Health and Social Care in the Community (2016) 24(4), 411–419
doi: 10.1111/hsc.12221
Patient safety in primary care: incident reporting and significant event reviews in British general practice David Rea
PhD BA
and Sarah Griffiths
MSc BA
Department of Public Health and Policy Studies, College of Human and Health Sciences, Swansea University, Swansea, UK Accepted for publication 16 January 2015
Correspondence David Rea Department of Public Health and Policy Studies College of Human and Health Sciences Swansea University Singleton Park Swansea SA2 8PP, UK E-mail: [email protected]
What is known about this topic
• • •
Patient safety is of growing concern in healthcare. A large volume of errors is likely in primary care settings because this is where most patient contact occurs. There is significant under-reporting of patient safety errors in primary care.
What this paper adds
• •
•
General practitioners exercise selectivity over which errors/ incidents to report. General practitioners are reluctant to report errors which can be addressed within their own practice, but will report when frustrated by other agencies. General practitioners do not wish to undermine patient confidence, within a context of competition, and do not believe error reports are read, learnt from or acted upon.
Abstract Over the past 20 years, healthcare has adapted to the ‘quality revolution’ by moving away from direct provision and hierarchical control mechanisms. In their place, new structures based on contractual relationships are being developed coupled with attempts to create an organisational culture that shares learning and that scrutinises existing practice so that it can be improved. The issue here is that contractual arrangements require surveillance, monitoring, regulation and governance systems that can be perceived as antipathetic to the examination of practice and subsequent learning. Historically, reporting levels from general practice have remained low; little information is shared and consequently lessons are not shared across the general practice community. Given large-scale under-engagement of general practitioners (GPs) in incident reporting systems, significant event analysis is advocated to encourage sharing of information about incidents to inform the patient safety agenda at a local and national level. Previous research has concentrated on the secondary care environment and little is known about the situation in primary care, where the majority of patient contacts with healthcare occur. To explore attitudes to incident reporting, the study adopted a qualitative approach to GPs working in a mixture of urban and rural practices reporting to a Welsh Local Health Board. The study found that GPs used significant event analysis methodology to report incidents within their practice, but acknowledged under-reporting. They were less enthusiastic about reporting externally. A number of barriers exist to reporting, including insufficient time to report, lack of feedback, fear of blame, and damage to reputations and patient confidence in a competitive environment. If incident reporting processes are perceived as supportive and formative, and where protected time is allocated to discuss incidents, then GPs are willing to participate. They also need to know how the information is used, and whether lessons are being learnt from errors. Keywords: general practice, incident reporting, patient safety, primary healthcare, quality and governance, significant event analysis
Introduction Patient safety is an international priority in healthcare, and error reduction can be improved by reporting and learning from errors (WHO 2005). National guidance therefore emphasises the importance of © 2015 John Wiley & Sons Ltd
reporting systems in a culture where individuals are supported to identify and report errors without threat of punitive action and blame (Welsh Assembly Government 2001, DH 2006, National Audit Office 2007). While most attention has been paid to barriers to incident reporting in hospitals (Kingston et al. 2004, 411
D. Rea & S. Griffiths
Evans et al. 2006, Pfeiffer et al. 2010), patients receive most of their healthcare in primary care. Underreporting in primary care is particularly significant (House of Commons Health Committee 2009, NPSA 2010) as, although primary care may imply lower risks for patients, the large volume of contacts in this sector suggests that safety incidents can be expected to occur (Harmsen et al. 2010). While many errors are only mild or moderately serious, some pose significant threats to patients, and all can undermine patient confidence (Makeham et al. 2002, Britten et al. 2000). Despite this, research into incident reporting in this setting is scarce and largely outside the UK (Dovey et al. 2002, Elder et al. 2006, Mikkelsen et al. 2006). Generally, hospitals have well-established incident reporting systems (Dovey et al. 2002). Staff in the UK can report any safety incident using an e-form to describe the incident. These are uploaded to the local risk management system (DATIX) and automatically collated by the National Learning and Reporting System at the National Patient Safety Agency (NPSA). Less is known about reporting systems within general practice. General practitioners (GPs), unlike staff employed by a Local Health Board (LHB), cannot report directly onto the LHB electronic system due to licensing issues. Paperbased forms are therefore submitted for entering onto the LHB system by LHB staff. The ‘one size fits all’ nature of largely hospital-based reporting systems – because they do not take into account the different infrastructure and diversity of general practice services – is officially a significant cause of under-reporting, alongside ‘blame culture’ and fear of litigation (House of Commons Health Committee 2009). Blame and fear seemed unlikely to be a prime cause of under-reporting because most GPs are independent practitioners, contracted to provide services to the NHS on terms defined by the general medical services (GMS) contract. Therefore, the aim of this study was to explore GP attitudes, beliefs and perceptions of their under-reporting. For readers outside the UK, we should explain the GMS contract includes provision for voluntary participation in the Quality and Outcomes Framework (QOF) which rewards practices for the provision of ‘quality care’ and helps to fund further improvements in the delivery of clinical care. Significant Event Audits (SEA) are a structured means of learning from patient safety incidents and ‘near misses’, using guidelines issued by the NPSA, and have been included within the QOF since 2004 (NPSA 2008). NPSA guidelines state a practice should report (either 412
to the National Reporting and Learning Service, or via the primary care trust/healthcare organisation) those events where patient safety has, or could have been, ‘compromised’ and ‘strongly recommends that SEAs should be routinely undertaken by primary care teams’ [our emphasis] (NPSA 2009). In short, a significant event or a safety incident is defined by practice. For the purposes of this paper, the terms ‘event’ and ‘incident’ will only refer to those occurrences where GPs have acknowledged a risk to patient safety that merits analysis. Errors, as such, are not formally reported. Since 2009, health service delivery in Wales has been the responsibility of seven LHBs and three Trusts. LHBs are responsible for delivering all healthcare services within a geographical area. This includes ensuring the governance arrangements of services with which it holds contracts to provide general medical services. Knowledge from incident reporting systems can provide the LHB with assurance of the safety of these services. This paper aims to contribute to an understanding of how these governance and contractual arrangements may be interpreted within general practice, and how this affects decisions to record and analyse incidents and report to the LHB.
Method Study design Qualitative methods were suited to our purpose because we wanted to understand GPs’ perceptions of the barriers to incident reporting and whether the process of significant event analysis supports the reporting of incidents. Under-reporting is a potentially sensitive subject and reluctance to openly discuss this subject with someone unknown to them might be expected. The study was therefore limited to one LHB where trusting relationships were well established. The first-named author supervised the project and participated in data analysis. The second-named author, a governance manager within the LHB, conducted and transcribed the interviews. The challenge of separating the second author’s occupational and performance management roles from the research role was acknowledged from the outset. Any suggestion of bias was minimised because the interviews were detailed and lengthy. Although there was potential for role conflict, it was believed that GPs would not be willing to participate unless they were already in a familiar and trusting relationship with the interviewer. © 2015 John Wiley & Sons Ltd
Patient safety in primary care
Setting and sample As independent businesses, GPs are self-employed, so permission to access this group from the LHB or other organisations was not required. The LHB’s Research and Development Department, and the Local Research and Ethics Committee both agreed the proposal did not require their consideration. Instead, it was scrutinised by the appropriate Research Ethics Committee at our University and adapted following its recommendations. The research results were relevant to local governance arrangements, so the Locality Director was contacted and subsequently provided support to the research being undertaken. Following written confirmation of the purpose, the secretary of the Local Medical Committee also confirmed their support. Each Practice Manager received details of the research to cascade throughout their practice. These individuals were identified as ‘gatekeepers’ to access participants: not in providing permission to access their GPs but as a means of recruiting participants. This strategy proved successful in establishing links into the practices. Managers also proved useful in arranging suitable times for interviews, which aided the smooth running of all interviews due to dedicated time slots without interruption. Because the aim was to explore attitudes, beliefs and perceptions of under-reporting by GPs, we opted for a purposive sample of those GPs with low incident reporting rates (Denscombe 2007, Parahoo 2007). As of October 2011, 78 GPs were identified as partners in a practice and holding a contract with the LHB to provide general medical services. GPs who were locum practitioners and Registrar GPs were excluded. All 78 GPs were contacted via email with details that explained the study aims, and an invitation to participate in a confidential semi-structured interview lasting no longer than 60 minutes. Taking time out for interviews might impact significantly on consultations, so causing GPs to refuse to participate. Therefore, flexibility was needed in organising meetings, and the importance of providing clear information about the research and the maximum length of the interview was paramount. No direct refusals were received and 17 GPs agreed to participate. Workload pressures on general practice and the dynamics of the GP/LHB relationship suggested that a low response rate might be expected. The sample was purposely selected from the number of respondents to the invitation on the basis that they were best placed to answer the research questions by meeting the inclusion criteria: © 2015 John Wiley & Sons Ltd
1 The participants were practising GP partners within a practice in the Locality as of October 2011. 2 The practice had reported incidents to the LHB, but not within the 12 months prior to January 2012. 3 The practice took part in SEA for the QOF process. Of the 17 GPs who first agreed to participate, the selection meant that 9 were interviewed. Of these, four GPs came from practices with less than four partners; five of these came from practices having four or more partners. Three interviewees practised in rural areas and six in urban practices. They were interviewed over a 3-month period. This was a small sample but sufficient for this purpose (Guest et al. 2006), and was also diverse in terms of practice size and location. While recommended sample size for qualitative research varies, little extra data can be expected after 30 interviews and sample size should generally be kept small to ensure depth (Carpenter 1995). The choice of sample size in qualitative research is not a statistical concern (Silverman 2010). Data saturation principles were reached and overall the sample size exceeded recommended minimums for qualitative data (Bernard 1995, Cresswell 2009). Data collection Interviews were conducted within the workplace in a manner intended to ensure participants had a comfortable environment, so they did not feel restricted in sharing information (Turner 2010). Informality contributed to the rapport that developed during the interview. This is illustrated in the lengthy and candid accounts that reflected their personal beliefs and practices rather forming an acceptable ‘public’ account. Interviews were digitally recorded and transcribed in full following each interview and before the next, and although time consuming, this increased reliability in subsequent interviews because the interviewer’s approach and questioning could be improved (Denscombe 2007, Bryman 2008, Hansen 2008).
Data analysis A simple descriptive approach was taken to analysis of the data (Sandelowski 2010) where content analysis was performed for each transcript. An initial process of coding involving both researchers was undertaken after the interviews were completed, when the data were fresh. This entailed repeatedly reading and reflecting on each script to identify emerging themes. All scripts were analysed and 413
D. Rea & S. Griffiths
coded individually before comparing them to look for similarities and differences. These were categorised to identify common themes in the data. Themes were grouped into key themes and sub-categories and a table of themes was produced. To ensure that the coding process was reliable, the transcripts were revisited later and re-coded by the second researcher, independently of the first coding exercise. Agreement was found initially in over 90% of cases with a final consensus on all items. Creating and adhering to a standard coding scheme increased the validity of the research. This is a conventional approach to qualitative content analysis with any study designed to describe phenomena when existing theory or research literature is limited (Hsieh & Shannon 2005). Preconceived categories and theoretical perspectives were avoided, with the intent that categories would emerge from the data. Thematic analysis goes further than just analysing the frequency of codes by placing the analysis in the context of what was said and allows for in-depth exploration of themes in relation to that context (Joffe & Yardley 2004). Furthermore, using direct quotations from participants is a powerful way of bringing the findings alive and reinforcing their validity (Coombes et al. 2009).
Results Thematic analysis of the interviews resulted in identification of four themes and these are used to organise this section. The general practice reporting environment To understand the under-reporting of incidents to external bodies such as the LHB, it is important to understand how incidents are first reported, recorded and analysed within practices, and that practices exercise discretion over whether incidents are sufficiently serious to merit analysis, or to be reported to the LHB. GPs showed a positive attitude towards reporting incidents within their own practice. They all recorded incidents and could describe their practice’s reporting systems, but there was wide acknowledgement (n = 6) that not all incidents were reported. All knew how to access an incident reporting form and what to do with it when completed. The degree of formality of incident reporting systems varied across the sample. More (n = 5) GPs reported that they had specific procedures in place to review incidents, and quarterly meetings were the norm. Two GPs did not have a meeting structure in place, choosing to discuss incidents with colleagues 414
informally, while two GPs discussed significant events as part of weekly practice meetings. All GPs felt equipped with the appropriate skills to carry out an analysis of an incident. None of the GPs expressed concerns about their ability to report and analyse an incident/significant event. Incident reporting systems within practices were paper based, with electronic forms to be downloaded when required and other manual methods to capture incidents. One GP questioned the value of formal meetings as they required participation of busy practice colleagues, and should occur soon after an incident. Another regretted the lack of team involvement in his practice: . . .significant events are supposed to involve the whole of the practice but in this practice we do them as individuals. Most of them, I reflect on them myself, not with colleagues. We do discuss some at practice level but not as many as I would have liked to. (GP6)
While all GPs said they recorded incidents, three quarters (n = 6) of interviewees could not remember the last time they had reported an incident within their practice. They did not think an incident had occurred that warranted reporting. Only two of the interviewees had reported to the NPSA. There was some confusion about whose role it was in the practice or whether it was the LHB’s responsibility. For instance: It’s not up to me. You (LHB) ask for significant events and the most serious of them are sent to you. I don’t personally make the decision. The Practice Manager deals with all that. It would be up to her if they were reported further. (GP4)
Interviewees had a poor knowledge of the function of the NPSA and what types of incident should be reported to the NPSA. The common belief was that NPSA undertook a regulatory role and that only more serious incidents would need to be reported to the NPSA. A supportive and open working environment facilitated the reporting and discussion of incidents. There was an appreciation that the fear of being blamed for things going wrong could be perceived as a barrier to the reporting of incidents, but was not an issue for any of the interviewees questioned. They all stated that they felt able to discuss incidents within their practice, albeit exercising selectivity with whom they would share incidents with: It’s supposed to be a no blame exercise. Within this practice we do but in other practices it might be more critical. Different doctors might take a more critical slant – I’m sure they do, I can think of a few! Blame is not an issue here. (GP3) © 2015 John Wiley & Sons Ltd
Patient safety in primary care
Value of incident reporting There was complete agreement that incident reporting was important. This related to learning from deviations in care and was the over-riding purpose for reporting and analysing incidents and provided motivation for GPs to complete incident reports. The majority of respondents (n = 8) stated that they could always learn something from even the most trivial significant event. There was general acceptance that incidents will occur, usually relating to a mistake being made. GPs commented on the emotive value of the process enabling them to personally reflect on their behaviour: You might feel a bit guilty if you’ve made a mistake – how could I have been so stupid? So speaking it through with colleagues makes you feel a lot better. (GP7)
The requirement within appraisal to demonstrate that GPs have reflected on their practice to identify areas for improvement was identified as a motivator to produce reports by more than half of the interviewees (n = 6). The appraisal process was also attributed with encouraging a culture of openness, where incidents were discussed freely without recrimination. Incident reporting and significant event analysis was seen as useful for self-awareness and self-improvement, with some satisfaction derived from sharing it with colleagues for some, but for others, personal reflection was more important. Some GPs suggested there was value in providing patients with evidence that the practice was dealing with incidents in an appropriate manner. Others (n = 3) felt threatened by patients being aware of adverse incidents; the effect it would have on the doctor–patient relationship and concerns around patients understanding of the circumstances. Attitudes to incident reporting There was acceptance that incidents could be expected to occur, and unanimous agreement that incidents and significant events were the same: GPs did not differentiate between how they were to be dealt with. There was consensus that analysis was the important part of incident reporting and it was important to establish why they had occurred and what could be done to prevent recurrence. Most GPs acknowledged that incidents might involve all aspects of providing general medical care – administrative incidents as well as clinical incidents: © 2015 John Wiley & Sons Ltd
Anything that has gone a bit wrong that might affect patient safety. Something administrative, something clinical. (GP2)
Most interviewees were not clear what constituted an incident/significant event and they relied on personal judgement to interpret what should be reported: I’d only report those things if I made a mistake because of workload, if it was too much, then I would report it. (GP9) I would say that most people do not know what a significant event is. (GP4)
Most (n = 8) interviewees commented that reporting an incident depended on its severity and the extent of harm caused to the patient. They were more likely to report an incident that had a detrimental effect on the patient than where no harm was caused. Minor errors were discussed informally with the intention to inform other colleagues that an incident had taken place, rather than formal analysis. The severity of an incident would also determine how quickly the incident would be discussed. Less severe incidents would be collated in readiness for the next significant event meeting, whereas incidents that had a severe impact on a patient would be dealt with quickly: If it was something minor that wasn’t going to cause any harm, it might need discussing but wouldn’t need to go any further or anything. (GP3) Some mistakes where no harm has been done, you might think keep it quiet. No one wants trouble. You don’t want to be answering in front of a panel or things like that. (GP9) If it is an individual error, and you know you need to change yourself and if it hasn’t caused any harm then you just want to learn yourself, so no point in sharing with others. (GP1)
GPs were more inclined to report incidents of a clinical nature, including prescribing and diagnostic errors. Communication issues between primary and secondary care was cited as the most common incident type, particularly in relation to the information flow between the two sectors. The most commonly raised barrier relating to incident reporting concerned the heavy workload and lack of time that typifies general practice. The ‘busyness’ of general practice and the clinical role always having to take priority were given as reasons for not having enough time to report and analyse incidents properly. Finding the time for meetings posed practical difficulties for some, especially when GPs were part-time: 415
D. Rea & S. Griffiths
I’d love to have the time to have regular significant event meetings, but the workload in this practice is extremely high and things do get pushed to one side. (GP4) There isn’t any time to do all these things. There are so many things that take priority over the other. That’s the biggest obstacle for reporting incidents. (GP6)
Apart from the time and effort involved, GPs also talked about their fears. For a minority, fear of embarrassment of loss of face was identified as a barrier to reporting: I think it’s not wanting to look foolish, people think you are stupid. (GP2)
All interviewees acknowledged that fear of blame affected whether an incident would be reported. None of the GPs stated that fear would affect their own personal reporting decisions within the practice because the environment promoted a culture conducive to the sharing of incidents: I used to be a GP appraiser and lots of GPs felt that they could never, ever bring anything to significant event meets because they felt embarrassed, felt they would be shamed and shouted at. (GP4) I don’t think I would be blamed if I reported an error but I can understand the fear of a tarnished reputation. (GP8)
While it was evident that all GPs reported and discussed incidents and significant events within their practices, frequent reporting to the LHB was less common. The most common barriers relating to reporting to the LHB concerned a lack of understanding of who to report incidents to and lack of feedback, which had been exacerbated by re-organisation of the LHB and change of LHB personnel: We actually had to ask where to send them to, if you’re going to be wanting to report an event, then at the very least you need to know where it is to go to avoid any delays or misdirected reports. (GP2)
The absence of feedback on incidents reported to the LHB would affect a GP’s decision whether to report to the LHB again. GPs questioned the point of reporting incidents and whether it had any effect on services: I’ve not seen anything particularly positive come from the formal process . . . never hear anything else about them. . . After I’ve sent it, what does happen to it? It puts you off reporting. I think what’s the point? It seems absolutely futile. (GP5) I’ll tell you why I don’t report to the LHB, you get no feedback and you feel it’s just gone off and that’s it. Does it make any difference and I have a feeling it doesn’t. So why waste my time doing it? (GP4)
416
Impact of contracting arrangements All but two of the interviewees were aware of the LHB’s incident reporting requirements, and the format for reporting prescribed by the LHB. None of the interviewees were able to confirm what happened once an incident was reported to the LHB or what the information is used for. There were some concerns that the practice lost control over the incident by reporting to the LHB, and suspicion over what the LHB would do with the information: It’s a bit like turkeys voting for Christmas . . . If something is reported that is unprofessional, I don’t think the LHB would just stop at that and just let it go. (GP1)
The general perception was that as sub-contractors, GPs are responsible for what goes on within their practice; should be sorted out within the practice; and the LHB does not need to know, particularly if there were practice issues that could be resolved internally. Others perceived the role of the LHB as more than ‘policing’ and valued the input of the LHB if it was separate from the performance management process: There does need to be an open and safe atmosphere for reporting and that reporting is separate from the disciplinary side of things as far as reasonably possible. (GP2)
Most interviewees felt that the LHB could support practices in analysing incidents and implementing changes. The majority (n = 8) felt that the LHB could have a role in compiling incidents with a view to sharing the trends and lessons learnt. Providing peer support was also suggested as a role for the LHB. The independent organisational nature of general practice was identified as having an influence over why incidents are not reported. Others felt that the application of different business models affected how solutions put in place in other practices could be generalised: GPs have historically been very territorial and very secretive . . . we are in competition with each other as separate commercial organisations. Practices don’t want others to know we’ve pooped in our own nest and if they don’t tell each other, then they are certainly not going to tell the LHB. (GP1)
GPs exercised their own judgement when determining what kind of incidents would be reported. There was a perception that mundane practice issues would not be of interest to the LHB: I don’t think you need to know about every SEA because they are peculiar to the practice, so not all of them are anything that the LHB can do anything about. (GP5) © 2015 John Wiley & Sons Ltd
Patient safety in primary care
If they are a bit mundane practice issues, systems and things in a practice, or a blip in a chemist or something, you feel they lack the importance to go to a wider forum. (GP2)
Some GPs (n = 5) stated that they used the process to report incidents to the LHB that were beyond their control, with the rationale that the influence and support of the LHB could make service changes where the practice could not, particularly if those incidents involved other agencies: If it is something that is related to the services in the community which we can’t change ourselves, then yes, we would love the LHB’s support, for example an urgent suspected cancer patient has not been seen for 3 months then that is not acceptable. (GP2) I’ve used significant events to point out fragmentation of the service, whereas patients with quite severe mental illness have been left. (GP5) If there was a secondary care problem, I’d report as you need all the help you can get to change the system there. (GP4)
Similarly, one GP used the NPSA process for the same reason: We were experiencing poor relations between primary and secondary care and the quality of communications that were coming out of secondary care and we were all trying to push the message that there were issues that certain issues were not getting summaries back to us. (GP3)
The support of the LHB was valued in providing protected time to enable the practice to analyse incidents, although two respondents felt that not enough time was allocated. Sessions that took place within the practice provided an opportunity to hold incident reporting/SEA meetings, without the distraction of day-to-day practice issues. There was consensus the LHB could do more to support their analysis by discussing common incidents, enabling the sharing of experiences and potential solutions. The LHB could be of assistance in promoting significant event/incident reporting, including providing reminders to continue using the process; a short anonymous newsletter in an electronic format to illustrate common themes occurring from incidents reported; and co-ordinating the sharing of solutions between practices by providing external facilitation for meetings. All GPs participated in the QOF, including the two significant event analysis indicators. There was disagreement between the interviewees as regards how this process contributed to the reporting of incidents, and whether the QOF motivated them to carry out this activity. More GPs (n = 5) reported that they © 2015 John Wiley & Sons Ltd
would still carry out the process without the QOF incentive and that their initial GP training had instilled this way of working which was continued through the annual appraisal process: I don’t think QOF affects it really. We do SEA because it’s a professional thing to do. (GP3)
The QOF process added a structure to the process, providing an emphasis on a required number of significant events, but this meant that only the required number of significant events would be reported to the LHB consistent with the QOF year-end and not routinely throughout the year: If we didn’t have to keep count, then it would just be done more in informal discussions, but we would still do it of some sort, definitely. (GP8)
In contrast, some interviewees (n = 4) felt that QOF was the prime motivator due to its financial rewards, and only carried out a process of completing incident forms in order to satisfy contract monitoring requirements: I write them down because I am paid to do so and would probably not write them down otherwise. It’s not a high priority for me. (GP6)
Discussion Main findings This study has shown that incident reporting is carried out within general practices as SEA. GPs value the process, despite the time it takes, when they perceive benefits to themselves and their practice. GPs select whether something is worth recording as a significant event, and worth analysis and reporting, according to a range of factors identified in this study. Fear of disciplinary action and fear of litigation were not acknowledged as an obstacle to reporting, but potential embarrassment and losses to patient confidence were. GPs identified concerns about what the information would be used for and some distrusted the LHB’s motives. Again, fear of litigation and blame were not spoken of as much as the reputation of the practice, especially in a competitive environment. This concern was not among those acknowledged in the House of Commons report Patient Safety (House of Commons Health Committee 2009). There was a lack of knowledge of how, why and when incidents should be reported to external organisations. Furthermore, confusion about what informa417
D. Rea & S. Griffiths
tion and what type of incident/event should be reported were perceived as barriers to reporting as well as a lack of feedback. Only two GPs reported to the NPSA; all other GPs demonstrated poor awareness of the NPSA and its process. The QOF was identified as the main motivator for reporting significant events formally, although most interviewees suggested that in its absence, they would still identify and discuss incidents albeit without the structure and formality the QOF process provides. The GPs viewed the main function of the LHB as assisting the implementation of incident/significant event analysis simply by maintaining a focus on this activity, providing reminders and sharing of information between practices. The provision of protected time was seen as essential for practices to undertake the process effectively. This study has highlighted variation in how GPs manage the reporting of incidents and under-reporting, particularly to external organisations. Barriers to reporting within the practice included the time it took to report and discuss an incident, fear of embarrassment and ambiguity of definition. Barriers to external reporting included fear of blame, damage to reputation and patient confidence, lack of clarity over who to report to and lack of feedback. This research appears to suggest that GPs are selective about which staff members attend meetings. The nature of significant events is the main determinant of which other staff members would be involved. Greater levels of involvement by other staff groups in significant event discussions were reported only if required by the GP partners, with only two practices routinely discussing incidents in multidisciplinary staff meetings. Practice managers and receptionists were more likely to be involved in the analysis of ‘administrative’ significant events than GP colleagues. Similarly, GP colleagues of respondents were more likely to be involved in the analysis of ‘clinical’ significant events. Comparison with other studies Earlier research has indicated that SEA may facilitate learning and quality improvement, but lack of time, training or confidence has been barriers (Pringle et al. 1995). There is evidence however that the SEA method of analysis lacks reliability (Bowie et al. 2008). Evidence on reporting is less common. Reporting systems are generally unpopular, costly, unable to produce epidemiological data, and it is uncertain they lead to cultural change (Dovey & Wallis 2011). The study reported here shows also that reporting processes are hindered by lack of time, lack of ade418
quate technology, lack of appropriate reporting processes that could be understood by GPs, and that lack of feedback for learning was a significant barrier to them being seen as worthwhile. Furthermore, the study reported here suggests that designated significant event meetings were the most common forum for discussing incidents, while evidence suggests it was unusual earlier (Bowie et al. 2004). The implementation of structured meetings could be attributed to the requirement of the QOF process for significant events (including participants at the meeting) to be documented. Strengths and limitations The study design demonstrates that GPs are willing to discuss how they manage SEA and reporting in circumstances where an open and trusting relationship between interviewer and interviewee was established. This would be difficult to achieve in a larger study, although further work would be desirable in other study areas. Furthermore, general practice necessarily involves uncertainty over risks and consequences. There is no objective definition of when an event demands analysis: this is determined by GPs from their perspective alone.
Conclusion The study has shed light on GPs’ perceptions of the value of reporting, and the variation arising from selective reporting. It raises questions about the effectiveness of reporting processes. Specifically, this research found that incident reports were used to reinforce claims for change in other organisations. The GPs were candid about using the incident reporting process to highlight deficiencies in secondary care, over which they felt they had no control. Furthermore, we cannot assume that if GPs are prepared to analyse a significant event, they will be able to do so effectively and consistently (NPSA 2010, De Wet et al. 2011). Although GPs in this study claimed they are competent to carry out incident reporting and analysis, this is not necessarily the case as the degree of analysis varied. Despite the limitations, the implications of this study are clear: we cannot be assured that patient safety is optimal and risks reduced until GPs have robust data collection systems. We would also recommend that purchasers need to view reporting as a process where reports are clearly being read: the focus should be on learning, and this should be evident to GPs. © 2015 John Wiley & Sons Ltd
Patient safety in primary care
References Bernard H.R. (1995) Research Methods in Anthropology. AltaMira, Walnut Creek, CA. Bowie P., McKay J., Norrie J. & Lough M. (2004) Awareness and analysis of a significant event by general practitioners: a cross sectional survey. Quality and Safety in Health Care 13, 102–107. Bowie P., Pope L. & Lough M. (2008) A review of the evidence base for significant event analysis. Journal of Evaluation in Clinical Practice 14 (4), 520–536. Britten N., Stevenson F.A., Barry C.A., Barber N. & Bradley C.P. (2000) Misunderstandings in prescribing decisions in general practice: qualitative study. British Medical Journal 320, 484–488. Bryman A. (2008) Social Research Methods, 3rd edn. Oxford University Press, Oxford. Carpenter D.R. (1995) Phenomenology as method. In: H.J. Streubert & D.R. Carpenter (Eds) Qualitative Research in Nursing, pp. 72–96. Lippincott, Philadelphia, PA. Coombes L., Allen D., Humphrey D. & Neale J. (2009) Indepth interviews. In: J. Neale (Ed.) Research Methods for Health and Social Care, pp. 197–210. Palgrave MacMillan, Basingstoke. Cresswell J.W. (2009) Research Design – Qualitative, Quantitative and Mixed Methods Approaches, 3rd edn. Sage, London. De Wet C., Bradley N. & Bowie P. (2011) Significant event analysis: a comparative study of knowledge, process and attitudes in primary care. Journal of Evaluation in Clinical Practice 17, 1207–1215. Denscombe M. (2007) The Good Research Guide: For SmallScale Social Research Projects, 3rd edn. Open University Press, Maidenhead. DH (2006) Safety First – A Report for Patients, Clinicians and Healthcare Managers. Department of Health, London. Dovey S.M. & Wallis K.A. (2011) Incident reporting in primary care: epidemiology or culture change? BMJ Quality and Safety 20, 1001–1003. Dovey S.M., Meyers D.S., Phillips R.L. Jr, Green L.A., Fryer G.E., Galliher J.M. & Grob P. (2002) A preliminary taxonomy of medical errors in family practice. Quality and Safety in Health Care 11 (3), 233–238. Elder N., Pallerla H. & Regan S. (2006) What do family physicians consider an error? A comparison of definitions and physician perception. BMC Family Practice 7, 73. Evans S.M., Berry J.G., Smith B.J., Esterman A., Selim P., O’Shaughnessy J. & DeWit M. (2006) Attitudes and barriers to incident reporting: a collaborative hospital study. Quality and Safety in Health Care 15, 39–43. Guest G., Bunce A. & Johnson L. (2006) How many interviews are enough? An experiment with data saturation and variability. Field Methods 18 (1), 59–82. Hansen E.C. (2008) Successful Qualitative Health Research, A Practical Introduction. Open University Press, Maidenhead. Harmsen M., Gaal S., van Dulmen S., de Feijter E., Giesen P., Jacobs A. & Wensing M. (2010) Patient safety in Dutch primary care: a study protocol. Implementation Science 50
© 2015 John Wiley & Sons Ltd
(5). Available at: http://www.implementationscience.com/content/5/1/50 (accessed on 13/2/2015). House of Commons Health Committee (2009) Patient Safety, Sixth Report of Session 2008/09, Volume 1. House of Commons Health Committee, London. Hsieh H. & Shannon S.E. (2005) Three approaches to qualitative content analysis. Qualitative Health Research 15 (9), 1277–1288. Joffe H. & Yardley L. (2004) Content and thematic analysis. In: D.F. Marks & L. Yardley (Eds) Research Methods for Clinical and Health Psychology, pp. 56–68. Sage, London. Kingston M.J., Evans S.M., Smith B.J. & Berry J.G. (2004) Attitudes of doctors and nurses towards incident reporting: a qualitative analysis. Medical Journal of Australia 181 (1), 36–39. Makeham M., Dovey S., County M. & Kidd M. (2002) An international taxonomy for reporting general practice error in Australia and five other countries. Medical Journal of Australia 177, 68–72. Mikkelsen T.H., Sokolowski I. & Olesen F. (2006) General practitioners’ attitudes toward reporting and learning from adverse event: results from a survey. Scandinavian Journal of Primary Health Care 24, 27–32. National Audit Office (2007) Improving Quality and Safety – Progress in Implementing Clinical Governance in Primary Care: Lessons for the New Primary Care Trusts. National Audit Office, London. NPSA (2008) Significant Event Audit – Guidance for Primary Care Teams. National Patient Safety Agency, London. NPSA (2009) Seven Steps to Patient Safety in General Practice. National Patient Safety Agency, London. NPSA (2010) Extracting Data and Sharing Learning from Significant Event Audits in General Practice. National Patient Safety Agency, London. Parahoo K. (2007) Nursing Research, Principles, Process and Issues, 2nd edn. Palgrave MacMillan, Basingstoke. Pfeiffer Y., Manser T. & Wehner T. (2010) Conceptualising barriers to incident reporting: a psychological framework. Quality and Safety in Healthcare 19 (6). DOI:10.1136/ qshc.2008.030445 Pringle M., Bradley C.P. & Carmichael C.M. (1995) Significant Event Analysis: A Study of the Feasibility and Potential of Case-Based Auditing in Primary Medical Care. Royal College of General Practitioners, London. Sandelowski M. (2010) What’s in a name? Qualitative description revisited. Research in Nursing & Health 33 (1), 77–84. Silverman D. (2010) Doing Qualitative Research, 3rd edn. Sage, London. Turner D.W. (2010) Qualitative interview design: a practical guide for novice investigators. The Qualitative Report 15 (3), 754–760. Welsh Assembly Government (2001) Improving Health in Wales – A Plan for the NHS with its Partners. Welsh Assembly Government, Cardiff. WHO (2005) WHO Draft Guidelines for Adverse Event Reporting and Learning Systems. World Health Organisation, Geneva.
419
doi: 10.1111/hsc.12221
Patient safety in primary care: incident reporting and significant event reviews in British general practice David Rea
PhD BA
and Sarah Griffiths
MSc BA
Department of Public Health and Policy Studies, College of Human and Health Sciences, Swansea University, Swansea, UK Accepted for publication 16 January 2015
Correspondence David Rea Department of Public Health and Policy Studies College of Human and Health Sciences Swansea University Singleton Park Swansea SA2 8PP, UK E-mail: [email protected]
What is known about this topic
• • •
Patient safety is of growing concern in healthcare. A large volume of errors is likely in primary care settings because this is where most patient contact occurs. There is significant under-reporting of patient safety errors in primary care.
What this paper adds
• •
•
General practitioners exercise selectivity over which errors/ incidents to report. General practitioners are reluctant to report errors which can be addressed within their own practice, but will report when frustrated by other agencies. General practitioners do not wish to undermine patient confidence, within a context of competition, and do not believe error reports are read, learnt from or acted upon.
Abstract Over the past 20 years, healthcare has adapted to the ‘quality revolution’ by moving away from direct provision and hierarchical control mechanisms. In their place, new structures based on contractual relationships are being developed coupled with attempts to create an organisational culture that shares learning and that scrutinises existing practice so that it can be improved. The issue here is that contractual arrangements require surveillance, monitoring, regulation and governance systems that can be perceived as antipathetic to the examination of practice and subsequent learning. Historically, reporting levels from general practice have remained low; little information is shared and consequently lessons are not shared across the general practice community. Given large-scale under-engagement of general practitioners (GPs) in incident reporting systems, significant event analysis is advocated to encourage sharing of information about incidents to inform the patient safety agenda at a local and national level. Previous research has concentrated on the secondary care environment and little is known about the situation in primary care, where the majority of patient contacts with healthcare occur. To explore attitudes to incident reporting, the study adopted a qualitative approach to GPs working in a mixture of urban and rural practices reporting to a Welsh Local Health Board. The study found that GPs used significant event analysis methodology to report incidents within their practice, but acknowledged under-reporting. They were less enthusiastic about reporting externally. A number of barriers exist to reporting, including insufficient time to report, lack of feedback, fear of blame, and damage to reputations and patient confidence in a competitive environment. If incident reporting processes are perceived as supportive and formative, and where protected time is allocated to discuss incidents, then GPs are willing to participate. They also need to know how the information is used, and whether lessons are being learnt from errors. Keywords: general practice, incident reporting, patient safety, primary healthcare, quality and governance, significant event analysis
Introduction Patient safety is an international priority in healthcare, and error reduction can be improved by reporting and learning from errors (WHO 2005). National guidance therefore emphasises the importance of © 2015 John Wiley & Sons Ltd
reporting systems in a culture where individuals are supported to identify and report errors without threat of punitive action and blame (Welsh Assembly Government 2001, DH 2006, National Audit Office 2007). While most attention has been paid to barriers to incident reporting in hospitals (Kingston et al. 2004, 411
D. Rea & S. Griffiths
Evans et al. 2006, Pfeiffer et al. 2010), patients receive most of their healthcare in primary care. Underreporting in primary care is particularly significant (House of Commons Health Committee 2009, NPSA 2010) as, although primary care may imply lower risks for patients, the large volume of contacts in this sector suggests that safety incidents can be expected to occur (Harmsen et al. 2010). While many errors are only mild or moderately serious, some pose significant threats to patients, and all can undermine patient confidence (Makeham et al. 2002, Britten et al. 2000). Despite this, research into incident reporting in this setting is scarce and largely outside the UK (Dovey et al. 2002, Elder et al. 2006, Mikkelsen et al. 2006). Generally, hospitals have well-established incident reporting systems (Dovey et al. 2002). Staff in the UK can report any safety incident using an e-form to describe the incident. These are uploaded to the local risk management system (DATIX) and automatically collated by the National Learning and Reporting System at the National Patient Safety Agency (NPSA). Less is known about reporting systems within general practice. General practitioners (GPs), unlike staff employed by a Local Health Board (LHB), cannot report directly onto the LHB electronic system due to licensing issues. Paperbased forms are therefore submitted for entering onto the LHB system by LHB staff. The ‘one size fits all’ nature of largely hospital-based reporting systems – because they do not take into account the different infrastructure and diversity of general practice services – is officially a significant cause of under-reporting, alongside ‘blame culture’ and fear of litigation (House of Commons Health Committee 2009). Blame and fear seemed unlikely to be a prime cause of under-reporting because most GPs are independent practitioners, contracted to provide services to the NHS on terms defined by the general medical services (GMS) contract. Therefore, the aim of this study was to explore GP attitudes, beliefs and perceptions of their under-reporting. For readers outside the UK, we should explain the GMS contract includes provision for voluntary participation in the Quality and Outcomes Framework (QOF) which rewards practices for the provision of ‘quality care’ and helps to fund further improvements in the delivery of clinical care. Significant Event Audits (SEA) are a structured means of learning from patient safety incidents and ‘near misses’, using guidelines issued by the NPSA, and have been included within the QOF since 2004 (NPSA 2008). NPSA guidelines state a practice should report (either 412
to the National Reporting and Learning Service, or via the primary care trust/healthcare organisation) those events where patient safety has, or could have been, ‘compromised’ and ‘strongly recommends that SEAs should be routinely undertaken by primary care teams’ [our emphasis] (NPSA 2009). In short, a significant event or a safety incident is defined by practice. For the purposes of this paper, the terms ‘event’ and ‘incident’ will only refer to those occurrences where GPs have acknowledged a risk to patient safety that merits analysis. Errors, as such, are not formally reported. Since 2009, health service delivery in Wales has been the responsibility of seven LHBs and three Trusts. LHBs are responsible for delivering all healthcare services within a geographical area. This includes ensuring the governance arrangements of services with which it holds contracts to provide general medical services. Knowledge from incident reporting systems can provide the LHB with assurance of the safety of these services. This paper aims to contribute to an understanding of how these governance and contractual arrangements may be interpreted within general practice, and how this affects decisions to record and analyse incidents and report to the LHB.
Method Study design Qualitative methods were suited to our purpose because we wanted to understand GPs’ perceptions of the barriers to incident reporting and whether the process of significant event analysis supports the reporting of incidents. Under-reporting is a potentially sensitive subject and reluctance to openly discuss this subject with someone unknown to them might be expected. The study was therefore limited to one LHB where trusting relationships were well established. The first-named author supervised the project and participated in data analysis. The second-named author, a governance manager within the LHB, conducted and transcribed the interviews. The challenge of separating the second author’s occupational and performance management roles from the research role was acknowledged from the outset. Any suggestion of bias was minimised because the interviews were detailed and lengthy. Although there was potential for role conflict, it was believed that GPs would not be willing to participate unless they were already in a familiar and trusting relationship with the interviewer. © 2015 John Wiley & Sons Ltd
Patient safety in primary care
Setting and sample As independent businesses, GPs are self-employed, so permission to access this group from the LHB or other organisations was not required. The LHB’s Research and Development Department, and the Local Research and Ethics Committee both agreed the proposal did not require their consideration. Instead, it was scrutinised by the appropriate Research Ethics Committee at our University and adapted following its recommendations. The research results were relevant to local governance arrangements, so the Locality Director was contacted and subsequently provided support to the research being undertaken. Following written confirmation of the purpose, the secretary of the Local Medical Committee also confirmed their support. Each Practice Manager received details of the research to cascade throughout their practice. These individuals were identified as ‘gatekeepers’ to access participants: not in providing permission to access their GPs but as a means of recruiting participants. This strategy proved successful in establishing links into the practices. Managers also proved useful in arranging suitable times for interviews, which aided the smooth running of all interviews due to dedicated time slots without interruption. Because the aim was to explore attitudes, beliefs and perceptions of under-reporting by GPs, we opted for a purposive sample of those GPs with low incident reporting rates (Denscombe 2007, Parahoo 2007). As of October 2011, 78 GPs were identified as partners in a practice and holding a contract with the LHB to provide general medical services. GPs who were locum practitioners and Registrar GPs were excluded. All 78 GPs were contacted via email with details that explained the study aims, and an invitation to participate in a confidential semi-structured interview lasting no longer than 60 minutes. Taking time out for interviews might impact significantly on consultations, so causing GPs to refuse to participate. Therefore, flexibility was needed in organising meetings, and the importance of providing clear information about the research and the maximum length of the interview was paramount. No direct refusals were received and 17 GPs agreed to participate. Workload pressures on general practice and the dynamics of the GP/LHB relationship suggested that a low response rate might be expected. The sample was purposely selected from the number of respondents to the invitation on the basis that they were best placed to answer the research questions by meeting the inclusion criteria: © 2015 John Wiley & Sons Ltd
1 The participants were practising GP partners within a practice in the Locality as of October 2011. 2 The practice had reported incidents to the LHB, but not within the 12 months prior to January 2012. 3 The practice took part in SEA for the QOF process. Of the 17 GPs who first agreed to participate, the selection meant that 9 were interviewed. Of these, four GPs came from practices with less than four partners; five of these came from practices having four or more partners. Three interviewees practised in rural areas and six in urban practices. They were interviewed over a 3-month period. This was a small sample but sufficient for this purpose (Guest et al. 2006), and was also diverse in terms of practice size and location. While recommended sample size for qualitative research varies, little extra data can be expected after 30 interviews and sample size should generally be kept small to ensure depth (Carpenter 1995). The choice of sample size in qualitative research is not a statistical concern (Silverman 2010). Data saturation principles were reached and overall the sample size exceeded recommended minimums for qualitative data (Bernard 1995, Cresswell 2009). Data collection Interviews were conducted within the workplace in a manner intended to ensure participants had a comfortable environment, so they did not feel restricted in sharing information (Turner 2010). Informality contributed to the rapport that developed during the interview. This is illustrated in the lengthy and candid accounts that reflected their personal beliefs and practices rather forming an acceptable ‘public’ account. Interviews were digitally recorded and transcribed in full following each interview and before the next, and although time consuming, this increased reliability in subsequent interviews because the interviewer’s approach and questioning could be improved (Denscombe 2007, Bryman 2008, Hansen 2008).
Data analysis A simple descriptive approach was taken to analysis of the data (Sandelowski 2010) where content analysis was performed for each transcript. An initial process of coding involving both researchers was undertaken after the interviews were completed, when the data were fresh. This entailed repeatedly reading and reflecting on each script to identify emerging themes. All scripts were analysed and 413
D. Rea & S. Griffiths
coded individually before comparing them to look for similarities and differences. These were categorised to identify common themes in the data. Themes were grouped into key themes and sub-categories and a table of themes was produced. To ensure that the coding process was reliable, the transcripts were revisited later and re-coded by the second researcher, independently of the first coding exercise. Agreement was found initially in over 90% of cases with a final consensus on all items. Creating and adhering to a standard coding scheme increased the validity of the research. This is a conventional approach to qualitative content analysis with any study designed to describe phenomena when existing theory or research literature is limited (Hsieh & Shannon 2005). Preconceived categories and theoretical perspectives were avoided, with the intent that categories would emerge from the data. Thematic analysis goes further than just analysing the frequency of codes by placing the analysis in the context of what was said and allows for in-depth exploration of themes in relation to that context (Joffe & Yardley 2004). Furthermore, using direct quotations from participants is a powerful way of bringing the findings alive and reinforcing their validity (Coombes et al. 2009).
Results Thematic analysis of the interviews resulted in identification of four themes and these are used to organise this section. The general practice reporting environment To understand the under-reporting of incidents to external bodies such as the LHB, it is important to understand how incidents are first reported, recorded and analysed within practices, and that practices exercise discretion over whether incidents are sufficiently serious to merit analysis, or to be reported to the LHB. GPs showed a positive attitude towards reporting incidents within their own practice. They all recorded incidents and could describe their practice’s reporting systems, but there was wide acknowledgement (n = 6) that not all incidents were reported. All knew how to access an incident reporting form and what to do with it when completed. The degree of formality of incident reporting systems varied across the sample. More (n = 5) GPs reported that they had specific procedures in place to review incidents, and quarterly meetings were the norm. Two GPs did not have a meeting structure in place, choosing to discuss incidents with colleagues 414
informally, while two GPs discussed significant events as part of weekly practice meetings. All GPs felt equipped with the appropriate skills to carry out an analysis of an incident. None of the GPs expressed concerns about their ability to report and analyse an incident/significant event. Incident reporting systems within practices were paper based, with electronic forms to be downloaded when required and other manual methods to capture incidents. One GP questioned the value of formal meetings as they required participation of busy practice colleagues, and should occur soon after an incident. Another regretted the lack of team involvement in his practice: . . .significant events are supposed to involve the whole of the practice but in this practice we do them as individuals. Most of them, I reflect on them myself, not with colleagues. We do discuss some at practice level but not as many as I would have liked to. (GP6)
While all GPs said they recorded incidents, three quarters (n = 6) of interviewees could not remember the last time they had reported an incident within their practice. They did not think an incident had occurred that warranted reporting. Only two of the interviewees had reported to the NPSA. There was some confusion about whose role it was in the practice or whether it was the LHB’s responsibility. For instance: It’s not up to me. You (LHB) ask for significant events and the most serious of them are sent to you. I don’t personally make the decision. The Practice Manager deals with all that. It would be up to her if they were reported further. (GP4)
Interviewees had a poor knowledge of the function of the NPSA and what types of incident should be reported to the NPSA. The common belief was that NPSA undertook a regulatory role and that only more serious incidents would need to be reported to the NPSA. A supportive and open working environment facilitated the reporting and discussion of incidents. There was an appreciation that the fear of being blamed for things going wrong could be perceived as a barrier to the reporting of incidents, but was not an issue for any of the interviewees questioned. They all stated that they felt able to discuss incidents within their practice, albeit exercising selectivity with whom they would share incidents with: It’s supposed to be a no blame exercise. Within this practice we do but in other practices it might be more critical. Different doctors might take a more critical slant – I’m sure they do, I can think of a few! Blame is not an issue here. (GP3) © 2015 John Wiley & Sons Ltd
Patient safety in primary care
Value of incident reporting There was complete agreement that incident reporting was important. This related to learning from deviations in care and was the over-riding purpose for reporting and analysing incidents and provided motivation for GPs to complete incident reports. The majority of respondents (n = 8) stated that they could always learn something from even the most trivial significant event. There was general acceptance that incidents will occur, usually relating to a mistake being made. GPs commented on the emotive value of the process enabling them to personally reflect on their behaviour: You might feel a bit guilty if you’ve made a mistake – how could I have been so stupid? So speaking it through with colleagues makes you feel a lot better. (GP7)
The requirement within appraisal to demonstrate that GPs have reflected on their practice to identify areas for improvement was identified as a motivator to produce reports by more than half of the interviewees (n = 6). The appraisal process was also attributed with encouraging a culture of openness, where incidents were discussed freely without recrimination. Incident reporting and significant event analysis was seen as useful for self-awareness and self-improvement, with some satisfaction derived from sharing it with colleagues for some, but for others, personal reflection was more important. Some GPs suggested there was value in providing patients with evidence that the practice was dealing with incidents in an appropriate manner. Others (n = 3) felt threatened by patients being aware of adverse incidents; the effect it would have on the doctor–patient relationship and concerns around patients understanding of the circumstances. Attitudes to incident reporting There was acceptance that incidents could be expected to occur, and unanimous agreement that incidents and significant events were the same: GPs did not differentiate between how they were to be dealt with. There was consensus that analysis was the important part of incident reporting and it was important to establish why they had occurred and what could be done to prevent recurrence. Most GPs acknowledged that incidents might involve all aspects of providing general medical care – administrative incidents as well as clinical incidents: © 2015 John Wiley & Sons Ltd
Anything that has gone a bit wrong that might affect patient safety. Something administrative, something clinical. (GP2)
Most interviewees were not clear what constituted an incident/significant event and they relied on personal judgement to interpret what should be reported: I’d only report those things if I made a mistake because of workload, if it was too much, then I would report it. (GP9) I would say that most people do not know what a significant event is. (GP4)
Most (n = 8) interviewees commented that reporting an incident depended on its severity and the extent of harm caused to the patient. They were more likely to report an incident that had a detrimental effect on the patient than where no harm was caused. Minor errors were discussed informally with the intention to inform other colleagues that an incident had taken place, rather than formal analysis. The severity of an incident would also determine how quickly the incident would be discussed. Less severe incidents would be collated in readiness for the next significant event meeting, whereas incidents that had a severe impact on a patient would be dealt with quickly: If it was something minor that wasn’t going to cause any harm, it might need discussing but wouldn’t need to go any further or anything. (GP3) Some mistakes where no harm has been done, you might think keep it quiet. No one wants trouble. You don’t want to be answering in front of a panel or things like that. (GP9) If it is an individual error, and you know you need to change yourself and if it hasn’t caused any harm then you just want to learn yourself, so no point in sharing with others. (GP1)
GPs were more inclined to report incidents of a clinical nature, including prescribing and diagnostic errors. Communication issues between primary and secondary care was cited as the most common incident type, particularly in relation to the information flow between the two sectors. The most commonly raised barrier relating to incident reporting concerned the heavy workload and lack of time that typifies general practice. The ‘busyness’ of general practice and the clinical role always having to take priority were given as reasons for not having enough time to report and analyse incidents properly. Finding the time for meetings posed practical difficulties for some, especially when GPs were part-time: 415
D. Rea & S. Griffiths
I’d love to have the time to have regular significant event meetings, but the workload in this practice is extremely high and things do get pushed to one side. (GP4) There isn’t any time to do all these things. There are so many things that take priority over the other. That’s the biggest obstacle for reporting incidents. (GP6)
Apart from the time and effort involved, GPs also talked about their fears. For a minority, fear of embarrassment of loss of face was identified as a barrier to reporting: I think it’s not wanting to look foolish, people think you are stupid. (GP2)
All interviewees acknowledged that fear of blame affected whether an incident would be reported. None of the GPs stated that fear would affect their own personal reporting decisions within the practice because the environment promoted a culture conducive to the sharing of incidents: I used to be a GP appraiser and lots of GPs felt that they could never, ever bring anything to significant event meets because they felt embarrassed, felt they would be shamed and shouted at. (GP4) I don’t think I would be blamed if I reported an error but I can understand the fear of a tarnished reputation. (GP8)
While it was evident that all GPs reported and discussed incidents and significant events within their practices, frequent reporting to the LHB was less common. The most common barriers relating to reporting to the LHB concerned a lack of understanding of who to report incidents to and lack of feedback, which had been exacerbated by re-organisation of the LHB and change of LHB personnel: We actually had to ask where to send them to, if you’re going to be wanting to report an event, then at the very least you need to know where it is to go to avoid any delays or misdirected reports. (GP2)
The absence of feedback on incidents reported to the LHB would affect a GP’s decision whether to report to the LHB again. GPs questioned the point of reporting incidents and whether it had any effect on services: I’ve not seen anything particularly positive come from the formal process . . . never hear anything else about them. . . After I’ve sent it, what does happen to it? It puts you off reporting. I think what’s the point? It seems absolutely futile. (GP5) I’ll tell you why I don’t report to the LHB, you get no feedback and you feel it’s just gone off and that’s it. Does it make any difference and I have a feeling it doesn’t. So why waste my time doing it? (GP4)
416
Impact of contracting arrangements All but two of the interviewees were aware of the LHB’s incident reporting requirements, and the format for reporting prescribed by the LHB. None of the interviewees were able to confirm what happened once an incident was reported to the LHB or what the information is used for. There were some concerns that the practice lost control over the incident by reporting to the LHB, and suspicion over what the LHB would do with the information: It’s a bit like turkeys voting for Christmas . . . If something is reported that is unprofessional, I don’t think the LHB would just stop at that and just let it go. (GP1)
The general perception was that as sub-contractors, GPs are responsible for what goes on within their practice; should be sorted out within the practice; and the LHB does not need to know, particularly if there were practice issues that could be resolved internally. Others perceived the role of the LHB as more than ‘policing’ and valued the input of the LHB if it was separate from the performance management process: There does need to be an open and safe atmosphere for reporting and that reporting is separate from the disciplinary side of things as far as reasonably possible. (GP2)
Most interviewees felt that the LHB could support practices in analysing incidents and implementing changes. The majority (n = 8) felt that the LHB could have a role in compiling incidents with a view to sharing the trends and lessons learnt. Providing peer support was also suggested as a role for the LHB. The independent organisational nature of general practice was identified as having an influence over why incidents are not reported. Others felt that the application of different business models affected how solutions put in place in other practices could be generalised: GPs have historically been very territorial and very secretive . . . we are in competition with each other as separate commercial organisations. Practices don’t want others to know we’ve pooped in our own nest and if they don’t tell each other, then they are certainly not going to tell the LHB. (GP1)
GPs exercised their own judgement when determining what kind of incidents would be reported. There was a perception that mundane practice issues would not be of interest to the LHB: I don’t think you need to know about every SEA because they are peculiar to the practice, so not all of them are anything that the LHB can do anything about. (GP5) © 2015 John Wiley & Sons Ltd
Patient safety in primary care
If they are a bit mundane practice issues, systems and things in a practice, or a blip in a chemist or something, you feel they lack the importance to go to a wider forum. (GP2)
Some GPs (n = 5) stated that they used the process to report incidents to the LHB that were beyond their control, with the rationale that the influence and support of the LHB could make service changes where the practice could not, particularly if those incidents involved other agencies: If it is something that is related to the services in the community which we can’t change ourselves, then yes, we would love the LHB’s support, for example an urgent suspected cancer patient has not been seen for 3 months then that is not acceptable. (GP2) I’ve used significant events to point out fragmentation of the service, whereas patients with quite severe mental illness have been left. (GP5) If there was a secondary care problem, I’d report as you need all the help you can get to change the system there. (GP4)
Similarly, one GP used the NPSA process for the same reason: We were experiencing poor relations between primary and secondary care and the quality of communications that were coming out of secondary care and we were all trying to push the message that there were issues that certain issues were not getting summaries back to us. (GP3)
The support of the LHB was valued in providing protected time to enable the practice to analyse incidents, although two respondents felt that not enough time was allocated. Sessions that took place within the practice provided an opportunity to hold incident reporting/SEA meetings, without the distraction of day-to-day practice issues. There was consensus the LHB could do more to support their analysis by discussing common incidents, enabling the sharing of experiences and potential solutions. The LHB could be of assistance in promoting significant event/incident reporting, including providing reminders to continue using the process; a short anonymous newsletter in an electronic format to illustrate common themes occurring from incidents reported; and co-ordinating the sharing of solutions between practices by providing external facilitation for meetings. All GPs participated in the QOF, including the two significant event analysis indicators. There was disagreement between the interviewees as regards how this process contributed to the reporting of incidents, and whether the QOF motivated them to carry out this activity. More GPs (n = 5) reported that they © 2015 John Wiley & Sons Ltd
would still carry out the process without the QOF incentive and that their initial GP training had instilled this way of working which was continued through the annual appraisal process: I don’t think QOF affects it really. We do SEA because it’s a professional thing to do. (GP3)
The QOF process added a structure to the process, providing an emphasis on a required number of significant events, but this meant that only the required number of significant events would be reported to the LHB consistent with the QOF year-end and not routinely throughout the year: If we didn’t have to keep count, then it would just be done more in informal discussions, but we would still do it of some sort, definitely. (GP8)
In contrast, some interviewees (n = 4) felt that QOF was the prime motivator due to its financial rewards, and only carried out a process of completing incident forms in order to satisfy contract monitoring requirements: I write them down because I am paid to do so and would probably not write them down otherwise. It’s not a high priority for me. (GP6)
Discussion Main findings This study has shown that incident reporting is carried out within general practices as SEA. GPs value the process, despite the time it takes, when they perceive benefits to themselves and their practice. GPs select whether something is worth recording as a significant event, and worth analysis and reporting, according to a range of factors identified in this study. Fear of disciplinary action and fear of litigation were not acknowledged as an obstacle to reporting, but potential embarrassment and losses to patient confidence were. GPs identified concerns about what the information would be used for and some distrusted the LHB’s motives. Again, fear of litigation and blame were not spoken of as much as the reputation of the practice, especially in a competitive environment. This concern was not among those acknowledged in the House of Commons report Patient Safety (House of Commons Health Committee 2009). There was a lack of knowledge of how, why and when incidents should be reported to external organisations. Furthermore, confusion about what informa417
D. Rea & S. Griffiths
tion and what type of incident/event should be reported were perceived as barriers to reporting as well as a lack of feedback. Only two GPs reported to the NPSA; all other GPs demonstrated poor awareness of the NPSA and its process. The QOF was identified as the main motivator for reporting significant events formally, although most interviewees suggested that in its absence, they would still identify and discuss incidents albeit without the structure and formality the QOF process provides. The GPs viewed the main function of the LHB as assisting the implementation of incident/significant event analysis simply by maintaining a focus on this activity, providing reminders and sharing of information between practices. The provision of protected time was seen as essential for practices to undertake the process effectively. This study has highlighted variation in how GPs manage the reporting of incidents and under-reporting, particularly to external organisations. Barriers to reporting within the practice included the time it took to report and discuss an incident, fear of embarrassment and ambiguity of definition. Barriers to external reporting included fear of blame, damage to reputation and patient confidence, lack of clarity over who to report to and lack of feedback. This research appears to suggest that GPs are selective about which staff members attend meetings. The nature of significant events is the main determinant of which other staff members would be involved. Greater levels of involvement by other staff groups in significant event discussions were reported only if required by the GP partners, with only two practices routinely discussing incidents in multidisciplinary staff meetings. Practice managers and receptionists were more likely to be involved in the analysis of ‘administrative’ significant events than GP colleagues. Similarly, GP colleagues of respondents were more likely to be involved in the analysis of ‘clinical’ significant events. Comparison with other studies Earlier research has indicated that SEA may facilitate learning and quality improvement, but lack of time, training or confidence has been barriers (Pringle et al. 1995). There is evidence however that the SEA method of analysis lacks reliability (Bowie et al. 2008). Evidence on reporting is less common. Reporting systems are generally unpopular, costly, unable to produce epidemiological data, and it is uncertain they lead to cultural change (Dovey & Wallis 2011). The study reported here shows also that reporting processes are hindered by lack of time, lack of ade418
quate technology, lack of appropriate reporting processes that could be understood by GPs, and that lack of feedback for learning was a significant barrier to them being seen as worthwhile. Furthermore, the study reported here suggests that designated significant event meetings were the most common forum for discussing incidents, while evidence suggests it was unusual earlier (Bowie et al. 2004). The implementation of structured meetings could be attributed to the requirement of the QOF process for significant events (including participants at the meeting) to be documented. Strengths and limitations The study design demonstrates that GPs are willing to discuss how they manage SEA and reporting in circumstances where an open and trusting relationship between interviewer and interviewee was established. This would be difficult to achieve in a larger study, although further work would be desirable in other study areas. Furthermore, general practice necessarily involves uncertainty over risks and consequences. There is no objective definition of when an event demands analysis: this is determined by GPs from their perspective alone.
Conclusion The study has shed light on GPs’ perceptions of the value of reporting, and the variation arising from selective reporting. It raises questions about the effectiveness of reporting processes. Specifically, this research found that incident reports were used to reinforce claims for change in other organisations. The GPs were candid about using the incident reporting process to highlight deficiencies in secondary care, over which they felt they had no control. Furthermore, we cannot assume that if GPs are prepared to analyse a significant event, they will be able to do so effectively and consistently (NPSA 2010, De Wet et al. 2011). Although GPs in this study claimed they are competent to carry out incident reporting and analysis, this is not necessarily the case as the degree of analysis varied. Despite the limitations, the implications of this study are clear: we cannot be assured that patient safety is optimal and risks reduced until GPs have robust data collection systems. We would also recommend that purchasers need to view reporting as a process where reports are clearly being read: the focus should be on learning, and this should be evident to GPs. © 2015 John Wiley & Sons Ltd
Patient safety in primary care
References Bernard H.R. (1995) Research Methods in Anthropology. AltaMira, Walnut Creek, CA. Bowie P., McKay J., Norrie J. & Lough M. (2004) Awareness and analysis of a significant event by general practitioners: a cross sectional survey. Quality and Safety in Health Care 13, 102–107. Bowie P., Pope L. & Lough M. (2008) A review of the evidence base for significant event analysis. Journal of Evaluation in Clinical Practice 14 (4), 520–536. Britten N., Stevenson F.A., Barry C.A., Barber N. & Bradley C.P. (2000) Misunderstandings in prescribing decisions in general practice: qualitative study. British Medical Journal 320, 484–488. Bryman A. (2008) Social Research Methods, 3rd edn. Oxford University Press, Oxford. Carpenter D.R. (1995) Phenomenology as method. In: H.J. Streubert & D.R. Carpenter (Eds) Qualitative Research in Nursing, pp. 72–96. Lippincott, Philadelphia, PA. Coombes L., Allen D., Humphrey D. & Neale J. (2009) Indepth interviews. In: J. Neale (Ed.) Research Methods for Health and Social Care, pp. 197–210. Palgrave MacMillan, Basingstoke. Cresswell J.W. (2009) Research Design – Qualitative, Quantitative and Mixed Methods Approaches, 3rd edn. Sage, London. De Wet C., Bradley N. & Bowie P. (2011) Significant event analysis: a comparative study of knowledge, process and attitudes in primary care. Journal of Evaluation in Clinical Practice 17, 1207–1215. Denscombe M. (2007) The Good Research Guide: For SmallScale Social Research Projects, 3rd edn. Open University Press, Maidenhead. DH (2006) Safety First – A Report for Patients, Clinicians and Healthcare Managers. Department of Health, London. Dovey S.M. & Wallis K.A. (2011) Incident reporting in primary care: epidemiology or culture change? BMJ Quality and Safety 20, 1001–1003. Dovey S.M., Meyers D.S., Phillips R.L. Jr, Green L.A., Fryer G.E., Galliher J.M. & Grob P. (2002) A preliminary taxonomy of medical errors in family practice. Quality and Safety in Health Care 11 (3), 233–238. Elder N., Pallerla H. & Regan S. (2006) What do family physicians consider an error? A comparison of definitions and physician perception. BMC Family Practice 7, 73. Evans S.M., Berry J.G., Smith B.J., Esterman A., Selim P., O’Shaughnessy J. & DeWit M. (2006) Attitudes and barriers to incident reporting: a collaborative hospital study. Quality and Safety in Health Care 15, 39–43. Guest G., Bunce A. & Johnson L. (2006) How many interviews are enough? An experiment with data saturation and variability. Field Methods 18 (1), 59–82. Hansen E.C. (2008) Successful Qualitative Health Research, A Practical Introduction. Open University Press, Maidenhead. Harmsen M., Gaal S., van Dulmen S., de Feijter E., Giesen P., Jacobs A. & Wensing M. (2010) Patient safety in Dutch primary care: a study protocol. Implementation Science 50
© 2015 John Wiley & Sons Ltd
(5). Available at: http://www.implementationscience.com/content/5/1/50 (accessed on 13/2/2015). House of Commons Health Committee (2009) Patient Safety, Sixth Report of Session 2008/09, Volume 1. House of Commons Health Committee, London. Hsieh H. & Shannon S.E. (2005) Three approaches to qualitative content analysis. Qualitative Health Research 15 (9), 1277–1288. Joffe H. & Yardley L. (2004) Content and thematic analysis. In: D.F. Marks & L. Yardley (Eds) Research Methods for Clinical and Health Psychology, pp. 56–68. Sage, London. Kingston M.J., Evans S.M., Smith B.J. & Berry J.G. (2004) Attitudes of doctors and nurses towards incident reporting: a qualitative analysis. Medical Journal of Australia 181 (1), 36–39. Makeham M., Dovey S., County M. & Kidd M. (2002) An international taxonomy for reporting general practice error in Australia and five other countries. Medical Journal of Australia 177, 68–72. Mikkelsen T.H., Sokolowski I. & Olesen F. (2006) General practitioners’ attitudes toward reporting and learning from adverse event: results from a survey. Scandinavian Journal of Primary Health Care 24, 27–32. National Audit Office (2007) Improving Quality and Safety – Progress in Implementing Clinical Governance in Primary Care: Lessons for the New Primary Care Trusts. National Audit Office, London. NPSA (2008) Significant Event Audit – Guidance for Primary Care Teams. National Patient Safety Agency, London. NPSA (2009) Seven Steps to Patient Safety in General Practice. National Patient Safety Agency, London. NPSA (2010) Extracting Data and Sharing Learning from Significant Event Audits in General Practice. National Patient Safety Agency, London. Parahoo K. (2007) Nursing Research, Principles, Process and Issues, 2nd edn. Palgrave MacMillan, Basingstoke. Pfeiffer Y., Manser T. & Wehner T. (2010) Conceptualising barriers to incident reporting: a psychological framework. Quality and Safety in Healthcare 19 (6). DOI:10.1136/ qshc.2008.030445 Pringle M., Bradley C.P. & Carmichael C.M. (1995) Significant Event Analysis: A Study of the Feasibility and Potential of Case-Based Auditing in Primary Medical Care. Royal College of General Practitioners, London. Sandelowski M. (2010) What’s in a name? Qualitative description revisited. Research in Nursing & Health 33 (1), 77–84. Silverman D. (2010) Doing Qualitative Research, 3rd edn. Sage, London. Turner D.W. (2010) Qualitative interview design: a practical guide for novice investigators. The Qualitative Report 15 (3), 754–760. Welsh Assembly Government (2001) Improving Health in Wales – A Plan for the NHS with its Partners. Welsh Assembly Government, Cardiff. WHO (2005) WHO Draft Guidelines for Adverse Event Reporting and Learning Systems. World Health Organisation, Geneva.
419
