Including patient experiences from online sources in guidelines: A natural language processing study on scabies

OBJECTIVE: Including patients' experience-based knowledge in the development of clinical and public health guidelines has been shown to enhance the quality, relevance, and applicability of guidelines. However, the meaningful and methodologically sound inclusion of patient experiences remains a challenge. This study aimed to showcase the potential of NLP methods as an innovative tool for guideline development to gain insights into patients' experiential knowledge and to incorporate this into the guideline development process. METHODS: For the revision of the Dutch public health guideline for scabies, we analyzed patients' experiences with scabies infestation shared on "dokter.nl", the Netherlands' largest online health community, between December 4, 2014, and May 19, 2023. Structural topic modelling was performed to discern thematic clusters from these patient experiences. RESULTS: We obtained 5781 unique posts on scabies and identified 13 major themes raised in forum conversations. The most prevalent themes revolved around community support (11.2%), uncertainty about treatment plans (11.1%) and coping with itching (11%). Recognizing scabies, alternative remedies, and decontamination measures were also issues frequently raised. The analysis highlighted the burden of disease and treatment-particularly the psychosocial burden-associated with scabies. This offered guideline developers an unprecedented insight into patients' experiences resulting in alterations to the Dutch public health guideline for scabies. CONCLUSION: Previous studies have highlighted the benefits of the integration of experiential knowledge for guideline development. Our study provides a novel method to make this type of knowledge accessible and usable for medical guideline development, without additionally burdening patients.