SEARCH THIS WEBSITE
As the video points out, anything which encourages skin checks is useful but the limitations of apps should be understood.
The Covid-19 pandemic brought at least one type of application to the fore, those used to make video calls. Teledermatology had been in place for several years but necessity created a sudden expansion in use.
This was of value, although not ideal for every case and the pandemic did deflect patients from care. Research projects and clinical trials also struggled, although a few managed to continue remotely.
They adopted technology to suit, allowing them to continue monitoring participants. This also avoids travel costs, or no shows, dedicated apps for research are likely to blossom but what about those offering wider diagnosis.
The Current Position
As the video above shows, Google have developed and are in the process of testing an app called Derm Assist. You upload 3 photos of your skin issue, answer a few questions and receive results on possible conditions.
Google is quick to state that what the app provides is not a final diagnosis, simply a realistic set of possibilities. Their result comes from machine learning, based on millions of related images, not from a doctor.
Numerous apps dedicated to skin diagnosis are available on download sites, such as Google Play, or the App Store. As with most topics, the quality varies significantly and studies of their efficacy is still at an early stage.
Research To Date
A study which graded 44 dermatology apps on content, accuracy, and design found that only 9 of them scored well and they needed revision to improve effectiveness. Similar results come from collating smaller studies.
Apps to identify skin cancers have been to the fore and whilst they can be more developed than most, risk assessments raise concerns. Advances have been made but no apps have matched the ability of a specialist.
Research on eczema self-management and diagnostic apps found that 34% provided misleading information and none complied with all required criteria. Psoriasis apps did fare better, although mainly in generating recommendations.
A study assessing a number of skin health mobile apps gave them an average score of 2.77 out of 5, perhaps a fair reflection of where we are. They do not represent scientific validity but there are upsides.
Positive Factors
App providers have a tendency to compare results with general medical, or nursing staff and a piece of 2021 research followed that plan to a degree. With the apps being used as support tools for staff, rather than solely by patients.
Diagnosis improved for 1 in every 8 to 10 cases and they felt the potential was there to increase quality. This matters in regions where specialist input is not available to patients, or can add to existing remote access.
Medical students use apps to help them study, staff of all types adopt them to improve efficiency, or communication. In public use they increase awareness, a factor in early diagnosis, as long as the one provided by the app is not taken as a finite result.
Practical Reality
The tech industry producing the apps has a well worn phrase, test and iterate. We are Guinea pigs in a developing field, which may one day take over medical diagnosis but that time is a long way off.
In a sense, dermatologists wouldn’t mind if their focus was on treatment, making patients well is what they live for. This can only however happen following accurate diagnosis and apps can not currently provide this.
Even GPs, with many years of training and experience need to refer to specialists. How about the lesions a patient hasn’t seen, or subtle symptoms which give pointers to a different result, or underlying cause.
Add to this the difficulty some have when using the apps, or when taking good quality images of their skin, or wrongly labelled images being used to train the apps.
App technology can be used in the future, including for monitoring a patient’s condition. If they make you more aware of your skin, this also helps but apps serving as a definitive tool will not be feasible for a number of years.