Assessment in Persons with Dementia: Relationship Between Self-Report and Behavioral Observation Ann L.

Horgas, RN, PhD,A Amanda F. Elliott, ARNP, PhD,w and Michael Marsiske, PhDz OBJECTIVES: To investigate the relationship between self-report and behavioral indicators of pain in cognitively impaired and intact older adults. DESIGN: Quasi-experimental, correlational study of older adults. SETTING: Data were collected from residents of nursing homes, assisted living, and retirement apartments in northcentral Florida. PARTICIPANTS: One hundred twenty-six adults, mean age 83; 64 cognitively intact, 62 cognitively impaired.

MEASUREMENTS: Pain interviews (pain presence, intensity, locations, duration), pain behavior measure, Mini-Mental State Examination, analgesic medications, and demographic characteristics. Participants completed an activitybased protocol to induce pain. RESULTS: Eighty-six percent self-reported regular pain. Controlling for analgesics, cognitively impaired participants reported less pain than cognitively intact participants after movement but not at rest. Behavioral pain indicators did not differ between cognitively intact and impaired participants. Total number of pain behaviors was signi? antly related to self-reported pain intensity (b 5 0. 40, P 5. 000) in cognitively intact elderly people. CONCLUSION: Cognitively impaired elderly people selfreport less pain than cognitively intact elderly people, independent of analgesics, but only when assessed after movement. Behavioral pain indicators do not differ between the groups. The relationship between self-report and pain behaviors supports the validity of behavioral assessments in this population. These ? ndings support the use of multidimensional pain assessment in persons with dementia.

J Am Geriatr Soc 57:126–132, 2009. Key words: pain; dementia; measurement From the ADepartment of Adult and Elderly Nursing, University of Florida, College of Nursing, Gainesville, Florida; wDepartment of Ophthalmology, School of Medicine, University of Alabama at Birmingham, Birmingham, Alabama; and zDepartment of Clinical and Health Psychology, College of Public Health and Health Professions, University of Florida, Gainesville, Florida. Address correspondence to Ann Horgas, College of Nursing, University of Florida, PO Box 100197-HSC, 101 S.

Newell Drive, Room 2201, Gainesville, FL 32610. E-mail: ahorgas@u?. edu DOI: 10. 1111/j. 1532-5415. 2008. 02071. x ain, a persistent daily problem for many elderly adults, is associated with physical and social disability, depression, and poor quality of life. 1 Between 50% and 86% of older adults experience pain; 32% to 53% of those with dementia experience it daily. 2 The high prevalence is associated with proliferation of pain-related health conditions in late life, such as osteoarthritis, hip fractures, peripheral vascular disease, and cancer. Dementia complicates pain assessment, because it impairs memory, judgment, and verbal communication. Dementia is associated with central nervous system changes that alter pain tolerance4 but not pain thresholds (e. g. , minimum level at which a painful stimulus is recognized as pain). 5 No empirical evidence indicates that persons with dementia physiologically experience less pain; rather, they appear less able to recognize and verbally communicate the presence of pain. Findings that cognitively impaired older adults underreport pain relative to nonimpaired elderly people7 and are less likely to be treated for pain than their cognitively intact peers8,9 re? ect dif? culty assessing pain in this population. Self-report is considered the criterion standard of pain assessment. Despite recent studies supporting the reliability and validity of self-report in persons with dementia,7,10 healthcare providers and pain experts recognize that selfreport alone is insuf? cient for this population and that observational pain assessment strategies are needed.

In 2002, the American Geriatrics Society established comprehensive guidelines for assessing behavioral indicators of pain. 1 More recently, the American Society for Pain Management Nursing Task Force on Pain Assessment in the Nonverbal Patient (including persons with dementia) recommended a comprehensive, hierarchical approach that integrates selfreport and observations of pain behaviors. 11 Recently, tools to measure pain in persons with dementia have proliferated. In 2006, a comprehensive stateof-the-science review of 14 observational pain measures was completed.

The authors concluded that existing tools are still in the early stages of development and testing and that more psychometric work is needed before tools are recommended for broad adoption in clinical practice. 12 Others, including an interdisciplinary expert consensus P JAGS 57:126–132, 2009 r 2008, Copyright the Authors Journal compilation r 2008, The American Geriatrics Society 0002-8614/09/$15. 00 JAGS JANUARY 2009–VOL. 57, NO. 1 PAIN ASSESSMENT IN PERSONS WITH DEMENTIA 127 panel on pain assessment in older persons,13 have corroborated these conclusions. 4 In particular, these authors highlight the need for more evaluation of observational pain measures, including validation against the criterion standard of self-report in intact and impaired populations. Almost all research on measuring pain in persons with dementia has focused exclusively on persons with moderate to severe disease. There has been only one published study that compared pain behaviors and self-reported pain in persons with and without cognitive impairment, but it focused on postoperative patients undergoing rehabilitation and acute pain associated with physical therapy. 5 Thus, the purpose of this study was to investigate the relationship between self-report and behavioral indicators of pain in cognitively intact and impaired older adults with persistent pain. Speci? cally, this study evaluated whether cognitive status (intact or impaired) differentially in? uenced verbal and nonverbal expression of pain. It was hypothesized that self-reported pain would be lower in cognitively impaired elderly people than in those who were cognitively intact but that pain behaviors, because they are more re? exive and less reliant on verbal communication, would be equivalent in both groups.

The relationship between pain behaviors and self-reported pain was also evaluated in cognitively intact elderly people to validate whether behaviors measured are indicators of pain. The following research questions were asked. Does cognitive status in? uence self-reported pain? Does cognitive status in? uence observed pain behaviors? Are self-reported pain and observed pain behaviors related, and is the relationship different in cognitively intact and impaired elderly people? One hundred forty participants were enrolled and completed the baseline interview; 126 (90%) completed the protocol. Attrition analyses revealed no signi? ant differences between completers and noncompleters on demographic, residential status, health, or pain variables. The ? nal sample was predominantly female (81%), Caucasian (97%), and widowed (60%), with a mean age of 83 (range 5 65–98). Thirty-nine percent resided in nursing homes, 39% resided in assisted living, and 22% lived independently in retirement apartments. Participants’ average Mini-Mental State Examination (MMSE) raw score was 24 (range 5 7–30, median 5 27, mode 5 29). Based on 10th percentile education-adjusted MMSE norms as the cutoff,16,17 64 (50. 8%) were cognitively intact, and 62 (49. %) were impaired. See Table 1 for a description of the total sample and of cognitively intact and impaired subsamples. Groups differed only in residential status (cognitively Table 1. Sample Characteristics, Overall (N 5 126) and According to Cognitive Status Total Sample Cognitive StatusA Intact Impaired (n 5 64) (n 5 62) PValue Characteristic METHODS The University of Florida institutional review board approved this study. Informed consent was obtained from cognitively intact participants and from impaired elderly people’s legally authorized representatives, with assent from persons with dementia.

Design A quasi-experimental, correlational design was used to investigate pain in older adults with mild to moderate dementia, because dementia status cannot be experimentally manipulated. Cognitively intact elderly people functioned as a comparison group to examine behavioral indicators and self-reported pain in the two groups. If self-report and behaviors were related in cognitively intact persons, there would be some basis to infer that the same behaviors indicated pain in cognitively impaired elderly people. Participants One hundred ? ty-eight older adults were screened for enrollment from 17 assisted living facilities, nursing homes, and retirement communities in north central Florida. Inclusion criteria were aged 65 and older, English-speaking, able to stand up from a chair and walk in place, diagnosed osteoarthritis in the lower body, and adequate vision and hearing to complete the interview. Sex, n (%) Male 24 (19. 0) 12 (18. 8) 12 (19. 4) Female 102 (81. 0) 52 (81. 3) 50 (80. 6) Race, n (%) White 123 (97. 6) 63 (98. 4) 60 (96. 8) Black 1 (0. 8) 0 (0) 1 (1. 6) Other 2 (1. 6) 1 (1. 6) 1 (1. 6) Marital status, n (%) Married 37 (29. ) 21 (32. 8) 16 (25. 8) Unmarriedw 89 (70. 6) 43 (67. 2) 46 (74. 2) Education, n (%) ohigh school 11 (8. 7) 5 (7. 8) 6 (9. 7) graduate High school graduate 38 (30. 2) 17 (26. 6) 21 (33. 9) Some college or 31 (24. 6) 18 (28. 1) 13 (21. 0) equivalent College graduate or 34 (27. 0) 18 (28. 1) 16 (25. 8) more Residence Assisted living 49 (38. 9) 28 (43. 8) 21 (33. 9) Nursing home 47 (37. 3) 14 (21. 9) 33 (53. 2) Retirement apartment 30 (23. 8) 22 (34. 4) 8 (12. 9) Analgesics taken 579 ? 1,320 313 ? 699 853 ? 1,708 (in acetaminophen equivalents), mean ? SD Age, mean ? SD 82. 2 ? 7. 3 81. 9 ? 7. 83. 1 ? 7. 6 Number of medical 6. 7 ? 3. 1 6. 6 ? 2. 9 6. 9 ? 3. 4 diagnoses, mean ? SD .93 .59 .39 .84 .001z .02§ .55 . 63 A Cognitive status was computed using the following education-adjusted Mini-Mental State Examination scores as cutoffs: o8th grade education, 20; 9 to 11 years, 24; high school graduate or equivalent, 25; some college, 27; and college degree or higher 5 27. 16,17 w Unmarried 5 never married, widowed, separated, or divorced. z Chi-square 5 15. 2, degrees of freedom 5 2, P 5. 001. § t (124) 5 2. 22. SD 5 standard deviation. 128 HORGAS ET AL. JANUARY 2009–VOL. 57, NO. 1 JAGS mpaired elderly people were signi? cantly more likely to reside in assisted living or nursing home facilities). to use in elderly adults than the traditional visual analogue scale. 21 Procedures Participants completed a brief screening interview to con? rm study eligibility and to ascertain cognitive status. Those eligible were interviewed about their pain and completed an activity-based protocol designed to evoke pain behaviors in persons with persistent pain (described in more detail below). Activity Protocol Participants were asked to sit, stand, lie on a bed, walk in place, and transfer between activities.

Based on previous work, the activity protocol had several strengths for use with this population. First, it simulates performance of basic activities of daily living, thereby enhancing ecological validity of the tasks. Second, it was tested in other studies, and activities were shown to induce pain in persons with osteoarthritis and chronic low back pain, thus providing a naturalistic pain induction method. Third, use of these realworld tasks avoids undue health or safety risks for elderly adults and eliminates potential bias associated with arti? cially induced (e. g. , laboratory-based) pain induction techniques. 8,19 The protocol was simpli? ed by using only 1-minute activity intervals (to reduce complexity of directions and physical demands for frail or cognitively impaired participants) and substituted walking in place for walking across the room and back (to accommodate physical space limitations in residential care facilities where data were collected). Activities were conducted in random order to minimize order effects, and the entire 10-minute protocol was videotaped. Measures Self-Reported Pain The principal investigator (ALH) or a trained research assistant interviewed each participant in a private session about their pain experience.

Pain presence, intensity, locations, and duration were assessed. Pain Presence. Questions from the Structured Pain Interview (SPI)20 were used to assess presence of self-reported pain. During the pain screening interview, participants were asked ‘‘Do you have some pain every day or almost every day (daily pain)? ’’ Pain was also assessed immediately before the start of the activity protocol (‘‘Are you having any pain right now? ’’ (pre-activity)) and immediately after it (‘‘Did you experience any pain during these activities? ’’ (postactivity)).

Response choices to all three questions were yes (1) or no (0). Pain Intensity If participants responded ‘‘yes’’ to experiencing pain (daily, pre-activity, or postactivity), they were asked to rate the intensity using a numerical rating scale (NRS). The NRS was presented as a horizontal line with 0 5 no pain and 10 5 worst pain as anchors and equally spaced dashes representing pain intensity rating of numbers 1 through 9. The scale was printed in large, bold font on an 8. 5" A 11" paper to facilitate use with older adults who may have vision dif? culties. The NRS is considered valid, reliable and easier

Pain Duration Participants were asked to indicate how long (in months and years) they had experienced daily or almost daily pain. Responses were coded as less than 1 year, 1 to 5 years, 6 to 10 years, 11 to 15 years, or more than 15 years. Pain Locations The pain map from the McGill Pain Questionnaire22 was used to assess pain locations. Participants indicated areas on the body drawing in which they were currently experiencing pain. Total number of painful locations was summed. This widely used measure has been validated in several epidemiological studies and has high interrater reliability (average kappa 5 0. 2). 23 Observed Pain Behaviors Pain Behaviors A modi? ed version of the Pain Behavior Measure18 was used to measure behavioral indicators of pain. Based on standardized behavioral de? nitions, occurrence of the following speci? c pain behaviors was evaluated: rigidity, guarding, bracing, stopping the activity, rubbing, shifting, grimacing, sighing or nonverbal vocalization, and verbal complaint. Standardized de? nitions were adapted from previous work,18,19 modi? ed for use in this older, moreimpaired population, and pilot tested in a sample of nursing home residents with dementia. 4 This measure has adequate reliability and validity. 13 Pain Behavior Coding Independent raters, all registered nurses blind to participants’ cognitive status, scored the videotaped activity protocols. Coders completed extensive training in coding procedures until intrarater and interrater agreement (with the master coder (PI) and another rater) reached a kappa coef? cient of 0. 80 or greater, indicating good to very good reliability. 25 After coding reliability was attained, reliability checks were conducted on 10% of all videotapes to minimize rater drift.

Noldus Observer software was used to analyze digitized videotapes and code pain behaviors (Noldus Information Technology, Wageningen, the Netherlands). The following summary variables were created and used in the analyses: total number of pain behaviors observed, number of times each behavior (rigidity, guarding, bracing, stopping, rubbing, shifting, grimacing, sighing or nonverbal vocalization, and verbal complaint) was observed, and total numbers of pain behaviors observed during each activity state (e. g. , number of behaviors while walking, reclining, sitting, standing, and transferring).

Cognitive Status Cognitive status was assessed using the MMSE,26 an 11-item screening instrument widely used to assess general cognitive status in elderly adults. The following MMSE scores served as the cutoffs to classify participants as intact or impaired: less than 8th grade education, 20; 9 to 11 years, 24; high school graduate or equivalent, 25; some college, 27; and college degree or higher, 27. 16,17 JAGS JANUARY 2009–VOL. 57, NO. 1 PAIN ASSESSMENT IN PERSONS WITH DEMENTIA 129 Analgesic Medications Drug data for each participant were coded according to the American Hospital Formulary Service system.

All pain medications were identi? ed and converted to acetaminophen equivalents. 8,27 This standardized drugs and dosages to a common metric and facilitated comparison of analgesic dosing. To ensure that only analgesics actually taken would be controlled for, equianalgesic dosages were considered in these analyses only if they were taken within the standard therapeutic dosing window for each drug (e. g. , acetaminophen, every 4–6 hours) before the activity protocol. Data Analysis SPSS, version 15. 0 (SPSS Corp. , Chicago, IL) was used for data analysis.

Descriptive statistics, Pearson chi-square (w2) tests, and t-tests were used to describe sample characteristics and examine group differences. Analysis of covariance (ANCOVA) was used to test relationships between cognitive status, pain intensity, and pain behaviors. Logistic regression was used to predict pain presence. Multiple regression was used to predict pain intensity and number of pain behaviors, with a centered cognitive status–by–pain intensity interaction term to identify group differences; standardized regression coef? cients (b) are reported in the results.

RESULTS Self-Reported Pain The majority of participants (86. 5%) reported experiencing pain every day or almost every day. More than 65% reported experiencing pain for more than 1 year ( $ 40% indicated duration of 45 years). On average, participants reported pain in four body locations (range 5 1–25); usual pain intensity was 4. 3 (moderate) on a scale from 0 to 10. Immediately before the activity protocol, 45 (35. 7%) participants reported experiencing pain. Mean pain intensity was rated as 1. 7 (range 5 0–9). After the protocol, 79 (62. 7%) reported experiencing pain during the activities; mean pain intensity was 3. (range 5 0–9). Relationship Between Cognitive Status and Self-Reported Pain Chi-square analyses were conducted to examine the relationship between cognitive status (impaired vs intact) and presence of self-rated daily pain and pain duration at baseline. The baseline pain interview was not always conducted on the same day as the activity protocol, and analgesic use before the interview was not assessed. Thus, initial analyses are descriptive only and do not control for analgesic use. At baseline, 77. 4% of impaired and 95. 3% of intact participants reported experiencing pain every day (w2(1) 5 8. 6, P 5. 003).

Cognitively impaired elderly people also recalled shorter pain duration (w2(3) 5 16. 0, P 5. 001) than intact participants, but no signi? cant differences were reported in the number of pain locations. Logistic regression, controlling for acetaminophen equivalents, indicated that cognitive status was not signi? cantly predictive of pre-activity pain presence. Regression analyses, with pre-activity pain intensity as the dependent variable and cognitive status and analgesics as predictors, revealed no signi? cant difference between the two groups (Figure 1). Intact Impaired 16 14 12 Mean values 10 8 6 4 2 0 In te a * t ns y SR :P a re- cti v in Pa ng cing ing rbal aint sity pi b l n e ra uar ig Sh op rima Rub onv mp Inte B G R St G N al co ain P rb Ve activ tos :P SR b Pain indicators cin g n di g i id ty in ift g a tt Si g g g g g in din kin yin rrin l e n L sf a Wa St an Tr c Activity states Figure 1. Relationship between self-report and observed pain behaviors in cognitively intact and cognitively impaired elderly people (N 5 126). aMean self-reported (SR) pain intensity, controlling for acetaminophen equivalents taken. bMean number of behaviors observed for each pain indicator, controlling for acetaminophen equivalents taken. Mean number of behaviors observed during each activity state, controlling for acetaminophen equivalents taken. 130 HORGAS ET AL. JANUARY 2009–VOL. 57, NO. 1 JAGS At the end of the activity protocol, cognitive status was signi? cantly associated with the reported presence of pain, controlling for analgesics (b 5 1. 2, P 5. 002); cognitively impaired elderly people were less likely to report pain. Impaired participants also reported signi? cantly lessintense pain than intact participants after the activity protocol (3. 8 vs 2. 6; F (1) 5 A 5. 0, P 5. 03).

Paired t-tests indicated that pain intensity increased signi? cantly from start to end of the protocol for both groups (Figure 1). Table 2. Relationship Between Self-Reported Pain Intensity and Observed Pain Behaviors (N 5 126) Total Number of Behaviors Observed Model bA P-Value 1 Pre-activity pain intensity Analgesics taken Pain intensity A cognitive status R2 F 2 Postactivity pain intensity Analgesics taken Pain intensity A cognitive status R2 F Standardized regression coef? cient. R2 5 coef? cient of determination. A Relationship Between Cognitive Status and Observed Pain Behaviors On average, 21. pain behaviors per person (range 5 3–50, median 5 21, mode 5 16) were observed during the activity protocol. ANCOVA models, controlling for analgesics, revealed no signi? cant differences in mean number of pain behaviors observed between cognitively intact and impaired participants (covariate-adjusted means 5 21. 8 and 21. 3, respectively; F (1) 5 0. 08, P 5. 77). The number of occurrences of each of the eight behavioral indicators observed was summed. ANCOVA models, controlling for analgesics and using Bonferroni correction for multiple comparisons (P 5. 005), revealed no signi? ant differences between cognitively intact and impaired elderly people for any behavioral pain indicators investigated (Figure 1). Of the activity states observed during the protocol, transferring elicited the most frequent pain behaviors (mean 5 13. 4; range 5 2–43). No signi? cant differences were noted between cognitively intact and impaired participants in number of behaviors observed during any of the ? ve observed activity states. Relationship Between Self-Reported Pain and Observed Pain Behaviors Regression analyses were conducted to examine the relationship between elf-reported pain intensity and total number of pain behaviors observed, controlling for analgesics. Before the activity protocol, pain intensity was signi? cantly predictive of the pain behaviors sum score (b 5 0. 27, P 5. 002), but the relationship did not differ between cognitively intact and impaired participants. After the activity protocol, self-reported pain intensity was signi? cantly (and more strongly) related to number of pain behaviors observed (b 5 0. 40, P 5. 000), and the painby-cognitive status interaction was signi? cant (b 5 0. 22, P 5. 008). Thus, postactivity pain intensity and summed behavioral indicators were signi? antly related in intact but not impaired participants (Table 2). DISCUSSION It was found that cognitive impairment diminishes selfreported pain assessed at rest but only when analgesics are not controlled. At baseline, cognitively impaired elderly people were signi? cantly less likely than cognitively intact elderly people to report pain, consistent with reports in the literature,7 but when analgesics were controlled for, these differences disappeared. This ? nding highlights the need to control for analgesics taken when making group comparisons, which to the best of the authors’ knowledge, has not been previously done.

The few studies reporting medication use include drugs prescribed or number of doses taken 0. 27 0. 01 0. 09 0. 08 2. 9 0. 40 A 0. 03 . 22 . 18 6. 70 .003 . 99 . 30 . 02 . 00 . 75 . 01 . 000 (regardless of medication class), whereas the current study identi? ed analgesics in the subject’s body during the pain assessment protocol. After the activity-based protocol was completed, selfreported pain intensity increased for both groups, but cognitively impaired elderly people reported less-intense pain than their intact peers. This ? ding supports the usefulness of the protocol to exacerbate pain in those with painful conditions and highlights the importance of mobility-based pain assessments. 12,14 This ? nding held even when the amount of analgesics taken by participants was controlled for in the statistical analysis. Behavioral indicators of pain observed during activities were equivalent across both groups. This ? nding contradicts previous work15 and may re? ect that medication use was controlled for and that the focus of the current study was on persistent pain, as opposed to more-acute, postoperative pain. This research con? ms that reliance on selfreport alone is insuf? cient to assess pain in older adults with dementia, because the pain experience may be underestimated,11 and supports growing recognition that behavioral observation is a necessary and useful pain measure, particularly in subjects with cognitive impairment. Cognitively impaired elderly people took signi? cantly more pain medication than their intact peers. The difference was approximately 500 acetaminophen equivalents, approximately the dose of one extra-strength acetaminophen tablet. This ? nding, which contradicts previous work,8,9 warrants further investigation.

Post hoc analyses indicated that this difference was not attributable to residential status, number of medical conditions, or demographic characteristics. Thus, it may re? ect recent changes in prescriptive practice as a result of heightened focus on pain in older adults with dementia. Another important ? nding is the signi? cant relationship between self-reported pain intensity and observed pain behaviors in cognitively intact persons. This ? nding provided support for the validity of behavioral pain JAGS JANUARY 2009–VOL. 57, NO. 1 PAIN ASSESSMENT IN PERSONS WITH DEMENTIA 31 indicators against the criterion standard of self-report, as least in cognitively intact elderly people, and is consistent with other researchers’ ? ndings. 28 Because there is no evidence that cognitively impaired elderly people experience less pain, it is reasonable to infer that pain behaviors are a valid indicator of pain in persons with dementia, although this assumption cannot be directly tested unless biological tests are developed. 12,24 Pain is subjective, and pain behaviors can be dif? cult to interpret, be subject to bias, and lack speci? city. 14,29 It has been uggested that some behaviors may indicate anxiety or generalized distress, not pain, in those with advanced dementia. 29,30 Thus, pain behavior measurements should be used in conjunction with selfreport, not as a replacement, and in the context of a comprehensive pain assessment. 14,30 Study strengths are that cognitively intact and impaired elderly people participated, thereby facilitating comparison of assessment strategies in persons of differing cognitive abilities, that a careful analysis of analgesics used during the pain assessment was conducted, and that persistent pain was focused on.

Most related prior research has included only persons with advanced dementia and postoperative pain. The sample was limited, however, by being primarily Caucasian and by being restricted to individuals with mild to moderate dementia. This was likely because of inclusion criteria requiring that participants be able to rise, stand, and walk. Individuals with severe dementia are typically more immobilized and unable to follow directions, factors that would impair ability to complete the activity-based protocol in this study. Thus, generalizations are limited, and further study is needed.

This study contributes several important ? ndings to the discourse on pain assessment in persons with dementia. First, it was con? rmed that self-reported pain, although still attainable, may be less reliable in those with mild to moderate dementia than in cognitively intact elderly people, depending on when it is assessed. Second, assessment of pain during movement is supported. Cognitively intact and impaired elderly people both showed greater self-reported pain intensity after movement, indicating that static assessment may underestimate pain.

Third, results support the validity of behavioral pain assessment against the criterion standard of self-report and provide evidence of an association between summed pain behaviors and self-reported pain intensity. More work is needed to establish scale properties of pain behaviors in relation to pain severity before this approach can be translated to clinical practice. Fourth, ? ndings highlight the importance of carefully evaluating analgesics taken when measuring pain, since results indicate that cognitively intact and impaired elderly people with persistent pain are often medicated differently.

This ? nding may re? ect a change in prescriptive practice that warrants further investigation. (Dr. Horgas) and a John A. Hartford Foundation Building Academic Geriatric Nursing Capacity Pre-doctoral Scholarship (Dr. Elliott). Author’s Contributions: Dr. Horgas was responsible for scienti? c oversight of all aspects of the study reported in this manuscript, including study design, data collection, data management, data analyses, and manuscript preparation. Dr. Elliott provided critical review of the manuscript and contributed to the design and study methods, data collection, and data coding.

Dr. Marsiske provided critical review of the manuscript and contributed to the design and study methods, data management, and statistical analyses. All authors have approved the ? nal version of this manuscript that was submitted for publication. Sponsor’s Role: The National Institute of Nursing Research sponsored this study but had no role in the design, methods, subject recruitment, data collections, data analyses, or manuscript preparation. REFERENCES 1. American Geriatrics Society. Clinical practice guidelines: The management of persistent pain in older persons.

J Am Geriatr Soc 2002;50:S205–S224. 2. Shega JW, Hougham GW, Stocking CB et al. Pain in community-dwelling persons with dementia: Frequency, intensity, and congruence between patient and caregiver report. J Pain Symptom Manage 2004;28:585–592. 3. Helme RD, Gibson SJ. The epidemiology of pain in elderly people. Clin Geriatr Med 2001;17:417–431. 4. Benedetti F, Vighetti S, Ricco C et al. Pain threshold and tolerance in Alzheimer’s disease. Pain 1999;80:377–382. 5. Huffman JC, Kunick ME. Assessment and understanding of pain in patients with dementia. Gerontologist 2000;40:574–581. . Bachino C, Snow AL, Kumik M et al. Principles of pain assessment and treatment in non-communicative demented patients. Clin Gerontol 2001;23: 97–115. 7. Fisher SE, Burgio LD, Thorne BE et al. Pain assessment and management in cognitively impaired nursing home residents: Association of certi? ed nursing assistant pain report, Minimum Data Set pain report, and analgesic medication use. J Am Geriatr Soc 2002;50:152–156. 8. Horgas AL, Tsai PF. Analgesic drug prescription and use in cognitively impaired nursing home residents. Nurs Res 1998;47:235–242. 9.

Won A, Lapane K, Gambassi G et al. Correlates and management of nonmalignant pain in the nursing home. J Am Geriatr Soc 1999;47:936–942. 10. Pautex S, Michon A, Guedira M et al. Pain in severe dementia: Self-assessment or observational scales. J Am Geriatr Soc 2006;54:1040–1045. 11. Herr K, Coyne PJ, Key T et al. Pain assessment in the nonverbal patient: Position statement with clinical practice recommendations. Pain Manage Nurs 2006;7:44–52. 12. Herr K, Bjoro K, Decker S. Tools for assessment of pain in nonverbal older adults with dementia: A state-of-the-science review.

J Pain Symptom Manage 2006;31:170–192. 13. Hadjistavropoulos T, Herr K, Turk D et al. An interdisciplinary expert consensus statement on assessment of pain in older persons. Clin J Pain 2007;23(Suppl):S1–S43. 14. Stolee P, Hillier LM, Esbaugh J et al. Instruments for the assessment of pain in older adults with cognitive impairment. J Am Geriatr Soc 2005;53: 319–326. 15. Hadjistavropoulos T, LaChapelle DL, MacLeod FK et al. Measuring movementexacerbated pain in cognitively impaired frail elders. Clin J Pain 2000;16:54–63. 16.

Crum RM, Anthony JC, Bassett SS et al. Population-based norms for the MiniMental State Examination by age and education level. JAMA 1993;269: 2386–2391. 17. Cullen B, Fahy S, Cunningham CJ et al. Screening for dementia in an Irish community sample using MMSE: A comparison of norm-adjusted versus ? xed cut-points. Int J Geriatr Psychiatry 2005;20:371–376. 18. Keefe FJ, Block AR. Development of an observation method for assessing pain behavior in chronic low back pain patients. Behav Ther 1982;13: 363–375. 19. Weiner D, Pieper C, McConnell E et al.

Pain measurement in elders with chronic low back pain: Traditional and alternative approaches. Pain 1996;67: 461–467. ACKNOWLEDGMENTS Con? ict of Interest: The editor in chief has reviewed the con? ict of interest checklist provided by the authors and has determined that the authors have no ? nancial or any other kind of personal con? icts with this manuscript. This study was supported by Grant R01 NR05069 from the National Institutes of Health, National Institute of Nursing Research 132 HORGAS ET AL. JANUARY 2009–VOL. 57, NO. 1 JAGS 20. Weiner D, Peterson B, Keefe F.

Chronic pain-associated behaviors in the nursing home: Resident versus caregiver perceptions. Pain 1999;80:577–588. 21. Gagliese L, Melzack R. Age-related differences in the qualities but not the intensity of chronic pain. Pain 2003;104:597–608. 22. Melzack R. The McGill Pain Questionnaire: Major properties and scoring methods. Pain 1975;1:277–299. 23. Lichtenstein MJ, Dhanda R. , Cornell JE et al. Disaggregating pain and its effect on physical functional limitations. J Gerontol A Biol Sci Med Sci 1998; 53A:M361–M371. 24. Horgas AL, Nichols AL, Schapson CA et al.

Assessing pain in persons with dementia: Relationships between the NOPPAIN, self-report, and behavioral observations. Pain Manage Nurs 2007;8:77–85. 25. Gibson SJ, Helme RD. Cognitive factors and the experience of pain and suffering in older persons. Pain 2000;85:375–383. 26. Folstein MF, Folstein SE, McHugh PR. ‘‘Mini-mental state’’: A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189–198. 27. Allen RS, Thorn BE, Fisher SE et al. Prescription and dosage of analgesic medication in relation to resident behaviors in the nursing home.

J Am Geriatr Soc 2003;51:534–538. 28. Labus JS, Keefe FJ, Jensen MP. Self-reports of pain intensity and direct observations of pain behavior: When are they correlated? Pain 2003;102: 109–124. 29. Weiner DK. Pain in nursing home residents: What does it really mean, and how can we help? J Am Geriatr Soc 2004;52:1020–1022. 30. Kovach CR, Logan BR, Noonan PE et al. Effects of the serial trial intervention on discomfort and behavior of nursing home residents with dementia. Am J Alzheimers Dis Other Demen 2006;21:147–155.