UBC Faculty Research and Publications

A cluster randomized controlled trial comparing three methods of disseminating practice guidelines for… Johnson, David W; Craig, William; Brant, Rollin; Mitton, Craig; Svenson, Larry; Klassen, Terry P Apr 28, 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13012_2006_Article_10.pdf [ 317.29kB ]
JSON: 52383-1.0223949.json
JSON-LD: 52383-1.0223949-ld.json
RDF/XML (Pretty): 52383-1.0223949-rdf.xml
RDF/JSON: 52383-1.0223949-rdf.json
Turtle: 52383-1.0223949-turtle.txt
N-Triples: 52383-1.0223949-rdf-ntriples.txt
Original Record: 52383-1.0223949-source.json
Full Text

Full Text

ralssBioMed CentImplementation ScienceOpen AcceStudy protocolA cluster randomized controlled trial comparing three methods of disseminating practice guidelines for children with croup [ISRCTN73394937]David W Johnson*1, William Craig2, Rollin Brant3, Craig Mitton4, Larry Svenson5 and Terry P Klassen2Address: 1Department of Pediatrics, Faculty of Medicine, University of Calgary, Calgary Alberta, Canada, 2Department of Pediatrics, Faculty of Medicine, University of Alberta, Edmonton, Alberta, Canada, 3Department of Statistics, University of British Columbia, Vancouver, British Columbia, Canada, 4Faculty of Health and Social Development, University of British Columbia- Okanagan, Kelowna, British Columbia, Canada and 5Health Surveillance, Alberta Health and Wellness, Edmonton, Alberta, CanadaEmail: David W Johnson* - david.johnson@calgaryhealthregion.ca; William Craig - wcraig@cha.ab.ca; Rollin Brant - rollin@stat.ubc.ca; Craig Mitton - cmitton@exchange.ubc.ca; Larry Svenson - larry.svenson@gov.ab.ca; Terry P Klassen - terry.klassen@ualberta.ca* Corresponding author    AbstractBackground: The optimal management of croup – a common respiratory illness in young children – iswell established. In particular, treatment with corticosteroids has been shown to significantly reduce therate and duration of intubation, hospitalization, and return to care for on-going croup symptoms.Furthermore treatment with a single dose of corticosteroids does not appear to result in any significantadverse outcomes, and yields overall cost-savings for both families and the health care system.However, as has been shown with many other diseases, there is a significant gap between what we knowand what we do. The overall aim of this study is to identify, from a societal perspective, the costs andassociated benefits of three strategies for implementing a practice guideline that addresses themanagement of croup.Methods/designs: We propose to use a matched pair cluster trial in 24 Alberta hospitals randomizedinto three intervention groups. We will use mixed methods to assess outcomes including linkage andanalysis of administrative databases obtained from Alberta Health and Wellness, retrospective medicalchart audit, and prospective telephone surveys of the parents of children diagnosed to have croup. Theintervention strategies to be compared will be mailing of printed educational materials (low intensityintervention), mailing plus a combination of interactive educational meetings, educational outreach visits,and reminders (intermediate intensity intervention), and a combination of mailing, interactive sessions,outreach visits, reminders plus identification of local opinion leaders and establishment of local consensusprocesses (high intensity intervention). The primary objective is to determine which of the threeintervention strategies are most effective at lowering the rate of hospital days per 1,000 disease episodes.Secondary objectives are to determine which of the three dissemination strategies are most effective atincreasing the use of therapies of known benefit. An economic analysis will be conducted to determinewhich of the three intervention strategies will most effectively reduce total societal costs including allhealth care costs, costs borne by the family, and costs stemming from the strategies for disseminatingPublished: 28 April 2006Implementation Science2006, 1:10 doi:10.1186/1748-5908-1-10Received: 25 January 2006Accepted: 28 April 2006This article is available from: http://www.implementationscience.com/content/1/1/10© 2006Johnson et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the 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.Page 1 of 13(page number not for citation purposes)guidelines.Implementation Science 2006, 1:10 http://www.implementationscience.com/content/1/1/10BackgroundIntroductionCompared with biomedical research, few resources havebeen devoted to translating knowledge into practice.However health care researchers are increasingly aware ofthe persistent and substantial gap between what we knowand what we do. One of the most common strategiesadvocated for closing this gap is the development of prac-tice guidelines. It is now widely accepted, however, thatsimple dissemination whether by publication in medicaljournals or direct mailings to physicians of guidelines,does not result in a significant change in practice.More than 40 systematic reviews have been publishedwhich examine a range of different strategies for imple-menting guidelines. An overview of this literature hasbeen published by the Cochrane Effective Practice andOrganisation of Care (EPOC) Group.[1,2] This reviewconcludes that many of the primary studies focusing onimplementation strategies have weak designs, methodo-logical flaws, and that virtually none have included eco-nomic evaluations. The authors call for randomized trialswith head-to-head, multi-arm comparisons of differentlevels of interventions, that use an appropriate analysis,based on clusters, and include a comprehensive cost-effec-tive analysis.We have designed a study that addresses each of the pointsoutlined by the EPOC Group's members. Furthermore webelieve – for several reasons – that practice guidelinesaddressing the management of croup will provide a goodtest case for analyzing our proposed intervention.Croup is second only to asthma as a cause of respiratoryemergencies in young children, and accounts for 5% of allemergent admissions in this population. In the last 15years, as a result of the publication of a number of thera-peutic trials and systematic reviews, the scientific basis forcroup treatment has been clarified. In particular, opti-mally timed treatment with steroids substantially reducesthe frequency and duration of both hospitalization andairway intubation. [3-5]Published literature and pilot data collected for this pro-posed trial show that many children with croup do notreceive optimal therapy, and that there is a substantial var-iation in hospitalization rates.[6,7] Since hospitalizationlikely accounts for most health care spending on this dis-ease, improvement in prescribing practices and standard-ization of indications for hospital admission could bothimprove outcomes and substantially reduce health careexpenditures for children with croup.Strategies for disseminating and implementing clinical practice guidelinesThe "Gap"On-going societal investment in basic and clinicalresearch is ultimately only meaningful if this acquiredknowledge is translated into better care for patients.Unfortunately, in health care, there is a substantial gapbetween what is known and what is actually done forpatients. Though this has been recognized for some time,this gap has not disappeared. Some well-documentedexamples include the use of photocoagulation for diabeticretinopathy,[8] management of hypertension,[8] prophy-laxis for patients at high risk for venous thrombosis andpulmonary embolism,[8] and, most recently, thrombo-lytic therapy for patients with myocardial infarction. [8-10]Approaches to knowledge translationMany different broad approaches to knowledge transla-tion have been advocated including development and dis-semination of clinical practice guidelines, continuingmedical education, continuous quality improvement, andmore recently, computerized decision support sys-tems.[11,12]What constitutes a good practice guideline?Guidelines are formally defined as "systematically devel-oped statements to assist practitioners' and patients' deci-sions about appropriate care for specified clinicaloutcomes".[13] A number of organizations have devel-oped policies and standards for how guidelines should bedeveloped and evaluated. [12-17] It is desirable for guide-lines to be valid (i.e. if followed, guidelines should yieldthe health care gains and costs predicted); reproducible(i.e. if given the same evidence and methods, two commit-tees should develop comparable guidelines); and reliable(i.e. in the same clinical circumstances, clinicians shouldinterpret the guidelines in the same way).[9] Guidelinesare most likely to be valid if developed by national organ-izations with representatives from all key disciplines, andif formally linked to meta-analysis.[9]Guideline development is expensive and often falls shortDevelopment of guidelines can be very expensive, espe-cially those developed by national agencies which canrange up to US $1 million.[17] Furthermore, there aremore than 20,000 guidelines registered in the U.S.National Guideline Clearinghouse, and more than 2,000guidelines on the Canadian Medical Association Web-site.[18] Consequently the total amount of healthresource dollars now devoted to guideline development isconsiderable. For example, the United Kingdom spendsan estimated 1.5% of the annual health budget on thePage 2 of 13(page number not for citation purposes)development of national guidelines.Implementation Science 2006, 1:10 http://www.implementationscience.com/content/1/1/10A critical review of practice guidelines found that, of theguidelines published in a selection of journals in the lastdecade, none meet all methodological standards, and fewmet more than 70% of standards.[19] The methodologi-cal area in which guidelines needed the greatest improve-ment was the identification, evaluation, and synthesis ofscientific evidence.Dissemination alone doesn't workIt is now widely accepted that no matter how well-devel-oped practice guidelines are, simple dissemination doesnot result in a significant change in practice.[1,2] Anumber of different strategies to implement guidelineshave been published. Cochrane Effective Practice AndOrganisation Of Care Group http://www.abdn.ac.uk/hsru/epoc/ has systematically reviewed the published lit-erature, developed a taxonomy of interventions, anddeveloped methodological criteria for assessing the qual-ity of the evaluation of these interventions.Implementation strategies: what works and what doesn't workBased on published reviews by the Cochrane EPOCGroup and others, we can broadly categorize the differentimplementation strategies into three groups as showingconsistent, variable, or little or no effectiveness.[1,2]Those interventions that consistently have shown effec-tiveness include interactive educational meetings, educa-tional outreach visits, reminders (either manual orcomputerized), and multifaceted interventions (definedas a minimum of two combined interventions). Interven-tions that have shown a range of effectiveness includeaudit and feedback, the use of local opinion leaders, localconsensus processes, and patient-mediated interventions.Interventions that have consistently shown little or noeffect are didactic educational meetings (lecture-format)and educational materials (distribution of recommenda-tions for clinical care, including practice guidelines, audi-ovisual materials, and electronic publications).Management of children with croupClinical presentation and infectious etiologyCroup (acute laryngotracheo-bronchitis) is a commonrespiratory tract illness in young children. Parainfluenzavirus is, by far, the most common cause of croup. Otherimportant causes include influenza, adeno, and respira-tory syncytial viruses.[20,21] The illness is characterizedby a barking cough, hoarseness, inspiratory stridor, andoften severe respiratory distress that can occur suddenly inthe middle of the night. This can be frightening for par-ents. The barking cough is distinct for croup. It and theaccompanying inspiratory stridor are easily recognizableby clinicians and, once educated, by parents.Epidemiology & burden of the diseaseIt most commonly affects children between the ages of 6months to 3 years.[21] The disease peaks biannually, withthe largest peak between September and December and asecond smaller peak between January and March.[22]Croup accounts for a significant proportion of paediatricemergency department (ED) visits and hospitalizations.The proportion of children evaluated as outpatients whoare hospitalized range from 1–6% of those seen in privatephysician offices, to 7–31% of those evaluated in anED.[21,23-25] Hospitalizations of children with croup aretypically short, with the median duration being 48 hours.The potential for respiratory failure and death is whatdrives physicians to consider hospitalization. However,endotracheal intubation is uncommon (0.4–1.4% of hos-pitalized children), and death is exceptionally rare (0.5%of intubated children).[26]Indications for hospital admissionFew studies have examined which children are at risk forrespiratory failure. A retrospective cohort study of 527hospitalized children suggested that the presence orabsence of persistent sternal and chest wall indrawingmay be an important clinical factor in determiningrisk.[27] A small prospective study suggests that the insti-tution of a clinical pathway for croup in an ED – usingexplicit criteria for when it is safe to discharge childrenhome – can significantly reduce admission rates andlength of stay without any significant adverse events.[28]Variability in hospitalization ratesUsing the technique of small area analysis, To and Wen-nberg have reported a five and 17-fold difference, respec-tively, in the rate of hospitalization per capita amongcommunities. Wennberg noted that the rate for croup andseveral other common pediatric problems, relative toadult medical conditions, was significantly more varia-ble.[6,7]BaselineUtilisation data from AlbertaUtilizing provincial administrative data, we have estab-lished health care utilization rates including the a)number of disease episodes, b) number of physician visitsper disease episode, and c) number of hospital admis-sions and hospitals days per disease episode, for childrenwith croup in Alberta for 6 years. Among the 107 Albertahospitals, we found up to an 18 fold range in hospitaladmissions rates, ranging from16 to 287/1000 diseaseepisodes. In general, those hospitals that evaluated largernumbers of children with croup had lower utilizationrates. For example, the six largest volume hospitals diag-nosed a total of 86,711 cases, and averaged 21 hospitaladmissions/1,000 disease episodes; and the 30 smallestPage 3 of 13(page number not for citation purposes)volume hospitals diagnosed 13,750 cases and averaged116 hospital admissions/1,000 disease episodes.Implementation Science 2006, 1:10 http://www.implementationscience.com/content/1/1/10Costs associated with hospitalizationThough there are no comprehensive studies examiningthe health care costs of croup, one study suggests that thetotal 'costs' for hospitalization were almost three timesthat of the total costs for all ED visits.[23] In the case ofbronchiolitis – which likely has a similar costing structureto croup – 62% of all health care expenditures for the dis-ease are expended on the less than 1% of children who areadmitted to hospital.[29] Our baseline provincial data forcroup shows that the average number of MD visits per dis-ease episode is only 1.15, and that 3.6% of children withcroup are hospitalised. Given the relative greater costsassociated with hospital admission versus MD visits, orother costs such as medications, it is likely that hospitaladmissions are the principal determinant of health expen-ditures. Therefore a substantial reduction in hospitaliza-tion rates should substantially reduce health careexpenditures for the disease.Overview of current therapyTreatment for croup includes oxygen, mist, epinephrine,corticosteroids, and heliox.[30,31] Though mist has beenused for a long time, there is little evidence for or againstits benefit.[32] The evidence includes small randomisedand non-randomized trials, a study that used an animalmodel of uncertain applicability, and, recently, a well-masked randomised trial published by one of us (TPK).[33-36] None of these studies found mist to be beneficial.Epinephrine is clearly effective in the short-term, based onseveral randomised trials using clinical scores, and non-controlled trials using a range of 'objective' measures toassess degree of respiratory distress.[34,37-40] In patientswith severe respiratory obstruction, its use probably pre-vents or delays the need for endotracheal intuba-tion.[40,41] Nebulized epinephrine is generally thoughtto be safe.[30,31,42] Since the effect of epinephrine, how-ever, is short-lived and does not alter the natural historyof the disease,[38] there appears to be no real benefit to itsuse in children with mild symptoms.Heliox (helium mixed with oxygen) allows efficient lami-nar flow in narrower airways, and, at least in theory, candelay or even prevent endotracheal intubation. [43] Twosmall randomized trials in children with moderatelysevere croup have been published. One compared helioxto nebulized epinephrine and found the two interven-tions to be equivalent.[44] The other study comparedheliox to standard oxygen therapy, and found a trend infavor of heliox that was not statistically significant.[45]The clinical benefit of steroids is well documented. Inaddition to observations that steroids improve clinicalbation, [4] the amount of treatment required with epine-phrine,[3,5] the duration of hospitalization,[47] the rateof hospitalization, [5,48] and the rate of return to theED.[49] Steroid therapy for croup is generally thought tobe safe. [31]Unsubstantiated therapies: antibiotics, decongestants, & beta-agonistsOther types of therapies sometimes administered topatients with croup include antibiotics, decongestants,and nebulized β-agonists. No evidence exists for theirbenefit.Variability in treatment practicesA recent study reported significant differences in the pro-portion of children treated with mist, epinephrine, andsalbutamol based on whether physicians were pediatri-cians or not.[50] All physicians in these two urban hospi-tals, however, used corticosteroids in more than 90% ofchildren with croup. In contrast, our data from Albertashows that a much smaller proportion of children receivecorticosteroids, (e.g. only 22% at the 12 smaller hospi-tals). In these same small hospitals, the same proportionof children receive antibiotics or salbutamol; treatmentsfor which there is no evidence of efficacy.Why are guidelines that address the management of croup a good choice for testing implementation strategies?First, croup is an extremely common disease that is seenin small rural health hospitals as well as in tertiary centres.Second, a large body of published evidence clearly dem-onstrates that properly timed treatment with corticoster-oids can significantly reduce health care utilization,without any significant adverse events. Furthermore,understanding what constitutes best therapy is not diffi-cult to grasp. Third, it is a reasonable assumption that hos-pitalization is the primary determinant of health careexpenditures for the disease. Fourth, there is a substantialvariation in practice patterns and hospitalization ratesacross Alberta. Therefore effective implementation strate-gies should lower overall health care expenditures andincrease societal benefits.Relevant systematic reviewsStrategies for implementing practice guidelinesCochrane EPOC Group members have published two'overviews' of systematic reviews of knowledge translationinterventions.[1,2] In general, these systematic reviewsfound, that despite the relatively large number (~ 1,600)of published articles which address the topic, only a com-paratively small number (~ 100) have a semi-rigorousdesign (randomised, quasi-randomised, controlled beforeand after studies).[51] Many of these primary studies arePage 4 of 13(page number not for citation purposes)scores, randomized trials have shown a reduction in theduration of endotracheal intubation,[46] the rate of intu-methodologically flawed in that they randomised subjectsby clusters, but used standard statistical techniques forImplementation Science 2006, 1:10 http://www.implementationscience.com/content/1/1/10analyzing their results.[52] Only a comparatively smallnumber of trials have systematically examined the impactof implementation strategies on both medical practiceand clinical outcomes, and very few studies haveaddressed the cost effectiveness of clinical guidelines.[1]The Cochrane EPOC Group has established a taxonomyfor classifying different types of implementation strate-gies. Educational meetings (defined as participation ofhealth care providers in conferences, lectures, or work-shops) show a range of effectiveness.[1] Interactive work-shops have shown moderate to moderately large effects,whereas didactic forums alone were ineffective.[1] Educa-tional outreach visits, (defined as the use of a trained per-son who meets with providers in their practice settings toprovide information with the intent of changing provid-ers' performances), have shown small to moderate effectson behaviour, though the authors point out that the costeffectiveness is unclear. Reminders (defined as any inter-vention, manual or computerised, that prompts thehealth care provider to perform a clinical action) appearto be more effective than audit and feedback but theresults have not been striking.[1] The identification oflocal opinion leaders (defined as health professionalsnominated by their colleagues as 'educationally influen-tial') has shown mixed results in studies published todate, and the authors of this Cochrane systematic reviewsuggested that further research was required before thistechnique becomes widespread.[1,2,53] Probably themost consistent finding among systematic reviews is thatmulti-faceted interventions (> 2 interventions of any type)are more effective than single interventions.[1,2]Management of children with croupTwo meta-analyses of randomized controlled trials exam-ining the benefit of corticosteroids have been published.Kairys et al. published a meta-analysis in 1989 of 10 pub-lished randomized controlled trial's involving 1286patients.[4] The analysis indicated that the use of steroidsin children hospitalized with croup is associated with asignificantly increased proportion of children showingclinical improvement at 12 and 24 hours following treat-ment, and a significantly reduced incidence of endotra-cheal intubation. More recently, we (TPK and DWJ)performed a meta-analysis of 24 randomized controlledtrial's, and found that corticosteroid treatment was associ-ated with an improvement in croup score at 6, 12, and 24hours, a decrease in epinephrine treatments, a decrease inlength of time spent in the emergency department, and areduction in hospital stay by 16 hours.[3]) A Cochraneprotocol has been submitted to examine humidified airinhalation for treating croup, but has not been publishedyet. No systematic reviews have been published or regis-Aim and objectivesThe overall aim of this study is to identify, from a societalperspective, the costs and associated benefits of threestrategies (of low, intermediate and high intensity respec-tively) for disseminating and implementing a practiceguideline that addresses the management of croup.Primary objectiveTo determine which of the three intervention strategies aremost effective at lowering the rate of hospital days per1,000 disease episodes. The null hypothesis is that noneof the intervention strategies reduce hospital utilizationrates from baseline. The alternate hypothesis is that theintervention strategies will have a graded degree of effecton hospitalization rates, with the low intensity interven-tion having minimal to no effect, the intermediate inten-sity intervention having moderate but significant effect,and the high intensity intervention having the greatesteffect.Secondary objectiveTo determine which of the three intervention strategies aremost effective at increasing the use of therapies of knownbenefit.Other objectivesTo determine which intervention strategy will most effec-tively maintain or improve clinical outcomes and main-tain or reduce the family pyschosocial burden. Clinicaloutcomes assessed will include both uncommon severeevents, as well as average duration of clinical symptoms.The assessment of family psychosocial burden willinclude the number of hours of sleep missed by the child,and the stress experienced by the primary caregiver (mostcommonly the mother).Economic analysisTo determine which of the three intervention strategieswill most effectively reduce total societal costs includingall health care costs, costs borne by the family, and costsstemming from the strategies for disseminating guide-lines. The null hypothesis is that neither the intermediatenor the high intensity interventions will consume lessresources than the low intensity intervention. The alter-nate hypothesis is that the intermediate intervention willconsume fewer resources than either the low intensityintervention or the high intensity intervention.MethodsProposed trial designWe will conduct a cluster randomized controlled trial aftercompleting a baseline survey. In order to provide a com-prehensive answer as to whether or not our interventionsPage 5 of 13(page number not for citation purposes)tered for either epinephrine or heliox. are beneficial, we have obtained data from several differ-ent sources. The stages are as follows.Implementation Science 2006, 1:10 http://www.implementationscience.com/content/1/1/101. Baseline survey (currently near completion):▪ Utilizing administrative data, we have documentedAlberta health care utilization rates.▪ We are documenting severe adverse outcomes in Albertautilizing administrative data, and both medical examinerand records audit.▪ Using this data, we have rank ordered Alberta hospitalsbased on number of disease episodes and rates of hospi-talization.▪ We have enrolled the highest ranking 24 hospitals whoconsented to participation.▪ We are documenting practice patterns in these 24selected Alberta hospitals utilizing medical records audit.▪ Utilizing a prospective cohort questionnaire, we are doc-umenting the duration of clinical symptoms of childrenwith croup and the psychosocial burden on their families2. Development of clinical guidelines that address indica-tions for drug therapy and hospital admission/dischargecriteria. The process will include:▪ convening an guideline committee with a range of disci-plines and professions;▪ critically reviewing the published literature;▪ drafting guidelines which meet standard criteria fordeveloping guidelines;▪ review of the guidelines by both parents and a wide-range of users (family physicians, nurses, and respiratorytherapists);▪ obtaining approval of the guidelines from the CanadianPediatric Society, Canadian Association of EmergencyPhysicians, and the Alberta Medical Association; and▪ publishing the guidelines in journals such as the Cana-dian Medical Association Journal.3. Randomization of 24 Alberta hospitals to one of threeimplementation strategies.4. A follow-up survey whose purpose is to detect anychange from baseline by:▪ documenting severe adverse outcomes and utilizationrates in Alberta;▪ documenting practice patterns in the 24 Alberta hospi-tals surveyed at baseline;▪ documenting the duration of symptoms of children withcroup and the psychosocial burden on their familiesattending these 24 hospital emergency departments.Proposed sample sizeThe twenty-four participating hospitals were initiallyselected by calculating anticipated hospital specific effectsizes based on assuming that the effect of interventionwould be proportional to baseline admission rates takinginto account hospital-specific frequencies of disease epi-sodes. Hospitals are to be randomized to one of threeintervention arms, eight hospitals per arm, after stratifica-tion on baseline frequency of disease episodes. The powerof our study depends on the anticipated change in meanhospital days as shown in Figure 1, which provides powerestimates plotted against percentage decrease based onexamining hospital stays over the three baseline and threepost-intervention years. (Appendix A – Details of calcula-Illustration of Estimated Study PowerFigu e 1Illustration of Estimated Study Power.Power for Two Arm ComparisonPercent ChangePower15 20 25 30 350. 6 of 13(page number not for citation purposes)in the same 24 selected Alberta hospital emergencydepartmentstions) A Bonferroni correction for the two comparisons ofintermediate and high intensity intervention against lowImplementation Science 2006, 1:10 http://www.implementationscience.com/content/1/1/10intensity intervention has been included. The estimatedpower (overall alpha = .05) to detect a 25% change is 0.80for each comparison.RecruitmentRecruitment of hospitalsAll 24 hospitals have been approached and have con-sented to participate in our study.Recruitment of families to complete prospective surveyAs of Oct. 27th, 2002 we have recruited a total of 170 chil-dren into our baseline cohort and 95% have been con-tacted daily until resolution of symptoms. We willcontinue to recruit patients until April 30th, 2003. Givenour rate of enrolment to date, we anticipate enrollingbetween 450 and 500 patients in the baseline cohort. Theintervention and follow-up sample will be obtained overa total of 24 months, as opposed to the baseline in whichwe will have enrolled patients for only 13 months. There-fore, conservatively, enrolment should be ~ 900 patientsin the follow-up cohort.Practical arrangements for allocating participants to trial groupsProposed strata, rationale and method of allocationAs noted above, a total of 24 hospitals have agreed to par-ticipate in our study. We will stratify these hospitals intothree levels based on the average number of croup casesdiagnosed by affiliated physicians per year (< 100 cases/year = small volume hospitals; 100 to 700 cases per year =medium volume hospitals, and > 700 cases per year = large vol-ume hospitals). Non-stratified randomization of only 24hospitals into three intervention arms could easily resultin substantially more hospitals of one stratum in oneintervention arm than the other arms. Therefore, given thesignificant differences in total number of disease episodes,number of hospitalizations, hospitalization rates, andlikely other significant differences between the differentsized hospitals, non-stratified randomization could sig-nificantly bias our results. A statistician not otherwiseinvolved in our study will randomise hospitals withineach of the three strata to one of the three interventionarms by computer software.Planned trial interventionsDevelopment of croup clinical practice guidelinesDevelopment committeeThe Alberta Medical Association Clinical Practice Guide-line Program (Alberta Medical Association Clinical PracticeGuidelines Program) appointed DWJ as Chair, and thehealth care professionals representing pediatric emer-gency medicine (TPK, WC); pediatric pulmonary; pediat-ric infectious diseases; emergency medicine; rural familyCriteria for developing a practice guidelineWe have used both the criteria developed by Grimshawand Shaneyfelt as the basis for formulating our guide-lines.[1,2]Systematic review of the literatureUtilizing the Alberta Research Centre for Child HealthEvidence (ARCHE), we carried out a systematic literaturereview to ensure we had identified all primary publishedstudies addressing the management of children withcroup. We have completed a draft of the guidelines.Remaining steps prior to completion of croup clinical practice guidelineWhen the guidelines are in final form, a parents' groupwill critically review the guidelines for clarity and content.As well, the guidelines will be reviewed by a randomlyselected group of family physicians for clarity and clinicalrelevance. Once development of the guidelines is com-plete, we will seek endorsement by the Canadian PediatricSociety and the Canadian Association of Emergency Phy-sicians. We will also submit the guidelines to the Cana-dian Medical Association Journal for publication.The three intervention strategiesThe intervention strategies to be compared will be a) mail-ing of printed educational materials (PEMs) (low inten-sity intervention); b) mailing PEMs plus a combination ofinteractive educational meetings, educational outreachvisits, and reminders (intermediate intensity interven-tion); and c) mailing PEMs, interactive sessions, outreachvisits, reminders plus identification of local opinion lead-ers (high intensity intervention.Low intensity interventionThe low intensity intervention is the current strategy uti-lized by the Alberta Medical Association Clinical PracticeGuidelines Program for disseminating new guidelines. Thisintervention arm will serve as a "control". The CroupGuidelines will be mailed to each physician registeredwith the Alberta College of Physicians and Surgeons. TheGuidelines will also be available on the Alberta MedicalAssociation Clinical Practice Guidelines Website.Intermediate intensity interventionThe intermediate intensity intervention utilizes tech-niques that have consistently shown at least a moderateimpact on professional behavior. We will work throughexisting organizational structures in order to more closelysimulate how the Alberta Medical Association Clinical Prac-tice Guidelines Program might disseminate future guide-lines. More specifically, the Alberta Medical AssociationClinical Practice Guidelines program and the Croup clinicalPage 7 of 13(page number not for citation purposes)medicine; respiratory therapy; and nursing to a committeethat was charged with developing the guidelines.practice guideline committee will work with the Universi-ties of Calgary and Alberta Continuing Education Pro-Implementation Science 2006, 1:10 http://www.implementationscience.com/content/1/1/10grams to develop an interactive educational program forphysicians, nurses, and respiratory therapists. Utilizingthe existing Continuous Quality Improvement Staff ineach Health Region, each hospital's admitting physicians,acute care nurses and respiratory therapists will be invitedto participate in the interactive educational program car-ried out by one of the Croup Clinical Practice GuidelineCommittee members. A "local champion" will also beidentified. They will conduct educational outreach visitsto physicians, nurses, and respiratory therapists who didnot attend the interactive educational program.High intensity interventionThe high intensity, intervention utilizes a multi-facetedapproach described by Lomas.[3] This approach uses avariety of techniques, including interactive educationalmeetings, educational outreach visits, reminders, identifi-cation of local opinion leaders and establishment of alocal consensus process. The physician-focused multi-fac-eted dissemination strategy will involve identification ofphysician "opinion leaders" at each hospital randomizedto this arm. All physicians with admitting privileges to ahospital will be asked to complete a validated question-naire for the purpose of identifying the "local opinionleader".[4] All "opinion leaders" will be invited to partic-ipate in a workshop on the evidence for the practice guide-line's recommendations and on basic principles ofbehaviour change.Planned inclusion/exclusion criteriaHealth care utilization using administrative databasesAlberta Health and Wellness, a branch of the Alberta pro-vincial government, is the custodian of several datasources that are accessible for research purposes. Thesedatabases can be linked, and are considered to be of goodquality with reliable personal identifiers on more than95% of records. Through collaboration with the ProjectCoordinator, Health Surveillance Branch, Alberta Health& Wellness, we extracted the data specified below from theCanadian Institute for Health Information Hospital Inpa-tient database, Alberta Ambulatory Care ClassificationSystem, the Alberta Health Care Insurance Plan (AHCIP)Payment Database, and the AHCIP Registry Dataset.All children 0–6 years of age and 6–16 years of age whoare Albertan residents, and who have been evaluated anddiagnosed by an Albertan physician to have croup (ICD-9-CM 464.1 Acute Tracheitis, 464.2 Acute Laryngotracheitis,and 464.4 Croup, either as a primary or secondary diagnosis)have been included in our utilization data.Methodological assignment of physicians to hospitalEncrypted physician identifiers are available for eachset were assigned to a specific hospital based on following:first, the hospital to which they admitted children withcroup (62% of physicians); second, the hospital clinic oremergency department at which they diagnosed childrento have croup (82% of physicians); or, third, the hospitalto which a physician associate admitted or evaluated chil-dren with croup (93% of all physicians). (Physicians whoshared the same billing address where assumed to be asso-ciates.)Adverse outcomesWe have used administrative databases, selected medicalrecord audit, and review of medical examiner cases toestablish Alberta morbidity and mortality rates for the six-year baseline. Data gathered from these three sources willallow us to establish the incidence of rare but severe out-comes such as death, anoxic brain injury, and endotra-cheal intubation.Administrative databases were screened for any childrenwith a diagnosis of croup who died, were intubated (ICD-9-CM 96.04), or were also diagnosed during the sameadmission to have anoxic brain damage (ICD-9-CM348.1) or cerebral edema (ICD-9-CM 348.5).Retrospective record auditWe are auditing the medical records of children identifiedfrom the administrative databases who met the above cri-teria. The baseline review is almost complete, and we willrepeat the process in follow-up. To double check that ourreliance on administrative databases has not resulted inmissed patients, we have also reviewed the manual admis-sion logs for Alberta Children's Hospital and the Univer-sity of Alberta Hospital intensive care units, and willcontact the medical records departments at each of theregional medical centers (Medicine Hat, Lethbridge, RedDeer, Fort McMurray, Grande Prairie).Medical examiner's officeWe have also contacted the Alberta Medical Examiners'Offices and have arranged for review of children ≤ 16years of age that died of "respiratory causes".Practice patterns using retrospective medical record auditRetrospective record audit will be used to establish theextent to which physicians use effective therapies forcroup. The baseline review is almost completed, and willbe repeated for the follow-up period. Each of the 24 tar-geted hospitals were asked to generate a list of medicalrecord numbers using standard ICD9 codes as above thatidentify all children that were hospitalized and/or evalu-ated in the emergency department with croup. Medicalrecords pertaining to all hospitalized children were (willPage 8 of 13(page number not for citation purposes)health care encounter in which a child was diagnosed tohave croup. Each of the physicians appearing in our data-be) reviewed. Of those children evaluated in the emer-gency department, a maximum of 30 patients per yearImplementation Science 2006, 1:10 http://www.implementationscience.com/content/1/1/10were randomly selected and reviewed. For the follow-uprecord audit, since significant annual change is likely tooccur, we will audit up to a maximum of 100 records toallow greater accuracy for our annual estimates.Clinical outcomes & psychosocial burden using prospective questionnaireThe parents of all children evaluated in, or admitted toeach of the 24 hospitals with a diagnosis of croup betweenSeptember and May during the baseline, intervention, andfollow up periods will be approached to participate in afollow up telephone survey by a staff nurse. A log of allchildren diagnosed to have croup will be maintained toallow the comparison of those children enrolled to thosenot enrolled.Identification & enrolment of participating hospitalsAs noted above, Alberta hospitals were rank ordered basedon the number of disease episodes and rates of hospitali-zation for a six-year period. We then approached in thisorder each of the hospital administrators and clinical stafffor permission to include their hospital in our study untila total of 24 hospitals consented (the staff of 11 hospitalsrefused to participate for reasons ranging from "antici-pated closing of the hospital in the near future" to "tooheavy an administrative workload to participate in astudy").Proposed duration of treatment periodDetailed planning for the intervention strategies will takeplace between February and July 2003. We will initiateeach of the interventions in August, 2003 and concludethem in March, 2004.Duration of the baseline and follow-up periodsUtilization and adverse outcome data obtained from the administrative datasetsWe have extracted data from administrative databasesfrom April 1st, 1994 to March 31st, 2000, providing sixyears of data. In summer 2006, we will again extract datafrom administrative databases from April 1st, 2000 toMarch 31st, 2006. Extraction of data across these time peri-ods will allow us to extend the baseline utilization ratesfrom six to nine years. Most importantly, this will ensureno significant changes occur just prior to the study inter-ventions. Extraction of data through March 31st, 2006 willprovide utilization data for the intervention and two fol-low-up years.Practice pattern data obtained from medical record auditWe are just completing an audit of medical records of chil-dren diagnosed to have croup based on ICD9 coding fromApril 1st, 1994 to March 31st, 2000. By late 2006, we willMarch 31st, 2006. This timetable for extraction of dataprovides the above delineated advantages.Clinical outcome & psychosocial burden data obtained from prospective questionnaireDuration of enrolmentThe parents of children admitted to each of the 24 hospi-tals with a diagnosis of croup during "croup season" (Sept1st to May 31st) from November 1st, 2001 and finishingMarch 31st, 2006, will be asked by a staff nurse to partici-pate in a telephone survey.Duration & mechanism of telephone follow upThe enrolment and consent forms are then faxed to thestudy coordinator (JW). She assigns the case to one ofthree centrally located and trained study investigatorswho will contact the primary caretaker within 24 hours ofenrolment. The initial telephone interview takes 20 min-utes. Daily telephone interviews with the primary care-taker occur until the child has been symptom-free for 24hours.Proposed primary and secondary outcome measuresThe primary outcome will be the rate of hospital days per1,000 disease episodes. If differences in rate of hospitaldays between intervention groups are detected, we willexplore to what degree they are due to a change in rates ofadmission versus a change in hospital lengths of stay.The secondary outcomes will be utilization of appropriatetherapies including:▪ Proportion of patients treated in the emergency depart-ment and hospital with a corticosteroid.▪ Proportion of patients evaluated in emergency depart-ment for at least three hours after treatment with corticos-teroids before the decision to admit to hospital is made.▪ Time to treatment with corticosteroids in both emer-gency department and hospital patients.Other outcomesClinical outcomesTo determine which dissemination strategy will mosteffectively maintain or improve clinical outcomes:▪ by reducing or maintaining the number of uncommon,severe clinical events such as intubation, respiratory andcardiac arrest, anoxic brain injury, or death.▪ by reducing, (in a prospective cohort of childrenenrolled from each hospital) croup symptoms on days 1,Page 9 of 13(page number not for citation purposes)repeat an identical audit of medical records of childrendiagnosed as having croup between April 1st, 2000 to2, and 3 following assessment, as measured by a tele-phone follow-upImplementation Science 2006, 1:10 http://www.implementationscience.com/content/1/1/10▪ assessment tool (Telephone Out Patient (TOP) score).Psychosocial burdenTo determine which dissemination strategy will mosteffectively maintain or reduce the psychosocial burden onparents by▪ reducing, (in a cohort of children enrolled from eachhospital), the stress experienced by the primary caregiverdue to the child's illness on days 1–3 following assess-ment.▪ reducing, (in a prospective cohort of children enrolledfrom each hospital), the number of hours of sleep missedby the child due to croup on days 1–3 following assess-ment.How will the outcome measures be measured at follow up?Utilization outcomesPreviously discussed.Adverse outcomesPreviously discussed in part.Development of data abstraction sheetThe Principal Investigator and Study Coordinator, follow-ing an informal review of medical records, developed acustomized Access relational database. This draft databasewas pilot tested and revised before beginning formalreview.Auditing processBecause these cases are significantly more complex, onlythe study coordinator and another trained senior pediatricemergency nurse will perform these audits. The principalinvestigator and the two nurses will meet regularly duringthe baseline and follow up audits, to discuss cases and anydifficulties encountered in auditing. To establish thedegree of agreement between auditors, the three willreview a random selection of medical records.Case identification: multiple and overlapping methodsIn order to minimize the likelihood that any adverse out-comes will be missed, overlapping methods for identify-ing cases will be used. Specifically, we will obtainprovincial medical examiner records, and hospital admin-istrative data from the province, all the health regions, andall the hospitals in urban areas.Practice pattern outcomesDevelopment of data abstraction sheetAfter an informal review of medical records, a customizedAccess relational database was developed. This draft dataTraining & monitoring of medical record auditorsThe projector coordinator has trained two research assist-ants to review medical records and directly enter data in arelational database (Access). The project coordinator andeach of the research assistants will review a randomlyselected number of records through the study to checkaccuracy in data abstraction.Accuracy of ICD9 codingTo assess the accuracy of using ICD9 coding to identifychildren diagnosed with croup, we will examine whetherhealth record analysts coded records in the same way ateach of the 24 hospitals. We will determine what percent-age of children were coded as croup using the standardcodes met a formal definition for croup (acute onset ofstridor associated with a "seal-like" barking cough). Alsoto explore whether some hospitals use other ICD9 codesfor children with croup, we will generate a list of potentialalternative codes. Using this list of potential alternativeICD9 codes, we will review a random selection of medicalrecords using these codes.Clinical & psychosocial burden outcomesRMO databaseWe are maintaining a database of all children diagnosedas having croup whose parents refused, were missed, orare otherwise excluded. Demographics and disease char-acteristics of children and their families in this databasewill be compared to those families who are administeredthe survey to ascertain if the two populations are different.Telephone interviewing techniquesSeveral procedures should enhance the reliability of tele-phone data collection. a) One study investigator will com-plete all telephone follow up for a given family. b) At thetime of enrolment the primary caretaker of the child willbe identified, and every effort will be made to conduct allfollow-up calls with this person. c) The questions willhave standardized responses. d)A standard audio record-ing of a child with croup will be played to the caregiver toaid them in identifying the easily recognizable and dis-tinctive cough.Proposed analysesThe principal analysis will examine hospital stays arisingfrom individual disease episodes over the six-year base-line and study period. We will apply a linear mixed model(Laird-Ware approach) incorporating random effects forhospital, year within hospital, as well as fixed effects forintervention and year. We will also pursue a more detailedexplanatory analysis incorporating individual subjectcharacteristics (age, sex) as well as random effects foradmission and discharge physician. Additionally we willPage 10 of 13(page number not for citation purposes)sheet was piloted and revised, before beginning formalreview.develop similar models for length of stay restricting to dis-ease episodes leading to admissions, and mixed effectsImplementation Science 2006, 1:10 http://www.implementationscience.com/content/1/1/10logistic regression model for admissions. Secondary anal-yses relating to steroid use and Telephone OutPatient (TOP)scores will follow a similar format. The examination ofday 1- 3 TOP scores will require augmenting the basicmodel framework to include random subject effects and aserially correlated (within subject) residual error structure.No subgroup analyses are planned.Economic analysisAn economic analysis will be carried out from a societalperspective with costs classified as either payer (costs bornby the province) or non-payer (costs born by individualsor the families of children with croup). The analysis willinclude all health care costs, costs borne by the family,and costs stemming from the strategies for implementingguidelines.Potential benefitsPotential benefits will be considered to be the duration ofcroup symptoms and the absence of any severe adverseoutcomes among children.Overview of costing methodsCosting information will be gathered in sufficient detail toallow a comparison of costs and benefits. The form thatthe analysis will take will depend on the relative ratios ofcost versus benefit. For example, if either of the two exper-imental strategies (non-standard) achieve similar or betterclinical outcomes at decreased or equal costs as the stand-ard dissemination strategy, the experimental interven-tions will be judged as cost-effective. Alternatively, if theexperimental interventions yield increased benefits atincreased costs, a ratio of the incremental cost per out-come gained will be calculated. And last, if the experimen-tal interventions yield decreased benefits but areassociated with increased or equal costs, the standard dis-semination strategy will be judged to be the most cost-effective. (Please note we have specifically avoided usingany reference to either cost-effective analysis or cost-bene-fit analysis. [5,6])Health care costsPayer costs resulting from the provision of health care willbe estimated by first determining from the administrativedatabase the total number of physician visits, emergencydepartment visits, days of hospitalization, patient trans-ports, and days in an intensive care unit (Edmonton orCalgary only). Costs per physician visit will be based onthe Alberta physician fee schedule. Costs per emergencydepartment visit, and hospital day will be based on pub-lished standards for Alberta.[7] Costs for drugs, radio-graphs, and laboratory tests will be based on provincialstandards. For the Edmonton and Calgary hospitalsthereby allowing a unique cost estimate for each patient.Air transport costs will be estimated to be $8 per air mile.Implementation strategy costsCosts of the three-implementation strategies will be clas-sified as either as payer and non-payer based on whetheror not the province assumes responsibility for the cost.They will be estimated from the costs of publication andmailing of questionnaires or learning materials, and thetime spent in meetings, organising, travel, and preparingfor presentations by the Alberta Medical Association sup-port staff, Regional Quality Assurance staff, "local cham-pions", opinion leaders, and any associates. (The hourlywage will be assumed to be $100 per hour for physiciantime and to be the mean provincial wage for other healthcare professionals and administrators.)Costs to the familyThe costs borne by families will be considered to be non-payer. The amount of work missed (including both wagesand house work), travel time, parking and ambulancecosts by each family will be assumed to equal the meanvalues determined by the prospective survey of childrenwith croup. The value of the lost wages will be estimatedby applying a mean wage rate to missed work time,obtained from STATS Canada for Alberta. The values fortravel time, parking costs, and ambulance costs will betaken from published standards for Alberta.[7]Calculation of societal costsTotal societal costs (the sum of payer and non-payer costs)for each hospital will be calculated for the 9 baselineyears, the one-year of intervention, and the two years offollow-up. Given the year-to-year variability in the inci-dence of croup, the total societal costs for each year will beadjusted for disease incidence (total costs per year dividedby the number of episodes of croup diagnosed that yearby those physicians assigned to the hospital). An averagecost per year for the baseline period will be calculated. Sta-tistical assessment of whether or not costs differ betweenthe three intervention groups will be calculated by theLaird-Ware model with baseline costs as the covariate.(This type of analysis will allow us to simultaneouslycompare costs at three time points (the intervention yearand two years of follow-up), while adjusting for the base-line cost)Quality of Life (QOL) measurementsWe will not utilize any QOL measures in our study. None,to our knowledge, have been developed for use in chil-dren with acute, transient illnesses.Page 11 of 13(page number not for citation purposes)included in the randomised trial and for any intensivecare patients, specific hospital case costing is availableImplementation Science 2006, 1:10 http://www.implementationscience.com/content/1/1/10Data safety and monitoring committeeGiven that the study intervention does not pose any directrisk to patients or health care professionals, a Data Safetyand Monitoring Committee is not necessary.Proposed methods for protecting against sources of biasMasking of allocationAs noted above, a statistician not aware of our study ques-tion will allocate hospitals using a standard computer ran-domization software.Cluster randomizationGiven that our study interventions specifically target phy-sicians and other health care workers, one threat to thevalidity of our study is the possibility of communicationbetween health care workers randomized to differentintervention arms. To minimize the likelihood of this pos-sibility, we have chosen hospitals as our unit of randomi-zation and analysis. All hospitals but three are located indifferent cities and towns, each with a significant distancebetween them. However, three of the six hospitals withlarge volumes are located in Edmonton, so at least two ofthe Edmonton hospitals will be randomized to differentintervention arms.Inability to completely blind health care workers to study intentAnother potential threat to the validity of our study is that– as a result of needing to approach hospital clinical staffto obtain consent for their hospital to be included in ourstudy – hospital staff had to be informed of the intent ofthe study. We have tried to minimize further unnecessarycontamination, however, by instructing our study staff toavoid discussing the study's main purpose with hospitalstaff, and speaking only about the operational details oftheir particular intervention.Assessment of outcomesThe assessment of outcome measures will be masked asmuch as possible without awareness of study purpose orintervention assignment. The data analyst responsible forextracting and linking the administrative databases willnot be informed as to specific purpose of the study nor towhich intervention arm each hospital has been rand-omized. Medical record auditors will not be informed asto the specific purpose of the study nor to which interven-tion arm each hospital has been randomized. Parents ofchildren with croup will not be told of the overall purposeof the study nor the intervention arm to which their hos-pital is randomized.AnalysisThe health economist (CM) will review and clean the eco-nomic data without being aware of the specific hospital or(RB) will be provided with data segregated into the threeintervention arms labelled with an encrypted hospital andintervention identifier.Competing interestsThe author(s) declare that they have no competing inter-ests.Authors' contributionsDWJ conceived of the idea for the study, devised the studydesign, and wrote the majority of study protocol. WCreviewed the study protocol and provided critical feed-back. RB wrote the statistical analysis section of the proto-col. CM wrote the health economic section of theprotocol. LS provided advice on the availability of admin-istrative data from Alberta Health & Wellness. TKreviewed the study protocol and provided critical feed-back.Additional materialReferences1. Grimshaw J, Shirran L, Thomas R, Mowatt G, Fraser C, Bero L, GrilliR, Harvey E, Oxman A, O'Brien MA: Changing provider behavior:an overview of systematic reviews of interventions.  Med Care2001, 39(8 Suppl 2):2-II.2. Bero L, Grilli R, Grimshaw J, Harvey E, Oxman AD, Thomson MA:Getting research findings into practice: Closing the gapbetween research and practice: an overview of systematicreviews of interventions to promote the implementation ofresearch fundings.  BMJ 1998, 317:465-468.3. Ausejo M, Saenz A, Pham B, Kellner JD, Johnson DW, Moher D, Klas-sen TP: The effectiveness of glucocorticoids in treating croup:meta-analysis.  BMJ 1999, 319:595-600.4. Kairys S, Olmstead EM, O'Connor GT: Steroid treatment oflaryngotracheitis: a meta-analysis of the evidence from ran-domized trials.  Pediatrics 1989, 83(5):683-693.5. Johnson DW, Jacobson S, Edney PC, Hadfield P, Mundy ME, Schuh S:A comparison of nebulized budesonide, intramuscular dex-amethasone, and placebo in moderately severe croup.  N EngJ Med 1998, 339:498-503.6. Wennberg JE: Small area analysis and the medical case out-come problem.  In Conference Proceedings Research Methodology:Strengthening Causal Interpretations of Nonexperimental Data Rockville,MD: Agency for Health Care Policy and Research, US Department ofHealth; 1990. 7. To T, Dick P, Young W: Hospitalization rates of children withcroup in Ontario.  Paediatr Child Health 1990, 1:103-108.8. Haynes R: Some problems in applying evidence in clinicalpractice.  Ann NY Acad Sci 1993, 703:210-224. discussion 224–2259. Grimshaw J, Hutchinson A: Clinical practice guidelines–do theyenhance value for money in health care?  Br Med Bull 1995,51(4):927-940.10. Grimshaw J, Thomson M: What have new efforts to change pro-Additional File 1Appendix I – Power CalculationsClick here for file[http://www.biomedcentral.com/content/supplementary/1748-5908-1-10-S1.pdf]Page 12 of 13(page number not for citation purposes)intervention (for this process, hospital and interventionidentifiers will be removed). Furthermore the statisticianfessional practice achieved?  J R Soc Med 1998, 91(Suppl35):20-25.Implementation Science 2006, 1:10 http://www.implementationscience.com/content/1/1/1011. Haynes R: Of studies, summaries, synopses, and systems: the"4S" evolution of services for finding current best evidence.Evid Based Ment Health 2001, 4:37-39.12. Gallagher E: How well do clinical practice guidelines guide clin-ical practice?  Ann Emerg Med 2002, 40(4):394-8.13. Field M, Lohr K, eds: Clinical Practice Guidelines: Directions ofa New rogram.  Institute of Medicine: Washington, DC;1990:1-160. 14. Eddy D: A Manual for Assessing Health Practices and Design-ing Practice Policies: The Explicit Approach.  American Col-lege of Physicians: Philadelphia, PA; 1992:1-126. 15. Hayward R, Wilson MC, Tunis SR, Bass EB, Rubin HR, Haynes RB:More informative abstracts of articles describing clinicalpractice guidelines.  Ann Intern Med 1993, 118(9):731-737.16. Cluzeau F, Littlejohns P, Grimshaw JM, Feder G, Moran SE: Develop-ment and application of a generic methodology to assess thequality of clinical guidelines.  Int J Quality Health Care 1999,11(1):21-28.17. Grimshaw J, Eccles M, Russell I: Developing clinically valid prac-tice guidelines.  J Eval Clin Pract 1995, 1(1):37-48.18. Leone A: Medical practice guidelines are useful tools in litiga-tion.  Med Malpractice Law Strategy 1993, 10:1-6.19. Shaneyfelt , Mayo-Smith TM, Rothwangl J: Are guidelines followingguidelines? The methodological quality of clinical practiceguidelines in the peer-reviewed medical literature.  JAMA1999, 281(20):1900-1905.20. Denny FW, Collier AM, Henderson FW, Clyde WA Jr: Infectiousagents of importance in airways and parenchymal diseases ininfants and children with particular emphasis on bronchioli-tis.  Pediatr Res 1977, 11:234-236.21. Denny FW, Murphy TF, clyde WA Jr, Collier AM, Henderson FW:Croup: an 11-year study in a pediatric practice.  J Pediatr 1983,71(6):871-876.22. Marx A, Torok TJ, Holman RC, Clarke MJ, Anderson LJ: Pediatrichospitalizations for croup (laryngotracheobronchitis): bien-nial increases associated with human parainfluenza virus 1epidemics.  J Infect Dis 1997, 176:1423-1427.23. Henrickson KJ, Kuhn SM, Savatski LL: Epidemiology and cost ofinfection with human parainfluenza virus types 1 and 2 inyoung children.  Clin Infect Dis 1994, 18:770-779.24. McConnochie KM, Hall CB, Barker WH: Lower respiratory tractillness in the first two years of life: epidemiologic patternsand costs in a suburban pediatric practice.  Am J Public Health1988, 1988:34-39.25. Health Records: Emergency department visits and hospitalizations in Cal-gary Regional Health Authority Hospitals with a discharge diagnosis ofcroup. FY1992/3-1996/7 Calgary Regional Health Authority: Calgary,Alberta. 26. McEniery J, Gillis J, Kilhan H, Benjamin B: Review of intubation insevere laryngotracheobronchitis.  Pediatrics 1991, 87:847-853.27. Wagener JS, Landau LI, Olinsky A, Phelan PD: Management of chil-dren hospitalized for laryngotracheobronchitis.  Pediatr Pulmo-nol 1986, 2:159-162.28. Chin R, Brown GJ, Lam LT, McCaskill ME, Fasher B, Hort J: Effective-ness of a croup clinical pathway in the management of chil-dren with croup presenting to an emergency department.  JPaediatr Child Health 2002, 38:382-387.29. Langley JM, Wang EE, Law BJ, Stephens D, Boucher FD, Dobson S,McDonald J, MacDonald NE, Mitchell I, Robinson JL: Economicevaluation of respiratory syncytial virus infection in Cana-dian children: a Pediatric Investigators Collaborative Net-work on Infections in Canada (PICNIC) study.  J Pediatr 1997,131(1):113-117.30. Osmond M: Croup.  Clin Evid 2002, 7:297-306.31. Brown JC: The management of croup.  British Medical Bulletin2002, 61:189-202.32. Lavine E, Scolnik D: Lack of efficacy of humidification in thetreatment of croup: why do physicians persist in using anunproven modality?  Can J Emerg Med 2001, 1:209-212.33. Bourchier , Dawson DK, Fergusson D: Humidification in viralcroup: a controlled trial.  Aust Paediatr J 1984, 20:289-291.34. Lenney W, Milner AD: Treatment of acute viral croup.  Arch DisChild 1978, 53:704-706.35. Wolfsdorf J, Swift D: An animal model simulating acute infec-36. Neto G, Kentab O, Klassen TP, Osmond M: A randomized con-trolled trial of mist in the acute treatment of moderatecroup.  Acad Emerg Med 2002, 9(9):873-879.37. Westley C, Ross CEK, Brooks JG: Nebulized racemic epine-phrine by IPPB for the treatment of croup.  Am J Dis Child 1978,132(May):484-487.38. Taussig LM, Castro O, Beaudry PH, Fox WW, Bureau M: Treat-ment of laryngotracheobronchitis (croup).  Am J Dis Child 1975,129:790-793.39. Corkey CWB, Barker GA, Edmonds JF, Mok PM, Newth CJL: Radi-ographic tracheal diameter measurements in acute infec-tious croup: an objective scoring system.  Crit Care Med 1981,9(8):587-590.40. Fanconi S, Burger R, Maurer H, Uehlinger J, Ghelfi D, Muhlemann C:Transcutaneous carbon dioxide pressure for monitoringpatients with severe croup.  J Pediatr 1990, 117:701-705.41. Adair JC, Ring WH, Jordan WS, Elwyn RA: Ten-year experiencewith IPPB in the treatment of acute laryngotracheobronchi-tis.  Anesth Analg 1971, 50:649-655.42. Butte MJ, Nguyen BX, Hutchison TJ, Wiggins JW, Ziegler JW: Pedi-atric myocardial infarction after racemic epinephrineadministration.  Pediatrics 1999, 104:e9.43. McGee DL, Wald DA, Hinchliffe S: Helium-oxygen therapy in theemergency department.  J Emerg Med 1997, 15:291-296.44. Weber JE, Chudnofsky CR, Younger JG, Larkin GL, Boczar M, Wilk-erson MD, Zuriekat GY, Nolan B, Eicke DM: A randomized com-parison of helium-oxygen mixture (Heliox) and racemicepinephrine for the treatment of moderate to severe croup.Pediatrics 2001, 107:e96.45. Terregino CA, Nairn SJ, Chansky ME: The effect of Heliox oncroup: a pilot study.  Acad Emerg Med 1998, 5:1130-1133.46. Tibballs J, Shann FA, Landau LI: Placebo-controlled trial of pred-nisolone in children intubated for croup.  Lancet 1992,340:740-748.47. Godden CW, Campbell MJ, Hussey M, Cogswekk JJ: Double blindplacebo controlled trial of nebulised budesonide for croup.Arch Dis Child 1997, 76:155-158.48. Klassen T, Feldman ME, Watters LK, Sutcliffe T, Rowe PC: Neb-ulized budesonide for children with mild-to-moderate croup.N Engl J Med 1994, 331(5):285-289.49. Geelhoed GC, Turner J, Macdonald WBG: Efficacy of a small sin-gle dose of oral dexamethasone for outpatient croup: a dou-ble blind placebo controlled clinical trial.  BMJ 1996,313:140-142.50. Hampers LC, Faries SG: Practice variation in the emergencymanagement of croup.  Pediatrics 2002, 109:505-508.51. Grimshaw J, Russell I: Effect of clinical guidelines on medicalpractice: a systematic review of rigorous evaluations.  Lancet1993, 342(8883):1317-1322.52. Campbell M, Grimshaw J: Cluster randomised trials: time forimprovement: the implications of adopting a cluster designare still largely being ignored.  BMJ 1998, 317:1171-1172.53. Lomas J, Enkin M, Anderson GM, Hannah WJ, Vayda E, Singer J: Opin-ion leaders vs audit and feedback to implement practiceguidelines.  JAMA 1991, 265:2202-2207.54. Hiss RG, MacDonald R, David WR: Identification of physicianeducational influentials in small community hospitals.  ResMed Educ 1978, 17:283-288.55. Donaldson C, Shackley P: Economic evaluation, in Oxford Textbook ofPublic Health Edited by: Detels R, et al.. Oxford University Press:Oxford; 1997. 56. Donaldson C, Currie G, Mitton C: Cost effectiveness analysis inhealth care: contraindications.  BMJ 2002, 325:891-894.57. Jacobs P, Bachynsky J: An Alberta standard cost list for healtheconomics evaluations.  Institute of Health Economics: Edmonton,Alberta; 1997. Page 13 of 13(page number not for citation purposes)tive upper airway obstruction of childhood and its use in theinvestigation of croup therapy.  Pediatr Res 1978, 12:1062-1065.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items