A teledermatology roadmap: implementing safe and effective teledermatology triage pathways and processes

This roadmap supports systems to accelerate the roll out of teledermatology to help manage demand and reduce unnecessary outpatient attendances. The roadmap sets out practical steps to implementing an efficient, safe and effective teledermatology service. It identifies opportunities to use teledermatology in new and innovative ways to deliver more personalised and better integrated care and prepare for the introduction of artificial intelligence in skin lesion pathways.

Summary

Technology for both teledermatology and artificial intelligence (AI) is fast-moving. This updated roadmap considers recent developments.

This roadmap recommends how NHS systems can accelerate the roll out of teledermatology to support the delivery of dermatology services and prepare for the deployment of AI as appropriate. It is aimed at clinicians, managers and transformation leads working at a local, integrated care system (ICS) or regional level to support planning and delivery of transformation of dermatology services.

Teledermatology is the use of digital images to triage, diagnose, monitor or assess skin conditions without the patient being physically present. A high quality teledermatology service, as part of locally agreed pan-system redesign of dermatology services, can support transformation of services and elective recovery.

Digital images can be used to:

  • enable specialist advice (advice and guidance; A&G) as the ‘front door’ to dermatology services for adults and children with skin, hair or nail disorders, or non 2-week wait skin lesions
  • support the virtual urgent suspected skin cancer pathway (previously known as the 2-week wait skin cancer pathway); where clinically appropriate, a teledermatology interaction can replace a face-to-face appointment
  • help prioritise patients on long waiting lists.

AI is the use of digital technology to create systems capable of performing tasks previously requiring human intelligence. AI as a medical device (AIaMD), linked to teledermatology services, is demonstrating potential in the assessment of skin lesions at sites across England, and their experience, subject to independent evaluation, will inform future national support for the safe implementation of AI in dermatology at scale. AI as part of teledermatology services will require efficient, safe and effective teledermatology infrastructure.

The roadmap has been developed with stakeholders including Getting It Right First Time, clinicians, relevant professional bodies, such as the British Association of Dermatologists, Primary Care Dermatology Society, British Dermatological Nursing Group, the Institute of Medical Illustrators and organisations that represent the views and interests of people with lived experience, including the National Eczema Society, Vitiligo Support UK and Psoriasis Association UK.

Steps to accelerate the roll out of teledermatology

  • Review the dermatology service to identify where teledermatology fits in; link that to a review of low priority frameworks and referral guidance.
  • Prioritise patient safety considerations from the outset; in particular when replacing a face-to-face interaction with a teledermatology one.
  • Ensure patients are informed about their teledermatology care pathway, diagnosis and management plan, in a clear, compassionate and timely manner (principle 1).
  • Recognise that not all patients or dermatological lesions are suitable for teledermatology; for example, rashes and lesions may be more challenging to assess and diagnose in patients with brown and black skin. Involving local service users in any teledermatology pathway redesign should ensure it does not disadvantage patients or exacerbate health inequalities (principle 1).
  • Identify all workload for teledermatology to ensure that it does not add burden. The teledermatology service should reduce delays in the care pathway for patients without making extra work for healthcare professionals (principle 2).
  • Identify a project team to work with all partners across the local healthcare system (including primary, intermediate and secondary care) and engage with patients, to agree a model for teledermatology services (step 1).
  • Identify who will take high-quality images (including dermoscopic images) and at which location, and who will review them; consider a pan-ICS approach (step 2).
  • Identify the resources and funding required to design and implement the model, including capital, rollout and annual expenditure (step 3) and be clear about funding arrangements.
  • Maintain teledermatology pathways through continuous training across professional groups and care settings (step 4).
  • Accurately record and audit teledermatology activity (step 5) to:
    1. reflect the type of clinical contact
    2. demonstrate benefits
    3. learn from any drawbacks
    4. support sustainable funding.

Why teledermatology?

There were about 3 million dermatology outpatient appointments in England in the year to April 2022, and dermatology services receive more urgent referrals for suspected cancer than any other specialty.

About half the 1 million dermatology referrals per year are urgent suspected skin cancer referrals (previously known as the 2-week wait skin cancer pathway). Many of those referred on this pathway have benign (non-clinically concerning) skin lesions, but about 8% will be diagnosed with melanoma and squamous cell carcinoma (SCC).

Teledermatology can reduce the number of face-to-face interactions offered to people with low-risk benign lesions as skin lesions can often be accurately diagnosed from high quality, in focus images. With macroscopic images many benign skin lesions can be diagnosed and managed, and some SCC and many basal cell carcinomas (BCC) directly booked for surgery; but both dermoscopic and macroscopic images are required for accurate diagnosis of pigmented lesions and melanoma.

A pan-system, locally agreed dermatology pathway redesign integrating teledermatology and artificial intelligence (AI) (subject to satisfactory evaluation) should reduce the need to see people with benign skin lesions face-to-face and prioritise those with skin cancer. As a result, those with rashes and long-term skin conditions will be seen and treated more quickly.

A teledermatology service must also increase capacity in intermediate or secondary care for patients who need face-to-face consultations; the clinical time released must be more than that needed to evaluate images. This will improve equity of access for everyone with skin disease, so patients are seen in the right place, by the right person, at the right time.

Principle 1: patient centred care

A teledermatology service must provide high-quality timely care for the right patients. People must be given clear information about their care.

  • Patients must be given clear information about why images are needed, how they will be transferred and stored, and their future care, so that they can give informed consent for the use of the images (eg clinical care, teaching, research).
  • Pathways should enable patient choice. Clinicians must be sensitive to people’s concerns about being photographed and not coerce patients, directly or indirectly, to be photographed or share images of themselves. This is particularly important for skin conditions affecting intimate body sites where relevant guidance should be followed.
  • Patients must not be disadvantaged if they cannot provide images or do not wish to have images taken. People may not have access to a smart device, they may need someone else to take the photographs and dermoscopy images will usually be needed for pigmented lesions and melanoma. Therefore, patients must be offered the option to have images taken by appropriately qualified healthcare professionals in convenient settings. Equity of access must be considered when redesigning dermatology pathways.
  • The outcome of the referral must be communicated to the primary healthcare professional and patient in a manner agreed when the teledermatology service was set up and depending on the referral pathway.
  • Not all patients or dermatological lesions will be suitable for teledermatology; in particular, it may be more challenging to assess and diagnose conditions from photographs in patients with brown or black skin.
  • Patients with long-term inflammatory skin disease should usually only be managed via teledermatology provided their referring clinician has the facilities and clinical experience to provide ongoing support and review based on the skin care management plan from the reporting skin specialist.

Principle 2: avoid creating additional burden

Sustainable teledermatology pathways should use digital tools and be quick and simple, so as not to increase burden for healthcare professionals or create more steps in the patient journey. The following are important considerations:

  • Teledermatology services require careful modelling and pathway design to reduce hospital attendances and make productivity gains. Consider pan-region models to share learning and benefit from economies of scale.
  • Dedicated teledermatology platforms can add value by offering a direct patient interface, streamlining data capture and optimising the referral process. But they can be costly and time-consuming to implement and are not a prerequisite for initiating teledermatology services.
  • Administrative support is essential to realise efficiency gains with improved slot use and cost-effective use of clinical time.
  • There should be a threshold of disease severity for asking for dermatology advice – this should be equivalent to that when an outpatient referral would be made. Although this may be difficult to achieve, a threshold is important to avoid overwhelming services and creating unwarranted variation.
  • High quality, accurate images will reduce unnecessary ‘double activity’ of a teledermatology referral followed by a face-to-face consultation.
  • Images must be accompanied by relevant clinical information (template proformas are available on Future NHS [login required]) so that the reporting clinician has enough information to make a diagnosis and develop a management plan at the first teledermatology interaction.

Step 1: identify the role of teledermatology in the local dermatology pathway

Systems should review local pathways of care for people with skin conditions and agree with key stakeholders the role of teledermatology in the whole service.

This pan-system approach will require:

Consider including one or more of the following teledermatology models in the redesign of local dermatology services.

1.1 Use of digital images to enable specialist advice and guidance (A&G) as the front door to dermatology services

The greatest demand on dermatology services is for the diagnosis and management of skin lesions. Appropriate clinicians can often determine that a lesion is benign from good images supported by adequate clinical information.

The use of specialist A&G supported by images, including dermoscopic images for skin lesions, with the option of conversion to a referral, is therefore recommended as the ‘front door’ to all dermatology services, including for people with rashes, except urgent suspected skin cancer referrals (previously known as the 2-week wait skin cancer pathway). This will enable people to be triaged to the appropriate clinical service. It is particularly suited to those who have a single lesion of concern.

Specialist A&G supported by images is provided by secondary care specialist teams and GPs with an extended role (GPwERs/GPwSI) working in intermediate dermatology services (appropriate clinicians) reviewing teledermatology referrals into their service to decide whether the patient:

  • can be managed in primary care with A&G
  • should be seen urgently or routinely by the relevant clinician in the best setting for the patient, which might be intermediate care or secondary care
  • needs their referral updating to a suspected skin cancer 2-week wait appointment
  • would benefit from a remote consultation (video or telephone)
  • can be booked directly for further diagnostics or surgery into an appropriate service in a timely fashion
  • should be redirected to a more appropriate service such as plastic surgery, ophthalmology or maxillofacial, according to local agreements, pathways and skill mix.

This will streamline and prioritise patient care, and ensure face-to-face attendances are only offered when necessary. There will still be instances where it is not possible or appropriate for the patient to provide an image (see principle 1).

1.2 Virtual urgent suspected skin cancer pathway (previously known as the 2-week wait skin cancer pathway)

The need to appropriately manage the increase in referrals for suspected skin cancer requires efficient identification of people with early melanoma and SCC.

Latest guidance to optimise the urgent suspected skin cancer pathway (previously known as the 2-week wait skin cancer pathway) removes the requirement for a face-to-face consultation and instead proposes a teledermatology interaction for selected patients. This meets the requirements of the NHS Constitution. The guidance identifies the considerations and requirements for implementation of this pathway.

Although general practitioners with extended roles (GPwERs)/general practitioners with a special interest (GPwSIs) working in integrated intermediate community dermatology services may triage skin lesions, they will not review urgent suspected skin cancer referrals (previously known as the 2-week wait skin cancer pathway) unless they are members of the specialist dermatology urgent suspected skin cancer (2-week wait) service.

1.3 Validation and clinical prioritisation of patients on non-admitted waiting list

The guidance on the clinical prioritisation of the dermatology on-admitted (outpatient) waiting list is underpinned by the need for high quality, recent images. Prioritisation triage is done without direct patient interaction and requires information, including outcome of a recent patient discussion and the original referral letter, as well as recent images. Each patient is then prioritised to reflect need and urgency for a face-to-face consultation. If the patient’s condition changes while waiting, A&G teledermatology can highlight the change in status and support appropriate reprioritisation.

Step 2: design the teledermatology service

High quality images should be taken in an appropriate setting that is convenient for the patient, sent using secure platforms for review with relevant supporting clinical information. A suitably trained appropriate clinician should manage the outcome of this review. Further information on designing and setting up teledermatology services can be found in the Institute of Medical Illustrators’ (IMI) teledermatology toolkit.

Locally agreed solutions will be important; one size does not fit all.

2.1 Image taking services and collecting supporting clinical information

Who can take the images?

Anyone taking images and collecting the required supporting clinical information will require appropriate training, feedback and audit of performance. More detail relating to taking images using mobile devices, can be found in guidance on the use of mobile photographic devices in dermatology. The IMI is developing training packages as part of its teledermatology toolkit. These will be available in due course.

Primary care

Suitably trained primary care healthcare professionals may be well placed to take images in primary care settings, which are convenient locations for the patient. However, GPs or advanced clinical practitioners (ACPs) should only be expected to do this if they have the capacity. Primary care-based image taking hubs established across ‘places’ or primary care networks (PCNs) work well.

Clinical photographers

Clinical photographers working in specialist dermatology teams may also be able to work from community settings. This approach would mean images were of consistently high quality and accurate, improving clinical decision-making.

Clinical photography services can train suitable healthcare professionals, including nurses, technicians and healthcare support workers, to take images, including dermoscopic images.

Members of the dermatology specialist team

Suitably trained members of the specialist dermatology team, such as specialist dermatology nurses, can take images and are likely to know what clinical information is needed to support the images. They can also provide information and advice about ‘next steps’ and prepare patients for a possible surgical procedure.

Images taken by patients

Local agreement is required about whether patients’ own images are appropriate for the teledermatology pathways. The following need to be considered:

  • The quality of images taken by patients can be variable. If patients are being asked to take their own images, resources are available to guide them.
  • Secondary and community care-led teledermatology services can use patient communication platforms to send patients a link to instructions on how to take and share images. However, until image quality can be assured and appropriate dermoscopic devices have been developed for patients, this approach will not be suitable for suspected skin cancer pathways where a dermoscopic image is required.

Taking the images: where and what equipment?

Images should be taken in a convenient and accessible location for the patient, such as primary care facilities, community hospitals and community diagnostic centres. Convenience and accessibility considerations are particularly important where images are taken in a hospital setting, as this must confer benefit over a face-to-face appointment in the same setting.

The space should meet space and lighting standards to optimise the image quality and ensure privacy and dignity for the patient.

Required equipment will include dermatoscope attachments for cameras and smartphones, as well as access to information governance compliant technology solutions.

What images and clinical information is required?

The number and type of images for different clinical scenarios should be agreed locally; for example, for skin lesions one locator image and at least one close-up with a measurement scale and one without. Dermoscopic images are essential for the urgent suspected skin cancer pathway (previously known as the 2-week wait pathway). For lesions that are raised from the skin, it may be appropriate to take an extra close-up to show the lesion in profile. Resources to support image taking are available from the Primary Care Dermatology Society and the NHS.

Referrals should be accompanied by a referral proforma or a conventional referral letter, which should be agreed locally. It is likely that different proformas will be required for skin lesions and rashes; existing urgent suspected skin cancer (2-week wait suspected skin cancer) referral proformas can be modified so that their use can be extended to the virtual pathway.

2.2 Sending and transferring the images

All systems must comply with General Data Protection Regulations (GDPR), including cloud data storage systems. There should be clear processes for the capture and storage of any data alongside the images.

Specialist A&G with images is advocated as the main route to specialist opinion for primary care healthcare professionals. Specialist A&G can be requested via e-referral system (e-RS) or alternative commercial platforms, which can link to e-RS pathways should conversion to referral be required. 

GDPR compliant apps to support image sharing can be commissioned as standalone apps to allow images to be securely uploaded to e-RS, or alternatively as part of a potential alternative commercial A&G pathway. These allow images to be captured securely with personal mobile phones and transferred safely to e-RS or other clinical systems. Further information can be found in the dermatology digital playbook and on the British Association of Dermatologists website.

Other points of note:

  • There may be limitations to the file size and number of images that can be attached, and service users need to be aware of these.
  • Some commercially available GDPR compliant photo apps can automatically adjust the file size and resolution to support attachment to e-RS. Application-programme interfaces (APIs) have been developed that allow external applications to initiate referrals on e-RS and can be used to deliver similar functionality. However, this depends on external application suppliers co-operating with the APIs.
  • A&G requests supported by images can be converted to a referral, if pre-emptively authorised by the referrer. This functionality is available to A&G requests raised through e-RS only. APIs have also been developed that allow providers to manage e-RS A&G conversations and referral conversion within their own systems. System suppliers will need to develop their systems against these APIs.

Image storage and retrieval

There needs to be clarity for all involved about where the images are stored and how they will be readily retrieved by the reviewing clinicians.

  • Patient images supporting either a referral or A&G request should be accessible from provider hospital IT platforms, either in e-RS or transferred via a suitable GDPR compliant platform, and ideally stored directly in both the primary care and specialist service electronic patient records.
  • Receiving clinicians should be able to easily review images and patient information in a digital format and action them directly in the system.
  • Images received via e-RS referrals will remain archived in e-RS and accessible from both primary and secondary care. Patient images should be added to the electronic patient record or to a secure digital image management database.
  • National e-RS workstreams are in place to improve the interoperability between e-RS referrer and providers to facilitate transfer of A&G dialogue and images into primary and secondary care patient records, reducing the burden on administrative teams.
  • Patient images that are sent via email, even to a secure mailbox, need to be managed and stored properly in the patient record. Without a suitable archive system, the images may not be immediately accessible or require the clinician or administrator to perform many separate computer processes.
  • Many clinical photography departments manage such mailboxes and upload images to patient records so this is something that clinical photographers may be able to support.

2.3 Review of the images, management and communication of the outcome

Appropriate clinicians

Systems need to identify the appropriate clinicians who will review the images and communicate the outcome to the referrer and/or the patient in a timely fashion. The method of communication and who will communicate the outcome should be agreed as part of the design of the service. Capacity modelling is required to ensure that adequate resource is available to manage the expected activity, although good teledermatology should reduce the overall requirement for clinical time.

Appropriate clinicians include:

  • members of secondary care specialist teams; may include suitably trained specialist nurses
  • GPwERs/GPSIs, ideally working in consultant-led intermediate dermatology services
  • suitably trained private providers offering outsourced teledermatology services; working to agreed standards and ideally linked to the local dermatology service.

Usually, clinicians undertaking teledermatology activity would be expected to maintain their clinical skills by also doing face-to-face consultations. Except for specific situations, teledermatology is expected to be only one part of the clinical activity most clinicians undertake.

In face-to-face medicine it is considered poor practice for clinicians to work in isolation and the same applies to teledermatology. By working collaboratively clinicians can give rapid second opinions for – and have the opportunity to discuss – borderline or difficult cases, share learning and reduce the risk of mistakes. Skin cancer recognition requires a high degree of diagnostic accuracy and regular calibration of clinical practice; working alongside other clinicians helps maintain this and can identify early any deviation from safe practice.

The following considerations are also important:

  • Appropriate time for teledermatology should be included in the appropriate clinician’s weekly sessional timetable; there should be time to review the images and all supporting administration.
  • In secondary care, teledermatology should be included in the direct clinical care (DCC) activities for doctors, using the time previously spent on face-to-face appointments.
  • Creating teledermatology activity will inevitably take clinicians away from other clinical activity; this needs to be recognised and managed appropriately.
  • Guidance from the British Association of Dermatologists recommends the number of teledermatology interactions in a programmed activity.

Managing and communicating the outcome

The appropriate clinician should be able to manage appropriate outcomes in a timely fashion. The outcomes should be agreed during development of the teledermatology pathway, and important considerations include:

  • Outcomes should be linked to a locally agreed optimal pathway and will include options such as reassurance, advice about treatments, booking for surgery, arranging a routine or urgent face-to-face consultation.
  • Systems should agree reasonable timeframes for an appropriate clinician to respond to a specialist A&G teledermatology interaction.
  • For the urgent suspected skin cancer (previously known as the 2-week wait) virtual teledermatology pathway, systems must adhere to the relevant 28-day Faster Diagnostic Standard ‘clock stop’ guidance.
  • Agreeing clinical responsibility for timely and effective communication with the referring clinician and patient; where possible include links to relevant patient information resources.
  • The teledermatology interaction should provide the opportunity for shared learning between primary care referrers and clinicians reviewing the images.
  • Where teledermatology services are outsourced to private providers, systems must have clearly identified and agreed pathways to link activity to local dermatology services, which patients can access without advantage or disadvantage, should the outcome of the interaction require this. It is important that there is no ‘queue jumping’ and images are available on the relevant patient record.

Step 3: Accurately model and identify the resources required, including funding, to set up and maintain the service

Teledermatology funding and payments

Integrated care systems will fund teledermatology services using existing locally agreed funding arrangements, usually as part of block contracts. The expectation is that new arrangements will be developed in due course. Monitoring and recording teledermatology activity and costs is important to support the development of sustainable funding models.

3.1 Setting up the teledermatology service

Time should be allocated for all those involved in designing and setting up the teledermatology service, including short-term project management to:

  • Develop and administer the pan-health community stakeholder group and support clinician and user engagement.
  • Agree the model of teledermatology service, design the patient pathway and supporting resources, such as referral proformas, patient questionnaires, consent forms and letter templates.
  • Develop and implement the image taking service, including training needs for those taking the images.
  • Identify the platform to deliver the service and train users, including clinical and non-clinical staff, in all required technology.
  • Collect baseline data, develop outcome measures and ensure prospective collection of performance data so that the service can be constantly evaluated (including at a clinician level) and modified as required.
  • Run an initial pilot to identify issues, establish realistic expectations and ensure adequate job planning.

3.2 Primary care healthcare professionals

The burden on primary care healthcare professionals should be kept to a minimum. The following considerations are important:

  • A teledermatology ‘champion’ should be identified in a primary care network; time and resource should be allocated for this role.
  • Where possible, teledermatology systems should be well integrated with the practice system for future reference.
  • For staff and patients to be confident that the digital tools being used are clinically safe, digital technology assessment criteria (DTAC) standards should be met.
  • Systems should be simple to use and set up; implementation should be supported by training for all those involved.

3.3 Image taking services

The resources required to provide an image taking service will include:

  • Equipping an appropriate space with cameras or smart phones, dermatoscopes and adapters, laptops, lighting and internet access. The IMI provides clinical photography studio design and planning guidance.
  • Recruitment and training of an appropriate workforce to take and send the images with the necessary clinical information, funding for ongoing training development support and audit of service.
  • Any licensing costs for software packages.
  • Ability to securely link into and upload images to relevant secure systems.

3.4 Appropriate clinicians reviewing teledermatology referrals

The clinical time for appropriate clinicians to perform teledermatology activity should be carefully itemised and include time to perform the following:

  • Review the images and clinical information received.
  • Respond to the referring clinician (and patient where appropriate) using the agreed process for communication.
  • Perform any necessary administrative processes such as arranging skin surgery and sending the patient relevant information.
  • Undertake regular review and audit of teledermatology activity to quantify patient experience, short and long-term outcomes, service efficiency and benefits analysis.
  • Teach, train and support other appropriate clinicians.

Extra clerical and administrative support will need to be identified; booking patients into appointments, particularly as part of timed skin cancer pathways is more time consuming than for similar patients on an urgent pathway (previously known as a 2-week wait pathway).

Step 4: training and development

Continuous training and sharing learning are necessary to maintain teledermatology pathways as part of a sustainable, integrated multidisciplinary dermatology service.

Training on teledermatology processes should be regularly refreshed for primary, intermediate and secondary care teams, and the training needs considered for the entire clinical and administrative workforce. Any healthcare professional working within teledermatology services will be expected to meet all relevant nationally agreed continuing professional development standards appropriate to their roles.

Staff could miss software developments that may provide opportunities for greater productivity if they are not made aware of them or do not know how to use them. New staff joining a practice or department should be trained in the local teledermatology pathway as part of their induction. The training needs of specialist registrars in primary and secondary care need to be considered. Developing local teledermatology champions across primary and secondary care can help sustain this focus on training.

Training also needs to be provided for those taking and uploading high quality images. For skin lesions this will include taking high quality dermoscopic images.

Step 5: audit, metrics and quality assurance

The teledermatology service should be regularly reviewed, metrics agreed at the outset, and data collected and shared regularly with stakeholders. Baseline metrics are essential before the service is introduced and will include:

  • patient satisfaction with the service and the quality of advice provided; did they think their care was managed appropriately and in a timely fashion?
  • number and type of teledermatology interactions
  • outcomes of the interactions; for example, advice only, conversion to face-to-face interaction, upgrade to urgent suspected skin cancer (previously known as 2-week wait) referral
  • clinician time, capacity and demand modelling for the service
  • ‘burden’ impact on those involved
  • referral data from individual practices, including levels of unsatisfactory images received to support targeted educational interventions
  • audit of referrers against referral criteria, diagnosis and outcomes
  • impact on waiting list and waiting times
  • effect on overall referral rates (pre and post teledermatology)
  • A&G data at departmental and clinician level: monthly A&G, % returned with advice and % not requiring appointment within 6 months, % converted, turnaround time (<2 days, 3–5 days, 6–10 days)
  • urgent suspected skin cancer (previously known as 2-week wait) teledermatology: turnaround against Faster Diagnostic Standards, % returned with advice, % booked for face-to-face, % booked straight to surgery, peer review of outcome decisions to improve standardisation across teams
  • delayed diagnosis and surgery of skin cancers using measures such as re-referrals.

A benefits analysis of the time and resource invested in the development of the service is needed.

Audits are suggested in Primary Care Commissioning’s Quality standards for teledermatology and should include:

  • obtaining the views and feedback from service users and providers, including patients and staff
  • reviewing the quality of the teledermatology referrals, in particular image quality
  • assessment of effectiveness of communication between healthcare professionals and patients
  • organisation, storage and retrieval of data (information governance audit)
  • training and continuing professional development needs.

For clinicians reviewing teledermatology referrals:

  • audit their practice in respect of the numbers of patients converted to face-to-face appointments
  • GP and patient satisfaction
  • clinical diagnoses and outcomes, particularly of those with suspected skin cancer.

Going further

This roadmap outlines what all systems can do accelerate the rollout of teledermatology services to manage demand and reduce unnecessary patient attendances, thereby freeing up outpatient capacity.

Further opportunities have been identified for digital technology and teledermatology to be used in new and innovative ways to deliver more personalised and better integrated care:

  • Shared patient records give patients greater control over their condition. Images of the patient’s skin condition, results from blood tests and other information can be shared with the specialist team to be reviewed in their appointment or to support a personalised schedule for follow-up appointments.
  • Teledermatology can support integrated care between different secondary care and community providers by supporting access for patients in areas with particularly low numbers of consultants. It can also be used to allow virtual multidisciplinary specialist teams to manage complex patients.
  • Apps are available and in development that can help patients take pictures of their skin condition and monitor changes over time. This technology could be used to help patients monitor their own condition, and support dermatology pathways.

Artificial intelligence as a medical device (AIaMD)

AI solutions are being rapidly adopted across healthcare settings. Systems should be developing and implementing teledermatology services underpinned by high quality image taking services so that they will be ‘AI ready’ and able take advantage of AI solutions to upscale at pace.

The Regulatory Horizons Council recognised the importance of having an appropriate regulatory framework that permits the innovation of digital technologies without exposing people to AIaMDs that could do them harm. The NHS AI Lab is working with innovators as they define the purpose of their products, and guiding and developing recommendations to health and care professionals as they start using these technologies to assist them in providing care.

Currently all AIaMDs being used on NHS patients must have a UK Conformity Assessment (UKCA) medical device classification and be used accordingly. Class IIA medical devices have the required MHRA classification to be used for direct diagnosis, and at the time of publication the only UKCA Class IIA-certified AIaMD is one that is being used to triage lesions in urgent skin cancer teledermatology pathways in primary and secondary care at various centres across England. However, there are over 60 commercially available AI technologies in the UK focused on skin cancer diagnostics/triage, with variable and often limited published and independently verifiable data supporting their safety and effectiveness. Further peer reviewed data and evaluations are awaited.

AIaMDs have the potential to manage the increased demand for dermatology services by accurately recognising benign skin lesions in primary and specialist services. The current deployments include a consultant dermatologist review of all cases following categorisation by the AI, this builds confidence for the user in the AI tool, although the intended use does not require this second review to happen. But for AIaMD to be deployed at scale in urgent skin cancer pathways, clinical and patient confidence in the safety and effectiveness of the technology is critical, and to this end, in 2023/24 there is a focus on accelerating real-world independent evaluation to provide assurance of the sensitivity, specificity, generalisability, patient acceptability and system integration of any solutions.

Successful deployment of AlaMD in dermatology will need:

  • Robust and standardised teledermatology services supported by good quality images and relevant clinical information.
  • A clearly defined intended use of the AlaMD, deployed accordingly. The Medicines and Healthcare products Regulatory Agency has published guidance on clearly defining the intended use or purpose of the AIaMD.
  • Appropriate information provided to the patient about the intended purpose of the AIaMD technology, what data will be collected, how their data will be used/shared/stored, and that consent must be documented.
  • Appropriate image capture devices that comply with the necessary standards in terms of information governance.
  • Image quality to meet the minimum requirements for any given AIaMD; automated checks to ensure image quality is adequate.
  • Clearly defined inclusion and exclusion criteria that the AIaMD has been validated for, eg size or site of lesion such as the exclusion of mucosal, palmoplantar or subungual surfaces.
  • A pilot/study for an agreed number of cases to allow validation of performance and optimisation for safe onward deployment and upscaling for the given population, eg including a clinician second read.
  • Optimisation of data diversity through increased inclusion of underrepresented groups, such as those with brown or black skin and those with lower digital literacy levels, to avoid exacerbating pre-existing biases.
  • The tool to be inclusive and represent the diversity of the local population.
  • Agreement of a clear post-marketing surveillance plan before deployment. Regular performance reports should demonstrate ongoing clinical monitoring to ensure that the system is still meeting the required performance targets, and to identify algorithm drift or model decay; that is, the tendency for AI model performance to drop over time as data and patient characteristics change.
  • Local incident reporting tools to capture false negatives whereby a full root cause analysis should be completed, and a risk register compiled to identify system-wide issues or recurring themes.
  • Patients who are discharged to receive clear advice on how to monitor their skin, signs of skin cancer and how to access further care if they have future concerns.

In summary, AIaMD tools for dermatology are being evaluated and comprehensive data regarding the safety and effectiveness of the tools that are in use will be released in 2023/24. Systems should be developing and implementing teledermatology services underpinned by high quality image taking services so that they will be ‘AI ready’ and be able to upscale at pace.

Tools and further guidance

A range of resources are available on the dermatology pages on the Outpatient Recovery and Transformation FutureNHS platform (log in required).

Referral guidelines

Image taking

Other resources

Artificial intelligence

Publication reference: PR00029