Abstract
Background One of the tenets of general practice is that continuity of care has a beneficial effect on patient care. However, little is known about how continuity can have an impact on the diagnostic reasoning of GPs.
Aim To explore GPs’ diagnostic strategies by examining GPs’ reflections on their patients’ individual thresholds for seeking medical attention, how they arrive at their estimations, and which conclusions they draw.
Design and setting Qualitative study with 12 GPs in urban and rural practices in Germany.
Method After each patient consultation GPs were asked to reflect on their diagnostic reasoning for that particular case. Qualitative and quantitative analyses of consultations and interview content were undertaken.
Results A total of 295 primary care consultations were recorded, 134 of which contained at least one diagnostic episode. When elaborating on known patients, GPs frequently commented on how ‘early’ or ‘late’ in an illness progression a patient tended to consult. The probability of serious disease was accordingly regarded as high or low. This influenced GPs’ behaviour regarding further investigations or referrals, as well as reassurance and watchful waiting. GPs’ explanations for a patient’s utilisation threshold comprised medical history, the patient’s characteristics, family background, the media, and external circumstances.
Conclusion The concept of an individual threshold for the utilisation of primary care would explain how GPs use their knowledge of individual patients and their previous help-seeking behaviour for their diagnostic decision making. Whether the assumption behind this concept is valid, and whether its use improves diagnostic accuracy, remains to be investigated.
- clinical decision making
- continuity of patient care
- general practice
- physician–patient relations
- qualitative research
INTRODUCTION
Continuity of care is one of the defining features of general practice.1,2 There is some evidence that relational or personal continuity of care has a positive impact on the process and the outcomes of care including diagnosis.3–7 But little is known about the underlying mechanisms or how knowing or being familiar with a patient has an effect on care and its outcome. Diagnostic processes in primary care have been investigated intensively in the past. There are different strategies and heuristics enabling GPs to arrive at a diagnosis.8–11 However, little is known about how diagnostic processes are influenced by knowing patients in general practice.
In the authors’ study of cognitive processes relevant in GPs’ diagnosis,11 physicians were asked to reflect on their diagnostic reasoning. At the beginning of each interview they were asked to start with the beginning of the consultation and the complaint presented by the patient. This was the point when they elaborated on their knowledge of the patient, in particular patients’ individual thresholds for attendance and how this was important for the diagnostic work-up. To the authors’ knowledge, this phenomenon, which is the topic of this study, has not been described before.
The utilisation threshold arose while working on the data for the previous paper.11 Here, in this study further analyses regarding the utilisation threshold are presented.
METHOD
A qualitative and descriptive approach was used to investigate the cognitive decision processes used by primary care physicians. Observation of patient–doctor interaction in the real practice setting was thought to offer the best method to capture cognitive mechanisms. After each consultation, GPs were asked to reflect on their diagnostic reasoning in a semi-structured interview, resulting in data for qualitative and quantitative analysis.
Twelve experienced GPs working independently in their practices but associated with the teaching practice network of the University of Marburg in Germany were recruited for this study. This sample offered a broad range of GP characteristics with regards to age, sex, and location (rural or urban) (Table 1). All GPs were informed orally and in writing about the study objectives and the privacy regulations prior to the start of the investigation. Patients were included in order of their appearance in the practice and irrespective of their problems or complaints. Patients were informed by their GPs about the study prior to giving their written consent.
In order to help participating GPs reflect on their diagnostic reasoning, semi-structured interview guidelines were developed based on the relevant literature (Appendix 1).8,12 Trained research staff carried out the interviews. They began by asking GPs to recall the beginning of the consultation and to describe their first impressions and thoughts. The aim was to reconstruct the diagnostic process within each individual consultation. Interviews were performed at the GP’s office subsequent to each consultation and lasted between 2 and 18 minutes. Because there was no pre-specified hypothesis regarding the individual threshold for seeking medical attention, no questions were asked in this regard. Consultations and interviews were videorecorded remotely without the need for personal control in the consultation room. Tapes were transcribed verbatim using standard word-processing software. Transcripts were saved and analysed in anonymous form.
How this fits in
Continuity of care is widely accepted as a key principle in general practice and has proven beneficial in various studies. However, little is still known about how continuity of care affects diagnosis. Assuming an individual utilisation threshold for known patients represents one strategy for GPs to exploit their familiarity with their patients. GPs tailor the diagnostic process accordingly.
Transcribed data were imported into MAXQDA (version 11), software that supports qualitative data analysis. Data analysis was primarily based on a deductive–inductive approach,13 starting with previously described diagnostic strategies in primary care such as inductive foraging, triggered routines, and hypothesis testing.8,14 For exploration of the individual utilisation threshold, which emerged without predefined hypotheses, additional codes were developed for high and low thresholds. In an inductive process, explanations for different utilisation thresholds and the consequences for diagnosis and treatment arose and were consecutively added. Codes were defined explicitly with multiple memos. Two physician researchers conducted primary coding and analysis, and an in-depth analysis regarding the individual utilisation threshold was undertaken. Two independent raters (both physicians) evaluated the reliability of the coding tree. They agreed in 83% of cases. Conflicts in coding were resolved by discussion within the study team. A departmental working group on qualitative research also discussed conclusions.
Consultations and interviews were conducted in German. For this publication, quotations were translated by the authors and consecutively numbered. All authors work as part-time primary care physicians with a strong background in primary care research.
RESULTS
Utilisation thresholds
During sessions covered by the study team, 344 patients were asked to participate: 49 patients refused to take part, mainly due to lack of time, and therefore 295 patients were included in this study. The number of patients who participated varied between 18 and 38 in each practice. One patient later withdrew their consent.
In 134 consultations at least one diagnostic episode could be identified. See Table 2 for patient characteristics. The remaining consultations were for followup visits, chronic disease management, vaccination, or administrative issues. Only consultations with diagnostic content were considered for this study.
Without explicitly being asked, in 49 out of 134 (36%) consultations with diagnostic content, GPs reflected on the individual utilisation threshold. Nine of the 12 GPs (75%) did this at least once. In 125 out of 134 (93%) consultations GPs stated that they knew their patient well or very well. Although prompted to reflect on their diagnostic thinking in general, they were not asked ‘when does a patient usually consult’, that is, the utilisation threshold.
In 48 out of 49 (98%) patients where GPs reflected on their patients’ individual utilisation threshold, patients were previously well-known by GPs.
Different utilisation thresholds were mostly framed as patients consulting ‘early’ or ‘late’ in an illness episode. In the view of participating GPs, attending ‘late’, that is, having a high-utilisation threshold, was associated with a high probability of defined or even serious morbidity, and vice versa:
‘So I do know that she comes [into my surgery] rather early than late.’ (GP3 on patient 1, female, age 42 years) [Identifying information in the brackets after quotes refers, first, to the doctor and, second, to the patient; sex and age of the patient is given. See Appendix 2 for GP information].
‘He is actually a very anxious patient, consulting me promptly … So, if he has got something, he comes rather quickly than too late.’
(GP11 on patient 18, male, 40 years)
‘She is not somebody coming here three times a week, OK? She comes here rarely … [and] if she comes, something is really going on.’
(GP6 on patient 19, female, 82 years)
Additionally, GPs described their patients as consulting regularly with unimportant issues or only when they really worried in order to illustrate low- or high-utilisation thresholds, respectively:
‘She is somebody, consulting me frequently with minor issues.’
(GP1 on patient 6, female, 62 years)
‘She only comes into my surgery, in cases [where] she really is in trouble.’
(GP2 on patient 10, female 74 years)
In just one exception a GP reflected on the utilisation threshold of a patient not personally known by him. That was made possible by a textual comment of a colleague in the patient’s record.
GPs’ perceptions of why patients consult early or late
During interviews, GPs gave explanations or reasons as to what they perceived might cause their patients to consult promptly or late. Five factors were identified: medical history and life-changing (medical) events; emotional and behavioural characteristics; family background and social environment; the media; and external circumstances.
The medical history and life-changing (medical) events of the patient
GPs argued that severe or even life-threatening disease that their patients had experienced in the past played a role in their threshold for seeking medical help. This was mostly found in patients with a more frequent and ‘early’ consultation pattern:
‘She once had leukaemia and chemotherapy … and of course she is a bit anxious now.’
(GP3 on patient 1, female, 42 years)
‘First time he went to the men’s health check with his urologist he was 45 years old being diagnosed with prostate cancer … Now he is driven by fears of a relapse.’
(GP11 on patient 34, male, 54 years)
Emotional and behavioural characteristics of the patient
Some patients were characterised as being generally anxious and sensitive, leading to a low-utilisation threshold:
‘He has dreadful angst.’
(GP3 on patient 5, male, 19 years)
‘… because he is just anxious … and very sensitive.’
(GP11 on patient 34, male 54 years)
Sometimes patients were described as being socially withdrawn, resulting in a high-utilisation threshold:
‘I know that he is a withdrawn person, not willing to seek the doctor’s help … It’s difficult to get anything out of him … He is a close-lipped person, as all men are.’
(GP5 on patient 4, male, 72 years)
‘She never shows up … because she doesn’t want to interrupt my work. That’s what she said to me once.’
(GP7 on patient 15, female 76 years)
Some patients presented not only ‘late’, but had also previously missed follow-up appointments:
‘One or two years ago she did manage not to attend for an INR (anticoagulation check) for more than 3 months.’
(GP11 on patient 2, female 75 years)
The patient’s family background and social environment
GPs pointed out the importance of the patients’ family and/or their social environment as an influence on their patients’ consultation habits.
Issues of trouble and stress in their surrounding family were associated with both high- and low-utilisation thresholds, depending on how patients were affected by them:
‘There is a lot of trouble [in the family].’ (GP11 on patient 2, female 75 years) was a comment in a patient with usually ‘late’ consultations.
However, the following was stated for patients with ‘early’ consultation habits:
‘He really had hard times, after the death of his partner in an accident … [Since then] he comes quite often for minor colds etc, somehow he appears like a child looking for love.’
(GP3 on patient 15, male, 45 years)
‘Now, 10 days ago her elder sister, only a little bit older than herself, died from a heart attack. And of course that totally threw her off the rails.’
(GP7 on patient 12, female, 40 years)
Both events apparently led to frequent visits with a low likelihood of serious morbidity.
GPs complained about a sometimes weak lay system [‘lay system’ here refers to self-care strategies and family advice] for managing minor health problems in patients with low-utilisation thresholds:
‘Often I ask myself: “Why is he coming to my surgery?” If there was a grandmother around, she might say “Stay in bed, I’ll make a cup of tea for you and it’ll all turn out fine tomorrow.”’
(GP3 on patient 15, male, 45 years)
The media
One GP mentioned that the media influenced susceptible people, which lowered their utilisation threshold.
‘She picks up from the media what might happen.’
(GP3 on patient 1, female, 42 years)
The external circumstances of the patient
GPs reported simple external constraints leading to lower-utilisation thresholds such as needing sick notes from the first day of illness, which in Germany is required for blue-collar workers (blue collar refers to someone of working-class extraction, historically defined by hourly rates of pay and manual labour):
‘She was just in need of a sick note for today.’
(GP11 on patient 31, female, 28 years)
Pre-emptive consultations apparently also occurred, by which patients prepared themselves for exceptional events, such as holidays or examinations:
‘If she didn’t need to leave for Berlin tomorrow, she wouldn’t have shown up, I guess.’
(GP2 on patient 6, female, 53 years)
The cognitive, emotional, and behavioural consequences for GPs when reflecting on individual utilisation thresholds
Although the concept of the individual utilisation threshold was described almost identically by most GPs touching on this topic, their related beliefs, feelings, and actions varied widely, depending on the particular situation.
GP cognition
In patients with high-utilisation thresholds, GPs showed an elevated level of awareness and concentration. They reported making an extra effort not to miss the prospect of serious disease:
‘He is not frank … and I have to be attentive to find out everything.’
(GP7 on patient 15, female, 76 years)
‘I have known her for many years … and have known her lack of compliance. [So] I’m very alarmed … so I think I have to dig deeper into it, not to miss anything.’
(GP7 on patient 6, female, 84 years)
Furthermore, the GPs pondered about why patients are withdrawn and whether or not there was a serious problem behind a seemingly innocuous presentation:
‘Probably he just raises this [trivial] problem as a pretext for coming here … but he is not yet willing to tell me what the matter is.’
(GP5 on patient 4, male, 72 years)
With patients who had a low-utilisation threshold, GPs described struggling to take the patient and their complaint seriously:
‘I really need to concentrate on taking their complaints seriously, as the patient expects me to … This symptom wouldn’t bother me at all … but I want to make him feel … that he is being looked after.’
(GP6 on patient 10, female, 72 years)
GP emotion:
When being confronted with patients consulting ‘early’ (in an illness progression), some GPs grew irritated or even reacted aggressively:
‘She was getting on my nerves … I’m glad to see her leave again.’
(GP1 on patient 10, female, 83 years)
‘When he comes here, I’m a little annoyed.’
(GP7 on patient 16, male, 67 years)
However, they could see this group of patients in a more positive light when they became aware of circumstances impacting on their patients’ utilisation behaviour:
‘Knowing his background [of prostate cancer], I do understand [his behaviour of frequent visits for minor problems].’
(GP11 on patient 34, male, 54 years)
GP behaviour
In patients with low-utilisation thresholds, GPs did not expect there to be a severe disease outcome. Thus they stressed the benign course of the symptoms mentioned and tried to reassure their patients. Sometimes they performed a physical examination or arranged for followup visits purely for this purpose. However, they usually refrained from specialist referrals or invasive investigations:
‘I try to explain the situation to her. The purpose behind me examining her abdomen is calming this patient down, rather than expecting anything there.’
(GP3 on patient 1, female, 42 years)
One GP became irritated and referred their patient to a specialist, so that he did not have to deal with her any more:
‘I gave her a referral, so I will have my peace …’
(GP6 on patient 13, female, 34 years)
Of the GPs who were aware of individually varying thresholds, most adapted their diagnostic decision making as described earlier. Two GPs, however, stressed the risks that might arise from this kind of individualisation. They deliberately tried to ignore individual variation in consultation thresholds and followed the diagnostic procedures required for the symptoms or problem at hand:
‘I really had the feeling: “that was exaggerated!” but still I do my differential diagnosis and have a further look.’
(GP5, patient 9, female, 63 years)
‘However, we will check the blood … and a stool sample … so that we won’t miss anything.’
(GP6 on patient 10, female, 72 years)
In patients with high-utilisation thresholds, GPs put more time and effort into their diagnostic work, probing for more information or referring to specialists. They tried actively to convince their patients to keep appointments:
‘I know he has some [serious issue], when he drops into my surgery … [Though I didn’t find anything on my examination] I finally referred him to a urologist.’
(GP11 on patient 10, male, 47 years)
‘Sometimes my nurses let me know when she [the patient] is in to get herself a new prescription: “You haven’t seen her for a while.” So I go to the reception desk and ask her if she wants to see me. But mostly she refuses and leaves the practice.’
(GP7 on patient 15, female, 76 years)
The above is an example of a GP believing a patient might harm herself by not consulting, that is, having a high-utilisation threshold and possibly putting her health at risk by neglecting the presence of a serious illness. The GP not only considers the patient’s utilisation threshold, but in this case also actively intervenes to influence consultation behaviour.
DISCUSSION
Summary
In more than one-third of interviews, GPs spontaneously reflected on their patients’ individual utilisation thresholds. These were usually framed as patients tending to consult ‘early’ or ‘late’ in the course of their illness, that is, having a low- or high-utilisation threshold, respectively. Patients with high-utilisation thresholds caused increased awareness or even alarm when consulting. In patients with low-utilisation thresholds, however, the likelihood of serious disease was regarded as low. In these cases GPs provided reassurance and mostly refrained from further investigations. GPs elaborated on their patients’ utilisation thresholds early in the interview, without being prompted and in considerable detail. This therefore appears to be a highly relevant part of GPs’ diagnostic decision making in patients they know well.
Strengths and limitations
The fact that consultations were video-recorded may have influenced GPs’ and patients’ behaviour. Nevertheless, this approach is more suitable for capturing real-world decision making than other techniques such as standardised patients in laboratory settings or GPs’ written self-reports.
In Germany primary care practices are small with high rates of short consultations (on average 7 minutes).15 Familiarity with patients is perhaps greater than in large practices with a more developed division of labour. This may have an impact on the personal continuity of care and should be kept in mind when comparing these results with other healthcare systems.
Reflection on someone else’s thinking requires time and a quiet space. Interviewing during the busy practice day might impede this process. Thus the GP’s surgery schedule was generously planned to allow for sufficient time for interviewing and reflection directly after each consultation. Moreover, practice partners helped out and saw patients waiting to see a GP where possible in order to prevent time pressure during interviews.
The idea that individual utilisation thresholds are relevant for GPs’ diagnostic reasoning was not initially under consideration when planning the study. The authors therefore present a secondary post-hoc analysis. There was, on the other hand, no risk of influencing participating GPs’ elaborations in the direction of the findings presented here. It might therefore be assumed that active questioning would have resulted in an even higher proportion of GPs referring to utilisation thresholds of the kind described in this study. Some GPs, for example, might use the technique without mentioning it. And, if they had been asked ‘Do you take into account your patients’ usual utilisation threshold?’, that would probably have led to a higher proportion of GPs admitting to use this. But this is an assumption requiring verification.
Comparison with existing literature
Only a small proportion of people with medical complaints present to health professionals. Most health problems are managed in the lay sector16,17 (by patients themselves and their families). The decision to consult a physician is multifaceted and highly complex.18 The term ‘illness behaviour’ defines ‘the ways in which given symptoms may be differentially perceived, evaluated and acted (or not acted) upon by different kinds of persons’.19 This study shows that GPs reflect on their patients’ transition from the lay to the professional sector and adapt their decision making accordingly.
Explanations for the individual thresholds provided by participating GPs are in line with findings from direct research of health services utilisation: family, social, and cultural background;20–22 emotional and behavioural characteristics;23,24 medical history and individual experiences;25 media influences;26,27 and external constraints28 have been shown to impact on utilisation behaviour. The current study’s findings thus provide a link between these psychological and sociological surveys, on the one hand, and the clinical decision-making literature, on the other.
The group of patients attending ‘early’ in an illness progression includes those characterised as ‘frequent attenders’,29,30 or a minority even as ‘heartsink patients’.31,32 The factors mentioned in studies as lowering utilisation thresholds partly overlap with the characteristics of frequent attenders in primary care.33
Although the problem of frequent attenders is usually framed as a burden for individual practices and the healthcare system,34 the reflections of the GPs in this study shed a positive light on the phenomenon. The variability in patients’ utilisation patterns allows for conclusions regarding the probability of serious illness. Patients leading to a GP’s ‘heart sink’ were found in this study at some points, for both patients with low- and high-utilisation thresholds. This provides a link between the phenomenon of heartsink patients and patients with high- or low-utilisation thresholds. However, it cannot be extrapolated from this study that heartsink patients only belong to one of the two groups.
Implications for research and practice
Given the design constraints of this study, it might be assumed that GPs use their experience of patients’ previous utilisation behaviour even more often than in a third of cases. This may be a universal feature of diagnosis in settings where the proportion of known patients is high. Whether this is true remains to be investigated in future and larger confirmatory studies. Future studies could investigate different healthcare systems, settings such as emergency departments and out-of-hours services, or could deliberately include unknown patients in order to allow direct comparisons. Moreover, it would be interesting to learn about the patients’ perspective on this phenomenon.
In diagnostic accuracy studies, manipulating the threshold of a diagnostic test leads to different positive predictive values. All other parameters being equal, a high threshold results in a high positive predictive value, and vice versa.35 In a similar fashion, GPs estimate the likelihood of relevant disease as high or low, depending on the threshold defined by the patient. GPs seem to adapt their behaviour accordingly by, for example, providing further diagnostic work-up or reassurance.
Although GPs often pride themselves for the continuity of care they provide, in the area of diagnosis this notion has been somewhat vague in the past. The reflection on individual utilisation thresholds provides a mechanism for how knowing a patient reasonably well promotes diagnostic reasoning.
Whether decision making is improved by using this convenient and time-saving strategy, however, remains to be investigated further.
Acknowledgments
The authors would like to thank the physicians and patients who participated in this study. Thanks to George K Freeman (Emeritus Professor of General Practice, Imperial College London) and Genevieve Allum (GP at Brighton Health Centre) for helpful comments on a previous draft. Matthias Michiels-Corsten presented parts of this article at the COGITA working group of the European General Practice Research Network, Marburg, Germany, in March 2015 and at the Annual Conference of the German College of General Practitioners and Family Physicians (DEGAM), in Bolzano, Italy, September 2015.
Appendix 1. Interview guideline for research staff
Introduction at the beginning of the session
As we have mentioned before, the objective of this study is to find out how physicians arrive at their [diagnostic] decisions. We are interested to learn how you personally make your decisions; this is not about ‘right’ or ‘wrong’. We are going to videotape each consultation. After each consultation I am going to ask you a few questions. Please elaborate on all the thoughts, feelings, and impressions you had during the consultation. Try to avoid any judgement regarding an idea or impression as being ‘important‘ or ‘unimportant‘, even if seems as if it had no rational justification or appears unimportant. After the interview I will ask you to fill in a brief questionnaire on the medical background of the patient.
Interview after each consultation
We will go through the consultation together. Please try to tell me about all the the thoughts, emotions, and impressions you had [it had been made clear before that these referred to diagnostic reasoning]. Please do not judge; just tell me all that went through your mind, even if it seems irrational or not important.
1. Please go back to the beginning of the consultation and describe what happened initially. Why did the patient come to see you?
2. After the patient had mentioned the reason for his/her visit, did you have an explanation by then already, perhaps more than one [an explanation for the signs and symptoms of what constitutes a diagnosis]? If not, when did the first explanation (a hypothesis for a diagnosis) come to your mind?
Was there a particular piece of information (or several) inducing the hypothesis?
Please elaborate on the information (cues).
What was your final diagnosis? What made you favour this particular one?
If there were several explanations, when did you consider each [explanations for the patients’ complaints and symptoms]?
3. Try to describe the kind of hypotheses (specific versus general, somatic versus psychological).
4. What were your thoughts on how to proceed from there?
5. Was there anything else during this consultation that you have not mentioned but might be important here?
6. At which point in time did you decide that you had sufficient information?
7 Optional: If interviewer has observed any of the GP’s behaviour not addressed during interview, ask for elaboration.
Appendix 2. GP characteristics (N = 12)
GP (ID) | Sex | Age, years | Years in practice | Practicea | Settings |
---|---|---|---|---|---|
1 | M | 52 | 20 | Group | Urban |
2 | F | 51 | 23 | Single | Urban |
3 | M | 48 | 9 | Group | Urban |
4 | F | 51 | 12 | Group | Small-town |
5 | M | 62 | 28 | Group | Rural |
6 | M | 49 | 12 | Group | Rural |
7 | M | 59 | 25 | Group | Urban |
8 | M | 62 | 28 | Group | Urban |
9 | F | 58 | 30 | Group | Rural |
10 | F | 54 | 17 | Group | Urban |
11 | F | 57 | 26 | Group | Rural |
12 | M | 53 | 20 | Single | Small-town |
↵a Group practice or single-handed practice.
Notes
Funding
The German Research Foundation (DFG) funded this study (DFG DO 513/2-1).
Ethical approval
Approval was granted by the Ethics Commission of the Faculty of Medicine, University of Marburg (Study 39/10).
Provenance
Freely submitted; externally peer reviewed.
Competing interests
The authors have declared no competing interests.
Discuss this article
Contribute and read comments about this article: bjgp.org/letters
- Received November 9, 2016.
- Revision requested December 22, 2016.
- Accepted January 24, 2017.
- © British Journal of General Practice 2017