Skip to main content

Validity and reliability of the Evidence Utilisation in Policymaking Measurement Tool (EUPMT)

Abstract

Background

As a well-known theory in studying the effective factors on behaviour, the theory of planned behaviour (TPB) is frequently used in evaluating the health behaviour of people and healthcare providers, but rarely applied in studying the behaviour of health policymakers. The aim of the present study is to design and validate a TPB-based measurement tool for evidence utilisation in health policymaking (the EUPMT) through a mixed approach using confirmatory factor analysis.

Methods

The study population consisted of all the specialised units and their employees in the five deputies of Iran’s Ministry of Health and Medical Education in 2013. All those eligible were invited to participate in the study, which comprised 373 persons. The reliability of the EUPMT was determined through test-retest and internal consistency. Additionally, its validity was determined by face, content, convergent, discriminant and construct validities. SPSS-20 and LISREL-8.8 were employed to analyse the data. To assess the fitness of the measurement models, three groups of indices were used, i.e. absolute, relative and parsimonious.

Results

The content and face validities of the tool were 83% and 67%, respectively. Cronbach’s alpha of different constructs ranged from 0.7 to 0.9. In the test-retest method, the intra-class correlations were between 0.75 and 0.87. Confirmatory factor analysis showed that the penta-factorial structure of the experimental data had acceptable fitness with the TPB (GFI = 0.86, NFI = 0.94, RSMEA = 0.075).

Conclusion

TPB is able to explain the behaviour of evidence utilisation in health policymaking. The finalised TPB-based tool has relatively good reliability and validity to assess evidence utilisation in health policymaking. The EUPMT can be applied to determine the status quo of evidence utilisation in health policymaking, whilst designing interventions for its improvement and assessing their outcomes.

Peer Review reports

Key message

  • TPB is able to explain the behaviour of evidence utilisation in health policymaking.

  • The finalised TPB-based tool has relatively good reliability and validity to assess evidence utilisation in health policymaking.

  • The EUMPT can be applied to determine the status quo of evidence utilisation in health policymaking, whilst designing interventions for its improvement and assessing their outcomes.

Background

Health systems often fail to meet the optimal utilisation of research evidence, which itself may lead to inefficacy, reduced quality of life, the life expectancy of citizens and, as a result, reduced productivity [1]. At an international level, there is an increasing interest in health system policymakers’ and managers’ awareness of relevant and valid research results [2,3,4,5,6]. Among the existing challenges is the difficulty in measuring policymakers’ research-use actions before and after the interventions aimed at promoting the practical utilisation of evidence [7, 8]. In such cases, we are left with no choice but to employ subjective methods, i.e. self-reporting of intention and/or behaviour, such as those derived from social cognition theories [9, 10].

To study behaviour-related factors, several theories have been presented, including Implementation Intention, Common Sense Self-regulation Model, Operant Learning Theory, Stage Model, and Theory of Planned Behaviour (TPB). TPB has been the dominant theoretical approach to study health-related behaviours over the past three decades [11]. This theory, originally presented by Ajzen [12], includes three constructs consisting of Attitude Towards to the Behaviour (ATB), Subjective Norm (SN) and Perceived Behavioural Control (PBC), which shape individuals’ intention, and intention is considered predictive of future behaviour. Figure 1 shows the conceptual model of TPB. TPB has been widely used to study health-related behaviours and is effective in predicting the individual health-related behaviours [13, 14].

Fig. 1
figure1

Theory of planned behaviour

TPB has been extensively applied in psychology and healthcare disciplines, and its efficacy in predicting individual health-related behaviour has been shown in a couple of systematic reviews [10]. Based on the results of systematic reviews conducted in psychology, this theory has been able to explain approximately 39% of changes in intention and 27% of changes in behaviour [15]. Moreover, existing evidence supports the use of TPB in predicting research evidence utilisation (clinical guidelines) by health specialists [10]. According to a systematic review [15], prediction of behaviour on the basis of health specialists’ intentions is similar to the values reported in other behaviour domains. Therefore, it may be helpful in studying the behaviour of health policymakers.

The application of TPB on any behaviour requires the identification of affecting factors on the intended behaviour and the development of a specific tool to measure their effects on the behaviour. Like any other measurement tool, an evaluation of the psychometric properties of this tool, such as validity and reliability, is required [10]. There are two approaches for designing a tool for the study of a behaviour based on TPB, namely the direct and indirect approaches. In the direct approach, the respondents’ overall judgment on the theory’s constructs is questioned. On the other hand, in the indirect approach, the respondents’ beliefs on the theory’s constructs are first questioned, and then the importance of the beliefs or the effects of them on the intention/behaviour are evaluated. The advantage of the indirect approach over the direct one is the possibility of accurately identifying the factors affecting intention and behaviour, and designing interventions to improve them [14]. The only developed TPB-based tool to measure the use of evidence in policymaking is that by Boyko et al. [10]. Nevertheless, due to the small sample size of the study, only its reliability was examined. Further, the approach taken by Boyko et al. [10] to design the questions was direct, which lacks the necessary efficacy to design knowledge translation interventions. Additionally, a protocol was published to study the factors affecting evidence utilisation for health policymaking based on TPB [7]. In this protocol, Lavis et al. [7] investigated the effects of access to a database of systematic reviews related to health policymaking on the intention of policy analysts and advisers in the Ministry of Health of Ontario, Canada, using the improved tool presented by Boyko et al. [10].

Makkar et al. [16], in a systematic literature review, found six developed tools to assess the utilisation of research evidence by policymakers. They highlighted the weaknesses of these tools, including not following a clear conceptual framework, assessing utilisation of research evidence of a particular policy or in general, not considering the critical appraisal of research evidence, assessing the use of research evidence in a long period of time and hence increasing the possibility of recall bias, not using the triangulation methods in data collection and, finally, not considering the imposed use of the research evidence. By developing the Staff Assessment of engagement with Evidence (SAGE) tool [16], they tried to overcome the limitations of previous tools and to objectively measure the utilisation of evidence in health policymaking.

Brennan et al. [17] developed and validated the Seeking, Engaging with and Evaluating Research (SEER) tool to measure the capacity and use of evidence by policymakers. This tool has three scales, namely those of individual capacity, research engagement actions, and the use of research. The subscales of individual capacity are similar to constructs of the Evidence Utilisation in Policymaking Measurement Tool (EUPMT), in as much as the individual values subscale is similar to the construct of ATB, the subscale of organisational values is similar to the SN construct, and the subscales of self-efficacy and organisational tools and systems are similar to the sub-constructs of PBC (self-efficacy and controllability, respectively). Another scale of the SEER tool involves research engagement actions that are similar to the behaviour construct of the EUMPT, but their questions are graded on two choices (Yes/No) rather than on a five-point Likert scale.

The reliability of the SEER tool was examined by test-retest and internal consistency and its validity was tested by confirmatory factor analysis and criterion validity. The bivariate correlations of individual capacity subscales and corresponding constructs of the tool proposed by Brennan et al. [17] present its criterion validity (0.419–0.671).

The present study was conducted to design a tool to assess evidence utilisation in health policymaking through the indirect approach to design predicting constructs of TPB and to assess its validity, reliability and factorial structure. Our aim was to design a tool that may be employed to determine the status quo (current beliefs and intentions about using research to inform policy), as well as to design interventions and assess their outcomes. The ultimate aim was to help promote the utilisation of evidence in health policies and its transformation into an organisational culture in health policymaking organisations.

Methods

A cross-sectional study of factor analysis type was conducted to design and determine the validity, reliability and factorial structure of the tool evaluating evidence utilisation in health policymaking.

Study population

The study population consisted of specialised units and their employees in the Central Department of Iran’s Ministry of Health and Medical Education (MOHME), including the office director, officer-in-charge, Head of the National Health Plan and expert officers (Table 1). The Central Department of MOHME comprises five deputies for Health Affairs, Curative Affairs, Nursing, Traditional Medicine, and Food and Drugs. Each deputy consists of a number of general offices, and each general office consists of a number of specialised units. Usually, the specialised units are where the policy briefs are developed.

Table 1 Characteristics of the participants in the construct validity test

Sample size and selection

The inclusion criteria of the study were a minimum of 2 years’ professional experience in the Central Department of MOHME and participation in the development of policy briefs. Although some guidelines about sample size for confirmatory factor analysis (CFA) have been published so far, there is no consensus regarding this [18]. Many researchers recommend a minimum of 200 observations [19,20,21]. According to a rule of thumb, between 5 and 10 observations are required for every free parameter [22, 23]. Bearing in mind the 32 variables, a minimum of 160 samples was necessary. However, the number of eligible individuals was 373, which was about 12 times the number of the variables. Since the number of eligible individuals fell into the range of observations recommended for CFA, all of them were invited to participate in the study.

Designing the questions

There are two approaches to designing the questions of predicting constructs of TPB (ATB, SN and PBC), namely the direct and indirect approaches [14]. In the direct approach, the overall judgment of the respondents is questioned about each theory construct, for example, to measure the ATB construct, some questions arise such as ‘the use of evidence in developing the health policy briefs is … (answer: five-point scale from helpful to unhelpful).

In the indirect approach, a qualitative study is first designed and implemented based on TPB constructs. To measure the construct of the ATB, some open questions arise, e.g. ‘in your opinion, what are the advantages of using evidence in developing the policy briefs?’ If one of the stated advantages is, for example, ‘to avoid wasting the resources’, to measure this attitude, two questions are designed as follows: 1. Developing the evidence-based policy briefs prevents wasting the resources (answer: five-point scale from likely to unlikely); 2. Preventing wasting the resources is … for me (answer: five-point scale from quite desirable to quite undesirable). As it might be noted, the first question measures the likelihood of studied behavioural outcomes while the second measures its utility in respondent opinion. The indirect approach is superior to the direct approach such that, in the indirect approach, specific factors related to the behaviour are identified and measured; therefore, using a tool that has been designed based on this approach can be more helpful in accurately designing the interventions for improving the considered behaviour.

The required qualitative study has already been carried out and published to identify the effective variables on the use of evidence in health policymaking based on TPB [24]. In this study, 32 variables were identified based on the indirect approach, where for each identified variable, two questions, and therefore a total of 64 questions, were designed. To design the questions related to intention and behaviour constructs, three and four questions were designed through a direct approach, respectively. Thus, the total number of questions in the initial tool was 71 (Table 2).

Table 2 Initial version of Evidence Utilisation in Policymaking Measurement Tool

Ethical considerations

Before conducting the study, the necessary permissions were acquired from the Ethics Board of Iran University of Medical Sciences (IUMS: 93.105.352). To ensure the confidentiality of data, the completed questionnaires were received in sealed envelopes and opened on a certain day. Completion of the questionnaire was followed up three times. If lost, another questionnaire would be handed over to the participant. The researcher tried to gain the informed participation consent of the eligible individuals by introducing himself and explaining the aim and process of the study. In the case of expressing informed consent by participants, the questionnaires were given to them and a specific date was assigned for its delivery.

Tool validity

The validity of the tool was tested through face, content and construct validities, and subsequently improved. Content validity refers to the capability of selected questions in reflecting features of measured constructs. Face validity assessed to what extent the questions of a scale are similar to the issue that they are designed to measure. Both validities can be calculated for any of the questions of a scale and to the whole scale [25,26,27]. Quantitative and qualitative methods were used to evaluate and improve the content and face validities. The designed tool was handed out to three groups of individuals, which were 18 members overall, including five content experts (people familiar with TPB), eight lay experts (evidence-based policy brief developers in the Central MOHME), and five methodologists (people who are experienced in questionnaire design and have published related articles). These individuals assessed the content and face validities of the questions and rated them in a varying degree of ‘completely desirable’ to ‘completely undesirable’. Then, the respondents were asked for their opinions on how the face validity of questions could be improved. The quantitative results of the test were estimated in the form of content validity ratio (CVR) and content validity index (CVI) indicators. CVR is calculated by dividing the number of individuals who believe the relevancy of the question is desirable and strongly desirable (in a four-point scale from ‘not desirable’ to ‘strongly desirable’) by the total number of respondents. The CVI is calculated by dividing the mean validity of each question by the total number of questions in the tool [25]. The questionnaire was revised by three panels of experts, including two methodologists, two topic experts, one Persian Language specialist, a MOHME expert, and our research team using the questions’ CVR, and suggestions were made regarding the omission or improvement of the questions’ validities.

Construct validity shows the extent to which observed scores on the tool are predicted by the theory upon which the tool was based. Accumulation of evidence about the face, convergent and discriminant validities of a tool is used to provide evidence of construct validity [28]. It is desirable that constructs are sufficiently correlated, but not so much that they cannot be discriminated from each other. Evidence of convergent validity can be demonstrated by showing a positive and significant correlation between scores for constructs that are expected to be related. The threshold considered as demonstrating convergent validity was 0.5. Discriminant validity is the extent to which a construct distinguishes itself from the other constructs. Exploratory Factor Analysis (EFA) and SPSS-20 software were used to test discriminant validity and CFA, while LISREL-8.8 was used to test convergent and construct validities. The Kaiser-Meyer-Olkin measure of sampling adequacy and Bartlett’s sphericity test were used to ensure the adequacy of the sample size and the existence of a correlation between the studied variables in the CFA. Before analysing the data with LISREL, the assumptions of the CFA were assessed for the existence of missing data, outliers, normality and linearity of the associations between the constructs [29, 30].

Tool reliability

To evaluate the reliability of the direct questions, internal consistency was tested by calculating Cronbach’s alpha coefficient. Test-retest reliability of the indirect questions was tested by calculating the intraclass correlation coefficient (ICC) [14]. The time gap for the test-retest was approximately 2 weeks. The reliability test for the indirect questions was conducted in MOHME’s Deputy of Education; although it was not among the study’s target groups, it was very similar to the study population in terms of organisational settings. Next, 36 and 35 questionnaires were distributed in the test and retest phases, respectively. The response rates in the two phases were 94% and 91%, respectively.

To assess the goodness-of-fit of the measurement models, three groups of indices (absolute, relative and parsimonious) were used. Among the absolute fit indices, the χ2 to ‘degree of freedom’ ratio (acceptable < 3), the goodness-of-fit index (GFI) and the adjusted GFI (AGFI) (acceptable < 0.9) were used. The Bentler-Bonett normed fit index and the non-normed fit index were used from the relative fit indices, and the root mean squared error of approximation was employed as a parsimonious fit index. To ensure the accuracy of data entry, 10% of the data were imported to Excel software by another individual and the correspondence of data was evaluated with Epi-info software [31]. The discrepancy rate was less than 2%, which is acceptable. To examine the possibility of non-response bias, the characteristics of the individuals who refrained from completing the questionnaire, along with their reasons, were documented.

Results

Reliability

Upon testing the reliability of the indirect questions, the ICC of seven questions was lower than 0.6 and thus they were omitted along with their corresponding questions (questions: 12, 24, 30, 35, 41, 55 and 70). After removing these variables, the ICC of the behaviour, ATB, SN and PBC constructs, and that of the entire tool, were estimated to be 0.75, 0.83, 0.87 and 0.89, respectively. Cronbach’s alpha for the behaviour construct was smaller than 0.7, which improved after removing question 4. Cronbach’s alpha coefficients for the behaviour and intention constructs were 0.75 and 0.88, respectively.

Validity

The response rate of the construct validity test was 92.76%. Despite following the participants up to three times and extending the dates of completion for the questionnaires, 27 individuals did not return their questionnaire. Non-response bias was not tested, as the response rate was high. Table 1 shows the characteristics of the participants in the construct validity test. Content validity of two questions (10 and 11) was not approved. The content and face validity indices of the tool and its constructs are presented in Table 3.

Table 3 The content and face validity indices of EUPMT

The Kaiser-Meyer-Olkin value was 0.89, indicating the adequacy of sample size for the CFA. The χ2 value calculated in Bartlett’s sphericity test was significant at P < 0.001, at a degree of freedom of 253, which is justified for the CFA test.

The TPB-based measurement model of EUPMT for the standard estimate state is illustrated in Fig. 2.

Fig. 2
figure2

The measurement model of EUPMT for the standard estimate state

The results of the EFA revealed that most of the observed variables had a correlation higher than 0.5 with only one of the constructs, and their correlation with the other factors was lower than 0.5; hence, these results provided evidence of their discriminant validity. However, correlations predicted by the theoretical model for seven variables were not observed. The rotated factorial matrix output Varimax revealed six constructs instead of three. Upon comparing the factorial weight and the variables’ constructional associations, evidence of the discriminant validity of 17 variables was provided. One variable had no acceptable factorial weight in either of the constructs (questions 55 and 56). However, the remaining six variables may have represented the construct(s) that the theory was unable to explain. Thus, the status of these six variables was examined to see whether they could represent a new construct(s) or not. The remaining six variables were equally distributed among the three constructs supplementary to the theory. The variables ‘workload’ (Questions 60 and 61) and ‘crowded and noisy workplace’ (Questions 62 and 63) were placed in one construct. ‘The colleagues’ norm’ (Questions 32 and 33) and ‘selection of MOHME experts on the basis of their competencies’ were put in another construct. ‘Time-consuming nature of the process’ (Questions 8 and 9) and ‘selection of the managers of MOHME among the faculty members’ (Questions 47 and 48) were placed in the third construct.

The two variables ‘workload’ and ‘crowded and noisy workplace’ were evidently among the participants’ control beliefs and logically should have acquired suitable factorial weight in the PBC construct. Therefore, these variables were not able to build a new differentiated construct from the TPB constructs. ‘The colleagues’ norm’ and ‘selection of MOHME’ experts on the basis of their competencies’ did not have a mutual semantic association that could build a new distinctive construct from TPB constructs. Moreover, the former variable is evidently part of the SN construct, and the latter variable is part of the PBC construct. Therefore, logically, they should have gained good factorial weight in their corresponding constructs. The variables ‘time-consuming nature of the process’ and ‘selection of the managers of MOHME among the faculty members’ did not have a mutual semantic association such that they could form a new independent construct from TPB constructs. Furthermore, the participants considered ‘time-consuming nature of the process’ as a disadvantage of the evidence utilisation to develop a policy brief, which was a negative attitude toward the behaviour. Logically, it is expected that a good factorial weight will be found in its relevant construct, i.e. ATB. Most participants recognised ‘selection of the managers of MOHME among the faculty members’ as an environmental/organisational facilitator for developing evidence-based policy briefs, which is a positive control belief and should have acquired acceptable factorial weight in its relevant construct, i.e. PBC. Therefore, the surplus constructs of TPB derived from the EFA could not build new constructs to a better explanation for developing evidence-based policy briefs and thus they were removed along with their corresponding questions (questions 8, 32, 43, 47, 53, 60 and 62).

The factorial weight of the observed variables and the mean variance extracted for the other constructs was greater than 0.5; hence, these results provided evidence of their convergent validity. The mean extracted variance for each of the constructs was greater than the squared correlation coefficient (R2) of their variables with the other constructs; thus, these results provided evidence of their discriminant validity. Overall, evidence about the convergent and discriminant validity of 23 variables (40 questions) was provided.

Although the GFI and AGFI values slightly differ from the acceptable values (Table 4), the fitness indices give an acceptable fitness for EUPMT [29]. After refinement, the final version of the tool had 40 questions.

Table 4 Fitness indices for the assessment of the measurement model EUPMT

Discussion

Fitness of the measurement model

The validity and reliability of EUPMT were examined using factor analysis, test-retest and internal consistency. The CVI of the tool was 83% and acceptable. The face validity of the tool was 67% and slightly smaller than the acceptable value; thus, it was improved in panels of experts. Upon testing the reliability of the directly observed variables via internal consistency, Cronbach’s alpha was between 0.7 and 0.9, and acceptable. Additionally, via testing of the reliability of the indirectly observed variables through test-retest, the ICC of different constructs was between 0.75 and 0.87, and the ICC of the tool was 0.89 overall and acceptable. The CFA results showed the acceptable goodness-of-fit indices for the measurement model of EUPMT.

To assess the validity and reliability of the SEER CFA, internal consistency and test-retest were used. The strength of the SEER was the approval of its criterion validity through simultaneous measuring with a TPB-based tool [17]; however, the criterion validity of EUPMT was not assessed. Another advantage of SEER is its consideration of the type of research usage (conceptual, instrumental, tactical and imposed). The results of CFA presented an acceptable fit of EUPMT and SEER tools. The reliability results of the test-retest of the constructs of EUMPT were superior to the corresponding subscales in SEER, such that ICCs of the constructs of ATB, SN and PBC sub-constructs (self-efficacy and controllability) were 16%, 7%, 2% and 17% higher than the corresponding subscales in SEER, respectively. The response rate in the examination of the validity and reliability of EUPMT was 92.8%, while this rate was 54%–55% in SEER.

Among the factors that lead to high rates of response in EUPMT study were the fame and reputation of one of the researchers (MR) and the fact that a signed invitation with the name of the participants was sent out; the measures to protect the confidentiality of responses (collecting the questionnaires in a sealed envelope and opening all of them on a given day); the interesting subject matter in the participants’ opinion; the timing delivery of completed questionnaires by participants within a 2-week period; three follow up reminders to receive the completed questionnaire and redelivering of the questionnaire in case of loss; collection of the data by the researcher instead of hiring interviewers; and suggesting the importance of the study considering the researcher’s degree (PhD), which was effective in motivating the participants. The researcher could draw the informed participation of the participants followed by introducing himself and explaining the objectives and steps of the study. Explaining the steps of the study was effective in increasing motivation. Finally, the fact that the questionnaires were not provided through translation or reviewing of previous texts, but rather were prepared through in-depth interviews of the participants’ co-workers, was also viewed positively.

EUPMT has been designed based on one of the most famous and prestigious behaviour theories (TPB). Moreover, SEER used the framework of SPIRIT (Supporting Policy In health with Research), which was developed based on a literature review and interviews with policymakers [32]. While the effectiveness of interventions based on the TPB has been demonstrated in different studies, the present study suggests the design and implementation of positive interventions for the use of evidence in health policymaking based on SEER and EUPMT and compares their effectiveness.

The reliability of the tool proposed by Boyko et al. [10] (assessing the intent of health policymakers to use evidence based on TPB) was also investigated through internal consistency (Cronbach’s alpha coefficients and generalisability) and test-retest reliability (Pearson correlation coefficients and generalisability); however, due to the small sample size (62 people), its construct validity was not tested using CFA. The alpha coefficient range of the constructs in Boyko et al.’s [10] tool was within 0.68 to 0.89, which was slightly lower than that of EUPMT (0.72 to 0.91). The advantage of EUPMT compared to the Boyko et al. [10] tool was to assess the construct validity through CFA. Another advantage of EUPMT was to use the indirect approach to design questions and hence the possibility of designing more specific and precise interventions to improve the use of evidence in health policymaking.

Among the limitations of the study is the use of self-reported data. To prevent social desirability bias, the questionnaires were received anonymously and in sealed envelopes. However, the bias may still be affecting the data. Another limitation of the study is the lack of testing of the criterion validity, which could provide further evidence of the validity of the EUPMT. Although the use of objective tools such as SAGE provides a more detailed profile of using the evidence in health policymaking, its use is difficult in practice. Other limitations of SAGE may be neglecting the mechanisms, structures and systems that can help improve the capacity of policymakers to use evidence, as well as focusing on barriers in using the evidence and ignoring the facilitators [16].

Conclusion

The results showed that real data on utilisation of evidence in health policymaking support the TPB; in other words, the TPB is also capable of explaining the use of evidence in health policymaking. EUPMT has relatively good reliability and validity to assess evidence utilisation in health policymaking. This tool may be employed to determine the status quo of evidence utilisation in health policymaking, to design interventions for its improvement and to assess the outcomes of conducted interventions. The EUPMT can effectively help health policymakers promote the utilisation of evidence in the development of policy briefs and transform it into an organisational culture.

Abbreviations

AGFI:

Adjusted goodness-of-fit index

ATB:

Attitude towards to the behaviour

CFA:

Confirmatory factor analysis

CVI:

Content validity index

CVR:

Content validity ratio

EFA:

Exploratory factor analysis

EUPMT:

Evidence utilisation in policymaking measurement tool

GFI:

Goodness-of-fit index

ICC:

Intra-class correlation

MOHME:

Ministry of Health and Medical Education

PBC:

Perceived behavioural control

SAGE:

Staff Assessment of engagement with Evidence

SEER:

Seeking, Engaging with and Evaluating Research

SN:

Subjective norm

SPIRIT:

Supporting Policy In health with Research

TPB:

Theory of planned behaviour

References

  1. 1.

    Ellen ME, Léon G, Bouchard G, Lavis JN, Ouimet M, Grimshaw JM. What supports do health system organizations have in place to facilitate evidence-informed decision-making? A qualitative study Implement Sci. 2013;8:84.

    Article  PubMed  Google Scholar 

  2. 2.

    Oxman AD, Lavis JN, Lewin S, Fretheim A. SUPPORT Tools for evidence-informed health Policymaking (STP) 1: What is evidence-informed policymaking? Health Res Policy Syst. 2009;7 Suppl 1:S1.

    Article  PubMed  PubMed Central  Google Scholar 

  3. 3.

    Sheldon TA. Making evidence synthesis more useful for management and policy-making. J Health Serv Res Policy. 2005;10:S1.

    Article  Google Scholar 

  4. 4.

    Almeida C, Báscolo E. Use of research results in policy decision-making, formulation, and implementation: a review of the literature. Cad Saude Publica. 2006;22:S7–S19.

    Article  PubMed  Google Scholar 

  5. 5.

    Imani-Nasab M, Seyedin H, Yazdizadeh B, Majdzadeh R. A qualitative assessment of the evidence utilization for health policy-making on the basis of SUPPORT tools in a developing country. Int J Health Policy Manag. 2017;6(8):457–65.

    Article  Google Scholar 

  6. 6.

    Dehnavieh R, Hekmat SN, Ghasemi S, Mirshekari N. The vulnerable aspects of application of “Health Technology Assessment”. Int J Technol Assess Health Care. 2015;31(03):197–8.

    Article  PubMed  Google Scholar 

  7. 7.

    Lavis JN, Wilson MG, Grimshaw JM, Haynes RB, Hanna S, Raina P, Gruen R, Ouimet M. Effects of an evidence service on health-system policy makers’ use of research evidence: A protocol for a randomised controlled trial. Implement Sci. 2011;6:51.

    Article  PubMed  PubMed Central  Google Scholar 

  8. 8.

    Bastani P, Kavosi Z, Alipoori S, Imani-Nasab MH. Evidence-based policy and decision-making among health managers: a case of Shiraz University of Medical Sciences. Galen Med J. 2017;6(1):30–8.

    Google Scholar 

  9. 9.

    Mohr DC, Lukas CV, Meterko M. Predicting healthcare employees’ participation in an office redesign program: Attitudes, norms and behavioral control. Implement Sci. 2008;3:47.

    Article  PubMed  PubMed Central  Google Scholar 

  10. 10.

    Boyko JA, Lavis JN, Dobbins M, Souza NM. Reliability of a tool for measuring theory of planned behaviour constructs for use in evaluating research use in policymaking. Health Res Policy Syst. 2011;9:29.

    Article  PubMed  PubMed Central  Google Scholar 

  11. 11.

    Bonetti D, Johnston M, Clarkson JE, Grimshaw J, Pitts NB, Eccles M, Steen N, Thomas R, Maclennan G, Glidewell L. Applying psychological theories to evidence-based clinical practice: identifying factors predictive of placing preventive fissure sealants. Implement Sci. 2010;5:25.

    Article  PubMed  PubMed Central  Google Scholar 

  12. 12.

    Ajzen I, Manstead AS. Changing health-related behaviors: An approach based on the theory of planned behavior. In: Hewstone M, Schut H, de Wit J, Van Den Bos K, Stroebe M, editors. The Scope of Social Psychology: Theory and Applications. Hove: Psychology Press; 2007. p. 43–63.

    Google Scholar 

  13. 13.

    Armitage CJ, Conner M. Efficacy of the theory of planned behaviour: A meta‐analytic review. Br J Soc Psychol. 2001;40(4):471–99.

    CAS  Article  PubMed  Google Scholar 

  14. 14.

    Francis J, Eccles MP, Johnston M, Walker A, Grimshaw J, Foy R, Kaner EF, Smith L, Bonetti D. Constructing questionnaires based on the theory of planned behaviour. A Manual for Health Services Researchers. Newcastle upon Tyne: Centre for Health Services Research, University of Newcastle upon Tyne; 2004.

  15. 15.

    Wilson MG, Lavis JN, Grimshaw JM, Haynes RB, Bekele T, Rourke SB. Effects of an evidence service on community-based AIDS service organizations’ use of research evidence: a protocol for a randomized controlled trial. Implement Sci. 2011;6:52.

    Article  PubMed  PubMed Central  Google Scholar 

  16. 16.

    R. Makkar S, Brennan S, Turner T, Williamson A, Redman S, Green S. The development of SAGE: A tool to evaluate how policymakers’ engage with and use research in health policymaking. Res Eval. 2016;25(3):315–28.

  17. 17.

    Brennan SE, McKenzie JE, Turner T, Redman S, Makkar S, Williamson A, Haynes A, Green SE. Development and validation of SEER (Seeking, Engaging with and Evaluating Research): a measure of policymakers’ capacity to engage with and use research. Health Res Policy Syst. 2017;15:1.

    Article  PubMed  PubMed Central  Google Scholar 

  18. 18.

    MacCallum RC, Widaman KF, Zhang S, Hong S. Sample size in factor analysis. Psychol Methods. 1999;4(1):84.

    Article  Google Scholar 

  19. 19.

    Gefen D, Straub D, Boudreau M-C. Structural equation modeling and regression: Guidelines for research practice. CAIS. 2000;4(1):7.

    Google Scholar 

  20. 20.

    Hoe SL. Issues and procedures in adopting structural equation modeling technique. J Applied Quant Methods. 2008;3(1):76–83.

    Google Scholar 

  21. 21.

    Boomsma A, Hoogland JJ. The robustness of LISREL modeling revisited. In: Cudeck R, du Toit S, Sörbom D (eds). Structural Equation Models: Present and Future. A Festschrift in Honor of Karl Jöreskog. Lincolnwood, IL: Scientific Software International; 2001. pp. 139–168.

  22. 22.

    Iacobucci D. Structural equations modeling: Fit indices, sample size, and advanced topics. J Consumer Psychol. 2010;20(1):90–8.

    Article  Google Scholar 

  23. 23.

    Quintana SM, Maxwell SE. Implications of recent developments in structural equation modeling for counseling psychology. Counseling Psychol. 1999;27(4):485–527.

    Article  Google Scholar 

  24. 24.

    Imani-Nasab MH, Seyedin H, Majdzadeh R, Yazdizadeh B, Salehi M. Development of evidence-based health policy documents in developing countries: A case of Iran. Global J Health Sci. 2014;6(3):27.

    Article  Google Scholar 

  25. 25.

    DeVon HA, Block ME, Moyle‐Wright P, Ernst DM, Hayden SJ, Lazzara DJ, Savoy SM, Kostas‐Polston E. A psychometric toolbox for testing validity and reliability. J Nurs Scholarsh. 2007;39(2):155–64.

    Article  PubMed  Google Scholar 

  26. 26.

    Lawshe CH. A quantitative approach to content validity. Personnel Psychology. 1975;28(4):563–75.

    Article  Google Scholar 

  27. 27.

    Polit DF, Beck CT. The content validity index: are you sure you know what’s being reported? Critique and recommendations Res Nurs Health. 2006;29(5):489–97.

    Article  PubMed  Google Scholar 

  28. 28.

    Bagozzi RP, Yi Y, Phillips LW. Assessing construct validity in organizational research. Adm Sci Q. 1991;36(3):421–58.

    Article  Google Scholar 

  29. 29.

    Stapleton LM. Structural equation modeling: foundations and extensions. Struct Equ Model Multidiscip J. 2011;18(3):492–6.

    Article  Google Scholar 

  30. 30.

    Johnson RA, Wichern DW. Applied Multivariate Statistical Analysis, vol. 4. Englewood Cliffs: Prentice Hall; 1992.

    Google Scholar 

  31. 31.

    Centers for Disease Control Prevention. Epi Info 7 User Guide. https://www.cdc.gov/epiinfo/support/userguide.html. Accessed 2 Aug 2017.

  32. 32.

    Redman S, Turner T, Davies H, Williamson A, Haynes A, Brennan S, Milat A, O’Connor D, Blyth F, Jorm L. The SPIRIT Action Framework: A structured approach to selecting and testing strategies to increase the use of research in policy. Soc Sci Med. 2015;136:147–55.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

The authors would like to thank the directors of technical offices and their senior technicians in Iran’s Ministry of Health and medical education who volunteered their time to participate in this study and Dr. Katayoun Maleki for English editing of the paper and anonymous reviewers for valuable comments.

Funding

This study was part of a PhD thesis supported by Iran University of Medical Sciences (grant No: IUMS.SHMIS-2013.342).

Author information

Affiliations

Authors

Contributions

RM and HS participated in study design and interpretation the findings. BY participated in study design and analysis of data. MS participated in analysis of data. MHIN gathered the data and participated in other stages of the study. All authors read and approved the final manuscript.

Corresponding author

Correspondence to B. Yazdizadeh.

Ethics declarations

Ethics approval and consent to participate

The necessary permission was obtained from the Ethics Committee of Iran University of Medical Sciences (IUMS: 93.105.352). The oral consent of every participant was obtained.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Imani-Nasab, M.H., Yazdizadeh, B., Salehi, M. et al. Validity and reliability of the Evidence Utilisation in Policymaking Measurement Tool (EUPMT). Health Res Policy Sys 15, 66 (2017). https://doi.org/10.1186/s12961-017-0232-6

Download citation

Keywords

  • Knowledge translation
  • Health policymaking
  • Confirmatory factor analysis
  • Theory of planned behaviour