Background
Physicians need to update their medical knowledge continuously. Practice guidelines, usually regularly updated, are intended to facilitate this. They are mainly provided as diagnosis-and evidence-based information available on the internet but also as printed material e.g. in brochures or books. Lectures or discussions with colleagues are other ways to communicate the content of practice guidelines. However, the implementation of practice guidelines in primary care is a complex task [
1,
2]. General practitioners (GPs) are expected to have the overall responsibility for complex patient cases with multiple health problems. Still, the available guidelines are not designed to guide the comprehensive treatment of patients with multiple diseases but rather to guide treatment of separate diagnoses. It is therefore not surprising that previous research has shown that there is a considerable gap between what is carried out by physicians in clinical practice and what should be carried out to achieve the care and target levels stated in the guidelines [
1‐
5].
In the Swedish healthcare system a common directive in most practice guidelines is that relevant basal investigations should be performed and evaluated in primary care before the patients are referred to secondary care. There are high expectations of medical knowledge in many areas since the formal requirements to become a specialist in family medicine, e.g. a GP, in Sweden are equivalent as regard years in education and competence level to the requirements for any specialties in secondary care. However, the Swedish system to support continuous professional development (CPD) for physicians lacks formal requirements. Specialists, including GPs, are expected to keep themselves updated and to follow practice guidelines. Quality registers at local and national level are used in many disciplines to monitor the adherence to guidelines, and national comparisons aim to support increased adherence to the guidelines.
It is of great importance that CPD activities for GPs are based on principles of adult learning, and educational research indicates that methods which include interactive learning have the potential to increase knowledge and skills, and can change practice behaviour [
6]. To get an effective and useful CPD in the complex context of primary care is a special challenge. Several qualitative studies have elucidated broad issues such as “GPs views on use of guidelines” or “Attitudes towards evidence-based medicine in general” [
7,
8]. Issues like these have a tendency to elicit general statements rather than provide deep understanding of facilitating or hindering factors regarding the use of guidelines in routine care. In contrast, to focus on specific diseases might narrow the perspective too much and impair comparison between implementation strategies and/or clinical conditions. Further, practice guidelines per se are not supporting interactivity. They are mostly formed as written texts which physicians are supposed to read, learn from and use in their daily practice. It is of outmost importance that development of guidelines and other knowledge sources is based on a deep understanding of the influence of the context and the interaction between healthcare professionals in learning and performance. As far as we know, research is lacking on how GPs contextualize, interact, learn from and use practice guidelines in their day-to-day decision-making process.
The aim of the study was to explore how GPs approach, learn from and use practice guidelines in their day-to-day decision-making process in the primary care context.
Method
Study design
A qualitative approach with focus group interviews was chosen. Focus group interviews have proven to be a useful method to provide in-depth information and to explore thoughts and feelings underlying behaviour [
9,
10].
The Dual Process Theory was used to interpret and discuss the conceptualized themes. The theory is a dominant model for understanding the complex process that underlies human decision-making [
11‐
13].
The Dual Process Theory includes an intuitive (system 1) and an analytical (system 2) process. The theory is based on pattern recognition and the dual process this induces. If a patient presents with an illness that is recognized by the decision-maker e.g. a GP, the system 1 process will be engaged, which is fast but vulnerable to bias. If a patient presents with an illness that is not recognised by the decision-maker, the slower, more reliable system 2 process will be engaged, which requires more cognitive effort. There is a varying degree of interaction between system 1 and system 2 processes and the calibration between them depends on type of illness, the context and the experience of the decision-maker. The repeated appearance of similar situations using the system 2 process leads to a pattern recognition and relegation to the system 1 process. However, either system may override the other. A system 2 process override may result in decision-making characterized by control or surveillance. A system 1 process override may on the other hand result in a decision-making process where practice guidelines are overridden in favour of individual clinical judgment. There is tendency in the Dual Process Theory system to strive for a state requiring the least cognitive effort, the “cognitive miser” function.
The study followed the principles outlined in the Declaration of Helsinki, 1964 and was approved by the Regional Ethical Review Board in Stockholm, Sweden, Diary number: 2011/1071-32. Prior to the interviews all participants were informed both in writing and verbally about the study and that confidentiality was ensured. Written informed consent was obtained from all the participants.
Setting and sample
The GPs in this study were selected from 14 primary healthcare centres (PHCs) in the north-western part of Stockholm, Sweden in 2011. The 14 PHCs had earlier shown interest in developing practice guidelines. All the physicians that worked at the 14 PHCs during the study time, in total 132 (85 women), were invited to participate in the focus group interviews by e- mail, with two reminders.
Twenty-two GPs (16 women), representing seven of the PHCs agreed to participate in the study. The GPs had a median of seven years of experience as physicians in primary care (interquartile range 3-14 years). By definition, GPs are specialists in family medicine. In this study, for simplicity reasons, we included all physicians at the PHCs willing to participate in the definition of GPs. The definition thus included specialists as well as residents in family medicine. In total, 16 of the GPs were specialists in family medicine while six were doing their residency. The seven PHCs that were represented employed between 6-19 GPs. One of the PHCs had a vacancy; the other six were fully staffed regarding physicians.
All the PHCs had access to written guidelines, scientific literature and professional journals e.g. printed as well as internet-based material.
Data collection
In total, four focus group interviews were performed. Two interviews took place at the participants’ workplace while the other two interviews were performed in a conference room at the regional hospital. In the interviews performed outside the PHCs, only specialists participated; there were no residents (Table
1). The GPs that participated in the interviews performed at the PHCs were familiar with each other while the GPs participating in the other two interviews were not.
Table 1
Characteristics of the focus group interviews
1 | 4 | 4 | Outside the PHC |
2 | 8 | 1 | At the PHC |
3 | 7 | 1 | At the PHC |
4 | 3 | 3* | Outside the PHC |
Each interview lasted 1.5-2 hours. In three of the interviews (the two outside the PHCs and one performed at a PHC) there were lively discussions while the group dynamics were low in the last interview. One of the authors (MI), known to most of the participants through an earlier study and as a consulting specialist, conducted all the interviews. The second author (PBR), not known to the participants, acted as an observer, handled the tape-recorder and took field notes. Before the interview started, each GP answered a short questionnaire regarding their professional background, their use and access to guidelines and data regarding the number of GPs employed, and if there were any vacancies at their PHC. The results from the questionnaire are presented above. There were no incentives for the GPs to participate.
The interview guide consisted of an opening question, an introduction question, three core questions covering three important domains, and finally a closing question. The domains focused on how the GPs use practice guidelines in their decision-making, and factors that influence their decision how to approach a specific type of guideline. Furthermore, the interviews aimed to explore if and how the guidelines could encourage the learning process in the daily routine practice. The interview ended with a closing question where the participants were asked if they wanted to add something. This was done in order not to miss any data.
The interviews were semi-structured with open-ended questions [
14]. Follow-up questions were used for clarification. The interview guide was tested in the first pilot focus group interview. Since it worked out well, the pilot focus group interview was included in the final study.
Data analysis
Qualitative content analysis was performed [
15,
16]. The text was read several times to get a sense of the whole. The analysis started by identifying meaning units in the transcripts. The meaning units were then preliminarily sorted by content and meaning. In a second step, meaning units were condensed and labelled with codes while still preserving the meaning of the text. Codes that belonged together were grouped to form categories. In the next step the latent analysis was performed by analysing the content in the categories in a more abstract and interpretative way, and thus themes were created. The process of analysis is shown in Table
2.
Table 2
Process of analysis, with examples of meaning units, codes and categories corresponding to the themes
”… in a very structured way introduced in our PHC. .. What should I do when I have diagnosed a patient with hypertension? … We circulated and then everybody got to talk… There was suddenly time for reflection.” - Female
| Structured group meetings give stimulation, opportunity for reflection | - Feedback by peer-learning | Learning to use guidelines by interactive contextualized dialogues. |
”When the guidelines were to be updated all the GPs were invited to a meeting…It was a short presentation followed by a long dialogue…” –Female
| Cooperation between primary and secondary care regarding new guidelines. | - Feedback by collaboration, mutual learning and equality between specialties | |
“After a couple of years I felt that … I know how to do this and continued to work without checking the guidelines… suddenly something went wrong… I got scared and started to check again.” – Male
| Confirmation of knowledge assures quality of care. | - Confidence by confirmation | Learning that establishes confidence to provide high quality care. |
” “I used to work in Gotland so I usually call the hospital there since I know everyone… I am always familiar with the person that I am talking to…” -”Female
| Reliability from consulting familiar, competent colleagues. | - Confidence by reliability | |
” I would like to see results…I want follow-ups, personal follow-ups…If I never get feedback on what I am doing…why should I care??” – Female
| Evaluation of improvement encourages adherence to guidelines. | - Confidence by evaluation of own results | |
“It should not be too compact…arranged in a recognizable pattern so the information is easy to find.” -Female
| Pedagogic lay-out | - Design and layout visualizing the evidence | Learning by use of relevant evidence in the decision-making process. |
”The biggest problem is lack of time…Oh God there is so much information, where should I start.” - Female
| Poor search- function time-consuming | - Accessibility adapted to the clinical decision-making process | |
Rigour
We chose the focus group interview method since interactions in a group of people with similar background on a specific topic enhance data quality [
17]. To improve the judgment of transferability, the sample setting, data collection and analysis process were carefully described. The findings might therefore be transferrable to similar groups of GPs. In the fourth interview, much of the information had already been obtained in the earlier interviews, thus saturation was assumed. All the interviews were performed within seven months, which should contribute to trustworthiness by decreasing the risk of inconsistency [
18]. The interviews were recorded, transcribed verbatim by one of the authors (MI) and checked against the audio-recording for accuracy. In order to ensure credibility, all the authors, who had different backgrounds and perspectives, performed the content analysis process and discussed categories and themes until consensus was obtained.
Quotations from the original interviews were selected in order to illustrate the results and to improve the credibility of the study.
Discussion
The possibility for learning to use guidelines by interactive contextualized dialogues and learning that establishes confidence in providing high quality care was emphasized by the participating GPs as important aspects to consider in their approach to the practice guidelines. A prerequisite for using the guidelines was that they should allow access to relevant evidence in the decision-making process.
Making correct clinical decisions for each individual patient is a central goal for all physicians. The findings in our study have highlighted the importance of the usability of the guidelines in the decision-making process. The themes conceptualized in this study could be linked in different ways to the Dual Process Theory [
11‐
13].
In the first theme “Learning to use guidelines by interactive contextualized dialogues” the main finding was the importance of dialogue in decision-making in clinical practice. Pattern recognition is central in this dialogue. Regardless of whether the dialogue is obtained by peer-learning or by collaboration between primary and secondary care, it enhances the repetitive operation of the system 2 process and thus increases the process of pattern recognition. The opposite, strictly detailed instructions or flowcharts with few possibilities for personal reflection instead lead to the system 2 process overriding system 1, which will hinder the probability of pattern recognition.
The importance of networks in decision-making has also been pointed out by Mascia et al. who showed that physicians’ attitude towards evidence-based medicine (EBM) is strongly correlated with professional networks [
19]. Their conclusion was to avoid marginalization and stimulate integration and continuity of care both within and across the boundaries between different healthcare providers. The importance of dialogue in decision-making has been emphasised by Norman in a review article [
20]. He concluded that there is no single best way to solve a problem in the decision-making process. The components of knowledge and skills required to achieve the goal of effective care are complex and multidimensional. Feedback or dialogue is essential for this process. Another aspect of peer-learning by dialogue was that group discussions of patients with rare diseases could improve the knowledge as well as decrease the risk of diagnostic errors for GPs [
21]. Practice guidelines designed as rigid, strictly detailed information were in our study perceived as assignments instead of being part of a continuous learning process since they did not offer any possibilities for personal reflections or dialogue. Thus, they were considered as barriers instead of facilitators in the decision-making process. These findings are in line with several other studies discussing barriers to following clinical practice guidelines [
2,
4,
22].
In the second theme, “Learning that establishes confidence to perform a high quality of care”, the main finding was the importance of the GPs feeling confidence in the practice guidelines and having the possibility for evaluation of their decisions during the diagnostic process. The decision-making process is a balance between systems 1 and 2 that involves earlier experience and context on one side, and the use of practice guidelines, consultations or evaluation of one’s own results on the other side.
Confidence could be obtained by confirmation of knowledge, preferably by immediate feedback, but only if the source was perceived as reliable. Lack of confidence could result in the system 1 process overriding system 2, a dysrational override that would increase the risk of diagnostic errors.
Thus, if the GPs felt more confident by having the possibility to evaluate their own results they could switch between system 1 and 2 as needed. This would, as well as facilitating decision-making and decreasing the risk of diagnostic errors, also strengthen the autonomy of the GPs. To get feedback and to feel confident in competence would strengthen intrinsic motivation and thus autonomy. This, in turn, would promote behavioural changes as well as a positive attitude towards the use of the guidelines [
23].
In the third theme, “Learning by use of relevant evidence in the decision-making process”, the main finding was that the most important aspect of using practice guidelines was to find relevant information. If the practice guidelines are easily accessible with a clear design and lay-out, the use of the system 2 process will be encouraged. This, in turn, will make it easier for the GPs to switch between system 1 and 2 during decision-making. If the practice guidelines do not fulfil these criteria there is a risk that the GPs will make their decisions based on their own experiences without considering the practice guidelines due to lack of time. This insufficient use of the type 2 system would, in turn, mean an increased risk of diagnostic errors in the decision-making process. This theory is supported by the study of Duran-Nelson et al. [
24] where speed, trust and portability were identified as the biggest drivers for resource selection, while information overload appeared to be one of the biggest barriers. Previous studies have also pointed out guideline format as an important factor affecting whether GPs use guidelines or not. These studies concluded that practice guidelines should be short, simple and preferably should include patient leaflets [
25,
26].
We found that the “cognitive miser” strategy was necessary for the GPs in their daily routine work. The challenge when using the “cognitive miser” strategy is to keep a balance in the use of cognitive effort. Too little cognitive effort leads to high efficiency but a higher risk of diagnostic errors, while too much cognitive effort leads to a low risk of diagnostic errors but would probably not be compatible with the time constraints in clinical practice. GPs need to have relevant information at hand to be able to make a decision. Yet, they must have some pre-understanding to know which diagnoses to look for in the practice guidelines. Otherwise the “cognitive miser” strategy will result in the system 1 process overriding system 2. An example of how the “cognitive miser” strategy can lead to diagnostic errors is described in the study of Schmidt et al. [
27]. In this study he showed a substantial availability bias causing a significant higher rate of diagnostic errors among experienced physicians compared to a control-group after exposure to media-provided information about a disease. In the next phase of the study when the exposure was the same but the physicians were given instructions (guidelines) and time for reflection there was no significant difference in diagnostic errors between the groups.
Previous studies [
28,
29] have shown that diagnostic errors are common. Most common were diagnostic reasoning errors when the physicians were not aware that their actions were incorrect. The conclusion was that better understanding of the decision-making process could help to improve the process and thus improve patient care. We think that it is important to consider the “cognitive miser” strategy when practice guidelines are elaborated and designed. Increasing the usability of the guidelines could thus reduce the risk of diagnostic errors.
There are several studies, mostly quantitative but also qualitative, exploring mainly the barriers but also the facilitating aspects of following practice guidelines [
2,
4,
5,
7,
8,
22]. In our study all participants worked at group practices, which may have influenced the results. The influence of the size of the primary health care centre in the decision-making process has been investigated in a systematic review by Damiani et al [
30]. This review showed strong evidence that group practice performed better drug prescription according to the guidelines than single-handed practice. Further there was moderate evidence that GPs working in a group practice were more satisfied with their compensation and that group practice improved the utilization of information technology. The findings in this review are in line with our findings regarding the importance of dialogue and feedback in the learning process, which is easier to obtain in a group practice. Further our findings that accessibility to practice guidelines is a prerequisite for use are well in line with the findings in the review where group practice improved the utilization of information technology. Their conclusion was that there seems to be organizational advantages in group practice compared to single-handed practices in using the practice guidelines in the decision-making process. In our study the number of GPs varied between 6-19 GPs. Further studies are needed to investigate the optimal size of a group practice to obtain high quality of care.
Decision-making is central for a physician in clinical practice. The strength of this study is therefore that it focuses on exploring how GPs approach, learn from and use practice guidelines in their decision-making process, and discusses the results in the light of the Dual Process Theory.
However, given the design of the study, interviews can only give data regarding what the GPs say they do; not what they actually do in practice. Nevertheless, a relationship could be anticipated between what people say and do. One way to increase the validity between “saying and doing” is to ask questions that reflect the participants’ way of acting [
31]. In our interviews we frequently used follow-up questions where we asked for concrete examples. The construction of the groups and the settings was initially planned to be equal in all four focus group interviews but this did not work out for practical reasons. Nevertheless, in three of the interviews there were lively discussions in which all the GPs participated equally. In the remaining interview the energy in the discussions was low and it was difficult to motivate two of the participants to speak. This lack of engagement could have been associated with the fact that this was the interview where the participants had the widest range of experience as physicians in primary care.
To provide optimal healthcare it is necessary that we understand how clinical evidence can be effectively transferred and applied in the context of primary care, and thus appreciate the importance of effective continuing education or professional development of physicians [
32]. We have identified learning strategies that could be considered when developing and implementing new practice guidelines. The most important issue in CPD is to take the context into account, to arrange the learning activities to facilitate interaction between healthcare professionals, and finally to include feedback on performance based on guidelines and other knowledge sources. Regular meetings when GPs discuss pros and cons with different guidelines in primary care have a potential to facilitate this contextualization and interaction. Further concerned GPs would preferably be involved in the process of developing new practice guidelines aimed for use in primary care setting.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MI, PBR and AK participated in the design of the study, analysed the data and drafted the manuscript. MI collected the data and performed the interviews together with PBR. All the authors read and approved the final manuscript.