The review of the forward–backward translation addressed several improvements to enhance the accuracy of the Malay-translated version as the original version without compromising the validity and reliability. Throughout the discussion and review with expert panels and a potential user group, all items (n = 26) in the four domains were retained as they were deemed important and appropriate.
Several words and phrases including strok for stroke, stres for stress, aktiviti sederhana berat for moderately intense activity, alkohol for alcohol and kolesterol buruk for bad cholesterol were selected after considering the usage of the words and phrases in the Malaysian scenario to define the correct meaning.
Moreover, the domains ‘perceived benefits and intentions to change’ and ‘healthy eating intentions’ have been revised to simplify the sequence of the questions, thereby making them more comprehensible and balanced. The three items from the domain ‘perceived benefits and intentions to change’, items 17, 18 and 21, were joined together with the three items in the domain ‘healthy eating intentions’. As a result, the third domain is renamed as ‘perceived benefits’ with four items, and the last domain is the ‘intention to change’ with six items. The details of the translation are attached in the appendix.
The CVI and FVI of the ABCD-M risk questionnaire were 0.94 and 0.99, respectively (Tables 1 and 2). Both parameters indicate that all items in the questionnaire are relevant to the domain, clear and understandable for the intended study population.
Table 1
Content validity index: the relevance ratings on the item scale by six experts
Item
|
Ea1
|
E2
|
E3
|
E4
|
E5
|
E6
|
I-CVIb
|
Q1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q2
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q3
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q4
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q5
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q6
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q7
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q8
|
0
|
1
|
1
|
1
|
1
|
1
|
0.83
|
Q9
|
1
|
0
|
1
|
1
|
0
|
1
|
0.67
|
Q10
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q11
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q12
|
1
|
0
|
1
|
1
|
1
|
1
|
0.83
|
Q13
|
1
|
0
|
1
|
1
|
1
|
1
|
0.83
|
Q14
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q15
|
1
|
1
|
1
|
1
|
0
|
1
|
0.83
|
Q16
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q17
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q18
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q19
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q20
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q21
|
1
|
0
|
1
|
1
|
1
|
1
|
0.83
|
Q22
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q23
|
0
|
1
|
1
|
1
|
1
|
0
|
0.67
|
Q24
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q25
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q26
|
1
|
1
|
1
|
1
|
1
|
0
|
0.83
|
Content validity index average
|
0.94
|
Note: a=expert; b=item-level content validity index |
Table 2
Face validity index: the clarity and comprehension ratings on the item scale by 10 users
Item
|
Ra1
|
R2
|
R3
|
R4
|
R5
|
R6
|
R7
|
R8
|
R9
|
R10
|
I-FVIb
|
Q1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q2
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q3
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q4
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q5
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q6
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q7
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q8
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q9
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
0
|
1
|
1
|
0.9
|
Q10
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q11
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q12
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q13
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q14
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q15
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q16
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q17
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q18
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q19
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q20
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q21
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q22
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q23
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q24
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q25
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Q26
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
1
|
Face validity index average
|
0.99
|
Note: a=respondent; b=item-level face validity index |
The construct validity and reliability testings were conducted using 179 samples from the targeted population who responded to the adapted and translated questionnaire. The age of the respondents ranged from 18 to 66 years, with a mean of 36.82 (SD, 12.17). Most of the respondents were married (n = 133, 75.14%), followed by the statuses single (n = 39, 22.03%) and divorced (n = 5, 2.82%). The respondents were predominantly female (n = 122, 68.16%). More than half of the respondents attended up to the secondary level of education (n = 91, 51.12%), followed by the tertiary (n = 82, 46.07%) and primary (n = 5, 2.81%) levels of education. The highest number of respondents worked for the government (n = 56, 31.28%), followed by self-employed (n = 39, 21.79%), housewife (n = 31, 17.32%), working in a private sector (n = 25, 13.97%), student (n = 16, 8.94%) and unemployed (n = 12, 6.7%). The distribution of the monthly income showed that 50.31% (n = 81) of the respondents earned RM1001 to RM3000 per month, 25.47% (n = 41) earned less than RM1000 per month, 13.66% (n = 22) earned RM3001 to RM5000 per month and 10.56% (n = 17) more than RM5000.
The overall result of the ABCD-M revealed that the range was between 6 and 77 (a total of 80 marks) with a mean of 49.26 (61.58%). All domains recorded more than 50% of the full marks, except for the ‘perceived risk of heart attack/stroke’. The domain exhibited a mean of 14.97 (SD, 6.01) (46.78% of the total marks). The sociodemographic characteristics of the respondents are listed in Table 3.
Table 3
Sociodemographic characteristics of the respondents
Variables
|
n (%)
|
Mean (SD)
|
Range
|
Age
|
-
|
36.82 (12.17)
|
18 – 66
|
Gender
Male
Female
|
57 (31.84)
122 (68.16)
|
-
|
-
|
Status
Single
Married
Divorced
|
39 (22.03)
133 (75.14)
5 (2.82)
|
-
|
-
|
Education
Primary
Secondary
College/Uni
|
5 (2.81)
91 (51.12)
82 (46.07)
|
-
|
-
|
Occupation
Unemployed
Student
Housewife
Self-employed
Government
Private
|
12 (6.70)
16 (8.94)
31 (17.32)
39 (21.79)
56 (31.28)
25 (13.97)
|
-
|
-
|
Income
<RM1000
RM1001 – RM3000
RM3001 – RM5000
>RM5000
|
41 (25.47)
81 (50.31)
22 (13.66)
17 (10.56)
|
-
|
-
|
Awareness (80 marks)
Knowledge (8 marks)
Perceived Risk (32 marks)
Perceived Benefits (16 marks)
Intention to Change (24 marks)
|
49.26 (9.31)
5.82 (1.58)
14.97 (6.01)
12.26 (2.86)
16.21 (3.43)
|
6 – 77
-
-
-
-
|
The EFA iteration confirmed the FLs and reliabilities as reported in Table 4. Most of the items had good FLs (>0.50) and low complexity, except for item 15 under the domain ‘perceived risk of heart attack/stroke’, which had low FL (0.26) and communality (0.122) and high complexity (2.06). In addition, almost all items had good communality values ranging from 0.32 to 0.94 [17, 40].
The internal consistency reliability of the structure was measured using Cronbach’s alpha. The α values of the domains ‘perceived risk of heart attack/stroke’ (eight items), ‘perceived benefits’ (four items) and ‘intention to change’ (six items) were 0.876, 0.806 and 0.696, respectively. Thus, all domains were above the minimum threshold of reliability of 0.70 [29].
On the other hand, Raykov’s rho of the CFA for each domain was good [32, 41], which ranged from 0.643 to 0.879. The α values of the domains ‘perceived risk of heart attack/stroke’, ‘perceived benefits’ and ‘intention to change’ were 0.885, 0.766 and 0.643, respectively. The correlations between the domains were <0.85, which indicated the absence of multicollinearity between the items [38]. Hence, the parallel analysis showed that three domains as used in the original questionnaire would be retained. The FLs of each item and correlations between the domains are illustrated in Figure 2.
Table 4
Factor analysis, internal consistency and composite reliability
Domain
|
Item
|
Factor Loading
|
Communality
|
Cronbach’s Alpha
|
Raykov’s Rho
|
Perceived Risk
|
9
10
11
12
13
14
15
16
|
0.747
0.882
0.939
0.822
0.761
0.690
0.263
0.539
|
0.589
0.774
0.860
0.670
0.600
0.510
0.122
0.320
|
0.876
|
0.885
|
Perceived Benefits
|
17
18
19
20
|
0.710
0.666
0.659
0.668
|
0.512
0.465
0.439
0.449
|
0.806
|
0.766
|
Intention to Change
|
21
22
23
24
25
26
|
0.777
0.722
0.970
0.738
0.728
0.892
|
0.609
0.533
0.940
0.544
0.529
0.797
|
0.696
|
0.643
|
Correlation:
Perceived Risk ↔ Perceived Benefits r = 0.141
Perceived Risk ↔ Intention to Change r = 0.107
Perceived Benefits ↔ Intention to Change r = 0.045
|
The fit indices from the CFA resulted in good goodness of fit (SRMR, 0.054; RMSEA, 0.029; CFI, 0.99; TLI, 0.99) (Table 5). The SRMR and RMSEA values were clearly below the cut-off value of 0.08, while the support from good CFI and TLI values were more than 0.95. The p-value for the chi-square statistic (χ2 [98] = 119.81) was not significant (p-value = 0.16), indicating the good model fit. The translated version (model 1) was compared with the modified version (model 2) with the value of the chi-square (χ2 [9] = 10.37, p-value = 0.32). Model 2 consisted of 17 items because of the elimination of item 15 under the domain ‘perceived risk of heart attack/stroke’. Both were fitted well models, with model 2 having a slightly lower AIC of 6307.1 and BIC of 6511.1 compared with model 1.
Table 5
Fit indices of the models
Model
|
χ2 (df)
|
p-value
|
SRMR
|
RMSEA
|
90% CI
|
CFI
|
TLI
|
AIC
|
BIC
|
1
|
119.81 (98)
|
0.16
|
0.054
|
0.029
|
0, 0.052
|
0.99
|
0.99
|
6851.8
|
7084.5
|
2
|
101.30 (89)
|
0.18
|
0.050
|
0.029
|
0, 0.053
|
0.99
|
0.99
|
6307.1
|
6511.1
|
Note: Model 1: Original version of 18 items constitutes of domains; 'Perceived Risk of Heart Attack/Stroke,' 'Perceived Benefits,' and 'Intention to Change.' Model 2: Modified version of 17 items with the elimination of item-15th under the 'Perceived Risk of Heart Attack/Stroke' domain. |
Model-to-model comparison: χ2 diff. (df) = 10.37 (9), p-value = 0.32. |
SRMR = standardized root mean square residual; RMSEA = root mean square error of approximation; CI = confidence interval; CFI = comparative fit index; TLI = Tucker-Lewis fit index; AIC = Akaike information criterion; BIC: Bayesian information criterion. |