BRIEF METHODOLOGICAL REPORTS
Values of the Minimal Clinically Important Difference for the Neuropsychiatric Inventory Questionnaire in Individuals with Dementia Hui-Fen Mao, MS,* Chun-An Kuo, MS,* Wen-Ni Huang, PhD,† Jeffrey L. Cummings, MD,‡ and Tzung-Jeng Hwang, MD§
OBJECTIVES: To estimate the minimal clinically important difference (MCID) for the Neuropsychiatric Inventory Questionnaire (NPI-Q), a widely used measure of behavioral and psychological symptoms of dementia (BPSDs) and associated caregiver stress. DESIGN: Ten registered nurses rated the severity of BPSDs and caregiver distress using the NPI-Q during six monthly assessments and an external reference, a 7-point Likert-type global rating of BPSDs change during five monthly assessments from the second to the sixth month. An anchor-based (global ratings of change) approach and a distribution-based (standard error of measurement) approach were used to determine the MCID for the NPI-Q severity and distress subscales. SETTING: Long-term care facility. PARTICIPANTS: Nonbedridden residents with dementia (n = 45) and registered nurses (n = 10). MEASUREMENTS: NPI-Q (severity and caregiver distress subscales) and global ratings of changes in BPSDs on a 7-point Likert-type scale. RESULTS: The NPI-Q MCID ranges were 2.77 to 3.18 for severity and 3.10 to 3.95 for distress. Residents in the highest NPI-Q tertile at baseline had higher MCID severity (3.62) and distress (5.08) scores than those in the lowest tertile (severity (2.40), distress (3.10)). CONCLUSION: This study provides an estimate of the MCID for severity and distress subscales of the NPI-Q, which can help clinicians and researchers determine whether NPI-Q change scores within a group of individuals with dementia are beyond measurement error and are clinically important. J Am Geriatr Soc 63:1448–1452, 2015. From the *School of Occupational Therapy, College of Medicine, National Taiwan University, Taipei, Taiwan; †Department of Physical Therapy, I-Shou University, Kaohsiung City, Taiwan; ‡Cleveland Clinic, Las Vegas, Nevada; and §Department of Psychiatry, National Taiwan University Hospital and College of Medicine, National Taiwan University, Taipei, Taiwan. Address correspondence to Tzung-Jeng Hwang, Department of Psychiatry, National Taiwan University Hospital, 7 Chung-Shan S. Rd., Taipei 100, Taiwan. E-mail: [email protected] DOI: 10.1111/jgs.13473
JAGS 63:1448–1452, 2015 © 2015, Copyright the Authors Journal compilation © 2015, The American Geriatrics Society
Key words: behavioral and psychological symptoms of dementia (BPSDs); dementia; minimal clinically important difference; Neuropsychiatric Inventory Questionnaire (NPI-Q)
B
ehavioral and psychological symptoms of dementia (BPSDs), such as psychosis, mood disturbances, and other neuropsychiatric symptoms, may occur in up to 80% of long-term care residents with dementia.1–3 These BPSDs are not only distressing for individuals with dementia, but also significantly increase job-related stress, burnout, demands on staff, and staff turnover.4 The severity and frequency of BPSDs are common outcome indicators for pharmacological and nonpharmacological management of BPSDs in long-term care facilities. Many studies have reported significant decreases in the severity of BPSDs, supporting the efficacy of interventions,5–7 but improvement that individuals and caregivers perceive may not be clinically significant.8 The concept of minimal clinically important difference (MCID), defined as “the smallest difference in score in the domain of interest which individuals perceive as beneficial and would mandate, in the absence of troublesome side effects and excessive cost, a change in the patient’s management,”9 was introduced to measure clinically meaningful or important change. A reliable, valid, clinically accessible assessment with defined MCIDs is required to sensitively monitor changes in BPSDs and accurately follow intervention responses. In this study, the MCID for the Chinese version of the Neuropsychiatric Inventory Questionnaire (NPIQ),10 a widely used assessment,11 was estimated.
METHODS Study Design This observational study was conducted in a long-term care facility in Taipei City, Taiwan, from January through
0002-8614/15/$15.00
JAGS
JULY 2015–VOL. 63, NO. 7
June 2012. Ten registered nurses rated the severity of BPSDs and caregiver distress using the NPI-Q during six monthly assessments and an external reference, global rating of BPSD change, during five monthly assessments from the second to sixth months. The National Taiwan University Hospital ethics committee approved the study, and all participants or their caregivers provided written informed consent.
Subjects Sixty-four residents were invited to participate. Forty-five who met the following criteria were included in the study: diagnosed with dementia by psychiatrists or neurologists based on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision;12 not bedridden; and without concurrent major or significant psychiatric disorders such as schizophrenia, bipolar disorder, major depressive disorder, generalized anxiety disorder, posttraumatic stress disorder, or obsessive-compulsive disorder according to physician chart records and direct clinical observation by nurses.
Rating Scales NPI-Q The NPI-Q, a short version of the Neuropsychiatric Inventory (NPI),13 is a brief, self-administered questionnaire on BPSDs and associated caregiver stress. It contains screening questions that cover 12 core symptoms: hallucinations, delusions, agitation and aggression, dysphoria and depression, anxiety, irritability, disinhibition, euphoria, apathy, aberrant motor behavior, sleep and night-time behavior change, and appetite and eating change. Symptoms were checked, and severity was rated on a scale of 0 to 3 (total score 0–36). Caregiver distress was rated on a scale of 0 to 5 for each symptom domain (total score 0–60).
Seven-Point Likert-Type Global Rating of BPSD Change During the follow-up monthly assessment, nurses were asked to rate the question “Has there been any global change in BPSD compared with previous assessments?” on a 7-point Likert scale ( 3 = a great deal worse, 2 = a good deal worse, 1 = a little worse, 0 = no change, 1 = a little better, 2 = a good deal better, 3 = a great deal better). This global rating served as an external criterion for a significant difference in NPI-Q scores.
Raters Ten registered nurses who spent at least 4 workdays per week (mean 40.2 3.5 h/wk) with the individuals were recruited as valid raters. The nurses’ mean age was 42.5 9.4, and they had a mean of 7.6 4.2 years of experience in caring for individuals with dementia. Seventy percent had at least a college degree. Before the beginning of the study, the nurses underwent 3 hours of training on the use of NPI-Q and the 7-point Likert-type global rating of BPSD change. The interrater reliability
MCID OF NPI-Q IN DEMENTIA
1449
was acceptable (NPI-Q severity subscale intraclass correlation coefficient (ICC) = 0.73; distress subscale ICC = 0.67). The monthly assessment was performed after regular shift-change meetings and collection of all relevant information including chart recording and information provided by nursing aids.
Procedure and Data Analysis Studies have suggested that MCIDs are best estimated using the anchor- and distribution-based approaches14,15 and as a range rather than a single value.16,17 The current study used these two approaches to estimate the MCID for the NPI-Q.
Anchor-Based Approach: Global Ratings of Change The anchor-based approach compares changes in NPI-Q measures with other clinical measures such as a global rating of change. The global rating of change is one of the most common methods used to provide the external criterion and a good measure of the significance of difference from an individual perspective.14 The NPI-Q MCID was defined as scores corresponding to 1 (a little better) or 1 (a little worse) on the 7-point Likert scale.18 Because each resident might have more than one score of 1 or 1 during the second to the sixth assessment, only the first scores of 1 or 1 were used for data analysis to avoid the repeated-measures problem. Mean NPI-Q MCID values were calculated and served as the first estimate of MCID for NPI-Q.
Distribution-Based Approach: Standard Error of Measurement Method The distribution-based approach is based on the statistical characteristics of the obtained sample, such as sample variation and measurement precision.14,16 The standard error of measurement (SEM) method, based on measurement precision, considers the possibility that some of the observed change may be due to random measurement error. The SEM was calculated as the square root of the within-subject error variance using test–retest reliability data. Test–retest reliability was determined on a sample of 35 randomly selected participants to whom the assessments were administered twice within 5 to 8 days. An ICC was used to estimate the reliability of the total scores of the severity and distress subscales. The minimal detectable change (MDC) was the amount of change that exceeded the random measurement error of two repeated measures.17 Exceeding the measurement error was considered clinically meaningful change. MDC was calculated as 1.96 9 SEM 9 √2, where 1.96 was the two-sided table z-value for the 95% confidence interval (CI) and √2 accounted for the difference between variances of the two measurements.16,17,19,20 When the change between two repeated measures is greater than or equal to the MDC, it is with 95% confidence that the observed change is real and not caused by measurement error. The calculated MDC served as the second estimate of MCID for NPI-Q. In the sensitivity analysis, the effect of the initial baseline score of the severity subscale on the MCID was
1450
MAO ET AL.
JULY 2015–VOL. 63, NO. 7
JAGS
Table 1. Demographic and Clinical Characteristics of Individuals in the Anchor- and Distribution-Based Method Groups Anchor-Based Method (n = 45)
Characteristic
Distribution-Based Method (n = 35)
Age, mean SD 82.2 6.5 Male, % 62.2 Diagnosis, n (%) Alzheimer’s disease 29 (64.4) Vascular dementia 13 (28.9) Dementia with Lewy bodies 3 (6.7) Years since dementia diagnosed, mean SD 5.5 3.0 Residence in the facility, years, mean SD 4.0 2.7 Mini-Mental State Examination score, mean SD (range 0–30) 9.5 7.6 Activity of daily living score, mean SD (range 0–100) 42.3 26.6 Initial Neuropsychiatric Inventory Questionnaire score, mean SD (range) Severity (range 0–36) 5.20 4.79 (0–23) Distress (range 0–60) 5.93 6.67 (0–33)
81.3 6.3 60.0 23 11 1 4.9 3.6 9.0 37.2
(65.7) (31.4) (2.9) 2.8 2.5 8.0 25.3
5.40 5.15 (0–21) 6.11 7.13 (0–27)
SD = standard deviation.
Table 2. Estimates of Minimal Clinically Important Difference (MCID) of the Severity and Distress Subscales of the Neuropsychiatric Inventory Questionnaire Severity Subscale
Distress Subscale
3.18 (2.15–4.20)
3.95 (2.63–5.27)
1.00 (0.71–1.41)
1.12 (0.85–1.69)
2.77 (1.98–3.91)
3.10 (2.35–4.70)
Method
Anchor-based, MCID estimate (95% CI), n = 40 Distribution-based, n = 35 Standard error of measurement (95% CI) Minimal detectable change (95% CI) CI = confidence interval.
estimated by defining two subgroups. The change scores of individuals with low initial baseline scores (lowest tertile) were compared with the change scores of individuals with high initial baseline scores (highest tertile). SPSS 17.0 for Windows was used for all statistical analyses (SPSS Inc., Chicago, IL).
RESULTS Forty-five residents with dementia were recruited, including 35 involved in the test–retest evaluations. Demographic and clinical characteristics of the participants in the
anchor- and distribution-based methods are summarized in Table 1.
Anchor-Based Method: Global Ratings of Change Of the 225 data items of the global rating of BPSDs and NPI-Q change scores collected from these 45 residents during the five assessments (assessments 2–6), nine data items from three residents were missing because they were hospitalized for other medical conditions. A total of 216 nurse ratings of the NPI-Q were reviewed: 120 “no-change” items (global rating 0), 79 “a little better” items (n = 26, global rating 1) or “a little worse” items (n = 53, global rating 1), 16 “a good deal better” items (n = 3, global rating 2) or “a good deal worse” items (n = 13, global rating 2), and one “a great deal worse” item (n = 1, global rating 3). Forty-one of the 45 patients were rated 1 or 1 at least once during the five global ratings of BPSDs. Data from one of these 41 residents were not analyzed because the resident had only one 1 change score, and it was 3 standard deviations (SDs) greater than the mean change scores of the other 40. Because there were no significant differences in change scores between the 1 (a little worse) and 1 (a little better) data (severity score P = .31, distress score P = .67), these two types were combined for MCID analysis. For the first estimate of the MCID for NPI-Q, the mean NPI-Q change scores were 3.18 (95% CI = 2.15–4.20) for severity and 3.95 (95% CI = 2.63–5.27) for distress (Table 2).
Table 3. Sensitivity Analysis of the Influence of Initial Severity Scores on the Minimal Clinically Important Difference (MCID) of Neuropsychiatric Inventory Questionnaire Initial Severity Subscale Score Mean Score and MCID
0–2 (n = 15)
3–6 (n = 12)
7–23 (n = 13)
Initial severity score, mean (95% CI) Severity subscale, MCID (95% CI) Distress subscale, MCID (95% CI)
1.20 (0.77–1.63) 2.40 (1.15–3.65) 3.40 (1.28–5.52)
4.27 (3.53–5.01) 2.27 (0.625–3.92) 3.17 (1.25–5.08)
10.79 (8.15–13.42) 3.62 (1.79–5.44) 5.08 (2.20–7.96)
BRIEF METHODOLOGICAL REPORTS
Values of the Minimal Clinically Important Difference for the Neuropsychiatric Inventory Questionnaire in Individuals with Dementia Hui-Fen Mao, MS,* Chun-An Kuo, MS,* Wen-Ni Huang, PhD,† Jeffrey L. Cummings, MD,‡ and Tzung-Jeng Hwang, MD§
OBJECTIVES: To estimate the minimal clinically important difference (MCID) for the Neuropsychiatric Inventory Questionnaire (NPI-Q), a widely used measure of behavioral and psychological symptoms of dementia (BPSDs) and associated caregiver stress. DESIGN: Ten registered nurses rated the severity of BPSDs and caregiver distress using the NPI-Q during six monthly assessments and an external reference, a 7-point Likert-type global rating of BPSDs change during five monthly assessments from the second to the sixth month. An anchor-based (global ratings of change) approach and a distribution-based (standard error of measurement) approach were used to determine the MCID for the NPI-Q severity and distress subscales. SETTING: Long-term care facility. PARTICIPANTS: Nonbedridden residents with dementia (n = 45) and registered nurses (n = 10). MEASUREMENTS: NPI-Q (severity and caregiver distress subscales) and global ratings of changes in BPSDs on a 7-point Likert-type scale. RESULTS: The NPI-Q MCID ranges were 2.77 to 3.18 for severity and 3.10 to 3.95 for distress. Residents in the highest NPI-Q tertile at baseline had higher MCID severity (3.62) and distress (5.08) scores than those in the lowest tertile (severity (2.40), distress (3.10)). CONCLUSION: This study provides an estimate of the MCID for severity and distress subscales of the NPI-Q, which can help clinicians and researchers determine whether NPI-Q change scores within a group of individuals with dementia are beyond measurement error and are clinically important. J Am Geriatr Soc 63:1448–1452, 2015. From the *School of Occupational Therapy, College of Medicine, National Taiwan University, Taipei, Taiwan; †Department of Physical Therapy, I-Shou University, Kaohsiung City, Taiwan; ‡Cleveland Clinic, Las Vegas, Nevada; and §Department of Psychiatry, National Taiwan University Hospital and College of Medicine, National Taiwan University, Taipei, Taiwan. Address correspondence to Tzung-Jeng Hwang, Department of Psychiatry, National Taiwan University Hospital, 7 Chung-Shan S. Rd., Taipei 100, Taiwan. E-mail: [email protected] DOI: 10.1111/jgs.13473
JAGS 63:1448–1452, 2015 © 2015, Copyright the Authors Journal compilation © 2015, The American Geriatrics Society
Key words: behavioral and psychological symptoms of dementia (BPSDs); dementia; minimal clinically important difference; Neuropsychiatric Inventory Questionnaire (NPI-Q)
B
ehavioral and psychological symptoms of dementia (BPSDs), such as psychosis, mood disturbances, and other neuropsychiatric symptoms, may occur in up to 80% of long-term care residents with dementia.1–3 These BPSDs are not only distressing for individuals with dementia, but also significantly increase job-related stress, burnout, demands on staff, and staff turnover.4 The severity and frequency of BPSDs are common outcome indicators for pharmacological and nonpharmacological management of BPSDs in long-term care facilities. Many studies have reported significant decreases in the severity of BPSDs, supporting the efficacy of interventions,5–7 but improvement that individuals and caregivers perceive may not be clinically significant.8 The concept of minimal clinically important difference (MCID), defined as “the smallest difference in score in the domain of interest which individuals perceive as beneficial and would mandate, in the absence of troublesome side effects and excessive cost, a change in the patient’s management,”9 was introduced to measure clinically meaningful or important change. A reliable, valid, clinically accessible assessment with defined MCIDs is required to sensitively monitor changes in BPSDs and accurately follow intervention responses. In this study, the MCID for the Chinese version of the Neuropsychiatric Inventory Questionnaire (NPIQ),10 a widely used assessment,11 was estimated.
METHODS Study Design This observational study was conducted in a long-term care facility in Taipei City, Taiwan, from January through
0002-8614/15/$15.00
1452
MAO ET AL.
help clinicians and researchers determine whether change scores on the NPI-Q are beyond measurement error and are clinically important.
ACKNOWLEDGMENTS The authors thank all of the participants and staff of the St. Joseph Home for Alzheimer’s Disease and Related Dementia for their participation and kind assistance. Dr. T-J Hwang has received honoraria from AstraZeneca, Taiwan Otsuka Pharmaceutical Co., and Eli Lilly and Company. Dr. Cummings owns the copyright of the NPI and the NPI-Q and has provided consultation to Abbvie, Acadia, ADAMAS, Alzheon, Anavex, AstraZeneca, Avanir, Biogen-Idec, Biotie, Boehinger-Ingelheim, BristolMyers Squibb, Chase, Eisai, Forum, Genentech, Grifols, Impax, Lilly, Lundbeck, Merck, Neurotrope, Novartis, Nutricia, Otsuka, Pfizer, Prana, QR Pharma, Resverlogix, Roche, Sonexa, Suven, Takeda, and Toyoma companies. Conflict of Interest: The editor in chief has reviewed the conflict of interest checklist provided by the authors and has determined that the authors have no financial or any other kind of personal conflicts with this paper. Author Contributions: Mao: study concept, statistical analysis, first draft of manuscript. Kuo: data acquisition, statistical analyses, interpretation of data. Huang, Hwang, Cummings: study concept, revision of article for important intellectual content. Sponsor’s Role: This study was part of the Smart Sustainable Human-Centric Home project and was supported by Grant NSC 97–2627-E-002–001 from the Taiwan National Science Council. The sponsor did not play any active role in the scientific investigation and reporting of this study, and all authors had complete independence in this study.
REFERENCES 1. Cummings JL, Victoroff JI. Noncognitive neuropsychiatric syndromes in Alzheimer’s disease. Neuropsychiatry Neuropsychol Behav Neurol 1990;3:140–158. 2. Selbæk G, Kirkevold Ø, Engedal K. The prevalence of psychiatric symptoms and behavioural disturbances and the use of psychotropic drugs in Norwegian nursing homes. Int J Geriatr Psychiatry 2007;22:843–849. 3. Zuidema SU, Derksen E, Verhey FR et al. Prevalence of neuropsychiatric symptoms in a large sample of Dutch nursing home patients with dementia. Int J Geriatr Psychiatry 2007;22:632–638. 4. Wetzels R, Zuidema S, Jansen I et al. Course of neuropsychiatric symptoms in residents with dementia in long-term care institutions: A systematic review. Int Psychogeriatr 2010;22:1040–1053. 5. Choi AN, Lee MS, Cheong KJ et al. Effects of group music intervention on behavioral and psychological symptoms in patients with dementia: A pilotcontrolled trial. Int J Neurosci 2009;119:471–481. 6. Chung J, Lai C, Chung P et al. Snoezelen for dementia. Cochrane Database Syst Rev 2002;(4):CD003152. 7. Kim SY, Yoo EY, Jung MY et al. A systematic review of the effects of occupational therapy for persons with dementia: A meta-analysis of randomized controlled trials. NeuroRehabilitation 2012;31:107–115.
JULY 2015–VOL. 63, NO. 7
JAGS
8. Kelleher CJ, Pleil AM, Reese PR et al. How much is enough and who says so? BJOG 2004;111:605–612. 9. Jaeschke R, Singer J, Guyatt GH. Measurement of health status: Ascertaining the minimal clinically important difference. Control Clin Trials 1989;10:407–415. 10. Fuh JL, Liu CK, Mega MS et al. Behavioral disorders and caregivers’ reaction in Taiwanese patients with Alzheimer’s disease. Int Psychogeriatr 2001;13:121–128. 11. Kaufer DI, Cummings JL, Ketchel P et al. Validation of the NPI-Q, a brief clinical form of the Neuropsychiatric Inventory. J Neuropsychiatry Clin Neurosci 2000;12:233–239. 12. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th Ed. Washington, DC: American Psychiatric Association, 2000. 13. Cummings JL, Mega M, Gray K et al. The Neuropsychiatric Inventory: Comprehensive assessment of psychopathology in dementia. Neurology 1994;44:2308–2314. 14. Copay AG, Subach BR, Glassman SD et al. Understanding the minimum clinically important difference: A review of concepts and methods. Spine J 2007;7:541–546. 15. Beaton DE, Boers M, Wells GA. Many faces of the minimal clinically important difference (MCID): A literature review and directions for future research. Curr Opin Rheumatol 2002;14:109–114. 16. Crosby RD, Kolotkin RL, Williams GR. Defining clinically meaningful change in health-related quality of life. J Clin Epidemiol 2003;56:395– 407. 17. van der Roer N, Ostelo RW, Bekkering GE et al. Minimal clinically important change for pain intensity, functional status, and general health status in patients with nonspecific low back pain. Spine 2006;31:578– 582. 18. Goligher EC, Pouchot J, Brant R et al. Minimal clinically important difference for 7 measures of fatigue in patients with systemic lupus erythematosus. J Rheumatol 2008;35:635–642. 19. Stratford PW, Goldsmith CH. Use of the standard error as a reliability index of interest: An applied example using elbow flexor strength data. Phys Ther 1997;77:745–750. 20. Beckerman H, Roebroeck M, Lankhorst G et al. Smallest real difference, a link between reproducibility and responsiveness. Qual Life Res 2001;10:571–578. 21. H€agg O, Fritzell P, Nordwall A. The clinical importance of changes in outcome scores after treatment for chronic low back pain. Eur Spine J 2003;12:12–20. 22. Ringash J, O’Sullivan B, Bezjak A et al. Interpreting clinically significant changes in patient-reported outcomes. Cancer 2007;110:196–202. 23. Boada M, Cejudo J, Tarraga L et al. Neuropsychiatric Inventory Questionnaire (NPI-Q): Spanish validation of an abridged form of the Neuropsychiatric Inventory (NPI). Neurologia 2001;17:317–323. 24. Matsumoto N, Ikeda M, Fukuhara R et al. Validity and reliability of the Japanese version of the Neuropsychiatric Inventory Caregiver Distress Scale (NPI D) and the Neuropsychiatric Inventory Brief Questionnaire Form (NPI-Q). Brain Nerve 2006;58:785–790. 25. Wu L, Wang Y, Li C et al. Reliability and validity of the Chinese version of the Neuropsychiatric Inventory Questionnaire (NPI-Q) in patients with Alzheimer’s disease. Chinese Mental Health J 2010;24:103–107. 26. Forester BP, Oxman TE. Measures to assess the noncognitive symptoms of dementia in the primary care setting. Prim Care Companion J Clin Psychiatry 2003;5:158–163. 27. Ruperto N. Is minimal clinically important difference relevant for the interpretation of clinical trials in pediatric rheumatic diseases? J Rheumatol 2007;34:463–465. 28. Wells G, Beaton D, Shea B et al. Minimal clinically important differences: Review of methods. J Rheumatol 2001;28:406–412. 29. Norman GR, Stratford P, Regehr G. Methodological problems in the retrospective computation of responsiveness to change: The lesson of Cronbach. J Clin Epidemiol 1997;50:869–879. 30. Cook CE. Clinimetrics Corner: The minimal clinically important change score (MCID): A necessary pretense. J Man Manip Ther 2008;16: E82–E83.
Values of the Minimal Clinically Important Difference for the Neuropsychiatric Inventory Questionnaire in Individuals with Dementia Hui-Fen Mao, MS,* Chun-An Kuo, MS,* Wen-Ni Huang, PhD,† Jeffrey L. Cummings, MD,‡ and Tzung-Jeng Hwang, MD§
OBJECTIVES: To estimate the minimal clinically important difference (MCID) for the Neuropsychiatric Inventory Questionnaire (NPI-Q), a widely used measure of behavioral and psychological symptoms of dementia (BPSDs) and associated caregiver stress. DESIGN: Ten registered nurses rated the severity of BPSDs and caregiver distress using the NPI-Q during six monthly assessments and an external reference, a 7-point Likert-type global rating of BPSDs change during five monthly assessments from the second to the sixth month. An anchor-based (global ratings of change) approach and a distribution-based (standard error of measurement) approach were used to determine the MCID for the NPI-Q severity and distress subscales. SETTING: Long-term care facility. PARTICIPANTS: Nonbedridden residents with dementia (n = 45) and registered nurses (n = 10). MEASUREMENTS: NPI-Q (severity and caregiver distress subscales) and global ratings of changes in BPSDs on a 7-point Likert-type scale. RESULTS: The NPI-Q MCID ranges were 2.77 to 3.18 for severity and 3.10 to 3.95 for distress. Residents in the highest NPI-Q tertile at baseline had higher MCID severity (3.62) and distress (5.08) scores than those in the lowest tertile (severity (2.40), distress (3.10)). CONCLUSION: This study provides an estimate of the MCID for severity and distress subscales of the NPI-Q, which can help clinicians and researchers determine whether NPI-Q change scores within a group of individuals with dementia are beyond measurement error and are clinically important. J Am Geriatr Soc 63:1448–1452, 2015. From the *School of Occupational Therapy, College of Medicine, National Taiwan University, Taipei, Taiwan; †Department of Physical Therapy, I-Shou University, Kaohsiung City, Taiwan; ‡Cleveland Clinic, Las Vegas, Nevada; and §Department of Psychiatry, National Taiwan University Hospital and College of Medicine, National Taiwan University, Taipei, Taiwan. Address correspondence to Tzung-Jeng Hwang, Department of Psychiatry, National Taiwan University Hospital, 7 Chung-Shan S. Rd., Taipei 100, Taiwan. E-mail: [email protected] DOI: 10.1111/jgs.13473
JAGS 63:1448–1452, 2015 © 2015, Copyright the Authors Journal compilation © 2015, The American Geriatrics Society
Key words: behavioral and psychological symptoms of dementia (BPSDs); dementia; minimal clinically important difference; Neuropsychiatric Inventory Questionnaire (NPI-Q)
B
ehavioral and psychological symptoms of dementia (BPSDs), such as psychosis, mood disturbances, and other neuropsychiatric symptoms, may occur in up to 80% of long-term care residents with dementia.1–3 These BPSDs are not only distressing for individuals with dementia, but also significantly increase job-related stress, burnout, demands on staff, and staff turnover.4 The severity and frequency of BPSDs are common outcome indicators for pharmacological and nonpharmacological management of BPSDs in long-term care facilities. Many studies have reported significant decreases in the severity of BPSDs, supporting the efficacy of interventions,5–7 but improvement that individuals and caregivers perceive may not be clinically significant.8 The concept of minimal clinically important difference (MCID), defined as “the smallest difference in score in the domain of interest which individuals perceive as beneficial and would mandate, in the absence of troublesome side effects and excessive cost, a change in the patient’s management,”9 was introduced to measure clinically meaningful or important change. A reliable, valid, clinically accessible assessment with defined MCIDs is required to sensitively monitor changes in BPSDs and accurately follow intervention responses. In this study, the MCID for the Chinese version of the Neuropsychiatric Inventory Questionnaire (NPIQ),10 a widely used assessment,11 was estimated.
METHODS Study Design This observational study was conducted in a long-term care facility in Taipei City, Taiwan, from January through
0002-8614/15/$15.00
JAGS
JULY 2015–VOL. 63, NO. 7
June 2012. Ten registered nurses rated the severity of BPSDs and caregiver distress using the NPI-Q during six monthly assessments and an external reference, global rating of BPSD change, during five monthly assessments from the second to sixth months. The National Taiwan University Hospital ethics committee approved the study, and all participants or their caregivers provided written informed consent.
Subjects Sixty-four residents were invited to participate. Forty-five who met the following criteria were included in the study: diagnosed with dementia by psychiatrists or neurologists based on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision;12 not bedridden; and without concurrent major or significant psychiatric disorders such as schizophrenia, bipolar disorder, major depressive disorder, generalized anxiety disorder, posttraumatic stress disorder, or obsessive-compulsive disorder according to physician chart records and direct clinical observation by nurses.
Rating Scales NPI-Q The NPI-Q, a short version of the Neuropsychiatric Inventory (NPI),13 is a brief, self-administered questionnaire on BPSDs and associated caregiver stress. It contains screening questions that cover 12 core symptoms: hallucinations, delusions, agitation and aggression, dysphoria and depression, anxiety, irritability, disinhibition, euphoria, apathy, aberrant motor behavior, sleep and night-time behavior change, and appetite and eating change. Symptoms were checked, and severity was rated on a scale of 0 to 3 (total score 0–36). Caregiver distress was rated on a scale of 0 to 5 for each symptom domain (total score 0–60).
Seven-Point Likert-Type Global Rating of BPSD Change During the follow-up monthly assessment, nurses were asked to rate the question “Has there been any global change in BPSD compared with previous assessments?” on a 7-point Likert scale ( 3 = a great deal worse, 2 = a good deal worse, 1 = a little worse, 0 = no change, 1 = a little better, 2 = a good deal better, 3 = a great deal better). This global rating served as an external criterion for a significant difference in NPI-Q scores.
Raters Ten registered nurses who spent at least 4 workdays per week (mean 40.2 3.5 h/wk) with the individuals were recruited as valid raters. The nurses’ mean age was 42.5 9.4, and they had a mean of 7.6 4.2 years of experience in caring for individuals with dementia. Seventy percent had at least a college degree. Before the beginning of the study, the nurses underwent 3 hours of training on the use of NPI-Q and the 7-point Likert-type global rating of BPSD change. The interrater reliability
MCID OF NPI-Q IN DEMENTIA
1449
was acceptable (NPI-Q severity subscale intraclass correlation coefficient (ICC) = 0.73; distress subscale ICC = 0.67). The monthly assessment was performed after regular shift-change meetings and collection of all relevant information including chart recording and information provided by nursing aids.
Procedure and Data Analysis Studies have suggested that MCIDs are best estimated using the anchor- and distribution-based approaches14,15 and as a range rather than a single value.16,17 The current study used these two approaches to estimate the MCID for the NPI-Q.
Anchor-Based Approach: Global Ratings of Change The anchor-based approach compares changes in NPI-Q measures with other clinical measures such as a global rating of change. The global rating of change is one of the most common methods used to provide the external criterion and a good measure of the significance of difference from an individual perspective.14 The NPI-Q MCID was defined as scores corresponding to 1 (a little better) or 1 (a little worse) on the 7-point Likert scale.18 Because each resident might have more than one score of 1 or 1 during the second to the sixth assessment, only the first scores of 1 or 1 were used for data analysis to avoid the repeated-measures problem. Mean NPI-Q MCID values were calculated and served as the first estimate of MCID for NPI-Q.
Distribution-Based Approach: Standard Error of Measurement Method The distribution-based approach is based on the statistical characteristics of the obtained sample, such as sample variation and measurement precision.14,16 The standard error of measurement (SEM) method, based on measurement precision, considers the possibility that some of the observed change may be due to random measurement error. The SEM was calculated as the square root of the within-subject error variance using test–retest reliability data. Test–retest reliability was determined on a sample of 35 randomly selected participants to whom the assessments were administered twice within 5 to 8 days. An ICC was used to estimate the reliability of the total scores of the severity and distress subscales. The minimal detectable change (MDC) was the amount of change that exceeded the random measurement error of two repeated measures.17 Exceeding the measurement error was considered clinically meaningful change. MDC was calculated as 1.96 9 SEM 9 √2, where 1.96 was the two-sided table z-value for the 95% confidence interval (CI) and √2 accounted for the difference between variances of the two measurements.16,17,19,20 When the change between two repeated measures is greater than or equal to the MDC, it is with 95% confidence that the observed change is real and not caused by measurement error. The calculated MDC served as the second estimate of MCID for NPI-Q. In the sensitivity analysis, the effect of the initial baseline score of the severity subscale on the MCID was
1450
MAO ET AL.
JULY 2015–VOL. 63, NO. 7
JAGS
Table 1. Demographic and Clinical Characteristics of Individuals in the Anchor- and Distribution-Based Method Groups Anchor-Based Method (n = 45)
Characteristic
Distribution-Based Method (n = 35)
Age, mean SD 82.2 6.5 Male, % 62.2 Diagnosis, n (%) Alzheimer’s disease 29 (64.4) Vascular dementia 13 (28.9) Dementia with Lewy bodies 3 (6.7) Years since dementia diagnosed, mean SD 5.5 3.0 Residence in the facility, years, mean SD 4.0 2.7 Mini-Mental State Examination score, mean SD (range 0–30) 9.5 7.6 Activity of daily living score, mean SD (range 0–100) 42.3 26.6 Initial Neuropsychiatric Inventory Questionnaire score, mean SD (range) Severity (range 0–36) 5.20 4.79 (0–23) Distress (range 0–60) 5.93 6.67 (0–33)
81.3 6.3 60.0 23 11 1 4.9 3.6 9.0 37.2
(65.7) (31.4) (2.9) 2.8 2.5 8.0 25.3
5.40 5.15 (0–21) 6.11 7.13 (0–27)
SD = standard deviation.
Table 2. Estimates of Minimal Clinically Important Difference (MCID) of the Severity and Distress Subscales of the Neuropsychiatric Inventory Questionnaire Severity Subscale
Distress Subscale
3.18 (2.15–4.20)
3.95 (2.63–5.27)
1.00 (0.71–1.41)
1.12 (0.85–1.69)
2.77 (1.98–3.91)
3.10 (2.35–4.70)
Method
Anchor-based, MCID estimate (95% CI), n = 40 Distribution-based, n = 35 Standard error of measurement (95% CI) Minimal detectable change (95% CI) CI = confidence interval.
estimated by defining two subgroups. The change scores of individuals with low initial baseline scores (lowest tertile) were compared with the change scores of individuals with high initial baseline scores (highest tertile). SPSS 17.0 for Windows was used for all statistical analyses (SPSS Inc., Chicago, IL).
RESULTS Forty-five residents with dementia were recruited, including 35 involved in the test–retest evaluations. Demographic and clinical characteristics of the participants in the
anchor- and distribution-based methods are summarized in Table 1.
Anchor-Based Method: Global Ratings of Change Of the 225 data items of the global rating of BPSDs and NPI-Q change scores collected from these 45 residents during the five assessments (assessments 2–6), nine data items from three residents were missing because they were hospitalized for other medical conditions. A total of 216 nurse ratings of the NPI-Q were reviewed: 120 “no-change” items (global rating 0), 79 “a little better” items (n = 26, global rating 1) or “a little worse” items (n = 53, global rating 1), 16 “a good deal better” items (n = 3, global rating 2) or “a good deal worse” items (n = 13, global rating 2), and one “a great deal worse” item (n = 1, global rating 3). Forty-one of the 45 patients were rated 1 or 1 at least once during the five global ratings of BPSDs. Data from one of these 41 residents were not analyzed because the resident had only one 1 change score, and it was 3 standard deviations (SDs) greater than the mean change scores of the other 40. Because there were no significant differences in change scores between the 1 (a little worse) and 1 (a little better) data (severity score P = .31, distress score P = .67), these two types were combined for MCID analysis. For the first estimate of the MCID for NPI-Q, the mean NPI-Q change scores were 3.18 (95% CI = 2.15–4.20) for severity and 3.95 (95% CI = 2.63–5.27) for distress (Table 2).
Table 3. Sensitivity Analysis of the Influence of Initial Severity Scores on the Minimal Clinically Important Difference (MCID) of Neuropsychiatric Inventory Questionnaire Initial Severity Subscale Score Mean Score and MCID
0–2 (n = 15)
3–6 (n = 12)
7–23 (n = 13)
Initial severity score, mean (95% CI) Severity subscale, MCID (95% CI) Distress subscale, MCID (95% CI)
1.20 (0.77–1.63) 2.40 (1.15–3.65) 3.40 (1.28–5.52)
4.27 (3.53–5.01) 2.27 (0.625–3.92) 3.17 (1.25–5.08)
10.79 (8.15–13.42) 3.62 (1.79–5.44) 5.08 (2.20–7.96)
BRIEF METHODOLOGICAL REPORTS
Values of the Minimal Clinically Important Difference for the Neuropsychiatric Inventory Questionnaire in Individuals with Dementia Hui-Fen Mao, MS,* Chun-An Kuo, MS,* Wen-Ni Huang, PhD,† Jeffrey L. Cummings, MD,‡ and Tzung-Jeng Hwang, MD§
OBJECTIVES: To estimate the minimal clinically important difference (MCID) for the Neuropsychiatric Inventory Questionnaire (NPI-Q), a widely used measure of behavioral and psychological symptoms of dementia (BPSDs) and associated caregiver stress. DESIGN: Ten registered nurses rated the severity of BPSDs and caregiver distress using the NPI-Q during six monthly assessments and an external reference, a 7-point Likert-type global rating of BPSDs change during five monthly assessments from the second to the sixth month. An anchor-based (global ratings of change) approach and a distribution-based (standard error of measurement) approach were used to determine the MCID for the NPI-Q severity and distress subscales. SETTING: Long-term care facility. PARTICIPANTS: Nonbedridden residents with dementia (n = 45) and registered nurses (n = 10). MEASUREMENTS: NPI-Q (severity and caregiver distress subscales) and global ratings of changes in BPSDs on a 7-point Likert-type scale. RESULTS: The NPI-Q MCID ranges were 2.77 to 3.18 for severity and 3.10 to 3.95 for distress. Residents in the highest NPI-Q tertile at baseline had higher MCID severity (3.62) and distress (5.08) scores than those in the lowest tertile (severity (2.40), distress (3.10)). CONCLUSION: This study provides an estimate of the MCID for severity and distress subscales of the NPI-Q, which can help clinicians and researchers determine whether NPI-Q change scores within a group of individuals with dementia are beyond measurement error and are clinically important. J Am Geriatr Soc 63:1448–1452, 2015. From the *School of Occupational Therapy, College of Medicine, National Taiwan University, Taipei, Taiwan; †Department of Physical Therapy, I-Shou University, Kaohsiung City, Taiwan; ‡Cleveland Clinic, Las Vegas, Nevada; and §Department of Psychiatry, National Taiwan University Hospital and College of Medicine, National Taiwan University, Taipei, Taiwan. Address correspondence to Tzung-Jeng Hwang, Department of Psychiatry, National Taiwan University Hospital, 7 Chung-Shan S. Rd., Taipei 100, Taiwan. E-mail: [email protected] DOI: 10.1111/jgs.13473
JAGS 63:1448–1452, 2015 © 2015, Copyright the Authors Journal compilation © 2015, The American Geriatrics Society
Key words: behavioral and psychological symptoms of dementia (BPSDs); dementia; minimal clinically important difference; Neuropsychiatric Inventory Questionnaire (NPI-Q)
B
ehavioral and psychological symptoms of dementia (BPSDs), such as psychosis, mood disturbances, and other neuropsychiatric symptoms, may occur in up to 80% of long-term care residents with dementia.1–3 These BPSDs are not only distressing for individuals with dementia, but also significantly increase job-related stress, burnout, demands on staff, and staff turnover.4 The severity and frequency of BPSDs are common outcome indicators for pharmacological and nonpharmacological management of BPSDs in long-term care facilities. Many studies have reported significant decreases in the severity of BPSDs, supporting the efficacy of interventions,5–7 but improvement that individuals and caregivers perceive may not be clinically significant.8 The concept of minimal clinically important difference (MCID), defined as “the smallest difference in score in the domain of interest which individuals perceive as beneficial and would mandate, in the absence of troublesome side effects and excessive cost, a change in the patient’s management,”9 was introduced to measure clinically meaningful or important change. A reliable, valid, clinically accessible assessment with defined MCIDs is required to sensitively monitor changes in BPSDs and accurately follow intervention responses. In this study, the MCID for the Chinese version of the Neuropsychiatric Inventory Questionnaire (NPIQ),10 a widely used assessment,11 was estimated.
METHODS Study Design This observational study was conducted in a long-term care facility in Taipei City, Taiwan, from January through
0002-8614/15/$15.00
1452
MAO ET AL.
help clinicians and researchers determine whether change scores on the NPI-Q are beyond measurement error and are clinically important.
ACKNOWLEDGMENTS The authors thank all of the participants and staff of the St. Joseph Home for Alzheimer’s Disease and Related Dementia for their participation and kind assistance. Dr. T-J Hwang has received honoraria from AstraZeneca, Taiwan Otsuka Pharmaceutical Co., and Eli Lilly and Company. Dr. Cummings owns the copyright of the NPI and the NPI-Q and has provided consultation to Abbvie, Acadia, ADAMAS, Alzheon, Anavex, AstraZeneca, Avanir, Biogen-Idec, Biotie, Boehinger-Ingelheim, BristolMyers Squibb, Chase, Eisai, Forum, Genentech, Grifols, Impax, Lilly, Lundbeck, Merck, Neurotrope, Novartis, Nutricia, Otsuka, Pfizer, Prana, QR Pharma, Resverlogix, Roche, Sonexa, Suven, Takeda, and Toyoma companies. Conflict of Interest: The editor in chief has reviewed the conflict of interest checklist provided by the authors and has determined that the authors have no financial or any other kind of personal conflicts with this paper. Author Contributions: Mao: study concept, statistical analysis, first draft of manuscript. Kuo: data acquisition, statistical analyses, interpretation of data. Huang, Hwang, Cummings: study concept, revision of article for important intellectual content. Sponsor’s Role: This study was part of the Smart Sustainable Human-Centric Home project and was supported by Grant NSC 97–2627-E-002–001 from the Taiwan National Science Council. The sponsor did not play any active role in the scientific investigation and reporting of this study, and all authors had complete independence in this study.
REFERENCES 1. Cummings JL, Victoroff JI. Noncognitive neuropsychiatric syndromes in Alzheimer’s disease. Neuropsychiatry Neuropsychol Behav Neurol 1990;3:140–158. 2. Selbæk G, Kirkevold Ø, Engedal K. The prevalence of psychiatric symptoms and behavioural disturbances and the use of psychotropic drugs in Norwegian nursing homes. Int J Geriatr Psychiatry 2007;22:843–849. 3. Zuidema SU, Derksen E, Verhey FR et al. Prevalence of neuropsychiatric symptoms in a large sample of Dutch nursing home patients with dementia. Int J Geriatr Psychiatry 2007;22:632–638. 4. Wetzels R, Zuidema S, Jansen I et al. Course of neuropsychiatric symptoms in residents with dementia in long-term care institutions: A systematic review. Int Psychogeriatr 2010;22:1040–1053. 5. Choi AN, Lee MS, Cheong KJ et al. Effects of group music intervention on behavioral and psychological symptoms in patients with dementia: A pilotcontrolled trial. Int J Neurosci 2009;119:471–481. 6. Chung J, Lai C, Chung P et al. Snoezelen for dementia. Cochrane Database Syst Rev 2002;(4):CD003152. 7. Kim SY, Yoo EY, Jung MY et al. A systematic review of the effects of occupational therapy for persons with dementia: A meta-analysis of randomized controlled trials. NeuroRehabilitation 2012;31:107–115.
JULY 2015–VOL. 63, NO. 7
JAGS
8. Kelleher CJ, Pleil AM, Reese PR et al. How much is enough and who says so? BJOG 2004;111:605–612. 9. Jaeschke R, Singer J, Guyatt GH. Measurement of health status: Ascertaining the minimal clinically important difference. Control Clin Trials 1989;10:407–415. 10. Fuh JL, Liu CK, Mega MS et al. Behavioral disorders and caregivers’ reaction in Taiwanese patients with Alzheimer’s disease. Int Psychogeriatr 2001;13:121–128. 11. Kaufer DI, Cummings JL, Ketchel P et al. Validation of the NPI-Q, a brief clinical form of the Neuropsychiatric Inventory. J Neuropsychiatry Clin Neurosci 2000;12:233–239. 12. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th Ed. Washington, DC: American Psychiatric Association, 2000. 13. Cummings JL, Mega M, Gray K et al. The Neuropsychiatric Inventory: Comprehensive assessment of psychopathology in dementia. Neurology 1994;44:2308–2314. 14. Copay AG, Subach BR, Glassman SD et al. Understanding the minimum clinically important difference: A review of concepts and methods. Spine J 2007;7:541–546. 15. Beaton DE, Boers M, Wells GA. Many faces of the minimal clinically important difference (MCID): A literature review and directions for future research. Curr Opin Rheumatol 2002;14:109–114. 16. Crosby RD, Kolotkin RL, Williams GR. Defining clinically meaningful change in health-related quality of life. J Clin Epidemiol 2003;56:395– 407. 17. van der Roer N, Ostelo RW, Bekkering GE et al. Minimal clinically important change for pain intensity, functional status, and general health status in patients with nonspecific low back pain. Spine 2006;31:578– 582. 18. Goligher EC, Pouchot J, Brant R et al. Minimal clinically important difference for 7 measures of fatigue in patients with systemic lupus erythematosus. J Rheumatol 2008;35:635–642. 19. Stratford PW, Goldsmith CH. Use of the standard error as a reliability index of interest: An applied example using elbow flexor strength data. Phys Ther 1997;77:745–750. 20. Beckerman H, Roebroeck M, Lankhorst G et al. Smallest real difference, a link between reproducibility and responsiveness. Qual Life Res 2001;10:571–578. 21. H€agg O, Fritzell P, Nordwall A. The clinical importance of changes in outcome scores after treatment for chronic low back pain. Eur Spine J 2003;12:12–20. 22. Ringash J, O’Sullivan B, Bezjak A et al. Interpreting clinically significant changes in patient-reported outcomes. Cancer 2007;110:196–202. 23. Boada M, Cejudo J, Tarraga L et al. Neuropsychiatric Inventory Questionnaire (NPI-Q): Spanish validation of an abridged form of the Neuropsychiatric Inventory (NPI). Neurologia 2001;17:317–323. 24. Matsumoto N, Ikeda M, Fukuhara R et al. Validity and reliability of the Japanese version of the Neuropsychiatric Inventory Caregiver Distress Scale (NPI D) and the Neuropsychiatric Inventory Brief Questionnaire Form (NPI-Q). Brain Nerve 2006;58:785–790. 25. Wu L, Wang Y, Li C et al. Reliability and validity of the Chinese version of the Neuropsychiatric Inventory Questionnaire (NPI-Q) in patients with Alzheimer’s disease. Chinese Mental Health J 2010;24:103–107. 26. Forester BP, Oxman TE. Measures to assess the noncognitive symptoms of dementia in the primary care setting. Prim Care Companion J Clin Psychiatry 2003;5:158–163. 27. Ruperto N. Is minimal clinically important difference relevant for the interpretation of clinical trials in pediatric rheumatic diseases? J Rheumatol 2007;34:463–465. 28. Wells G, Beaton D, Shea B et al. Minimal clinically important differences: Review of methods. J Rheumatol 2001;28:406–412. 29. Norman GR, Stratford P, Regehr G. Methodological problems in the retrospective computation of responsiveness to change: The lesson of Cronbach. J Clin Epidemiol 1997;50:869–879. 30. Cook CE. Clinimetrics Corner: The minimal clinically important change score (MCID): A necessary pretense. J Man Manip Ther 2008;16: E82–E83.
