SEARCH

SEARCH BY CITATION

Keywords:

  • breast care;
  • evidence-based practice;
  • knowledge transfer;
  • RCT;
  • exercise and breast cancer;
  • health promotion;
  • knowledge transfer;
  • research dissemination

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. BACKGROUND
  4. AIM OF STUDY
  5. METHODS
  6. RESULTS
  7. DISCUSSION
  8. IMPLICATIONS AND CONCLUSIONS
  9. References

Background: In the United Kingdom (UK), it was documented that a problem of knowledge transfer existed within the speciality of breast-cancer care, thus depriving patients of receiving optimal care. Despite increasingly robust research evidence indicating recommendation of whole body exercise for people affected by breast cancer, commensurate changes to practice were not noted amongst breast-care nurses (BCNs).

Aim: To evaluate the effect of a targeted booklet, Exercise and Breast Cancer: A Booklet for Breast-Care Nurses, on changes in knowledge, reported practice, and attitudes of BCNs in the UK.

Method: A prospective, experimental approach was used for designing a pre- and post-test randomised controlled study. Comparisons of knowledge, reported practice, and attitudes based on responses to a questionnaire were made at two time-points in two groups of BCNs (control and experimental). The unit of randomisation and analysis was hospital clusters of BCNs. The sample comprised 92 nurses from 62 hospitals. Analysis consisted of descriptive statistics and clustered regression techniques: clustered logistic regression for knowledge items, clustered linear regression for knowledge scores, ologit for attitude and reported practice items, and clustered multiple regression for paired and multiple variable analysis.

Results: A statistically significant increase in knowledge and changes in reported practice and attitudes were found. Robust variables affecting knowledge acquisition were: promotion of health, promotion of exercise, and understanding how exercise can reduce cancer-related fatigue.

Discussion: The study has shown that evidence-based printed material, such as an information booklet, can be used as an effective research dissemination method when developed for needs, values, and context of a target audience.

Conclusions: This practical approach to research dissemination could be replicated and applied to other groups of nurses.


BACKGROUND

  1. Top of page
  2. ABSTRACT
  3. BACKGROUND
  4. AIM OF STUDY
  5. METHODS
  6. RESULTS
  7. DISCUSSION
  8. IMPLICATIONS AND CONCLUSIONS
  9. References

Nursing care based on research evidence is a clinical and professional imperative that has globalimplications. Historically, nurses have struggled to introduce research-based interventions into routine clinical practice. Reasons for this difficulty are varied and complex ranging from poor communication between clinical and academic based nurses (Estabrooks et al. 2003) to individual and organisational barriers that obstruct the implementation of research evidence into practice (Parahoo 2000; Kirshbaum et al. 2004). Conclusions from an extensive body of literature show many of the same difficulties associated with the way research is communicated, the skills of nurses, the research culture of the organisation, and the quality of the research itself (Walsh 1997; Dunn et al. 1998; Kajermo et al. 1998; Rutledge et al. 1998; Retsas & Nolan 1999; Closs et al. 2000; Parahoo 2000; Kirshbaum et al. 2004). The main barriers to research utilisation, such as: not understanding research reports, insufficient time to read research, and not having the authority to change practice, appear to remain consistent, universal, and deeply rooted in nursing culture and practice the world over. These barriers cannot be addressed easily or sufficiently in the short term. Multiple, systematic, and innovative strategies are required to remove barriers and to strengthen facilitating aspects.

The call for a pragmatic and systematic approach to promoting research utilisation and evidence-based practice has been expressed by many (Closs & Cheater 1994; Luker & Kendrick 1995; Pryjmachuk 1996; Mulhall et al. 1998; Thompson et al. 2001; Tolson et al. 2006) and is an objective that continues to require attention. Within the context of the UK's National Health Service (NHS) but applicable worldwide, it appears that for an overall strategy to succeed, a balance of attention and resources is required to address three requirements of clinical effectiveness: obtaining evidence, implementing evidence, and evaluating the effect of changed practice (NHS Executive 1996). Because of what is known about the obstacles noted above, ways of improving research synthesis, access, and understanding through effective dissemination strategies need empirical study.

Typically, clinical information is communicated to health care professionals through various written and electronic formats such as guidelines, protocols, evidence synthesis bulletins, meta-analyses, systematic reviews, and journal articles. It has been assumed naively and erroneously that once a research report is produced, it is accessed, evaluated for quality and relevance, and then used to “inform” practice. However, this process of direct and seemingly effortless translation of research into clinical application is not a true reflection of reality (Oxman et al. 1995; NHS Centre for Reviews and Dissemination 1999).

The Cochrane Effective Practice and Organisational Care Group (CEPOCG) was established to review the effectiveness of dissemination interventions on predominately medical practitioners. In addition, critical literature includes: two comprehensive Effective Health Care Bulletins (NHS Centre for Reviews and Dissemination 1994, 1999) and several reviews on gathering, appraising, and synthesising literature on effective and ineffective methods of promoting research-based changes in clinical practice (e.g., Grimshaw & Russell 1993; Closs & Cheater 1994; Thomson et al. 1998; Lock et al. 1999; Grimshaw et al. 2001; Thomas et al. 2003).

Universal agreement is found in both medical and nursing literature that providing practitioners with valid, reliable, credible, authoritative, effective, and ultimately “useful” research evidence to inform clinical practice is highly complex. Many varied approaches to communicating clinically relevant information exist, yet few have been empirically evaluated in nursing populations. Findings from the array of published reviews (listed above) indicate that:

  • • 
    Written materials such as information packs and booklets are associated with improvements in specific, topical knowledge and have the potential to have an effect on reported practice.
  • • 
    Educational methods provided within acute health care institutions are varied with some programmes combining several types of educational interventions along with practical applications. However, no generalisations about which programmes are most effective can be made.
  • • 
    A collaborative and facilitative approach to dissemination of research evidence may prove to be suitable for nurses; however, insufficient data are available about how this can be optimally achieved.

Printed educational material is a relatively low cost and widely used strategy for disseminating clinical information, but there is an interesting inconsistency between the medical and nursing literature. In a Cochrane Review of 11 studies of physician practice it was concluded that standardised printed educational materials did not produce a change in clinical practice when used on their own (Freemantle et al. 2001). The conclusion of this review was that a largely passive mode of information exchange was insufficient. Given that accessibility of research findings remains a barrier for some groups of nurses (MacGuire 1990; Thompson et al. 2001; Kirshbaum et al. 2004), finding out why printed educational materials were ineffective would seem to be important. One possibility is that a more personalised and targeted version of synthesised educational materials might meet the information needs of nurses.

In contrast to the conclusions reached in the medical literature, two independent, well-conducted experimental studies evaluating personalised and targeted strategies for disseminating research-based educational material to nurses found different results (Luker & Kendrick 1995; Williams et al. 1997). These two studies provided substantial evidence in support of the effectiveness of printed materials in improving nurses' knowledge of leg ulcer care (Luker & Kendrick 1995) and continence care (Williams et al. 1997). Shared elements of their successful strategies relevant to nursing include: a preliminary assessment of the target population, an identification of local barriers and limitations, and the identification and participation of local opinion leaders as potential early adopters of a research-based change in practice.

Nurses share many values and priorities with their medical colleagues, yet they subscribe to a profession that has a distinctly different focus, history, culture, educational preparation, status, and social dimension (Kneafsey 2000; Farrell 2001; Watts et al. 2001). Therefore the qualities for successful research dissemination could also be different. It cannot be assumed that the conclusions reached from the findings of predominately medically oriented reviews and studies will be directly transferable to nursing. Until more is understood about the complex relationship between research knowledge and nursing practice, patient care may be detrimentally affected. Empirically investigating research dissemination from a nursing perspective could improve knowledge in this area.

In the UK, it was reported that a problem of dissemination and utilisation of research-based knowledge existed within the specialty of breast cancer (Kirshbaum et al. 2004), thus depriving individuals of receiving optimum care. Despite increasingly robust research evidence showing the numerous benefits of aerobic exercise for people affected by breast cancer (Courneya et al. 2003; Mock et al. 2005), commensurate changes to practice were not occurring. To assist in addressing this deficiency, a three-stage study was designed to identify the barriers to research utilisation and preferred methods of research dissemination; and develop and evaluate a dissemination intervention for BCNs. The preliminary stages have been published elsewhere (Kirshbaum et al. 2004; Kirshbaum 2005a,b, 2007). The evaluation stage will be presented here.

AIM OF STUDY

  1. Top of page
  2. ABSTRACT
  3. BACKGROUND
  4. AIM OF STUDY
  5. METHODS
  6. RESULTS
  7. DISCUSSION
  8. IMPLICATIONS AND CONCLUSIONS
  9. References

The aim of the study was to evaluate the effect of a targeted booklet: Exercise and Breast Cancer: A Booklet for Breast-Care Nurses, on changes in knowledge, reported practice, and attitudes of BCNs.

METHODS

  1. Top of page
  2. ABSTRACT
  3. BACKGROUND
  4. AIM OF STUDY
  5. METHODS
  6. RESULTS
  7. DISCUSSION
  8. IMPLICATIONS AND CONCLUSIONS
  9. References

Design

A longitudinal, prospective, experimental, clustered approach was chosen to evaluate a research-dissemination intervention. A randomised, controlled pre-test/post-test design (Campbell & Stanley 1963) enabled comparisons to be made at two time intervals between two groups of BCNs based on their responses to an analytical survey questionnaire. Each cluster, the unit of analysis, consisted of all BCNs working within a single hospital.

The Target Population

BCNs in the UK were the target population for the study. Because no national register was available that could be used as a sampling frame, a systematic approach was used to construct an up-to-date and comprehensive list. A database was compiled over several months by contacting regional breast-care nursing groups, consulting a national guide to breast-care services (Cancer Relief Macmillan Fund 1996) for telephoning each centre directly, and encouraging known BCNs to identify newly appointed or unlisted BCNs. At the time this research began, formal approval from a local research ethics committee was not required because study participants were professional nurses.

Sample Size

The size of the sample is determined by the outcome measures to be used (Wilson & Rose 1998). This study used a series of outcome measures, matched to the type of data (e.g., categorical, continuous, ordinal) and the objectives of statistical tests. The determination of the sample size was based on the two sample t-test because it is the appropriate test to compare two means from continuous data such as the comparison of knowledge scores between experimental and control groups; a key objective in the study.

A clustering approach was selected, which required the sample size to be calculated based on hospitals, rather than on individuals (see Donner et al. 1981). The estimate for sample size was based on detecting a change of one unit on the ranking scales used in the measurement tool, which would approximate one standard deviation. According to Machin and Campbell (1987, p. 87) the required number of participants per study group under these circumstances is 22; however, because the calculation was based on nonparametric ranking scales, a slight increase in sample size is recommended. As a result, the target number of hospitals per group was increased slightly to 24; a sample size that was deemed possible to attain even when accounting for nonresponding participants.

The sample consisted of a subset from the national population of BCNs approached previously in a national survey (Kirshbaum et al. 2004). The inclusion criterion was that the participant was currently a BCN in one of the northern geographic regions of England. Previous contact between the researcher and BCNs within professional networks in the northern regions was recognised as a potential factor in maximising rapport and trust: 137 BCNs working in 76 hospitals responded.

Randomisation

A stratified block cluster randomisation was undertaken in which the unit of randomisation and analysis was the hospital at which the BCNs worked. This was done to avoid unit-analysis error (Whiting-O'Keefe et al. 1984; Bero et al. 1998). It is important for the researcher to isolate and identify all possible experimental effects. If more than one clinician is providing the intervention under study, differences in outcome may be because of characteristics of the provider, not necessarily the intervention. Similarly, if the patient, rather than the provider of the intervention, is identified as the unit of randomisation and unit of analysis and is allocated to an experimental group, a unit-analysis error may result. Under these conditions, the significance of observed effects may be overestimated (Bero et al. 1998). It was recognised that BCNs frequently work with other BCNs in the same hospital and usually in the same office. The potential influence caused by their usual day-to-day interaction was acknowledged. Differences in outcome between the experimental and control group could have been affected not just by the intervention (the information booklet) but because of other methods of dissemination such as informal discussions or team meetings. Selecting the hospital as the unit of randomisation and analysis ensured that all nurses working at the same hospital were allocated to the same study group (experimental or control), thus decreasing the likelihood of error.

Findings from a national survey of BCNs (Kirshbaum et al. 2004), indicated that compared with those who are based at district general hospitals, nurses who worked in teaching hospitals or specialist hospitals had more formal education and made more frequent use of research. This finding indicated type of hospital (general or specialist) as a variable that might influence outcome. Therefore, every hospital was categorized as general or specialty before randomisation. As part of randomisation the sample was stratified for type of hospital.

The unit of randomisation throughout was the hospital coded by hospital number. All nurses working at the same hospital were allocated to the same group. A block randomisation procedure described by Altman (1999) was used to reduce bias between study groups and promote high levels of internal validity.

The Intervention

A critical review was undertaken to identify, assess, and synthesise empirical data about breast cancer and physical exercise (Kirshbaum 2007). In the exercise booklet, research evidence to meet the interests and requirements of the target audience was described. Specific attributes and characteristics of the experimental dissemination method (the intervention) were derived from the results of (1) a national survey of barriers to research utilisation of breast-care nurses (Kirshbaum et al. 2004) and (2) the development of a conceptual framework used for selecting a targeted intervention (Kirshbaum 2005b).

Specific information requirements of the intended audience were also identified. Information was gathered to describe the targeted group, identify important criteria concerning innovations or interventions and the applicability of the determinants of behavioural change (Ajzen & Madden 1986; Ajzen 1991; Prochaska et al. 1992; Rogers 1995; Tones & Tilford 2001). Descriptive data derived from the national survey of BCNs were integral to understanding the target group. Information collected included demographic details, data about perceived views of barriers and facilitators of research utilisation, preferences for dissemination methods, and BCNs' comments about their authority to change practice, autonomy in the workplace, and associations within networks and multidisciplinary teams. It was proposed that the dissemination method needed to be accessible, understandable, time efficient, communicated clearly, research-based, critical, interesting, relevant, and practical. To optimise effectiveness, the intervention would ideally acknowledge possible reservations and limitations of implementing recommended changes and integrate positive change-agent characteristics (e.g., credibility, perceived expertise, rapport, respect, and trustworthiness).

The aim for the information booklet was to address proposed attributes of an effective dissemination method (as listed above). The booklet was written by the researcher who achieved credibility and expertise in breast-care nursing throughout 15 years of clinical practice and research, and had participated in regional, national, and international conference presentations and publication activities. The personalised cover letters were deliberately aimed at achieving rapport and trustworthiness by specifying the background and purpose of the research, who was leading it, what it entailed, and the importance of participating to further evidence-based practice.

The booklet indicated the special experience of the target group (BCNs), had professional terminology, and indicated in the title that it was intended for the use of BCNs. The booklet began with a summary of the physical and psychological needs of breast-cancer patients and the text was structured into eight sections: Introduction, The Challenges of Breast Cancer, The Benefits of Exercise, What Type of Exercise is Best?, Implications for Nursing Practice, Summary and Implications, References, and Table of Empirical Studies.

A panel of experts working in the field of breast cancer, including a surgeon, clinical psychologist, and lecturer in cancer nursing were asked to review and confirm clinical accuracy of the booklet. Recommendations from these experts were incorporated into the final version, which consisted of 18 pages of text, 6 pages of references and a 3-page table that showed details of 18 empirical studies on the benefits of exercise for breast-cancer patients. The product was posted to all participating nurses and therefore highly accessible.

Data Collection Instrument

The Exercise and Breast Cancer Questionnaire was developed to test the hypotheses that a booklet designed for BCNs could facilitate changes in knowledge, reported practice, and attitude outcomes. Because no validated or applicable measure was identified in the literature, a questionnaire was developed following a structured approach for the purpose, type, order, and wording of questions to be included (Oppenheim 1992). Questionnaire items were developed after a critical review of the literature.

The Exercise and Breast Cancer Questionnaire was organised into four sections: (1) demographic data, (2) questions about reported practice, (3) questions to test knowledge, and (4) questions about attitudes. Nurses were asked to use a 5-point Likert scale (1 = always to 5 = never) to record how often they would recommend exercise to their patients for each of 12 common quality of life problems such as weight gain, insomnia, and fatigue.

The third section consisted of 17 research-based statements with three options: “True,”“False,” and “Don't know.” A “knowledge score” was calculated to indicate the number of correctly answered knowledge questions (maximum score = 17). Areas of knowledge selected for inclusion were determined by the magnitude of their clinical importance (e.g., contraindications to exercise, patient benefit) as assessed by the researcher and clinical experts consulted throughout the development phase of the questionnaire. Every statement was supported by undisputed, empirical evidence. The order of statements was selected randomly and followed no particular pattern associated with the level of difficulty, correct answer (“true” or “false”), or focus.

In the last section, a 5-point Likert scale was used to elicit attitudes and beliefs surrounding the topic area. The pre-test questionnaire contained nine statements and the post-test version included two additional questions to record if respondents had actually received and read the booklet. The attitude statements were conceived with one overarching objective in mind, namely, to explore BCNs' views of promoting exercise as part of their professional role.

Drafts of the questionnaire were reviewed by academic colleagues and other experts in the field of breast cancer who confirmed content validity. The questionnaire was pilot tested on nurses who were not BCNs and research nurses at an oncology centre. Minor changes were made to improve the clarity of questions.

Data Collection Procedures

All 137 participants received a pre-intervention (baseline) questionnaire that was sent through the post along with a cover letter and self addressed, stamped return envelope. A reminder was posted to all nonresponding BCNs after 3 weeks. In total, 112 BCNs responded and these were randomised into two study groups: experimental (n= 56) and control (n= 46). After 6 weeks from the initial mailing date, participants assigned to the experimental group were sent the booklet; 2 months passed before the post-intervention (follow-up) questionnaire was sent to nurses in the control and experimental groups. After 4 weeks, a reminder was posted to all nonresponding BCNs. After another 4 weeks, a copy of the booklet was posted to all nurses assigned to the control group for information and use. Data was collected during 2002–2003.

Statistical Methods

Hospital cluster, identified as the unit of randomisation, also served as the unit of analysis. A series of clustered regression techniques based on the adaptation of standard regression methods for clustering models, known as the estimation of robust standard errors, was chosen for this study because it allows for the adjustment of both individual- and cluster-level covariates while testing for the effects of an intervention (Donner 1998).

Clustered regression analysis was applied (STATA 2001a). This technique allows users to determine if there is a relationship between variables and uses the notion of a straight line to develop a prediction equation to describe particular types of patterns in the empirical data. “An estimate of the intra-cluster correlation coefficient statistic is used to account for the variation within clusters” (Campbell 2001, p. 82). The estimated regression coefficient is the change in the dependent variable for a change in 1.0 in the independent variable.

Clustered linear regression (for nurses clustered within hospitals) was used as a model to estimate the relationship between an interval dependent variable (e.g., knowledge score) and one or more interval or binary independent variables (e.g., study group).

To analyse correctly clustered ordinal data obtained for the practice and attitude items on the questionnaire, ologit with “bootstrapping” was advised by statisticians. Ologit is a technique of ordered logistical estimation available in the STATA statistical software package (STATA 2001b). The method of calculation is similar to logistic regression but allows relationships between ordered dependent variables and a set of independent variables to be estimated (STATA 1997).

The bootstrap is a “data-based simulation method for statistical inference, which can be used to study variability of estimated characteristics of the probability distribution of a set of observations and provide confidence intervals for parameters in situations where they are difficult or impossible to derive in the usual way” (Everitt 1995, p. 32). In this study the only way to arrive at a statistic of variability for the responses within clusters for the ordinal data was through the use of multiple random samples. The bootstrap procedure utilised random samples from the original data to provide a reliable estimate for the set of observations.

RESULTS

  1. Top of page
  2. ABSTRACT
  3. BACKGROUND
  4. AIM OF STUDY
  5. METHODS
  6. RESULTS
  7. DISCUSSION
  8. IMPLICATIONS AND CONCLUSIONS
  9. References

Participants

In all, 104 baseline questionnaires were returned from eligible BCNs working in 63 hospitals—a 76% response rate; 92 follow-up questionnaires were returned, for an overall response rate of 69%.

Data from the national survey were used to test external validity of the evaluation study (Kirshbaum et al. 2004). Five variables were selected to compare responses of BCNs from the Northern counties with those from other areas of the UK (Table 1).

Table 1.  Comparison of BCNs from northern counties to BCNs from other areas in the UK
VARIABLESNORTHERN AREAS n= 152 (%)OTHER AREAS IN UK n= 111 (%)CHI SQUARE
Type of hospitalDistrict General Hospital (DGH)TeachingDGHTeachingχ2= 3.00 df= 1 p= 0.083
 81 (53.3) 71 (46.7)71 (64.0)40 (36.0) 
Degreeyesnoyesnoχ2= 0.09 df= 1 p= 0.769
 38 (25.0)114 (75.0)26 (23.0)85 (76.6) 
Breast-care courseyesnoyesnoχ2= 7.25 df= 1 p= 0.007
114 (75.0) 38 (25.0)98 (88.3)13 (11.7) 
Research courseyesnoyesnoχ2= 0.32 df= 1 p= 0.574
 26 (17.1)126 (82.9)22 (19.8)89 (80.2) 
Work with other BCNsyesnoyesnoχ2= 1.23 df= 1 p= 0.268
113 (74.3) 39 (25.7)89 (80.2)22 (19.8) 
Years in nursingmeanSDmeanSDt =−.27 df= 261 p= .783
19.08.119.28.5 

Of the five variables used in the comparison, only one (breast-care course) indicated a significant difference. From these data it is apparent that overall, the two groups were similard but that fewer BCNs from the Northern counties had attended a breast-care nursing course. The potential effect of this difference was noted but considered negligible because of the overwhelming similarities between the two groups; also no documented evidence existed to indicate that the breast-course variable would exert a strong influence.

The sample of 92 BCNs consisted of 52 respondents from general hospitals and 40 from specialist hospitals (Table 2). Fifty-two BCNs (56.6%) had 5 or more year's experience in their current role; 86 BCNs (93.5%) worked with at least 1 other BCN; of these, 18 (19.5%) worked with 3 or more BCNs. Ninety nurses (97.8%) completed a specialist BCN course and 52 (56.5%) completed a separate research course or module.

Table 2.  Demographic details of breast-care nurses, n= 92
INDEPENDENT VARIABLESEXPERIMENTAL GROUP n (%)CONTROL GROUP n (%)COMBINED n (%)
Type of hospital
 General27 (52.9)25 (61.0)52 (56.5)
 Specialist24 (47.1)16 (39.0)40 (43.5)
Years in breast-care nursing
 Less than 2 yrs.3 (5.9) 5 (12.2)8 (8.7)
 2-4 yrs18 (35.3)14 (34.1)32 (34.8)
 5-10 yrs20 (39.2)13 (31.7)33 (35.9)
 More than 10 yrs10 (19.6) 9 (22.0)19 (20.7)
Work with other BCNs
 Yes48 (94.1)38 (92.7)86 (93.5)
 No3 (5.9)3 (7.3)6 (6.5)
Work with how many others
 1 or 235 (73.0)33 (87.0)68 (79.1)
 3 or more13 (27.0) 5 (13.0)18 (20.9)
Completed breast-care course
 Yes 51 (100.0)39 (95.0)90 (97.8)
 No0 (0)2 (5.0)2 (2.2)
Completed research course
 Yes28 (54.9)24 (58.5)52 (56.5)
 No23 (45.1)17 (41.5)40 (43.5)
Total514192

After randomisation, responses from BCNs in both study groups were compared to observe equivalence based upon baseline demographic details (Table 2). Using these data the numeric values of the study groups were similar for all but one characteristic; in the intervention group, 13 BCNs worked with 3 or more colleagues, whereas only 5 BCNs in the control group reported this characteristic.

Reported Practice, Knowledge, and Attitude Item Scores at Baseline

Baseline responses to the questionnaire were to describe the reported practice, level of knowledge, and attitudes of the sample concerning the topic of exercise and breast cancer. Comparisons were made between experimental and control groups based on mean scores to determine equivalence. No significant difference between groups was observed.

Effect of the Intervention

Clustered regression analyses were conducted to determine the effect of the study group (i.e., receiving the booklet or not) on the knowledge, reported practice, and attitudes of the sample using baseline and follow-up data. For each group of questionnaire items, a summary table of unclustered data is shown first followed by results from the clustered regression analyses. The experimental or control group was selected as the primary predictor (independent) variable. Responses to reported practice, knowledge, and attitude items were the outcome (dependent) variables.

Effect of Study Group on Knowledge

The intervention booklet appeared to markedly improve knowledge of the intervention group; this was shown by a greater number of correct responses to each of the 17 knowledge items compared to those by the control group (Figure 1).

image

Figure 1. Comparison of study groups for knowledge items at follow-up (Control n= 41, Experimental n= 51).

Download figure to PowerPoint

Following the analysis of unclustered data, 11 out of 17 knowledge items showed a significant predictive relationship between clustered units and responses to knowledge statements (Table 3). The odds ratio indicates the multiplicative change in the odds of a correct answer when the study group is changed from experimental to control. Low odds ratios (less than 1) are associated with improvement in knowledge because the experimental group was coded as Group 1 and the control group as Group 2. These results indicate that the study group was as a strong predictor variable associated with improving knowledge.

Table 3.  Effect of study group on responses to knowledge items (clustered logistic regression)
OUTCOME VARIABLEODDS RATIO95% CONFIDENCE INTERVALPROBUST STANDARD ERROR
Knowledge 1
 Exercise during the months of cytotoxic chemotherapy treatment will tend to make nausea more severe. (False)0.1200.040–0.417<0.0010.078
Knowledge 2
 If a structured exercise programme is followed for 10 weeks, breast-cancer patients can expect to experience less fatigue. (True)0.0940.034–0.259<0.0010.049
Knowledge 3
 For breast-cancer survivors, swimming has been shown to be the most beneficial form of physical exercise. (False)perfect predictionperfect prediction<0.001perfect prediction
Knowledge 4
 Whatever form of exercise is selected, anxiety and depression will only be reduced when there are cardiovascular benefits. (False)0.3200.143–0.714 0.0050.131
Knowledge 5
 IV chemotherapy within the previous 24 hours is a contraindication to exercise for breast-cancer patients. (True)0.2900.120–0.695 0.0060.129
Knowledge 6
 The main reason why women with breast cancer begin an exercise programme is the same as for healthy women. (False)0.2510.103–0.611 0.0020.114
Knowledge 7
 People who are obese, smoke, or are elderly have a high risk of dropping out from exercise programmes. (True)0.4230.175–1.021 0.0560.190
Knowledge 8
 Breast-cancer patients who participate in an exercise programme can expect to lose weight. (False)0.8980.395–2.042 0.7980.376
Knowledge 9
 Physical activity provides increased protection against breast cancer for post-menopausal women. (False)0.2260.071–0.723 0.0120.134
Knowledge 10
 Physical activity provides increased protection against breast cancer for post-menopausal women. (False)0.3960.156–1.005 0.0510.188
Knowledge 11
 Patients who want to continue their established exercise routines throughout breast- cancer treatments should be encouraged to continue without modification. (False)0.5320.236–1.201 0.1290.221
Knowledge 12
 Dizziness and vomiting within the previous 36 hours are contraindications to exercise for breast-cancer patients. (True)0.3910.156–0.982 0.0470.184
Knowledge 13
 To maintain a new behaviour such as regular exercise, it is necessary to have positive encouragement during and after the behaviour. (True) .6130.170–2.214 0.4550.402
Knowledge 14
 Breast-cancer patients who participate in an exercise programme can expect to increase their self-esteem. (True)0.7830.221–2.769 0.7040.505
Knowledge 15
 If a structured exercise programme is followed for 10 weeks, breast-cancer patients can expect to decrease their functional peak capacity (VO2Max). (False)0.3770.149–0.954 0.0340.179
Knowledge 16
 Vigorous repetitive upper body exercises should not be encouraged since this will probably result in arm lymphoedema. (False)0.2660.119–0.592 0.0010.109
Knowledge 17
 Women should be advised to wait at least 2 weeks after surgery before doing any form of physical exercise. (False)0.3600.153–0.852 0.0200.158

Effect of the Study Group on Reported Practice

The study group was examined for its effect on each reported practice item. Participants were asked the same questions as at baseline. Figure 2 shows the comparison between study groups using unclustered data.

image

Figure 2. Comparison of study groups for reported practice at follow-up. Control n= 41, Experimental n= 51 (unclustered data).

Download figure to PowerPoint

In the clustered analyses using ologit, the intervention was shown to significantly affect 3 out of the 12 practice items (Table 4). Significant results were found for the individual items in relation to: nausea, loss of appetite, and fatigue.

Table 4.  Can study group predict reported practice for follow-up responses?
PREDICTOR VARIABLEOUTCOME VARIABLEODDS RATIO95% CONFIDENCE INTERVAL
Study groupPractice 1: weight gain1.550.73–3.03
Study groupPractice 2: insomnia1.460.64–3.60
Study groupPractice 3: loss of libido1.920.94–3.64
Study groupPractice 4: panic attacks2.230.89–5.75
Study groupPractice 5: altered body image1.620.67–3.82
Study groupPractice 6: headaches2.410.98–5.42
Study groupPractice 7: nausea2.542.53–13.20
Study groupPractice 8: loss of appetite3.671.82–8.76
Study groupPractice 9: fatigue2.441.12–5.99

Effect of the Study Group on Attitude

The study group was examined for its effect on each reported attitude item. Participants were asked the same questions as at baseline. Figure 3 shows the comparison between study groups using unclustered data.

image

Figure 3. Comparison between study groups on attitude items at follow-up.

Download figure to PowerPoint

In testing for the effect of the intervention on changing attitudes using ologit (see Table 5), the results indicate that responses of the clustered units could be predicted for two items: I would promote exercise to my patients if I knew more about the associated benefits and limitations (Attitude 3) and I believe exercise is important for women with breast cancer (Attitude 9).

Table 5.  Can study group predict attitude for follow-up responses
PREDICTOR VARIABLEOUTCOME VARIABLEODDS RATIO95% CONFIDENCE INTERVAL
Study groupAttitude 1: Promoting health is an important part of the breast-care nurse's role.2.320.63–11.25
Study groupAttitude 2: The promotion of exercise is an important part of the breast-care nurse's role.0.780.31–1.90
Study groupAttitude 3: I would promote exercise to my patients if I knew more about the associated benefits and limitations.0.261.57–10.28
Study groupAttitude 4: I understand how exercise can reduce cancer-related fatigue.0.890.35–2.03
Study groupAttitude 5: Any intervention than can improve a patient's sense of control should be promoted.0.790.87–2.14
Study groupAttitude 6: It is insensitive to suggest increasing physical activity to someone who is emotionally distraught.1.390.58–3.22
Study groupAttitude 7: I have enough time to promote exercise to my patients.0.900.38–2.27
Study groupAttitude 8: Regular exercise is an important part of my lifestyle.2.230.94–5.47
Study groupAttitude 9: I believe exercise is important for women with breast cancer.2.690.14–0.92

The control group indicated a higher agreement with the belief that if they knew more about exercise, they would promote it to their patients. The results also show that study group had a strong predictive effect on believing that exercise is important for women with breast cancer.

DISCUSSION

  1. Top of page
  2. ABSTRACT
  3. BACKGROUND
  4. AIM OF STUDY
  5. METHODS
  6. RESULTS
  7. DISCUSSION
  8. IMPLICATIONS AND CONCLUSIONS
  9. References

The findings indicate that an information booklet, produced specifically in response to the expressed requirements of a targeted population, can be used to enhance clinically relevant knowledge. Within the context of promoting exercise to breast-cancer patients, attitude statements that indicated health promotion, exercise promotion, and the reduction of cancer-related fatigue were strongly associated with greater acquisition of knowledge; this outcome was independent from the experimental variable (the booklet). In addition, findings indicate that relationships among knowledge, reported practice, and attitudes “resist” linear interpretation.

This study has been set within the paradigm of evidence-based practice. As Kitson (2004) states, evidence-based medicine (EBM) incorporates many of the critical attributes of populations postulated by Rogers in his classic diffusion of innovations studies (Rogers 1995). Nurses might share with doctors a commitment to providing effective health-related care and use the output from EBM to inform clinical care (Harrison et al. 2002), but are less advanced in developing ways to promote practice that is grounded in evidence. The findings from this study have shown that nurses do indeed respond positively to written educational materials (e.g., targeted information booklet).

Knowledge

The targeted information booklet was clearly effective in improving knowledge. As an intervention, it was intended to provide a credible resource for BCNs, capable of conveying the simple message that exercise should be promoted to people with breast cancer. Given the recognition that information is critical to changing nursing practice (White et al. 1998), it is vital that clinicians be encouraged to obtain current research findings in an accessible format and to bring practice and theory together for patients' benefit (Crane 1995; Thomson et al. 1998).

The results indicated that the minimum goal of achieving basic awareness of the importance of exercise far exceeded expectations. When responses of the two groups of clustered BCNs were compared, a significant predictive relationship was shown between receiving the booklet and correctly answering 10 out of 17 knowledge questions. Of the seven questions that did not show a significant effect, three were answered correctly by over 90% of the total sample. Retrospectively, it would appear that these three questions reflected general rather than special knowledge and were probably easier to answer correctly.

Not surprising is that educational interventions formed the basis for many research utilisation programmes (Stetler 1994; Stetler et al. 1995; Lacey 1996; Titler et al. 1999; Rodgers 2000; Howell et al. 2001). In these examples, the broad spectrum of approaches is apparent. The programme described by Stetler et al. (1995) has succeeded by following cohorts of nurses from introductory research awareness sessions through active implementation and evaluation of innovations in practice at the bedside. In the context of a university setting, Lacey (1996) demonstrated that an introductory research course could enhance research-based practice and that the effect could continue to have an effect 6 months after the students completed the course. In common with the current study, Luker and Kendrick (1995) and Williams et al. (1997) demonstrated previously that “packages” of research findings in the form of handbooks or information booklets were also effective in improving nurses' knowledge of a specific body of evidence. The current study indicates support for the findings from these two earlier studies but also advances contributions by indicating the effect of information on reported practice and attitudes.

Reported Practice

Initiating changes to clinical practice is a major challenge both to achieve (Kanouse & Jacoby 1988) and evaluate (Waddell 2002). The booklet was found to be a predictive indicator for change in reported practice. A statistically significant increase was recorded in promoting exercise for the purpose of relieving nausea, fatigue, and loss of appetite. Relief of the first two of these symptoms had been supported by empirical studies cited in the booklet, but not the loss of appetite. An explanation for this anomaly in outcome might be that the nurses might understandably have made an association between loss of appetite and depressed mood state. Their experience and knowledge of the linkage between these two conditions might have led them to assume that exercise would affect both disturbances in the same positive way. Another possible though unsubstantiated explanation may be that the BCNs were incorporating aspects of their own “life experience” into practice by making their own links between exercise and a “healthy” appetite.

Earlier evaluations of written educational materials, where the respondents were mainly medical doctors (Freemantle et al. 2001), found a lack of effectiveness. By contrast, the booklet on exercise and breast cancer has been shown to influence a moderate change in the reported practice of nurses. A reason for the disparity may be that different sorts of risks are involved in changing practice by different types of practitioners. Innovations that involve altering a medical treatment or surgical procedure may require more robust and convincing attributes than the health promotion message advocated to BCNs in this study.

It is acknowledged within social psychology (Prochaska et al. 1992; Prochaska & DiClemente 1994) and health promotion (Tones & Tilford 2001) that developing a positive attitude toward a change in behaviour is essential to eventual adoption of the change. Encouraging changes, such as cessation of smoking, involves a complex interplay among the quality of information, previous personal experience, and the credibility of the health professional exists. Using the terminology of innovation diffusion theory, it follows that the change agent (the researcher), the message (evidence contained in the booklet) and the attitudes of the sample are all significant indicators for the proposed changes in practice.

Attitude

The findings indicate that the booklet affected slight changes in attitudes of the BCNs, which is congruent with the literature about relationships between attitude change and information exchange (Gass 1998; Howell et al. 1998; Messmer et al. 1998; White et al. 1998). Changing attitudes is difficult because they are derived from values that are often salient and deep-seated (Tones & Tilford 2001). In the context of research dissemination and utilisation in nursing, this remains a key point if the profession is to advance. In their systematic review of individual determinants of research utilisation, Estabrooks et al. (2003) found that most factors, for example, education, information seeking, involvement in research activities, did not show any consistent positive effect with the exception of beliefs and attitudes towards research. To change an attitude, communication must be highly persuasive and delivered by a credible source. Because no direct contact occurred between the respondents and researcher, it was not unexpected to observe that the intervention influenced only one attitude item when tested as a single-outcome variable. However, the finding is crucial in terms of clinical significance. An increase in agreement with the statement, I would promote exercise to my patients if I knew more about the associated benefits and limitations, combined the desire for knowledge with a desire to introduce an evidence-based change into practice. In this case the booklet was influential in increasing BCNs' awareness of the exemplar topic; a result that indicates a considerable accomplishment.

Strengths and Limitations

The information booklet was developed to be relevant and interesting to BCNs and to be a stimulus for changing practice. In addition, the best available research-based evidence was included. However, the content of the booklet was limited by the effect size of many of the studies available at the time that have included exercise programmes involving cancer patients. Extensive and robust research leading to unequivocal conclusions about the benefits of exercise for breast-cancer patients (e.g., Segal et al. 2001; Courneya et al. 2003) would have been preferable, but were not available at the time the booklet was written.

Ideally, all instruments used in the study would have had face, content, criterion, and construct validity and confirmed reliability. However, because no measurement tool existed that could be used to assess changes in knowledge, reported practice, and attitude for the promotion of exercise for breast-cancer patients, developing a questionnaire was necessary. The questionnaire helped to assess the effectiveness of the information booklet. Content and face validity was established via a systematic literature search and consensus of experts in related disciplines. Measurement of attitude was accomplished using a simplified approach, following techniques described by Oppenheim (1992).

Limitations of the instrument were that it could only be used to collect information about reported rather than actual practice and that BCNs in the experimental group had access to the booklet while they completed the questionnaire. The researcher did not control for the possibility that some respondents might have referred directly to the booklet while answering the knowledge questions, which could have affected outcome. The Hawthorne effect (Roethlisberger & Dickson 1939) may have also influenced some of the responses, as could the fact that the same questionnaire was used at both intervals. Participants in the control group might have acquired answers to some of the knowledge questions during the interval between the pre- and post-intervention assessments. Correspondingly, there is no way of knowing if the BCNs who received the information booklet actually read and absorbed the information. The marked increase in knowledge, reported change in practice, and response to attitude statements could have been affected by external influences. However, realistic maturation, historical, and motivational threats to internal validity were recognised during the design phase and were addressed by imposing the shortest reasonable time between the pre- and post-test assessments, having a control group, and further reducing bias through coded assessment forms. Participants were not able to influence whether they received a booklet or not and coded data forms allowed single blinding.

Further depth of knowledge could have been obtained within the positivist paradigm through the inclusion of additional outcome measures. It might have been worthwhile to determine how BCNs promoted exercise, duration of behaviour change, and patients' perspectives. Direct inquiries about documented changes in physical activity for patients and any measurable declines in reported patient problems could have been pursued. To some extent, aspects of the three phases (Medical Research Council 2000) of the standard clinical trial framework were achieved, but not as fully as they could have been: (1) preclinical (theoretical); (2) modelling; and (3) exploratory trial.

IMPLICATIONS AND CONCLUSIONS

  1. Top of page
  2. ABSTRACT
  3. BACKGROUND
  4. AIM OF STUDY
  5. METHODS
  6. RESULTS
  7. DISCUSSION
  8. IMPLICATIONS AND CONCLUSIONS
  9. References

The main conclusion is that in contrast to findings from the medical literature (Freemantle et al. 2001), it would appear that written materials can be effective in transmitting research evidence to specific groups of nurses. In addition to influencing changes in knowledge, as shown previously in nursing literature (Luker & Kendrick 1995; Williams et al. 1997), the booklet was also associated with changing reported practice and related attitudes.

The success of the information booklet was attributed to its suitability to the identified task through presenting a relevant research-based message targeted to meet the expressed needs, professional qualities, and social context of the audience. The booklet appeared to be well received by BCNs, increased their awareness of the benefits of exercise for their patients, and influenced reported changes in clinical practice.

This study showed not only a successful choice of format for disseminating “the message” but that nurse researchers can be effective change agents in the process. Direct access to the research evidence was ensured because the booklet was posted directly to BCNs' workplaces.

To further the clinical contribution of the current study, the benefits of exercise could be promoted strategically within the context of relieving cancer-related fatigue. This type of fatigue is known to be the most prevalent symptom for cancer patients (Richardson 1995) and also one that can be reduced by physical exercise in many circumstances (Mock et al. 2005). The current study indicated that BCNs would be more likely to promote exercise to their patients if they knew more about its associated benefits and limitations. Further publicizing the benefits of exercise for breast-cancer patients could be achieved through presentations to BCN networks and conferences as well as publication of articles in popular and academic journals. A consumer-oriented version of the information could be written to address interests of the general public, and Web-based links to authoritative and popular cancer information networks could be publicized.

The approach evaluated here can be replicated and applied to other groups of nurses. Similar booklets could be produced and distributed at relatively low cost.

References

  1. Top of page
  2. ABSTRACT
  3. BACKGROUND
  4. AIM OF STUDY
  5. METHODS
  6. RESULTS
  7. DISCUSSION
  8. IMPLICATIONS AND CONCLUSIONS
  9. References
  • Altman D.G. (1999). Practical statistics in medical research. London : Chapman and Hall.
  • Ajzen I. (1991). The Theory of Planned Behavior. Organisational Behavior in Human Decision Processes, 50, 179211.
  • Ajzen I. & Madden T.J. (1986). Prediction of goal directed behaviour: Attitudes, intentions and perceived behavioural control. Journal of Experimental Social Psychology, 22, 453474.
  • Bero L.A., Grilli R., Grimshaw J.M., Harvey E., Oxman A.D. & Thomson M.A. (1998). Getting research practice into practice. Closing the gap between research and practice: An overview of systematic reviews of interventions to promote the implementation of research findings. British Medical Journal, 317, 465468.
  • Campbell M.J. (2001). Statistics at square two. London : BMJ Books.
  • Campbell D.T. & Stanley J.C. (1963). Experimental and quasi-experimental designs for research. Boston : Houghton Mifflin.
  • Cancer Relief Macmillan Fund. (1996). Raising standards in breast cancer care. The Macmillan directory of breast cancer services in the U.K. (2nd ed.). London : Author.
  • Closs S.J., Baum G., Bryar R., Griffiths J. & Knight S. (2000). Barriers in research implementation in two Yorkshire hospitals. Clinical Effectiveness in Nursing, 4(1), 310.
  • Closs S.J. & Cheater F.M. (1994). Utilization of nursing research: Culture, interest and support. Journal of Advanced Nursing, 19, 762773.
  • Courneya K.S., Mackey J.R., Bell G.J., Jones L.W., Field C.J. & Fairey A.S. (2003). Randomised controlled trial of exercise training in postmenopausal breast cancer survivors: Cardiopulmonary and quality of life outcomes. Journal of Clinical Oncology, 21, 16601668.
  • Crane J. (1995). The future of research utilisation. Nursing Clinics of North America, 30(3), 565577.
  • Donner A. (1998). Some aspects of the design and analysis of cluster randomisation trials. Applied Statistics, 47(Part 1), 95113.
  • Donner A., Birkett N. & Buck C. (1981). Randomisation by cluster: Sample size requirements and analysis. American Journal of Epidemiology, 114(6), 906914.
  • Dunn V., Crichton N., Roe B., Seers K. & Williams K. (1998). Using research for practice: A U.K. experience of the BARRIERS Scale. Journal of Advanced Nursing, 26, 12031210.
  • Estabrooks C.A., Floyd J.A., Scott-Findlay S., O'Leary K.A. & Gushta M. (2003). Individual determinants of research utilisation: A systematic review. Journal of Advanced Nursing, 43(5), 506520.
  • Everitt B.S. (1995). The Cambridge dictionary of statistics in the medical sciences. Cambridge , UK : Cambridge University Press.
  • Farrell G.A. (2001). From tall poppies to squashed weeds: Why don't nurses pull together more. Journal of Advanced Nursing, 35(1), 2633.
  • Freemantle N., Harvey E.L., Wolf F., Grimshaw J.M., Grilli R. & Berro L.A. (2001). Printed educational materials to improve the behaviour of health professionals and patient outcomes (Cochrane Review). In The Cochrane Library, Issue 1. Oxford , UK : Update Software.
  • Gass J.P. (1998). The knowledge and attitudes of mental health nurses to electro-convulsive therapy. Journal of Advanced Nursing, 27(1), 8390.
  • Grimshaw J.M. & Russell I.T. (1993). Effect of clinical guidelines on medical practice: A systematic review of rigorous evaluations. Lancet, 342, 13171322.
  • Grimshaw J.M., Shirran L., Thomas R., Mowatt G., Fraser C., Bero L., Grilli R., Harvey E., Oxman A. & O'Brien M.A. (2001). Changing provider behaviour: An overview of systematic reviews of interventions. Medical Care, 39(Suppl. 2), 112145.
  • Harrison S., Downswell G. & Wright J. (2002). Practice nurses and clinical guidelines in a changing primary care context: An empirical study. Journal of Advanced Nursing, 39(3), 299207.
  • Howell D., Butler L., Vincent L., Watt-Watson J. & Stearns N. (2001). Influencing nurses' knowledge, attitudes and practice in cancer pain management. Cancer Nursing, 23(1), 5563.
  • Howell S.L., Nelson-Marten P., Krebs L.U., Kaszyk L. & Wold R. (1998). Promoting nurses' positive attitudes toward cancer prevention/screening. Journal of Cancer Education, 13, 7684.
  • Kajermo K.N., Nordstrom G., Krusebrant A. & Bjorvell H. (1998). Barriers and facilitators of research utilization as perceived by a group of registered nurses in Sweden. Journal of Advanced Nursing, 27, 798807.
  • Kanouse D.E. & Jacoby I. (1988). When does information change practitioners' behaviour? International Journal of Technology Assessment in Health Care, 4, 2733.
  • Kirshbaum M. (2005a). The case for promoting physical exercise in breast cancer care. Nursing Standard (Arts & Science), 19(41), 4148.
  • Kirshbaum M. (2005b). A conceptual framework for targeting research dissemination interventions to meet the needs of breast cancer patients in the United Kingdom. Oncology Nursing Forum, 32(1), 164.
  • Kirshbaum M. (2007). A review of the benefits of whole body exercise during and after treatment for breast cancer. Journal of Clinical Nursing, 6(1), 104121.
  • Kirshbaum M., Beaver K. & Luker K. (2004). Perspectives of breast care nurses on research dissemination and utilisation. Clinical Effectiveness in Nursing, 8, 4758.
  • Kitson A. (2004). The state of the art and science of evidence-based nursing in the U.K. and Europe. Worldviews on Evidence-Based Nursing, 1, 68.
  • Kneafsey R. (2000). The effect of occupational socialization on nurses' patient handling practices. Journal of Clinical Nursing, 9(4), 585593.
  • Lacey A. (1996). Facilitating research based practice by educational intervention. Nurse Education Today, 16, 296301.
  • Lock C.A., Kaner E.F.S., Heather N., McAvoy B.R. & Gilvarry E. (1999). A randomised trial of three marketing strategies to disseminate a screening and brief alcohol intervention programme to general practitioners. British Journal of General Practice, 49, 695698.
  • Luker K.A. & Kendrick M. (1995). Towards knowledge based practice: An evaluation of a method of dissemination. International Journal of Nursing Studies, 32(1), 5967.
  • Machin D. & Campbell M.J. (1987). Statistical tables for the design of clinical trials. Oxford , UK : Blackwell Scientific Publications.
  • MacGuire J.M. (1990). Putting nursing research findings into practice: Research utilisation as an aspect of the management of change. Journal of Advanced Nursing, 15, 614620.
  • Medical Research Council. (2000). A framework for development and evaluation of RCTs for complex interventions to improve health. MRC Health Services and Public Health Research Board. Retrieved June 4, 2004 from http://www.mrc.ac.uk/index/publications/publications-ethics_and_best_practice/publications-clinical_trials_guidelines.htm.
  • Messmer P.R., Jones S., Moore J., Taggart B., Parchment Y., Holloman F. & Quintero L.M. (1998). Knowledge, perceptions, and practice of nurses toward HIV+/AIDS patients diagnosed with tuberculosis. Journal of Continuing Education in Nursing, 29(3), 117125.
  • Mock V., Frangakis C., Davidson N.E., Ropka M.E. & Pickett M. (2005). Exercise manages fatigue during breast cancer treatment: A randomized controlled trial. Psycho-oncology, 14, 464477.
  • Mulhall A., Alexander C. & Le May A. (1998). Appraising the evidence for practice: What do nurses need? Journal of Clinical Effectiveness, 3(2), 5458.
  • National Health Service Centre for Reviews and Dissemination. (1994). Implementing clinical practice guidelines. Effective Health Care Bulletin, 1(8).
  • National Health Service Centre for Reviews and Dissemination. (1999). Getting evidence into practice. Effective Health Care Bulletin, 5(1).
  • National Health Service Executive. (1996). Promoting clinical effectiveness. A framework for action in and through the NHS. Leeds , UK : Author.
  • Oppenheim A.N. (1992). Questionnaire design, interviewing and attitude measurement. London : Continuum.
  • Oxman A.D., Thomson M.A., Davis D.A. & Haynes R.B. (1995). No magic bullets: A systematic review of 102 trials of interventions to improve professional practice. Canadian Medical Association Journal, 153, 14231431.
  • Parahoo K. (2000). Barriers to and facilitators of research utilisation among nurses in Northern Ireland. Journal of Advanced Nursing, 31, 8998.
  • Prochaska J.O. & DiClemente C.C. (1984). Transtheoretical approach: Traditional boundaries of therapy. San Francisco , CA : Dow Jones Irwin.
  • Prochaska J.O., DiClemente C.C. & Norcross J.C. (1992). In search of how people change: Applications to addictive behaviours. American Psychologist, 47(9), 11021114.
  • Pryjmachuk S. (1996). Pragmatism and change: Some implications for nurses, nurse managers and nursing. Journal of Nursing Management, 4, 201205.
  • Retsas A. & Nolan M. (1999). Barriers to nurses' use of research: An Australian hospital study. International Journal of Nursing Studies, 36, 335343.
  • Richardson A. (1995). Fatigue in cancer patients: A review of the literature. European Journal of Cancer Care, 4(1), 2032.
  • Rodgers S.E. (2000). A study of the utilisation of research in practice and the influence of education. Nurse Education Today, 20, 279287.
  • Roethlisberger F.J. & Dickson W.J. (1939). Management of the worker. Cambridge , MA : Harvard University Press.
  • Rogers E.M. (1995). Diffusion of innovations (4th ed.) New York : Free Press.
  • Rutledge D.N., Ropka M., Greene P.E., Nail L. & Mooney K.H. (1998). Barriers to research utilization for oncology staff nurses and nurse manager/CNS. Oncology Nursing Forum, 25(3), 497506.
  • STATA. (1997). STATA reference manual, release 5 (Vol. 3). College Station , TX : Author.
  • STATA. (2001a). STATA reference manual, release 7 (Vol. 3, pp 564591). College Station , TX : Author.
  • STATA. (2001b). STATA reference manual, release 7 (Vol. 2, pp 451-459). College Station , TX : Author.
  • Segal R., Evan W., Johnson D., Smith S., Colletta S., Gayton J., Woodward S., Wells G. & Reid R. (2001). Structured exercise improves physical functioning in women with Stages I and II breast cancer: Results of a randomised controlled trial. Journal of Clinical Oncology, 19, 657665.
  • Stetler C. (1994). Refinement of the Stetler/Marram Model for application for research findings to practice. Nursing Outlook, 42, 1525.
  • Stetler C.B., Bautista C., Vernale-Hannon C. & Foster J. (1995). Enhancing research utilization by clinical nurse specialists. Nursing Clinics of North America, 30(3), 457473.
  • Thomas L., Cullum N., McColl E., Rousseau N. & Steen N. (2003). Guidelines in professions allied to medicine (Cochrane Review). In The Cochrane Library, Issue 3. Oxford , UK : Update Software.
  • Thompson C., McCaughan D., Cullum N., Sheldon T.A., Mulhall A. & Thompson D.R. (2001). The accessibility of research-based knowledge for nurses in United Kingdom acute care settings. Journal of Advanced Nursing, 36(1), 1122.
  • Thomson M.A., Freemantle N., Wolf F., Davis D.A. & Oxman A.D. (1998). Educational meetings, workshops and preceptorships to improve the practice of health professionals and health care outcomes (Cochrane Review). The Cochrane Library Issue 1. Oxford , UK : Update Software.
  • Titler M.G., Mentes J.C., Rakel B.A., Abbott L. & Baumler S. (1999). From book to bedside: Putting evidence to use in the care of the elderly. Journal of Quality Improvement, 25(10), 545556.
  • Tolson D., Schofield I., Booth J., Kelly T. & James, L. (2006). Constructing a new approach to developing evidenced-based practice with nurses and older people. Worldviews on Evidence-Based Nursing, 3(2), 6272.
  • Tones K. & Tilford S. (2001). Health promotion: Effectiveness, efficiency and equity (3rd ed.). Cheltenham , UK : Nelson Thornes.
  • Waddell C. (2002). So much research evidence, so little dissemination and uptake: Mixing the useful with the pleasing. Evidence Based Nursing, 5, 3840.
  • Walsh M. (1997). How nurses perceive barriers to research implementation. Nursing Standard, 11(29), 3439.
  • Watts T., Jones M., Wainwright P. & Willams A. (2001). Methodologies analysing individual practice in health care: A systematic review. Journal of Advanced Nursing, 35(2), 238256.
  • White K., Eagle J., McNeil H., Dance S., Evans L.A., Harris H. & Reid M.J. (1998). What are the factors that influence learning in relation to nursing practice. Journal for Nurses in Staff Development, 14(3), 147153.
  • Whiting-O'Keefe Q.E., Henke C. & Simborg D.W. (1984). Choosing the correct unit of analysis in medical experiments. Medical Care, 22(12), 11011114.
  • Williams K.S., Crichton N.J. & Roe B. (1997). Disseminating research evidence: A controlled trial in continence care. Journal of Advanced Nursing, 25, 691698.
  • Wilson K. & Rose K. (1998). Patient recruitment and retention strategies in randomised controlled trials. Nurse Researcher, 6(1), 3546.