Theme 1: access to information and data

Participants reported an overwhelming number of performance measures which did not reflect patient safety but were considered a mechanism for remuneration: “You get fixated on depression because that's what you're being paid for…So you tend to ignore other mental health co-morbidities because depressions the one you're focusing on” (GP1). Individuals working in general practice reported not knowing which harms they should be evaluating to monitor safety: “If they set out really clearly, ‘we believe that these five things would really improve patient safety and so we want you to report to us, every single medication error, every single needle stick injury’…We could then do that. I suppose that's the problem, it's just so wide at the moment in primary care that we're never really sure” (PM2).

GPs simultaneously spoke of too little and too much information on discharge summaries (“Lots and lots of information about various tests that the patients have had but it's information that I actually don't have the expertise to interpret” CCG3). Not knowing what occurs to patients after a referral to district nursing was a concern (“Referral to district nursing—it's like dropping it into a black hole. You don't know if the nurse has ever seen the patient or whether what you've asked to be done has been done” CCG3). There were issues around receiving letters/communication in a format incompatible with the system holding the patient records in the practice, in which case pertinent information was manually entered into the patient's record by practice staff which created a lot of opportunity for error: “If there's ten or fifteen medications which is not uncommon with patients, that could be a really big problem…every possible error, from transcription error on names of medication or dosages, lengths of time that the patient's expected to be on the medication—be it permanent or short-term—loads of room for error on that” (GP7).

Participants from NHS England reported that no core metrics were routinely analysed for safety monitoring (“It's very underdeveloped…The honest answer is we don't have a set of metrics that we look at” NHSE1). Instead, the accessible data were manually scanned for red flags. These data may be discussed at operational group meetings, which take place every 1 or 2 months, but these meetings were described as fixated on trying to get through the information collected through secondary care quality and safety indicators: “We're trying to look at those. There's hundreds. There's literally about two hundred. Three hundred” (CCG7). Participants from management organisations (NHS England, CCG governing boards and the Commissioning Support Units (CSU)) tended to report that safety data (such as serious incidents) and complaints were distributed across and within a number of organisations (“It's distributed across NHS England: the revalidation team, the performance list team, the contract managing team, and so on.”) with recommendations to collate this information in the future.

Theme 2: clarity of policies and guidelines

Across participants, there was no consensus on the meaning of the term ‘patient safety’ in relation to primary care because the concept was considered vague or they described it as everything in the medical process: “It could mean all sorts of things…So it's everything actually. Patient safety is everything we do” (GP5). Additionally, it was not clear what constituted a serious incident, whether reporting was mandatory and where to report them. Participants explained that serious harm and never events had an acute focus and that general practice was comparatively safer: “You think ‘well, compared to that, our risk is zero’ so it feels like an overreaction to follow some of this process” (PM2). Different methods for reporting patient safety incidents (such as emailing or ringing up a local or national team at NHS England, completing an incident report form from NHS England, or anonymously reporting through the National Reporting and Learning System (NRLS)) appeared to result in confusion about which agency the information was received by and which of these methods satisfied mandatory reporting requirements, even among those at the CCG governing board level: “I struggle when I ask the question to get any sense of the mechanisms by which GPs might report, or anybody in general practice might report, the mechanisms by which patients might report their concerns…I have no idea. And my suspicion would be that nobody has any idea” (CCG6).

Some GPs reported having carried out informal safety monitoring evaluations or audits in the past. This type of monitoring was optional, variable and time-intensive: “Looking at your prescribing rates compared to somebody else…GPs are having to do that by hand and that's why they might do it one year, skip it another year…you might be looking at methotrexate and all the anti-tumour drugs that might be prescribed. So you're covering so many areas you do not have time to do every single one. If somebody could do that and just present the data” (GP2). Multiple GPs mentioned that they needed more guidance about which drugs to monitor and the frequency of medication reviews (especially for long-term medications and high-risk drugs) for repeat prescribing. Specifically, they were interested in: “All the drugs that patients take where monitoring is recognised and recommended and then, what are the monitoring intervals and what are the ranges that are acceptable?” (CCG2).

Theme 3: responsibility and action

At the management level, there were conflicting responses about whose responsibility it was to monitor patient safety in primary care. CCG governing body members generally reported that monitoring safety was outside of their remit and lay with NHS England, whereas participants from NHS England saw themselves as part of a collaborative effort with CCG governing bodies and the practice networks. There was mention of the fact that the CCGs do “have this vague responsibility for quality [improvement] in general practice, whatever that's supposed to mean” (CCG3). Participants from NHS England and the CCG governing bodies also reported conflicting responses about who monitors patient safety in urgent care centres and for out-of-hours services, with some GPs stating that it appeared that nobody was monitoring these services: “And urgent care centres are making huge amounts of money but the quality of care—who's questioning that?…Do we have any data on the safety of prescribing or drug errors or prescribing errors in urgent care centres, is anyone looking at that—even out of hours?” (GP6).

Participants from CCG governing bodies and NHS England spoke at length about how GP practices managed their incidents locally and made use of networks or peer groups (of 6–11 practices meeting monthly) to check-up on each other and share information. However, participants working in GP practices reported using these meetings to make sense of recent changes to policies instead of discussing patient safety. Additionally, poor performing surgeries may work together with similar others in order to avoid detection, referred to as “collusion” (CCG8). Almost all participants reported that they had local knowledge about which practices were and were not safe and that the focus of patient safety monitoring should be on these types of practices, with some identifying single-handed or two-handed practices as a cause for concern: “There's a pattern of poor performance in men, over 50, who trained abroad, who didn't train in the UK, and who are single handed, small, very small practice and probably have got poor premises. They're high indicators of underperformance…the trouble is the trained abroad stuff, is politically very sensitive” (NHSE4).

Participants who worked in practices frequently described the difficulty of managing time pressures on an average working day. GPs reported not having the time to fill in incident report forms or conduct safety audits: “Although there are areas where we are asked to collect data, we're just so busy and so stretched that we don't really do it…We hold the minor surgery service and in theory we try to run an audit of, or we try to keep a record of if there's post-operative infections. But actually to do that properly, it's really difficult, so we don't do it properly” (CCG2). Many GPs explained that it was difficult to be safe in a 10–15 min appointment in which patients often brought multiple serious health and social concerns to the same appointment due to having access to care issues: “But I think primary care's really dangerous right now to be honest. I am getting quite near to the feeling that I don't want to carry on doing it” (CCG2).

For a number of reasons, the recommended protocol for dealing with a potentially serious incident was not always followed: “So what I should be doing is logging it on that, sending it off to them. To be honest, almost never happens” (CCG3). Other than lack of time, participants feared blame, organisational and personal repercussions that were amplified if the potentially serious incidents involved a senior GP. Multiple GPs reported the belief that NHS England would not or could not act on the evidence, and that this deterred them: because “the onus is on that GP and so on top of your normal workload, and for the fear of being isolated and victimised, who's going to do that? It's easier to walk away from it” (GP6). The failure to report incidents outside of the practice was attributed to a number of factors, including a workplace culture that mistakes were deemed to be “within acceptable limits even though in fact if one was to have the hard evidence and comparative with what's going on on a national basis, you might find that you are a complete outlier” (GP7).

Summary

Participants’ report of the barriers to monitoring patient safety in primary care are outlined in box 3 and their recommendations for the future are provided in box 4.