Reliability and Validity of the Chinese Version of Mild Behavioral Impairment Checklist in Mild Cognitive Impairment and Mild Alzheimer’s Disease

INTRODUCTION

Neuropsychiatric symptoms (NPS), common to all types of neurocognitive disorders [1], are linked to poorer quality of life [2], greater caregiver burden [3], higher rates of institutionalization, more rapid functional decline, added disease burden [4], greater healthcare costs and higher rates of mortality [5], and higher burden of neuropathologic markers of dementia [6]. There is an increasing acknowledgment that NPS form an intrinsic aspect of prodromal dementia and may be an early marker of cognitive decline that can precede or accompany the onset of cognitive symptoms and clinical diagnosis [7, 8]. The concept of mild behavioral impairment (MBI) describes the emergence of NPS in later life and has been proposed as an at-risk state for incident cognitive decline and dementia, and the early manifestation of neurodegenerative disease for some [9, 10].

The NPS Professional Interest Area of the International Society to Advance Alzheimer’s Research and Treatment (ISTAART), a subgroup of the Alzheimer’s Association (AA), formalized the assessment of later life onset NPS as an at-risk state for cognitive decline and dementia with the development of proposed research diagnostic criteria for MBI [9]. MBI is characterized by late-life emergent and persistent NPS in the following domains: 1) decreased motivation, interest and drive (apathy); 2) emotional or affective dysregulation (mood and anxiety symptoms); 3) impulse dyscontrol (agitation, aggression, and abnormal reward salience); 4) social inappropriateness (impaired social cognition); and 5) abnormal thoughts and perception (psychotic symptoms). Importantly, the ISTAART MBI criteria made explicit the relationship between MBI and mild cognitive impairment (MCI) in that MBI can occur following, in concert with, or even in advance of MCI at the subjective cognitive decline (SCD) or normal cognition stages. Notably, MBI could potentially identify an earlier stage of the disease, measured using NPS changes, as opposed to cognitive changes [11].

MBI is common in clinical populations as evidenced by recent studies in cognitive neurology [4] and psychiatric outpatient clinics [12]. Importantly, longitudinal evidence has demonstrated that MBI can be distinguished from psychiatric conditions in older adults, and has a greater risk for incident dementia than chronic and/or recurrent psychiatric disorders in late life [10]. In addition, Johansson et al. found that MBI could be an important early clinical manifestation associated with tau pathology in preclinical AD [13].

In the absence of an MBI-specific case ascertainment instrument, the Mild Behavioral Impairment Checklist (MBI-C) was developed in accordance with the ISTAART-AA MBI criteria, with language and questions geared toward functionally independent community dwelling older adults [14]. The MBI-C includes 34 items, representing the five MBI domains, and can be completed by patient, close informant, or clinician. It was specifically designed to: 1) operationalize the MBI concept; 2) measure a selected list of NPS which may help identify prodromal or preclinical disease; and 3) help predict risk of several dementias, including, but not exclusive to AD. Several studies have validated the MBI-C in MCI, SCD, and normal cognition [15–17]. Mallo et al. found that the MBI-C total scores correlated with NPI-Q and determined the cut-off of defining MBI with 6.5 points in MCI and 8.5 points in SCD [15, 17]. Creese et al. explored the structure of the MBI-C and confirmed the five-factor model was appropriate for its application among cognitively normal older adults [16].

At present, awareness of dementia among Chinese older adults remains low [18]. Mild behavioral problems have been overlooked or underdiagnosed. Identifying NPS in memory clinics where persons with MCI and mild AD are often referred to has been considered essential for timely diagnosis. Therefore, exploring the psychometric properties of MBI-C in MCI and mild AD would be necessary to supplement its application in memory clinics. However, the reliability and validity of the Chinese version of the MBI-C for MCI and mild AD has not been fully investigated.

Our study hypothesized that, when examining the mild behavioral problems in older adults with MCI and mild AD, the Chinese version of the MBI-C in general has satisfactory psychometric properties except for a few culturally relevant items. Therefore, the present study aimed to, firstly, examine the psychometric properties of the Chinese version of the MBI-C in MCI and mild AD; and secondly, to explore the structure of the five factors of the MBI-C in Chinese culture.

METHODS

RESULTS

Demographic and general clinical information of the study participants

A total of 68 MCI (n = 33) and mild AD (n = 35) patient-caregiver dyads completed the assessment with the MBI-C. The two groups were matched for age, gender, and education. As summarized in Table 1, there were no between-group differences in the total scores of NPI-Q, as well as the NPI. The CDR-SOB was significantly higher in mild AD than MCI. Mild AD had poorer performance than MCI on MMSE and MoCA.

OPEN IN VIEWER

Table 1

Demographic and clinical characteristics of study participants

	Total (n = 68)	MCI (n = 33)	Mild AD (n = 35)	t/χ2	p
Age (y)	75.93±6.584	75.09±6.302	76.71±6.837	1.016	0.313
Gender
Male	22 (32.4%)	8 (24.2%)	14 (40.0%)	1.899	0.201
Female	46 (67.6%)	25 (75.8%)	21 (60.0%)
Education (y)	12.99±3.122	13.67±2.890	12.34±3.235	1.775	0.080
NPI-Q total score	3.76±4.293	3.69±4.748	3.82±3.888	0.128	0.899
NPI total score	7.90±11.339	7.67±11.902	8.11±10.951	0.162	0.872
CDR-SOB	1.77±1.176	0.88±0.613	2.61±0.930	9.014	< 0.001
MMSE	23.75±3.565	25.88±2.282	21.80±3.428	5.671	< 0.001
MoCA	18.21±4.430	20.94±3.047	15.71±4.026	5.944	< 0.001

MCI, mild cognitive impairment; Mild AD, mild Alzheimer’s disease; NPI-Q, Neuropsychiatric Inventory Questionnaire; NPI, Neuropsychiatric Inventory; CDR-SOB, Clinical Dementia Rating–Sum of box; MMSE, Mini-Mental State Examination; MoCA, Montreal Cognitive Assessment.

The Cronbach’s alpha coefficient was 0.895 for all participants. As shown in Supplementary Table 1, the inter-item correlation coefficients were satisfactory. Additionally, the MBI-C total score was positively correlated with five domains, including decreased motivation, affective dysregulation, impulse dyscontrol, social inappropriateness, and abnormal perception or thought content (r = 0.577∼0.840, p < 0.001, Table 2).

OPEN IN VIEWER

Table 2

Inter-domain correlation coefficients of the five domains and total scores of MBI-C^*

	DMot	EmD	ImD	SocI	AbPT	Total score
DMot	–
EmD	0.482	–
ImD	0.358	0.620	–
SocI	0.794	0.503	0.564	–
AbPT	ns	0.500	0.446	ns	–
Total MBI-C score	0.698	0.806	0.840	0.767	0.577	–

^*Correlation coefficients reaching statistical significance were presented. All p < 0.01 (2-tailed). ns, non-significant. MBI-C, Mild Behavioral Impairment Checklist; DMot, decreased motivation; ImD, impulse dyscontrol; EmD, emotional dysregulation; SocI, social inappropriateness; AbPT, abnormal perception or thought content.

There was no significant correlation between the MBI-C total score and the CDR-SOB (p > 0.05). As shown in Fig. 1, the frequency of MBI symptoms was similar in MCI and mild AD groups. The severity of the symptoms did not differ between two groups. Among the 34 items, item #16 ([Does the person display sexually disinhibited or intrusive behavior, such as touching (themselves/others), hugging, groping, etc., in a manner that is out of character or may cause offence?] - no response at all) and item #18 ([Does the person display a new recklessness or lack of judgement when driving (e.g., speeding, erratic swerving, abrupt lane changes, etc.)?] – 52.9% response) had low rate of response.

OPEN IN VIEWER

Fig.1

The frequency of MBI-C symptoms presented in the study participants.

The total score of the MBI-C was significantly correlated with the total scores of NPI-Q (r = 0.714, p < 0.001) and NPI (r = 0.749, p < 0.001). The five domain scores of MBI-C correlated significantly with the domain scores of NPI-Q (r = 0.312∼0.673, p < 0.001) and NPI (r = 0.389∼0.673, p < 0.001, see Table 3).

OPEN IN VIEWER

Table 3

Correlations of five domains and total score of MBI-C with the total scores of NPI-Q and NPI^*

	NPI-Q						NPI
MBI-C	Apathy	Emotional problems	Irritability	Disinhibition	Psychosis	Total score	Apathy	Emotional problems	Irritability	Disinhibition	Psychosis	Total score
DMot	0.397	0.307	0.373	ns	ns	0.312	0.514	ns	ns	0.463	ns	0.389
EmD	0.470	0.586	0.550	0.378	0.350	0.626	0.435	0.517	ns	0.427	0.340	0.673
ImD	0.344	0.638	0.605	0.430	0.274	0.673	0.370	0.521	0.300	0.500	0.275	0.640
SocI	0.355	0.412	0.618	0.253	ns	0.450	0.489	ns	ns	0.524	ns	0.450
AbPT	0.265	0.332	0.274	0.256	0.753	0.479	0.265	0.414	0.295	ns	0.745	0.606
total score	0.482	0.662	0.622	0.427	0.323	0.714	0.532	0.553	0.309	0.544	0.323	0.749

^*Correlation coefficients reaching statistical significance were presented. All p < 0.05 (2-tailed). ns, non-significant. MBI-C, Mild Behavioral Impairment Checklist; NPI-Q, Neuropsychiatric Inventory Questionnaire; NPI, Neuropsychiatric Inventory; DMot, decreased motivation; ImD, impulse dyscontrol; EmD, emotional dysregulation; SocI, social inappropriateness; AbPT, abnormal perception or thought content

For the exploratory factor analysis, item #16 and #18 were removed from the prediction model as both items had poor response rate. The components of each factor in Chinese version of MBI-C were slightly different from those of the designing concept (KMO measure of sampling adequacy = 0.695, Bartlett’s Test of Sphericity: Chi-Square=1818.202, df = 496, p < 0.001, Table 4).

OPEN IN VIEWER

Table 4

Exploratory factor analysis of Chinese version of MBI-C and the comparison with the conceptual structure

	Chinese MBI-C					Conceptual structure of MBI-C
Items	DMot	ImD	EmD	SocI	AbPT	DMot	ImD	EmD	SocI	AbPT
Lost motivation on obligations	0.805					X
Talking openly about private matters in public	0.763								X
No longer care about anything	0.761					X
Lack curiosity	0.759					X
Lost interest in friends	0.758					X
Lack social judgement to say or behave	0.745								X
Less spontaneous and active	0.714					X
Less affectionate	0.642					X
Unreasonably argumentative		0.828					X
More impulsive		0.825					X
Agitated/Aggressive		0.772					X
View self as a burden		0.678						X
More stubborn or rigid		0.662					X
Less concerned about words/actions		0.622							X
Unrealistic beliefs about power, wealth or skills		0.373								X
Sadness/Low spirits			0.734					X
Less able to experience pleasure			0.694					X
Change in eating behaviors			0.621				X
Trouble in regulating smoking etc.			0.601				X
Discouraged about future			0.563					X
No longer find food tasteful/enjoyable			0.520				X
Easily frustrated/impatient			0.510				X
Say rude things/sexual remarks				0.769					X
More anxious and worried				0.649				X
Talk to strangers/Intrude strangers				0.490					X
Simple repetitive behaviors				0.480			X
Hoard objects				0.292			X
Complain about seeing things that are not there					0.830					X
Hearing voices/Talking to imaginary people					0.771					X
Suspiciousness about intentions or motives of others					0.662					X
Beliefs of danger/harm					0.659					X
Felt tense / inability to relax					0.593			X

DMot, decreased motivation; ImD, impulse dyscontrol; EmD, emotional dysregulation; SocI, social inappropriateness; AbPT, abnormal perception or thought content.

DISCUSSION

The present study demonstrated the validity and reliability of the MBI-C in MCI and mild AD among Chinese people. Though our study confirmed five domains of the MBI-C, the components of each domain among Chinese participants were slightly different from the original version of the MBI-C. The items concerning sexual disinhibition and judgement of driving had poor response rates and may not be ideal in Chinese culture.

The Cronbach’s alpha indicated that the MBI-C has good structure validity. The finding is consistent with the previous study by Cui et al. [36]. In that study, however, the MBI-C was assessed for utility as test completed by family members as a screening tool to detect AD. The MBI-C inversely correlated with MMSE and MoCA scores, and an MBI-C score cutpoint of 6/7 detected AD with a sensitivity of 87.0% and a sensitivity of 86.0%. In our study, which included MCI and mild AD, the score of MBI-C strongly correlated with the score NPI-Q and NPI demonstrating the ability to detect NPS in this patient population. To our knowledge, only one other study has assessed the MBI-C in Chinese patients without dementia. Fan et al. utilized the Chinese MBI-C in a sample of primary care patients with at most MCI. In this study, MBI-C score predicted frailty, and both frailty and MBI-C were associated with lower MMSE and MoCA scores [37]. Combining all studies, then, the Chinese MBI-C appears to be a suitable tool to detect MBI in MCI populations, and NPS in those with mild AD.

Two MBI-C items had poor response rates. Participating informants did not complete sexual disinhibition question item #16: [Does the person display sexually disinhibited or intrusive behavior, such as touching (themselves/others), hugging, groping, etc., in a manner that is out of character or may cause offence?]. This lack of response might be attributable to Chinese cultural beliefs and values precluding discussions about the sexual behaviors among elderly. Similarly, item #18 [Does the person display a new recklessness or lack of judgement when driving (e.g., speeding, erratic swerving, abrupt lane changes, etc.)?] also had a poor response rate. As driving in old age is uncommon in China, this question may not have good cultural sensitivity for the current generation of elderly (although this may change in the future). The driver’s license of people aged 70 and above is reviewed annually along with the physical health status and tests of memory, judgement, and responsiveness. Whether people with suspected cognitive impairment are permitted to drive remains controversial in China, and more research is required in this area. When comparing our findings with other translations, such as Spanish and Italian, the items related to sexual disinhibition and driving behaviors may not be applicable in Chinese culture [17, 38].

Factor analysis found that most of the factors were consistent with the original concept, but with slight differences. For example, item #26 [talking openly about private matters in public] and item #28 [lacking social judgement to say or behave] belong to the social inappropriateness domain in the original concept [9], but better loaded onto the decreased motivation factor in our study. Social inappropriateness can be a first clinical sign of a neurodegenerative process, especially in behavioral variant frontotemporal dementia [39] and apathy or decreased motivation is also a notable behavior in prodromal dementia of all etiologies [40]. Item structure of emotional dysregulation and impulse dyscontrol had some overlap as well. The items of abnormal perception or thought content remain similar to the original concept. These results differ from a factor analysis from the UK in a large sample of cognitively normal individuals, in which the factor structure of the MBI-C was comparable to the a priori defined domains [16]. The reason for the difference in findings is unclear; however, the populations do not overlap as the UK study was in an online sample of cognitively normal older adults, and our sample consisted of a memory clinic sample of patients with MCI and AD. Future studies will need to explore these issues.

The research findings need to be interpreted with caution as there are several limitations. First, our study sample of MCI was aMCI and probably due to AD, rather than due to other type of neurodegenerative diseases. Social inappropriateness, delusions, and hallucinations may not be prominent features for mild AD. Thus, future studies to investigate the cognitive impairment due to diverse etiology are warranted. Second, while MBI was developed as a construct to capture NPS in preclinical and prodromal populations, our study included prodromal and AD patients, which may confer a different profile of MBI-C items. Further studies on cognitively normal older adults might provide more insight into the validity of the MBI-C in local culture. Besides, there were few differences between the MCI and AD groups with respect to NPS, and in fact, for some NPS, frequency was higher in the MCI group than it was in the AD group. Most studies have found that NPS frequency is higher in AD than in MCI [41], and thus our finding is unusual. This may be due to sample size, or due to the nature of patients referred to clinic. Further exploration is warranted. Finally, mild behavioral impairment may emerge independent of the presence of mild cognitive impairment. Therefore, our study was limited to those overlapping with mild cognitive impairment. It would be important to examine the preclinical phenotype of MBI in future studies.

CONCLUSION

In general, this study supports the validity and reliability of the Chinese version of the MBI-C. The main psychometric properties of the MBI-C confirm its application in clinical practice for behavioral impairment assessment of prodromal dementia in the Chinese population. However, items of sexual disinhibition and driving behaviors need adaptation to sociocultural context. The cultural appropriateness of these behaviors needs further investigation.

AVAILABILITY OF DATA AND MATERIALS

The databases are not freely available to the public because of a lack of specific consent in the ethics committee approval. Still, they will be available from the corresponding author on reasonable request.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE

The institutional review board of Peking University Institute of Mental Health (Sixth Hospital) approved the study. Written informed consent was obtained from each patient-caregiver dyad. The patient and his/her legal guardian both provided written consent for the patient to participate in the study.