Background
Failure to treat to target, or prescribing that is not concordant with guidelines are increasingly being referred to as clinical inertia (CI). Phillips et al. [
1] first coined the term, which they defined as a failure to initiate or intensify therapy when indicated, or a failure to act despite recognition of the problem [
1]. Alongside patient non-adherence to treatments, CI is believed to be a major factor that contributes to inadequate management of chronic conditions [
2,
3]. It has been suggested that CI related to management of diabetes, hypertension and lipid disorders may contribute up to 80% of heart attacks and strokes [
4]. As it associates with poor control of risk factors known to cause long-term health problems, CI has an economic impact alongside medical consequences [
5-
7].
O’Connor et al. [
2] postulated three classes of factors leading to CI: factors related to (i) providers, (ii) patients, and (iii) the system, with an estimated relative contribution of 50%, 30% and 20% respectively [
2]. Other authors report up to 75% provider-related factors [
8]. The three provider-related factors that were initially defined by Phillips et al. [
1] are assumed to be the most common contributors to CI [
2,
4,
9]: (i) providers’ overestimation of the care they give; (ii) providers’ use of ‘soft’ reasons to avoid therapy; (iii) providers’ lack of education, training or organisation for achieving therapeutic goals.
This list of factors is of little help in overcoming CI in practice. Indeed, practitioners need to be helped to overcome CI rather than systematically blamed for inaction [
10], the latter being otherwise occasionally appropriate. Indeed, as summarised by Reach [
11], clinical inaction may be called “true” CI only if: (i) a recommendation exists; (ii) the provider knows the recommendation; (iii) the provider believes the recommendation applies to the patient; (iv) the provider has the necessary resources to apply the recommendation; (v) the provider does not apply the recommendation for a particular patient, even though the conditions 1 to 4 are present [
11]. Following this definition, non-adherence to guidelines may correspond to appropriate inaction as a result of good clinical reasoning. As shown in our recent literature review [
12], actual CI is therefore difficult to observe and distinguish from appropriate inaction. It should not be evidenced without a careful investigation of a practitioner’s reasoning underlying their decisions. Moreover, in order to help anticipate the risk of CI in practice, it is necessary to understand which modifiable and non-modifiable factors underlie CI. Our study aimed to elucidate GPs’ beliefs regarding the risk of CI in their own practice, and personal modifiable factors associated with the risk of CI.
Methods
We conducted an exploratory qualitative study through group interviews among a sample of general practitioners (GP) in the Wallonie-Bruxelles Region (the French speaking part of Belgium). All participants were members of the same scientific society (
Société Scientifique de Médecine Générale - SSMG). Over 40% of all GPs in the French speaking part of Belgium are members of SSMG, thus ensuring truthworthiness of our study. As part of their vocational training, members of SSMG from a same geographical area meet regularly around topics of interest to their practice. We used the opportunity of these existing practice-sharing encounters to conduct our interviews. In accordance with the local group coordinators, the GPs were informed beforehand of the topic of the discussion through a letter co-signed by the first author (IA) and the medical coordinator of SSMG. The practitioners who accepted to participate in our study were asked to inform the coordinator of their group that they would participate. We opted for interviews during such formal natural groups [
13], rather than purposively sampled groups, as we hypothesized that the GPs would feel more comfortable discussing the issues surrounding CI in the familiar setting of their regular meetings. Our sample involved a total of 114 GPs in 8 group interviews, between October and December 2012. After the 5 first focus groups, a meeting of the steering committee (co-authors) was organised in order to refine the emerging themes and discuss implications for practice. After 3 more focus groups, as no new themes emerged, descriptive saturation of data was reached. The participants’ characteristics are shown in Table
1. Although sampling representativeness may be a less important issue in qualitative research, it is worth stressing that the male/female ratio among the participants in our sample (58.9/41.2) was the same than the male/female ratio among the members of SSMG (59/41).
Table 1
Participants’ characteristics (n = 114)
Gender | | |
Male | 67(58,8) | |
Female | 47(41,2) | |
Type of practice | | |
Single-handed practice | 73(64) | |
Group practice | 34(29,5) | |
Mixed (people working in more than one practice) | 7(6,5) | |
Setting | | |
Urban | 66(58) | |
Rural | 48(42) | |
Work experience | | |
Median duration of work experience | | 31,3(2–40) |
GPs with less than 10 years of practice | 11(9,6) | |
GPs with over 35 years of practice | 17(15,5) | |
Data collection
A standardised procedure was adopted for the data collection in every group. Two researchers were present at every meeting, and moderated the focus group discussions: The first author would introduce the topic with a brief review of the literature on IC. The group discussion would then be moderated by the second author. The research interviews were guided by two categories of open-ended questions consistent with the objectives of the study: (i) GPs’ beliefs regarding the risk of CI in their own practice, (ii) factors associated with CI. The duration of the discussions was limited to two hours. At the end of every focus group, there was a debriefing between the two researchers to discuss the most important themes that had emerged, and possible similarities and differences to other focus groups.
Data analysis
All focus groups were audio-taped and transcribed verbatim for analysis. As part of a process of respondent validation and error reduction, every participant received a synthesis of their group’s discussion, and was invited to share comments or further information directly with the researchers, either over the phone or by e-mail. We received feedback from six (6) participants, representing 5 different groups. Three (3) participants thanked the researchers for the discussion and the follow-up, and formally validated the synthesis. Two participants explained how the ideas shared during the focus group were further discussed within their groups, one stressing that it had made him personally more aware of the risk of CI in his own practice. The last participant informed the researchers of a TV programme relevant to the topic of CI. The participants’ comments did not add much to our analysis in terms of thematic contents, but confirmed the relevance of addressing the sensitive issue of CI with GPs.
The transcripts were analysed using an integrated approach of categorisation: A start list of two codes reflecting the research objectives was used as a framework to organise the thematic categories that were inductively created through careful revision of the transcribed material [
14,
15]. To ensure validity of the findings and interpretations, the first and second authors independently coded the transcripts and met at regular intervals to discuss emerging categories. As already stated, the members of the steering committee (co-authors) were invited to reflect and comment on the analysis process after 5 focus groups. The study was stopped after 8 groups, as no new themes were emerging, meaning that descriptive saturation was achieved. Our results hereafter are illustrated with individual quotations from the various focus-groups [
16].
Ethics
Our study did not involve any patients nor patients’ relatives, nor did it require that patient data be shared with the researchers. Our study does therefore not fall within the scope of the Belgian Law of 7 May 2004 on Human experiments, and did therefore not require the approval of an ethics committee, nor that informed consent forms be signed by the participants.
Ethical considerations were present however at all stages of the project: Every GP was personally informed of the study and invited to knowingly join (or not) the focus group interview several weeks before the encounter; the interview process was conducted with a non-judgmental attitude and was very respectful of the GPs’ perceptions and self-reported experiences regarding the phenomenon of CI in their own practice; last but not least, every GP within a group was sent a synthesis of the discussion within that group, and was thus given the possibility to comment personally, either by e-mail or over the phone, on the synthesis of their interview.
Discussion
The results of our group interviews show that non-adherence to guidelines in general practice should not be systematically labeled CI. The Belgian health-care system is characterized by fee-for-service, free access to all levels of care, mostly small single-handed practices in primary care, no compulsory guidelines (which are available but are not strictly “followed”), no real quality assurance, and a tradition of hospital-centered disease-oriented care, all of which can make GPs frustrated about working conditions. Despite this, the GPs in our study displayed great commitment to their work, and believed that prevalent definitions of CI did not take sufficiently into account the active role of the patient in the consultation and medical decision-making process, nor the breadth and complexity of their role, which includes tackling sociocultural determinants of health as much as medical and pharmacological ones.
When situations of CI were actually acknowledged, these were more often connected with a lack of investigation to timely diagnose and treat a problem than with failing to prescribe more. Insufficient delivery of health-promotion and patient-education messages were considered a form of CI as well. Across a variety of cases of CI in which patient and system factors were involved, feelings of being overwhelmed and disempowered was a common issue that emerged as an important factor of CI. Again, appropriate decisions not to adhere to guidelines in specific situations, for specific patients, at specific times are not to be mistaken for CI. As far as prescriptions are concerned, the participants in our study saw their decisions as appropriate most of the time although they were not always concordant with guidelines. Hence, the definition by Philips et al. [
1], which emphasizes prescription failures, was considered too narrow, and irrelevant to general practice.
At the heart of evidence-based medicine is the providers’ capacity to integrate individual clinical expertise with the best available clinical evidence from systematic research [
17]. Although the concept of CI arises in the context of evidence-based practices, attempts to define it tend to overlook the importance of adjusting care in individual cases. Indeed the main ways of measuring CI, which are based on target, timeframe, and decision to intensify therapy (or not) [
2,
4] are not sufficient to determine whether individual decisions to increase or not increase therapy might be appropriate for a specific patient [
18,
19]. As reflected by the participants in our study, in order to adequately assess CI, it is necessary to define
intermediate outcomes that incorporate information on, and justification of treatment decisions that are made [
20]. In the absence of such information, CI may be only apparent, and actually reflect good clinical practice [
18].
Application of clinical guidelines in the real world is somewhat problematic because there are practitioner, patient and system variables at play. However, clinicians only have direct control over the first. Finally, the use of the word ‘inertia’ to define a concept in Medicine (which would be better described as inaction) may initially have resulted from the mistaken belief that ‘inertia’ in Physics actually implies some obstacle or reluctance in changing motion. Describing the choice to maintain treatment without any change as ‘inert’ can give the incorrect impression of lethargy, and is rather debatable semantically.
The main strengths of our study lie in the number of group interviews and the total number of participants across the country, aiming for saturation of the data and internal validity. Independent coding by two researchers and discussions at several stages with a scientific committee including representatives of several medical specialities and two GPs, are another strength of our study. Regarding the researchers’ characteristics that might have influenced the study, whereas one researcher (first author) is very experienced in conducting qualitative research in the field of health care and clinical communication, the other researcher (second author) is a research nurse with extensive clinical experience and thorough knowledge of evidence-based practice. A strength of the study therefore lies in their complementarity to conduct the study. Moreover, none of them is a medical doctor. Although it might be argued that professions other than physicians are not given enough credence by physicians in other circumstances, we believe that this was a strength for our study, as we approached the physicians without a priori representations of how they should be dealing with the risk of CI in their practice. The main limitation lies with our choice of group interviews. Although it enabled us to collect useful information on physicians’ opinions regarding the much debated concept of CI, in-depth individual interviews might have yielded richer data regarding personal experience of CI.
Acknowledgements
The authors wish to thank all participating GPs for their willingness to cooperate, as well as Hans Vandenberghe, MD, and Annelies Vankeirsbilck, Lead Medical Affairs and Medical Affairs Manager respectively, at AstraZeneca Belgium & Luxembourg. We would like to thank them not only for the financial support received, but also for the very positive and open collaboration from study design to preparation of the manuscript. Special thanks are also expressed to Eric Legrand at Université Catholique de Louvain for invaluable administrative support.
Competing interests
The study received financial support from AstraZeneca Belgium & Luxembourg.
Authors’ contributions
IA: Designed the study, conducted the interviews, analysed the results, and prepared the manuscript. PJ: Conducted the interviews, analysed the results, and contributed to the preparation of the manuscript. MH: Discussed preliminary findings in several occasions and contributed to the preparation of the manuscript. ER, AS, PT: Discussed preliminary findings in several occasions and critically revised the manuscript. ED: Contributed to study design, discussed preliminary findings in several occasions and critically revised the manuscript. JW: Gave methodological advice during the study, discussed preliminary findings in several occasions and supervised the preparation of the manuscript. All authors read and approved the final manuscript.