http://orcid.org/0000-0003-4594-1202
Figures
Abstract
Introduction
The use of generic medicines is a cost-effective policy, often dictated by fiscal restraints. To our knowledge, no fully validated tool exploring the students’ knowledge and attitudes towards generic medicines exists. The aim of our study was to develop and validate a questionnaire exploring the knowledge and attitudes of M.Sc. in Health Care Management students and recent alumni’s towards generic drugs in Greece.
Materials and methods
The development of the questionnaire was a result of literature review and pilot-testing of its preliminary versions to researchers and students. The final version of the questionnaire contains 18 items measuring the respondents’ knowledge and attitude towards generic medicines on a 5-point Likert scale. Given the ordinal nature of the data, ordinal alpha and polychoric correlations were computed. The sample was randomly split into two halves. Exploratory factor analysis, performed in the first sample, was used for the creation of multi-item scales. Confirmatory factor analysis and Generalized Linear Latent and Mixed Model analysis (GLLAMM) with the use of the rating scale model were used in the second sample to assess goodness of fit. An assessment of internal consistency reliability, test-retest reliability, and construct validity was also performed.
Results
Among 1402 persons contacted, 986 persons completed our questionnaire (response rate = 70.3%). Overall Cronbach’s alpha was 0.871. The conjoint use of exploratory and confirmatory factor analysis resulted in a six-scale model, which seemed to fit the data well. Five of the six scales, namely trust, drug quality, state audit, fiscal impact and drug substitution were found to be valid and reliable, while the knowledge scale suffered only from low inter-scale correlations and a ceiling effect. However, the subsequent confirmatory factor and GLLAMM analyses indicated a good fit of the model to the data.
Citation: Domeyer PJ, Aletras V, Anagnostopoulos F, Katsari V, Niakas D (2017) Development and validation of a tool to assess knowledge and attitudes towards generic medicines among students in Greece: The ATtitude TOwards GENerics (ATTOGEN) questionnaire. PLoS ONE 12(11): e0188484. https://doi.org/10.1371/journal.pone.0188484
Editor: Mohammad Shahid, Aligarh Muslim University, INDIA
Received: August 23, 2015; Accepted: November 8, 2017; Published: November 29, 2017
Copyright: © 2017 Domeyer et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information file.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Generic drugs are a means of cost-effective medical treatment favoring both patients and the healthcare system. Doctors are often urged to prescribe generic drugs as far as possible [1], although multiple strategies are necessary to change doctor prescribing habits [2]. The global deterioration of the economic situation and demographic change makes this urge more prominent nowadays, given the increased healthcare systems economic burden along with the imposed economic restraints. In particular, Greece is a country severely hit by the economic crisis. The imperative need for fiscal reform has led to fiscal austerity and budgetary cuts in many sectors, including healthcare expenditure. Strikingly, the pharmaceutical spending during the years 2010, 2011 and 2012 was sharply reduced (annual decrease of 10.1%, 9.0% and 18.4% respectively) and dropped from the 2009 level of 5.6billion euros to 2.37billion in 2013 [3,4]. In order to achieve those targets, policies have been implemented leading to drug expenditure rationalization and the introduction of a recommended list of medicines focusing on generic drugs.
However, despite the obvious cost-effectiveness of generic drugs considerable mistrust has been noted worldwide regarding their safety and efficacy and several studies indicate difficulties associated with treatment adherence or compliance [5–10]. According to Keenum et al [11], although generic medications are considered less expensive and of better value than their counterparts, fewer than 50% of the prescribers chose generic drugs for themselves. It seems that respondents’ attitudes towards generics are shaped from personal beliefs and knowledge–the latter being dissipated from mass media, friends and family-, previous experience with generics and trust to the prescriber [12].
Healthcare students are the future prescribers, dispensers and patients' advisors regarding medications, which makes them an important study target. Several studies have been published regarding their knowledge and perceptions regarding generic medicines [13–21]. Three of the abovementioned studies [17, 20, 21] are based on the only available validated questionnaire, although no validation study has been published. The use of such a validated tool would definitely increase the reliability and validity of studies on this important issue and would allow the direct comparability of studies. To our knowledge, only one study by Figueiras et al [22], in which a generic medicines scale was also created, was recently validated among the Malaysian population [23]. However, the scale underwent partial validation with the use of exploratory factor analysis only, without the use of any Structural Equation Modeling (SEM) technique.
To our knowledge, no study has emphasized on students and alumni of M.Sc. programs of Health Care Management, which are future health care leaders and pharmaceutical policy shapers. The future market share of generics will largely depend on the policy makers’ strategy, which is in turn influenced by their knowledge and attitude towards generics. The aim of our study was to develop and validate a questionnaire exploring the knowledge and attitudes of M.Sc. in Health Care Management students and that of recent alumni’s towards generic drugs.
Materials and methods
Study sample
The study sample comprised of all students and recent (up to 1 year from graduation) alumni of the M.Sc. program of Health Care Management at the Hellenic Open University. The Hellenic Open University provides distance education in both undergraduate and postgraduate levels. Contrary to the majority of other Public Universities, the students should pay fees for the cost of their studies. Students are selected without exams, by a public electronic lottery. Priority is given to candidates being at least 23 years old. For the abovementioned reasons, most students are usually older, with mean age being 32 years [24] and employed.
Instrument creation
The MEDLINE, EMBASE, Cochrane and CINAHL databases were scanned for full articles published between January 2000 and November 2013, regarding knowledge/attitudes/perceptions/behaviours/opinions/views of doctors, pharmacists, consumers, patients and students towards generic drug use. Only full original research papers written in English were included. Title and abstract fields were searched for publications containing the terms: generic; knowledge; attitude; perception; behaviour; opinion; view; healthcare professional; physicians; doctors; pharmacists; consumers; patients; students. Boolean operators (AND/OR) were used to combine search components and to yield as many search results as possible. MeSH (Medical Subject Headings) terms were used where possible.
The databases' search returned a total of 1446 articles, the majority of which (873) related to consumers/patients/students; the PubMed searches proved the most successful, returning 972 articles. After the removal of duplicate findings, the reading of titles and abstracts of the remaining articles resulted in excluding those whose content was not relevant to this study. The identification of such articles was due to the design of the searches, using the selected search terms, in a deliberate attempt to ‘capture’ as many published articles in the field as possible. As a result, 60 articles were retained, eleven of which were not in English. Therefore, 49 papers [5–10, 11, 13, 14, 22, 25–63] were reviewed for the design of the study's questionnaire. Questions related to the aims of our study that commonly appeared in these studies and questions that yielded statistically significant differences among groups were retained. Accordingly, items were created and carefully phrased in order to avoid ambiguity, biases, double-barreled and double negative questions. Five sociodemographic questions regarding age, sex, marital and professional status and profession were added.
A preliminary version of the instrument was handed over to 10 researchers and experts in the fields of generics, attitude measurements and questionnaire construction. Face to face interviews were conducted by two of the researchers and recorded, during which they were asked to complete the questionnaire and highlight ambiguous or problematic items. They were also asked to propose additional items regarding aspects of a person’s attitude towards generic medicines not already captured by the instrument and to delete any trivial items, thus resulting in enhanced content validity and reliability. All necessary modifications, based on divergence of opinions between experts, were made to the instrument, which was subsequently handed over to 20 subjects randomly selected from the students’ and alumni pool. Again the same procedure was followed, which resulted in only minor textual changes and in the final version of the instrument to be handed over. This final version of the ATtitude TOwards GENerics (ATTOGEN) questionnaire consisted of close-ended Likert-type questions, apart from sociodemographic questions, with possible answers ranging from strongly agree (= 1) to strongly disagree (= 5).
Instrument administration
The study took place between March and August 2014. The instrument was uploaded to Google Drive (Google Inc, California, USA). A web link was sent to all participants, who were asked to complete the questionnaire. In order to avoid missing data, all fields were marked as required and the questionnaire could not be submitted unless all questions had been answered. Monthly reminders were sent from the M.Sc. Director to the participants, in order to maximize the response rate.
Ethics
Approval of the study by the Review Board of the Hellenic Open University was granted. Participants were reassured that their anonymity would be respected. Although the importance of achieving a high response rate was highlighted, participants were free to decide whether to complete the questionnaire or not. Furthermore, it was made clear that no recriminations would occur irrespective of their decision. There was no compensation for the participants. All participants provided their written informed consent to participate in this study. The informed consent form was sent electronically to the participants, who signed it and returned it by email or fax.
Statistical analysis
The instrument created was psychometrically tested and subscales were created with the use of exploratory factor analysis. Subsequently, confirmatory factor analysis and generalized linear latent and mixed models were used to test the fit of the models to our data.
Exploratory factor analysis.
To begin, the sample was randomly split into two halves. In the first sample, the exploratory factor analysis (EFA) was used for the creation of multi-item scales. The first step towards this direction was to check our data for normality, since factor analysis requires a multivariate normal distribution of the data. Although some degree of skewness was noted to our data, for large samples, like in our study, the multivariate Central Limit Theorem can be relied upon for a good approximation of normality, thus enabling the use of factor analysis [64]. Furthermore, we inspected the polychoric correlation matrix, in order to check whether our data were suitable for exploratory factor analysis.
Cronbach’s alpha coefficient was used to test for internal consistency [65]. Polychoric correlations were used since responses were expressed on a Likert (ordinal) scale [66,67]. The polychoric correlation matrix was created and ordinal versions of the alpha coefficients were also calculated. We also calculated Spearman’s coefficient, as an additional tool for the assessment of inter-item correlations, which is useful for ordinal data and robust to outliers, causing potential skewness.
We subsequently run the Bartlett’s test for sphericity to test the appropriateness of the factor model. We also calculated the Kaiser-Meyer-Olkin measure of Sampling Adequacy, which tests whether the partial correlations among variables are small; levels more than 0.5 are necessary for a satisfactory factor analysis to proceed [68], whereas values between 0.8 and 0.9 are considered great [69].
The exploratory factor analysis was carried out in the first half of the sample with the use of the polychoric correlation matrix. In order to further minimize the effect of non-normally distributed data, the principal axis factoring method was chosen. Following other research work, we also performed a typical factor analysis based on Pearson correlations, for comparative reasons, since the Likert scale could be treated as an interval or ratio scale [69]. The promax rotation was initially performed, assuming that the factors are correlated. In order to decide on the number of factors to retain, we used the Kaiser’s criterion, according to which all factors with eigenvalues<1 are dropped and also assessed the scree plot, as well as the results of the confirmatory factor analysis that followed [69]. Given that the Kaiser’s criterion is known for its tendency to overextract factors [70], we ultimately decided to adopt a less stringent approach, in order to reach an informative but relatively parsimonious model. The rotated factor loadings were computed; the commonly used threshold value of 0.5 was adopted for factor extraction [71,72]. The factor correlation matrix was calculated. According to Tabachnick & Fidell (2007) [73], if the absolute value of correlations exceed 0.32, then there is 10% (or more) overlap in variance among factors, enough variance to warrant oblique rotation. Finally, the predicted scoring coefficients were calculated and Cronbach’s alpha coefficients were computed for each factor separately and assessed according to the literature’s rule of thumb [74].
Structural equation modeling technique.
Data from the second sample were used to implement the SEM technique, in order to test the fit of the models to our data. The confirmatory factor analysis and the Generalized Linear Latent and Mixed Model (GLLAMM) analysis were both used, as each method adopts a different statistical approach.
Confirmatory factor analysis.
During the confirmatory factor analysis, we tried to assess the fit of the data to the model. Adjustments of the items contained in each factor were made based on post-estimation absolute and relative fit indices. Fit indices which were assessed included: i) the chi-squared test [75], ii) the Root Mean Square Error of Approximation (RMSEA), whose values of 0.06 or less are indicative of an acceptable model fit [76], iii) the Root Mean Square Residual (RMSR), whose values of 0.08 or less are indicative of an acceptable model fit [77] and iv) the Tucker Lewis Index (TLI) and the Comparative Fit Index (CFI), whose values of 0.9 or more are indicative of an acceptable model fit, with values of 0.95 and over being considered as a much better fit [78,79]. The squared multiple correlations (R2) for each observed variable of the model were also calculated [80].
Test-retest reliability.
The sensitivity of the scale to external factors was assessed by examining test-retest reliability. The test-retest reliability was assessed using polychoric correlation coefficients, which were calculated for each subscale.
Construct validity.
Multi-trait analysis was used to assess the construct validity of our tool; convergent and discriminant validity were therefore examined. Convergent validity refers to the degree to which two construct measures that theoretically should be related are truly related. Therefore, the item scale correlation between an item and the scale in which it allegedly belongs to should be expected to be higher than its correlations with irrelevant scales. According to Fornell and Lacker [81], average variance extracted (AVE) estimates should be 0.50 or above. On the other hand, discriminant validity refers to the degree to which two construct measures that theoretically should be unrelated are truly unrelated. In order that evidence for discriminant validity exists, the shared variances between paired constructs should be lower than the AVE of the individual constructs. Moreover, item-scale and inter-scale correlations were calculated. Due to the ordinal nature of the data, we computed the Pearson, Spearman and polychoric correlations, corrected for overlap, as per our previously published work [82].
Ceiling and floor effects.
Furthermore, ceiling and floor effects were examined by each created subscale. The presence of ceiling and floor effects compromises the content validity of the scale. In particular, the ceiling and floor effects are encountered when several subjects’ responses consistently fall into the highest and lowest category, respectively.
Generalized linear latent and mixed model analysis.
In the case of the GLLAMM analysis, two models, the partial credit model (PCM) [83] and the rating scale model (RSM) [84] were tested for each factor separately. The PCM is an extension of the Rasch model to polytomous items with ordered response categories. The RSM is a special case of the PCM, used when response categories have the same meaning for all items and under the assumption that the differences in the step difficulties for different categories are the same for all items [85]. In other words, the response scale is fixed for all items, i.e. the response threshold parameters are assumed to be identical across items.
To compare the various models’ output in both confirmatory factor analysis and GLLAMM analysis, a likelihood ratio test was performed. Finally, estimates and predictions were made based on the chosen GLLAMM model.
Statistical analysis was performed using Stata version 13.1 (Stata Corp., Texas, USA). Two-tailed p values <0.05 were considered statistically significant.
Results
Among 1402 persons contacted, 986 persons completed the questionnaire (response rate = 70.3%). The mean age of the participants was 39.8±7.11 years. The sociodemographic characteristics of our study sample are presented in Table 1.
The questionnaire that was handed over to the participants of our study comprised 24 items, apart from the sociodemographic questions. Of the 24 items, 2 items (“I know if my current medications include generic medicines”) and (“I have noticed substantial differences between brand name and generic medicines”) were related to the intake of medications and were excluded since the majority of the participants (665/986, 67.4%) were not under medications. On inspecting the polychoric correlation matrix, two variables (“I am satisfied with the information I get regarding generic medicines” and “Among two generic medicines, I would trust more the one that would be manufactured in Greece”) were found to have low correlation coefficients with the rest of the variables, were considered unsuitable for factor analysis and were hence excluded. The remaining items of the questionnaire that underwent statistical analysis, along with their descriptive statistics, are presented in Table 2.
Exploratory factor analysis
Cronbach’s alpha and ordinal Cronbach’s alpha coefficients for the entire questionnaire were both high and did not differ substantially (0.871 and 0.893 respectively). The Bartlett’s test for sphericity was statistically significant (p<0.0001) and the Kaiser-Meyer-Olkin measure of sampling adequacy was 0.858. The factor analysis based on polychoric correlations with subsequent promax rotation yielded 6 factors, which account for 96.1% of the total variance and are shown in Table 3. Due to negative factor loadings, item 9 was inversely recoded. The factor correlation matrix was also calculated (Table 4). Most correlations exceeded 0.32; therefore, oblique rotation was performed.
Confirmatory factor analysis
A confirmatory factor analysis was subsequently run to explore the strength of the proposed 20-item scale, which yielded moderate fit results (RMSEA = 0.061, RMSR = 0.072, TLI = 0.912, CFI = 0.928). The 6-scale approach seemed to have the best fit. The increase of factors retained from 6 to 7 (matching the scree plot results) or the decrease to 5 or 4 factors (nearly matching or matching the results of the Kaiser’s criterion) provided even worse fit results at the subsequent confirmatory factor analysis. After examining the exclusion of item(s) with the weakest factor loadings and the highest uniqueness scores on the 6 factor model, it was found that deleting items 14 and 15, thus reducing the increased number in scale 1 from 6 to 4, would provide a better fit (RMSEA = 0.059, RMSR = 0.063, TLI = 0.927, CFI = 0.942). In an attempt to further ameliorate the model fit, the free estimation of the covariance between the error terms of items 1 and 2 as well as items 17 and 18 was allowed. These specific pairs of items were chosen, because the variables within each pair are measuring similar attitudes and have very strong correlation with each other. The resulting model provided the best fit (RMSEA = 0.050, RMSR = 0.048, TLI = 0.948, CFI = 0.960). Indeed, this newly introduced correlation was statistically significant according to the likelihood ratio test (p<0.0001). Table 5 presents the squared multiple correlations (R2) for each observed variable of the model. Most of them ranged from acceptable to good, apart from items 1, 2 and 7. By introducing the correlated error term for items 1 and 2, the R2 of those variables decreased. This is fairly rational, since the covariance between these two items no longer contributes to explaining the factor “knowledge”, given that the covariance between the error terms δ1 and δ2 is now freely estimated, as suggested by Kolenikov.[86] Regarding variable 7, its low squared multiple correlation is consistent with its comparatively lower factor loading in explanatory factor analysis.
The Cronbach’s alpha coefficients were calculated for each generated scale andwere found to be good for scales 1, 2 and 3 (0.821, 0.805 and 0.847 respectively), acceptable for scales 5 and 6 (0.777 and 0.719) and questionable for scale 4 (0.655). Table 6 presents the internal consistency reliability of the 6 scales generated from the factor analysis, including Cronbach’s alpha coefficients, mean and median values.
Test-retest reliability
Thirty randomly selected participants were contacted for a second completion of the questionnaire, which took place two weeks after the initial administration; among them twenty eight completed it (response rate 93.3%). The test-retest reliability results were also good, with polychoric correlation coefficients all above 0.80.
Construct validity
The results of the multi-trait analysis using polychoric correlations are shown in Table 7. Pearson and Spearman correlations were also calculated, with similar results. Inter-scale correlations were calculated and all of them were found highly statistically significant (p<0.0001). An exception was noted for the “Knowledge” scale, regarding inter-scale correlations with “State audit”(p = 0.199), “Fiscal impact” (p = 0.183) and “Drug substitution” (p = 0.866) scales. The AVE estimates were computed and were found to be above 0.5, except for scale 4 (0.496), and higher than the corresponding estimated squared correlations. Therefore, there was adequate overall evidence for convergent and discriminatory validity, respectively.
Ceiling and floor effects
Regarding ceiling effects, the percentages of observations falling into the best attitude score category for the six summated scales were 2.03%, 6.29%, 2.84%, 9.94%, 44.62% and 2.43%. The respective percentages for the flooring effects were 1.83%, 2.43%, 4.87%, 0.41%, 0.20% and 19.41%. Only the “Knowledge” scale presented a ceiling effect.
Generalized linear latent and mixed model analysis
The GLLAMM analysis was also performed on each scale separately. Both the partial credit model and the rating scale model were tested. Since the likelihood ratio test was highly insignificant in all scales (p>0.9) and the rating scale model is more parsimonious, the latter was chosen. According to this model, prediction estimates were calculated and category probability curves (CPCs) were created. S1–S3 Figs show CPCs for each of the three items of the “fiscal impact” scale, which indicate good model fit and no category disordering. Corresponding graphs were also created for the other scales, with similar results.
Discussion
The ATTOGEN questionnaire was developed and validated in order to assess the knowledge and attitude of students towards generic medicines. The questionnaire initially contained 24 items. Although two questions related to the intake of medications were excluded, the content validity of our questionnaire does not seem to be significantly compromised because of the existence of items 1 and 2, which ask participants whether they know what generics are as well as the differences between brand name and generics. The combination of exploratory and confirmatory factor analysis revealed items with low correlation coefficients and items that substantially worsened the model fit, which were also excluded from the composite scales. The final instrument consisted of 6 scales, namely trust (4 items), drug quality (3 items), state audit (3 items), fiscal impact (3 items), knowledge (3 items) and drug substitution (2 items). Because the criteria regarding the number of factors to be retained during the exploratory factor analysis were discordant–Kaiser’s criterion suggested 4 factors and the scree plot 7 factors -, the “trial and error” approach was adopted, using confirmatory factor analysis to test the fit of various models. The choice of the final model was based on the goodness of fit of the models, while respecting the content validity of the questionnaire.
The ATTOGEN questionnaire was found to have good overall internal consistency. Regarding the psychometric properties of the scales, satisfactory internal consistency reliability, test-retest reliability and construct validity were noted, with the exceptions of scale 5 (knowledge) with low inter-scale correlations. A ceiling effect was only present in the knowledge scale, which is an effect that seems to appear quite commonly in the literature [77,87]. Further studies may be needed to examine a potential improvement or exclusion of the knowledge scale without jeopardizing content validity. Nevertheless, the confirmatory factor analysis and the GLLAMM analysis both indicated a good overall fit. We believe that the confirmation of this finding with the concomitant use of two different SEM techniques is particularly meaningful.
In the study by Figueiras et al [22], in which a generic medicines scale was also created and recently validated among the Malaysian population [23], the initial number of items was not too different (18 items vs 24 items in our study). However, the exploratory factor analysis led to two factors only, namely efficacy and similarity, instead of six factors in our study. Moreover, the fit of this model to the data was not assessed, which could have altered the final results.
The strengths of the present study include the creation of a validated questionnaire regarding generics, which fills an important gap in the international literature, with the use of a large number of participants who completed the questionnaire with no missing values. Another strength was the concomitant use of two SEM techniques, both confirming the goodness of fit of our results. Our study, however, bears certain limitations that should be addressed. Firstly, our sample may not be representative of Greek students in conventional universities; health management students may express different views compared to other students, because of their specific academic background and because the majority of them are already working in clinical practice as healthcare providers. On the other hand, students from the Hellenic Open University belong to various age, economic and sociodemographic groups, because of the nature of the study environment (distance learning), which may somewhat increase the generalizability of the study’s results. In addition, due to the lack of any other similar questionnaire translated in Greek, the convergent validity of the questionnaire was not fully assessed. Finally, the low inter-scale correlations and the ceiling effect noted in scale 5 (knowledge) may deteriorate the psychometric properties of the questionnaire.
Conclusions
To conclude, the use of exploratory and confirmatory factor analysis as well as GLLAMM modeling technique revealed five reliable and valid multi-item scales: trust, drug quality, state audit, fiscal impact and drug substitution. A sixth scale (knowledge) was found to suffer from relatively low internal consistency reliability and low inter-scale correlations. However, both the confirmatory factor and the GLLAMM analyses indicated an adequate fit, suggesting that the uncompromised use of all six scales could be justified. Further studies should be elaborated in order to further test the questionnaire to large representative samples of the Greek population, in order to ascertain its generalizability.
Supporting information
S1 Fig. Category probability curves for item 7 under the rating scale model.
https://doi.org/10.1371/journal.pone.0188484.s001
(TIF)
S2 Fig. Category probability curves for item 10 under the rating scale model.
https://doi.org/10.1371/journal.pone.0188484.s002
(TIF)
S3 Fig. Category probability curves for item 11 under the rating scale model.
https://doi.org/10.1371/journal.pone.0188484.s003
(TIF)
S1 Table. Sociodemographic characteristics of the study sample.
https://doi.org/10.1371/journal.pone.0188484.s005
(DOCX)
S2 Table. Descriptive statistics of the items.
https://doi.org/10.1371/journal.pone.0188484.s006
(DOCX)
S3 Table. Factor analysis with polychoric correlations (rotated factor loadings).
https://doi.org/10.1371/journal.pone.0188484.s007
(DOCX)
S4 Table. Factors correlation matrix in Exploratory Factor Analysis with oblique rotation.
https://doi.org/10.1371/journal.pone.0188484.s008
(DOCX)
S5 Table. Squared multiple correlations (R2) for each observed variable of the model.
https://doi.org/10.1371/journal.pone.0188484.s009
(DOCX)
S6 Table. Internal consistency reliability and scale scores.
https://doi.org/10.1371/journal.pone.0188484.s010
(DOCX)
S7 Table. Multi-trait analysis with polychoric correlations.
https://doi.org/10.1371/journal.pone.0188484.s011
(DOCX)
Author Contributions
- Conceptualization: PJD, DN.
- Data curation: PJD, VK.
- Formal analysis: PJD.
- Investigation: PJD, VK, DN.
- Methodology: PJD, DN.
- Project administration: DN.
- Resources: DN.
- Supervision: DN.
- Validation: VA, FA.
- Visualization: PJD.
- Writing – original draft: PJD.
- Writing – review & editing: PJD, VA, FA, DN.
References
- 1. Godman B, Bishop I, Finlayson AE, Campbell S, Kwon HY, Bennie M. Reforms and initiatives in Scotland in recent years to encourage the prescribing of generic drugs, their influence and implications for other countries. Expert Rev Pharmacoecon Outcomes Res. 2013 Aug; 13(4):469–82. pmid:23977975
- 2.
Organisation for Economic Co-operation and Development. OECD Health Statistics 2014: How does Greece compare? Available: http://www.oecd.org/els/health-systems/Briefing-Note-GREECE-2014.pdf Accessed 10 August 2015.
- 3.
World Health Organisation. Greek crisis fallout is an opportunity for health. Available: http://www.who.int/bulletin/volumes/92/1/14-030114/en/ Accessed 10 August 2015.
- 4. Barbui C, Conti V. Adherence to generic v. brand antidepressant treatment and the key role of health system factors. Epidemiol Psychiatr Sci. 2014 Dec 18:1–4.
- 5. Hakonsen H, Eilertsen M, Borge H, Toverud EL. Generic substitution: additional challenge for adherence in hypertensive patients? Curr Med Res Opin. 2009; 25:2515–21. pmid:19708764
- 6. Roman B. Patients’ attitudes towards generic substitution of oral atypical antipsychotics: a questionnaire-based survey in a hypothetical pharmacy setting. CNS Drugs. 2009; 23:693–701. pmid:19594198
- 7. Bulsara C, McKenzie A, Sanfilippo F, Holman CD, Emery JE. ‘Not the full Monty’: a qualitative study of seniors’ perceptions of generic medicines in Western Australia. Aust J Prim Health. 2010; 16:240–5. pmid:20815994
- 8. Hakonsen H, Toverud EL. Special challenges for drug adherence following generic substitution in Pakistani immigrants living in Norway. Eur J Clin Pharmacol. 2011; 67:193–201. pmid:21161197
- 9. Sharrad AK, Hassali MA. Consumer perception on generic medicines in Basrah, Iraq: preliminary findings from a qualitative study. Res Social Adm Pharm. 2011; 7:108–12. pmid:21397885
- 10. Hensler K, Uhlmann C, Porschen T, Benecke R, Rösche J. Generic substitution of antiepileptic drugs–A survey of patients’ perspectives in Germany and other German-speaking countries. Epilepsy Behav. 2013; 27:135–9. pmid:23416284
- 11. Keenum AJ, Devoe JE, Chisolm DJ, Wallace LS. Generic medications for you, but brand-name medications for me. Res Social Adm Pharm. 2012 Nov-Dec; 8(6):574–8. pmid:22357268
- 12. Allenet B, Golay A. What are patients' attitudes towards generic drugs? The example of metformin. Rev Med Suisse. 2013 May 8; 9(385):1005–9. pmid:23750394
- 13. Hassali MA, Kong DC, Stewart K. A comparison between senior medical students' and pharmacy pre-registrants' knowledge and perceptions of generic medicines. Med Educ. 2007 Jul; 41(7):703–10. pmid:17614892
- 14. Hassali MA, Stewart K, Kong DC. A national survey on knowledge and perceptions of senior medical students in Australia about generic medicines. Med J Aust. 2008 Jan 21; 188(2):123–4. pmid:18205593
- 15. Hassali MA, Kong DC, Stewart K. Knowledge and perceptions of recent pharmacy graduates about generic medicines. Pharmacy Education. 2007 Jan; 7(1):89–95.
- 16. Siam MKS, Khan A, Khan TM. Medical and pharmacy students’ knowledge and perceptions about generic medicines in Bangladesh. J. Pharm. Health Serv. Res. 2013;4(1):57–61.
- 17. Jamshed SQ, Ibrahim MI, Hassali MA, Sharrad AK, Shafie AA, Babar ZU. Understanding and perceptions of final-year Doctor of Pharmacy students about generic medicines in Karachi, Pakistan: a quantitative insight. Adv Med Educ Pract. 2015 May 15;6:359–66. pmid:26028981
- 18. Othman GQ, Abdulghani MAM. Assessment of knowledge and perceptions of generic medicines among pharmacy students in Yemeni universities. Pharmacy Education. 2015 April;15(1):93–98.
- 19. Othman GQ, Abdulghani MAM, AlHyi HA, Alkhazzan A, Al-Qadsi F. Knowledge and perceptions of final-year medical students in Yemeni Universities about generic medicines. Yemeni J Med Sci. 2016;10:16–24.
- 20. Gyawali S, Hassali MA, Saha A. A survey exploring the knowledge and perceptions of senior medical students in Nepal toward generic medicines. SAGE Open Med. 2016;4:2050312116662570. pmid:27551423
- 21. Shankar PR, Herz BL, Dubey AK, Hassali MA. Assessment of knowledge and perceptions toward generic medicines among basic science undergraduate medical students at Aruba. Indian J Pharmacol. 2016 Oct;48(Suppl 1):S29–S32. pmid:28031604
- 22. Figueiras MJ, Alves NC, Marcelino D, Cortes MA, Weinman J, Horne R. Assessing lay beliefs about generic medicines: Development of the generic medicines scale. Psychol Health Med. 2009 May; 14(3):311–21. pmid:19444709
- 23. Wong ZY, Hassali MA, Saleem F, Yahaya AHM. Translation and validation of the Malaysian version of Generic Medicine Scale (GMS). J Med Market. 2014; 14(1)32–40.
- 24.
OECD thematic review on recognition of non-formal and informal learning. Country background report: Greece. Athens, 2007.
- 25. Allenet B, Barry H. Opinion and behaviour of pharmacists towards the substitution of branded drugs by generic drugs: survey of 1,000 French community pharmacists. Pharm World Sci. 2003; 25:197–202. pmid:14584225
- 26. Babar ZU, Grover P, Stewart J, Hogg M, Short L, Seo HG, et al. Evaluating pharmacists’ views, knowledge, and perception regarding generic medicines in New Zealand. Res Social Adm Pharm. 2011; 7:294–305. pmid:21272551
- 27. Babar ZU, Stewart J, Reddy S, Alzaher W, Vareed P, Yacoub N, et al. An evaluation of consumers’ knowledge, perceptions and attitudes regarding generic medicines in Auckland. Pharm World Sci. 2010; 32:440–8. pmid:20559730
- 28. Berg MJ, Gross RA, Haskins LS, Zingaro WM, Tomaszewski KJ. Generic substitution in the treatment of epilepsy: patient and physician perceptions. Epilepsy Behav. 2008; 13:693–9. pmid:18589000
- 29. Chong CP, Hassali MA, Bahari MB, Shafie AA. Exploring community pharmacists’ views on generic medicines: a nationwide study from Malaysia. Int J Clin Pharm. 2011; 33:124–31. pmid:21365404
- 30. Chua GN, Hassali MA, Shafie AA, Awaisu A. A survey exploring knowledge and perceptions of general practitioners towards the use of generic medicines in the northern state of Malaysia. Health Policy. 2010; 95:229–35. pmid:20044165
- 31. Fabiano V, Mameli C, Cattaneo D, Delle Fave A, Preziosa A, Mele G, et al. Perceptions and patterns of use of generic drugs among Italian family pediatricians: first round results of a web survey. Health Policy. 2012; 104:247–52. pmid:22217861
- 32. Gossell-Williams M. Generic substitutions: a 2005 survey of the acceptance and perceptions of physicians in Jamaica. West Indian Med J. 2007; 56:458–63. pmid:18303760
- 33. Halme M, Linden K, Kaaria K. Patients’ preferences for generic and branded over-the-counter medicines: an adaptive conjoint analysis approach. Patient. 2009; 2:243–55. pmid:22273245
- 34. Hamann J, Mendel R, Kissling W, Leucht S. Psychiatrists’ decision making between branded and generic drugs. Eur Neuropsychopharmacol. 2013; 23:686–90. pmid:23021566
- 35. Hassali MA, Kong DCM, Stewart K. Generic medicines: Perceptions of general practitioners in Melbourne, Australia. J Generic Med. 2006; 3:214–25.
- 36. Hassali MA, Shafie AA, Awaisu A, Ibrahim MI, Chong Chee P, Jamshed S. Physicians’ views on generic medicines: A narrative review. J Generic Med. 2010; 7:30–9.
- 37. Hassali MAA, Shafie AA, Jamshed S, Ibrahim MIM, Awaisu A. Consumers’ views on generic medicines: A review of the literature. Int J Pharm Pract. 2009; 17:79–88. pmid:20214255
- 38. Heikkila R, Mantyselka P, Ahonen R. Do people regard cheaper medicines effective? Population survey on public opinion of generic substitution in Finland. Pharmacoepidemiol Drug Saf. 2011; 20:185–91. pmid:21254290
- 39. Heikkilä R, Mäntyselkä P, Ahonen R. Why people refuse generic substitution: a population survey of public opinion on generic substitution in Finland. Drugs Ther Perspect. 2012; 28:24–6.
- 40. Himmel W, Simmenroth-Nayda A, Niebling W, Ledig T, Jansen RD, Kochen MM, et al. What do primary care patients think about generic drugs? Int J Clin Pharmacol Ther. 2005; 43:472–9. pmid:16240704
- 41. Jamshed SQ, Hassali MA, Ibrahim MI, Babar ZU. Knowledge attitude and perception of dispensing doctors regarding generic medicines in Karachi, Pakistan: a qualitative study. J Pak Med Assoc. 2011; 61:80–3. pmid:22368910
- 42. Jamshed SQ, Ibrahim MI, Hassali MA, Masood I, Low BY, Shafie AA, et al. Perception and attitude of general practitioners regarding generic medicines in Karachi, Pakistan: A questionnaire based study. Southern Med Review. 2012; 5:22–30. pmid:23093896
- 43. Kersnik J, Peklar J. Attitudes of Slovene general practitioners towards generic drug prescribing and comparison with international studies. J Clin Pharm Ther. 2006; 31:577–83. pmid:17176362
- 44. Kjoenniksen I, Lindbaek M, Granas AG. Patients’ attitudes towards and experiences of generic drug substitution in Norway. Pharm World Sci. 2006; 28:284–9. pmid:17111247
- 45. Kobayashi E, Karigome H, Sakurada T, Satoh N, Ueda S. Patients’ attitudes towards generic drug substitution in Japan. Health Policy. 2011; 99:60–5. pmid:20685003
- 46. Kohli E, Buller A. Factors influencing consumer purchasing patterns of generic versus brand name over-the-counter drugs. South Med J. 2013; 106:155–60. pmid:23380752
- 47. Maly J, Dosedel M, Kubena A, Vlcek J. Analysis of pharmacists’ opinions, attitudes and experiences with generic drugs and generic substitution in the Czech Republic. Acta Pol Pharm. 2013; 70:923–31. pmid:24147373
- 48. McAuley JW, Chen AY, Elliott JO, Shneker BF. An assessment of patient and pharmacist knowledge of and attitudes toward reporting adverse drug events due to formulation switching in patients with epilepsy. Epilepsy Behav. 2009; 14:113–7. pmid:18768168
- 49. Ngo SN, Stupans I, McKinnon RA. Generic substitution in the treatment of epilepsy: patient attitudes and perceptions. Epilepsy Behav. 2013; 26:64–6. pmid:23220461
- 50. Olsson E, Kalvemark SS. Pharmacists’ experiences and attitudes regarding generic drugs and generic substitution: two sides of the coin. Int J Pharm Pract. 2012; 20:377–83. pmid:23134096
- 51. Omojasola A, Gor B, Jones L. Perceptions of generic drug discount programs among low-income women: a qualitative study. Womens Health Issues. 2013; 23:e55–60. pmid:23218868
- 52. Papsdorf TB, Ablah E, Ram S, Sadler T, Liow K. Patient perception of generic antiepileptic drugs in the Midwestern United States. Epilepsy Behav. 2009; 14:150–3. pmid:18845274
- 53. Patel A, Gauld R, Norris P, Rades T. “This body does not want free medicines”: South African consumer perceptions of drug quality. Health Policy Plan. 2010; 25:61–9. pmid:19726560
- 54. Patel A, Gauld R, Norris P, Rades T. Quality of generic medicines in South Africa: perceptions versus reality—a qualitative study. BMC Health Serv Res. 2012; 12:297. pmid:22943592
- 55. Pereira JA, Holbrook AM, Dolovich L, Goldsmith C, Thabane L, Douketis JD, et al. Are brand-name and generic warfarin interchangeable? A survey of Ontario patients and physicians. Can J Clin Pharmacol. 2005; 12:e229–39. pmid:16278495
- 56. Quintal C, Mendes P. Underuse of generic medicines in Portugal: an empirical study on the perceptions and attitudes of patients and pharmacists. Health Policy. 2012; 104:61–8. pmid:22024369
- 57. Rathe J, Larsen P, Andersen M, Paulsen M, Jarbol D, Thomsen J, et al. Associations between generic substitution and patients’ attitudes, beliefs and experiences. Eur J Clin Pharmacol. 2013; 69:1827–36. pmid:23765409
- 58. Sewell K, Andreae S, Luke E, Safford MM. Perceptions of and barriers to use of generic medications in a rural African American population, Alabama, 2011. Prev Chronic Dis. 2012; 9:E142. pmid:22935144
- 59. Shrank WH, Cadarette SM, Cox E, Fischer MA, Mehta J, Brookhart AM, et al. Is there a relationship between patient beliefs or communication about generic drugs and medication utilization? Med Care. 2009; 47:319–25. pmid:19194329
- 60. Shrank WH, Cox ER, Fischer MA, Mehta J, Choudhry NK. Patients’ perceptions of generic medications. Health Aff (Millwood) 2009; 28:546–56.
- 61. Shrank WH, Liberman JN, Fischer MA, Girdish C, Brennan TA, Choudhry NK. Physician perceptions about generic drugs. Ann Pharmacother. 2011; 45:31–8. pmid:21205953
- 62. Tsiantou V, Zavras D, Kousoulakou H, Geitona M, Kyriopoulos J. Generic medicines: Greek physicians’ perceptions and prescribing practices. J Clin Pharm Ther. 2009; 34:547–54. pmid:19744010
- 63. Williams AE, Chrystyn H. Survey of pharmacists’ attitudes towards interchangeable use of dry powder inhalers. Pharm World Sci. 2007; 29:221–7. pmid:17242855
- 64.
Cureton EE, D’Agostino RB. Factor analysis: An applied approach. Lawrence Erlbaum Associates, 1993.
- 65. Cronbach LJ. Coefficient alpha and the internal structure of tests. Psychometrika 1951; 16:297–334.
- 66. Carroll J. B. The nature of data, or how to choose a correlation coefficient. Psychometrika 1961; 26:347–372.
- 67. Jöreskog KG. New developments in LISREL: analysis of ordinal variables using polychoric correlations and weighted least squares. Qual Quant 1990; 24:387–404.
- 68.
Pett MA, Lackey NR, Sullivan JJ. Making Sense of Factor Analysis: The Use of Factor Analysis for Instrument Development in Health Care Research. Thousand Oaks: Sage Publications; 2003.
- 69.
Hutcheson G, Sofroniou N. The multivariate social scientist: Introductory statistics using generalized linear models. Thousand Oaks, CA: Sage Publications, 1999.
- 70.
Bandalos DL, Boehm-Kaufman MR. Four common misconceptions in exploratory factor analysis. In: Lance CE, Vandenberg R J (eds) Statistical and Methodological Myths and Urban Legends: Doctrine, Verity and Fable in the Organizational and Social Sciences. New York, NY: Taylor & Francis; 2008. p. 61–87.
- 71.
Field A.P. Factor analysis on SPSS. In: Discovering statistics using SPSS. 2nd ed. London, UK: Sage, 2005.
- 72.
Stevens J. Factor analysis. In: Applied multi varied statistics for the social sciences. Hillsdale, NJ: Lawrence Erlbaum Associates, 1986.
- 73.
Tabachnick BG, Fidell LS. Using multivariate statistics (5th ed.). Upper Saddle River, NJ: Pearson Allyn & Bacon. 2007.
- 74.
George D, Mallery P. SPSS for Windows step by step: A simple guide and reference. 11.0 update (4th ed.). Boston: Allyn & Bacon, 2003.
- 75.
Gatignon H. Confirmatory Factor Analysis. In: Gatignon H (ed) Statistical analysis of management data. 2nd ed. New York, NY: Springer; 2010, p.59–122.
- 76. Kelley K, Lai K. Accuracy in parameter estimation for the Root Mean Square Error of Approximation: sample size planning for narrow confidence intervals. Multivariate Behavioral Research 2011; 46:1–32. pmid:26771579
- 77. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling 1999; 6(1):1–55.
- 78.
Jöreskog KG, Sörbom D, Lisrel VI. Analysis of linear structural relationships by maximum likelihood, instrumental variables, and least squares methods. Mooresville, Indiana: Scientific Software, 1984.
- 79. Tucker LR, Lewis C. A reliability coefficient for maximum likelihood factor analysis. Psychometrika 1973; 38:1–10.
- 80.
Munro BH. Statistical methods for healthcare research. Lippincott, 2005.
- 81. Fornell C, Larcker DF. Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research 1981; 18(1): 39–50.
- 82. Aletras VH, Kostarelis A, Tsitouridou M, Niakas D, Nicolaou A. Development and preliminary validation of a questionnaire to measure satisfaction with home care in Greece: an exploratory factor analysis of polychoric correlations. BMC Health Serv Res. 2010 Jul 5; 10:189. pmid:20602759
- 83. Masters GN. A Rasch model for partial credit scoring. Psychometrika 1982; 47: 149–174.
- 84.
Wright BD, Masters GN. Rating Scale Analysis. Chicago: MESA Press, 1982.
- 85. Zheng X, Rabe-Hesketh S. Estimating parameters of dichotomous and ordinal item response models with gllamm. The Stata Journal 2007; 7(3):313–333.
- 86. Kolenikov S. Confirmatory factor analysis using confa. The Stata Journal 2009; 9(3):329–373.
- 87. Alexander M, Berger W, Buchholz P, Walt J, Burk C, Lee J, et al. The reliability, validity, and preliminary responsiveness of the Eye Allergy Patient Impact Questionnaire (EAPIQ). Health Qual Life Outcomes 2005 Oct 31; 3:67. pmid:16259630