196 CHAPTER 8 for their proper assessment and presented a 25-item checklist for the evaluation of imagebased AI algorithm reports in dermatology [66]. Also, the use of AI chatbots (such as Chat Generative Pre-trained Transformer (ChatGPT, developed by OpenAI) or Med-Palm (developed by Google)) are currently extremely popular, also in healthcare. We expect that these AI chatbots can play a major role in digital (dermatology) consultation in the future. In dermatology home consultation, for example, AI chatbots can aid GPs in formulating “standardized” or “tailored”’ understandable responses to the patients about their skin condition and suggested management plan. An AI chatbot can formulate within a few seconds a response to patients seeking dermatology care including a simple and patient-friendly vocabulary and possibly references to relevant evidence-based websites. The GP can verify these answers on correctness, can supplement the answer if necessary and send the adapted chatbot response to the patient instead of formulating the answer from scratch. In teledermatology and teledermoscopy, chatbots might likewise prompt GPs in asking relevant questions about the patient’s skin condition before submitting the consultation request to the TD. Furthermore, GPs and TDs can use the AI chatbots to access recent dermatology literature related to a specific skin condition or medical guidelines supporting them in making well-informed decisions. However, not all AI chatbots are designed for providing medical advice, have biases in the training datasets or lack specialized dermatology knowledge and terms [67,68]. Therefore, before these AI chatbots are validated and guidelines and regulations exist, it is important that GPs and TDs verify answers provided by these chatbots in their correctness when used in dermatology care. Likewise, the Royal Dutch Medical Association (Dutch: Koninklijke Nederlandse Maatschappij tot bevordering der Geneeskunst (KNMG)) predicts that in 2040 physicians will intensively use AI applications, but that physicians will also be the “human interface” for patients to interpret the data and information and will remain responsible for the diagnosis and treatment [69]. CONCLUSION This thesis has shown that teledermatology and teledermoscopy are integrated into Dutch general practice and support GPs in assessing and diagnosing patients’ skin lesions and in deciding whether a referral of the patient to the dermatologist is necessary. This thesis confirmed that teledermatology and teledermoscopy save patients’ travel and wait time compared to conventional dermatology care, and reduce the number of physical referrals to the dermatologist and healthcare costs. Additionally, by continued use of teledermoscopy and receiving TDs’ advice over years, GPs slightly improve their expertise and dermatology knowledge about diagnosing skin lesions. Furthermore, GPs consider the provided TD responses helpful and instructive. Another added value of availability of teledermoscopy is that GPs changed their initial referral decision after TD assessment in half of the consultations. This means that teledermoscopy aids in immediate referral of patients with (pre)malignant or complex skin lesions whom GPs initially did not
RkJQdWJsaXNoZXIy MTk4NDMw