Understanding the problem

6.2. Understanding the problem

Understanding why things go wrong is subtly different to knowing how things go wrong. None of the published QI initiatives for CAP appear to have been informed by studies of why certain aspects of process are not achieved. While multidisciplinary teams may think that they understand a problem and know how to correct it, such teams usually consist of local experts or senior members of staff who may be removed from, or not participating in, the day- to-day delivery of acute medical care. In the United Kingdom, for example, the “gate-keepers” of the initial delivery of appropriate antibiotics in acute medi- cine are usually junior members of the medical and nursing team, who may not have the necessary knowledge and experience to achieve satisfactory process of care. Involving these healthcare professionals in the development process is therefore essential to understanding why process is not achieved and to design- ing successful interventions. Failure to explore this question is more likely to lead to an unsuccessful intervention and, in the future, routine or research- based QI programmes are less likely to be funded if this question is not studied or already understood.

A variety of methods are available to explore why things go wrong in healthcare process. Prior to TAYCAPP, we suspected that there was a lack of understanding of the link between severity assessment and appropriate antibiotic therapy amongst physicians, which then resulted in suboptimal process of care. We initially explored this by performing a structured survey of junior and middle grade physicians’ attitudes and knowledge. A previously published model of clinician adherence with clinical practice guidelines (Cabana et al., 1999) and other published evidence (Halm et al., 2000; Tunis et al., 1994), informed the design and content of this survey. This showed that suboptimal undergraduate and postgraduate training experiences, poor working environment, lack of familiarity with the BTS guidelines, and lack of knowledge regarding severity assessment and antibiotic prescribing were all potential barriers to the efficient delivery of appropriate antibiotics. In con- trast, attitudes towards guidelines in general and CAP appeared to be positive.

We went on to explore the identified barriers in more detail by performing

a small number of qualitative in-depth interviews. Qualitative methods are ide- ally suited to exploring complex phenomena, such as the efficient delivery of antibiotics. In order to ascertain a range of views and experiences, physicians were purposively sampled from the cohort of physicians who had completed the quantitative survey. The aim was to gain a better understanding of phenom- ena that were related to the efficient delivery of appropriate antibiotics and to identify reasons for non-adherence with local and national guidance and potential interventions. A number of other qualitative methods (e.g., focus

Pneumonia Guidelines in Practice

55 groups or observational methods) could have been used. For example, it would

have been more appropriate to perform a case-based or critical-incident analy- sis (i.e., the interviewing of healthcare professionals involved in the care of patients receiving delayed or inappropriate antibiotic therapy). However, prob- ably because of the long and intensive shifts on the acute medical admissions ward, physicians working in this environment were reluctant to be interviewed. In contrast, physicians who had recently finished this part of their post, and who were working in less busy areas of the Medical Unit, were more enthusiastic.

It is also legitimate to use qualitative enquiry to inform the domains and design of questions for a quantitative survey, which is then administered to a larger number of respondents. If qualitative methods are to be used, because of the interviewing and analysis skills required, it is vital to seek expert advice. Good starting points include the Scottish Consensus Statement on Qualitative Research in Primary Health Care (Dowell et al., 1995), Greenhalgh (2001), Pope and Mays (2000), and Pope et al. (2000).

The combination of quantitative and qualitative enquiry allowed the development of a local model of the delivery of antibiotic therapy for CAP. One of the interesting and surprising findings was that respondents’ confidence and perception in their ability to manage CAP appeared to relate to their previous clinical experience. For example, when starting as a house officer confidence in their ability to manage CAP was low. As their experience of seeing CAP increased so did respondents’ confidence and their perception that they were able to manage it well. However, this is a double-edged sword. While respon- dents’ confidence may lead to a timely diagnosis and early delivery of antibi- otics, the appropriateness of therapy depends on the knowledge acquired during their clinical experience. For example, some respondents’ believed themselves to be managing CAP according or close to recommended practice. When their answers in the quantitative survey, which were performed before the in-depth interviews, were checked, this was not always the case.

This finding suggests that CAP guidance is most likely to be used by junior or inexperienced physicians and highlights the importance of ongoing reflec- tion in clinical practice. These data were subsequently used in interactive educational sessions with physicians and other healthcare professionals to emphasise these issues. Additionally, we have been able to concentrate on other identified correctible barriers (e.g., knowledge about the link between severity assessment and appropriate antibiotic therapy) rather than barriers that we perceived to be important. The model has also identified barriers generic to acute medicine (e.g., ward organisation) that cannot be resolved as part of TAYCAPP. This means that a “ceiling effect” is likely to exist on the extent to which certain process of care measures can be achieved.

56 Gavin Barlow

Organisational factors

Team support systems

Clinical Experience

Work intensity

Quality of Perception

Use of

Guidance

CAP Care Priority of CAP

Human error

Difficult diagnosis (Some patients only) UG/PG

Awareness/ education

familiarity with

Guideline

local/national

factors

guidance Figure 3. The TAYCAPP model of the micro-determinants of the delivery and appropriate-

ness of antibiotic therapy in the acute medical admissions unit. The model is based on the qualitative and quantitative surveys described above. It should be noted, however, that sample sizes were small and the model was designed to inform TAYCAPP alone. Boxes show barriers to either the delivery or appropriateness of antibiotic therapy. Shaded boxes appear to be the more important barriers in Dundee.