JAC Advance Access originally published online on December 6, 2007
Journal of Antimicrobial Chemotherapy 2008 61(2):442-451; doi:10.1093/jac/dkm462
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Original research |
Identifying barriers to the rapid administration of appropriate antibiotics in community-acquired pneumonia
1 Castle Hill Hospital, Hull and East Yorkshire Hospitals NHS Trust, Cottingham, East Yorkshire, UK 2 Ninewells Hospital and Medical School, Tayside University Hospitals NHS Trust, Dundee, UK 3 Health Informatics Centre, Mackenzie Building, Kirsty Semple Way, University of Dundee, Dundee, UK
* Corresponding author. E-mail: p.g.davey{at}chs.dundee.ac.uk
Received 18 July 2007; returned 11 October 2007; revised 21 October 2007; accepted 29 October 2007
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
Background: Despite multiple guidelines for management of community-acquired pneumonia (CAP), barriers to guideline use are rarely evaluated.
Methods: We performed quantitative and qualitative surveys of junior doctors before implementation of a CAP management pathway. After implementation, we identified patient-related determinants of pathway adherence by multivariate analysis.
Results: We surveyed 83 (77%) of the 108 junior doctors working in acute medicine between August 2001 and July 2002 and selected 8 for in-depth interview. We identified five main themes that influence pathway adherence. First, education (recognized to be insufficient on antimicrobial therapy) and experience: increasing clinical experience was associated with greater knowledge of pathway content, but decreasing likelihood of consulting the pathway. Second, attitudes to CAP: doctors recognized that they had not treated CAP with respect early on. Third, work intensity and lack of senior support were barriers to good practice. Fourth, guideline factors: they need to be simple enough to be easy to use while containing enough information to be useful. Fifth, CAP is sometimes difficult to diagnose on admission. Notably, when given three clinical scenarios only six (7%) of respondents assessed CAP severity correctly. In the intervention study, early administration of antibiotics was associated (P < 0.05) with indicators of increased severity of illness (pulse, respiratory rate, oxygenation and temperature) in addition to being admitted to the intervention hospital (P < 0.001).
Conclusions: Some of the identified barriers could be overcome by undergraduate and postgraduate education. However, equally important are organizational barriers that can only be overcome by systems redesign.
Keywords: quality improvement , implementation , guidelines , care pathway
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
Several studies have demonstrated wide variation in clinical practice in the management of hospitalized patients with community-acquired pneumonia (CAP).1–3 This has occurred despite the development of national and specialist society guidelines for the management of CAP over the last 15 years.4–6 We wanted to improve processes of care observed in a previous study of patients admitted to hospital in Dundee with CAP.3 At the time that we planned the study there were few data, and none from the UK, about why hospital physicians do not adhere to recommended guidelines for the management of CAP or other antibiotic policies. Our primary objective was to increase the proportion of patients receiving appropriate antibiotics within 4 h of admission to hospital. We designed an intervention based on a systematic review, which identified seven major barriers to clinicians' adherence to guidelines under three main headings: knowledge, attitudes and behaviour.7 In addition to these barriers, patient and/or environmental factors may also prevent adherence. It is recommended that such barriers are evaluated prior to guideline implementation,8 although less than a quarter of evaluations include any pre-implementation assessment of barriers to change.9
In order to design our intervention, we identified barriers to the early delivery of appropriate antibiotics with two pre-implementation surveys of junior and middle grade doctors. We also collected post-implementation qualitative field-notes. In a controlled before and after analysis, we showed that implementation of a management pathway using posters, small group educational sessions and feedback resulted in a significant improvement (from 33% to 56%) in the proportion of patients receiving appropriate antibiotics within 4 h of admission to hospital.10 We now present analyses of doctor and patient characteristics that may have influenced the success of our intervention in order to provide a fuller understanding of the determinants of appropriate antibiotic prescribing in hospitalized patients. We have compared our data with the results of five recently published studies about clinician adherence to antimicrobial guidelines from Canada,11 the Netherlands,12,13 Republic of Ireland14 and Spain.15
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
Ethics
The protocols for qualitative and quantitative interviews and the intervention study were approved by the Tayside Local Research Ethics Committee. Written informed consent was obtained from all interviewees.
A detailed description of the study has been published.10 A controlled before and after study was performed. Baseline data were collected from all CAP patients admitted to an intervention and a control site between 1 November 2001 and 31 April 2002. The intervention [management pathway, available as Supplementary data at JAC Online (http://jac.oxfordjournals.org/)] was implemented in September and October 2002 and post-intervention data were collected between 1 November 2002 and 31 April 2003. To inform the design and implementation of our intervention, a structured survey and semi-structured (in-depth) interviews of medical staff were performed before implementation. To inform the feedback process and help us understand our results, qualitative field-notes were collected and univariate and multivariate analyses performed after implementation. The methods for each of these studies are detailed below.
A structured survey was administered to two consecutive cohorts of junior and middle grade medical staff working at Ninewells Hospital, Dundee between August 2001 and July 2002. To optimize response rate, prevent revision and ensure understanding of the questions, the survey was administered in-person. Questions included physician-related demographic data and open, 5-point ordinal (Likert) scale16 and ‘real-life’ clinical scenario questions.16a Definitions of appropriate antibiotic therapy and adverse prognostic indicators were taken from the 2001 BTS CAP guidelines.5 Data were analysed with SPSS for Windows. Results are presented as descriptive statistics. Associations between variables were analysed using a variety of non-parametric tests (
2 test, 2 for trend, Spearman's rank correlation and Kendall's rank correlation) depending on the nature of the data (i.e. nominal, ordered nominal, ordinal, dichotomous etc.). Statistical significance was defined as a P value of 
0.05.
Semi-structured (in-depth) interviews
The aims of the semi-structured interviews were:
- To identify reasons why hospitalized patients with CAP receive delayed and/or inappropriate antibiotic therapy.
- To identify existing barriers to adherence with available local and national CAP guidance.
- To identify potential barriers to the successful implementation of a CAP management pathway.
Because previous experience is likely to be an important influence on use of guidelines, respondents were sampled according to the primary dimension of grade (junior house officers, inexperienced senior house officers and experienced senior house officers/specialist registrars). Secondary dimensions included factors likely to influence experiences and views (e.g. non-Dundee medical school graduate and previous infectious disease and/or respiratory medicine experience) and responses to the quantitative survey. Potential respondents were approached in person or by telephone. At the time of interview, respondents were aware that a quality improvement project for CAP was in development, but they had not been exposed to any educational materials. Written consent was obtained prior to interview.
Interviews were semi-structured according to a pre-developed topic guide. This was based on the research questions, Cabana's model7 and the experience and knowledge of the researchers. The following topics were covered: previous CAP training experiences; awareness/familiarity with local and national CAP guidance; attitudes towards guidelines; external barriers to guideline adherence; the respondent's previous experiences of managing CAP; and the reasons why CAP patients receive delayed or inappropriate antibiotic therapy. This guide was adapted between interviews in light of emerging themes and the success of certain questions. Interviews were tape-recorded and lasted 
1 h. Tape-recorded interviews were transcribed verbatim and analysed using ‘Framework’ methodology.17
During the collection of the quantitative data for the main evaluation of the project, qualitative field-notes were made on the same data collection sheet. The aim of this was to identify barriers to adherence with door to appropriate antibiotics within 4 h of admission to hospital not easily captured by the quantitative data collection and/or identified in the pre-implementation surveys. When appropriate, extracts of case-notes were recorded verbatim. These data were organized into major themes and subthemes according to whether the identified reasons for non-adherence were likely to be amenable to an educational intervention or not.
Univariate and multivariate analyses to identify determinants of door to antibiotic time and appropriate antibiotic prescribing
A detailed description of the methods used to collect the data has been published.10 Patients were identified prospectively by review of admission records at the intervention and control sites of the controlled before–after study described in the Study design section in the Methods section. Patients were included if they were receiving an antibiotic for a lower respiratory tract infection and had either a new infiltrate on the chest radiograph or a clinical diagnosis of CAP, made by a specialist registrar or consultant, documented in the case-notes. Patients were excluded if they were under 16 years old, taking immunosuppressive drugs, HIV positive or neutropenic (neutrophil count <1.0 x 109/L), or had aspiration or hypostatic pneumonia, progressive malignancy or if the diagnosis was changed prior to discharge from hospital.
Statistical analyses were performed using SPSS for Windows (version 10). To identify potential independent determinants of door to antibiotic time, univariate linear regression was performed using door to antibiotic time as the dependent variable and a wide-range of potential predictor variables. Predictor variables found to have a statistical association with door to antibiotic time at the P 0.1 level were then entered into a multivariate linear regression model in order to identify independent predictors. The same methods were used to identify potential determinants of appropriate antibiotic prescribing except that logistic regression was used.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
Quantitative survey
Of the 108 junior and middle grade staff working in the Medicine and Cardiovascular Directorate between August 2001 and July 2002, 83 (77%) were surveyed. Of those interviewed, 43 (52%) were junior house officers, 30 (36%) were senior house officers and 10 (12%) were specialist registrars. Just over half (52%) had worked at the hospital for more than 6 months. One-third (34%) of respondents had previously or currently worked on a respiratory medicine unit and 29% had previously or currently worked on an infectious diseases unit. Most respondents (72%) had either not taken or not passed Membership of the Royal College of Physicians (MRCP), UK part 1 or 2 examinations; 14.5% had passed both parts.
Respondent-reported experiences of formal and informal training, familiarity with local and national CAP guidelines, attitudes towards guidelines and CAP, and experience of the working environment on the AMAU are summarized in Table 1. In an open question, when asked: while admitting a patient in ward 15 (AMAU), if you required information about what antibiotic regimen to prescribe for a patient with CAP, from what or who would you seek advice? 78% of respondents stated that they would consult the hospital's sepsis protocol. The second commonest response (12%) was to seek advice from either a senior house officer or specialist registrar.
|
Knowledge of severity assessment
When the 83 doctors were asked in an open question to state the top four criteria they would use to assess severity of illness in patients with CAP, only 3 (4%) correctly stated all four of the core prognostic (CURB) criteria recommended by the BTS. In contrast, when asked in a Likert scale question how they rated the importance of 12 criteria in the assessment of severity in CAP, all four of the BTS core criteria were in the top five answers. When given three clinical severity assessment scenarios, however, only six (7%) of doctors were able to assess all three cases correctly (NB probability of guessing all three cases correctly = 4%).
Knowledge of recommended antibiotic therapy
Based on the answers given when respondents were asked to suggest an antibiotic regimen for a patient with severe CAP (open question), 56% of patients would receive a BTS adherent regimen, although 78% would receive antibiotics that covered both Streptococcus pneumoniae and ‘atypical’ bacteria. When asked for an antibiotic regimen for a patient with non-severe CAP (oral route available if asked), 46% stated a BTS adherent regimen, 48% antibiotics without activity against atypical bacteria, 34% a regimen with unnecessarily broad-spectrum activity and 4% intravenous antibiotics.
Ability to assess severity in a clinical scenario
When respondents were asked to grade the severity of a patient with CAP in a clinical scenario,16a 7% correctly answered all three scenarios, 57% correctly answered two scenarios, 31% one scenario and 5% scored zero (NB probability of guessing all three cases correctly = 4%, two cases = 7%, one case = 14% and none = 29%).
Associations between grades of staff, respondent-reported familiarity with local and national antibiotic guidance and performance
Respondents' performance in stating the BTS core severity criteria was statistically significantly associated with their grade (higher grades performed better), self-reported familiarity with the BTS guidelines (the more familiar the respondent was, the better the performance) and current or past experience of working in respiratory medicine or infectious diseases (those who had, performed better). In contrast, self-reported familiarity with the local sepsis protocol was not associated with performance in answering severity-based questions. No associations were identified between any of the above variables and respondents' ability to correctly assess the three case scenarios (Table 2).
|
Respondents' ability to prescribe a BTS adherent antibiotic regimen for CAP was significantly associated with their grade (higher grades performed better; P = 0.015 for severe and P = 0.002 for non-severe), familiarity with both the local sepsis protocol (the more familiar the respondent was, the better the performance; P = 0.003 for severe and P < 0.001 for non-severe) and BTS guidelines (P = 0.005 for severe and P = 0.01 for non-severe), and (for severe, but not non-severe CAP) their current or past experience of working in a respiratory or infection unit (those who had, performed better; P = 0.006).
Semi-structured (in-depth) interviews
Five major themes were identified as being of importance to the research questions:
Theme 1: knowledge, experience and confidence. Some respondents commenced their first house officer posts with a lack of key clinical skills to manage CAP [my undergraduate experience didn't really teach me anything about severity of pneumonia (008, 796–8); I didn't know how to put up IVs (010, 241–9)] and poor knowledge about antibiotic therapy [Certainly when I was a lot more junior, I found a lot of difficulty remembering which antibiotics were appropriate (043, 117–20); it's not emphasized ... the importance of early antibiotic treatment ... throughout medical school (037, 679–83)]. Some respondents thought the latter was because they ‘didn't get enough teaching about therapeutics’ (001, 447–8). Others reflected on a lack of hands-on clinical experience [We didn't get to see any (pneumonia) patients (008, 36–7)].
Following graduation, some respondents commented on their lack of formal and informal CAP training. Others had clearly had better experiences, which appeared to be associated with a positive training culture in that environment. There was a strong preference for clinically orientated and small-group interactive teaching. Respondents were either unaware of or unfamiliar with the BTS CAP guidelines [I'm not aware of anything about that (020, 279); I'm not familiar with them at all, I have to say (043, 190)]. This appeared to be due to the availability of any easily accessible and well-respected local protocol [...it's not easily accessible (the BTS guidelines), the sepsis management protocol is easily accessible (037, 269–71)], the lack of user friendliness of the BTS guidelines [they're so un-user friendly (001, 327–8)], the failure to ‘spoon-feed’ (008, 436) respondents and ‘focusing in on your own little world’ (010, 129–33).
There was a strong relationship between respondents' confidence in their ability to manage CAP [It's a reasonably common thing that I've managed before, so I have a good grasp of what I'm doing (043, 895–6)] and the likelihood of a respondent using guidance [I'm quite comfortable managing pneumonia without having to look at guidelines (025, 1016–8)]. There was also a strong relationship between clinical experience, confidence and achievement of processes of care. For example, one respondent reflected how, when she first started as a house officer, her inexperience and lack of faith in her own ability resulted in waiting for the chest radiograph to come back before confirming a diagnosis of CAP and commencing treatment. As her experience and confidence increased, the likelihood of waiting for a chest radiograph before commencing treatment decreased [just as I went along I learned to spot it ... so may be in the first month I would have waited for the X-ray, but now I wouldn't because I believe in what I'm doing ... that's just about experience (010, 601–9)]. Another respondent reflected on how she had felt confident to manage CAP as a house officer, but ‘retrospectively you realize that perhaps things were sub-optimal’ (001, 72–4).
As respondents' clinical experience and confidence increased, the likelihood of using available guidance decreased [Yes, I mean if it (a CAP guideline) had been there when I was less experienced, I would certainly have used it ... I probably wouldn't use it so much now (043, 899–905)]. There was evidence to suggest that the shift from using guidance regularly to using guidance less often occurred after just a few occasions [And you don't look back again after doing it twice or thrice ... you don't need to go back to protocols always (020, 563–8); eventually you won't have to look at it, it's the first couple of times you need to look at it (037, 375–7)].
Theme 2: attitudes to pneumonia and guidelines and influence of using guidelines on future practice. Although all respondents thought CAP was an important condition, some reported a ‘blasé’ (010, 533) attitude towards it with a lower priority status in comparison with other clinical presentations [...it's not quite as prioritized as much as like chest pain, which is much faster (010, 227–8); ...we can do a lot for it (myocardial infarction) and minutes mean muscle to coin a phrase, whereas in pneumonia it's just pneumonia ... It's the attitude to it that I see (043, 442–8); it's just not seen in folks eyes as being exciting as, something like chest pain, which you can thrombolyse (038, 648–50)]. There was some evidence of an ‘old man's friend’ (043, 575) attitude with one respondent reflecting his experiences of the intensive care unit refusing to take older patients with CAP who he considered ‘treatable’ (025, 705).
There appeared to be a relationship between increasing clinical experience and an increased recognition of the importance of CAP. For example, for one respondent ‘pneumonia didn't feature that highly at that point [when starting her first house job] in my mind’ (001, 75–6) whereas at the time of interview she considered it to be important and ‘in need of treatment quickly’ (001, 94–5). This may have been because some respondents had seen enough pneumonia patients deteriorate ‘despite everything’ (001, 599), to avoid becoming ‘cocky’ (001, 601) and to recognize the importance of taking it ‘seriously and not (to) delay treatment’ (008, 785–6).
All respondents had positive attitudes towards guidelines, but reflected on the need for clinical autonomy when ‘other factors’ (037, 149) in an individual patient (e.g. co-morbidity) necessitated the need for variance. There was evidence of a relationship between increasing clinical experience and increasing scepticism about guidelines as a result of experiencing such situations [you realize once you've had more experience ... that the guidelines are not the best thing and there is a better form of management (037, 867–9)]. Trust in guidelines was not a strong theme, but when asked, determinants of why respondents trusted guidelines were: observing that the recommended practice worked, knowing that the developers had ‘looked into it, researched it, used evidence-based practice’ (008, 893–4), observing more senior doctors using guidelines, and if the guideline was recent (i.e. in-date).
Theme 3: environmental factors. An important theme was the high intensity of workload for nursing and medical staff on the acute medical admissions unit (AMAU) [you feel that your life blood is being drained out of you (001, 862–3); it's knackering, very tiring (008, 910); your thinking power is not always there with you, sometimes you'll be doing things mechanically (020, 840–1)] and how this could result in the delayed diagnosis and management of CAP [...the patient comes in and you've got the delay of the doctor actually seeing the patient (010, 220–3); Busy day, busy nurses, lots to do, time for antibiotics to be made up ... other things to run through (043, 386–7)]. Intensity was compounded by organizational issues such as a lack of beds and the need to house patients in the corridor [I've had someone sitting in the corridor that I've walked past and ‘Oh, she's fine, she's sitting there quietly,’ she's sitting there quietly because she's breathing her last (001, 1015–7)], an inadequate triage system, a lack of porters to take patients to the X-ray department, the lack of a system to review returned chest radiographs and the effect of the shift system on medical staff [the medicine is interesting, but the shifts are nightmares ... they are a long shift and it can be hard to get through them (043, 768–71)].
Respondents commented on their experiences of and the potential impact of poor communication and teamwork between medical and nursing staff and other hospital departments (e.g. the differing priorities of doctors and nurses and the refusal of many Accident and Emergency Department doctors to prescribe antibiotics prior to transfer to the AMAU) and the lack of approachable senior support for inexperienced junior staff [when you're starting, it's really intimidating to try and find the Reg. on-call about what he thinks about their patient, so therefore you kind of leave it and wait for the ward round (010, 314–7)]. One respondent overcame some of the potential delays by ensuring she always wrote up a ‘once off’ (038, 345) dose and putting it ‘in front of somebody's nose, that's what I've done, I've done it as a once off’ (038, 346–8).
Theme 4: guideline factors. There were strong relationships between the likelihood of a respondent using guidance and the accessibility [...provided you put the guidelines everywhere (020, 183); ...if they're presented in a logical manner that you can understand, easy to find and not ambiguous (008, 824–6)] and usability [fairly simple and straightforward (001, 1062); ...some of them are really well set-out and some of them are just pages of text and I mean it's all good stuff, but it's not easy to flick through (010, 54–5)] of that guidance. The sepsis protocol was often used as an example of a local guideline that was both easily accessible and useful. Some respondents thought that a portable version of CAP guidance would be useful. Others emphasized the importance of including enough detail to make it useful in clinical practice (e.g. severity assessment criteria and alternative antibiotics for patients allergic to penicillin).
Theme 5: human and patient factors. While respondents stated that ‘the (diagnostic) clues are usually there’ (043, 235), examples were given of cases in which the diagnosis of CAP had initially been difficult [His chest was clear and even when I had the chest X-ray with pneumonia on it, I listened to his chest again—still sounded fine (043, 219–22); ...couldn't find the source of sepsis so didn't start the antibiotics and it turned out to be quite severe pneumonia (008, 712–6); ...it's just a few crackles ... and sometimes we tend to wait on the X-ray (020, 475–9)]. Respondents rarely mentioned lone human error as a cause of delayed or inappropriate antibiotic therapy, although occasionally patients ‘just get forgotten’ (010, 223–6) or ordered investigations (e.g. the chest radiograph) are not followed-up.
Post-implementation qualitative field-notes
Before implementation, 33% (60/181) of patients at the intervention site and 32% (19/60) of patients at the control site received appropriate antibiotic therapy within 4 h of admission to hospital. After implementation, the proportions increased to 56% (118/209) at the intervention site and 36% (19/53) at the control site. At the intervention site, 91 patients (44%) in the post-implementation cohort did not receive appropriate antibiotics within 4 h of admission. Identified reasons for variance were considered to be amenable to education alone in 24 (26%) cases (Table 3). Failure to prescribe an immediate dose of antibiotics in the once-only section of the drug chart was the most common reason for variance that would be amenable to education. In 55 cases (60%), at least one reason was unlikely to be amenable to education. A clear reason could not be identified in the remaining 12 cases (13%).
|
Multivariate analyses of determinants of door to antibiotic time and appropriate antibiotic prescribing
Of the 503 patients included in the controlled before and after study, 489 patients were included in this study, the remaining 14 patients being excluded because their diagnosis was changed to a non-pneumonia diagnosis between discharge from the AMAU to a ward and subsequent discharge from hospital.
Referral to the intervention hospital was the only pre-admission characteristic that was associated with being more likely to receive antibiotics within 4 h of admission and appropriate antibiotic therapy (Table 4). In addition, mention of lower respiratory tract infection in the general practitioner letter was associated with earlier administration of antibiotics, but not with appropriateness of antibiotic choice (Table 4). Four patient characteristics that may be associated with greater severity of illness (increasing pulse and respiratory rate, decreasing pulse oximetry and a temperature of <36 or >38°C) were associated with significantly shorter door to antibiotic time (Table 4).
|
In contrast, only pulse >125/min was significantly associated with being more likely to receive appropriate antibiotics, but patients with new confusion were significantly less likely to receive appropriate antibiotics (Table 4).
Using the study of Cabana et al. as a foundation,7 the results of four studies were used to develop a model of the determinants of the rapid delivery of appropriate antibiotics to patients admitted to hospital with CAP in Tayside (Figure 1). The relationships between barriers are predominantly informed by the in-depth interviews with the results of the structured survey, post-implementation qualitative field-notes and multivariate analyses ‘triangulating’ these data.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
The summation of the four studies presented in this paper is the most comprehensive assessment of barriers to the implementation of a CAP guideline or other antibiotic policy. Previously published studies have used qualitative interviews in representative samples of 10–22 doctors11,12,14 or reported on the doctor or patient characteristics associated with guideline adherence,15 but no previous study has applied all of these methods simultaneously.13
Comparison with previous studies
We have focused on studies about influences on adherence with antibiotic policies11–13,15 as opposed to guidelines about other aspects of CAP management such as discharge of low-risk patients.18 Our results are consistent with many of the results of these previous studies. For example, in the qualitative studies, junior doctors in Ireland also said that undergraduate education about antimicrobials needs to be more practical and easier to apply to on-ward situations (Theme 1), that they relied more on personal experience as their career progressed (Theme 2) and that hospital guidelines need to be presented in a user-friendly format (Theme 4).14 Doctors in Canada identified local organizational barriers to implementing guideline recommendations (Theme 3).11 Doctors in the Netherlands again stressed the need for effective dissemination (Theme 4) and also questioned the need for guidance because they did not perceive any problems with antimicrobial prescribing (Theme 2).12
In the quantitative studies from the Netherlands13 and Spain,15 indicators of increasing severity of patients' illness were also positively associated with shorter door to antibiotic time. In the Netherlands,13 antibiotic therapy before admission was significantly associated with lower likelihood of receiving guideline antibiotics (OR 0.46; 95% CI 0.26–0.80), which was also true in our study (Table 4), although the association was not statistically significant (P = 0.12). In Spain,15 a very large (n = 1288) multi-hospital study was able to show a positive association between specialist training in Infectious Diseases or Respiratory Medicine and guideline adherence. Our study was not designed to investigate this association, but we did show that specialist experience was significantly associated with better knowledge about guideline content (Table 2). Differences between our results and previous qualitative studies probably reflect differences in the sampling strategy and the healthcare environment. Our strategy was designed to investigate differences between levels of experience among junior doctors, which was similar to the study from Ireland,14 whereas the other qualitative studies contrasted junior staff with senior staff. They revealed additional themes about the perceived credibility and flexibility of guidelines among senior staff11 and the role of junior doctors as independent decision makers.12
Influence of surveys results on our intervention strategy
Both our pre-implementation surveys clearly showed that the doctors thought they knew the criteria for assessment of severity of CAP (Table 1) and the more senior were in fact able to state them correctly (Table 2), but they were very poor at actually putting this information into practice (Table 2). Consequently education about guideline content was unlikely to change behaviour unless it was accompanied by regular feedback about doctors' adherence to key guideline recommendations.7 The lack of priority status given to CAP compared with myocardial infarction (MI) was a key finding from the qualitative interviews, which was not revealed by the structured survey. This is likely to be due to the longstanding door to needle time initiative in the UK for MI, which is widely recognized by junior and middle grade medical staff. In implementing the management pathway, therefore, we tried to create a door to antibiotic time ethos by emphasizing the high morbidity and mortality associated with CAP, and the evidence linking early antibiotic therapy with improved patient outcomes.19,20
The post-implementation qualitative field-note data were vital in identifying that the failure to write a once-only dose of antibiotics in the appropriate section of the drug chart at the commencement of therapy could result in considerable delay. This potential barrier, although seemingly obvious in hindsight, had not been identified pre-implementation except that it had been mentioned by one respondent in the semi-structured interviews. These data were fed-back to clinicians and the achievement of door to antibiotic time improved following this.10 This insight also led to a potentially more important organizational change, which was to agree with senior Accident and Emergency staff that antibiotics should be administered in the department when patients with CAP presented there.
Wider implications of our study results
Our insights into doctors' misplaced confidence in their ability to assess CAP severity correctly have resulted in changes to both undergraduate and postgraduate education. Specific to CAP we now instruct undergraduates and postgraduates not to try to remember the CURB-65 criteria as they will inevitably confuse them with other clinical scores. Instead, we tell them that the criteria are readily accessible through posters in the wards and on-line in the antibiotic policy. We have embedded this approach into on-line teaching resources for undergraduate medical students and Foundation Year doctors in Scotland.21,22 We stress that good doctors check their facts and only rely on their memories when they are absolutely certain. Moreover, they can tell the difference because they actively seek data about their management of patients and reflect on how they are doing.
The need to present evidence that is useful in a simple format has led to the development of a care-bundle for CAP. The concept of a care-bundle has two key features.23,24 First, all of the recommendations are separately linked to outcome via good quality evidence. Second, all of the elements can be delivered by a single clinical team, which helps to overcome some of the organizational barriers that we identified (Theme 3). The problems with early identification of CAP (Theme 5) have led to a broader initiative to improve the turn-around time for chest radiographs in acute admissions. Finally, the importance of identifying individual and organizational barriers to change has supported the application of rapid tests of change as a core method for improving quality in NHS Tayside.25
The two pre-implementation surveys and the post-implementation qualitative field-notes were all performed to inform implementation of our quality improvement intervention rather than as stand-alone research. This means that although our findings are likely to have relevance to other healthcare systems, the results may not be completely generalizable to all acute medical environments. Similarly, the importance of different themes and the relationships between themes may be different in other environments. Barriers may also exist that we were unable to identify. The results are broadly consistent, however, with the relatively small amount of previously published data and should be used to inform the investigation of barriers at a local level.25
The two pre-implementation surveys were relatively small. Not all views and experiences may have been ascertained by the in-depth interviews, although no new themes were emerging by the last interview. Nevertheless, the results from a relatively small number of interviews can still be useful in identifying potential barriers to recommended practice and informing quality improvement. During the in-depth interviews, respondents were asked to recall a recently managed or memorable real-life case of CAP, which then prompted subsequent questioning. Although this proved to be a rich seam of data, it may have been more appropriate to use a case-based or critical-incident study design and interview respondents recently involved in suboptimally managed CAP cases. Such respondents could have been compared with those involved in well-managed cases, thereby allowing the identification of, for example, concepts and typologies directly related to the efficiency of delivery and appropriateness of antibiotics in a ‘real-world’ environment. For the reasons outlined in the Methods section, this proved difficult. Although attempts were made to put respondents at-ease, reduce the junior/senior divide and ensure confidentiality, it is recognized that these and similar issues may have biased respondents' answers in both surveys. For example, concern that consultants may gain access to the data may have reduced criticism of senior staff and the system. This would appear to be unlikely, however, given the vigorous critique by respondents of the AMAU environment. Additionally, the results of our previous audit1 and the other results reported in this paper appear to be consistent with the in-depth interview findings.
This is the first study to combine quantitative and qualitative methods to investigate barriers to implementation of CAP guidelines and then analyse this information alongside results of an intervention to change practice. The similarity between our results and studies that have only applied one of these methods implies that our findings may be generalizable. The combination of techniques provided rich information that had a major impact on the design of our intervention and probably contributed to its success by directing us towards organizational changes in addition to educational strategies and by including regular feedback of results to front-line staff.
|
Funding |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
The original study was funded by a Training Fellowship for G. B. from the Scottish Council for Postgraduate Dental and Medical Education and by a grant from the Chief Scientist Office of the Scottish Executive. The authors are not receiving any ongoing funding for this project.
|
Transparency declarations |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
Within the last 2 years, G. B. has received support to attend conferences/meetings from Astra-Zeneca, Sanofi-Aventis, Gilead, Pfizer and Chiron, and received honoraria for speaking/DVD development from Sanofi-Aventis.
Within the last 2 years, D. N. has served on Advisory Boards for Janssen Cilag (UK Anti-Infectives), Wyeth (UK tigecycline), Novartis (UK daptomycin) and Optimer (Global, Optima-80), and received honoraria for speaking from Novartis, Johnson & Johnson, Pfizer and Wyeth.
Within the last 2 years, P. D. has served on Advisory Boards for Johnson & Johnson (Global Anti-infectives) and Wyeth (UK tigecycline), received honoraria for speaking from Johnson & Johnson, Optimer, Pfizer and Wyeth, and received research funding from Boehringer Ingelheim, GlaxoSmithKline and Pfizer.
The other authors have none to declare.
|
Supplementary data |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
The management pathway is available as Supplementary data at JAC Online (http://jac.oxfordjournals.org/).
|
Acknowledgements |
|---|
We would like to acknowledge the help and support of John Winter, Consultant Physician in Respiratory Medicine, William Morrison, Consultant in Accident and Emergency Medicine, Michael Jones, Consultant Physician in Acute Medicine, and Peter Slane, Acute Liaison Principal/General Practitioner.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
1 Gilbert K, Gleason PP, Singer DE, et al. Variations in antimicrobial use and cost in more than 2,000 patients with community-acquired pneumonia. Am J Med (1998) 104:17–27.[CrossRef][ISI][Medline]
2 Meehan TP, Fine MJ, Krumholz HM, et al. Quality of care, process, and outcomes in elderly patients with pneumonia. JAMA (1997) 278:2080–4.[Abstract]
3 Nathwani D, Williams F, Winter J, et al. Use of indicators to evaluate the quality of community-acquired pneumonia management. Clin Infect Dis (2002) 34:318–23.[CrossRef][ISI][Medline]
4
Woodhead M, Macfarlane J. Local antibiotic guidelines for adult community-acquired pneumonia (CAP): a survey of UK hospital practice in 1999. J Antimicrob Chemother (2000) 46:141–3.
5
British Thoracic Society. BTS guidelines for the management of community acquired pneumonia in adults. Thorax (2001) 56(Suppl 4):1–64.
6 Mandell LA, Wunderink RG, Anzueto A, et al. Infectious Diseases Society of America/American Thoracic Society consensus guidelines on the management of community-acquired pneumonia in adults. Clin Infect Dis (2007) 44:S27–72.[CrossRef][ISI][Medline]
7
Cabana MD, Rand CS, Powe NR, et al. Why don't physicians follow clinical practice guidelines? A framework for improvement. JAMA (1999) 282:1458–65.
8 MRC Health Services and Public Health Research Board. A Framework for Development and Evaluation of RCTs for Complex Interventions to Improve Health. (2004) 1–18. http://www.mrc.ac.uk/pdf-mrc_cpr.pdf (8 November 2004, date last accessed).
9 Grimshaw JM, Thomas RE, MacLennan G, et al. Effectiveness and efficiency of guideline dissemination and implementation strategies. Health Technol Assess (2004) 8:iii–72.[Medline]
10
Barlow G, Nathwani D, Williams F, et al. Reducing door to antibiotic time in community-acquired pneumonia: controlled before and after evaluation and cost-effectiveness analysis. Thorax (2007) 62:67–74.
11 Majumdar SR, Simpson SH, Marrie TJ. Physician-perceived barriers to adopting a critical pathway for community-acquired pneumonia. Jt Comm J Qual Saf (2004) 30:387–95.[Medline]
12 Mol PG, Rutten WJ, Gans RO, et al. Adherence barriers to antimicrobial treatment guidelines in teaching hospital, the Netherlands. Emerg Infect Dis (2004) 10:522–5.[ISI][Medline]
13
Schouten JA, Hulscher ME, Kullberg BJ, et al. Understanding variation in quality of antibiotic use for community-acquired pneumonia: effect of patient, professional and hospital factors. J Antimicrob Chemother (2005) 56:575–82.
14
De Souza V, MacFarlane A, Murphy AW, et al. A qualitative study of factors influencing antimicrobial prescribing by non-consultant hospital doctors. J Antimicrob Chemother (2006) 58:840–3.
15
Menendez R, Torres A, Zalacain R, et al. Guidelines for the treatment of community-acquired pneumonia: predictors of adherence and outcome. Am J Respir Crit Care Med (2005) 172:757–62.
16
Tunis SR, Hayward RS, Wilson MC, et al. Internists' attitudes about clinical practice guidelines. Ann Intern Med (1994) 120:956–63.
16a Barlow G. Improving the delivery and appropriateness of antibiotic prescribing in community-acquired pneumonia http://www.dundee.ac.uk/medden/groups/group23 (26 November 2007, date last accessed). PhD Thesis, University of Dundee, UK.
17 Ritchie J, Spencer L. Qualitative analysis for applied policy research. In: Analysing Qualitative Data—Bryman A, Burgess RG, eds. (1994) London: Routledge.
18
Halm EA, Atlas SJ, Borowsky LH, et al. Understanding physician adherence with a pneumonia practice guideline: effects of patient, system, and physician factors. Arch Intern Med (2000) 160:98–104.
19
Battleman DS, Callahan M, Thaler HT. Rapid antibiotic delivery and appropriate antibiotic selection reduce length of hospital stay of patients with community-acquired pneumonia: link between quality of care and resource utilization. Arch Intern Med (2002) 162:682–8.
20
Houck PM, Bratzler DW, Nsa W, et al. Timing of antibiotic administration and outcomes for Medicare patients hospitalized with community-acquired pneumonia. Arch Intern Med (2004) 164:637–44.
21
Davey P, Garner S. Professional education on antimicrobial prescribing: a report from the Specialist Advisory Committee on Antimicrobial Resistance (SACAR) Professional Education Subgroup. J Antimicrob Chemother (2007) 60(Suppl 1):i27–32.
22
Nathwani D, Christie P. The Scottish approach to enhancing antimicrobial stewardship. J Antimicrob Chemother (2007) 60(Suppl 1):i69–71.
23 Resar R, Pronovost P, Haraden C, et al. Using a bundle approach to improve ventilator care processes and reduce ventilator-associated pneumonia. Jt Comm J Qual Patient Saf (2005) 31:243–8.[Medline]
24
Pronovost P, Needham D, Berenholtz S, et al. An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med (2006) 355:2725–32.
25 Langley GL, Nolan KM, Nolan TW, et al. The Improvement Guide: A Practical Approach to Enhancing Organizational Performance (1996) 1st edn. San Francisco: Jossey-Bass Publishers.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J D Chalmers, A Singanayagam, and A T Hill Systolic blood pressure is superior to other haemodynamic predictors of outcome in community acquired pneumonia Thorax, August 1, 2008; 63(8): 698 - 702. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||