The Evidence-Based Practice Beliefs and Implementation Scales: Psychometric Properties of Two New Instruments

Evidence Digest

The Evidence-Based Practice Beliefsand Implementation Scales: PsychometricProperties of Two New Instruments

Bernadette Mazurek Melnyk, RN, PhD, CPNP/NPP, FAAN, FNAP, Ellen Fineout-Overholt, RN, PhD, FNAP, Mary Z. Mays, PhD

ABSTRACTBackground: Implementation of evidence-based practice (EBP) by health professionals is a key strategy

for improving health care quality and patient outcomes as well as increasing professional role satisfac-tion. However, it is estimated that only a small percentage of nurses and other health care providers areconsistently using this approach to clinical practice.

Aim: The aim of this study was to report on the development and psychometric properties of twonew scales: (1) the 16-item EBP Beliefs Scale that allows measurement of a person’s beliefs about thevalue of EBP and the ability to implement it, and (2) the 18-item EBP Implementation Scale that allowsmeasurement of the extent to which EBP is implemented.

Methods: Nurses (N = 394) attending continuing education workshops volunteered to complete thescales. Data were analysed to evaluate reliability and validity of both instruments.

Results: Cronbach’s alpha was > .90 for each scale. Principal components analysis indicated that eachscale allowed measurement of a unidimensional construct. Strength of EBP beliefs and the extent of EBPimplementation increased as educational level increased (p < .001) and as responsibility in the workplaceincreased (p < .001).

Conclusion: In this study, initial evidence was provided to support the reliability and validity of theEBP Beliefs and Implementation Scales in a heterogeneous sample of practicing nurses.

Evidence to Action: Use of the scales in future research could generate evidence to guide EBP imple-mentation strategies in practice and education. Results could establish the extent to which EBP is beingimplemented and its effect on clinician satisfaction and patient outcomes.

KEYWORDS evidence-based practice, measurement, reliability, validity, nursing, beliefs

Worldviews on Evidence-Based Nursing 2008; 5(4):208–216. Copyright ©2008 Sigma Theta Tau International

Evidence-based practice (EBP) is a problem-solvingapproach to the delivery of care that incorporates

the best evidence from well-designed studies in combina-tion with a clinician’s expertise and patients’ preferenceswithin a context of caring (Sackett et al. 2000; Melnyk &Fineout-Overholt 2005b). Because there has been supportfor better outcomes as a result of evidence-based versus

Bernadette Mazurek Melnyk, Dean and Distinguished Foundation Professor in Nurs-ing, Ellen Fineout-Overholt, Clinical Associate Professor and Director, Center for theAdvancement of Evidence-Based Practice, Mary Z. Mays, Associate Professor, ArizonaState University College of Nursing & Healthcare Innovation, Phoenix, Arizona.

Address correspondence to Bernadette Mazurek Melnyk, Arizona State UniversityCollege of Nursing & Healthcare Innovation, 500 North 3rd Street, Phoenix, AZ85004; [email protected]

Accepted 4 October 2007Copyright ©2008 Sigma Theta Tau International1545-102X/08

tradition-based care (Heater et al. 1988; Grimshaw et al.2006), national and federal organisations as well as re-cent summit meetings held by the Institute of Medicineand leaders of health care professions have recommendedits teaching and implementation (Greiner & Knebel 2003;Melnyk et al. 2005). However, the uptake by health careprofessionals of research findings into practice and imple-mentation of EBP has been slow and inconsistent because ofmultiple intra-personal and environmental barriers (Eccleset al. 2005; Fineout-Overholt et al. 2005; Parahoo 2000).

Findings from a recent survey to assess nurses’ readinessto engage in EBP conducted by the Nursing InformaticsExpert Panel of the American Academy of Nursing witha United States’ (U.S.) sample of 1,097 randomly selectedregistered nurses indicated that: (1) almost half were notfamiliar with the term “EBP,” (2) more than half reportedthat they did not believe their colleagues use research

208 Fourth Quarter 2008 �Worldviews on Evidence-Based Nursing

Evidence-Based Practice Beliefs and Implementation Scales

findings in practice, (3) only 27% of the respondents hadbeen taught how to use electronic databases, (4) mostdo not search information databases (e.g., Medline andCINAHL) to gather practice information, and (5) thosewho search these resources do not believe they have ade-quate searching skills (Pravikoff et al. 2005). As a result ofthese findings, along with other major barriers to EBP (e.g.,lack of EBP mentors, negative attitudes toward research,inadequate resources at the point of care, competing prior-ities, perceived lack of time to implement EBP) (Kajermoet al. 2000; Fineout-Overholt et al. 2005; Hutchinson &Johnson 2006), it is estimated that it takes approximately17 years to translate research findings into clinical practiceto improve patients’ outcomes (Balas & Boren 2000).

Furthermore, there is a crisis in nursing workforce short-ages in several countries throughout the world. The crux ofthis shortage is largely two-fold: (1) too few nurses, and (2)dissatisfaction with work environments that do not supportprofessional nursing practice.

The demands placed upon nurses as a result of the short-age have led to increasing reports of job dissatisfaction andintent to leave the profession. Research has shown thatwhile most nurses are committed to their profession, theyare highly dissatisfied with their work environments (U.S.General Accounting Office 2001). In one study, 23% ofnurses reported intending to leave the profession with an-other 37% uncertain of their future (Larrabee et al. 2003).High turnover rates are very costly to the health care systemand negatively affect patient outcomes (Vahey et al. 2004).The Nurse Reinvestment Act (NRA, PL 107–205[h1]) in-dicates that nurse dissatisfaction contributes to the nursingshortage, and that retention might be increased and patientoutcomes improved by nurse involvement in evidence-based clinical decision making.

Although only a few studies have provided direct evi-dence, clinicians who use research evidence in their prac-tices are more satisfied with their role and their patientshave better clinical outcomes (Ciliska et al. 1996; Retsas2000; Maljanian et al. 2002). In addition, anecdotal re-ports from nurses support that providing evidence-basedcare (EBC) renews the professional spirit of the nurse, akey variable in professional satisfaction. EBP also supportsnurses to become “strong patient advocates, focused onimproving the quality of the care given to patients” (Strout2005, p. 39).

Future investigations of the effect of EBP on clinical careand job satisfaction could be facilitated by instruments tomeasure: (1) the strength of beliefs in EBP, and (2) the ex-tent to which EBP is implemented. Therefore, the purposeof this study was to begin validation of two new instru-ments: (1) a 16-item EBP Beliefs Scale that allows mea-surement of an individual’s beliefs about the value of EBP

and the ability to implement it; and (2) an 18-item EBP Im-plementation Scale that allows measurement of the extentof actual EBP implementation.

The Theoretical Model that Guided Developmentof the Two EBP ScalesThe transtheoretical model of health behaviour change(Prochaska & Velicer 1997), which is now being extendedto the field of organisational change, and the ARCC (Ad-vancing Research and Clinical practice through close Col-laboration) model (Melnyk & Fineout-Overholt 2002)guided development of the two EBP scales. The trans-theoretical model has five stages that show the processthrough which an individual makes a change in behaviour(i.e., pre-contemplation, contemplation, preparation, ac-tion, and maintenance; Prochaska & Velicer 1997). Withinthis model are 10 processes that might produce a changein actual behaviour, including three that show the follow-ing cognitive beliefs: (1) appreciating that the change isimportant to one’s success (i.e., self-re-evaluation), (2) be-lieving that a change can succeed and making a firm com-mitment to the change (i.e., self-liberation), and (3) ap-preciating that the change will have a positive effect onthe work environment (i.e., environmental re-evaluation;Prochaska et al. 2001). From this theory, the literature andthe ARCC model, a 52-item descriptive survey was devel-oped by the first two authors and used in a previous studywith 160 nurses who were attending EBP workshops inthe Northeast region of the U.S. (Melnyk et al. 2004a). The52-item survey tapped nurses’ demographic variables aswell as their knowledge and beliefs about EBP, the extentto which they implemented EBP, and barriers and facili-tators to EBP implementation. Seven items on the 52-itemsurvey specifically assessed nurses’ knowledge, beliefs, andability to implement EBP. Findings from this survey indi-cated that nurses who held stronger beliefs about the valueof EBP and their ability to implement it reported a higherlevel of implementation than did nurses with weaker EBPbeliefs. Therefore, in an attempt to more fully study EBPbeliefs and their relationship to EBP implementation, threeitems from the 52-item descriptive survey that tapped be-liefs about EBP served as the foundation for an expanded16-item EBP Beliefs instrument.

The EBP Beliefs Scale also was adapted from previousbeliefs scales developed by Melnyk (1994) and Melnyket al. (2001, 2006) and was specifically designed to mea-sure a clinicians’ beliefs about the value of EBP and theirbeliefs/confidence in implementing it in practice. The con-structs of self-re-evaluation, self-liberation, and environ-mental re-evaluation from the transtheoretical model ofchange also were incorporated into the scale. The relation-ship between an individual’s beliefs/confidence about his

Worldviews on Evidence-Based Nursing �Fourth Quarter 2008 209


or her ability to implement certain behaviours and actualimplementation of those behaviours has been empiricallysupported in many studies (Melnyk 1995; Melnyk et al.2001; Arora et al. 2005; Reynolds & Magnan 2005; Mel-nyk et al. 2006; Rhodes & Plotnikoff 2006).

The ARCC model was first conceptualised in 1999 asa system-wide implementation model of EBP. ARCC isguided by control theory (Carver & Scheier 1982, 1998),which contends that a discrepancy between a standard orgoal (e.g., system-wide implementation of EBP) and a cur-rent state (e.g., the extent to which an organisation is im-plementing EBP) will motivate behaviours to reach thestandard or goal. However, in health care organisations,many barriers exist that inhibit nurses and other healthcare professionals from implementing EBP. In the ARCCModel, strategies are implemented (e.g., EBP mentors whoassist clinicians with implementation of EBP and providingEBP skills building workshops) to remove barriers so thatnurses and other health care providers can implement EBCin order to achieve system-wide implementation.

The first step to system-wide implementation of EBCin ARCC is an organisational assessment of readiness andthe context for EBP so that the strengths and barriers in asystem can be identified. Inherent to successful advance-ment of EBP in the system is the key role of an EBP mentor,an advanced practice nurse who has in-depth skills in EBPas well as working knowledge of how to remove barri-ers to organisational change. Mentorship with direct-carenurses on clinical units by the EBP mentors throughout thesystem is important in strengthening nurses’ beliefs aboutthe value of EBC and their ability to implement it. Evi-dence from previous research has supported the premisethat when nurses have stronger beliefs about their abilityto implement EBP, implementation of EBC is higher thanwhen they have weaker beliefs (Melnyk et al. 2005). Con-tinued research on these issues will be facilitated by instru-ments that allow systematic measurement of EBP beliefsand implementation.

Congruent with the theoretical model, beliefs in EBPwere defined as “endorsement of the premise that EBP im-proves clinical outcomes and confidence in one’s EBPknowledge/skills.” Implementation of EBP was defined as“engaging in relevant behaviours,” including: (1) seeks andappraises scientific evidence, (2) shares evidence or datawith colleagues or patients, (3) collects and evaluates out-come data, and (4) uses evidence to change practice. Thetwo new instruments were created to measure these con-structs on a continuum from strong disagreement to strongagreement in the case of EBP beliefs and from never to dailyfor EBP implementation.

Validation of an instrument is a long-term process ofaccumulating a persuasive body of evidence (Nunnally &

Bernstein 1994; Goodwin 2002; Worthington & Whitaker2006). This study is the first in a series of planned studiesdesigned to refine and validate two new instruments (i.e.,the EBP Beliefs and EBP Implementation Scales) that areused to measure EBP beliefs and implementation, and pro-vide initial evidence of their reliability, construct validity,and criterion validity.

METHODS

ParticipantsThe sample comprised 394 nurses from five states in theU.S. who attended continuing education workshops onEBP provided by the first two authors during 2005 and2006. Age of the nurses ranged from 21 to 69 years (M =45, SD = 10). The sample was composed predominantlyof women (96%). Participants’ self-identified race/ethnicityas White (89%), African American (5%), Asian/Pacific Is-lander (2%), Native American (2%), Hispanic (2%), orOther (< 1%). They were residents of Arizona, Colorado,New Jersey, Ohio, and Texas in the U.S. The majority of par-ticipants were employed full-time (82%) and the majorityworked the day shift (78%). Educational achievement ofthe participants ranged from diploma nurses with less than5 years of experience since attaining the diploma to doctor-ally prepared nurses with more than 20 years experiencesince attaining the doctorate. The majority of participantshad a bachelor’s degree or higher: 5% doctorate, 33% mas-ter’s, 41% bachelor, 16% associate, and 4% diploma.

ProcedureEach group of participants was asked to complete the pa-per and pencil scales prior to commencement of the EBPworkshops they were attending. Participants received anexplanation about how their data would be used to es-tablish the psychometric properties of the EBP Beliefs andImplementation scales. Completion of the scales by theparticipants served as their consent to participate in thestudy. All scales were completed anonymously and partic-ipants completed the two scales in less than 15 minutes.The institutional review board of Arizona State Universityapproved the study.Item development. The EBP Beliefs Scale (Melnyk &Fineout-Overholt 2002), was an outgrowth of earlier work(Melnyk 1994, 1995, Melnyk et al. 1997, 2004a, 2006).On the scale, clinicians are asked to respond to each of16 items (Table 1) on a 5-point Likert-scale that rangesfrom 1 (strongly disagree) to 5 (strongly agree). Scoring ofthe instrument consists of reverse scoring two negativelyphrased items (Table 1) and then summing responses tothe 16 items for a total score that ranges between 16 and80.



TABLE 1Percentage endorsement (N = 333) and factor loadings on items of the EBP beliefs scale

STRONGLYAGREE

ITEM OR AGREE LOADING

8. I am sure that I can implement EBP in a time efficient way. 48% 0.773. I am sure that I can implement EBP. 66% 0.766. I believe that I can search for the best evidence to answer clinical questions in a time efficient way. 62% 0.7315. I am confident about my ability to implement EBP where I work. 46% 0.727. I believe that I can overcome barriers in implementing EBP. 61% 0.7110. I am sure about how to measure the outcomes of clinical care. 44% 0.6714. I know how to implement EBP sufficiently enough to make practice changes. 35% 0.6712. I am sure that I can access the best resources in order to implement EBP. 57% 0.679. I am sure that implementing EBP will improve the care that I deliver to my patients. 94% 0.654. I believe that critically appraising evidence is an important step in the EBP process. 91% 0.632. I am clear about the steps of EBP. 53% 0.615. I am sure that evidence-based guidelines can improve clinical care. 96% 0.561. I believe that EBP results in the best clinical care for patients. 94% 0.5316. I believe the care that I deliver is evidence-based. 47% 0.5213. I believe EBP is difficult. (reverse scored) 47% 0.4311. I believe that EBP takes too much time. (reverse scored) 56% 0.38

The EBP Implementation Scale was developed from areview of literature on the essential components and stepsof EBP. Participants were asked to respond to each of the 18items on a 5-point frequency scale (Table 2) by indicating

TABLE 2Percentage implementation (N = 319) and factor loadings on items of the EBP implementation scale

5 OR MORETIMES IN

THE LAST 8ITEM WEEKS LOADING

16. Shared the outcome data collected with colleagues. 17% 0.836. Shared evidence from a study/ies in the form of a report or presentation to > 2 colleagues. 25% 0.838. Shared an EBP guideline with a colleague. 16% 0.8310. Shared evidence from a research study with a multidisciplinary team member. 18% 0.8114. Used an EBP guideline or systematic review to change clinical practice where I work. 11% 0.7917. Changed practice based on patient outcome data. 16% 0.7915. Evaluated a care initiative by collecting patient outcome data. 17% 0.787. Evaluated the outcomes of a practice change. 20% 0.7818. Promoted the use of EBP to my colleagues. 23% 0.781. Used evidence to change my clinical practice. 24% 0.779. Shared evidence from a research study with a patient/family member. 17% 0.7611. Read and critically appraised a clinical research study. 31% 0.764. Informally discussed evidence from a research study with a colleague. 37% 0.752. Critically appraised evidence from a research study. 33% 0.743. Generated a PICO question about my clinical practice. 13% 0.735. Collected data on a patient problem. 34% 0.7013. Accessed the National Guidelines Clearinghouse. 12% 0.6812. Accessed the Cochrane database of systematic reviews. 15% 0.60

how often in the past 8 weeks they performed the item. Thescale ranges from 0 meaning “0 times” to 4, meaning “>8times.” Scoring consisted of summing responses to the 18items for a total score that could range from 0 to 72.



During item development, the face and content validityof early drafts of the instruments were assessed in conve-nience samples of practicing staff nurses (N = 15) and EBPsubject-matter experts (N = 8) who reviewed the two ques-tionnaires for content and clarity. Final versions of the in-struments were scored for readability: Flesch-Kincaid read-ing grade level was 8.0 for the EBP Beliefs Scale and 9.6 forthe EBP Implementation Scale.Statistical procedures. Reliability was assessed using: (1)the Cronbach procedure for measuring internal consis-tency, and (2) the equal-length, split-half Spearman-Brownprocedure for measuring intra-scale correlation. Reliabilitystatistics were calculated after reverse coding of items wascompleted.

The goal of the instrument development process wasto develop two brief scales, each measuring a singleglobal factor. Attainment of this goal was evaluated witha principal components analysis (PCA) of each scale. Thevalidity of the two scales was analysed separately be-cause, congruent with their different purposes, they usedresponse-rating scales that were semantically and numer-ically distinct, precluding their combination into a singlePCA.

The PCA method was chosen over other factor-analysismethods such as “principal axis factoring” or “common fac-tor analysis” for three reasons. First, this initial analysis waslargely exploratory, rather than confirmatory, designed toidentify essential items for measuring the construct, ratherthen the latent structure of the instrument (Worthington& Whitaker 2006). Second, PCA allows users to take intoaccount all of the variance in items, not simply the sharedvariance, and to identify factors in order of the percent-age of variance accounted for; that is, the first factor inPCA accounts for the maximum amount of total variability.Third, PCA identifies only orthogonal factors, when mul-tiple factors are identified. Thus, PCA can quickly high-light items not contributing to the construct, which canthen be considered for deletion from the scale. The inter-item correlation matrix was used as the basis for the PCA,because the sample was a broad cross-section of the tar-get audience. List-wise deletion of cases was used in eachprocedure.

Differences in attitudes and behaviours among sub-groups were evaluated to determine whether the scalescould distinguish between the subgroups, an indication ofcriterion validity. An independent samples t test was usedto compare two subgroups formed by degree of prior expo-sure to EBP. One-way analysis of variance (ANOVA) wasused to compare five subgroups defined by educationallevel, four defined by nursing roles, and five defined byage. Results were considered statistically significant whenp < .05.

TABLE 3Reliability coefficients of the EBP beliefs and implementationscales

Spearman-Survey N Cronbach α Brown r

EBP Beliefs Scale 330 .90 0.87EBP Implementation Scale 319 .96 0.95

RESULTS

Response PatternsThe percentage of participants who responded stronglyagree or agree is shown in Table 1 for each item on the EBPBeliefs scale. Items with high levels of endorsement werefocused on beliefs about the positive effect of EBP. Itemswith low levels of endorsement were focused on confidencein implementing EBP.

The percentage of participants who responded five ormore times in the last 8 weeks is shown in Table 2 for eachitem of the EBP Implementation scale. The most commonimplementation item was critical appraisal of scientific ev-idence. The least common implementation item was ac-cessing or using a published EBP guideline.

ReliabilityCronbach α and Spearman-Brown r reliability coefficientsexceeded 0.85 for each of the scales (Table 3).

Construct Validity: Exploratory PCAList-wise deletion resulted in a sample of 333 participantsfor PCA of the EBP Beliefs scale. The scree plot of eigen-values indicated that the major discontinuity occurred be-tween the first and second factors. The first factor had aneigenvalue of 6.44 and accounted for 40% of the variancein the scale. Three other factors had eigenvalues > 1.0(1.8, 1.3, and 1.1, respectively). They accounted for 11%,8%, and 7% of the variance in the scale, respectively, notmeaningfully more than the 6.3% variance that could beaccounted for by any single item (Steger 2006). Thus, asingle-factor solution was the most parsimonious interpre-tation of the results. Factor loadings for each item in thesingle-factor solution are shown in Table 1. The unidimen-sional nature of the scale is supported by the high factorloadings. Every item on the scale has a factor loading >

0.35; it is noteworthy that the only items with factor load-ings < 0.50 are the two reverse-scored items. These highloadings combined with the high Cronbach α (.90) indicatethat a single construct is being measured (Linn 1968; Hak-stian et al. 1982; Nunnally & Bernstein 1994; Steger 2006;



TABLE 4Comparison of prior exposure subgroups

N M (SD) p

EBP beliefs scaleNo exposure to EBP 133 54.68 (12.45) =.67Prior exposure to EBP 197 55.46 (18.60)

EBP implementation scaleNo exposure to EBP 133 8.60 (10.74) <.001Prior exposure to EBP 183 18.27 (16.60)

Worthington & Whitaker 2006). However, given that thisorthogonal factor accounts for <50% of the variance in to-tal scores, it is possible that a shorter version of the scalecould be developed.

List-wise deletion resulted in a sample of 319 partici-pants for PCA of the EBP Implementation scale. The screeplot indicated a single factor that the major discontinuityoccurred between the first and second factors. The first fac-tor had an eigenvalue of 10.53 and accounted for 59% of thevariance in the scale. One other factor had an eigenvalue >

1.0; at 1.5, it accounted for only 8% of the variance in thescale, not meaningfully more than the 5.5% variance thatcould be accounted for by any single item (Steger 2006).Again, the single-factor solution appeared to be the best fitto the data. Factor loadings for each item in the single fac-tor solution are shown in Table 2; all items have loadings>0.60. These high loadings combined with the high Cron-bach α (.96) indicate that a single unidimensional con-struct is being measured (Linn 1968; Hakstian et al. 1982;Nunnally & Bernstein 1994; Steger 2006; Worthington &Whitaker 2006).

Criterion Validity: Known Groups ComparisonsPrior exposure. Participants who indicated that they hadbeen exposed to the principles of EBP in school, duringcontinuing education coursework, or by reading profes-sional literature were compared to those who indicated thatthey had little or no prior exposure to EBP. Average scoreson the scales are shown for each subgroup in Table 4. In-dividuals who had prior exposure to EBP had beliefs aboutEBP that were similar to those who had no prior exposure.However, those with prior exposure scored twice as highon average on the Implementation scale.

The nature of this relationship is further indicatedby the different correlations between scores on the Be-liefs and Implementation scales for the two groups. Al-though both correlations were significantly different from0 (p < .001), the correlation was significantly higher(p = .05) for participants who had prior exposure to EBP

(r = 0.51) than for participants with little or no experi-ence with EBP (r = 0.35). That is, when participants hadprior exposure to EBP through formal training, their be-liefs in EBP were more strongly related to the frequencywith which they implemented EBP.

This pattern of relationships shows that, in a sample ofindividuals seeking new or additional training in EBP, indi-viduals might have strong beliefs in EBP that are not basedon formal training in EBP (the two groups were similar inbeliefs), but lack of training in EBP might be a barrier toactual implementation of EBP (the two groups were dis-similar in implementation and implementation was morestrongly correlated with beliefs among those with priortraining).Education level. Participants were divided into five sub-groups on the basis of the highest level of educationachieved. The strength of beliefs in EBP significantly in-creased with the level of education, F(4, 344) = 7.03,p < .001: Participants with associate degrees scored low-est (M = 49.70, SD = 19.95), while those with doctoraldegrees scored highest (M = 64.06, SD = 9.14). Similarly,scores on the Implementation scale significantly increasedwith level of education, F(4, 331) = 7.46, p < .001: Partic-ipants with associate degrees scored lowest (M = 8.37, SD= 12.96), while those with doctoral degrees scored highest(M = 25.50, SD = 21.08).Nursing roles. Four subgroups of participants had rolesthat should systematically vary in the number of opportu-nities for implementing EBP: staff nurse (N = 111), nursemanager (N = 33), clinical nurse specialist (N = 27), andeducator/faculty (N = 59). The strength of beliefs in EBPsignificantly increased from staff nurse to educator/faculty,F(3, 233) = 9.34, p < .001: Staff nurses scored lowest (M =48.72, SD = 21.63) and educator/faculty scored highest (M= 61.50, SD = 8.51). The frequency of implementing EBPsignificantly increased from staff nurse to educator/faculty,F(3, 226) = 6.97, p < .001: Staff nurses scored lowest (M= 10.36, SD = 13.54) and educator/faculty scored highest(M = 20.85, SD = 18.71).Age. Participants were divided into five subgroups on thebasis of age decades. The strength of beliefs in EBP signifi-cantly increased with age, F(4, 337) = 5.60, p < .001: Theyoungest participants, age 21 to 30 years, scored lowest(M = 48.35, SD = 23.87), while the oldest, age 61 to 70years scored highest (M = 59.75, SD = 4.74). Scores onthe Implementation scale did not vary significantly withage. Age was not meaningfully associated with role, time incurrent position, employment status, shift worked, hoursworked per week, or educational level. Thus, it appearsthat unidentified chronological or secular trends operateto increase endorsement of EBP, but not implementation ofEBP across the life span.



DISCUSSION

Both the EBP beliefs and implementation scales had goodpsychometric properties. Response patterns on individualitems and total scores indicated that the scales were sen-sitive to a wide range of attitudes and behaviours. HighCronbach coefficients indicated that the internal consis-tency of the scales was excellent. Although the choiceof factor analysis procedure in exploratory factor analy-sis and the best method of extracting factors have beenwidely debated (Linn 1968; Hakstian et al. 1982; Steger2006; Worthington & Whitaker 2006), the pattern of re-sults seen in this study clearly support the premise thatthe scales are measuring unidimensional constructs. Thecriterion validity of the scales was supported by the dif-ferences seen between subgroups. The strength of beliefsin EBP was strongly associated with the frequency of im-plementing EBP and this relationship was strongest amongparticipants who had prior training in EBP, suggesting thatwhile formal training in EBP was not a prerequisite to be-liefs about EBP, training facilitated implementation of EBP.Similarly, level of education was strongly associated withbeliefs about EBP and implementation of EBP, suggestingthat graduate education increases appreciation of the pos-itive impact of EBP and instills a desire to use EBP to im-prove patient outcomes. Like educational level, role wassignificantly associated with EBP beliefs and implementa-tion with nurse educators and faculty having significantlystronger beliefs in EBP and implementing EBP significantlymore frequently than did staff nurses.

IMPLICATIONS FOR RESEARCH,EDUCATION, AND PRACTICE

This initial validation study had some limitations. Test-retest reliability was not measured, so the instruments’ sta-bility is unknown. Cross-validation studies are needed toconfirm the factor structure of the scales, to derive the op-timal scale length, and to establish the reliability and valid-ity of the scales in other populations. A longitudinal studyshould be conducted to determine the predictive validityof the EBP Beliefs scale (i.e., whether the strength of EBPbeliefs allows predictions of EBP implementation rates).Such a study could also allow reassessment of the relation-ship of age to EBP beliefs and implementation rates anddefining how the relationship of beliefs to implementationis affected by chronological and secular trends. Finally, acontrolled intervention study should be conducted to as-sess the instruments’ sensitivity to changes in beliefs andimplementation as a function of educational intervention.

The results of this study were consistent with priorwork suggesting that beliefs are amenable to change witheducational interventions (Melnyk 1995; Melnyk et al.

1997; Melnyk & Feinstein 2001; Melnyk et al. 2004a;Melnyk et al. 2006). Therefore, strategies such as edu-cational and skills-building sessions on EBP with nursesshould strengthen their beliefs about the value of EBP andtheir ability to implement it, which should, in turn, in-crease its implementation rate. Concepts of EBP can beintroduced into educational curricula at all levels. In as-sociate degree education, nursing students can be exposedto the EBP paradigm, helped to develop a spirit of inquiry,and taught the importance of data-driven decision mak-ing. In addition to these EBP foundations, baccalaureatecurricula should emphasise how to use valid research toinform EBP (American Association of Critical Care Nurses2004; Ciliska 2005). Furthermore, baccalaureate programsshould teach students the steps of EBP so they can use thisproblem-solving approach to deliver the highest quality ofcare. Graduate education programs should emphasise theneed to expand the science of EBP principles and measurethe effect of EBP on clinical outcomes.

In order for EBP to be consistently implemented inhealth care organisations, a culture of best practice needsto be established, in which all nursing professionals, re-gardless of educational preparation, have an important rolein advancing EBC. The development of EBP mentors maybe key to implementing and sustaining an EBP culture(Fineout-Overholt et al. 2005; Melnyk 2007). In a recentrandomised controlled pilot study with staff nurses in aVisiting Nurse Service in the Northeastern U.S., presenceof an ARCC EBP mentor led to higher EBP beliefs in nurses,which in turn, led to greater EBP implementation (Levinet al. 2007). Subsequently, greater EBP implementation bythe nurses who had an ARCC mentor versus an advancedpractice nurse who taught physical assessment skills led tohigher group cohesion, which is a predictor of nurse jobsatisfaction and retention (Levin et al. 2007). Therefore,ARCC advanced practice mentors might be central to notonly advancing EBC, but also to enhancing nurse satisfac-tion and retention at a time in which several countries arefacing serious nursing shortages.

CONCLUSION

The EBP Beliefs and EBP Implementation scales are psycho-metrically sound instruments that can be used to systemat-ically study the effect of EBP educational and mentorshipprograms on EBP skills, clinical care, job satisfaction, andretention.

ReferencesAmerican Association of Critical Care Nurses. (2004).

Position statement on nursing research. Retrieved May28, 2008, from http://www.aacn.nche.edu/Publications/positions/NsgRes.htm.



Arora N.K., Avanian J.Z. & Guadagnoli E. (2005). Exam-ining the relationship of patients’ attitudes and beliefswith their self-reported level of participation in medicaldecision-making. Medical Care, 43(9), 865–872.

Balas E.A. & Boren S.A. (2000). Managing clinical knowl-edge for healthcare improvements. In V. Schattauer(Ed.), Yearbook of medical informatics (pp. 65–70). NewYork: Stuttgart.

Carver C.S. & Scheier M.F. (1982). Control theory: A use-ful conceptual framework for personality-social, clini-cal, and health psychology. Psychological Bulletin, 92,111–135.

Carver C.S. & Scheier M.F. (1998). On the self-regulation ofbehavior. Cambridge, UK: Cambridge University Press.

Ciliska D. (2005). Educating for evidence-based practice.Journal of Professional Nursing, 21(6), 345–350.

Ciliska D., Chambers L., Hayward S., James M. & Un-derwood J. (1996). Increasing evidence-based decisionsabout public health interventions. Canadian Journal ofPublic Health, 87(6), 376, 410.

Eccles M., Grimshaw J., Walker A., Johnston M. & PittsN. (2005). Changing the behavior of healthcare profes-sionals: The use of theory in promoting the uptake of re-search findings. Journal of Clinical Epidemiology, 58(2),113–116.

Fineout-Overholt E., Melnyk B.M. & Schultz A. (2005).Transforming health care from the inside out: Advancingevidence-based practice in the 21st century. Journal ofProfessional Nursing, 21(6), 335–344.

Goodwin L.D. (2002). Changing conceptions of measure-ment validity: An update on the new standards. Journalof Nursing Education, 41, 100–106.

Greiner A. & Knebel E. (2003). Health professions educa-tion: A bridge to quality. Washington, DC: National Aca-demic Press.

Grimshaw J., Eccles M., Thomas R., MacLennan G., Ram-say C., Fraser C. & Vale L. (2006). Toward evidence-based quality improvement. Evidence (and its limita-tions) of the effectiveness of guideline dissemination andimplementation strategies 1966–1998. Journal of Gen-eral Internal Medicine, 21(Suppl. 2), S14–S20.

Hakstian A.R., Rogers W.D. & Cattell R.B. (1982). Thebehavior of numbers of factors rules with simulated data.Multivariate Behavioral Research, 17, 193–219.

Heater B., Becker A. & Olsen R. (1988). Nursing interven-tions and patient outcomes: A meta-analysis of studies.Nursing Research, 37, 303–307.

Hutchinson A.M. & Johnston L. (2006). Beyond the barri-ers scale: Commonly reported barriers to research use.Journal of Nursing Administration, 36(4), 189–199.

Kajermo K.N., Nordstrom G., Krusebrant A. & BjorvellH. (2000). Perceptions of research utilization: Compar-

isons between health care professionals, nursing stu-dents and a reference group of nurse clinicians. Journalof Advanced Nursing, 31(1), 99–109.

Larrabee J.H., Janney M.A., Ostrow C.L., Withrow M.L.,Hobbs G.R. Jr. & Burant C. (2003). Predicting regis-tered nurse job satisfaction and intent to leave. Journalof Nursing Administration, 33, 271–283.

Levin R., Vetter M.J., Fineout-Overholt E., Melnyk B.M.& Barnes M. (2007). Advancing research and clinicalpractice through close collaboration (ARCC): A pilottest of an intervention to improve evidence-based careand patient outcomes in a community health setting. Pa-per presented at the 8th Annual National/InternationalEvidence-Based Practice Conference, Translating Re-search into Best Practice with Vulnerable Populations.Phoenix, Arizona, February 23.

Linn R.L. (1968). A Monte Carlo approach to the numberof factors problem. Psychometrika, 33, 37–71.

Maljanian R., Caramanica L., Taylor S.K., MacRae J.B. &Beland D.K. (2002). Evidence-based nursing practice,Part 2: Building skills through research roundtables.Journal of Nursing Administration, 32(2), 85–90.

Melnyk B.M. (1994). Coping with unplanned childhoodhospitalization: Effects of interventions on mothers andchildren. Nursing Research, 43(1), 50–55.

Melnyk B.M. (1995). Coping with unplanned childhoodhospitalization: The mediating functions of parental be-liefs. Journal of Pediatric Psychology, 20(3), 299–312.

Melnyk B.M. (2007). The evidence-based practice mentor.A promising strategy for implementing and sustainingEBP in healthcare systems. Worldviews on Evidence-BasedNursing, 4(3), 123–125.

Melnyk B.M. & Feinstein N.F. (2001). Mediating functionsof maternal anxiety and participation in care on youngchildren’s posthospital adjustment. Research in Nursing& Health, 24, 18–26.

Melnyk B. & Fineout-Overholt E. (2002). Putting researchinto practice. Reflections on Nursing Leadership, 28, 22–25.

Melnyk B.M. & Fineout-Overholt E. (2005a). Evidence-based practice in nursing & healthcare. A guide to bestpractice. Philadelphia: Lippincott, Williams & Wilkins.

Melnyk B.M. & Fineout-Overholt E. (2005b). Outcomesand implementation strategies from the first U.S.Evidence-Based Practice Leadership Summit. World-views on Evidence-based Nursing, 2, 185–193.

Melnyk B.M., Alpert-Gillis L.J., Hensel P.B., Cable-BillingR.C. & Rubenstein J. (1997). Helping mothers cope witha critically ill child: A pilot test of the COPE Intervention.Research in Nursing and Health, 20, 3–14.

Melnyk B.M., Alpert-Gillis L., Feinstein N.F., FairbanksE., Schultz-Czarniak J., Hust D., Sherman L., LeMoine



C., Moldenhauer Z., Small L., Bender N. & SinkinR.A. (2001). Improving cognitive development of LBWpremature infants with the COPE program: A pilotstudy of the benefit of early NICU intervention withmothers. Research in Nursing and Health, 24, 373–389.

Melnyk B.M., Alpert-Gillis L., Feinstein N.F., Crean H.,Johnson J., Fairbanks E., Small L., Rubenstein J., SlotaM. & Corbo-Richert B. (2004a). Creating opportunitiesfor parent empowerment (COPE): Program effects onthe mental health/coping outcomes of critically ill youngchildren and their mothers. Pediatrics, 113(6). AccessedDecember 17, 2004, from http://www.pediatrics.org/cgi/content/full/113/6/e597-e607.

Melnyk B.M., Fineout-Overholt E., Feinstein N., Li H.,Small L., Wilcox L. & Kraus R. (2004b). Nurses’ per-ceived knowledge, beliefs, skills, and needs regardingevidence-based practice: Implications for acceleratingthe paradigm shift. Worldviews on Evidence-based Nurs-ing, 1, 185–193.

Melnyk B.M., Fineout-Overholt E., Stetler C. & Allan J.(2005). Outcomes and Implementation Strategies fromthe First U.S. EBP Leadership Summit. Worldviews onEvidence-Based Nursing 2(3), 113–121.

Melnyk B.M., Feinstein N.F., Alpert-Gillis L., Fairbanks E.,Crean H.F., Sinkin R., Stone P.W., Small L., Tu X. & GrossS.J. (2006). Reducing premature infants length of stayand improving parents’ mental health outcomes withthe COPE NICU program: A randomized clinical trial.Pediatrics, 118(5), 1414–1427.

Nunnally J.C. & Bernstein I.H. (1994). Psychometric theory(3rd Ed.). New York: McGraw-Hill.

Parahoo K. (2000). Barriers to, and facilitators of, researchutilization among nurses in Northern Ireland. Journal ofAdvanced Nursing, 31(1), 89–98.

Pravikoff D.S., Pierce S.T. & Tanner A. (2005). Evidence-based practice readiness study supported by academynursing informatics expert panel. Nursing Outlook, 53,49–50.

Prochaska J.O. & Velicer W.F. (1997). The transtheoreticalmodel of health behavior change. American Journal ofHealth Promotion, 12(1), 38–48.

Prochaska J.M., Prochaska J.O. & Levesque D.A. (2001).A transtheoretical approach to changing organizations.Administration and Policy in Mental Health, 28(4), 247–261.

Retsas A. (2000). Barriers to using research evidence innursing practice. Journal of Advanced Nursing, 31(3),599–606.

Reynolds K.E. & Magnan M.A. (2005). Nursing atti-tudes and beliefs toward human sexuality: Collabora-tive research promoting evidence-based practice. Clini-cal Nurse Specialist, 19(5), 255–259.

Rhodes R.E. & Plotnikoff R.C. (2006). Understandingaction control: Predicting physical activity intention-behavior profiles across 6 months in a Canadian sample.Health Psychology, 25(6), 292–299.

Sackett D.L., Straus S.E., Richardson W.S., Rosenberg W.& Hayes R.B. (2000). Evidence-based medicine: How topractice and teach EBM. London: Churchill Livingstone.

Steger M.F. (2006). An illustration of issues in factor extrac-tion and identification of dimensionality in psycholog-ical assessment data. Journal of Personality Assessment,86(3), 263–272.

Strout T.D. (2005). Curiosity and reflective thinking: Re-newal of the spirit. Excellence in Nursing Knowledge. In-dianapolis, IN: Sigma Theta Tau International.

U.S. General Accounting Office. (2001). Nursing workforce:Emerging nurse shortages due to multiple factors. Reportto the Chairman, Subcommittee on Health, Committee onWay and Means, House of Representatives [GAO-01–944].Washington, DC: Author.

Vahey D.C., Aiken L.H., Sloane D.M., Clarke S.P. & VargasD. (2004), Medical Care, 42(2 Suppl): II 57–66.

Worthington R.L. & Whitaker T.A. (2006). Scale develop-ment research a content analysis and recommendationsfor best practices. The Counseling Psychologist, 34(6),806–838.


Documents

The Evidence-Based Practice Beliefs and Implementation Scales: Psychometric Properties of Two New Instruments