Patient safety skills in primary care: a national survey of GP educators

General Practice specialty training (GPST) across the four postgraduate medical deaneries in Scotland is 3 to 4 years in length. Of this, 18 months are spent in a GP setting with the remainder of the time in a variety of hospital posts. Each trainee has a named GPES throughout the length of their training programme. The curriculum for GPST is provided by the RCGP [17] and must be covered during training and evidenced within the trainees’ e-portfolio. Patient safety and quality of care are part of the curriculum but there are no defined methods of teaching or assessing this important area, thereby introducing significant potential for variation in educational provision, with possible implications for patient care.

A previously published questionnaire of safety skills formed the basis of this study [13]. This original questionnaire was developed via online survey and a subsequent face-to-face focus group of ten experienced patient safety researchers (backgrounds in medicine, surgery and psychology) to elicit their views of key safety skills. In order to avoid altering the underlying constructs in any way, the 73 safety skills proposed by participants were used verbatim to form items for the questionnaire, which sought responses regarding perceived importance and trainability of each item on a 5-point Likert scale.

For this study, the questionnaire was initially reviewed and content validated during a 90-minute medical education workshop involving all 24 trained GPES from a single specialty training group in the west of Scotland. This phase was important to ensure that the subsequent questionnaire used for stage two contained the appropriate safety skills that would be relevant to GPs - particularly as the previous iteration was intended for use by hospital doctors. All participants were provided with the original questionnaire of 73 items and asked to independently rate the relevance of each item to being a safe GP. A four-point Likert scale [22] was used to indicate level of relevance (1 = not relevant; 2 = unable to assess relevance without item revision or item in need of such revision that it would no longer be relevant; 3 = relevant but needs minor alteration; 4 = very relevant and succinct). Free-text comments against each item were also encouraged from participants.

Data were analysed using SPSS v19.0. For each item, a Content Validity Index (CVI) was calculated to assess the level of agreement between the 24 GPES, based on the proportion with a rating of 3 or higher (on the 4-point scale). The CVI is a method to quantitatively summarise expert views, expressed on a numerical scale, on a topic – in our case on the elements that make up a safe GP. A cut-off of 0.80 is typically set in research contexts, such that items that score >0.80 across a pool of experts are considered to have reached agreement (i.e. on their importance, or relevance) and hence they are maintained, whereas those that fail to do so are discarded, or amended [22]. Using this criterion, items with a CVI < 0.80 were removed from the questionnaire as these were perceived by the experienced GPES to be of less relevance to being a safe GP [23]. Two independent researchers also conducted a qualitative analysis of free-text comments. Items that had a high CVI > 0.80 and received strong support through the qualitative analysis were reviewed by a multi-disciplinary panel consisting of the original questionnaire developers, patient safety experts, GPES and senior educationalists to form the final content-validated questionnaire to be used in stage two.

All 691 GPES across Scotland were identified from the NHS Education for Scotland organisational databases and invited to participate in the study. The modified questionnaire from stage one was administered electronically using ‘QuestBack’ software (http://​www.​questback.​co.​uk). For each listed skill/attribute, participants were asked to indicate (i) its perceived importance/relevance to primary care and (ii) its perceived trainability. Both questions were answered on 5-point Likert scales. Participants were also invited to suggest additional skills/attributes not listed within the questionnaire. Questionnaire responses were anonymous and participation was voluntary, with subjects free to exit the survey at any point. Two-weekly reminders were sent between November 2011 and February 2012 to non-respondents. Ethical approval was not required for this study.

Descriptive statistics were calculated using SPSS v 19.0 (proportion agreeing (score = 4 or 5), disagreeing (score = 1 or 2) and neither agreeing/disagreeing (score = 3) for each item). For the purposes of data analysis the importance and trainability data for each questionnaire item (i.e., skill) were grouped into 17 broad categories (of the original questionnaire developed with hospital doctors). As per the secondary objective of this study, this enabled comparisons with the original hospital study. This latter data-set comprised questionnaire responses from 50 physicians and surgeons. Additional items suggested by respondents were analysed for emerging themes by two independent clinical reviewers.

The content validation process is summarised in Figure 1. Initial content validation of the original 73-item safety skills questionnaire resulted in 6 items being removed (CVI < 0.80). Qualitative analysis of respondents’ comments by two researchers resulted in 48 items being included with no modifications; 17 items being modified via multidisciplinary consensus, and 2 items being removed. One item that had been initially removed (‘Having good leadership skills’) was reinstated as the multidisciplinary panel felt it would be important to explore the views of the broader GPES community given national interest in medical leadership and management [24]. The final content validated questionnaire for safety skills in primary care therefore comprised 66 skills/attributes grouped into 17 broad categories (Additional file 1).

A total of 348 GPES completed the survey (response rate = 50%) between November 2011 and February 2012, with all respondents indicating levels of agreement with all statements to provide a full data set.

There was generally good agreement between what GPESs in this study, and hospital doctors of the recent study [13] perceive as core qualities/attributes of a safe practitioner. As in this study, technical skills and honesty were perceived to be two of the most important skills/attributes of a safe doctor in the hospital study (for both items, 98% respondents agreed or strongly agreed). ‘Awareness of patient’ (includes empathy) was ranked relatively low in importance in both studies (78% respondents agreed/strongly agreed in this study, 82% in the hospital study). In contrast, whilst leadership was perceived to be important by only 36% of respondents in this study, 93% of doctors in the hospital study believed it to be important to being a safe doctor.

Regarding skill trainability, technical/clinical skills were considered the most trainable in both studies. Skills relating to team-working, preparedness, vigilance and situational awareness also featured highly in perceived trainability across both studies, which are all well-recognised non-technical skills [12]. In contrast, ‘softer’ skills and attributes such as confidence, humility and common sense were ranked relatively lower in both studies.

GPESs in this study recognised a broad range of skills/attributes associated with being a safe GP. Importantly, many of these were judged to be trainable. Not surprisingly, technical/clinical skills were considered important and the most trainable. Least trainable were attributes that would constitute professional values; yet these are crucial to the effective performance of any clinical practitioner, the learning culture in the practice and therefore patient safety [17],[25]. Interestingly, leadership was considered the least important skill/attribute to being a safe GP, which contrasts sharply with the views of hospital doctors [13]. This may reflect a difference in the way that GPs understand and interpret the concept of leadership in respect of safety, as compared to hospital doctors. For GPs, safety may reside much more in individual practice and care, rather than in the leadership of a team and they may conflate ‘leadership’ with ‘management’. In contrast, hospital doctors often have more clearly defined leadership roles as part of clinical teams at the frontline and also as part of clinical directorates at managerial level. This may make leadership in respect to safety a more recognised construct within secondary care.

The additional novel skills/attributes proposed by respondents further suggest that there are important, albeit subtle differences in the perceived necessary safety skill-set between primary and secondary care doctors. For example, the fundamental feature of general practice in coordinating the effective provision of holistic, long-term patient care may be reflected in the additional item suggested ‘taking ownership/accountability’. This is in contrast to care in the hospital sector where the majority of patient contacts and follow-up are on a relatively short-term basis, with patients ultimately discharged back to the care of their GP. In a similar vein, that GPs are often the patients’ first port of call for medical advice (and the onus that this places on GPs) may be reflected in the additional item suggested ‘intuition – listening to that worrying little inner voice’. Interestingly, ‘awareness of patient’ was ranked relatively low by GPESs in both importance and perceived trainability. It is possible that this item is not fully understood. GPES and trainees are focused on ‘patient-centred’ consulting and determining the patient’s agenda, which is what is required to pass the RCGP membership examination. The proposed alternative terms ‘patient-focused’ and ‘patient advocacy’ reflect this specific context. In a similar vein, ‘negotiation skills’ (as a subset of teamwork and communication) was suggested as an additional item, perhaps emphasising this critical GP skill in negotiating a management plan with patients (as their advocates), and with other healthcare professionals, as ‘gatekeepers’ to secondary care.

To the best of our knowledge, this is the first national survey to explore GPES’ views regarding the key skills and attributes necessary to being a safe GP. The views of GPs regarding what constitutes ‘patient safety’ have been previously explored, but the specific individual skills and attributes necessary to be a safe GP were largely overlooked when compared to team and systems level priorities [26]. Most of our study findings compare favourably with existing competency frameworks [17], but provide further insight into the potential prioritisation of competencies and what is trainable, as perceived by GPES from a patient safety perspective. For example, training in some safety-critical skills such as team-working and communication is widely available and researched. However, perhaps an added dimension is required to emphasise how these skill-sets interact and impact on the safety of the immediate workplace and the wider NHS. From this perspective there is now growing interest in educating the NHS workforce in human factors science [27]. A possible educational gap, therefore, is the need to move beyond a limited focus on the importance of ‘non-technical skills’ to patient safety to an appreciation of a more holistic understanding of the role and consequences of human-system interactions in the clinical workplace. Developing educational interventions for patient safety that focus on raising awareness of the workplace as a complex socio-technical system will be necessary [28].

The survey response rate of 50% was moderate, but comparable to other internet-based surveys in healthcare [29]. The large number of respondents was sufficient for analysis and inference. We did not capture demographic data in an attempt to reduce the questionnaire completion time and because we judged that, although important, this information was not particularly critical to the study purpose and interpretation of findings. We were therefore unable to compare the characteristics of responders and non-responders so the possibility of response bias cannot be excluded. Our findings may not therefore be generalisable from a GP educator perspective, or representative of the non-training GP community. Moreover, we acknowledge that this particular group of GPs may not be considered ‘experts’ and therefore the possibility of ‘unconscious incompetence’ exists, with GPES having differing interpretations of safety and a potential lack of knowledge/insight of safety systems and their role in individual errors when compared to patient safety ‘experts’. Related to this point is the technical limitation that using the CVI as a numerical cut-off for survey items may have resulted in items being discarded based on respondent consensus, whereas in fact safety experts would have scored them differently and potentially retained them. We acknowledge that there is inherent subjectivity in methods that rely on human judgement. Indeed, the additional items suggested by respondents may indicate a gap in the question set and/or interpretational issues on the part of the participants. Nonetheless we believe this study adds to our knowledge and understanding of this under-researched topic, with GPES well-placed to identify the safety attributes of a safe GP. All GPES must be deemed safe consulters prior to being accepted as educational supervisors, and patient safety competences both at individual (through medical appraisal) and system (practice) level are assessed at re-accreditation visits. Moreover, the GPES are all trained to assess and make judgements on trainees’ ability to practice (including safely) and can be called to account for these judgements.

These findings provide a basic educational platform on how to go about improving patient safety in primary care through targeted training of GPs in those safety skills/attributes which were perceived as a higher priority and judged to be trainable by experienced and informed frontline educators and clinicians. Importantly, this can help develop educational programmes in priority safety skills at different levels of training and career progression for GPs. Additionally, we may be able to agree a skill-set on which to base the development and alignment of assessment tools for testing the safety competencies of GPs, particularly those in training. Given increasing calls for primary care clinicians to engage in and lead in safety improvement efforts [30] our findings suggest a need to encourage GPs into more visible leadership positions and also emphasise how GPs are leaders in their own practices, although many may not explicitly recognise this. This is particularly prudent given the renewed moves towards primary care-led commissioning in England [31] and health and social care integration in Scotland [32]. Finally, this study also gives weight to the argument for extending GP specialty training to enable the acquisition of skills in patient safety, leadership and quality improvement in parallel to core clinical competencies - necessary for delivering safe, effective care for the 21^st century [17]. Future research should further explore GPs’ views on the concept of leadership as a safety skill in primary care. Repeating this study to explore the views of GP trainees and the wider primary care team could also provide invaluable insight into potential differences in views and may help to shape targeted training interventions for the whole primary care team in the future.