Clinical risk in remote consultations in general practice: findings from in-COVID-19 pandemic qualitative research

Risk domain 1: Practice set-up and organisation

During the first wave of the COVID-19 pandemic, infection control measures to minimise virus transmission (for example, one-way flows and consulting room closures) often reduced the number of available appointments (risk 1A in Table 2), especially for face-to-face assessment. Everyone seeking care was required to go through triage and pre-assessment, risking inefficiency owing to ‘double-touches’ with patients (risk 1B). Premises that were too small to support infection control measures comfortably had knock-on consequences for patients and staff. This was exemplified in one practice where patients had to speak through an intercom at the front door before entering the waiting room, creating a risk to confidentiality.

As the COVID-19 pandemic eased, this situation improved somewhat, but in the interviews with patients some felt they were not a priority candidate for care, and did not even try to contact their practice (risk 1D).

Those who did attempt to access care were sometimes unsuccessful (risk 1C), for a variety of reasons relating to the following: digital exclusion, including inability to use the required technology; blocked phone lines, owing to the high call volumes (and unwillingness to persist in attempts to connect); limitations on patients’ telephone data package (lack of phone minutes or bandwidth for video consultation); and not having a lifestyle that accommodates a remote callback service, as the following quote illustrates:

'They [patients experiencing homelessness] don’t hang around and wait for the phone call at [name of hostel] ... And of course you can’t reach them on a mobile because they normally don’t have one. So there was this concern that we were going to miss this group of patients, but actually it did start to work out all right. And then the patients realise they can still come and knock on the door. […] It helped them a bit and it wasn’t so daunting. But it was a learning curve for all of us.' (RBD practice manager SR)

Notably, some vulnerable groups — particularly people with physical disabilities and some with mental health problems — were positive about changing to remote consulting, reducing the need to travel and wait in crowded waiting rooms. One nurse interviewee reported that she had reviewed long-term conditions in some people who had not attended for in-person reviews for many years, implying that the need to travel to and wait in the surgery had been a barrier to engagement with care.

The technologies for remote triage and consulting were more or less reliable depending on the practice’s digital maturity (risk 1E). For some, video consultations were not available at all. In others, video infrastructure was in place but if a glitch occurred, some staff struggled to cope and created workarounds, causing stress for patients and staff alike (see below, risk 6A). In practices where staff were able to switch between technologies when necessary, this could add value to consultations but swallowed up time:

'There was a patient who phoned with abdominal pain and fever and she couldn’t speak English very well. So, we eventually got a ... we got an interpreter to do a three-way conversation who spoke to her over the phone, navigated her how to get into Near Me and then we did a video assessment of her at home. She had a ... I think one of her family members was there who was able to do some self-examination or directed examination for us to feel her stomach . And it was very time consuming, but it ended up ... we ended up admitting her as well through the ... using Near Me.' (HFVC33 GP Scotland)

Risk domain 2: Communication and the clinical relationship

The dataset contained many examples of incomplete, inaccurate, or misunderstood communication between patient and clinician as a result of consulting remotely (risk 2A). In the context of telephone consultations using the LanguageLine interpreting service, with no visual clues available, some patients struggled to describe their condition and GPs were sometimes uncertain about whether they had understood a patient’s problems correctly. On occasions, GPs reported long waits for a LanguageLine interpreter, leaving them feeling rushed once the consultation started.

GPs talked of the consultation becoming less rich, more transactional, and more awkward in nature, characterised by less active listening and less attention to the emotional dynamics of the interaction (risk 2B) as the following comment illustrates:

'… And I think it’s all a wee bit false and artificial … on the video … probably a wee bit more conscious of myself and my body language and all that sort of stuff . And that there’s less … I think there’s less informal chats.' (HFVC06 GP)

There were other accounts of the pressures faced by patients during remote consultations, which risked the quality or accuracy of information communicated to the clinician:

‘I have a patient who is deaf, she is 30 years old. We called her husband. We have got consent that we can contact him. So once we called him up to speak about a problem, and he also has autism and he just couldn’t cope with my questioning. I was asking too many … it sent him to a little overload and he couldn’t provide the answers.' (Practice, RBD practice nurse GI)

A related, more long-term risk is that without the depth of information and understanding gained from an initial face-to-face consultation, building and maintaining the therapeutic relationship becomes more difficult (risk 2C). While some GPs reported experiencing a good rapport with patients through remote consultations, this could be harder for patients with complex health or social situations. An initial in-person consultation was considered important to assess patients holistically, including the family and home situation, in order to establish rapport. As one GP explained:

‘… generally if it’s complex patients who were deteriorating, I would do a telephone, I would do some investigations, bloods, and then I would do a home visit myself. And that didn’t change it, because I couldn’t … really … I did it with video a couple of times. But … I did a home visit, because I wanted to do a holistic assessment of them and examine them myself. And also just that building that trust and relationship in the first instance.' (RBD GP KS)

Risk domain 3: Clinical care

The majority of remote consultations in general practice occurred by phone. Some GPs reported that the time needed to explain to patients how to join a video call was not worth the additional information gained, so they abandoned this medium. Those who did use video commented that it sometimes provided clinically crucial information.

Clinicians were also concerned about missed diagnoses owing to loss of non-verbal cues, which could alert them to clinical problems (risk 3A) or safeguarding concerns (risk 3B) that went unreported by patients. To some extent these risks could be mitigated by taking a full clinical history. Several interviewees argued that the changes wrought by the COVID-19 pandemic had restored the art and the importance of history-taking, including checking for ‘red flag’ symptoms indicating serious acute problems:

'I thought they had back pain and they’d given me a generally sensible history of back pain with … so, no leg weakness, no urinary symptoms, no bowel symptoms; I would say, "Look, you know, it sounds like you’ve got back pain; it should get better after a couple of weeks … However ..." — and this is where I do the red flag, "If this, this and this happens ... so leg numbness, bowel weakness, bladder weakness", and in the main I’d follow that up with a text message and a patient information leaflet, so that they had the information with them. We would follow that up and make sure it was all documented just to make sure.' (HFVC19 GP Wales)

The clinical consultation is a dialogue, in which every utterance of one party is in response to, and in anticipation of, an utterance from the other party. For this reason, the asynchronous nature of online consultations, while effective for some clinical problems, posed a risk of inadequate assessment (and hence incorrect diagnosis). It was not uncommon to find that the patient’s given reason for consulting differed from their most important clinical symptoms and detailed probing — over the phone — was needed to identify their main problem:

'[On the text or email] maybe a vague symptom that doesn’t really allow you to respond … and it sort of forces you to phone up and say “well, actually sorry I didn’t quite follow what you’re saying about that tummy ache”, and then you realise that actually there was a bit more going on there and you didn’t respond to that with that text about the eczema because that’s not really what they wanted to talk about, it’s that hidden agenda thing. That is no small part of what we do as GPs and it’s all a bit unspoken and difficult to quantify.' (RBD GP CK)

The loss of precise appointment times noted above led to patients being called in workplaces and shops, on public transport or at home surrounded by other people. In some cases this created a risk that patients might withhold or distort the clinical or personal information provided (risk 2A). Several people also raised safeguarding concerns about situations where abusive partners or family members were present but out of sight during a consultation, controlling the information given to the doctor (risk 3B).

An obvious risk of a remote consultation is that elements of care, which need physical co-presence, particularly physical examination, become impossible or logistically awkward, leaving the clinician without some key information needed to reach a diagnosis. As the COVID-19 pandemic progressed, some practices reinstated face-to-face appointments and invited patients in for a physical examination if needed. Others continued with fewer or no face-to-face appointments, filling gaps in information they would have obtained from physical examination through blood tests and other investigations or through referral to a specialist. Such an approach had mixed success, as the following quote illustrates:

'Thinking about the last two or three cancers that we’ve picked up. Quite honestly, I asked myself had I examined this patient face to face would I have picked up the cancer. And the answer is no. You know what, once [in a different patient] I did, yes, the patient had an enlarged liver. And that prompted me to do a two-week wait referral and … then it was an ovarian carcinoma with metastasis. But a lot of things I can think of that it is investigation that picks up so and fills in some gaps to the lack of face to face. Having said that, of course you will … you will jump into the examination if you feel that the history prompts you to do an examination then you have to do an examination.' (RBC GP SQ)

This quote raises the possibility that invasive tests and imaging may be invoked as a substitute for a face-to-face examination (risk 3C), as another GP reflected:

'So I will give you one example, which is, if I am managing somebody and I can’t go and see them, or if I can’t send them to hospital, I would if you would ask me, hypothetically … how would you manage? I’d say, I would love an investigation.' (RBD GP KS)

Remote care also brought a risk of overtreatment — notably, the tendency to prescribe more antibiotics for suspected infections, partly for medicolegal reasons — and over-referral to specialists (risk 3C). One GP put it thus:

'You were dealing with so many unknowns. You were having to deal with patients remotely. Patients that you haven’t seen face-to-face. The uncertainty, the unknown level, the stress level was quite high. And it did worry me about the medicolegal issues. The medicolegal risks ... I think that if you ask most people, they were prescribing more than they would normally do. Because you haven’t seen the patient you are going to prescribe amoxycillin antibiotics for a cough in case it’s a chest infection. It might not be a chest infection, but if it is and you don’t prescribe and they go on to have pneumonia then that’s not good.' (RBD GP IP)

Risk domain 4: The patient’s role in their own care

The dataset included many examples of patients who struggled to make judgements about whether or not to seek care; navigate often complex care pathways with digital entry-points; convey their symptoms in the digital environment; and monitor their own illness and use equipment such as blood pressure monitors or oximeters at home (risk 4A). While patients who were less sick and less impaired were often able to achieve these tasks, many patients were not. A patient with several long-term conditions, for example, described the stress of finding the right words to describe her symptoms, particularly if she was having a bad day:

'If I’m non compos mentis then it’s hard to make it clear … I have foggy brain sometimes. I have fatigue most of the time. Today I’ve prepared myself [for the research interview]; I’ve had enough sleep; I’m articulate. If this was a situation where I need support from a GP, I may not be as articulate.' (RBD patient FD )

Some patients reported feeling stressed when asked to examine themselves during a consultation and report back to the GP. Clinicians described having developed ways to guide patients; for example, how to look at their tonsils in the mirror, but it could turn into a prolonged and complicated process, which added to the burden of illness.

Another risk was the disruption of the care routines that involved explaining to patients what was wrong in an adaptive and supportive way (risk 4B). One GP, for example, described how they drew pictures of the bowel to explain diverticulitis to an older man with poor hearing so they could understand what was causing their symptoms. The GP missed this possibility with remote consulting.

Risk domain 5: Population and public health

Population health interventions, such as screening and immunisations, stopped dead at the start of the COVID-19 pandemic (risk 5A). Cervical smears — which can only be conducted in person — were restored at different rates in different practices. Where screening was offered, some patients remained too scared of catching COVID-19 to attend for a preventive check.

In relation to diabetic foot checks, which would normally require in-person assessment, some practices developed methods to undertake a remote, modified check by telephone, taking a history and talking patients through a foot self-inspection and examination. Healthcare assistants and nurses were trained to conduct these assessments. Their perception was that while this did not allow for checking foot pulses, they were otherwise a reasonable alternative to an in-person examination. The finding that remote checks are perceived to be safe accords with other research, but there is as yet no evidence on the effectiveness or safety of this model.¹⁴

Remote consulting also created risks for maintaining the holism of general practice and the role of medical generalists in assessing and treating patients within their family and social context (risk 5B). The transactional nature of remote consulting described above, frequently negated accessing wider contextual information about patients. And while younger GP interviewees with less experience of following patients through their life course expressed less concern about this, older GPs feared losing a wider understanding of patients’ situations.

Risk domain 6: Professional development and wellbeing

As mentioned in risk domains 1 and 4, some clinicians were stressed by the uncertainties of both the COVID-19 pandemic and the remote medium (risk 6A), describing the laborious and inefficient process of getting a full and clear verbal history from a patient remotely, and combing through notes to fill gaps in their understanding about a patient who they had not been able to see. In the quote below, a nurse describes the physical and psychological effects of many months of remote consulting:

'I’m trying to try my best in my consultations, I’m really focusing. I’m really tired from listening on the phone, because it is exhausting. I’m really tired from these Zoom meetings, because it’s very different to have this sort of like overload of, you know, concentration and physically trying to do it, overload of information.' (RBD practice nurse IM)

Stress and burnout have been shown to clearly carry a direct risk to clinical performance,¹⁵ but also increase the chance that a clinician will retire or leave the profession early, creating critical gaps in the workforce.

Even in those who did not report burnout, clinicians faced reduced opportunities for meeting colleagues face to face and seeking their advice about difficult consultations, limiting opportunities for learning from each other (risk 6B):

'… the partners were saying, you know, “try to stay in your room at work as much as possible” … So it was, you know, I don’t want to spread anything, I don’t want to get anything. So I’d come into … and I just stay in my room … And then you feel more isolated because you’re not getting that general little chat about a case or … about how things are.' (RBD practice nurse IM)