RESEARCH REPORT. Clarifying the factors associated with progression of cases in the GMC s Fitness to Practise process

RESEARCH REPORT Clarifying the factors associated with progression of cases in the GMC s Fitness to Practise process 1. Background 2. Objectives 3. Methods 3.1 Study design 3.2 Source of data and sample 3.3 Procedures 3.4 Variables included in the analysis 3.5 Method of analysis 3.6 Sample size 4. Results CONTENTS 4.1 Descriptive statistics 4.2 Associations with high impact outcomes 4.3 Discussion 4.4 Limitations of the study 4.5 Conclusions 5. Activities 6. Outputs 7. Impacts 8. Future research priorities 9. erences 10. Annexes Table 1: Characteristics of cohort of enquiries, according to place of qualification and ethnicity Table 2: Outcome of initial triage (Decision point A) Table 3: Outcome of investigation (Decision point B) Table 4: Outcome of adjudication (Decision point C) Table 5: Risk of high impact outcome at Decision Point A (promotion for further investigation by the GMC) Table 6: Risk of high impact outcome at Decision point B (referral for adjudication) Table 7: Risk of high impact outcome at Decision point C (doctor erased or suspended) 13

1. Background The General Medical Council (GMC) is the independent regulator for doctors in the UK. Set up in 1858, its statutory purpose is to protect, promote and maintain the health and safety of the public by ensuring proper standards in the practice of medicine. The GMC does this by controlling entrance to the medical register and setting the educational standards for medical schools. It also determines the principles and values that underpin good medical practice and has a responsibility to deal firmly and fairly with doctors whose fitness to practise is in doubt. The standards of competence, care and conduct expected of doctors are set out in the GMC s statement of principles Good Medical Practice. Box 1: Dimensions and standards of Good Medical Practice Dimensions Standards Doctors must Good clinical care provide good standards of clinical care, practice within the limits of their competence and ensure that patients are not put at unnecessary risk Maintaining good medical keep up to date with developments in their field, maintain their skills practice and audit their performance Relationships with patients develop and maintain successful relationships with their patients, by respecting patients autonomy and other rights Working with colleagues work effectively with their colleagues Teaching and training ( where they have teaching responsibilities) develop the skills, attitudes and practices of a competent teacher Probity be honest and trustworthy Health not allow their own health conditions to endanger patients Source: Adapted from The Meaning of Fitness to Practise, available at www.gmc-uk.org There are currently around 250,000 doctors on the GMC register, of whom approximately 150,000 are in active practice. The GMC receives approximately 5,000 separate enquiries each year about identifiable individual doctors. Four fifths of these 14

enquiries come from individual members of the public; the rest are mostly referrals from public organisations such as the NHS, the police or other regulatory bodies. All enquiries are assessed on receipt to see whether they raise potential issues about a doctor s fitness to practise. Those that do not are closed or referred back to the doctor s employer for local investigation. The minority of enquiries (approximately one third) that do appear to raise concerns about impaired fitness to practice - through deficient performance, misconduct, criminal behaviour, or physical or mental ill health - are investigated further by the GMC. The investigation may include obtaining further documentary evidence, witness statements and expert reports and assessing the doctor s performance or their health. Once investigations are completed, the case is reassessed. Most cases are concluded at this stage without further action. Of the remainder, some doctors receive a warning, some are asked to agree specific undertakings about their future practice, and some are referred to a Fitness to Practise panel hearing. In 2007, a total of 256 Fitness to Practise hearings took place. As a result of these, 60 doctors were permanently removed from the medical register, 79 were suspended, 68 received warnings or had restrictions imposed upon their practice; and 49 cases were closed with no further action. The GMC has a general duty under the Race Relations Act 1976, amended by the Race Relations (Amendment) Act 2000, to have due regard to the need to eliminate unlawful racial discrimination and to promote equality of opportunity and good relations between people of different racial groups. The Government s 2007 White Paper on the regulation of health professionals (Secretary of State, 2007) states that in the future the GMC will be required, as part of its report to Parliament, to provide information on equality issues relevant to the regulation of doctors. And the Council has made a public commitment to 15

ensure that its processes and procedures are fair, objective, transparent and free from unlawful discrimination. There has been concern for some while about possible inequalities and unfairness within the GMC s Fitness to Practise process in respect of ethnic minority doctors and doctors qualified outside the UK. Concerns about the volume of ethnic minority doctors within the General Medical Council s (GMC s) performance procedures, and the possibility of racism, first emerged in the early 1990s (Esmail and Everington, 1994). Since that time, a number of studies have been commissioned by the GMC to investigate these issues. An investigation by the Policy Studies Institute (PSI) (Allen, 2000) of the GMC s handling of complaints, found that the proportion of overseas qualified doctors (OQDs) complained about to the GMC matched their presence in the overall population of doctors in the UK. However, there were some differences between the OQDs and UK-qualified doctors dealt with by the GMC OQDs were older and more likely to be GPs, and the nature of the complaints made against them was different in some respects. This study also found that a greater proportion of complaints about OQDs came from public bodies than from other sources, and complaints by public bodies were more likely to progress at each stage of GMC procedures than those received from other sources. It was suggested that these differences might help explain the further finding that OQDs were over-represented, as compared to UK qualifiers, at later stages in the GMC s procedures. The PSI study found no evidence of racism in the GMC s processes, though it noted that the lack of transparency in the decision making processes then in place made this difficult to assess, and the possibility could not therefore be ruled out. 1 1 Since that time, detailed and explicit criteria have been adopted for decision making at each stage of GMC processes. A recent independent audit of a random sample of documented decisions at the initial triage and investigations stages of GMC procedures found that decisions were adequately justified and there was no evidence of inconsistency in the application of decision criteria (Hughes et al, 2007). 16

A further study by the York Health Economics Consortium (West et al, 2006) of GMC Fitness to Practise data on complaints received in 2005 also found that a disproportionately high number of the referrals made by public bodies concerned doctors qualified outside the UK (particularly those qualified outside Europe and, to a lesser extent, those qualified in Europe). The same pattern of over-representation was also evident at a later stage in the GMC s procedures, though to a lesser extent. Both the above studies were constrained by limitations in the quality of the data available and neither of them undertook multivariate analysis to explore the relative or combined effects of a range of different variables on the Fitness to Practise processes. Without more detailed information of this kind, the findings of these studies remain difficult to interpret. Importantly, also, interpretation of their results was hampered by the paucity of information held at the time about doctors ethnic status. In the PSI study, which was commissioned to explore the possibility of racial bias in GMC procedures, place of qualification was used as a proxy for ethnicity. Subsequently, a modelling exercise commissioned by the GMC to determine whether place of qualification could justifiably be used for this purpose found that it could not. Extrapolation from data on the minority of doctors about whom new complaints were received in 2005 for whom the GMC did hold ethnicity data, showed only a limited association between ethnicity and country of qualification (West, 2006). Nevertheless the tendency to elide coming from abroad with ethnic minority status in discussions in this area has continued. This is unhelpful because, while there is clearly overlap both within some individuals and in the issues raised, the experience and impact of migration and racism are not entirely the same and the causes and policy implications of problems associated with each are also different. 17

Growing appreciation of the challenges facing overseas-trained healthcare staff working in the UK (see, for example Allan et al 2004), continuing concerns about racism and discrimination in medicine (McManus et al 1998; Esmail et al 2003; BMA 2005) and increased attention at national level to equality matters in professional regulatory bodies (Secretary of State 2007) all point to the need for better understanding of ethnicity and migration related patterns of medical regulatory risk. In response, the ESRC Public Services Programme funded three linked studies to further explore these issues, including: a literature review (RES-153-25-0102) aimed at drawing together what is already known about the experiences and circumstances of migrant/bme doctors and considering how this might influence doctors performance or vulnerability to regulatory censure; an investigation of measures of workplace discrimination and attitudes towards diversity in healthcare organisations (RES-153-25-0103); and the study reported here - a new analysis of outcomes in the GMC s Fitness to Practise process. In 2007/08 the GMC undertook a major exercise to improve the quality and coverage of its ethnicity data, through a survey of all doctors on the Medical Register. The quality of the overall Fitness to Practise documentation and data management systems has also significantly improved following the introduction and consolidation of the new Siebel database in 2005. These two developments together provide the opportunity for a more detailed and thorough investigation than was previously possible. 2. Objectives The aim of this study was to improve understanding of factors associated with increased risk of high impact outcomes in the GMC s Fitness to Practise process. High impact outcomes include progression to the next stage of the process, and professional or legal sanctions. Specific objectives were: 18

i) To test the hypothesis that doctors qualified outside the UK and/or doctors from Black and Minority Ethnic (BME) backgrounds are more likely to receive high impact decisions at various stages in the Fitness to Practise process. ii) To evaluate the alternative hypothesis that high impact decisions are associated with other demographic or professional factors or characteristics of the complaints received that are independently associated with place of qualification or ethnic status. 3. Methods 3.1 Study design We undertook a secondary analysis of data held by the GMC about doctors going through the Fitness to Practise process. In our investigation we focused on outcomes at three different decision points in this process: Decision point A: Decision point B: Decision point C: Initial triage of enquiries received by the GMC Outcome of investigation stage Outcome of adjudication stage For the purposes of our analysis, outcomes at each stage were grouped by the seriousness of their implications for the doctor concerned as high impact, intermediate impact 2 or low impact. The main outcomes of interest were high impact decisions at each of the three decision points (Figure 1). 2 Intermediate impact outcomes may include warnings, conditions or undertakings. A warning may be deemed appropriate where the doctor s fitness to practise is not impaired, but there has been a significant departure from the principles set out in Good Medical Practise. Undertakings are an enforceable agreement between the GMC and the doctor about the doctor s future practise. They might include restrictions on the doctor s practise or behaviour, or commitments to have medical supervision or retraining. 19

Figure 1: Flow chart showing progression of enquiries through GMC Fitness to Practise process Decision points used in this analysis Enquiries received by GMC about identified doctors Triage stage Outcome of triage (Decision point A) Enquiry promoted for further investigation by GMC Enquiry referred for local investigation only No further action by GMC Investigation stage Outcome of investigation (Decision point B) Enquiry referred for adjudication Doctor given warning or undertakings No further action by GMC Outcome of adjudication (Decision point C) Adjudication stage Doctor erased or suspended Doctor given conditions or undertakings No further action by GMC High impact outcomes Intermediate impact outcomes Low impact outcomes 3.2 Source of data and sample The GMC holds basic demographic and professional information on its Siebel database about all doctors on the UK medical register. When the GMC receives an enquiry about a doctor, additional information is collected about the doctor s employment and the source and nature of the enquiry. Progress through the Fitness to Practise process and outcomes of enquiries are also recorded on the database. We were granted access to anonymised doctor-related and enquiry-related data for all enquiries received by the GMC about identified doctors between 1 April 2006 and 31 March 2008. 20

3.3 Procedures Ethical approval for the study was granted by the King s College London Research Ethics Committee on 31 July 2008. The study dataset was extracted by GMC internal staff from the database in April 2009 - thus allowing for at least one year to have passed since receipt of the most recent enquiries initiated in the sample period. Data anonymisation was done by GMC staff, who supplied the anonymised data to us on excel spread sheets. A doctor can be the subject of more than one distinct enquiry at the same or different times. Enquiries about the same doctor may be dealt with together or separately, and can have different outcomes. We decided that we could not use the individual doctor as our unit of analysis since, for those with several enquiries during the sample period, there would be no clear grounds for selecting which enquiry or outcome to include. Instead, our analysis is based on unique enquiries, each of which has a range of doctor-related and enquiry-related characteristics (see below). We defined as unique, all enquiries received on separate dates about the same or different doctors. In a few cases, more than one enquiry had been received on the same date about the same doctor. Where these had evidently been handled separately (i.e. allocated differently at triage), each one was treated as separate in the analysis. Where several enquiries received about the same doctor on the same date had been handled together (i.e. allocated to the same route at triage), we treated them as a single event. Any doctor who has been erased from the medical register can apply for their registration to be restored (after a period of at least five years has elapsed). Each year, the GMC receives a small number of such applications, which are investigated and then referred for consideration at a hearing. For this analysis, we excluded any enquiries of this type, 21

categorised by the GMC as restoration applications, since they are not part of the main Fitness to Practise process. 3.4 Variables included in the analysis Following a preliminary assessment of the relevance and quality of the data available, we identified five doctor-related variables (gender, ethnicity, place of qualification, time since primary medical qualification and a simple breakdown of practice specialty GP/not GP) and three enquiry-related variables (source of enquiry, type of enquiry, presence/absence of the main categories of allegations clinical care, probity, relationships with patients, working with colleagues, doctor s health) that were relevant and useable for our analysis. Other potentially relevant items in the database included doctor s age, a more detailed breakdown of speciality, qualifications, employment status and employment sector. However, the data on these items was found to be too incomplete to include in the analysis. The GMC s intensive recent efforts to improve the quality and coverage of the data it holds about doctors ethnicity have made a considerable difference, however the information held on ethnicity is still incomplete. At the date of extraction, ethnicity data were available for only 63% of enquiries within our dataset. On detailed inspection it became apparent that these data were skewed, with more missing values among doctors whose enquiries had progressed to further stages in the GMC process and had received higher impact outcomes. The GMC s survey of all doctors on the UK medical register requesting information about their ethnicity was carried out in the autumn/winter of 2007/08. This was towards the end of the two-year period from which our dataset was taken, and it means that the majority of doctors in our sample were asked to provide ethnicity information during or soon after their involvement in the Fitness to Practise 22

process. It is perhaps unsurprising to find that those who had progressed further through the process were less likely to comply. While acknowledging the limitations of a situation where ethnic status is known only for those doctors who had responded to the GMC s request for that information, it remained important, given the study aims, to include ethnicity in our analysis. Since one specific aim was to gain a better understanding of the possible interaction and/or separate impacts of ethnicity and coming from abroad, we created a new, composite variable combining known ethnicity (using a simple two-way split of White/BME 3 ) and country of qualification (a two-way split of UK/non-UK) to enable analysis of the possible influence of these two factors in different combinations. We used this composite variable for the main analysis whose findings are presented in this report. The policy context and arrangements for working in the UK differ in some respects for doctors coming from other parts of the EU/EEA as against those from outside the EEA. We therefore also undertook a separate analysis (not reported here), excluding ethnicity (because of the data limitations) but including a three-way split of country of qualification (UK/rest of EU and EEA/outside EEA). 3.5 Method of analysis Outcomes were modelled separately for the three decision points - triage, investigation and adjudication. At each decision point, a multinomial logistic model was fitted with the decision outcome as the dependent variable. This had the categories: high impact outcome; intermediate impact outcome; low impact outcome; no decision yet. Explanatory variables were gender, years since qualification, enquiry type, enquiry source, 3 We acknowledge the limitations of this simple binary division. However, the complexity of the Fitness to Practise process and the limited numbers of enquiries available mean that it was impracticable to use a more detailed breakdown by ethnicity in the analysis 23

specialty of doctor and (in selected models) allegation content. We used the robust option to allow for clustering of enquiry outcomes by doctor. 3.6 Sample size This study took place a relatively short time after the GMC s new database was introduced. The sample of enquiries for potential inclusion was necessarily limited to those that had been both initiated and completed in the intervening period. In practice, this meant focusing on enquiries initiated between April 2006 and March 2008. We provisionally anticipated a sample of 8,000 enquiries, including 2,000 concerning doctors qualified outside the UK. If 40% of those enquiries received high impact decisions at Decision Point A, there would be more than 90% power to detect a difference in proportion of such decisions between groups of about 4.5%. If 600 cases (including 200 concerning doctors qualified outside the UK) progressed through the Fitness to Practise process to Decision Point C, and 50% of those cases received high impact decisions at that stage, then there would be more than 90% power to detect a difference in proportion of unfavourable decisions between groups of 15%. 4. Results 4.1 Descriptive statistics There were 7526 unique enquiries received between 1 April 2006 and 31 March 2008, involving 6954 individual doctors. Of these, 35% were promoted for further investigation by the GMC, 7% were subsequently referred for adjudication; and 2% resulted in erasure or suspension of the doctor concerned. Four hundred enquiries had reached no final outcome by April 2009 when the dataset was extracted (199 of these were still awaiting a decision following investigation and 201 were awaiting the outcome of adjudication). 24

Table 1 shows the descriptive statistics for this cohort of enquiries. Some comparisons can be made with NHS workforce data 4 on the general population of doctors employed in the NHS in 2008 (though it should be noted that the two groups are not strictly comparable, since not all doctors referred to the GMC work for the NHS): 80% of the enquiries in the cohort involved male doctors, while 58% of doctors employed in the NHS were male; 46% of the enquiries involved doctors in general practice, while 28% of doctors employed in the NHS were GPs; 62% of the enquiries involved doctors who had been qualified for more than 20 years, while 39% of doctors employed in the NHS were aged 45 or older (and therefore likely to have qualified that long ago); and 37% of the enquiries involved doctors who qualified outside the UK, while 33% of doctors employed in the NHS had been trained in other countries. Within the cohort of enquiries, there were some differences between UK-qualified and non UK-qualified doctors on the doctor-related variables, and also some variation by ethnicity within these groups. For example: 77% of UK-qualified White doctors and 85% of non UK-qualified BME doctors were male; 62% of UK-qualified BME doctors, 45% of non UK-qualified White doctors, and 32% of UK-qualified White doctors had qualified in the past 20 years; and 30% of non UK-qualified White doctors and 42% of UK-qualified White doctors were GPs. 4 Published by NHS Information Centre at: www.ic.nhs.uk/statistics-and-datacollections/workforce/nhs-staff-numbers/ 25

There were also variations between these groups with regard to the enquiry related variables. For example: 77% of enquiries involving UK-qualified White doctors and 60% of those involving non UK-qualified BME doctors came from individual members of the general public (rather than from organisational sources); 87% of enquiries involving UK-qualified White doctors and 72% of those involving non UK-qualified BME doctors were categorised by the GMC as complaints (rather than referrals, determinations 5 or criminal convictions 6 ); and 13% of the enquiries involving UK-qualified White doctors, 18% of those involving UK-qualified BME doctors and 19% of those involving non UKqualified BME doctors included allegations concerned with probity. 4.2 Associations with high impact outcomes Tables 2, 3 and 4 show patterns of progression through the Fitness to Practise process and outcomes at Decision Points A, B and C. We estimated odds ratios for high impact outcomes at each decision point for each variable listed, using the low impact category closed no further action for reference. Allegation categories were excluded from the regression analysis at Decision Point A, since allegations are not recorded on the GMC s database for enquiries that get closed at triage. All the variables were included in the multilogistic regression analysis at Decision Points B and C. Tables 5, 6 and 7 show the adjusted odds ratios at Decision Points A, B and C. After adjustment, the following factors were associated with high impact outcomes: 5 This represents a determination by a body in the UK responsible under any enactment for the regulation of a health or social care profession to the effect that his/her fitness to practise as a member of that profession is impaired, or a determination by a regulatory body elsewhere to the same effect. 6 A criminal conviction is a conviction or caution in the British Isles for a criminal offence, or a conviction elsewhere for an offence which, if committed in England or Wales, would constitute a criminal offence. 26

At the triage stage (Decision point A): Enquiries received from organisational sources Enquiries categorised as referral or criminal conviction Enquiries about male doctors. Enquiries involving doctors qualified outside the UK (regardless of ethnicity) At the investigation stage (Decision point B): Enquiries received from organisational sources Enquiries categorised as referral, criminal conviction or determination Enquiries including allegations about clinical care, probity, relationships with patients, working with colleagues or doctors health Enquiries about male doctors. Enquiries involving doctors qualified outside the UK (regardless of ethnicity) At the adjudication stage (Decision point C): Enquiries received from organisational sources Enquiries categorised as referral or determination Enquiries including allegations about probity, relationships with patients, or doctors health After adjustment, enquiries involving UK-qualified doctors showed no association between ethnicity and outcome at any of the three decision points. Nor was there any association by specialty or time since qualification. We did not find convincing evidence of an association of non-uk qualification (or male gender) with adjudication outcomes at Decision Point C. However, in view of the more limited number of observations, our 27

study did not have sufficient power to detect small effects which might nevertheless be considered important. 4.3 Discussion Our findings showed, as might be expected, that the enquiry-related variables were the strongest predictors of outcomes. In relation to the main hypothesis about ethnicity and place of qualification, our findings have confirmed and reinforced the findings of earlier studies, but have also refined our understanding of the issues. In particular, the study has begun to help distinguish between the impacts of ethnicity and coming from abroad our findings suggest that the latter is an important risk factor for high impact outcomes 7, irrespective ethnicity, and that among those qualified within the UK ethnicity is not a source of additional risk. The major limitation on drawing firm conclusions in this respect is, as mentioned earlier, that the ethnicity data available for this analysis were still significantly incomplete. Accepting that place of qualification is a significant influence on outcomes in its own right (independent of other doctor-related factors, or the provenance, nature or content of the enquiries), why might that be? Are some foreign doctors receiving harsher treatment than is appropriate or, alternatively, are some UK-qualified doctors getting off more lightly than they deserve? Are some foreign doctors less able than those trained in the UK (perhaps because of fewer resources and connections, less confidence or external support) to defend themselves or challenge decisions? Are enquiries involving foreign doctors being assessed in some way as more serious than those involving a UK doctor with the same enquiry details? Similar questions might be asked about the influence of 7 This is confirmed by the separate regression analysis which we also undertook comparing UK-qualified doctors with those trained in other EU/EEA countries and countries outside the EEA, which showed associations with high impact outcomes for both non-uk qualified groups. 28

doctors gender on enquiry outcomes, since this study has also shown a greater risk of high impact outcomes for enquiries involving male doctors. The next step in pursuing these questions would be to look in much more detail at particular cases and decisions. A parallel project recently completed within the ESRC Public Services Programme offers useful insights about the scope for more detailed analysis of the information on the Siebel database. 8 Our study confirmed the findings of earlier studies that enquiries to the GMC involve a greater proportion of male doctors and doctors qualified longer ago than in the general (NHS-employed) medical population in the UK. We also found that GPs were overrepresented compared to doctors in other specialties. These patterns in the overall profile of regulatory referral are similar to those found in analyses of NHS referrals to the National Clinical Assessment Service about doctors with performance concerns (NCAS, 2006; NCAS, 2009). In our study there were proportionately very slightly more enquiries about doctors qualified outside the UK than in the general (NHS employed) population. There were also differences by place of qualification and ethnicity regarding the categorisation and content of enquiries and the routes by which they reached the GMC; with both non UK-qualification and BME status associated with higher risk enquiry characteristics. Again, these findings provoke questions as to why this might be. 9 In the literature review carried out in parallel to this study, we identified four domains where non-locally qualified and/or BME doctors may encounter additional challenges or disadvantages: medical education and the context of professional practice; the 8 Lloyd-Bostock: An an analysis of data on registration and fitness to practise case held by the GMC in the context of risk-based approaches to medical registration. (RES-153-25-0087) 9 Humphrey, Cohen and Esmail Challenges encountered by ethnic minority and migrant doctors, healthcare workers and related groups and the implications for performance regulation (RES-153-25- 0102) 29

circumstances and opportunities of their working lives; the attitudes and behaviour of other people; and their personal circumstances outside work. The literature suggested a range of ways in which such challenges might impact negatively on doctors careers and experiences, but we found no papers that specifically considered how coming from abroad or BME status might affect doctors performance (except in relation to academic performance) or their vulnerability to regulatory censure. However, a brief look at the wider literature outside medicine identified racism as a factor in BME groups experience of regulatory censure in several other fields, and it is possible to see how this might also be the case for doctors. In the other linked study 10 we have been exploring the potential for measuring organisational factors in NHS bodies (such as their competence in relation to equality and diversity issues) that might influence their prior handling of performance issues among doctors or patterns of referral or enquiry to the GMC. 4.4 Limitations of the study The new Siebel database and the GMC s ethnicity survey between them ensured sufficient improvements in data quality about the Fitness to Practice process to make this study worthwhile. However we knew from the outset that there would still be some problems with the data available. We envisaged the study therefore as a prototype analysis, designed to explore what could be done with the data and to help clarify questions and problems rather than offer any definitive answers. The main limitations we encountered were as follows: The number of enquiries included in the study was relatively small. This, and the fact that some enquiries had still not reached a final outcome at the time the 10 Esmail, Humphrey and Cohen Measuring organisational attitudes to workplace discrimination, prejudice and diversity: an exploratory study of NHS organisations (RES-153-25-0103) 30

dataset was extracted, limited the power of the study to detect associations, especially at Decision Point C. The available data on ethnicity were both incomplete and skewed, which means that the conclusions about the significance of ethnicity must be treated with some caution. At the time of writing, in November 2009, we understand that coverage has improved, with ethnicity data now held by the GMC for 74% of doctors on the register, and this figure will continue to rise over time as new doctors join the register Some of the other potentially relevant variables, such as employment status, were also incomplete, and therefore could not be included in the analysis. The data came from the GMC s internal management database. The extraction of these data in useable form by the GMC s internal staff and the subsequent process of data sorting and cleaning by the research team, were both considerably more laborious and time consuming activities than had been anticipated either by us or the GMC. The result was a very considerable delay (nine months) before we achieved a final, useable dataset for analysis. Consequently, within the time frame allotted to the project we were unable to progress as far as we had hoped to in developing a model that unified all three decision outcomes in a single analysis. We have continued the analysis since the funded period came to an end, and will pursue this further in the next few months, exploring and comparing different approaches to modelling the GMC processes using a single model to summarise outcomes observed at each of the decision points. For any future study of this sort using data from the Siebel database, both the research team and the GMC itself will need to plan in more time and resources to cover the data preparation work. We plan to consider with GMC staff how the database could 31

be developed to improve availability of a more robust and user-friendly minimum dataset for research. 4.5 Conclusions The findings of this study support the hypothesis that doctors qualified outside the UK are more likely to receive high impact decisions at various stages in the Fitness to Practise process. This association is partially explained, but cannot be fully accounted for, by interactions with other doctor-related and enquiry-related characteristics that are themselves associated with high impact outcomes. Better understanding of the risk associated with coming from abroad would require more detailed investigation of the GMC decision making process, including the categorisation and content of enquiries and allegations, and of doctors experiences within that process. After adjusting for other doctor-related and enquiry-related characteristics, we did not find an association at any stage between ethnicity and outcome among UK-qualified doctors. Among doctors qualified outside the UK, risks were increased among both White and BME groups (though slightly more so among the latter). The findings in respect of ethnicity must be treated with caution because of data limitations. With this caveat, our findings do not support the hypothesis that ethnicity is a risk factor for high impact outcomes in and of itself. Our study adds some insights into the separate and combined significance of ethnicity and coming from abroad, but further understanding would require a larger study with more fine-grained analysis of both categories. Such a study could be undertaken in 32

future, once more cases have had time to accrue and the ethnicity data are more complete. 5. Activities Updates on the research presented to a wide range of non-academic stakeholders at two meetings (10 th September 2008 and 4 th June 2009) of the General Medical Council s Equality and Diversity Research Forum. Research discussed at Public Services Programme Practical Aspects of Medical Regulation Subgroup meeting Bristol 28 th November 2008 Plans presented at a conference on poorly performing doctors Managing performance - a hard nut to crack held at the Royal Society of Medicine, 9 th July 2008. 6. Outputs A poster on the project will be displayed at the ESRC Public Services Programme End of Programme event Public Services in the 2010s: Prosperity, austerity and recovery in London, 11 th December 2009. Findings from the study will be presented at a conference on overseas doctors organised by the Association for the Study of Medical Education in February 2010. 7. Impacts 33

Findings will be presented to the GMC Council alongside further presentations to the GMC s Equality and Diversity Research Forum, which includes representatives from organisations such as the Equality and Human Rights Commission and the BMA. 8. Future Research Priorities Further research is needed as follows: When more data have had time to accrue on the Siebel database, a larger study with greater power to show associations and the capacity for more fine-grained statistical analysis. A more detailed, qualitative study investigating the factors that influence how decisions are made about the categorisation and handling of enquiries. A study exploring how the detailed content of allegations may vary between different groups of doctors. Studies of the factors influencing decision making processes in organisations outside the GMC, which result in the submission of enquiries about doctors 34

9. erences Allan H, Larsen JA, Bryan K, Smith P. The social reproduction of institutional racism: Internationally recruited nurses experiences of the British Health Service Diversity in Health and Social Care 2004;1(2):117-26. Allen I. The handling of complaints by the GMC: a study of decision-making outcomes London, Policy Studies Institute, 2000. British Medical Association. Tackling racism in medical careers: the role of consultants London, BMA Central Consultants and Specialists Committee, 2005. Esmail A, Everington S. (1994) Complaints may reflect racism. British Medical Journal 208:1374. Esmail A, Abel P, Everington S. Discrimination in the discretionary points award scheme: comparison of white with non-white consultants and men with women. British Medical Journal 2003;326:687-8. Hughes J, Locke R, Humphrey C. (2007) An independent audit of decisions in the investigation stage of the GMC s Fitness to Practise process. London: King s College London. McManus IC, Esmail A, Demetrious M. Factors affecting the likelihood of applicants being offered a place in medical schools in the United Kingdom in 1996 and 1997: retrospective study. British Medical Journal 1998;317:1111-7. Secretary of State for Health. Trust, assurance and safety the regulation of health professionals in the 21 st century. Cm 7013. London, The Stationery Office, 2007. West P. Potential use of proxy measures for studies of the impact of GMC Fitness to Practise procedures on doctors from different ethnic groups York Health Economics Consortium, University of York, 2006. 35

West P, Bending M, Chaplin S. A descriptive analysis of Fitness to Practise data for 2005 complaints. York Health Economics Consortium, University of York, 2006. 36

10. Annexes Table 1: Characteristics of cohort of enquiries, according to place of qualification and ethnicity Table 2: Outcome of initial triage (Decision point A) Table 3: Outcome of investigation (Decision point B) Table 4: Outcome of adjudication (Decision point C) Table 5: Risk of high impact outcome at Decision Point A (promotion for further investigation by the GMC) Table 6: Risk of high impact outcome at Decision point B (referral for adjudication) Table 7: Risk of high impact outcome at Decision point C (doctor erased or suspended) 37

Table 1: Characteristics of cohort of enquiries, according to place of qualification and ethnicity UK-qualified Not UK -qualified All All UK UK White UK BME UK Ethnicity unknown 1646 (100) frequency (column %) 4702 (100) 2648 (100) 408 (100) 2814 (100) 488 (100) 1230 (100) Male 3666 (78) 2035 (77) 319 (78) 1312 (80) 2311 (82) 384 (79) 1050 (85) 877 (80) 5985 (80) Years since qualification: </=10 586 (12) 251 (9) 108 (26) 227 (14) 388 (14) 59 (12) 172 (14) 157 (14) 974 (13) 11 to 20 1137 (24) 599 (23) 147 (36) 391 (24) 738 (26) 162 (33) 282 (23) 294 (27) 1875 (25) 21 to 30 1686 (36) 1019 (39) 115 (28) 552 (33) 704 (25) 164 (33) 316 (26) 224 (21) 2392 (32) 30 to 40 1051 (22) 689 (26) 35 (9) 327 (20) 756 (27) 84 (17) 377 (30) 295 (27) 1808 (24) >40 241 (5) 90 (3) 3 (1) 148 (9) 225 (8) 18 (4) 82 (7) 125 (11) 468 (6) Specialty: GP 2270 (48) 1125 (42) 166 (41) 979 (60) 1191 (42) 145 (30) 493 (40) 553 (51) 3461 (46) Source of enquiry: Public (individual) 3406 (72) 2031 (77) 274 (67) 1101 (67) 1570 (56) 312 (64) 735 (60) 523 (47) 4980 (66) Public (organisation) 451 (10) 195 (7) 57 (14) 199 (12) 484 (17) 77 (16) 191 (16) 216 (20) 937 (12) Person acting in a public capacity 343 (7) 157 (6) 25 (6) 161 (10) 460 (16) 54 (11) 165 (13) 241 (22) 807 (11) Other doctor 412 (9) 223 (8) 46 (11) 143 (9) 251 (9) 35 (7) 120 (10) 96 (9) 663 (9) Other source 83 (2) 39 (2) 6 (2) 38 (2) 48 (2) 10 (2) 18 (1) 20 (2) 131 (2) Type of enquiry: Complaint 3935 (84) 2310 (87) 323 (79) 1302 (79) 1893 (67) 357 (73) 882 (72) 654 (60) 5832 (78) erral 357 (8) 150 (6) 47 (12) 160 (10) 553 (20) 75 (15) 217 (18) 261 (24) 910 (12) Criminal conviction 276 (6) 118 (5) 32 (8) 126 (7) 202 (7) 27 (6) 91 (7) 84 (8) 483 (6) Determination 19 (0) 5 (0) 1 (0) 13 (1) 30 (1) 1 (0) 4 (0) 25 (2) 49 (1) Other 115 (2) 65 (2) 5 (1) 45 (3) 136 (5) 28 (6) 36 (3) 72 (6) 252 (3) Allegation content*: Clinical care 1981 (42) 1144 (43) 168 (41) 669 (41) 1184 (42) 220 (45) 552 (45) 412 (38) 3165 (42) Probity 676 (14) 333 (13) 74 (18) 269 (16) 516 (18) 66 (14) 238 (19) 212 (19) 1192 (16) Relationships with patients 689 (15) 389 (15) 61 (15) 239 (15) 403 (14) 71 (15) 175 (14) 157 (14) 1092 (15) Working with colleagues 170 (4) 85 (3) 23 (6) 62 (4) 139 (5) 22 (5) 64 (5) 53 (5) 309 (4) * An enquiry may involve several different allegations All not UK Not UK White Not UK BME Not UK Ethnicity unknown 1096 (100) 7526 (100) 38

Table 2: Outcome of initial triage (Decision point A) Low impact outcome Intermediate impact outcome High impact outcome All n [row%] Closed - no further action 2195 [29] Promoted - for local investigation only 2668 [35] Promoted for investigation by the GMC 2663 [35] 7526 [100 ] Gender Male 1681 [28] 2072 [35] 2232 [37] 5985 [100] Female 514 [33] 596 [39] 431 [28] 1541 [100] Place of primary medical qualification and ethnicity combined (10 missing) UK Known White 874 [33] 1086 [41] 688 [26] 2648 [100] UK- Known BME 111 [27] 150 [37] 147 [36] 408 [100] UK - Not known 499 [30] 584 [35] 563 [34] 1646 [100] Non-UK - Known White 135 [28] 161 [33] 192 [39] 488 [100] non-uk - Known BME 280 [23] 403 [33] 547 [44] 1230 [100] Non-UK - Not known 286 [26] 284 [26] 526 [48] 1096 [100] Years since primary medical qualification (9 missing) </= 10 263 [27] 234 [24] 477 [49] 974 [100] 11 to 20 532 [28] 697 [37] 646 [34] 1875 [100] 21 to 30 705 [29] 954 [40] 733 [31] 2392 [100] 31 to 40 535 [30] 664 [37] 609 [34] 1808 [100] > 40 151 [32] 119 [25] 198 [42] 468 [100] Medical specialty (6 missing) General practice 923 [27] 1501 [43] 1037 [30] 3461 [100] Other specialty 1266 [31] 1167 [29] 1626 [40] 4059 [100] Source of enquiry (8 missing) Public (individual) 1558 [31] 2428 [49] 994 [20] 4980 [100] Public (organisation) 128 [14] 59 [6] 750 [80] 937 [100] Person acting in public capacity 269 [33] 29 [4] 509 [63] 807 [100] Other doctor 204 [31] 115 [17] 344 [52] 663 [100] Other source 36 [27] 31 [24] 64 [49] 131 [100] Type of enquiry Complaint 1765 [30] 2657 [46] 1410 [24] 5832 [100] erral 88 [10] 11 [1] 811 [89] 910 [100] Criminal Conviction 79 [16] - 404 [84] 483 [100] Determination 13 [27] - 36 [73] 49 [100] Other 250 [99] - 2 [1] 252 [100] 39

Table 3: Outcome of investigation (Decision point B) Low impact outcome Intermediate impact outcome High impact outcome No decision yet All n [row%] Closed - no further action 6414 [85] Given warning or undertakings 401 [5] erred for adjudication 512 [7] 199 [3] 7526 [100] Gender Male 5020 [84] 336 [6] 450 [8] 179 [3] 5985 [100] Female 1394 [90] 65 [4] 62 [4] 20 [1] 1541 [100] Place of primary medical qualification and ethnicity combined (10 missing) UK Known White 2414 [91] 98 [4] 97 [4] 39 [1] 2648 [100] UK- Known BME 345 [85] 31 [8] 22 [5] 10 [2] 408 [100] UK - Not known 1415 [86] 90 [5] 109 [7] 32 [2] 1646 [100] Non-UK - Known White 410 [84] 26 [5] 33 [7] 19 [4] 488 [100] Non-UK - Known BME 972 [79] 96 [8] 126 [10] 36 [3] 1230 [100] Non-UK - Not known 848 [77] 60 [5] 125 [11] 63 [6] 1096 [100] Years since primary medical qualification (9 missing) </= 10 714 [73] 128 [13] 109 [11] 23 [2] 974 [100] 11 to 20 1606 [86] 98 [5] 127 [7] 44 [2] 1875 [100] 21 to 30 2102 [88] 96 [4] 128 [5] 66 [3] 2392 [100] 31 to 40 1581 [87] 66 [4] 111 [6] 50 [3] 1808 [100] > 40 402 [86] 13 [3] 37 [8] 16 [3] 468 [100] Medical specialty (6 missing) General practice 3051 [88] 130 [4] 194 [6] 86 [2] 3461 [100] Other specialty 3357 [83] 271 [7] 318 [8] 113 [3] 4059 [100] Source of enquiry (8 missing) Public (individual) 4776 [96] 49 [1] 114 [2] 41 [1] 4980 [100] Public (organisation) 517 [55] 163 [17] 192 [20] 65 [7] 937 [100] Person acting in public capacity 524 [65] 82 [10] 138 [17] 63 [8] 807 [100] Other doctor 486 [73] 95 [14] 56 [8] 26 [4] 663 [100] Other 103 [79] 12 [9] 12 [9] 4 [3] 131 [100] Type of enquiry Complaint 5498 [94] 91 [2] 176 [3] 67 [1] 5832 [100] erral 454 [50] 147 [16] 218 [24] 91 [10] 910 [100] Criminal Conviction 186 [39] 162 [34] 99 [21] 36 [7] 483 [100] Determination 25 [51] 1 [2] 18 [37] 5 [10] 49 [100] Other 251 [100] - 1 [-] - 252 [100] Content of allegations (enquiry includes allegations in these categories) Clinical care - yes 2997 [95] 21 [1] 41 [1] 106 [3] 3165 [100] Probity - yes 945 [79] 29 [2] 119 [10] 99 [8] 1192 [100] Relations with patients -yes 1012 [93] 6 [1] 28 [3] 46 [4] 1092 [100] Working with colleagues - yes 258 [84] 3 [1] 15 [5] 33 [11] 309 [100] Doctor s health - yes 205 [73] 18 [6] 26 [9] 33 [12] 282 [100] 40

Table 4: Outcome of adjudication (Decision point C) n [row%] Low impact outcome Closed - no further action 7110 [94] Intermediate impact outcome Given warning or undertakings 51 [1] High impact outcome Doctor erased or suspended 164 [2] No decision yet 201 [3] All 7526 [100] Gender Male 5615 [94] 46 [1] 146 [2] 178 [3] 5985 [100] Female 1495 [97] 5 [-] 18 [1] 23 [1] 1541 [100] Place of primary medical qualification and ethnicity combined (10 missing) UK Known White 2577 [97] 8 [-] 22 [1] 41 [2] 2648 [100] UK- Known BME 388 [95] 2 [-] 8 [2] 10 [2] 408 [100] UK - Not known 1557 [95] 19 [1] 39 [2] 31 [2] 1646 [100] Non-UK - Known White 461 [94] 6 [1] 10 [2] 11 [2] 488 [100] Non-UK - Known BME 1125 [91] 12 [1] 30 [2] 63 [5] 1230 [100] Non-UK - Not known 992 [91] 4 [-] 55 [5] 45 [4] 1096 [100] Years since primary medical qualification (9 missing) </= 10 893 [92] 15 [2] 37 [4] 29 [3] 974 [100] 11 to 20 1766 [94] 13 [1] 46 [2] 50 [3] 1875 [100] 21 to 30 2288 [96] 12 [1] 41 [2] 51 [2] 2392 [100] 31 to 40 1718 [95] 9 [1] 31 [2] 50 [3] 1808 [100] > 40 436 [93] 2 [-] 9 [2] 21 [4] 468 [100] Medical specialty (6 missing) General practice 3303 [95] 18 [1] 58 [2] 82 [3] 3461 [100] Other specialty 3801 [94] 33 [1] 106 [3] 119 [3] 4059 [100] Source of enquiry (8 missing) Public (individual) 4894 [98] 10 [-] 21 [-] 55 [1] 4980 [100] Public (organisation) 774 [83] 20 [2] 77 [8] 66 [7] 937 [100] Person acting in public capacity 693 [86] 9 [1] 46 [6] 59 [7] 807 [100] Other doctor 620 [94] 9 [1] 18 [3] 16 [2] 663 [100] Other 121 [92] 3 [2] 2 [2] 5 [4] 131 [100] Type of enquiry Complaint 5698 [98] 17 [-] 36 [1] 81 [1] 5832 [100] erral 722 [79] 12 [1] 78 [9] 98 [11] 910 [100] Criminal Conviction 403 [83] 19 [3] 41 [8] 20 [4] 483 [100] Determination 36 [73] 2 [4] 9 [18] 2 [4] 49 [100] Other 251 [100] 1 [-] - - 252 [100] Content of allegations (enquiry includes allegations in these categories) Clinical care - yes 2997 [95] 21 [1] 41 [1] 106 [3] 3165 [100] Probity - yes 945 [79] 29 [2] 119 [10] 99 [8] 1192 [100] Relations with patients -yes 1012 [93] 6 [1] 28 [3] 46 [4] 1092 [100] Working with colleagues - yes 258 [84] 3 [1] 15 [5] 33 [11] 309 [100] Doctor s health - yes 205 [73] 18 [6] 26 [9] 33 [12] 282 [100] 41

Table 5: Risk of high impact outcome at Decision Point A (promotion for further investigation by the GMC) Adjusted relative risk ratio 95% confidence interval Significance Gender Male 1.41 1.18 to 1.68 p<0.001 Female Place of primary medical qualification and ethnicity combined UK Known white UK - Known BME 1.29 0.95 to 1.75 ns Non-UK - Known White 1.64 1.23 to 2.20 p<0.001 Non-UK - Known BME 1.88 1.53 to 2.31 p<0.001 Years since primary medical qualification </= 10 0.91 0.72 to 1.15 ns 11 to 20 1.06 0.88 to 1.27 ns 21 to 30 31 to 40 1.03 0.86 to 1.23 ns > 40 0.97 0.73 to 1.29 ns Medical specialty General practice Other specialty 0.88 0.77 to 1.01 ns Source of enquiry Public (individual) Public (organisation) 4..06 3.13 to 5.26 p<0.001 Person acting in public capacity 1.43 1.09 to 1.87 p<0.01 Other doctor 1.96 1.58 to 2.43 p<0.001 Other source 2.42 1.53 to 3.83 p<0.001 Type of enquiry Complaint erral 6.22 4.68 to 8.28 p<0.001 Criminal Conviction 3.36 2.48 to 4.56 p<0.001 Determination 1.55 0.77 to 3.12 ns 42

Table 6: Risk of high impact outcome at Decision point B (referral for adjudication) Adjusted odds ratio 95% confidence interval Significance Gender Male 1.49 1.09 to 2.05 p<0.013 Female Place of primary medical qualification and ethnicity combined UK Known white UK - Known BME 1.26 0.74 to 2.15 ns Non-UK - Known White 1.78 1.10 to 2.87 p<0.019 Non-UK - Known BME 2.35 1.67 to 3.30 p<0.001 Years since primary medical qualification </= 10 1.06 0.74 to 1.50 ns 11 to 20 0.93 0.68 to 1.28 ns 21 to 30 31 to 40 1.04 0.75 to 1.43 ns > 40 0.98 0.58 to 1.65 ns Medical specialty General practice Other specialty 0.98 0.76 to 1.26 ns Source of enquiry Public (individual) Public (organisation) 3.14 2.05 to 4.79 p<0.001 Person acting in public capacity 2.85 1.85 to 4.38 p<0.001 Other doctor 2.39 1.58 to 3.61 p<0.001 Other 2.61 1.27 to 5.34 P<0.009 Type of enquiry Complaint erral 3.62 2.50 to 5.26 p<0.001 Criminal Conviction 3.50 2.25 to 5.36 p<0.001 Determination 7.18 3.35 to 15.38 p<0.001 Content of allegations (enquiry includes allegations in these categories) Clinical care allegation 2.70 2.08 to 3.50 p<0.001 No clinical care allegation Probity allegation 9.41 7.32 to 12.10 p<0.001 No probity allegation Relationships with patients allegation 2.60 1.90 to 3.60 p<0.001 No relationships with patients allegation Working with colleagues allegation 1.70 1.07 to 2.60 p<0.025 No working with colleagues allegation Doctor s health allegation 9.30 6.02 to 14.34 p<0.001 No doctor s health allegation 43