• diabetes;
  • osteoporosis health beliefs;
  • reliability;
  • validity


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Conflict of Interest
  10. References


The aims of this study were to translate and examine the psychometric properties of the Malaysian version of the Osteoporosis Health Belief Scale (OHBS-M) among type 2 diabetes patients (T2DM) and to assess the correlation between osteoporosis knowledge, health belief and self-efficacy scales, as well as assess the osteoporosis risk in the sample population using quantitative ultrasound measurement (QUS).


A standard ‘forward–backward’ procedure was used to translate OHBS into the Malay language, which was then validated with a convenience sample of 250 T2DM. Bone mineral density (BMD) measurements were carried out using QUS at the calcaneus.


The mean score of OHBS-M was 158.31 ± 20.80. The Fleiss' kappa, content validity ratio range and content validity index were 0.99, 0.75–1.00 and 0.88, respectively. Seven factors of the OHBS-M were identified using exploratory factor analysis and were confirmed through confirmatory factor analysis. Internal consistency and test–retest reliability values were 0.89 and 0.555, respectively. In addition, only 22% had a normal BMD (low risk of abnormal BMD), while osteopenia and osteoporosis were 57.6% and 20.4% (considered as high risk of abnormal BMD), respectively.


The results showed that the OHBS-M is a reliable and valid instrument for measuring health belief toward osteoporosis in diabetic patients. In addition, it is an appropriate tool to identify patients needing a bone health-promoting intervention regarding lifestyle behavior changes in a clinical setting. Moreover, the sample population showed high risk of osteoporosis and would subsequently benefit from dual-energy x-ray absorptiometry scanning for definite evaluation and treatment.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Conflict of Interest
  10. References

Diabetes mellitus (DM) is a chronic metabolic disorder with substantial morbidity and mortality.[1] This disorder and its complications require continuing medical care and education to prevent and reduce the risk of long-term complications. The incidence of DM is escalating around the world and this is especially the case in Asia.[2] The prevalence of type 2 diabetes (T2DM) has rapidly increased during the last 20 years with a staggering prevalence of 14.9% in the adult population according to the Malaysian Third National Health and Morbidity Survey (NHMS III).[3]

On the other hand, osteoporosis is a silent progressive skeletal disease that constitutes a great socioeconomic threat, with a negative impact on health outcome.[4, 5] In Malaysia, osteoporosis prevalence was reported as high as 24.1%.[6] Osteoporosis continues to be an underestimated problem in diabetic patients, and remains undetected until fractures occur, even though several studies have shown that diabetes is a common risk factor for osteoporosis.[7-9]

The overall osteoporosis prevalence in the Asian population is higher than Western countries due to the fact that the Asian population has lower body mass index and shorter height.[10] Furthermore, a low dietary calcium intake and lack of physical activity have been reported to be among the risk factors for osteoporosis in the Asian population.[11] The longer life expectancy in people with diabetes due to improvement of medical care may increase the incidence of osteoporosis in such patients.

One of the most widely used instruments to assess osteoporosis health belief is the Osteoporosis Health Belief Scale (OHBS). Although the OHBS had been previously validated in Caucasian populations,[12] its validation among South-East Asian populations is not confirmed. In addition, no studies have been carried out in a special population such as diabetic patients. The aims of this study were to examine the validity, internal consistency, as well as reliability of the instruments for the assessment of OHBS-M in a linguistically distinct, Malaysian population. Furthermore, this study aimed to assess osteoporosis risk in the sample population and the correlation between osteoporosis knowledge, health belief and self-efficacy scales.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Conflict of Interest
  10. References

Participants and setting

A cross-sectional study was conducted for 250 outpatient T2DM patients at Penang General Hospital (PGH). Data were collected for the period May–July 2011. Face-to-face interviews included the administration of the translated OHBS-Malaysian version (OHBS-M) and the collection of socio-demographic data. After screening, 312 diabetic patients were accepted to participate in this study, but 62 patients were ineligible due to incomplete response. In addition, 40 patients from the sample (= 250) were randomly selected and agreed to undertake a 1–2 week reliability test–retest analysis. Only 30 patients completed the test–retest. This study was approved by the Clinical Research Centre (CRC), Penang General Hospital and Medical Research Ethics Committee (MREC) of the Ministry of Health, Malaysia. All subjects provided a written consent form prior to participation.

The original OHBS was developed and validated in 201 postmenopausal women in the USA aged 35–95 years.[12] The OHBS tool was developed based on the Health Belief Model (HBM) theory. The HBM states that perception of a health behavior threat is influenced by beliefs about vulnerability to the disease and the consequences of a health problem. From that, the clinician or educators often use a person's perceived susceptibility to the problem as an explanation for bad outcomes. Therefore, the OHBS tool is used to determine the relationship between health beliefs and osteoporosis preventive health behaviors, including calcium intake and exercise.

The OHBS has 42 items grouped in seven subscales: perceived susceptibility, perceived seriousness, benefits of exercise, benefits of calcium intake, barriers to exercise, barriers to calcium intake, and health motivation.The OHBS is rated using a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree). The possible range of scores for each subscale is 6–30 with a possible total score range of 42–210.

In this study, the Osteoporosis Knowledge Tool (OKT-M) and Osteoporosis Self-efficacy Scale (OSES-M) (Malaysian version; Abdulameer S.A., Syed Sulaiman S.A., Hassali M.A., Subramaniam K., Sahib M.N., unpublished data) questionnaires, which are valid and reliable, were administered prior and following the OHBS-M to assess patients' osteoporosis knowledge and self-efficacy, respectively. The original OKT and OSES were previously developed and published elsewhere.[13] The OKT (24-item) and OSES (12-item) each have two subscales: OKT-exercise and OKT-calcium, OSES-exercise and OSES-calcium, respectively.

Instrument translation methodology

A forward–backward–forward translation method was used to translate the 42-item OHBS-M, as previously mentioned.[14, 15] The translation process was undertaken by experts at the School of Languages, Literacies & Translation, University Sains Malaysia (USM), Penang.

The translation process included the following steps.

  • The initial translation (forward translation) of the original questionnaire was undertaken by two different qualified, independent linguistic translators from English to the Malay language to produce a version that was semantically and conceptually as close as possible to the original questionnaire.
  • Reconciliation of the two-version of the forward translation through the meeting of the research group to review, reconcile and harmonize the two-translation version with the original one to create the reconciled forward translation which met all demands of the conceptual equivalence with the original English questionnaire.
  • Back translation (reverse translation) of the reconciled forward translation from Malay to English was carried out by two different translators, experts in both Malay and English. After repeated discussion between the professional forward translators and the research group, both were reviewed and judged semantically equivalent to the original English questionnaire. Inconsistencies were resolved in a consensus meeting and a final harmonized version, ready for testing, was generated.
  • Pre-test cognitive interviews (debriefing) were conducted to ensure the feasibility, understandability, interpretation and cultural relevance of the translation. First, eight clinical pharmacists, experts in the pharmacy field, were invited to review and provide feedback on the translated questionnaire. In addition, pre-testing of the translated questionnaire (as a pilot study) was distributed to 40 Malaysian non-academic staff at the School of Pharmaceutical Science (USM) and 30 T2DM out-patients to test degree of difficulty and clarity of questions, appropriateness and comprehensiveness of the items and acceptance of the items by asking participants to rephrase the items using their own words immediately after answering the items. This allows the researchers to assess whether participants understand the items in total. The questionnaire was modified based on feedback from the pilot study before being used and those individuals were not included for any statistical procedures in the final study outcomes.
  • The final version of the Malaysian questionnaire was completed and made available for the validity and reliability study. The questionnaire takes about 15–20 min to complete.


Face validation

The face validity of a measurement item refers to the degree to which the measurement item appears, on its face value, to measure the construct that it intends to measure.[16] Eight experts in the pharmacy field were invited as subject-matter experts (SME) to review, provide feedback and judge the face validity of the translated questionnaire. Each expert was asked to complete the two tasks. For task A, the eight experts were provided with the operational definitions for each of the dimensions and a random listing of the 42 measurement items to guide them in classifying the measurement items into no more than one dimension. The sorting results from task A were then used to compute Fleiss' kappa, an index of beyond-chance agreement among different ‘judges’ for the overall task, after chance agreement has been removed.[17]

The SME were then asked to evaluate, in task B, how adequately each measurement item measures the dimension to which it had been assigned. For each measurement item, the experts were asked to respond to a seven-point Likert scale from 1 (very inaccurate) to 7 (almost perfect). From the experts' input for task B, the average adequacy score and the standard deviation of adequacy scores for each individual measurement item were computed and evaluated.

Content validation

The content validity is the extent to which a measurement tool captures the different facets of an entire construct domain.[18] The content validity of the items was quantified by applying a content validity ratio (CVR), which is commonly used to facilitate ‘the rejection or retention of specific items’.[19] In this study, the internal validation was carried out by asking eight experts whether or not the defined 42 items in OHBS-M scale were ‘1 = essential’, ‘2 = useful but not essential’ or ‘3 = not necessary’. Responses from all panelists were pooled and the number indicating ‘essential’ for each item was determined. The content validity index (CVI) represents the average percentage of overlap between the test items and the performance domain. It was calculated using significant CVR values of the OHBS-M.[19]

Construct validity

The construct validity of the OHBS-M was examined by using both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) to establish and confirm the factor structures of OHBS-M (Malay version). In EFA, the factor structure was established using a principal axis factoring method for extraction with varimax rotations. In interpreting the rotated factor, an item was supposed to load on a given component if the factor loading was 0.40 or greater for that component, and was less than 0.40 for the other.[20] Factor analysis appropriateness was assessed by Kaiser–Meyer–Olkin (KMO) value (> 0.5), with a significant level < 0.05 for Bartlett's test of sphericity. The decision about the number of factors to retain was based on a combination of methods including Kaiser's criterion (eigenvalue ≥ 1.0), scree plot and parallel analysis (PA), as well as conceptual meaningfulness of the rotated factors.

The present study employed the Structural Equation Modeling (SEM) to evaluate the OHBS-M factor structure by using the CFA approach using maximum likelihood estimates (ML) with a bootstrapping method. As well, the CFA was used to test validity and reliability of the measurement model (i.e., convergent, discriminant validity and composite reliability). Convergent validity is the degree to which multiple attempts to measure the same concept are in agreement. Discriminant validity is the degree to which the measures of different concepts are distinct and can be examined by comparing the squared correlations between constructs and average variances extracted (AVE) for a construct. The AVE, which reflects the overall amount of variance in the indicators accounted for by the latent construct, should exceed the recommended value of 0.5. The composite reliability (CR) values, which depict the degree to which the construct indicators (latent variable) indicate the latent construct, should exceed the recommended level of 0.7.[21]

Several indices were employed to judge whether the model tested fits to the data. The common indices selected were: the chi-square statistics (χ2); normed chi-square/degrees of freedom score (χ2/d.f.); Bentler's comparative fit index (CFI); goodness of fit index (GFI); adjusted goodness-of-fit index (AGFI); root mean square error of approximation (RMSEA); Tucker-Lewis index (TLI); parsimonious normed fit index (PNFI); parsimony goodness of fit index (PGFI); and Akaike information criteria (AIC).[22-25] A good model fit is indicated by values of 0.90 or higher for the CFI, GFI and TLI, and for the RMSEA, values of 0.05 or lower indicate a close fit, while values < 0.08 indicate an acceptable fit; in addition, the one with the lowest AIC is preferred in model comparison.[26]


Reliability is the consistency of a measurement item with a minimum acceptable criterion above 0.5, and range from 0.0 to 1.0.[27] Internal consistency was assessed using Cronbach's alpha and corrected item-total correlations between the scales and their corresponding items. A correlation of < 0.20 is considered poor.[28] Pearson's correlation coefficient was used to assess test–retest reliability.

Quantitative ultrasound (QUS) measurements

Bone mineral density (BMD) measurements were carried out using QUS (SONOST 3000; OsteoSys Co., Ltd., Seoul, Korea) at the calcaneus. Subjects were classified as normal, osteopenic or osteoporotic based on the QUS T-scores using the World Health Organization (WHO) criteria as cut-off values.[29] The QUS technology is less expensive, portable and also has the advantage of not using ionising radiation, so it is safer than dual-energy X-ray absorptiometry (DEXA).

Statistical analysis

Predictive analytics software (PASW) version 18.0 and Analysis of Moment Structures (AMOS) version 6.0 (both SPSS Inc., Chicago, IL, USA) were used to analyze data in this study. Descriptive statistics, including percentages, mean ± standard deviation (± SD) were used to describe demographic characteristics as well as instrument findings. A bivariate correlation analysis was performed to examine the relationships between scales.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Conflict of Interest
  10. References

Sociodemographic data

Employing the recommended scoring method, the mean ± SD of for OHBS-M scores was 158.31 ± 20.80. The respondents were predominantly male (58%). The mean age was 61.88 ± 9.86 years and the majority were Chinese (47.6%). The characteristics of the diabetic patients are shown in Table 1. According to QUS measurements, only 22% had a normal BMD (low risk of abnormal BMD), while osteopenia and osteoporosis were considered as 57.6% and 20.4% (high risk of abnormal BMD), respectively. Surprisingly, 68.8% (9.27 ± 3.16) and 71.6% (639.13 ± 148.53) of the sample population had low knowledge and self-efficacy toward osteoporosis, respectively.

Table 1. Demographic characteristics of diabetic patients; data were expressed as mean ± SD or frequency (%)
CharacteristicsTotal sample (= 250)Low osteoporosis health belief (= 163)High osteoporosis health belief (= 87)
42 item OHBS-M score158.31 ± 20.80147.28 ± 16.60178.98 ± 8.43
OHBS-M Susceptibility21.90 ± 7.6220.33 ± 8.0224.86 ± 5.78
OHBS-M Seriousness22.84 ± 6.8620.99 ± 7.7626.30 ± 2.07
OHBS-M Benefit Exercise25.44 ± 3.6225.06 ± 3.9426.16 ± 2.82
OHBS-M Benefit Calcium Intake25.81 ± 2.8225.46 ± 3.1426.46 ± 1.93
OHBS-M Barrier Exercise21.49 ± 8.4518.36 ± 8.9127.34 ± 2.06
OHBS-M Barrier Calcium Intake16.16 ± 7.7413.42 ± 6.0121.28 ± 8.04
OHBS-M Health Motivation24.67 ± 5.9123.66 ± 6.7326.57 ± 3.17
< 4411 (4.4)5 (3.1)6 (6.9)
45–5452 (20.8)36 (22.1)16 (18.4)
55–6484 (33.6)50 (30.7)34 (39.1)
≥ 65103 (41.2)72 (44.2)31 (35.6)
Male145 (58)98 (60.1)47 (54)
Female105 (42)65 (39.9)40 (46)
Malay67 (26.8)40 (24.5)27 (31)
Chinese119 (47.6)81 (49.7)38 (43.7)
Indian64 (25.6)42 (25.8)22 (25.3)
Educational levels (years)
< 12173 (69.2)111 (68.1)62 (71.3)
≥ 1277 (30.8)52 (31.9)25(28.7)
Marital status
Single25 (10)16 (9.8)9 (10.3)
Married208 (83.2)135 (82.8)73 (83.9)
Separated/divorced/widowed17 (6.8)12 (7.4)5 (5.7)
Monthly income in ringgit (RM)
≤ 1999196 (78.4)128 (78.5)68(78.2)
> 200054 (21.6)35 (21.5)19 (21.8)
Employment status
Private51 (20.4)41 (25.2)10 (11.5)
Not employed107 (42.8)67 (41.1)40 (46)
Retired60 (24)38 (23.3)22 (25.3)
Government32 (12.8)17 (10.4)15 (17.2)
Body mass index (BMI) (kg/m2)
Underweight (< 18.49)5 (2)4 (2.5)1 (1.1)
Normal (18.5–24.9)104 (41.6)62 (38)42 (48.3)
Overweigh (25–29.9)91 (36.4)65 (39.9)26 (29.9)
Obesity (≥ 30)50 (20)32 (19.6)18 (20.7)
Diabetic duration (years)
3–6 years103 (41.2)66 (40.5)37 (42.5)
7–1075 (30)54 (33.1)21 (24.1)
11–1436 (14.4)23 (14.1)13 (14.9)
≥ 1536 (14.4)20 (12.3)16 (18.4)
Glycosylated haemoglobin (HbA1c)
< 6.542 (16.8)27 (16.6)15 (17.2)
≥ 6.5208 (83.2)136 (83.4)72 (82.8)
Living place
Rural33 (13.2)23 (14.1)10 (11.5)
Urban217 (86.8)140 (85.9)77 (88.5)
Family history of osteoporosis
No233 (93.2)151 (92.6)82 (94.3)
Yes17 (6.8)12 (7.4)5 (5.7)
Family history of fracture
No232 (92.8)150 (92)82 (94.3)
Yes18 (7.2)13 (8)5 (5.7)
Cigarette use
No203 (81.2)128 (78.5)75 (86.2)
Yes47 (18.8)35 (21.5)12 (13.8)
Alcohol use
No213 (85.2)139 (85.3)74 (85.1)
Yes37 (14.8)24 (14.7)13(14.9)


Face and content validation

In task A, Fleiss' kappa was computed to be 0.99, which represents excellent agreement. For task B, all the measurement items were deemed to have face validity with high average adequacy scores (> 3) and small standard deviations of adequacy scores (≤ 1.00). According to Lawshe, with a panel of eight respondents, the minimum value of CVR needs to be at least 0.75 in order for it to be acceptable. This study showed that all 42 variables had a significant CVR with values ≥ 0.75, varying 0.75–1.00. The CVI value for the OHBS-M was 0.88. Thus, it was inferred that all 42 variables were strongly valid for this research and could be included in the final form of the questionnaire.

Construct validity

Prior to conducting the primary analyses, the data were examined for accuracy, missing values, outliers and multivariate assumptions.[30] No outliers were identified when using Mahalanobis distance (d-squared, with a cut-off of 0.001). The multivariate kurtosis and critical ratio values were 181.89 and 23.65, respectively, which indicated a departure from normality.[31]

The adequacy and reliability of the research data for factor analysis was tested according to the sample size. For factor analysis, a sample of a minimum ratio of five subjects per variable (item) is adequate, or at least 100 participants was considered essential for factor analysis.[9] In this study, a ratio of 5 : 1 with a total size of 250 participants was used.

Exploratory factor analysis (EFA)

In this study, upon examination of the correlation matrices, the majority of the results showed correlation larger than 0.3. The KMO value in the present analysis was 0.9, which was within the range of ‘superb’ value and the data set was suitable for factor analysis as it was greater than 0.5.[20] The final criterion to check the adequacy of data for factor analysis was a significant value in Bartlett's Test of Sphericity which was found to be highly significant (χ2(861) = 12311.138; = 0.000). This result allowed us to identify the factor model using the EFA approach.[32, 33] In addition, the requested analysis yielded seven factors, with eigenvalues greater than 1, which explained 78.53% of the variance, as shown in Table 2. As a whole, the EFA data were adequate for factor analysis with a final result of seven latent variables.

Table 2. Component matrix of exploratory factor analysis for Osteoporosis Health Belief Scale Malay version (OHBS-M)
ItemRotated factor matrixCommunalities
Factor 1Factor 2Factor 3Factor 4Factor 5Factor 6Factor 7
  1. Extraction method: principal axis factoring; Rotation method: Varimax with Kaiser normalization; Factor 1 = OHBS-Susceptibility; Factor 2 = OHBS-Seriousness; Factor 3 = OHBS-Benefits Exercise; Factor 4 = OHBS- Benefits Calcium Intake; Factor 5 = OHBS-Barriers Exercise; Factor 6 = OHBS-Barriers Calcium Intake; Factor 7 = OHBS-Health Motivation; items comprising each factor are in bold.

Question 1 0.864       0.760
Question 2 0.917       0.855
Question 3 0.937       0.886
Question 4 0.941       0.894
Question 5 0.932       0.884
Question 6 0.918       0.852
Question 7  0.881      .831
Question 8  0.899      0.836
Question 9  0.900      0.856
Question 10  0.895      0.828
Question 11  0.911      0.855
Question 12  0.920      0.881
Question 13   0.730     0.610
Question 14   0.777     0.684
Question 15   0.760     0.647
Question 16   0.834     0.726
Question 17   0.840     0.752
Question 18   0.792     0.694
Question 19    0.597    0.570
Question 20    0.651    0.609
Question 21    0.621    0.549
Question 22    0.537    0.392
Question 23    0.682    0.582
Question 24    0.725    0.600
Question 25     0.914   0.884
Question 26     0.935   0.919
Question 27     0.905   0.876
Question 28     0.914   0.866
Question 29     0.952   0.928
Question 30     0.953   0.934
Question 31      0.863  0.789
Question 32      0.923  0.886
Question 33      0.821  0.706
Question 34      0.924  0.887
Question 35      0.949  0.928
Question 36      0.868  0.802
Question 37       0.821 0.712
Question 38       0.855 0.768
Question 39       0.915 0.870
Question 40       0.901 0.825
Question 41       0.938 0.901
Question 42       0.915 0.868
% variance17.91512.6267.8282.60318.89210.0318.633Total = 78.528
Cronbach's α0.9700.9690.9270.8710.9810.9650.963Total = 0.89
Confirmatory factor analysis (CFA)

A seven-factor model was tested based on Health Belief Model (HBM) theory[34] and the findings from the EFA in our study. CFA was subsequently used to examine the construct validity (composite reliability, convergent and discriminant validity) of a seven-factor model. Convergent validity was assessed based on factor loading, composite reliabilities and variances extracted.[21] The factor loading for all items exceeds the recommended level of 0.6.[35] CR values ranged from 0.86 to 0.98. The AVE were in the range between 0.52 and 0.89. The results of the convergent validity were shown in Table 3. The results from discriminant validity showed that the correlations for each construct were less than the AVE by the indicators measuring that construct, as shown in Table 4. In summary, the measurement model of the OHBS-M tool demonstrated adequate reliability, convergent and discriminant validity.

Table 3. Convergent validity of the Osteoporosis Health Belief Scale Malay version (OHBS-M)
ConstructItemCronbach's alphaConvergent validity
Factor loadingsAVECR
  1. †Average variance extracted (summation of the square of the factor loadings)/(summation of the square of the factor loading) + (summation of error variances). ‡Composite reliability (square of the summation of the factor loadings)/(square of the summation of the factor loadings) + (summation of error variances).

Benefit exerciseOHBSQ130.9270.7350.670.93
Benefit calcium intakeOHBSQ190.8710.7970.520.86
Barrier exerciseOHBSQ250.9810.9430.890.98
Barrier calcium intakeOHBSQ310.9650.8850.830.97
Health motivationOHBSQ370.9630.8280.820.96
Table 4. Discriminant validity of the Osteoporosis Health Belief Scale Malay version (OHBS-M)
SubscalesAVEHealth motivationBarrier calicumBarrier exerciseBenefit calicumBenefit exerciseSeriousnessSusceptibility
  1. Note: Diagonal elements (in bold) are the square root of the average variance extracted (AVE). Off-diagonal elements are the correlations among the constructs. For discriminant validity, diagonal elements should be larger than off-diagonal elements.

Health motivation0.82 0.90       
Barrier calicum0.83−0.02 0.91      
Barrier exercise0.89−0.0310.35 0.94     
Benefit calicum0.520.324−0.049−0.059 0.72    
Benefit exercise0.670.150.010−0.0690.674 0.82   
Seriousness0.840.1670.1660.1670.2060.170 0.92  
Susceptibility0.85−0.009−0.0230.0160.0950.0150.165 0.92

The calculations of the goodness-of-fit indices for the seven items of the OHBS-M scale produced a good fit and these indices include the observed χ2/d.f. score for the measurement model which was 1.31 (χ= 1041.48; d.f. = 795; = 0.000) and was still an acceptable value (< 3).[36] The RMSEA was 0.035 (95% CI: 0.029–0.041) which gave evidence of good fit. The CFI (0.98) and TLI (0.978) have values greater than 0.9, which exceeds the recommended cut-off value. The AGFI of the hypothesized model was 0.82 which also indicated a good fit. The PNFI (0.85) and PGFI (0.905) values were > 0.60 which is generally acceptable.[37] As a whole, these results indicated a good evidence of fit to the measurement model with seven factor dimensions for the OHBS-M version.[26]


The OHBS-M and its subscales showed excellent reliabilities for the questionnaire. The Cronbach's alpha values were equal to: 0.89 for OHBS-M; 0.970 for OHBS-susceptibility; 0.969 for OHBS-seriousness; 0.927 for OHBS-benefit exercise; 0.871 for OHBS-benefit calcium intake; 0.981 for OHBS-barrier exercise; 0.965 for OHBS-barrier calcium intake; and 0.963 for OHBS- health motivation. The corrected item-total correlation values ranged from 0.204 to 0.574, as shown in Table 5. Test–retest reliabilities of the OHBS-M and its subscales demonstrated significant positive relationships in a sample of 30 T2DM patients (< 0.01) (= 0.555 for OHBS-M scale; = 0.764 for OHBS-susceptibility; = 0.827 for OHBS-seriousness; = 0.596 for OHBS-benefit exercise; = 0.700 for OHBS-benefit calcium intake; = 0.582 for OHBS-barrier exercise; = 0.755 for OHBS-barrier calcium intake; and = 0.693 for OHBS-health motivation). An initial Cronbach's alpha result for the OHBS-M test–retest group was 0.77, and was 0.74 after approximately 2 weeks. The findings demonstrated that the OHBS-M and its seven subscales were reliable and stable.

Table 5. Reliability test of 42 questions of the Osteoporosis Health Belief Scale Malay version (OHBS-M)
OHBS question no.Mean ± SDCorrected item – total correlationCronbach's α if item deleted
  1. Cronbach's α was 0.89 for the total scale.

Question 13.64 ± 1.4360.3410.888
Question 23.65 ± 1.2870.3190.888
Question 33.59 ± 1.3060.3510.888
Question 43.60 ± 1.2730.3660.887
Question 53.68 ± 1.3470.3380.888
Question 63.74 ± 1.5000.3380.888
Question 73.84 ± 1.2690.5740.884
Question 83.78 ± 1.2140.5070.885
Question 93.77 ± 1.1720.5460.885
Question 103.82 ± 1.2790.4900.885
Question 113.84 ± 1.1710.5000.885
Question 123.79 ± 1.2610.5370.885
Question 134.22 ± 0.7120.2410.889
Question 144.21 ± 0.6440.2500.889
Question 154.28 ± 0.7080.2400.889
Question 164.22 ± 0.7040.2410.889
Question 174.26 ± 0.7490.2500.889
Question 184.26 ± 0.7100.2890.888
Question 194.32 ± 0.6010.3280.888
Question 204.26 ± 0.5830.3230.888
Question 214.34 ± 0.6480.242.889
Question 224.35 ± 0.6040.2150.889
Question 234.24 ± 0.5580.2420.889
Question 244.30 ± 0.6150.2190.889
Question 253.62 ± 1.4380.4710.886
Question 263.65 ± 1.5110.4980.885
Question 273.58 ± 1.4770.5010.885
Question 283.46 ± 1.3770.4890.885
Question 293.55 ± 1.5020.4590.886
Question 303.62 ± 1.5320.4930.885
Question 312.59 ± 1.4320.4580.886
Question 322.63 ± 1.3270.4730.886
Question 332.70 ± 1.4680.3450.888
Question 342.76 ± 1.3770.4640.886
Question 352.71 ± 1.3880.4630.886
Question 362.78 ± 1.3840.4740.886
Question 374.07 ± 0.9410.2040.889
Question 384.11 ± 0.9960.3080.888
Question 394.15 ± 1.1760.3310.888
Question 404.11 ± 1.0760.2970.888
Question 414.10 ± 1.1310.2900.888
Question 424.13 ± 1.0900.3350.888

Relationships between scales and subscales

Correlation between OKT-M, OSES-M and OHBS-M scores in the calcium and exercise subscales were examined (Table 6). There were correlations between scores on osteoporosis knowledge and some of the health belief and self-efficacy subscales. Patients who were most knowledgeable about exercise (OKT-M exercise) were also most aware of the benefits and more self-confident to exercise (OSES-M exercise). As a consequence, there was a negative relation between OSES-M exercise and perceived barriers to exercise but this result was insignificant. In addition, the patients with greater OKT-M calcium subscale scores perceived more confidence and benefit in calcium intake. The osteoporosis self-efficacy of calcium was negatively related with perceived barriers to calcium intake.

Table 6. Correlation between osteoporosis knowledge tool (OKT-M), osteoporosis self-efficacy scale (OSES-M) and osteoporosis health believe scale (OHBS-M) scores in the calcium and exercise sub-scales
 OKT Exercise-MOSES-M ExercisePerceived benefits of exercise
OSES-M Exercise0.242*0.225*
Perceived benefits of exercise0.201*0.225*
Perceived barriers to exercise0.073−0.013−0.030
 OKT Calcium-MOSES-M CalciumPerceived benefits of calcium
  1. OKT-M Exercise, Osteoporosis Knowledge Tool Exercise sub-scale Malay version; OKT-M Calcium, Osteoporosis Knowledge Tool Calcium sub-scale Malay version; OSES-M Exercise, Osteoporosis Self-Efficacy Scale Exercise sub-scale Malay version; OSES-M Calcium, Osteoporosis Self-Efficacy Scale Calcium sub-scale Malay version. *P-value < 0.01.

OSES-M Calcium0.286*0.018
Perceived benefits of calcium0.167*0.018
Perceived barriers to calcium−0.116−0.253*−0.189*


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Conflict of Interest
  10. References

The diabetes educator's role is to improve the quality of patients' education and behavior. However, the educator must have a valid and reliable tool to evaluate the effectiveness of the teaching and learning that are done. This tool in conjunction with OKT-M and OSES-M can identify and reinforce the patient's present and past successes in their osteoprotective behaviors.

This was the first study conducted to translate and validate the OHBS questionnaire tool in a clinical setting. A multistep approach was taken, involving forward–backward translation, pilot testing of the preliminary instrument and psychometric testing of the revised instrument in T2DM patients. Awareness of a particular health problem (osteoporosis) in a special population (diabetic patient) represents a significant component of an effective preventive program. In osteoporosis research, there is a paucity of literature regarding the use of this instrument for assessing osteoporosis health beliefs in clinical practice. Furthermore, the validity and reliability of this instrument must be high enough for its use and interpretation linguistically across diverse groups (i.e. cultural adaptation). The results of the present study were an important first step before attempting to apply the instrument as a tool in clinical research.

The content validity was carefully reviewed using quantitative approaches (CVR), which offer practicality in terms of time and cost, as well as being quick and easy in order to evaluate and validate the OHBS-M scale.[38] Considering the limited use of a quantitative approach to assess content validity of a scale in clinical studies, this study illustrates the practicality of such an approach when evaluating and validating the OHBS-M scale. In this study, all items in the OHBS-M scale represented the best possible pool of items to retain at this stage, with a good content validity level and conceptual fit.

The OHBS-M has a stable factor structure. The EFA establishes that the seven factors related to health belief subscales accounted for 78.53% of the variance, which was considered higher than other studies.[12, 39] The extensive CFA confirmed that the OHBS-M items can be represented by seven subscales as the developers suggest, which indicated a successful theoretical framework of the osteoporosis health belief hypothesis even within different cultures.

The 42-item OHBS-M was internally reliable with an excellent overall Cronbach's alpha (0.89). In addition, the degree of consistency for OHBS-M was comparable to the original developed study.[12] Similar results were found in a Persian study, which showed that the OHBS had good validity with good test–retest reliability in Iranian women.[39] The item-total correlation results showed that OHBS-M scale scores were considered sufficiently reliable. The test–retest reliability and Cronbach's alpha for OHBS-M and its subscales over 1–2 weeks showed good results. These findings show a good reproducibility of values over time, with precision of measurements, and could be used in longitudinal studies to assess the change in osteoporosis health beliefs. The Cronbach's alpha value after 1–2 weeks was lower than the initial Cronbach's alpha value for the test–retest group, indicating that patients may need a continuous education program. The results of validity and reliability showed successful cultural adaptation of the translated tool from the English version in a Malaysian population.

Using QUS T-scores is valuable to identify patients at risk of osteoporosis who would subsequently benefit from DEXA scanning for definite evaluation and treatment. Although we did not confirm our diagnoses using DEXA scan, our study revealed low QUS values among T2DM outpatients. Therefore, knowledge about patients' health beliefs toward osteoporosis and preventive health behaviors should aid their treating physicians in adequately targeting those in need of further testing and/or treatment.

The perceived benefit of exercise was significantly correlated to OKT-M and OSES-M exercise sub-scales. Thus, diabetic patients with low OKT-M levels are at a higher risk of osteoporosis than others, so more attention should be focused on populations besides post-menopausal women and the elderly. Therefore, this result highlighted the need of an exercise educational program to those special populations as regular physical activity increases muscle and bone strength, increases lean muscle, enhances psychological well-being, and improves diabetes and lipid control.[40, 41] Similar results were found with the perceived benefit of calcium. Therefore, patient education needs to highlight the fact that consuming calcium-rich foods not only improves the nutritional quality of the diet, but can also enhance weight loss,[42] reduce blood pressure,[43] and improve lipid and diabetic control, which are often significant comorbidity risks in diabetic patients.[44, 45]

Moreover, the subjects in this study considered themselves susceptible toward the development of osteoporosis. These results were important from a behavioral point of view, as perceptions of personal susceptibility and belief in the seriousness of a disease are important for influencing behavioral changes in disease prevention programs.[46] It is well known that perceived barriers to risk-reducing behavior and perceived susceptibility to the disease appear to be the most powerful health belief model components in terms of bringing about behavioral change.[47] It is arguable that health promotion programs which address only knowledge and ignore the health beliefs model will fail to initiate an appreciable increase in risk-reducing behaviors in their target audience.[48] Therefore, the attitude of the sample population in this study suggests that there is an opportunity to improve the effectiveness of future osteoporosis prevention programs.

Since this was a cross-sectional design with a convenient sample, our findings may not therefore, be representative of all diabetic populations in Penang/Malaysia and require reinforcement through further work. Unequal numbers of ethnic, age and gender groups may impact the final results if we compared this with a randomized controlled study. Although DEXA is the current preferred method (gold standard) for diagnosing osteoporosis, QUS methods were used in our study as an alternative in the evaluation of bone status. This technology is less expensive, portable and also has the advantage of not using ionizing radiation, so it is safer. Even though it would be beneficial to conduct routine osteoporosis screening, it is not feasible to do it in developing countries due to cost constraints and insufficient availability of DEXA.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Conflict of Interest
  10. References

In conclusion, the OHBS-M was a valid and reliable instrument for assessing diabetic patients' health beliefs toward osteoporosis in the Malaysian setting. Therefore, it can be used to identify individuals in need for educational interventions and to assess the effectiveness of educational efforts as a part of osteoporosis management.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Conflict of Interest
  10. References

S.A. Abdulameer gratefully acknowledges the Universiti Sains Malaysia, Penang, Malaysia, for granting her the USM Postgraduate Student Fellowship.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Conflict of Interest
  10. References
  • 1
    Sahib MN, Abdulameer SA, Aziz NA, Hassan Y (2009) Pathogenesis of diabetic kidney disease: review of cellular aspects of renal lesions. Afr J Pharm Pharacol 3, 50714.
  • 2
    Wild S, Roglic G, Green A, Sicree R, King H (2004) Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care 27, 104753.
  • 3
    Letchuman G, Wan Nazaimoon W, Wan Mohamad W et al. (2010) Prevalence of diabetes in the Malaysian national health morbidity survey III 2006. Med J Malaysia 65, 1739.
  • 4
    Hadjidakis D, Mylonakis A, Sfakianakis M, Raptis A, Raptis S (2005) Diabetes and premature menopause: is their co-existence detrimental to the skeleton? Eur J Endocrinol 152, 43742.
  • 5
    Reyes B, Moreno O (2005) Prevalence of osteopenia and osteoporosis in postmenopausal women. Aten Primaria 35, 3427.
  • 6
    Lim PS, Ong FB, Adeeb N et al. (2005) Bone health in urban midlife Malaysian women: risk factors and prevention. Osteoporos Int 16, 206979.
  • 7
    Thrailkill K, Lumpkin C Jr, Bunn R, Kemp S, Fowlkes J (2005) Is insulin an anabolic agent in bone? Dissecting the diabetic bone for clues. Am J Physiol Endocrinol Metab 289, E73545.
  • 8
    Abdulameer SA, Syed Sulaiman SA, Hassali MA, Subramaniam K, Sahib MN (2012) Is there a link between osteoporosis and Type 1 diabetes? Findings from a systematic review of the literature. Diabetol Int 3, 11330.
  • 9
    Abdulameer SA, Syed Sulaiman SA, Hassali MA, Subramaniam K, Sahib MN (2012) Osteoporosis and type 2 diabetes mellitus: what do we know, and what we can do? Patient Prefer Adherence 6, 43548.
  • 10
    Loh KY, Shong HK (2007) Osteoporosis: primary prevention in the community. Med J Malaysia 62, 3558.
  • 11
    Lau EM, Woo J, Lam V, Hong A (2001) Milk supplementation of the diet of postmenopausal Chinese women on a low calcium intake retards bone loss. J Bone Miner Res 16, 17049.
  • 12
    Kim K, Horan M, Gendler P, Patel M (1991) Development and evaluation of the osteoporosis health belief scale. Res Nurs Health 14, 15563.
  • 13
    Kim K, Horan M, Gendler P (1991) Osteoporosis Knowledge Tests, Osteoporosis Health Belief Scale, and Osteoporosis Self-Efficacy Scale. Grand Valley State University, Allendale, MI.
  • 14
    Wild D, Grove A, Martin M et al. (2005) Principles of good practice for the translation and cultural adaptation process for Patient Reported Outcomes (PRO) measures: report of the ISPOR Task force for translation and cultural adaptation. Value Health 8, 94104.
  • 15
    Beaton DE, Bombardier C, Guillemin F, Ferraz MB (2000) Guidelines for the process of cross-cultural adaptation of self-report measures. Spine 25, 3186.
  • 16
    DeVon HA, Block ME, Moyle-Wright P et al. (2007) A psychometric toolbox for testing validity and reliability. J Nurs Scholarsh 39, 15564.
  • 17
    Fleiss JL (1971) Measuring nominal scale agreement among many raters. Psychol Bull 76, 37882.
  • 18
    Rungtusanatham M (1998) Let's not overlook content validity. Decision Line 29, 103.
  • 19
    Lawshe CH (1975) A quantitative approach to content validity. Pers Psychol 28, 56375.
  • 20
    Fabrigar LR, Wegener DT, MacCallum RC, Strahan EJ (1999) Evaluating the use of exploratory factor analysis in psychological research. Psychol Methods 4, 27299.
  • 21
    Fornell C, Larcker DF (1981) Evaluating structural equation models with unobservable variables and measurement error. J Mark Res 18, 3950.
  • 22
    Bentler PM (1990) Comparative fit indexes in structural models. Psychol Bull 107, 23846.
  • 23
    Steiger JH (1998) A note on multiple sample extensions of the RMSEA fit index. Struct Equ Modeling 5, 4119.
  • 24
    Tucker LR, Lewis C (1973) A reliability coefficient for maximum likelihood factor analysis. Psychometrika 38, 110.
  • 25
    Akaike H (1987) Factor analysis and AIC. Psychometrika 52, 31732.
  • 26
    Hu L, Bentler PM (1999) Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Struct Equ Modeling 6, 155.
  • 27
    Cronbach L (1951) Coefficient alpha and the internal structure of tests. Psychometrika 16, 297334.
  • 28
    Streiner DL, Norman GR (2008) Health Measurement Scales: A Practical Guide to Their Development and use. Oxford University Press, New York.
  • 29
    Kanis J (1994) Assessment of fracture risk and its application to screening for postmenopausal osteoporosis: synopsis of a WHO report. Osteoporos Int 4, 36881.
  • 30
    Peat JK, Barton B (2005) Medical Statistics: A Guide to Data Analysis and Critical Appraisal. Blackwell Publishing, Oxford.
  • 31
    Mardia KV (1970) Measures of multivariate skewness and kurtosis with applications. Biometrika 57, 51930.
  • 32
    Bartlett MS (1950) Tests of significance in factor analysis. Br J Math Stat Psychol 3, 7785.
    Direct Link:
  • 33
    Bartlett MS (1951) A further note on tests of significance in factor analysis. Br J Math Stat Psychol 4, 12.
    Direct Link:
  • 34
    Rosenstock I (1966) Why people use health services. Milbank Mem Fund Q 44, 94127.
  • 35
    Chin WW, Gopal A, Salisbury WD (1997) Advancing the theory of adaptive structuration: the development of a scale to measure faithfulness of appropriation. Inf Sys Res 8, 34267.
  • 36
    Segars AH, Grover V (1993) Re-examining perceived ease of use and usefulness: a confirmatory factor analysis. MIS Quarterly 17, 51725.
  • 37
    Chau PYK, Hu PJH (2001) Information technology acceptance by individual professionals: a model comparison approach. Decision Sci 32, 699719.
  • 38
    Tojib DR, Sugianto LF (2006) Content validity of instruments in IS research. JITTA 8, 5.
  • 39
    Baheiraei A, Ritchie JE, Eisman JA, Nguyen TV (2005) Psychometric properties of the Persian version of the osteoporosis knowledge and health belief questionnaires. Maturitas 50, 1349.
  • 40
    Schneider SH, Khachadurian AK, Amorosa LF, Clemow L, Ruderman NB (1992) Ten-year experience with an exercise-based outpatient life-style modification program in the treatment of diabetes mellitus. Diabetes Care 15, 180010.
  • 41
    Kelley G, Kelley K (2007) Effects of aerobic exercise on lipids and lipoproteins in adults with type 2 diabetes: a meta-analysis of randomized-controlled trials. Public Health 121, 64355.
  • 42
    Davies KM, Heaney RP, Recker RR et al. (2000) Calcium intake and body weight. J Clin Endocrinol Metab 85, 46358.
  • 43
    Ascherio A, Hennekens C, Willett WC et al. (1996) Prospective study of nutritional factors, blood pressure, and hypertension among US women. Hypertension 27, 106572.
  • 44
    Pereira MA, Jacobs DR Jr, Van Horn L, Slattery ML, Kartashov AI, Ludwig DS (2002) Dairy consumption, obesity, and the insulin resistance syndrome in young adults. JAMA 287, 20819.
  • 45
    Sabherwal S, Bravis V, Devendra D (2010) Effect of oral vitamin D and calcium replacement on glycaemic control in South Asian patients with type 2 diabetes. Int J Clin Pract 64, 10849.
  • 46
    Janz N, Becker M (1984) The health belief model: a decade later. Health Educ Behav 11, 147.
  • 47
    Abood DA, Black DR, Feral D (2003) Nutrition education worksite intervention for university staff: application of the health belief model. J Nutr Educ Behav 35, 2607.
  • 48
    Von Hurst PR, Wham CA (2007) Attitudes and knowledge about osteoporosis risk prevention: a survey of New Zealand women. Public Health Nutr 10, 74753.