ORIGINAL CONTRIBUTION

Barriers and Facilitators to Intraosseous Access in Adult Resuscitations When Peripheral Intravenous Access Is Not Achievable Warren James Cheung, MD, Hans Rosenberg, MD, and Christian Vaillancourt, MD, MSc

Abstract Objectives: Studies suggest that intraosseous (IO) access is underutilized in adult resuscitations, despite recommendations from advanced trauma and cardiac life support guidelines. The objective was to determine factors associated with IO access use by physicians during adult resuscitations when intravenous (IV) access is not immediately achievable. Methods: This study was an online survey among physicians purposefully recruited from various clinical care areas at three teaching hospitals. Questions were generated from the qualitative results of 20 iterative interviews, verified for internal validity, and piloted. The interview guide was based on the constructs of the Theory of Planned Behavior (TPB), which elicits salient attitudes, social influences, and control beliefs that potentially influence intention to use IO access. Recruitment took place in September 2012 until reaching more than 100% of the required sample size (n = 200). Internal consistency was measured using Cronbach’s alpha, and the effect of TPB constructs and specific beliefs were assessed with regression analyses. Results: For the 205 respondents, the mean age was 35 years (range = 20 to 66 years), and 53.3% were male. Participants’ departmental affiliations were 50.3% emergency medicine (EM), 16.9% internal medicine, 14.9% anesthesia, 10.8% general surgery, and 7.2% critical care. Residents comprised 60.7% of the sample, and 39.3% were attending physicians. Median intention to use IO access when IV is not immediately achievable was 4.67 (interquartile range [IQR] = 4 to 5) out of 5 (5 highest) and predicted by the following TPB constructs: attitudes (AdjCoefficients = 0.504; 95% confidence interval [CI] = 0.334 to 0.673), social influences (AdjCoefficients = 0.285; 95% CI = 0.172 to 0.398), and control beliefs (AdjCoefficients 0.217; 95% CI = 0.113 to 0.320). Physicians were more likely to use IO access if they believed that it provided rapid vascular access for delivering large volumes of fluids, could prevent delays in care, and was associated with a low complication rate. Conversely, the perception that nurses are not familiar or supportive of IO access and a lack of physician confidence regarding the appropriate indications for IO access were barriers to use. Conclusions: These data are an important step in the knowledge-to-action process, as they identify specific factors associated with physician use of IO access. Interventions addressing these actionable facilitators and barriers are likely to have a positive effect on increasing the appropriate physician use of this potentially life-saving technique in adult patients requiring emergent vascular access. ACADEMIC EMERGENCY MEDICINE 2014; 21:250–256 © 2014 by the Society for Academic Emergency Medicine

From the Department of Emergency Medicine, University of Ottawa (WJC, HR, CV), Ottawa, Ontario; and the Clinical Epidemiology Program, Ottawa Hospital Research Institute (CV), University of Ottawa, Ottawa, Ontario, Canada. Received May 17, 2013; revisions received July 22, August 22, and September 27, 2013; accepted October 7, 2013. Presented at the Canadian Association of Emergency Physicians Conference, Niagara Falls, ON, Canada, June 2012, and Canadian Association of Emergency Physicians Conference, Vancouver, BC, Canada, June 2013. Funding provided by the Department of Emergency Medicine, University of Ottawa. The authors have no potential conflicts of interest to disclose. Supervising Editor: Clifton W. Callaway, MD, PhD. Address for correspondence and reprints: Warren James Cheung, MD; e-mail: [email protected].

250

ISSN 1069-6563 PII ISSN 1069-6563583

© 2014 by the Society for Academic Emergency Medicine doi: 10.1111/acem.12329

ACADEMIC EMERGENCY MEDICINE • March 2014, Vol. 21, No. 3 • www.aemj.org

O

btaining vascular access is a top priority in the management of critically ill patients. In most cases, peripheral intravenous (IV) cannulation is rapidly achievable. When peripheral IV access has failed or is delayed, central venous catheterization has been the traditional alternative.1 However, obtaining central venous access can be time-consuming and may result in serious complications, including pneumothorax and inadvertent arterial catheterization.2,3 Intraosseous (IO) vascular access takes advantage of the noncollapsible intramedullary space of cancellous bone and has been demonstrated to provide a rapid, reliable, and safe method of delivering medications, fluids, and blood products.4–6 Recognizing its merits, IO access is now endorsed by the American College of Surgeons Advanced Trauma Life Support course and recommended as the first alternative to failed or delayed IV access by the American Heart Association and the International Liaison Committee on Resuscitation.7–9 Despite these recommendations, a study of American emergency medicine (EM) residency training programs demonstrated that in unstable patients requiring emergent vascular access, central venous catheterization was the prevailing second-line alternative to peripheral IV access, and IO insertion was only considered the technique of choice if a fourth attempt was required.1 In addition, several European studies suggest that IO is infrequently used by physicians when indicated in adult resuscitations.10–12 To develop interventions to increase the appropriate use of this potentially life-saving technique in adult resuscitations, a better understanding of the barriers and facilitators to IO access use by physicians is required. The Theory of Planned Behavior (TPB) proposes that an individual’s intention to use IO access during adult resuscitations when peripheral IV access is not achievable, and his or her perceived control over adopting that behavior, are direct determinants of actually engaging in this behavior.13–16 Measurement of an individual’s intention has shown to correlate well with engaging in the behavior.15 This has been illustrated by the development of successful interventions for smoking cessation,17 colorectal screening,18 and healthy eating and physical activity.19 The TPB further posits that intention is influenced by three predictive constructs: attitudes or behavioral beliefs, subjective norms or social pressures, and control beliefs (Figure 1).14,15 These constructs can be used as a framework to explore theoretically derived determinants of behavior that can later be mapped to specific behavior change techniques.20 The purpose of our study was to identify determinants of behavior that influence physicians’ intentions to use IO access in adult resuscitations when peripheral IV access is not achievable. METHODS Study Design and Population We conducted an electronic survey evaluating the barriers and facilitators to performing IO access during adult resuscitations when peripheral IV access is not achievable. We based this survey on the constructs of the TPB (Figure 1).14,16,21 There is a growing body of literature in social sciences supporting the identification of evi-

251

Behavioral Beliefs Attitude towards the behavior

Subjective Norms Social pressure to perform behavior

Intention

Behavior

Control Beliefs Perceived control over the behavior

Figure 1. The Theory of Planned Behavior.15

dence-based competencies required to effectively change behavior in health care related interventions.22–24 Over the years, the TPB has emerged as a useful, systematic tool to identify barriers to and facilitators of change, aiding in the design of appropriate forms of interventions.16,25–30 We recruited residents and attending physicians at The Ottawa Hospital (consisting of three hospital campuses) from the Departments of Emergency Medicine and Anesthesia, and Divisions of General Internal Medicine, General Surgery, and Critical Care—the population of physicians most likely to be involved in adult resuscitations. We obtained research ethics approval from The Ottawa Hospital Research Ethics Board, Ottawa, Ontario (2011162-01H) prior to initiating the study. Participation in the study was voluntary and completion of the electronic survey implied consent to participate. Survey Content and Administration We created our survey questions based on the “themes” (Data Supplement S1, available as supporting information in the online version of this paper) that emerged from iterative, semistructured, qualitative interviews of a purposive sample of 20 residents and attending physicians.31 Interviews consisted of open-ended questions and were conducted, recorded, and analyzed until data saturation was achieved.32 Two independent reviewers (WJC, HR) performed inductive analysis to identify emerging themes (e.g., “Intraosseous access is painful for the patient”) within the framework of the TPB constructs. Similar themes were combined by way of consensus, and all themes were subsequently ranked according to how frequently they were expressed by interviewees. Themes representing the top 75 percentile within each TPB construct were retained for inclusion as a question/formulated statement in the final survey.15 All authors reviewed the initial draft survey to ensure content and face validity. We then piloted the survey with 10 participants and made minor adjustments for clarity. Our electronic survey instrument consisted of a small number of demographic related questions, as well as 51

252

statements quantified using a five-point Likert scale ranging from “strongly disagree” (score of 1) to “strongly agree” (score of 5) with “neither agree nor disagree” at the center (score of 3). A copy of the final survey instrument is presented in Data Supplement S2 (available as supporting information in the online version of this paper). We intentionally phrased selected statements negatively throughout the survey to avoid yea-saying bias (tendency to respond to each statement with the same answer). We designed questions to measure respondents’ intention to perform the behavior of interest: inserting an IO needle in adult patients during resuscitation when peripheral IV access is not achievable. We quantified the overall effect of each TPB construct on the intention to perform the behavior (main outcome measure) using “general” predictive statements.33 For example, we used general statements such as “I am confident I could use an IO if I wanted to” to assess the predictive value of the “control belief” TPB construct. We also quantified the “specific” effect of individual factors on the intention to use IO access by pairing theme-based beliefs generated from the interview process (e.g., “Intraosseous access is painful for the patient”) with a statement of effect (e.g. “If intraosseous access is painful for the patient, I will not use it”). As suggested in the TPB manual by Francis et al.,33 we used the product of the scores of the theme-based and effect statements as a “weighted theme-based” quantitative predictive measure of intention to perform the behavior. We distributed the electronic survey using an online link via the respective e-mail distribution lists of the departments and divisions. We were granted approval by residency program directors and department or division heads prior to distribution. We sent out followup reminder e-mails 1 and 2 weeks after the initial distribution. Data Analysis We performed all data analyses using SAS (Version 9.2, Cary, NC) and report respondent characteristics using descriptive statistics. We report intention to insert an IO during an adult resuscitation when IV access is not achievable as a median with interquartile range (IQR), alongside the actual frequency of responses. We measured the survey’s internal consistency using Cronbach’s alpha. Once this was verified, we averaged the scores for all three statements measuring behavioral intention and for all general statements measuring the predictive effect of each of the three TPB constructs (two general statements measuring attitudinal beliefs, three for normative beliefs, and two for control beliefs). To determine which of the three TPB constructs was most predictive of the respondents’ intention to use IO access, we used a multivariate stepwise linear regression model. We also evaluated the following variables in consideration for inclusion in the modeling exercise: age, sex, specialty, number of IOs inserted, having receiving didactic instructions on IO use, previous hands-on training experience, “real-life” resuscitation experience, attending status versus resident, years of experience as attending, and years of experience as resident. First, we performed a series of univariate analy-

Cheung et al. • BARRIERS TO INTRAOSSEOUS ACCESS

ses, only selected those variables with two-tailed levels of significance of p ≤ 0.25, and excluded the following variables as a result: age, sex, attending status, and years of experience as attending. Second, we used a stepwise approach to determine which remaining variables should be included in the final linear regression model. SAS uses the F-statistic at a significance level of 0.15 to make this selection. After a variable is added into the model, this stepwise method looks back at all the variables already included in the model and deletes any variable that does not produce a significant F-statistic. A new variable is only added once this is verified, and this process ends when none of the remaining variables outside the model has an F-statistic that is significant. Five variables were introduced sequentially into what became the final model: attitudinal beliefs, normative beliefs, control beliefs, real-life resuscitation experience, and specialty. Finally, before reporting on the adjusted influence of each variable included in the final model on participants’ intention to use IO access, we verified that basic model assumptions were met, that residuals were normally distributed, that there were no significant outliers or confounders, and that acceptable homoscedasticity was present. We report results from the linear regression as adjusted coefficients with 95% confidence intervals (CIs) and the model goodness-of-fit using r-square statistics. In this case, these adjusted linear regression coefficients represent the increase in a respondent’s intention to perform IO insertion for every 1-unit increase in response for independent predictive variables. To identify which individual factors should be targeted in interventions to improve appropriate use of IOs, we performed a series of univariate regression analyses using weighted theme-based beliefs (independent variables) that could influence one’s decision to use IO or not (dependent variable). We chose this univariate approach because we wanted to identify not only unique or “adjusted” variance, but “any” variance that is associated with each potential predictive weighted theme. To discriminate between high and low intention, we used a dichotomized Likert scale based on “median intention” to use an IO. We report results from these univariate logistic regression analyses as unadjusted ORs with 95% CI. It is estimated that a sample size of 80 participants is sufficient for most TPB surveys, assuming at least a moderate effect size.33,34 To be conservative, we aimed to recruit 200 participants. The sample of 200 is adequate for providing a sufficient number of cases per variable in the final linear regression model, five to 15 cases/covariate being a common standard. RESULTS We collected a total of 225 responses, and 85.3% of respondents completed all elements of the survey. Characteristics of the 205 survey respondents who provided demographic data are presented in Table 1. Seventyfour percent of respondents had prior IO insertion training. This was highest among emergency physicians and lowest in the internal medicine group. Similarly, experience with prior IO insertion was highest in the

ACADEMIC EMERGENCY MEDICINE • March 2014, Vol. 21, No. 3 • www.aemj.org

Table 1 Characteristics of 205 Survey Respondents Characteristic Age (yr) Male Physician type Attending Experience (yr) Resident Residency training level (yr) Specialty EM Internal medicine General surgery Anesthesia Critical care Previous IO access training EM Internal medicine General surgery Anesthesia Critical care All specialties Previous IO insertion experience EM Internal medicine General surgery Anesthesia Critical care All specialties Previous IO insertions EM Internal medicine General surgery Anesthesia Critical care All specialties

n (%) or Mean (range) 35 (20–66) 104 (53.3)

253

Table 2 Evaluation of the Predictive Value of Theory of Planned Behavior Constructs for Respondents’ Intention to Perform IO Insertion when Peripheral IV Access Is Not Achievable

IO Insertion Independent Variables

Adjusted Linear Regression Coefficients

95% CI

0.504 0.285

0.334 to 0.673 0.172 to 0.398

0.217 –0.117

0.113 to 0.320 –0.199 to –0.036

–0.102

–0.187 to –0.017

77 12 119 3

(39.3) (1–40) (60.7) (0–10)

98 33 21 29 14

(50.3) (16.9) (10.8) (14.9) (7.2)

Attitude statements Normative statements Control statements Specialty other than EM Having more real-life resuscitation experience

88 17 14 18 8 145

(89.8) (51.5) (66.7) (62.1) (57.1) (74.4)

R2 statistic = 0.60. These adjusted linear regression coefficients represent the increase in a respondent’s intention to perform IO insertion for every 1-unit increase in response for independent predictive variables. IO = intraosseous.

87 13 11 11 10 132

(88.8) (39.4) (52.4) (37.9) (71.4) (67.7)

7.3 0.9 1.1 0.7 3.2 4.3

(0–60) (0–3) (0–5) (0–3) (0–10) (0–60)

IO = intraosseous.

EM group and lowest among general internists and anesthesiologists. In total, 68% of all participants had prior experience inserting an IO needle. Each group of questions achieved an acceptable Cronbach’s alpha statistic after removing one disparate attitude statement and one disparate control question from the survey (Cronbach’s alpha ranging between 0.77 and 0.89). Median intention to use IO access in an adult resuscitation when peripheral IV access is not achievable was 4.67 (IQR = 4 to 5) on a five-point Likert scale, and the frequencies of responses were 1 (2.4%), 2 (9.2%), 3 (9.2%), 4 (25.9%), and 5 (53.4%). The predictive effect of each of the three TPB constructs on intention to use IO access is presented in Table 2. All three constructs were associated with intention to engage in the behavior of interest with attitudinal beliefs demonstrating the greatest predictive effect, followed by normative beliefs and control beliefs. Specialties other than EM were less likely to intend to use IO access, and having more reallife resuscitation experience also appeared to be inversely associated with one’s intention to use an IO. Weighted theme-based beliefs that were predictive of intention to use IO access in adult resuscitations when IV access is not achievable are presented in Table 3. Five of seven attitudinal beliefs, all three normative beliefs, and only one of four control beliefs were predic-

Table 3 Predictive Value of Specific Weighted Theme-based Beliefs for Respondents’ Intention to Perform IO Insertion When Peripheral IV Access Is Not Achievable Statements Specific attitude statements There are low complication rates associated with inserting an IO needle Inserting an IO is an easy procedure IO insertion provides rapid vascular access IO insertion can prevent delays in fluid resuscitation in patients without IV access IO access allows for infusion of large volumes of fluid IO needle insertion is a temporary form of vascular access Inserting an IO needle causes the patient pain Specific normative statements Allied health professionals are not familiar with the utility of IO access Nurses are supportive of me using IO There is a negative stigma associated with inserting an IO (failure of IV or central access) Specific control statements I know the indications and contraindications of inserting an IO needle IO equipment is accessible Prior hands-on experience is important IO insertion protocols are helpful

OR

95% CI

1.16

1.11–1.25

1.14 1.13 1.10

1.08–1.21 1.03–1.23 1.04–1.16

1.07

1.00–1.13

1.05

0.97–1.13

0.99

0.91–1.06

1.16

1.09–1.23

1.15 1.12

1.08–1.22 1.06–1.18

1.12

1.03–1.22

1.04 1.03 1.01

0.96–1.11 0.97–1.09 0.97–1.05

IO = intraosseous.

tive of respondents’ intention to engage in IO access during adult resuscitations. Median response scores for all unweighted beliefs are also presented in Data Supplement S2.

254

DISCUSSION The results of our study indicate that, to increase physicians’ use of IO access in adult resuscitations when peripheral IV access is not achievable, educational interventions need to address their attitudinal, normative, and control beliefs. More specifically, we demonstrated that physicians are more likely to engage in the behavior if they know that IO access is easy to obtain, that it provides rapid vascular access for delivering large volumes of fluids, that it can prevent delays in care, and that it is associated with a low complication rate. Consequently, behavioral change techniques targeting these attitudinal beliefs are most likely to have an effect on use of IO access. We also demonstrated that social pressures from allied health professionals might influence the use of IO access by physicians. In our survey, physicians who perceived that nurses were not familiar with or not supportive of IO insertion were less likely to report intent to use this form of access in adult resuscitations. This is in keeping with our finding that physicians who perceived a negative stigma toward IO access were less likely to intend to perform the behavior. Accordingly, interventions addressing these normative barriers need to be developed. Regarding control beliefs, lacking confidence in the appropriate indications and contraindications for IO access was associated with low intention to perform the behavior. This finding suggests that educational interventions addressing the “knows” and “knows how” levels of Miller’s pyramid of clinical competence35 are necessary foundations to increasing appropriate use of IO access. Several studies have identified similar barriers to the use of IO access in emergency situations and have suggested methods of addressing these difficulties.2,5,36,37 Voigt et al.36 acknowledge the important role of nurses in achieving and maintaining vascular access and their potential influence on driving appropriate use of IO infusions within the emergency department setting. Similarly, in their consensus paper, Phillips et al.5 recommend that, to increase appropriate use of IO access, the procedure should become a core component of nursing school curricula. In keeping with our findings, nursing staff who are knowledgeable about the appropriate indications and maintenance of IO infusions may help to reduce some of the negative social pressures perceived by physicians. Others suggested that a shift in clinical practice that encourages appropriate use of IO access can be achieved through the development of hospital-based protocols that outline procedures for insertion, maintenance, and removal of IO devices, as well as through the establishment of guidelines from physician specialty societies.2,5,36,37 Although our study suggests that IO access protocols are not predictive of intention to use IO access in emergent situations, difficult vascular access algorithms incorporating IO infusions have been successfully implemented in the past.2 These hospital-based protocols may empower physicians to use IO access by reminding them of the appropriate indications for use,36 a control belief barrier identified in our study. Furthermore, hospital-based protocols and specialty society guidelines may help reduce the negative stigma associated with IO access by

Cheung et al. • BARRIERS TO INTRAOSSEOUS ACCESS

normalizing its use. Hospital and specialty society endorsement may additionally help to promote more widespread teaching of the technique among physicians.10,38 Future research should be targeted at the implementation and evaluation of these interventions to determine their effect on increasing the appropriate utilization of IO access. LIMITATIONS This study has several limitations that are common to any survey research. Our study involved voluntary participation, which may have introduced selection bias. Study participants may have had more experience or stronger beliefs about the use of IO access than nonparticipants. We found that almost 75% of participants had some form of prior IO access training and 68% had prior experience inserting an IO needle. To reduce the potential for selection bias, we purposefully recruited physicians from a variety of specialties. Furthermore, the response rate among emergency physicians, who represented half of respondents, was high (87.5%). Although the overall response rate for our study is not known, analyses of the TPB are primarily driven by sample size.16 Another limitation is that the results of our study may not be generalizable to other settings as the survey was conducted at three local hospitals in Ottawa, Canada. However, several European and American studies have identified similar barriers to appropriate IO access use and support our findings.5,36,38 Finally, an inherent limitation is found in the TPB, which proposes that an individual’s intention is correlated with actually engaging in the behavior of interest. This association has been previously demonstrated, however,15 and the TPB is a widely recognized and accepted theoretical framework for studying complex health care interventions.16–18,21 CONCLUSIONS The apparent underutilization of intraosseous access described in the literature represents a breakdown in knowledge translation. We have contributed an important step in the knowledge-to-action process by applying the Theory of Planned Behavior framework to identify actionable facilitators and barriers to intraosseous access that were not previously understood. Interventions addressing these barriers at a multidisciplinary level are likely to have a positive effect on increasing the appropriate use of intraosseous access in adult resuscitations when peripheral intravenous access is not achievable. References 1. Bloch S, Bloch AJ, Silva P. Adult intraosseous use in academic EDs and simulated comparison of emergent vascular access techniques. Am J Emerg Med 2013;31:622–4. 2. Gazin N, Auger H, Jabre P, et al. Efficacy and safety of the EZ-IOTM intraosseous device: out-of-hospital implementation of a management algorithm for difficult vascular access. Resuscitation 2011;82:126–9.

ACADEMIC EMERGENCY MEDICINE • March 2014, Vol. 21, No. 3 • www.aemj.org

3. Taylor RW, Palagiri AV. Central venous catheterization. Crit Care Med 2007;35:1390–6. 4. Leidel B, Kirchhoff C, Bogner V, Braunstein V, Biberthaler P, Kanz KG. Comparison of intraosseous versus central venous vascular access in adults under resuscitation in the emergency department with inaccessible peripheral veins. Resuscitation 2012;83:40–5. 5. Phillips L, Brown L, Campbell T, Miller J, Proehl J, Youngberg B. Recommendations for the use of intraosseous vascular access for emergent and nonemergent situations in various health care settings: a consensus paper. Crit Care Nurse 2010;30:e1–7. 6. Rosenberg H, Cheung WJ. Intraosseous access. CMAJ 2012;185:E238. 7. American College of Surgeons. Advanced Trauma Life Support for Doctors (Student Course Manual), 8th ed. Chicago, IL: American College of Surgeons, 2008. 8. American Heart Association. American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care: part 7.2– management of cardiac arrest. Circulation 2005;112 (Suppl 24):IV58–66. 9. International Liaison Committee on Resuscitation. International consensus of cardiopulmonary resuscitation and emergency cardiovascular care science with treatment recommendations: advanced life support. Resuscitation 2005;67(Pt 4):213–47. 10. Lavis M, Vaghela A, Tozer C. Adult intraosseous infusion in accident and emergency departments in the UK. J Accid Emerg Med 2000;17:29–32. 11. Hallas P, Brabrand M, Folkestad L. Reasons for not using intraosseous access in critical illness. Emerg Med J 2012;29:506–7. 12. Molin R, Hallas P, Brabrand M, Schmidt TA. Current use of intraosseous infusion in Danish emergency departments: a cross-sectional study. Scand J Trauma Resusc Emerg Med 2010;18:37. 13. Ajzen I. The direct influence of attitudes on behaviour. In: Gollwitzer PM, Bargh JA, eds. The Psychology of Action: Linking Cognition and Motivation to Behavior. New York, NY: Guilford Press, 1996:385– 403. 14. Vaillancourt C, Charette M, Kasaboski A, et al. Barriers and facilitators to CPR knowledge transfer in an older population most likely to witness cardiac arrest: a theory-informed interview approach. Emerg Med J 2013; [Epub 2013 May 1]. 15. Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process 1991;50:179–211. 16. Godin G, Kok G. The theory of planned behavior: a review of its applications to health-related behaviors. Am J Health Promot 1996;11:87–98. 17. Michie S, Churchill S, West R. Identifying evidencebased competences required to deliver behavioural support for smoking cessation. Ann Behav Med 2011;41:59–70. 18. Michie S, Collins V, Halliday J, Marteau TM. Likelihood of attending bowel screening after a negative genetic test result: the possible influence of health professionals. Genet Test 2002;6:307–11.

255

19. Michie S, Abraham C, Whittington C, McAteer J, Gupta S. Effective techniques in healthy eating and physical activity interventions: a meta-regression. Health Psychol 2009;28:690–701. 20. Michie S, Johnston M, Francis JJ, Hardeman W, Eccles M. From theory to intervention: mapping theoretically derived behavioural determinants to behaviour change techniques. Appl Psychol 2008;57:660–80. 21. Conner M, Sparks P. The theory of planned behaviour and health behaviour. In: Conner M, Norman P, eds. Predicting Health Behaviour. Berkshire, UK: Open University Press, 1996:121–62. 22. Michie S, Hyder N, Walia A, West R. Development of a taxonomy of behaviour change techniques used in individual behavioral support for smoking cessation. Addict Behav 2011;36:315–9. 23. Michie S, Johnston M, Abraham C, Lawton R, Parker D, Walker A. Making psychological theory useful for implementing evidence based practice: a consensus approach. Qual Saf Health Care 2005;14:26–33. 24. Webb T, Sniehotta F, Michie S. Using theories of behaviour change to inform interventions for addictive behaviours. Addiction 2010;105:1879–92. 25. Sheeran P, Orbell S. Do intentions predict condom use? Meta-analysis and examination of six moderator variables. Br J Soc Psychol 1998;37:231–50. 26. Sutton S. Predicting and explaining intentions and behavior: how well are we doing? J Appl Soc Psychol 1998;28:1317–38. 27. Godin G. Determinants of nurses compliance to universal precautions. Am J Infect Control 2000;28:359– 64. 28. Godin G, Myers T, Lambert J, Calzavara L, Locker D. Understanding the intention of gay and bisexual men to take the HIV antibody test. AIDS Educ Prev 1997;9:31–41. 29. Bunce D, Birdi K. The theory of reasoned action and the theory of planned behaviour as a function of job control. Br J Health Psychol 2003;3:265–75. 30. Conner M, Heywood-Everett S. Addressing mental health problems with the theory of planned behaviour. Psychol Health Med 1998;3:87–95. 31. Cheung WJ, Rosenberg H, Vaillancourt C. A survey of factors associated with the use of intraosseous access in adult patients when peripheral intravenous access is not available [abstract]. CJEM 2013;15(Suppl 1):S14. 32. Francis JJ, Johnston M, Robertson C, et al. What is an adequate sample size? Operationalising data saturation for theory-based interview studies. Psychol Health 2010;25:1229–45. 33. Francis JJ, Eccles M, Johnston M, et al. Constructing questionnaires based on the Theory of Planned Behavior: a manual for health services researchers. Newcastle, UK: Center for Health Services Research, 2004. 34. Cohen J. Statistical Power Analyses, 2nd ed. Hillsdale, NJ: Erlbaum, 1988. 35. Miller G. The assessment of clinical skills/competence/performance. Acad Med 1990;65:S63–7.

256

36. Voigt J, Waltzman M, Lottenberg L. Intraosseous vascular access for in-hospital emergency use: a systematic clinical review of the literature and analysis. Pediatr Emerg Care 2012;28:185–99. 37. Fowler R, Gallagher JV, Isaacs SM, Ossman E, Pepe P, Wayne M. The role of intraosseous vascular access in the out-of-hospital environment (resource document to NAEMSP position statement). Prehosp Emerg Care 2007;11:63–6. 38. Lo TY, Reynolds F. To use intraosseous access or not to use intraosseous access: determinants of

Cheung et al. • BARRIERS TO INTRAOSSEOUS ACCESS

trainees’ decision in paediatric emergencies. Eur J Emerg Med 2009;16:301–4. Supporting Information The following supporting information is available in the online version of this paper: Data Supplement S1. Results of semi-structured interviews. Data Supplement S2. Final survey instrument.