Article Text
PDF
Abstract
Physician burnout has reached epidemic levels in many countries, contributing to adverse personal, patient and service outcomes. Adverse socioeconomic conditions, such as the economic downturn in the Ireland post 2008, contribute to a situation of increased demand but inadequate resources. Given a recent unprecedented increase in referrals to Irish child and adolescent mental health services (CAMHS), coupled with a fragmented and poorly resourced service, it is important to reflect on consultant child psychiatrists’ well-being.
Objectives To report on the level of burnout among consultants working in CAMHS in Ireland using a cross-sectional design.
Setting Community CAMHS in Ireland.
Participants An online questionnaire was sent to all consultant child psychiatrists registered with the Irish Medical Council (n=112). Fifty-two consultants replied (46% response rate).
Primary outcome measures Questions assessed demographic and occupational details, career satisfaction and perceived management, government and public support. The Copenhagen Burnout Inventory measured personal, work and patient-related burnout.
Results The prevalence of moderate or higher levels of work-related and personal burnout was 75% and 72.3%, respectively. Fewer (n=14, 26.9%) experienced patient-related burnout. There was a strong correlation between work burnout and personal (r=0.851, n=52, p<0.001) and patient-related burnout (r=0.476, n=52, p<0.001). Lack of confidence in government commitment to investment in CAMHS (p<0.001) and perceived ineffective management by health authorities (p=0.002) were associated with higher burnout scores. Few consultants (n=11, 21%) felt valued in their job. The majority (n=36, 69%) had seriously considered changing jobs, and this was positively associated with higher burnout (p<0.001). Higher burnout scores were present in those (n=15, 28.8%) who would not retrain in child psychiatry (p=0.002).
Conclusion The high level of burnout reported by respondents in this study, and ambivalence about child psychiatry as a career choice has huge professional and service implications. Urgent organisational intervention to support consultant psychiatrists’ well-being is required.
- burnout
- stress
- CAMHS
- mental health
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Strengths and limitations of this study
Use of a validated instrument to measure personal, work and patient-related burnout.
Relevance of the study to the current staffing and resources crisis in child and adolescent mental health services in Ireland.
Low response rate (46%).
Cross-sectional design does not allow causality to be inferred.
Risk of responder recall bias due to self-report nature of the survey.
Background
Occupational stress is among the most prevalent work-related illnesses reported by workers in the European Union (EU), with significant personal and societal costs.1 In the Ireland, the most recent data available show that in 2015, 18% of the Irish workforce experienced work-related stress.2 Although less than the European average, rates of occupational stress more than doubled between 2001 and 2015, with workers in Ireland experiencing the steepest increase compared with workers in other Western European countries.2 3
Burnout (BO) is a term used to describe long-term, unresolved, work-related stress, with feelings of physical and emotional exhaustion, depersonalisation and a reduced sense of personal accomplishment as central features.4 In this context, BO is thought to occur when demand exceeds capability or availability. In conditions where there is increased demand but reduced or inadequate resources, this imbalance will be most evident, presenting a ripe breeding ground for occupation stress. Such a scenario is familiar to many working in health services: a recent study of the Irish workforce shows those working in the health sector reported the highest rates of BO.2 High BO rates among doctors is a very worrying statistic, given the implications for patient safety and service delivery.5 High BO rates have also been described in other clinical6 7 and non-clinical8 groups.
Currently in the Irish healthcare context, as with many other EU countries, it is recognised that demand far exceeds provision, leading to long waiting lists in many specialities and many doctors working in excess of European Working Time Directive requirement.9 In a recent survey of hospital doctors in Ireland (n=1749, 55% response rate), 82% reported significant workplace stress, with one in three doctors reporting BO.9 This seems to be particularly relevant for mental health services: the proportion of the Irish health budget devoted to mental health has decreased since 2008 and currently stands at 6%, lower than in many other countries.10 An even smaller percentage goes to child and adolescent mental health services (CAMHS), which remains neglected, despite a surge in referral rates (up 26% from 2012) and recognition of CAMHS staffing being well below recommendations.11
In addition to reduced funding, consultants working in psychiatry may be a particularly vulnerable group, as they are exposed to and expected to manage stress from numerous sources.12 They must be able to manage their own stress levels, support their teams and treat stress and mental illness in their patients. Although not unique to psychiatry, compassion fatigue, negative transference and vicarious traumatisation have been implicated in increased risk of BO.12 Given the unprecedented increase in referrals to CAMHS, coupled with a fragmented and poorly resourced service, it is important to reflect on consultant child psychiatrists’ well-being. This imbalance between shrinking resources and increasing demand in a specialty where clinician compassion and leadership is key led us to examine BO levels in consultants leading CAMHS teams. To date, no study has investigated BO in this specific population.
Objectives
This study aimed to report on the prevalence of BO among consultant psychiatrists working in CAMHS in Ireland using a cross-sectional design.
Method
Following ethics exemption, a survey was designed using the SurveyMonkey platform. A survey invitation and link was sent by the College of Psychiatry, Ireland (CPI) to all consultant child and adolescent psychiatrists who were registered with the Irish Medical Council, and members of the CPI (n=112). Participants were invited to complete the questionnaire, which took approximately 10–15 min. For ease of completion, responses to questions used a Likert Response Scale, respondents rating their level of agreement from always/often/sometimes/seldom/never. Two reminder follow-up emails were sent at 3 monthly intervals (May–August 2017). All responses were anonymised. The strengthening the reporting of observational studies in epidemiology cross-sectional checklist was used when writing the report.13
Patient and public involvement
The focus of this study was on consultant stress levels. Although there was no direct involvement of patients or the public, the CPI assisted in the design and delivery of the study. They also sent the questionnaire directly to participants, collected and anonymised the responses, ensuring data privacy.
Study questionnaire
The study-specific questionnaire included basic questions on demographic details such as years in practice, work setting and team staffing levels. Questions assessed consultants’ confidence in the public healthcare system, managed by the Health Service Executive (HSE) and funded by the government. Participants were also asked about public perception of CAMHS and satisfaction with their chosen career. To minimise identification, minimal personal data, such as age or gender, were collected.
A well-validated BO questionnaire, the Copenhagen Burnout Inventory (CBI),14 was used. This scale allows an examination of the level of personal stress or BO, along with the degree of BO that is perceived to be related to work and client/patient population. Scores fall along three domains: work-related BO (seven questions), personal BO (six questions) and patient-related BO (six questions). The CBI has been found to predict future sick leave, relevant to work settings, and correlates well with other frequently used scales.15 All 19 questions are scored using a 5-point Likert Scale; 12 items have frequency responses (always/often/sometimes/seldom/never) and seven items intensity responses (ranging from ‘a very low degree’ to ‘a very high degree’). All responses are positively skewed. Scores of 50–74 are considered moderate levels of BO, 75–99 high and 100 severe BO.15 The CBI has good psychometric properties.14 16 In the current study, there was high internal reliability as evidenced by high Cronbach alphas coefficients (work-related BO α=0.942, personal BO α=0.874 and patient-related BO α=0.833).
Analysis
For the purpose of analysis, Likert scales were considered as interval data. Kolmogorov-Smirnov tests confirmed normality for total and subscale CBI scores and hence parametric analysis was carried out. Descriptive data were calculated for each of the subscales, along with the proportion of respondents scoring moderate or higher levels of BO. Relationships between the subscales were evaluated using Pearson correlation coefficients. Mean CBI scores were compared with other binary variables using t-tests. Participants were divided into groups based on responses to (1) having considered changing jobs in the last year (‘never’ or ‘rarely’/‘sometimes’/‘yes, definitely’), and (2) whether they would retrain in child psychiatry (no/maybe/yes). One-way analysis of variance (ANOVA) was used to compare mean CBI scores across these dimensions. There were minimal missing sociodemographic and study-specific questionnaire data; the number of respondents for each variable is represented in each table by ‘n’. All respondents completed the CBI in full.
Results
Sample profile
Fifty-two consultant child and adolescent psychiatrists replied (46% response rate). Forty-nine participants provided their specialty, with the majority (n=42, 81%) working in CAMHS outpatient community settings (table 1). A smaller number (n=7, 13%) worked exclusively in an inpatient setting (five in a specialist CAMHS inpatient unit and two in a paediatric hospital). Three did not indicate their work environment. Respondents (n=51, 98%) were, in general, seasoned clinicians with on average 14.5 years as a consultant (range 1–30 years).
Professional characteristics
Working conditions
Forty-four (84.6%) consultants responded to a question evaluating their service based on the Irish national mental health strategy document, a Vision for Change, which sets out the direction and staffing planning for mental health services in Ireland.17 Twenty-seven (61%) were staffed at 50% or less of the recommended levels. Although two respondents (one each from an outpatient and inpatient setting) reported full staffing levels, nine (20.5%) had levels of staffing 30% below recommended levels. The majority of consultants (n=37, 90%) had sought additional funding directly from the HSE and a smaller number (n=10, 19%) had advocated publically for such funding (table 1). Consultants were also asked what percentage of their time they spent on work considered outside of their area of responsibility or non-core work. The median value was 30%, with a range from 0% (n=5, 15.5%) to more than half of their time (n=5, 15.5%).
Work satisfaction
The majority of respondents had seriously considered changing jobs (n=36, 69%). Almost a third (n=15, 29%) would not choose child psychiatry again as a specialty if retraining, with a further 18 (36%) undecided (table 2). Few consultants felt valued in their job, almost half stating this was ‘rarely or never’ their experience (n=21, 40%). A similar number believed staff on their teams rarely or ever experienced job satisfaction (n=19, 37%). The majority believed that their workload (n=46, 88%) or degree of job satisfaction (n=29, 57%) interfered (often or always) with their ability to engage in new initiatives, training or academic work.
Study-specific questions
Perceived management, government and public support
Consultants were asked about the HSE and the government’s commitment to planning and investing in child psychiatry (table 2). Although the majority of respondents had in fact sought (although not necessarily received) additional funding in the previous 2 years (n=37, 90%), the majority (n=38, 73%) believed commitment to investment was ‘never or rarely’ present. They also expressed low confidence in the ability of the recent Seanad Committee (senate of Ireland) hearing on mental health18 to have a meaningful impact on their lives as clinicians (n=12, 25%) or those of their patients (n=15, 31%). Only one respondent felt the government was serious about a commitment to children’s mental health or effectively planning services. There was a general sense of cynicism regarding the ability of political involvement to make meaningful change (n=19, 39%). Consultants feared minimal change in public opinion towards CAMHS in recent years (n=31, 60% stated no or minimal change) and one-third perceived the public to have very unrealistic expectations about what CAMHS could or should deliver (n=17, 33%). Participants also felt that, when compared with other medical specialities, CAMHS was viewed less favourably by the public and was less well resourced (n=40, 77%).
Burnout
Fifty-two respondents completed all questions on the CBI. The overall prevalence of respondents reporting moderate degrees of BO or higher was 59.6% (table 3). Rates varied between the CBI subdomains, with very high levels being reported in both work and personal domains. Thirty-nine (75%) experienced moderate or higher levels of work-related BO, and 35 (72.3%) consultants experienced moderate or higher levels of personal BO. Three consultants (5.8%) rated their level of personal BO as severe. The overall prevalence of patient-related BO was much lower, with 14 (26.9%) experiencing moderate or higher levels of work-related BO. The majority (n=38, 73%) reported no or minimal patient-related BO.
Scores, cut-offs and reliability of the Copenhagen Burnout Inventory (CBI)
There was a strong positive correlation between all CBI subscales, strongest for work and personal BO, r=0.851, n=52, p<0.001. High levels of work-related BO were also associated with high patient-related BO, r=0.476, n=52, p<0.001. Although there was a significant association with personal and patient-related BO, this was not as strong, r=0.322, n=52, p=0.02. Using the coefficient of determination, work-related BO helped explain 72.4% of the variance on the personal subscale and 22.7% of the variance on the patient BO scale, while the shared variance between personal and patient BO was only 10.3% (table 4).
Pearson product-moment correlations between burnout domains
An independent t-test was used to compare the total BO scores for clinicians working in CAMHS outpatient settings as compared with an inpatient setting. There was no significant difference in mean BO scores between outpatient and inpatient consultants. Although mean BO scores were higher in those with fewer resources, this was not significant (table 5). However, significantly higher BO scores were found in those who lacked confidence in the HSE’s ability to effectively manage services and the degree of government commitment to CAMHS (table 5).
Independent samples test
A one way between group ANOVA identified a significant relationship between BO and consultants’ consideration of changing jobs, F (2, 49)=12.198, p≤0.001, η2=0.332. Post hoc comparisons using Tukey’s HSD test indicated that the mean BO score for those who had not/rarely wished to change jobs was significantly lower (M=30.263, SD=15.85) than those who responded ‘sometimes’ (M=47.862, SD=14.60) or ‘yes, definitely’ (M=57.20, SD=16.972). There was no significant difference between the latter two groups. Similarly, an ANOVA revealed significant BO differences between those who would and would not consider retraining in child psychiatry, F (3, 47)=7.359, p=0.002, η2==0.264. Post hoc comparisons indicated that the mean BO score for those who would not retrain in child psychiatry (M=59.04, SD=17.13) or were undecided (M=58.49, SD=9.14) was significantly higher than the no group (M=40.64, SD=16.86).
Discussion
Fifty-two consultant child and adolescents psychiatrists responded to this survey with the majority (59.6%) reporting moderate or higher levels of BO, as measured by the CBI. Both work-related and personal BO levels were high, with 75% and 72.3% scoring in the moderate or higher levels, respectively. The corresponding means (work related=59.41, personal=55.93) are much higher than the referenced norms.19 In a large (n=1914) Danish prospective study, which recruited participants from seven different types of work (including hospital employees), mean work-related BO (M=33.0) and personal BO (M=35.9) were much lower than values found in this study.19 More recent studies using the CBI allow a direct comparison with other clinical groups. In a sample of almost 1500 senior doctors and dentists in New Zealand, their reported rates of BO were lower, with corresponding prevalence rates of 42.1% and 50.1%.20 In a sample of 1037 Australian midwives, equivalent prevalence rates were also much lower (43.8% for work-related BO and 64.9% for personal) than that found among child psychiatry consultants practicing in Ireland.21 Two other studies examining doctors working in Irish hospitals reported rates of BO considered by the authors to be higher than their European colleagues, with junior doctors being most at risk.9 22
There is some evidence to suggest that BO is further elevated in the specialty of psychiatry and among trainees. A recent systematic review concluded that psychiatrists, particularly woman and those in training, suffered from higher levels of BO and psychological distress than other medical groups.23 Hayes et al found that younger age, higher effort–reward imbalance and overcommitment were associated with higher BO rates among Irish hospital doctors.9 The ‘BOSS’ study, which was a large international study on BO (n=1980) and included both psychiatrists and trainees, found 39% of their sample reported severe BO, high levels of depression and suicidal thinking. In fact, among the Irish trainees recruited in the BOSS study, although their rate of depression was lower than their peers (10% vs 16%, using the Patient Health Questionnaire-9), stressful working conditions and lack of supervision contributed to their higher BO scores.24 Although age and gender were not collected in our study, to help protect respondent identity, there was no association with years working as a consultant and stress levels, suggesting that in our study seniority and experience did not confer advantage.
There is some evidence in the literature of a higher rate of BO in clinicians working with children. Margiotta et al found that paediatricians in Ireland scored higher than other groups,22 and in a large Finnish study (n=3313), while psychiatrists had a higher prevalence of BO than other specialities, the highest rates were found in those working in child psychiatry and in the community.25 Even when biological markers are used to measure ‘stress’, levels among nurses in a CAMHS inpatient setting were found to be higher than in a neonatal intensive unit.26
High rates of personal and work-related BO have been associated with more negative feelings towards patients12 and higher rates of cynicism.27 Although respondents in our study reported very high rates of BO in both work and personal domains, rates of patient-related BO was much lower. The majority (n=38, 73%) reported no or minimum BO, suggesting an absence of compassion fatigue, and the majority continued to advocate for their patient group. However, given the therapeutic salience of the relationship psychiatrists have with their patients, and their need to be emotionally available and empathic, high personal or work-related stress may impede this. Maslach and Leiter emphasise that in addition to risks associated with lack of support from colleagues and management, contact with patients and families who present with psychological trauma, have unrealistic service expectations, and/or may be demanding or challenging, present unique challenges for the psychiatrist.12
There were ample evidence in our study of a sense of lack of support from both management (HSE) and government, cynicism about the possibility of change, coupled with unrealistic public expectations of CAMHS and a sense of a lack of parity of child psychiatry. These were all factors outside of the control of the consultant. In a qualitative study of Irish consultants, Walsh et al found that the top three concerns contributing to BO, again extrinsic factors, were ‘quick fix’ management planning, staff shortages and inadequate resources.28
The issue of under-resourcing was also prominent in this study and most respondents indicated their services were understaffed based on A Vision for Change recommendations.17 Although 15% indicated that their staffing levels were at or above 75% recommended levels, most (n=22, 50%) had services operating at 50% staffing levels, with one service having only 20% of recommended resources. However, unlike in other studies, respondents operating with lower than recommended resourcing levels did not report higher BO levels.28 Certain perceptions rather than resourcing were associated with higher BO levels. These included not having confidence in either the HSE’s ability to manage the services effectively (n=37, 71%) or the government’s commitment to invest (n=38, 73%), also reflected in other research.28 These results resonate with the words of author and doctor, Rachel Clarke’s in her book ‘Your Life in My Hands’, with reference to a ‘a desperately overstretched workplace that breeds corner-cutting, indifference and excessive haste’.29 Tireless and dedicated service delivery, despite poor work conditions, as seen in this study, is a known response by care providers, as they prioritise the needs of others over their own, with the resultant risk of professional inefficacy, cynicism and personal BO.
What is perhaps most alarming is that 65% of consultant in this study, if given the opportunity, had reservations about training in child psychiatry. Seventy per cent had seriously considered changing jobs in the last 6–12 months and this was significantly related with high levels of BO. Other studies have also reported that higher BO rates are linked with career choice dissatisfaction.24 This high rate of career ambivalence among consultant child psychiatrists is higher than rates in other specialities.9 This may contribute to understanding the current and unprecedented difficulty with staff recruitment and retention in CAMHS in Ireland, with recent media and government reference to large numbers of unfilled consultant child psychiatry posts, and many more filled with occupants not on the specialist register, raising serious concerns in relation to patient safety.30
The consequences of systematic failure to address medical under-resourcing have been recognised in the UK (Francis Report)31and in Ireland (MacCraith Report),32 with parallel attempts to address these (MacCraith Working Group). However, progress has been slow in Ireland, where the negative impact of the 2008 financial crisis and the associated 30% pay reduction continue to have effect. Many of the recommendations made by the MacCraith report, especially around retention, have not been successful.33 34 Poorer working conditions, unequal remuneration and type of contract were found to adversely affect BO levels among consultants in Ireland, where a two tier consultant contract exists.22 Working conditions of hospital doctors in Ireland were also considered less favourable than in other EU countries.9
Current consultant psychiatry numbers (per 100 000) in Ireland are less than half the EU average, and a vicious cycle is perpetuated between under-resourcing, poor work conditions, high work stress, long waiting lists and unacceptable patient risk. Retention of existing staff and recruitment of new staff are a top priority, yet difficult to attain, with both Irish trained and non-EU trained doctors considering leaving.35 In our study, as with other studies, new entrants to consultant posts are already showing signs of high BO, with many newly trained doctors now choosing to emigrate. The Medical Council of Ireland Intelligence Report and the Royal College of Surgeons Ireland (RCSI) Health Workforce examined reasons for emigration among Irish medical graduates in 2015. They identified ‘push’ factors to include ‘stressful conditions’ compounded by low staffing and ‘lack of designated and supervised training’,36 37 findings which have been echoed in other studies .12 28 If clinical standards are to remain high, and children and families are to receive expert quality care, then urgent attention needs to be given to both work conditions and physicians’ mental health.
If clinical standards are to remain high, and children and families are to receive expert quality care, then urgent attention needs to be given to both work conditions and physicians’ mental health.
Retention and attracting newly qualified doctors into psychiatry are essential to ensure that we have the necessary manpower required. Such concerns are not unique to Ireland, and have been raised in the UK and in the USA.3 38 39 Psychiatry was found to be a less-favoured specialty by UK-trained graduates than other career choices40; attitudes towards a career in psychiatry were more polarised than views about other specialties and graduates who initially chose psychiatry as a career did not stay.41 In addition to resources and working conditions, negative perceptions regarding the scientific nature of the specialty, effectiveness of interventions and public attitudes all increase the risk of trainee drift, some of which were apparent in our study.
Collectively these factors contribute to a worldwide concern that physician BO has reached epidemic levels, and that in particular psychiatrists risk becoming an ‘endangered species’.39 42 43 Given the reluctance of the consultants in our study to consider retraining, and the impact of BO on their ability to engage in training of junior staff, this is a real concern for child psychiatry in Ireland.
Recognising and reducing work and personal BO are now as much a public concern as lack of investment in services. The negative effects of physician ill-health are well recognised both for patient, service and themselves.39 43 44 Doctors’ well-being is integral to professionalism and central to patient care, and has been proposed as a central missing quality indicator.45 46 A systems level response is necessary, given contributors to BO include personal, service and culture issues within healthcare organisations.47 All should be tackled in order to reduce physician BO, as reported by a recent systematic review and meta-analysis.43
Long waiting lists are a particular problem facing CAMHS in Ireland, and require innovative solutions. One example associated with positive outcomes, both in physician BO and patient well-being, was the recruitment of medical assistants to help prepare, manage and triage referrals worthy of exploration within Irish CAMHS.48 Individually targeted approaches, such as personal engagement in mindfulness, small group discussions and stress reduction programme, have also been identified as being effective.43 Attendance at Schwartz rounds has been associated with a statistically significant improvement in well-being and resulted in positive changes in clinician practice,49 50 and is currently being rolled out in Ireland.51 Balint groups, regular facilitated meetings with physicians, have long been known to improve physician well-being and protect against BO.52 They allow for a safe and unpressured reflection on everyday clinical encounters and the relationship between doctor and patient, and the emotions it engenders. The CPI has now mandated participation of all trainees in Balint groups and support to attend such interventions, even on line, might well be considered for consultants. However, when asked, doctors opt for primary-level interventions aimed at bringing about system and cultural change, rather than focusing at the individual level.28 Avoiding ascribing blame to the clinician for being responsible for their own ill-health is paramount. A combination of individually targeted and system-wide interventions might be the most effective and sustainable approach.
Strengths and limitations
This is the first study examining levels of BO among consultant child and adolescent psychiatrists in Ireland, set against a background of unprecedented consultant vacancies and increasing CAMHS referrals. Although the response rate (46%) was lower than hoped, it is comparable to that found in studies among consultants in Ireland (42%–55%),9 22 studies within psychiatry (26%)27 and studies using CBI (22.5%–40%).20 21 It is unclear whether non-response would preferentially apply to busy overworked clinicians with high BO levels, too exhausted to participate, or if respondents are more likely to reflect those who had indeed experienced BO and wanted to have their voice heard. However, given there were no significant outliers, and the data were normally distributed, it is hoped that the results offer an informative insight into current BO levels in the target group.
Given the cross-sectional design and self-report nature of the survey, causality cannot be inferred and there is a risk of responder recall bias. The study used a validated instrument to measure personal, work and patient-related BO, but minimal data were gathered on personal characteristics such as gender, age, marital status, thus limiting analysis on factors already known to be associated with higher BO rates. This was a necessary step, given the small sample size and the possibility of responder anonymity being lost.
Conclusion
Prevalence of personal and work-related BO among child and adolescent psychiatrists in Ireland is unacceptably high. Consultants are working with chronically understaffed teams, feel undervalued and have low levels of job satisfaction. They expressed significant ambivalence regarding remaining in post, retraining in the specialty, and a limited ability to contribute to academic work and training. Coupled with this is an external locus of control to effect change, with low confidence in either the HSE or the government, and a sense of unrealistic public expectations. These are worrying statistics and finding a solution is imperative. With the number of vacant consultant posts rising in both psychiatry and other medical professionals in Ireland and elsewhere, we need to heed the rising levels of work related and personal BO and implement change in order to ensure the well-being of our patients, professionals and indeed specialty, so that we do not become a ‘speicis atá i mbaol’ (Irish (Gaeilge) for endangered species).
Acknowledgments
The authors express gratitude to all the consultants who completed the study and to the College of Psychiatry, Ireland for assistance with the study.
References
Footnotes
Contributors All contributors have meet at least of one the criteria recommended by the ICMJE. FM conceived and designed the study and wrote the first draft of manuscript. SS assisted with data analysis. SS, CO and EB contributed to subsequent drafts and revisions of paper.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval This study received ethical exemption by chairman action OLCHC ethics committee. May 2017.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request. The anonymous data set is held with the PI, FM. Any request to access the data should be made to fiona.mcnicholas@ucd.ie and will be considered.