The patient path to diagnosis of atrial fibrillation: a qualitative study in primary care

Symptom appraisal

Most participants experienced symptoms before diagnosis. Participants described symptoms such as breathlessness, reduced exercise tolerance, dizziness, fatigue, and palpitations, and they varied in severity. The range of symptoms described by patients is presented in Supplementary Box S5. Patients with acute onset of symptoms perceived them as important and presented to accident and emergency (A&E) where AF was detected. On the other hand, the majority of patients who experienced chronic or non-specific symptoms such as palpitations, dizziness, reduced exercise tolerance, or breathlessness, attributed them to old age or pre-existing conditions and therefore did not act on them:

‘Oh I was feeling terribly dizzy. I couldn’t even stand up, even if I did stand up I fell back on the settee, and I was like that for about 2 hours. And of course we had to get the ambulance out.’ (Patient [P]011, female, aged 83 years)

‘Yeah, I used to get out of breath a little bit. Not every time, but I thought well I’m mid 70s or whatever, I’ve got to expect it. It didn’t bother me. So I just rested and it corrected itself … it didn’t cause me any trouble and I could still carry on.’ (P035, male, aged 77 years)

Despite being aware of the onset of symptoms many did not infer illness or a need to see a healthcare professional:

‘Why bother? I didn’t feel ill or anything like that so you’re mithering people if you do. I’m not that kind, not with people from medics. I won’t do it.’ (P055, male, aged 80 years)

Some patients took the opportunity to mention the symptoms when attending the GP for another reason:

‘I was seeing my GP about something else … I had an attack of gout in the summer, first time ever, so I was having a chat to the GP about that and did I want medication … And I just thought oh while I’m here I’ll just mention I’ve been getting rather breathless.’ (P012, female, aged 73 years)

Lack of awareness

Lack of awareness of AF was a key bottleneck in the appraisal process and a driver of delay in seeking help from a healthcare professional. Most participants had not heard of AF before diagnosis and did not have knowledge of the signs or symptoms. Patients did not associate symptoms such as breathlessness and reduced fitness with the heart, and chest-related symptoms did not align with their understanding of heart conditions.

There was also a lack of awareness of the significance of an irregular pulse, for example, one patient had observed their blood pressure monitor flash ‘irregular heartbeat’ for a couple of years without taking action. Another patient was aware of having an irregular pulse as described below:

‘I feel my pulse at certain times. I’m aware that it just doesn’t feel constant. But then maybe that’s normal, I don’t know.’ (P035, male, aged 77 years)

Lack of patient awareness also emerged in clinician interviews and a couple of GPs described how this impacts the consultation:

‘And I think language is important too because not everybody knows what palpitations are. I’ll ask them [patients] if they have any palpitations and I can just tell they don’t have a clue what that means. Some of them may say “I don’t know what you mean Doc”. Some of them will just say “No” because they don’t want to be embarrassed. I think some — again, it’s important to explain to the patient what you mean by a palpitation.’ (HCP04)

Patients’ experiences in primary care

Interactions with primary care was a central theme in patients’ narratives. Patients described two types of interaction: instances where they had initiated a GP appointment as a result of their symptoms and instances where AF had been detected incidentally. Patients who initiated a GP appointment were motivated by concern that the symptoms could be serious because they kept recurring and/or were having an impact on their day-to-day activities:

‘The last year I had a fit of feeling quite uncomfortable. I had quite uncomfortable, well chest discomfort really is the best way I can describe it. And then took a couple of aspirin and that went away but for the next few weeks I had some really tiresome problems [that] would go on for, oh 3 or 4 hours, which was unheard of for me, and I also felt really tired and breathless and that made me go to the doctor.’ (P030, female, aged 74 years)

Patients’ experiences when they presented to a GP with symptoms ranged from the GP immediately suspecting and detecting AF, to instances where GPs initially attributed symptoms to comorbid conditions. Some patients reported feeling ‘fobbed off’ or ‘dismissed’, and one patient consequently consulted a cardiologist privately.

The pathway for patients with paroxysmal AF was particularly complex; patient narratives describe ‘churn’ in the search for a diagnosis involving several visits to the GP. One patient who was being treated as having AF did not have a definitive diagnosis as their Holter test did not detect AF and they were awaiting further monitoring. Electrocardiogram (ECG) tests were done in the practice, but all patients who required Holter tests were referred to secondary care, where they would encounter wait times of up to 6 months. There were no differences in symptoms reported by male and female patients, but all the patients who reported feeling dismissed or who experienced a churn in their journey to diagnosis were female.

Patients also described incidental detection of symptomatic and asymptomatic AF during routine appointments in primary care. Incidental detection usually related to annual hypertension reviews; however, one diabetic patient’s AF was detected by a podiatrist during a routine foot examination. There were also examples of missed opportunities to detect AF in primary care. One patient reported a previous irregular pulse detected in primary care that was not investigated.

The majority of cases of AF were detected in primary care, but a few patients had their AF incidentally detected in secondary care during pre-operative assessments or during an inpatient stay. For example, two patients were diagnosed on the ward following a stroke. There were a few participants under the care of a cardiologist for pre-existing cardiac conditions such as valvular heart disease who were diagnosed incidentally following routine investigations.

Clinician knowledge and practice

Interviews with GPs demonstrated that GPs were knowledgeable about AF, its symptoms, and the NICE guidelines on detection and management. Generally, GPs felt they had the knowledge required to detect and manage AF, and kept up to date with changes in the NICE guidelines. Regardless, they reported not having received any formal training on the detection of AF additional to that received in medical school. GPs had contrasting views as to whether they required additional training for AF, with the majority feeling that this was neither required nor feasible. A minority felt training would be useful, and there was no consensus how this should be delivered, with some preferring videos or virtual sessions while others preferred face-to-face sessions.

GPs reported varying practices regarding opportunistic screening through pulse palpation; a few did this routinely while others did this only when they suspected AF. Consequently, the frequency of detecting AF varied, with one GP reporting they detected new AF once every 2 weeks; some GPs hardly detected AF; and one GP’s experience of AF was only through reports from secondary care or incidental findings on ECG. The nurse, on the other hand, reported routine pulse palpation as part of patients’ annual reviews, with a yield of finding a new case of AF weekly to fortnightly:

‘I mean I’d probably hazard a guess of once every three months directly myself. And most often that would be because I’ve felt an irregular pulse, which might be a completely different — the patient might have come there with something else completely different and I’ve felt their pulse, or it may be that patients have come in with more sort of classic symptoms and obviously I’ve gone down that diagnostic route.’ (HCP04)

There were also variations in GPs’ practice in terms of referrals for a Holter test, and the presence of symptoms and/or risk factors often determined which patients were referred for ambulatory ECG monitoring, even though many of the GPs perceived AF to be largely asymptomatic:

‘Yeah, again, I think that’s a problem, because it cannot always be permanent, can it? So if they’re symptomatic, so palpitations usually, I would refer to cardiology for a 24-hour tape or something like that, 72-hour tape, whatever. So persistent palpitations with a normal ECG I’d refer to cardio. But I suppose if the symptoms are a bit more subtle, there’s a danger that they could be reassured without further monitoring if the ECG is normal, you know, you’d kind of rule it out.’ (HCP01)

‘I think for that particular case, because they were older, there were risk factors and I had felt an irregular pulse, I would be doing 24-hour tape without question. I wouldn’t think about that, because I’d want to try and be assured that I wasn’t missing an AF. So those ones are pretty easy for me to decide. They’d be referred.’ (HCP04)

Digital technologies

Pathways to diagnosis included AF detected by digital technologies, most commonly smartwatches. The few patients who had smartwatch-detected AF were also symptomatic, and the smartwatch was the deciding factor in seeing a healthcare professional. Patients with smartwatch-detected AF had a positive response from healthcare professionals, and AF was confirmed with an ECG:

‘I’ve had a Fitbit for a long time, and I changed to a new one because the old one died on me. And it has a heart monitor. And it started sending me messages ... I was a bit hesitant about saying “My Fitbit is telling me”. So I was a bit worried because I thought she [GP] might think “I did five years at medical school thank you very much” but she was lovely and she had a look at it … they also got me an ECG. And so again, that confirmed it. And so then I was in the system and they’ve now sent me a letter to say they’re offering me a cardioversion at the end of this month.’ (P033, female, aged 72 years)

These accounts of smartwatch-detected AF aligned with the experiences of primary care clinicians who reported a trend of patients presenting with smartwatch-detected AF, particularly in the more affluent areas. GPs with such encounters reported that they usually request ECGs, which usually confirm AF, although one GP was wary about taking all smartwatch-detected AF seriously. The nurse reported that patients regularly came in with smartwatch-detected AF. One patient and two GPs reported using KardiaMobiles to detect AF. The patient was advised to purchase a KardiaMobile to monitor for AF as ECG had not detected AF and there was a long wait for a Holter test in their area. Both GPs found it very useful to have the KardiaMobile at hand during consultations and used their personal devices. A third GP mentioned they were trying to convince GP partners at their practice to purchase a cardio monitor:

‘I’m also a big fan of the cardio monitors. I use that daily, certainly weekly, because I think it’s a handy tool. And the number of patients I’ve identified with atrial fibrillation, probably in the last 6 months, has probably been about four, five patients. And that’s incidental atrial fibrillation.’ (HCP03)

One of the practices was using a novel digital technology, FibriCheck, an app that measures the heart rhythm by placing the finger on the camera of a smartphone.

Constraints of AF detection in primary care

GPs acknowledged that non-specific symptoms, such as palpitations and breathlessness, were associated with a range of conditions that could include AF:

‘Yeah, I mean something like shortness of breath, because it’s such a wide range of causes, ECG isn’t normally, if it’s just shortness of breath, isn’t normally near the top of my list as a test to organise.’ (HCP07)

‘These days we see complex comorbidity, so somebody with COPD [chronic obstructive pulmonary disease], with shortness of breath that’s got worse isn’t uncommon. And in amongst that I suppose you’ve got to think “Is this just their COPD getting worse? Is there an infection?” or actually “Is there something completely new that’s happening that we don’t know about?”’ (HCP04)

GPs’ accounts alluded to variations in diagnostic capability and capacity. Some practices had an external organisation interpret their ECGs whereas some practices had to interpret them in-house, and were not always confident of their interpretation skills:

‘I think from a practice point, having that easy access to the diagnostics, whether we need training around those particular diagnostics, particularly the ECGs. Many GPs are probably very deskilled in reading ECGs … Yeah. So we have a machine but we don’t have a member of staff trained.’ (HCP06)

In addition, a lack of access to Holter tests in primary care led to delays in diagnosis, leading some GPs to start anticoagulation while patients waited for their Holter tests. It was suggested having access to the Holter test in primary care would facilitate AF detection:

‘Nowadays it can be several months [until a patient has access to a Holter test] which is not ideal if you’ve got somebody flipping in and out of AF. It’s not ideal at all. I have to say — I’ve got a KardiaMobile.’ (HCP08)

‘I think we need something, Primary Care, it would be useful if we could fit more 24-hour ECGs, 72-hour ECGs. Because we can do it for blood pressure machines. So we can send patients home with a blood pressure machine attached for 24, 48 hours. But we can’t with ECG.’ (HCP05)

Interviews revealed the widespread use of remote consultations in primary care meant GPs were not able to examine patients opportunistically. In addition, there was concern that some practices were conducting patients’ annual hypertension reviews remotely using self-reported readings. Other limitations included using mechanical blood pressure devices as opposed to a manual blood pressure machine with a stethoscope, and difficulty getting GP appointments. Healthcare assistants were reported to be taking on more responsibility such as blood pressure checks, again reducing clinician opportunities to detect AF. Some primary care clinicians reported they had trained their healthcare assistants to do manual pulse palpation, but this did not appear to be standard practice. Workload pressures meant it was not always possible to do an ECG immediately, thereby missing the window of opportunity when patients are in AF.