ABSTRACT
Objective:
This study aimed to determine the validity and reliability of the Turkish version of the Osteoporosis Awareness Scale (OAS) developed to assess people’s osteoporosis awareness.
Materials and Methods:
This methodological study was designed to investigate the validity and reliability of the “OAS” in Turkish society. This study included healthcare workers who worked in Community Health and Family Health Centers (n=346). Numbers, percentages, t-test, correlation analysis, Cronbach’s alpha reliability coefficient, and factor analysis methods were used to analyze the data.
Results:
The Cronbach’s alpha reliability coefficient was 0.94 for the overall OAS and 0.86, 0.82, 0.88, 0.88, 0.85 for its bone physiology, preventive behaviors, risk factors, exercise, and characteristics of osteoporosis sub-dimensions, respectively. The exploratory factor analysis demonstrated that the scale explained 66.16% of the total variance. According to the confirmatory factor analysis, factor loadings of the scale ranged between 0.64 and 0.82, and CMIN(x2)/df was 3.384, goodness-of-fit index was 0.85, normed fit index was 0.83, comparative fit index was 0.88, and RMSEA was 0.08.
Conclusion:
Results indicate that the Turkish version of the OAS highly valid and reliable, which can be administered in future studies.
Introduction
Osteoporosis is an irreversible skeletal system disease that usually progresses without any symptoms, affects bone density and quality, and thus causes an increase in bone fragility. Nowadays, with the increase in life expectancy, osteoporosis has become an important public health problem due to its increasing incidence (1,2).
Osteoporosis, which causes bone fragility in individuals as they age, is also responsible for the increase in morbidity and mortality rates in old age, which is also considered as a social problem, because it requires costly investment allocated to osteoporosis treatment and causes workforce losses (3).
While sex (female), ethnicity (Asian, Spanish), advanced age, family history of osteoporosis or fractures are among the non-modifiable risk factors leading to osteoporosis, inadequate calcium intake, obesity, physical inactivity, smoking, alcohol use, vitamin D deficiency and caffeine intake are among the modifiable risk factors (3,4). Osteoporosis is a preventable disease if risk factors are identified early and changes in lifestyle are made (5). Due to osteoporosis and its complications, an individual’s activities of daily living are restricted and the individual may experience psychological problems as well. Therefore, the individual’s quality of life is gradually affected adversely. Women are at a higher risk than are men, and thus women should be informed about the early diagnosis of osteoporosis and how to prevent it from a young age by providing them with continuous and regular training (6,7). Health professionals such as physicians, midwives and nurses who have served the same population for many years in primary health care services can significantly contribute to efforts aimed at determining the risk groups and raising people’s awareness of this issue.
In the literature, various scales have been developed on osteoporosis. These scales are used to determine individuals’ osteoporosis knowledge levels (8-10), diagnosis of the disease (11), their health beliefs about osteoporosis (12), and their perception of self-efficacy towards osteoporosis (13). Populations of studies conducted on osteoporosis are mostly women and students (14-16). On the other hand, the number of studies conducted with health professionals who assume significant responsibilities in the diagnosis, evaluation and treatment of osteoporosis is rather limited (17-19). In this context, the aim of this study was to evaluate osteoporosis awareness for health professionals and to determine the validity and reliability of the Turkish version of the Osteoporosis Awareness Scale (OAS).
Materials and Methods
Study Design and Sample
The population of this methodological study planned to perform the validity and reliability study of the Turkish version of the OAS in society comprised of healthcare professionals working in Community Health and Family Health Centers in Karşıyaka district of İzmir Province affiliated to the İzmir Provincial Health Directorate. The study was conducted between March and June 2019 in İzmir. A sample size of 5-10 times the number of items in the scale is recommended for scale validity and reliability studies (20). Since OAS consists of 31 items, a sample size of 155 and 310 participants was expected. According to the recommandations, a sample size of more than ten times the number of items was aimed, and 346 health professionals were included in this study. The inclusion criterion of the study is as follows: working as a health worker in the units providing primary health care services in Karşıyaka district of İzmir province. The data were collected by face-to-face interview technique. Data collection was applied by using face-to-face interview technique.
Data Collection Tools
The study data were collected using the Healthcare Professional Information Form and the Turkish version of the OAS. It took each participant about 10-15 minutes to fill in the tools.
Healthcare Professional Information Form
The form developed by the researchers based on the relevant literature (3,12). includes 10 items questioning the healthcare professionals’ socio-demographic characteristics and their knowledge of general health.
Osteoporosis Awareness Scale
The scale Developed by Choi et al. (3) in 2008 in English consists of 31 items and 5 sub-dimensions. The responses given to the items of the OAS are rated on a 4-point Likert type scale ranging from 1 to 4 (4= I know very well, 3= I know, 2= I know a little, 1= I do not know at all). The minimum and maximum possible scores to be obtained from the OAS are 31 and 124 respectively. The higher the mean score obtained from the overall scale is the higher the level of awareness of osteoporosis is. The scale does not have reverse-scored items and cut-off point. While Cronbach’s alpha (a) reliability coefficient of the scale was found to be 0.948, Confirmatory Factor Analysis (CFA) was not performed in the original study.
Statistical Analysis
In the analysis of the study data, descriptive statistics were given in the arithmetic mean, standard deviation, numbers and percentages. Whether the data were normally distributed was determined using the Shapiro-Wilk normality test. The data obtained were analyzed using the IBM SPSS Version 24.0 statistics package program. P-values less than 0.05 were considered statistically significant.
Validity and Reliability Analysis
Language Validity
In order to make the language equivalence, the scale was translated from English to Turkish by experts in the field of health sciences who know both languages. After the translations were made, the translations made by the expert researchers were analyzed and the most appropriate expressions to be used in the Turkish version of the scale were determined and the translated text on which consensus was reached was created. Then the final form of Turkish version of the scale produced was translated back to English from Turkish by another expert with a good command of both languages and the approval of the author of the original scale was obtained.
Content and Construct Validity Analyses
To find out whether the overall OAS and its items adequately define the feature to be measured, experts were consulted to obtain their opinions. The experts’ opinions were evaluated through the Davis (21) technique. The content validity index (CVI) of the OAS and its items was calculated by dividing the number of experts who rated an item as “appropriate” or “needs slight revision” by the number of all the experts. A CVI higher than 0.8 indicates that the items of the scale are sufficient in terms of content validity (22). The content validity of the measurement tool was scored by five experts who have a good knowledge of the field of science for which the scale was prepared.
The exploratory factor analysis (EFA) was conducted to investigate the relationship between the scale items and factors. Before performing the factor analysis, whether the data set provided multivariate normality was investigated using the Bartlett’s test of Sphericity (BTS), and the adequacy level of the sample size was evaluated with the Kaiser-Meyer-Olkin (KMO) test. While the EFA was performed, the principal components analysis, and of the rotation methods, the Orthogonal Rotation and Varimax rotation were used to reveal the implicit structure of the OAS.
CFA was performed in order to find out whether the model that was previously tested and determined with EFA fitted the model to be adapted to the culture of the society. The IBM SPSS AMOS (Statistical Package for the Social Sciences-Analysis of Moment Structures) 25.0 analysis program was used to perform the CFA.
Reliability Analysis
In the present study, the stability of the scale over time was checked with the intermittent method in which 40 people included in the study sample underwent the test-retest application at a 4-week interval. To determine the internal consistency of the scale, the Cronbach’s a reliability coefficient and the item-total score analysis were used.
Ethical Considerations
In order to perform the validity and reliability study of the Turkish version of the OAS in primary health care workers, the permission was obtained from Choi et al. (3) who developed the OAS via e-mail. To conduct the study, ethical approval from a Ege University Medical Research Ethics Committee (decision no: 19-2T/23, date: 06.02.2019) was obtained. From the healthcare professionals who agreed to participate in the study, verbal and written consent was obtained. The participants who were administered the data collection tools using the face-to-face interview technique marked the options of the items they chose in the tools.
Results
Sample Characteristics
The research was completed with 346 healthcare workers who met the inclusion criteria. Their mean age was 46.7±9.9 years. The distribution of the participants according to their professions is as follows: 34.7% were physicians, 37% were midwives, 4.3% were nurses and 4% were from other professions (dentist, dietician, psychologist etc.). Their mean length of service in the profession was 23.1±9.5 years ranging from 1 to 49 years. Of the healthcare professionals, 8.1% were senior high school graduates, 17.9% had an associate’s degree, 3.4% had a bachelor’s degree, and 35.5% had a post-graduate degree (Table 1).
Validity Analysis
Language and Content Validity
In the present study, while opinions of three experts were obtained to ensure the language validity of the measurement tool, opinions of five experts were obtained to evaluate the content validity. The CVI was 0.95 for the overall Turkish version of the OAS and varied between 0.60 and 1 for its items.
Construct Validity
Exploratory Factor Analysis
According to the results of the KMO and Bartlett Sphericity tests conducted to evaluate the adequacy of the sample size and to determine whether the data were homogeneously distributed before the EFA was conducted, the following values were determined: KMO: 0.921; c (351): 6110.786; BTS: p<0.05.
After the analysis performed during the validity and reliability study of the Turkish version of the OAS, the number of the items was reduced to 27. Of the items, in the measurement tool, item 18 whose factor loading was low and items 22, 23 and 29 which overlapped were excluded from the analysis. The remaining 27 items within the scope of the analysis were collected under five sub-dimensions: bone physiology, preventive behaviors, risk factors, exercise and characteristics of osteoporosis (Table 2).
The factor loadings of the OAS range between 0.509 and 0.820. These factors account for 66.165% of the total variance. As for the sub-dimensions, bone physiology, preventive behaviors, risk factors, exercise and characteristics of osteoporosis accounted for 14.28%, 14.25%, 13.53%, 12.22% and 11.89% of the total variance respectively (Table 2).
Confirmatory Factor Analysis
According to the CFA, the structural equation modeling results of the scale were significant at the level of p<0.05, and that the 27 items and five sub-dimensions constituting the scale were related to the scale structure (Figure 1, Table 3).
The model was modified. During the modification, variables that decreased the fit were determined, and new covariates were created for those with high covariance among residual values (e1-e2; e5-e6; e8-e13; e9-e13; e23-e24). According to the renewed fit indices after the modification, root mean square error of approximation (RMSEA) (0.08), goodness-of-fit index (GFI) (0.85), comparative fit index (CFI) (0.88), root mean squared residual (RMR) (0.036) and c2 (3.384) (p<0.05) values indicated the fit of the model was at an acceptable level (Figure 1, Table 3).
Reliability Analysis
According to the reliability analysis of the OAS and its sub-dimensions, Cronbach’s a reliability coefficients were as follows: 0.949 for the overall OAS, 0.866 for the bone physiology sub-dimension, 0.862 for the preventive behaviors sub-dimension, 0.882 for the risk factors sub-dimension, 0.866 for the exercise and 0.858 for the characteristics of osteoporosis sub-dimension. The analysis of the item-total score correlations of the OAS revealed that the distribution of the item total correlation values of the scale items ranged between 0.497 and 0.739 (p<0.05) (Table 2).
Internal Consistency of the Subscales
A statistically positive and significant relationship was determined between the mean scores for the sub-dimensions of the OAS obtained from the test-retest application. The dependent samples t-test conducted to determine whether there was a difference between the mean scores obtained from the scale and its sub-dimensions demonstrated that the difference between the two application scores was not statistically significant (p>0.05).
Disscussion
Content Validity of the OAS
If the CVI is above 0.80, then the measurement tool can be said to have content validity (21,23,24). While the CVI of the sub-dimensions of the Turkish version of the OAS ranged between 0.60 and 1.00, the CVI of the overall OAS was 0.95.
Construct Validity of the Scale
According to the literature, KMO values should be between 0.5 and 1.0, and values below 0.5 indicate that the sample is not sufficient for the factor analysis (25). In order to perform the factor analysis, whether the data set provides the multivariate normality is assessed by the BTS. A p-value less than 0.05 in the Bartlett’s test of sphericity means that the data provide multivariate normal distribution (25). In the present study, KMO was 0.921; c (351) was 6110.786; BTS p-value was less than 0.05 These results indicate that the sample size was sufficient for the EFA and the distribution of the data set was homogeneous.
In the literature, in multi-factor measurement tools, if item factor loadings explain 40% to 60% of the total variance, this is accepted as sufficient (26,27). According to studies, the higher the variance ratio is, the stronger the factor structure of the scale is. In the present study, it was determined that the Turkish version of the OAS explained 66.165% of the total variance and had five sub-dimensions with an eigenvalue above 1. That the variance explained was high in the present study indicates that the scale accurately measured the property to be measured. In addition, these results indicate that the Turkish version of the OAS is consistent with the construct of the original scale which has five sub-dimensions.
In the literature, it is stated that if the loading value of an item of a measurement tool is less than 0.32 in more than one sub-dimension and that if the difference between the factor loadings of the item in two sub-dimensions is less than 0.10, which indicates that the items overlap, then this item should be removed from the scale (25,28).
It is stated that if an item is to be included in any factor of a measurement tool, then its factor loading value should be at least 0.32 (25). Therefore, based on the results of the EFA performed in the present study, the items 18, 22, 23, 29 were removed from the scale because the item 18 had a factor loading less than 0.32 and the difference between the factor loading values of the items 22, 23, 29 in two factors was less than 0.10.
CFA was not performed in the original study. In the CFA conducted in the present study, a value of ≤0.08 for RMSEA indicates acceptable fit, while ≤0.05 indicates perfect fit (29).
According to the first level multifactorial CFA results, in the indexes of the Turkish version of the OAS, RMSEA was 0.08, which indicated the fit was acceptable. Although there are different citations in the literature regarding the criteria of goodness of fit, values of 0.80≤CFI≤0.90, 0.80≤ normed fit index ≤0.95, 0.85≤GFI≤0.90, and 0≤RMR≤0.080 indicate acceptable fit (29,30). According to the goodness of fit indices of the Turkish version of the OAS, CFI was 0.88, GFI was 0.85, RMR was 0.036, c2 was 3.384 (p<0.05) and RMSEA was 0.08, which indicated that an acceptable level of fit was achieved (Table 3). After the CFA, it was determined that the factor loadings of the Turkish version of the OAS were over 0.50 and varied between 0.64 and 0.82.
Reliability of the OAS
The Cronbach’s a reliability coefficient ranges between 0 and 1. The fact that a scale’s Cronbach’ a coefficient is as close to 1 as possible means that the scale is highly reliable (31,32). If the Cronbach’s a coefficient is less than 0.40, the scale is not reliable, if it is between 0.40 and 0.59, its reliability is low, if it is between 0.60 and 0.79, it is regarded quite reliable, and if it is between 0.80 and 1.00, it is considered very reliable (23,29). The Cronbach’s a reliability coefficient of the OAS developed by Choi et al. (3) (2008) was 0.948.
In the current study, the Cronbach’s a reliability coefficient of the Turkish version of the OAS was 0.949 for the overall scale, 0.866 for the “bone physiology” sub-dimension, 0.862 for the “preventive behaviors” sub-dimension, 0.882 for the “risk factors” sub-dimension; 0.886 for the “exercise” sub-dimension, and 0.858 for the “characteristics of osteoporosis” sub-dimension (Table 2). That the Cronbach’s a values of the OAS and its sub-dimensions are over 0.80 in the present study indicates that both the overall OAS and its sub-dimensions are highly reliable.
In item selection, there is no certain standard indicating that the item’s reliability will be insufficient when the item’s correlation coefficient is lower than which value of the correlation coefficient. However, according to Büyüköztürk (20), Şencan (27), and Karasar (33), if the correlation coefficient is greater than 0.25, 0.30 and 0.50 respectively, the item is regarded reliable. It is recommended that items with low correlation should be removed from the measurement tool (27,28,33). In the present study, the correlation coefficients of the 27-item scale ranged between 0.497 and 0.739 and thus they were accepted as statistically significant for all the items of the OAS (Table 2).
The minimum required value for the item-total test correlation to be sufficient is 0.30 (34). In the current study, the item-total correlation was above 0.30.
In order to determine the distinctiveness of the items in the scale, the raw scores obtained from each factor were ranked in the ascending order from the lowest to the highest, and the mean scores of the groups in the bottom 27% and top 27% were compared with the independent samples t-test. The result of the comparison indicated a significant difference between the mean lower and upper group item scores for each sub-dimension in terms of all the items at the p<0.05 level. Therefore, it can be said that the sub-dimensions of the scale are distinctive in terms of measuring the desired quality.
These results indicate that all the items of the Turkish version of the OAS have adequate correlation with the total score of the five sub-dimensions, which suggests that the item reliability of the sub-dimensions of the scale is high (Table 2).
Strength and Limitations
One of the strengths of the present study is that in the study, the awareness of physicians, midwives, nurses and all other healthcare professionals who served the same population in primary health care for many years about osteoporosis was investigated. The main limitation of the study was that it was conducted only with healthcare professionals in a single center.
Conclusion
According to the results of the present study, The Turkish version of the OAS is a valid and reliable measurement tool and it can be used to assess “awareness of osteoporosis”. Evaluation of the awareness of students studying in health-related departments such as medicine, midwifery, nursing, etc. about osteoporosis from their school years and the elimination of any missing/incorrect information before graduation are very important. The quality of the healthcare services provided by healthcare workers will be improved through the in-service training programs to be organized to assess their awareness of osteoporosis. It may be recommended that the OAS should be administered to the healthy population in order to increase the society’s awareness of osteoporosis, the importance of early diagnosis of osteoporosis and prevention of osteoporosis, and if appropriate, it should be included in the screenings performed especially in primary care units.
Ethics
Ethics Committee Approval: To conduct the study, ethical approval from a Ege University Medical Research Ethics Committee was obtained (decision no: 19-2T/23, date: 06.02.2019).
Informed Consent: From the healthcare professionals who agreed to participate in the study, verbal and written consent was obtained.
Peer-review: Internally peer-reviewed.
Authorship Contributions
Concept: S.O.A., M.G.Ö., Design: S.O.A., M.G.Ö., Data Collection or Processing: S.O.A., R.M., Analysis or Interpretation: S.O.A., M.G.Ö., R.M., Literature Search: S.O.A., Writing: S.O.A., M.G.Ö., R.M.
Conflict of Interest: No conflict of interest was declared by the authors.
Financial Disclosure: The authors declared that this study has received no financial support.