Cummings et al.

BMC Health Services Research 2010, 10:168

http://www.biomedcentral.com/1472-6963/10/168

RESEARCH ARTICLE Open Access

The relationship between characteristics of context

Research article

and research utilization in a pediatric setting

Greta G Cummings*1,2, Alison M Hutchinson2,3, Shannon D Scott1, Peter G Norton4 and Carole A Estabrooks1

Abstract
Background: Research utilization investigators have called for more focused examination of the influence of context
on research utilization behaviors. Yet, up until recently, lack of instrumentation to identify and quantify aspects of
organizational context that are integral to research use has significantly hampered these efforts. The Alberta Context
Tool (ACT) was developed to assess the relationships between organizational factors and research utilization by a
variety of healthcare professional groups. The purpose of this paper is to present findings from a pilot study using the
ACT to elicit pediatric and neonatal healthcare professionals' perceptions of the organizational context in which they
work and their use of research to inform practice. Specifically, we report on the relationship between dimensions of
context, founded on the Promoting Action on Research Implementation in Health Services (PARIHS) framework, and
self-reported research use behavior.
Methods: A cross-sectional survey approach was employed using a version of the ACT, modified specifically for
pediatric settings. The survey was administered to nurses working in three pediatric units in Alberta, Canada. Scores for
three dimensions of context (culture, leadership and evaluation) were used to categorize respondent data into one of
four context groups (high, moderately high, moderately low and low). We then examined the relationships between
nurses' self-reported research use and their perceived context.
Results: A 69% response rate was achieved. Statistically significant differences in nurses' perceptions of culture,
leadership and evaluation, and self-reported conceptual research use were found across the three units. Differences in
instrumental research use across the three groups of nurses by unit were not significant. Higher self-reported
instrumental and conceptual research use by all nurses in the sample was associated with more positive perceptions of
their context.
Conclusions: Overall, the results of this study lend support to the view that more positive contexts are associated with
higher reports of research use in practice. These findings have implications for organizational endeavors to promote
evidence-informed practice and maximize the quality of care. Importantly, these findings can be used to guide the
development of interventions to target modifiable characteristics of organizational context that are influential in
shaping research use behavior.

Background findings in healthcare settings remains slow and unpre-

In 2006 and 2007 respectively, special issues of the Jour- dictable, despite considerable investment of public funds
nal of Continuing Education in the Health Professions into healthcare research,[1] describing numerous theo-
(JCEHP) and Nursing Research were dedicated to advanc- retical approaches to how knowledge transfer could be
ing the theory and science of knowledge translation and supported [2]. The Nursing Research special issue
research utilization by the healthcare professions with the advanced the science of KT by using the Promoting
ultimate aim of improving health outcomes. The JCEHP Action Research in Health Services (PARiHS) framework
issue emphasized that the transfer and uptake of research [3] to frame the analysis of relationships between various
dimensions of organization context in hospitals and
* Correspondence: gretac@ualberta.ca
1 Faculty of Nursing, 3rd Floor, Clinical Sciences Building, University of Alberta,
nurses' reported research use [4]. In contexts character-
Alberta, AB, T6G 2G3, Canada ized by better leadership, empowering work environ-
Full list of author information is available at the end of the article ments (culture), and open feedback on performance
2010 Cummings et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Com-
mons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduc-
tion in any medium, provided the original work is properly cited.
Cummings et al. BMC Health Services Research 2010, 10:168 Page 2 of 10
http://www.biomedcentral.com/1472-6963/10/168

(evaluation), nurses reported significantly greater research utilization is the direct or concrete application
research use [5], which was related to better patient out- of research findings, often identified as the application of
comes (reduced adverse patient events) [6]. In summary, clinical practice guidelines, procedures and clinical pro-
research to date suggests the need for a parsimonious, tocols. Conceptual research utilization occurs when
valid and reliable measure of healthcare context, which research serves an 'enlightenment' function [12,13]: that
was met by the development and assessment of the is, practitioners' become aware of research findings, and
Alberta Context Tool (ACT) [7]. the findings inform, broaden or alter their thinking and
practice in indirect ways [11]. Persuasive or symbolic
Aim research utilization is when research findings are used as
The aim of this paper is to report findings from a pilot a tool to advocate for a certain procedure or practice.
study assessing the ACT, an instrument designed to mea- Generally speaking, overall research utilization can be
sure modifiable dimensions of organizational context and defined as the use of research findings in any and all
self-reported research utilization. Specifically, we aspects of one's work [14].
describe the relationship between three dimensions of Poor uptake of research in clinical practice has largely
context (culture, leadership and evaluation) as defined by been attributed to individual characteristics such as a
the Promoting Action on Research Implementation in practitioner's inability to understand research (a lack of
Health Services (PARiHS) framework[3], and research research skills and inadequate educational preparation),
utilization behavior in neonatal and pediatric settings. age, or attitude toward research [15-17]. Consequently,
the majority of the research in the field has centred on
Literature trying to understand individual level barriers and facilita-
Promoting Action on Research Implementation in Health tors to research utilization [18-22]. Much of the work in
Services nursing has been in this vein and has explored individual
Since its first publication in 1998, the PARiHS framework determinants of research use [23] with the predominance
[3] of research implementation has gained increasing of studies employing bivariate statistical approaches.
attention. In this framework, three constructs are consid- These approaches do not permit interactional or causal
ered essential for the successful implementation of exploration. Furthermore, syntheses of findings of studies
research into practice: evidence, facilitation, and context. on the individual determinants of research utilization to
Evidence includes both codified and non-codified sources date are equivocal, and the majority of those currently
of knowledge including research evidence, clinical exper- studied (e.g., age, years of clinical experience) are ques-
tise, local data or information, and patient experience. An tionably modifiable [23].
underlying assumption of the conceptualization of evi-
dence is that these different evidence forms are melded, A Shift to Exploring Organizational Factors
through negotiation and shared understanding, within a Previous work demonstrates that even if clinicians have
complex and multifaceted clinical environment [8]. Con- adequate and recent research-based knowledge they do
text is understood to be "the environment or setting in not automatically use it in practice [24]. There is growing
which the proposed change is to be implemented" [[3], p. consensus that the challenges in transferring research
150]. Context is proposed to have three dimensions (cul- into clinical practice are often more due to organizational
ture, leadership and evaluation). Early PARiHS develop- factors than to attributes of individual clinicians or the
ment work suggested that research implementation methods by which research findings are disseminated
would be facilitated by a value-oriented culture that is [16,18,25]. Given that the majority of healthcare profes-
receptive to change; clear, transformational leadership sionals work in complex organizational environments,
that supports teamwork and staff involvement in decision redirecting research efforts toward understanding the
making; and multiple methods of evaluation at various influence of organizational context is warranted. Organi-
levels [9]. The final element, facilitation, involves "provid- zational context is believed to shape the utilization of
ing help and support to achieve a specific goal to enabling research in practice [26] through its influence on individ-
individuals and teams to analyze, reflect and change their ual and group behavioural norms and innovation.
own attitudes, behaviors and ways of working" [[10], p. Despite organizational context being consistently iden-
580]. tified as an important factor influencing nurses' use of
research, it has not been well studied [23,27]. Kitson and
Research Utilization colleagues [3,9,28] have begun to investigate more thor-
Research utilization, a specific form of knowledge utiliza- oughly the importance of organizational features (e.g.,
tion, is a complex process in which research findings are culture, leadership) in influencing research use. Despite a
transformed from one or more studies into instrumental, lack of research on how organizational context or work
conceptual or persuasive utilization [11]. Instrumental environment influences research utilization, some recent
Cummings et al. BMC Health Services Research 2010, 10:168 Page 3 of 10
http://www.biomedcentral.com/1472-6963/10/168

investigations have begun to shed light on the nature of Understanding the patterns of research utilization in
and processes inherent in organizational context. Using child health contexts is an important first step to help
qualitative methods Scott and colleagues [29] found that address these gaps.
uncertainty within an organizational context (e.g., incon-
sistent management) significantly hindered nurses' use of Methods
research in practice. Investigators have successfully iden- We used a cross-sectional survey design to elicit health
tified contextual features that influence research utiliza- professionals' opinions about aspects of the organiza-
tion and more sophisticated analytic work including the tional context in which they work and the extent to which
development and testing of models to demonstrate how they use research evidence to inform their practice.
these features work and interact, have started to emerge.
For instance, in a study underpinned by the PARIHS Sample and setting
framework, Wallin and colleagues [30] used multivariate Nurses, managers, clinical specialists, doctors and allied
procedures to derive a score of nurses' research utiliza- healthcare professionals working in one of three pediatric
tion and then demonstrate that degrees of context from units (one pediatric inpatient unit and two pediatric/neo-
low to high were significantly related to increasing natal critical care units) located in two university affili-
research utilization scores. Subsequently, Cummings and ated hospitals in Alberta, Canada were invited to
colleagues [6] used structural equation modeling to test a participate in this study. For this study we analyzed data
theoretical model of relationships among features of only from the nursing sample.
organizational context (e.g. responsive administration,
Measures
relational capital, and hospital size), nurses' research utili-
The Alberta Context Tool (ACT), conceptually framed by
zation scores and adverse patient events. They found that
the PARiHS framework, was used to collect the data. The
these organizational characteristics interacted with better
index version of the ACT was developed for use in acute
leadership, culture, and evaluation, to lead to reports of
care settings and comprises a suite of survey instruments
greater research use by nurses, which then led to fewer
designed to assess modifiable characteristics of organiza-
adverse patient events. These findings suggest that strate-
tional context and self-reported research use [41]. The
gies to improve dimensions of organizational context
refined ACT consists of 56 items reflecting the following
could potentially increase research utilization behaviors.
eight contextual dimensions: culture, leadership, evalua-
The haphazard nature of research utilization is frequently
tion, social capital, informal interactions, formal interac-
attributed to the unique contexts in which research is
tions, structural and electronic resources, and
implemented. Further sophisticated analytic work is
organizational slack (representing three sub-concepts -
needed in order to make theoretical advancements and to
time, space, human resources). Each dimension is mea-
identify contextual predictors of research use.
sured by a separate scale or set of items. For the culture,
Research Utilization in Child Health leadership and evaluation scales, each item in the instru-
Child healthcare settings are not immune to the chal- ment used a five-point Likert-type scale (1 = strongly dis-
lenges of applying the best available research evidence in agree, 2 = disagree, 3 = neither agree nor disagree, 4 =
clinical practice. Thus, ensuring research is used to agree, 5 = strongly agree). In this study, we used data
inform clinical practice is of central importance. Effective from the three PARiHS contextual dimensions (culture,
research utilization is fundamental to ensuring that the leadership and evaluation) due to the smaller sample size
best available research evidence informs the health and obtained in this pilot study and to replicate Cummings et
healthcare of infants, children, youth and families. al's [6] approach to differentiating contexts from high to
Research shows that provision of research-informed care low. Additionally we hypothesized that there were posi-
to children not only improves health outcomes, but also tive relationships between the combined core contextual
reduces healthcare utilization [31-35]. Yet, studies of factors of culture, leadership and evaluation and nurses'
adult populations in the USA and the Netherlands sug- research utilization.
gest that 30-40% of patients do not receive care that is Finally, measures of instrumental and conceptual
well supported by scientific evidence, and 20-25% of their research utilization, the dependent variables in this study
care is either not needed or potentially harmful [36,37]. were developed and validated by Estabrooks [14,30,42,43]
Similar findings have been demonstrated in Canadian and used in conjunction with the ACT in this pilot study.
child health research [38]. Four single items (not combined to form an index) mea-
To date, information about research utilization among sured four kinds of research use: instrumental, concep-
child healthcare professionals [39] is lacking. In fact there tual, persuasive, and overall. Each item was preceded by a
is growing recognition that health services research in definition and examples of that kind of research use. For
child health has been largely under-represented [40]. each kind of research use, respondents were asked to
Cummings et al. BMC Health Services Research 2010, 10:168 Page 4 of 10
http://www.biomedcentral.com/1472-6963/10/168

indicate how often they used research in this way on their On a weekly basis, we distributed colorful posters to
last typical workday. Items were scored on a 5-point scale report graphically the cumulative response rate for each
(from 10% or less to almost 100%). In this study we used unit. This strategy increased the study profile and
conceptual and instrumental research utilization as our prompted potential respondents to complete the survey.
dependent variables due to the small sample and because
nurses have identified that these are more directly linked Data collection
to their practice [44]. This study was undertaken in 2007. Data were collected
The ACT was pilot tested in adult acute care settings using paper-based and electronic survey methods. Uni-
and factor analytic procedures demonstrated that almost versity of Alberta Population Research Laboratory (PRL)
70% of the variance in the context construct was http://www.uofaweb.ualberta.ca/prl/index.cfm was con-
accounted for in a sample of nurses, clinical specialists, tracted to prepare and administer the electronic and
managers, doctors and allied healthcare professionals paper-based surveys. All healthcare professionals who
working in four major teaching hospitals [41]. The instru- were eligible to complete the survey were provided with a
ment was further refined based on psychometric and personalized survey package. The package contained a
bivariate analysis of these data. The resulting instrument letter to introduce the study, a business card providing
contained 56 items and was completed in approximately the Uniform Resource Locator (URL) for the survey and a
9 minutes. Principal Components Analysis produced a unique password, information detailing how to complete
13-factor solution that accounted for 59.26% of the vari- the survey online, a signed continuing education certifi-
ance in perceptions of organizational context. Acceptable cate, and a five-dollar gift certificate. PRL assembled the
internal reliability for the culture, leadership and evalua- packages, which were either hand delivered to each eligi-
tion dimensions of organizational context was found with ble staff member on the three units or placed in their
Cronbach's alpha scores of .72, .91 and .91 respectively work setting mailboxes. Consistent with a modified Dill-
[7]. Additional detail on group alphas and other psycho- man [46] approach, all eligible participants were sent a
metrics of the ACT are described elsewhere [45]. postcard reminder two weeks and four weeks after com-
For the current study some adaptation was made to the mencement of data collection. The non responding
ACT to ensure the language and content were appropri- nurses in the sample were also sent a paper copy of the
ate to the pediatric setting. Specifically, the word patient survey with this final reminder.
was substituted with patients and families; the examples
Analysis
which preceded questions were revised to ensure rele-
Drawing on the approach used by Wallin et al. [30] and
vance to the pediatric setting. The culture, leadership and
Cummings et al. [6], we used respondents' scores on the
evaluation dimensions of organizational context were
three context dimensions (culture, leadership and evalua-
again found to be reliable with Cronbach's alpha scores of
tion) to sort data into four conceptually distinct contex-
.71, .90 and .87 respectively [45].
tual groups (high, moderately high, moderately low, and
Ethical approval low) within their unit. To preserve the sample size for this
Ethical approval to conduct the study was received from pilot study, mean scale scores were calculated for all indi-
the University of Alberta Health Research Ethics Board. viduals who responded to at least two items in each scale.
Operational and administrative approval to conduct the If respondents only answered one question in a scale, that
study was also attained. The surveys were completed score was used. This allowing us to retain another 83
anonymously and confidentiality of the data was main- cases (23%) of the respondents, as 275 respondents had
tained. no missing data on the three context scales. Then analysis
of variance was used to compare the culture, leadership,
Recruitment evaluation and research use scores across the three units.
We used a modified Dillman [46] approach to recruit sur- In order to sort nurses into the four contextual groups,
vey respondents (reported in detail elsewhere [47]). The we needed to effectively divide the number of respon-
recruitment procedure was initiated four months before dents into high and low on each dimension and therefore
the data collection commenced, and involved a series of required an appropriate cut point between high and low
formal presentations and informal (one-on-one) interac- mean scores on each of the 5-point leadership, culture
tions with staff to familiarize them with study aims and and evaluation scales. In our earlier work [6,30] the 4-
processes of data collection. In particular, the presenta- point Likert scale from Strongly Agree to Strongly Dis-
tions included information about use of the Internet to agree made the choice of cut point simple - nurses who
access and complete the instrument. Additionally, elec- had reported Strongly Agree and Agree for a contextual
tronic and print materials were used to promote aware- dimension were categorized as high, and those reporting
ness of and communicate information about the study. Strongly Disagree and Disagree were categorized as low.
Cummings et al. BMC Health Services Research 2010, 10:168 Page 5 of 10
http://www.biomedcentral.com/1472-6963/10/168

In our current study, the 5-point Likert included a 3 = found between groups for perceptions of culture, leader-
neither agree nor disagree between 4 = Agree and 2 = ship and evaluation, and for self-reported conceptual
Disagree. Therefore we examined three possible cut research use, while no significant differences were found
points for nurses' individual means on the contextual in instrumental research use between the groups. On
dimensions; 3.0, 3.5, and 3.9. Both 3.0 and 3.9 provided average, nurses working on Unit A were the most positive
skewed groupings where insufficient numbers of nurses group in their perceptions of culture and leadership.
were found in low groups for 3.0 and in high groups for Nurses working on Unit B were the most positive in their
3.9. On this basis we selected 3.5 as the cut-off score to perceptions of evaluation.
categorize contextual dimensions as 'high' (an individual Table 3 presents mean scores and standard deviations
respondent's mean score > 3.5) or 'low' (an individual for instrumental and conceptual research use for all
respondent's mean score 3.5). Respondents' data were nurses in the sample, by unit, and by category of context.
grouped using their mean scores of the three contextual Higher self-reported instrumental and conceptual
scales (culture, leadership and evaluation) to create four research use by all nurses in the sample was associated
distinct groups. Hence, those who reported a high score with more positive perceptions of organizational context.
on all three organizational context scales were catego- For Unit B and when data for nurses from all three units
rized as working in a high context. Those who reported a were combined, self-reported instrumental use of
high score on any two context variables and a low score research is consistently higher in relation to more positive
on any single variable were categorized as working in a perceptions of organizational context. Whereas nurses on
moderately high context. Those who scored high on any Unit A who perceived their context as 'high' reported
single variable and low on the two remaining variables slightly lower instrumental research use than those who
were categorized as working in a moderately low context. reported context as 'moderately high'. Nurses on Unit C,
Finally, those who scored low on all three context vari- who perceived the organizational context as 'moderately
ables were categorized as working in a low context. low', reported higher instrumental research use when
Grouping the respondents in this manner enabled us to compared with their colleagues who perceived their con-
examine their self-reported research use behaviors in text as 'moderately high'. However, nurses who perceived
relation to their perceptions of organizational context. the organizational context as 'high' clearly reported the
highest instrumental use of research for Unit C.
Results The pattern for conceptual research use and organiza-
The instrument was offered to 362 nurses, including reg- tional context categories was slightly different. While
istered, graduate and licensed practical nurses. A nurses on Units A and C reported steadily higher concep-
response rate of 69% was achieved for a total sample size tual use of research in conjunction with more positive
of 248. Internal reliability for the culture, leadership and perceptions of the organizational context; conceptual
evaluation scales was acceptable with Cronbach's alpha research use reported by nurses on Unit B, and for all
scores of .688, .898, .849 respectively. nurses combined, did not follow this pattern. For all
The demographic characteristics of the nurses from nurses combined, those reporting the organizational con-
each of the three units are reported in Table 1. Approxi- text as 'low' reported fractionally higher conceptual use of
mately 96% of the nurse respondents were female. The research than did nurses perceiving the organizational
highest level of education reported was a Master's degree context as 'moderately low'. On Unit B, nurses who per-
(N = 1, 0.4% of the total sample); the majority of nurse ceived the organizational context as 'low' reported higher
respondents (N = 154, 61.4% of the total sample) reported conceptual research use than did their colleagues who
having obtained a Bachelor's degree. Across the units the perceived the organizational context as 'moderately low'
proportion of nurses who reported having a Diploma/ or 'moderately high'.
certificate was approximately equal. Similarly, the pro- The appropriateness of the unit of analysis at individual
portion of nurses who possessed a Bachelor's degree was or unit levels for research use was assessed using inter-
similar in each of the units. The mean number of years class correlations[48]. Table 4 depicts the ICC(1) scores
the nurses were employed in their current position was which indicate that instrumental research use is best ana-
10.1 (SD 9.6). For units B and C the mean number of lyzed at the individual level (all ICC(1) scores being less
years in the current position was 11.03 (SD 9.06) and than .10). Two ICC(1) scores for conceptual research use
11.58 (SD 10.27), respectively. This contrasted with an (CRU) in moderately high and low contexts suggest that
average of 5.79 (SD 7.5) years of work in their current CRU could be analyzed at the unit level, yet the scores for
position reported by nurses in Unit A. high and low contexts suggest individual unit of analysis
Table 2 presents the results of analysis of variance of is most appropriate. Our earlier analyses of ICC of the
mean unit scores for culture, leadership, evaluation and ACT contextual dimensions [41,45] support their aggre-
research use. Statistically significant differences were gation to the unit level.
Cummings et al. BMC Health Services Research 2010, 10:168 Page 6 of 10
http://www.biomedcentral.com/1472-6963/10/168

Table 1: Sample Demographics

Unit A Unit B Unit C Total (%)

N (%) 55 (22.2) 68 (27.4) 125 (50.4) 248

Gender Male 4 (7.3) 3 (4.4) 0 7 (2.8)
[N, (%)]
Female 51 (92.7) 63 (92.6) 124 (99.2) 238 (95.9)
Missing Values 0 2 (2.9) 1 (0.8) 3 (1.2)
Education Diploma/Certificate 26 (47.3) 32 (47.0) 62 (49.6) 120 (48.4)
[N, (%)]*
Bachelors Degree 35 (63.6) 43 (63.2) 76 (60.8) 154 (61.4)
Masters Degree 0 0 1 (0.8) 1 (0.4)
Missing Values 0 0 0 0
Years in current position 5.79 (7.5) 11.03 (9.06) 11.58 (10.27) 10.1 (9.6)
[mean (SD)]
* Respondents were asked to identify all that apply, and therefore the numbers are larger than the study n.

Figure 1 illustrates for each unit, the patterns described Discussion

above for nurses' self-reported instrumental and concep- These data illustrate a positive trend in the relationships
tual research use along with the 95% confidence interval between organizational context and research utilization.
according to the four context categories. For both instru- That is, more positive perceptions of organizational con-
mental and conceptual research use, a general pattern of text were generally associated with higher self-reported
increasing degrees of research use was associated with instrumental and conceptual research use. We have orga-
more positive perceptions of context across all three nized our discussion around three key points: 1) the
units. The highest levels were associated with high per- nature of the relationship between instrumental research
ceptions of context across all units with the exception use and organizational context, 2) new insights into the
instrumental research use in Unit A as previously potential appropriate 'level' of analysis (individual or unit)
described. of the concepts studied (research use, culture, leadership
and evaluation), and 3) the establishment of a method-

Table 2: Contextual Dimensions and Research Use Descriptives

Mean (SD) ANOVA

(N)

All Nurses Unit A Unit B Unit C F-Statistic p-value

Culture 3.49 (0.65) 3.88 3.25 3.46 16.484 .000

(N = 248) (.54) (.69) (.61)
(N = 55) (N = 68) (N = 125)
Leadership 3.51 (0.79) 3.97 3.22 3.46 15.767 .000
(N = 248) (.80) (.81) (.68)
(N = 55) (N = 68) (N = 125)
Evaluation 3.31 (0.72) 3.12 3.61 3.23 9.410 .000
(N = 246) (.80) (.64) (N = 68) (.68) (N = 125)
(N = 53)
Instrumental Research Use 3.28 (1.35) 2.94 (1.32) 3.30 3.45 (1.34) 2.510 .084
(N = 225) (N = 54) (1.36) (N = 105)
(N = 66)
Conceptual Research Use 3.55 (1.26) 3.36 (1.21) 3.03 3.91 (1.10) 12.227 .000
(N = 241) (N = 53) (1.38) (N = 122)
(N = 66)
Cummings et al. BMC Health Services Research 2010, 10:168 Page 7 of 10
http://www.biomedcentral.com/1472-6963/10/168

Table 3: Instrumental and conceptual research use by context category for each unit

Context All Nurses Unit A Unit B Unit C

[mean (SD)] [mean (SD)] [mean (SD)] [mean (SD)]
(N) (N) (N) (N)

Instrumental Conceptual Instrumental Conceptual Instrumental Conceptual Instrumental Conceptual

research use research use research use research use research use research use research use research use
High 3.69 (1.26) 4.12 (.79) 3.07 (1.27) 4.07 (.73) 3.80 (1.48) 3.90 (.88) 4.00 (1.06) 4.22 (.80)
(N = 48) (N = 51) (N = 14) (N = 14) (N = 10) (N = 10) (N = 24) (N = 27)
Moderately high 3.39 (1.32) 3.45 (1.26) 3.12 (1.31) 3.24 (1.23) 3.65 (1.04) 2.95 (1.05) 3.46 (1.53) 4.04 (1.25)
(N = 70) (N = 71) (N = 26) (N = 25) (N = 20) (N = 20) (N = 24) (N = 26)
Moderately low 3.29 (1.30) 3.36 (1.42) 2.75 (.96) 3.00 (1.15) 3.09 (1.35) 2.67 (1.52) 3.52 (1.29) 3.84 (1.20)
(N = 58) (N = 66) (N = 4) (N = 4) (N = 23) (N = 24) (N = 31) (N = 38)
Low 2.72 (1.38) 3.37 (1.30) 2.25 (1.39) 2.75 (1.49) 2.77 (1.59) 3.17 (1.70) 2.85 (1.29) 3.61 (1.02)
(N = 47) (N = 51) (N = 8) (N = 8) (N = 13) (N = 12) (N = 26) (N = 31)

ological 'cut off' point for distinguishing between 'low' these findings also add some level of justification to the
and 'high' contexts. derivation procedures for the outcome variable (research
First, our findings are consistent with Wallin et al. [30] utilization) in those earlier studies.
and Cummings et al. [6] who studied nurses from adult Our findings provide further empirical support for the
acute and general hospitals only, in that, nurses who context dimension of the PARiHS model [3,9,28]. Specifi-
worked in high levels of contexts reported higher levels of cally, both instrumental and conceptual research utiliza-
instrumental research use. Thus, these findings support tion scores were higher in association with more positive
our initial claims that in pediatric environments, context contextual conditions (culture, leadership and evalua-
shapes research utilization behaviors. In sum, nurses tion). Yet, these relationships are stronger for instrumen-
working in contexts characterized by strong leadership, tal research use. Although caution must be used in light
positive feedback and culture reported more instrumen- of the relatively small sample size, we believe based on
tal and conceptual research utilization than nurses work- this observation in this pilot study that the influence of
ing in less positive contexts. This study also adds to the culture, leadership and evaluation may align more closely
work of Wallin et al. [30] and Cummings et al. [6] by with instrumental research use. It is possible that the
using specific questions to directly measure instrumental questions contained within these scales are more congru-
and conceptual research utilization, in contrast to a ent with the act of 'doing' and, therefore, with instrumen-
derived research utilization score. Thus to some degree tal use of research. As well, these findings imply that

Table 4: Aggregation at Unit level by Context category

Context Category Variable F BMS WMS ICC1 ICC2 eta_2 Omega_2 p-value

Low IRU 0.5712 1.1060 1.9362 0.0000 0.0000 0.0253 0.0000 0.5690

CRU 1.6505 2.7000 1.6359 0.0425 0.3941 0.0643 0.0249 0.2027

Moderately Low IRU 1.8964 3.2092 1.6923 0.0446 0.4727 0.0634 0.0295 0.1596

CRU 5.8959 10.2428 1.7373 0.2032 0.8304 0.1556 0.1275 0.0045

Moderately High IRU 0.9789 1.7118 1.7487 0.0000 0.0000 0.0284 0.0000 0.3810

CRU 5.3239 7.5530 1.4187 0.1536 0.8122 0.1354 0.1086 0.0071

High IRU 2.6325 3.8920 1.4784 0.0929 0.6201 0.1047 0.0637 0.0829

CRU 0.6287 0.3994 0.6353 0.0000 0.0000 0.0255 0.0000 0.5376

Cummings et al. BMC Health Services Research 2010, 10:168 Page 8 of 10
http://www.biomedcentral.com/1472-6963/10/168

Figure 1 Nurses' instrumental and conceptual research use according to context category, by unit.

through concerted efforts to improve culture, leadership able action associated with the application of research
and evaluation, higher degrees of research utilization may findings. In today's healthcare environments, doing is val-
be achievable. Instrumental research utilization directly ued and nurses are expected to complete tasks at work
relates to "doing" things, that is, providing direct, observ- [26,49]. As a result, it makes sense that acute care pediat-
Cummings et al. BMC Health Services Research 2010, 10:168 Page 9 of 10
http://www.biomedcentral.com/1472-6963/10/168

ric nurses would see more direct relevance, in terms of tional endeavors to promote evidence-informed practice
their work, for instrumental research utilization. Addi- and maximize the quality of care. Importantly, these find-
tionally, nurses may more readily recall and report ings can be used to guide the development of interven-
instances of instrumental research use than other more tions to target modifiable characteristics of child health
covert forms of research use. organizational contexts that are influential in shaping
Second, our findings using the three contextual dimen- research use behavior.
sions aggregated to the unit level suggest that they behave
Competing interests
appropriately as unit level variables - both by the near lin- The authors declare that they have no competing interests.
ear trends in relationships with research use and the con-
sistently small standard deviations across units (Table 2). Authors' contributions
GGC led the design and interpretation of statistical analysis and development
Our aggregation analyses reported in Table 4 and the rel- of the manuscript. AH contributed to the interpretation of statistical analysis
atively larger standard deviations of both instrumental and manuscript development.
and conceptual research use across individual nurses SS contributed to the design and conduct of the study as the Alberta site
investigator, and through interpretation of statistical analysis and manuscript
within a unit, seen in Table 2 suggest that these variables development. PN contributed to manuscript development. CAE is the Alberta
are appropriately analyzed at the individual level. Con- lead investigator and as such secured resources, led the design and conduct of
ceptually, this assertion aligns with Estabrooks' research the study and provided guidance on manuscript development. All authors
approved the final manuscript.
[50]; that is, research utilization behaviors are individual
level phenomena that can then be shaped by the context Acknowledgements
in which one works. However, these findings suggest that "The Alberta Context Tool was developed as part of an Alberta-based study
funded by the Alberta Heritage Foundation for Medical Research (AHFMR) (Co-
more targeted methodological work needs to be com-
Principal Investigators: CAE and PN). The authors thank Dr. Dwight Harley for
pleted in order to determine whether research utilization his input into psychometric testing. We acknowledge the contributions of Lin-
scores from individual participants can be 'averaged' glong Kong and Sunghyun Kang for their statistical advice and analyses in this
across groups of individuals working in a similar context project. Dr. Cummings is supported by a New Investigator Award, Canadian
Institutes of Health Research, and a Population Health Investigator Award,
and still remain meaningful.
Alberta Heritage Foundation for Medical Research. At the time this study was
Third, the results of this pilot study lay the foundation conducted, Dr Hutchinson was a Postdoctoral Fellow with the Faculty of Nurs-
for future studies. Literature on dichotomizing continu- ing and the Knowledge Utilization Studies Program at the University of Alberta,
ous variables suggests the importance of establishing cut Edmonton, Canada. Dr Hutchinson was supported by CIHR and AHFMR Fellow-
ships. Dr Estabrooks is supported by a Canada Research Chair in Knowledge
points because prior work can be invalidated if cut points Translation. The authors also acknowledge the CIHR Team in Children's Pain for
change [51]. We established a methodological 'cut off' contributions to this study. Funding was provided by the Canadian Institutes of
point for distinguishing between 'low' and 'high' contexts. Health Research (CTP-79854 and MOP-86605).
This pilot work sets the foundation for future analyses in
Author Details
a variety of settings and professional groups, each of 1Faculty of Nursing, 3rd Floor, Clinical Sciences Building, University of Alberta,

which will provide opportunities for validation and reli- Alberta, AB, T6G 2G3, Canada, 2School of Nursing and Midwifery, Faculty of
ability procedures. We did not want to use the median Health Medicine Nursing & Behavioural Sciences, Deakin University, Australia,
3Cabrini-Deakin Centre for Nursing Research, Cabrini Institute, Cabrini Health,
score on each contextual dimension as it suggests relativ- Melbourne, Victoria, Australia and 4Department of Family Medicine, University
ity in scores across studies, disciplines and settings. of Calgary, UCMC North Hill, 1707, 1632 - 14 Avenue NW, Calgary, AB T2N 1M7,
Establishing an appropriate cut point for analyzing high Canada
and low aspects of contextual dimensions is an important Received: 7 December 2009 Accepted: 16 June 2010
study outcome [52]. Published: 16 June 2010
BMC

This
2010Health
is
article
an isServices
Cummings
Open etResearch
available
Access al;from:
article
licensee
http://www.biomedcentral.com/1472-6963/10/168
2010,
distributed
BioMed
10:168under
Centralthe
Ltd.
terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

References
Limitations 1. Straus SE, Graham ID, Mazmanian PE: Knowledge translation: resolving
The sample is drawn from university hospitals only and the confusion. The Journal of continuing education in the health
these hospitals are linked operationally. The sample professions 2006, 26(1):3-4.
2. Estabrooks CA, Thompson DS, Lovely JJ, Hofmeyer A: A guide to
includes healthcare professionals from pediatric intensive knowledge translation theory. J Contin Educ Health Prof 2006,
care units and therefore the results cannot be generalized 26(1):25-36.
beyond this setting. This was a pilot study hence the sam- 3. Kitson A, Harvey G, McCormack B: Enabling the implementation of
evidence based practice: A conceptual framework. Quality & Safety in
ple size is small and this has restricted the degree of anal- Health Care 1998, 13:149-158.
ysis that could be undertaken. 4. Estabrooks CA: Prologue: A program of research in knowledge
translation. Nursing Research 2007, 56(4 Supplement 1):S4-S6.
5. Adewale AJ, Hayduk L, Estabrooks CA, Cummings GG, Midodzi WK,
Conclusion Derksen L: Understanding hierarchical linear models: Applications in
Overall, the results of this study lend support to the claim nursing research. Nurs Res 2007, 56(4 Supplement 1):S40-46.
that more positive organizational contexts influence 6. Cummings GG, Estabrooks CA, Midodzi WK, Wallin L, Hayduk L: Influence
of organizational characteristics and context on research utilization.
greater use of research in practice in child health settings. Nurs Res 2007, 56(4 Suppl):S24-39.
These findings have important implications for organiza-
Cummings et al. BMC Health Services Research 2010, 10:168 Page 10 of 10
http://www.biomedcentral.com/1472-6963/10/168

7. Estabrooks CA, Squires JE, Cummings GG, Birdsell JM, Norton PG: 33. Rowe BH, Bretzlaff JA, Bourdon C, Bota GW, Camargo CA Jr: Magnesium
Development and assessment of the Alberta Context Tool. BMC Health sulfate for treating exacerbations of acute asthma in the emergency
Serv Res 2009, 15(9):234. department. Cochrane Database of Systematic Reviews (Online)
8. Kitson AL, Rycroft-Malone J, Harvey G, McCormack B, Seers K, Titchen A: 2000:CD001490.
Evaluating the successful implementation of evidence into practice 34. Rowe BH, Spooner CH, Ducharme FM, Bretzlaff JA, Bota GW:
using the PARiHS framework: Theoretical and practical challenges. Corticosteroids for preventing relapse following acute exacerbations
Implement Sci 2008, 3(1):. of asthma. Cochrane Database of Systematic Reviews (Online)
9. McCormack B, Kitson A, Harvey G, Rycroft-Malone J, Titchen A, Seers K: 2001:CD000195.
Getting evidence into practice: The meaning of 'context'. J Adv Nurs 35. Russell K, Wiebe N, Saenz A, Ausejo SM, Johnson D, Hartling L, Klassen TP:
2002, 38(1):94-104. Glucocorticoids for croup. Cochrane Database of Systematic Reviews
10. Harvey G, Loftus-Hills A, Rycroft-Malone J, Titchen A, Kitson A, McCormack (Online) 2004:CD001955.
B, Seers K: Getting evidence into practice: The role and function of 36. Schuster M, McGlynn E, Brook R: How good is the quality of health care
facilitation. J Adv Nurs 2002, 37(6):577-588. in the United States? Milbank Q 1998, 76(4):517-563.
11. Estabrooks CA: Will evidence-based nursing practice make practice 37. Grol R: Successes and failures in the implementation of evidence-based
perfect? Canadian Journal of Nursing Research 1998, 30(1):15-36. guidelines for clinical practice. Medical Care 2001, 39(8 Supplement
12. Hasenfeld Y, Rino P: The utilization of research in administrative 2):II-46-II-54.
practice. In Research Utilization in the Social Services Edited by: Grasso AJ, 38. Johnson DW, Williamson J: Management of croup: Variations in practice
Epstein I. New York: Haworth Press; 1992:221-239. among 21 Alberta hospitals. Pediatr Res 2004:24a.
13. Weiss C: The many meanings of research utilization. Public 39. McCleary L, Brown GT: Barriers to paediatric nurses' research utilization.
Administration Review 1979, 39(5):426-431. Journal of Advanced Nursing 2003, 42(4):364-72.
14. Estabrooks CA: The conceptual structure of research utilization. 40. Lohr K, Dougherty D, L S: Methodological Challenges in health services
Research in Nursing & Health 1999, 22(3):203-16. research in the pediatric population. Ambul Pediatr 2001, 1(1):36-38.
15. Lacey EA: Research utilization in nursing practice--a pilot study. Journal 41. Estabrooks CA, Squires JE, Adachi AM, Kong L, Norton PG: Utilization of
of Advanced Nursing 1994, 19(5):987-995. Health Research in Acute Care Settings in Alberta Technical Report.
16. Pettengill MM, Gillies DA, Clark CC: Factors encouraging and Edmonton: Faculty of Nursing, University of Alberta; 2008.
discouraging the use of nursing research findings. Image - the Journal 42. Estabrooks CA, Scott S, Squires JE, Stevens B, O'Brien-Pallas L, Watt-Watson
of Nursing Scholarship 1994, 26(2):143-147. J, Profetto-McGrath J, McGilton K, Golden-Biddle K, Lander J, et al.:
17. Rodgers S: An exploratory study of research utilization by nurses in Patterns of research utilization on patient care units. Implement Sci
general medical and surgical wards. Journal of Advanced Nursing 1994, 2008, 3:31.
20:904-911. 43. Estabrooks CA, Kenny DJ, Cummings GG, Adewale AJ, Mallidou AA: A
18. Brett JLL: Organizational integrative mechanisms and adoption of comparison of research utilization among nurses working in Canadian
innovations by nurses. Nursing Research 1989, 38(2):105-110. civilian and United States Army healthcare settings. Res Nurs Health
19. Brett JLL: Use of nursing practice research findings. Nursing Research 2007, 30(3):282-296.
1987, 36(6):344-349. 44. Squires JE, Adachi AM, Estabrooks C: Developing a Valid and Reliable
20. Champion VL, Leach A: Variables related to research utilization in Measure of Research Utilization. Edmonton: Faculty of Nursing,
nursing: An empirical investigation. J Adv Nurs 1989, 14(9):705-710. University of Alberta; 2008.
21. Kirchhoff KT: A diffusion survey of coronary precautions. Nursing 45. Hutchinson AM, Adachi A-M, Kong L, Estabrooks CA, Stevens B: Context
Research 1982, 31(4):196-201. and Research Use in the Care of Children: A Pilot Study. Edmonton:
22. Varcoe C, Hilton A: Factors affecting acute-care nurses' use of research Faculty of Nursing, University of Alberta; 2008.
findings. Can J Nurs Res 1995, 27(4):51-71. 46. Dillman DA: Mail and internet surveys: The tailored design method.
23. Estabrooks CA: Translating research into practice: Implications for 2nd edition. New York: John Wiley & Sons, Inc; 2000.
organizations and administrators. Canadian Journal of Nursing Research 47. Chizawsky LL, Estabrooks CA, Sales A: The feasibility of web-based
2003, 35(3):53-68. surveys as a data collection tool: A process evaluation. Applied Nursing
24. Stastny P, Ichinose T, Thayer S, Olson R, Keens T: Infant sleep positioning Research 2009. In Press
by nursery staff and mothers in newborn hospital nurseries. Nurs Res 48. Bliese PD: Within-group agreement, non-independence, and reliability:
2004, 53(2):122-129. Implications for data aggregation and analysis. In Multilevel Theory,
25. Crane J: Factors associated with the use of research-based knowledge Research and Methods in Organizations: Foundations, Extensions and New
in nursing. In Unpublished doctoral dissertation University of Michigan; Directions Edited by: Klein KJ, Kozlowski SWJ. San Francisco: Jossey-Bass;
1990. 2000:349-381.
26. Scott-Findlay S, Golden-Biddle K: Understanding how organizational 49. Cummings GG, Fraser K, Tarlier DS: Implementing advanced nurse
culture shapes research use. Journal of Nursing Administration 2005, practitioner roles in acute care: An evaluation of organizational
35(7-8):359-65. change. J Nurs Adm 2003, 33(3):139-145.
27. Estabrooks CA: Research utilization and qualitative research. In The 50. Estabrooks CA, Midodzi WK, Cummings GG, Wallin L: Predicting research
Nature of Qualitative Evidence Edited by: Morse JM, Swanson J, Kuzel A. use in nursing organizations: A multilevel analysis. Nurs Res 2007, 56(4
Thousand Oaks, CA: Sage; 2001:275-298. Suppl):S7-23.
28. Rycroft-Malone J, Kitson A, Harvey G, McCormack B, Seers K, Titchen A, 51. Streiner DL;: Breaking up is hard to do: The heartbreak of dichotomizing
Estabrooks C: Ingredients for change: Revisiting a conceptual continuous data. Canadian Journal of Psychiatry 2002, 47(3):262-266.
framework. Qual Saf Health Care 2002, 11(2):174-180. 52. Barth AE, Stuebing KK, Anthony JL, Denton CA, Mathes PG, Fletcher JM,
29. Scott SD, Estabrooks CA, Allen M, Pollock C: A context of uncertainty: Francis DJ: Agreement among response to intervention criteria for
how context shapes nurses' research utilization behaviors. Qual Health identifying responder status. Learning and Individual Differences 2008,
Res 2008, 18(3):347-357. 18(3):296-307.
30. Wallin L, Estabrooks CA, Midodzi WK, Cummings GG: Development and
validation of a derived measure of research utilization by nurses. Nurs Pre-publication history
Res 2006, 55(3):149-160. The pre-publication history for this paper can be accessed here:
31. Bjornson CL, Klassen TP, Williamson J, Brant R, Mitton C, Plint A, Bulloch B, http://www.biomedcentral.com/1472-6963/10/168/prepub
Evered L, Johnson DW: A randomized trial of a single dose of oral
dexamethasone for mild croup. N Engl J Med 2004, 351(13):1306-1313. doi: 10.1186/1472-6963-10-168
32. Johnson DW, Jacobson S, Edney PC, Hadfield P, Mundy ME, Schuh S: A Cite this article as: Cummings et al., The relationship between characteris-
comparison of nebulized budesonide, intramuscular dexamethasone, tics of context and research utilization in a pediatric setting BMC Health Ser-
and placebo for moderately severe croup. N Engl J Med 1998, vices Research 2010, 10:168
339(8):498-503.