Introduction  

The scale and the complexity of modern healthcare data have resulted in successive policy recommendations for investment and innovation in healthcare information technology.1–4 Following the 2017 Wachter review, the UK is embracing ‘big data’, ‘digital medicine’ and ‘data-driven healthcare’ through the Global Digital Exemplar Programme in hospitals, and Health Data Research UK (HDR-UK), the new national health informatics (HI) institute. HI in this context may be defined as ‘the intelligent use of information and technology to provide better care for patients’.5

Both in the UK and internationally, there is growing appetite to quantify and improve the digital maturity of health systems.6–9 The universal relevance of HI is exemplified by the WHO’s recent adoption of a digital health resolution, led by India,10 and the urgency is amplified by the ambition of recent investments in precision medicine11 12 and artificial intelligence.13 Despite acknowledgment that ‘upskilling’ and ‘increased capacity’ are necessary for ‘digital readiness’ across all health professions,4 7 14–16 there has been far less research and policy focus on training of doctors and other health professionals in HI, the variably defined discipline that has evolved into the application of information science and computer science to healthcare.17 Even in the USA, where a separate subspecialty accreditation programme in HI has been developed, there has been limited success with digital transformation.18

In the UK, there are approximately 113 500 doctors,19 of whom up to 65 700 are estimated to be in postgraduate training and less than 4000 are clinical academics.20 The Wachter review recognised that previous large-scale HI programmes in the National Health Service (NHS) have suffered from ‘a shortage of individuals with relevant skills’.4 The UK trains 7500 medical students per annum.21 Despite repeated General Medical Council (GMC) guidance,22 deficiencies in UK undergraduate HI education persist.23 Therefore, postgraduate HI training will be crucial in bridging the knowledge and skills gap in HI among doctors. The NHS Digital Academy will train chief clinical information officers (CCIOs),24 25 and the Faculty of Clinical Informatics (FCI) seeks to ‘professionalise’ HI accreditation, but these recently formed organisations will emphasise leadership. The Digital Academy will train only 300 digital leaders in the next 3 years, which cannot alone effect change in postgraduate training of the clinical or clinical academic workforce.

Although leadership will be important for digital transformation projects, all doctors will need to be trained in HI. However, despite successive attempts in the UK3 26 27 and across Europe,28 existing postgraduate curricula for HI have not been evaluated systematically and have not led to agreed HI standards for all doctors. There is evidence of substantial heterogeneity in HI training at undergraduate and postgraduate training across countries in Europe.29 Without a coordinated effort to map the use of these standards across all postgraduate medical specialties in a particular country, it will be difficult to develop robust, universal HI competencies for all doctors.30

We therefore sought to: (1) develop up-to-date, evidence-based, international HI standards for postgraduate medical training and (2) to evaluate HI competencies currently required across all 71 medical specialties in the UK, compared with these standards.

Methods

We conducted a mixed methods study with three components (scoping review, curricula content analysis and expert interviews) to iteratively develop universal HI standards for postgraduate medical training.

Results

Scoping review

From 2734 references, 21 curricula documents were eligible for inclusion (figure 1), ranging from 1998 to 2017. There were 12 papers from the USA30 37–47 with the remainder from Canada (n=3),48–50 Germany (n=2),51 52 the Netherlands (n=2),53 54 Australia (n=1)55 and a collaboration between the Netherlands, USA and Germany.56 Documents described academic (n=10), clinical (n=6), a combination of academic and clinical (n=2) and ‘comprehensive’ curricula (n=3) (developed by professional HI bodies not affiliated to one academic establishment or workplace, eg, Canada’s Health Informatics Association). Curricula were aimed at all HI professionals (n=10) doctors (n=6) and all healthcare professionals (n=4). One study45 did not specify the target audience. The intended learning setting was university (n=11), workplace (n=9) or online (n=1). Most documents did not specify methodology for formulating competencies (table 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Flow diagram of included curricula in scoping review.

Curricula documents were mainly for higher education qualifications, requiring active uptake by the student and too detailed for the frontline clinician.38 43 52 The competencies described varied widely even between MSc degrees with the same target audience, for example, focused on technical knowledge and computer science51 versus more comprehensive coverage of topics including ethics and communication.39 One curriculum document identified 10 core HI competencies for healthcare professionals by asking delegates attending the American Medical Association (AMA) 1999 Spring Conference to discuss their views on HI education in small group workshops.30 However, it does not specifically target doctors, and although suggested competencies are relevant, the broad themes overlook more recent topics, such as e-prescribing.

Mapping of documents to ‘IMIA+’ domains identified three missing competencies, leading to the iterative addition of domain 46: international developments,53 domain 47: medical physics51 and domain 48: teaching50 (table 2). Domains 5 (information literacy), 9 (information systems in supporting patients and the public) and 12 (structure design and analysis of the health record including data quality) were present in all documents, highlighting their importance. Health information science, chemoinformatics and nanoinformatics (domains 41–43) were not present in any documents, which is compatible with the ‘optional’ status given to these domains for BMHI specialists within the IMIA recommendations.35 Any competencies relating to health information science were usually mapped in an HI context, for example, domain 10: regional networking and shared care, present in 81% of documents. Document-derived domains 46 and 47 (medical physics and international developments, respectively) were both present in 10% (n=2) of documents. Domain 48 (teaching) was present in 33% (n=7) of documents. The three iteratively added domains demonstrate that the IMIA recommendations represent the majority, but not all, of the HI curriculum. Full mapping is in online supplementary appendix 2.

View this table:

• View inline
• View popup

Table 2

Fifty domains of competency in health informatics (from International Medical Informatics Association) including skill level and presence in 71 specialties

Wide variation in definitions of ‘core’ and ‘introductory’ competencies across the reviewed documents underscored the need for expert consultation. The majority of competencies across the documents were attributed as core: 41.2% (n=247); the remainder were classified as core standard: 6.3% (n=38), standard: 42.5% (n=255), standard specialist: 4.7% (n=28) and specialist: 4.7% (n=28). Mixed knowledge combining core–standard–specialist skills accounted for 0.7% (n=4): an uncommon situation where document competencies with multiple skill levels mapped to a single framework domain. Forty-three domains were identified as core skills ranging from 71% of mapped documents advocating domain 48 (teaching) as a core competency to 13% for domain 15 (biomedical modelling and simulation).

Curricula content analysis

A total of 71 curricula were analysed (online supplementary appendix 1). Child & Adolescent Psychiatry and Core Surgical Training curricula did not contain any of the keywords searched. No curricula contained the keyword e[-]Health. Of the remaining 69 curricula, 96% (n=66) contained Technolog*, 46% (n=32) IT System, 29% (n=20), Electronic [Health/Clinical/Medical/Patient] Record and 14% (n=10) Informatic*. Mapping of the curricula led to the addition of domains 49 (employ new technologies appropriately, including information technology) and 50 (proactive approach to new technology).

Out of the 50 domains, 29 were present across 71 curricula documents. Domains 3 (efficient and responsible use of information processing tools to support healthcare professionals’ practice and their decision making), 4 (use of personal application software for documentation, personal communication including internet access, for publication and basic statistics), 11 (appropriate documentation and health data management principles), 26 (basic IT skills) and 49 (employ new technologies appropriately, including information technology) were mapped in over 50% of curricula (table 2; online supplementary appendix 3). Chemical pathology and paediatric and perinatal pathology mapped the highest percentage of domains (32%; n=16), followed by histopathology, general practice (GP) and anaesthetic curricula all mapping 30% of domains (n=15). Surgical specialties mapped particularly poorly with 9/10 specialties mapping only 2% (n=1) of the domains (figure 2). Thirty per cent (n=21) of all curricula (n=71) mapped between 0 and 2 of the total 50 domains. Full curricular mapping is in online supplementary appendix 3.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Proportion of 50 health informatics competency domains mapped in postgraduate medical curricula of 71 specialities.

Expert consultation

HI competencies were unanimously felt to be essential for front-line clinicians. Experts found the IMIA framework useful as a guide, but not comprehensive, making iterations necessary. Introductory domains did not alone fulfil needs of a clinician; most of the introductory and intermediate competencies were identified as necessary skills. Many domains were felt to comprise a mixed skillset, for example, introductory domain 12 (structure, design and analysis of the health record). Analysis of the health record is a universally required skill, but system design is an advanced skill more appropriate to CCIO level for system procurement. Similarly, some advanced IMIA domains, such as domain 18 (use of informatics tools to support education), were considered universally relevant. Regarding skill levels, ‘universal’ competencies for all clinicians were felt to be more important than the IMIA framework’s IT user/BMHI specialist distinction, validating findings of the scoping review, where curricular documents identified core competencies across all IMIA skill levels.

Experts felt that the iteratively added domains had mixed utility. Domain 48 is a vital skill throughout clinical training and therefore naturally extends to HI in both teaching other clinicians and empowering patients to engage with HI resources. Domain 46 (international developments) was acknowledged as important but not considered a universal requirement. Domain 47 (medical physics) was not considered to be in the remit of HI. Experts agreed that domains 49 (employ new technologies appropriately, including information technology) and 50 (proactive approach to new technology) were too vague to have impact on HI skill dissemination throughout the workforce and were symbolic of a ‘tick-box attitude’ in curricula development.

All experts agreed that HI was grossly under-represented in postgraduate medical training. While the framework and iteratively added competencies covered the majority of HI education (at all levels, not limited to working clinicians), some areas were felt to assume more detail than was explicitly stated, including secondary use of data, digital communication, system and staff safety and patient empowerment. These areas were all identified in the recommended universal HI competencies, in addition to information governance/security, information and knowledge management and emerging technologies (figure 3). Experts involved in curricula development during their career felt that HI should be integrated into pre-existing competencies, since a dedicated HI subsection would neglect relevance of HI across healthcare.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Universal health informatics competency domains developed iteratively by scoping review, curricular content analysis and expert consultation.

Discussion

To our knowledge, this is the first review of HI curricula available for postgraduate medical training and the first comprehensive mapping across all specialties at country level, with four key findings of global relevance. First, existing curricula for HI are mostly developed in North America (15/21; 71%). Second, curricula have been developed without standardised methodology, often omitting key HI competencies. Third, across 71 UK postgraduate medical specialty curricula, the 50 competencies identified as important for HI training were poorly represented and absent in two specialties (Child & Adolescent Psychiatry and Core Surgical Training). Fourth, given poor adoption of exhaustive lists of HI competencies (eg, IMIA), we proposed a pragmatic list of six universal competency domains for all postgraduate doctors.

The emphasis of heads of government and the WHO on digital health, precision medicine and artificial intelligence highlights the need for critical awareness and skills in HI for every doctor.10 11 57 However, our scoping review shows that existing postgraduate curricula for HI are developed in a very small number of high-income countries. It is likely that existing digital competencies are neither tailored nor relevant to all doctors in all contexts. Middle-income or low-income countries are unrepresented in the literature. The ‘digital divide’ has been extensively described58–60 but is understudied with respect to education and training of doctors in HI, which may especially exacerbate inequities within and across countries. The USA has led in the field of HI, in terms of research, training and practice. The vast majority of EHR providers and other HI-related companies are based in North America, predominantly in the USA. Moreover, in the UK, recent reviews of NHS healthcare information technology have been led by US experts, including the ongoing Topol review of the role of artificial intelligence, genomics and digital medicine61 and the Wachter review.4 Although IMIA competencies are comprehensive and the most up-to-date international standards for HI, local adaptation is necessary for training, research and practice in all countries, including the UK, which requires local training and expertise.

Current HI curricula are stand-alone, require independent uptake by the clinician or focus on specialist HI skills. The current model of motivated clinicians seeking out short-term training or qualifications at masters or doctoral level in HI, often at considerable expense, is neither sustainable nor scalable. For example, of 7660 doctors registered to undertake NHS e-Learning for Healthcare modules between 2012 and 2016, only 250 doctors completed all five modules and 151 doctors completed one to four modules.62 ‘Medical education’ is often interchangeably used to signify both postgraduate and undergraduate education, meaning that several curricular documents had an unclear target audience.37 44 45 51 53 54 56 Institutions may be approaching HI education in the same way for undergraduate students and postgraduate doctors,53 despite differing educational63 and knowledge needs64 of the two groups. In a rapidly moving discipline, standardisation is essential, but curricula have not been produced with reproducible methodology, making the evidence base for existing standards questionable at best. Our method, adapted from a prior study, uses literature review, curricular content analysis and expert review to iteratively develop competency domains and could be used routinely in areas other than HI.

HI competencies in current UK training curricula for clinicians are limited, fragmented and often neglected. It is perhaps unsurprising that specialties such as pathology, general practice (GP) and anaesthetics mapped to the highest percentage of domains. These specialties have embraced digitisation,65 66 especially GP, which has near universal EHR use in the UK.67 Conversely, it is of concern that while the government is aiming for a paperless NHS by 2020,68 Child & Adolescent Psychiatry and Core Surgical Training curricula do not contain any of the HI keywords. Fifty-nine per cent of curricula contained the stock phrase ‘Employ new technologies appropriately, including information technology’, but 14% of curricula contained no other HI-related competencies. Such ‘token’ competencies lack specificity and are difficult to implement for both trainees and trainers. Creating ‘future-proof’ training at postgraduate level requires regular review and updating of competencies for doctors, but the curriculum is already overcrowded. Therefore, our list of six universal HI competency domains offers a solution, to which areas of increasing importance can be added iteratively, for example, the role of research, artificial intelligence and multiomics (which are largely ignored, even in IMIA competencies). HI skills must also be balanced with the non-analytical, humanistic aspects of medicine to produce an ‘even more effective physician for the future’.69

Although evidence that HI training improves the likelihood of success of digital implementation programmes is currently lacking, there is strong evidence that current IT programmes and current efforts to train clinicians in HI are suboptimal.14 70 Therefore, universal postgraduate training in HI represents the best way to create change at scale, alongside training of leaders and individuals seeking more specialist training. Generic, stand-alone competencies are likely to be ineffective because training curricula, trainee needs and working environments vary greatly. Lack of competency integration could explain the suboptimal implementation of NHS-endorsed HI skills to date. Competencies in evidence-based medicine71 and medical ethics72 have been successfully integrated into training of all doctors. Given current deficiencies in HI curricula, both in terms of literature review and mapping of UK postgraduate specialties, we provide a simpler framework of universal HI competencies for all postgraduate doctors. Improving ‘HI literacy’ among doctors is an international agenda that requires three actions. First, specialty organisations (eg, Royal College of Physicians) must identify opportunities to integrate HI competencies across multiple curricula, for example, record keeping, audit and research. Second, the validation and adaptation of our six universal competency domains must be overseen by postgraduate leads in medical education in collaboration with professional HI standards bodies (eg, FCI). Third, the optimal implementation plan for universal HI competencies for all postgraduate doctors will require consultation with a wide range of stakeholders, including regional/national leads in HI (eg, CCIOs), the HI research community (eg, HDR-UK), government (eg, NHS Digital and Department of Health) and professional regulators (eg, GMC).

The major strengths of this study are its comprehensive consideration of HI training across all postgraduate specialties; the use of previously used methodology and frameworks to iteratively develop HI standards for curricula; and independent expert feedback. A possible limitation was the subjective nature of determining keywords and mapping competencies between the IMIA+framework and curricular documents. Additional keywords could have been included in the curricula keyword search, uncovering further HI competencies. Bias was reduced by addressing inter-rater reliability with a second researcher when identifying papers for inclusion, involving experts in academic and clinical HI to determine keywords and consulting a range of experts. Another limitation is that not all competencies listed in the IMIA+framework are relevant to each specialty; for example, domain 29 (technical informatics/computer science) is only mapped in the Aviation & Space Medicine Curricula. HI competencies for training doctors were studied, but those for other clinicians, such as associate specialists, staff grade clinicians and consultants, have no national curricula, making it impossible to assess if HI skills are being promoted as part of their continuing professional development (CPD).

Conclusion

This study has identified that HI education for postgraduate doctors is not fit-for-purpose, partly due to inconsistencies in HI terminologies and scope within existing HI curricula. It is unsurprising that without agreement on what to teach, postgraduate training curricula often represent HI with a ‘token competency’ approach. Our method, adapted from a prior study, uses literature review, curricular content analysis and expert review to iteratively develop competency domains and could be used routinely in areas other than HI. We propose 20 universal HI competencies for integration into pre-existing curricula for doctors (see figure 3). A focus on ‘digital maturity index’ of hospitals and clinical institutions alone, without a ‘digitally-ready workforce’, will restrict the enormous potential benefits of digital healthcare. ‘Technology is as much an integrated tool of the practitioner as is the stethoscope’30 and should be recognised as such within clinician curricula for the digital healthcare era.