[Artificial intelligence to support telemedicine in Africa]

Telemedizin findet seit Jahrzehnten Anwendung im Alltag von Dermatologen. Insbesondere in afrikanischen Ländern mit begrenzter medizinischer Versorgung, zu überbrückenden geografischen Distanzen und einem zwischenzeitlich relativ gut ausgebauten Telekommunikationssektor liegen die Vorteile auf der Hand. Nationale und internationale Arbeitsgruppen unterstützen den Aufbau von teledermatologischen Projekten und bedienen sich in den letzten Jahren zunehmend KI(künstliche Intelligenz)-gestützter Technologien, um Ärzte vor Ort zu unterstützen. Vor diesem Hintergrund stellen ethnische Variationen eine besondere Herausforderung in der Entwicklung automatisierter Algorithmen dar. Um die Genauigkeit der Systeme weiter zu verbessern und globalisieren zu können, ist es wichtig, die Zahl der verfügbaren klinischen Daten zu erhöhen. Dies kann nur mit der aktiven Beteiligung der lokalen Gesundheitsversorger sowie der dermatologischen Gemeinschaft gelingen und muss stets im Interesse des einzelnen Patienten erfolgen.

Linkage between patients' characteristics and prescribed systemic treatments for psoriasis: a semantic connectivity map analysis of the Swiss Dermatology Network for Targeted Therapies registry

BACKGROUND

Several treatment options are currently available for the treatment of psoriasis.

OBJECTIVE

To explore the main associations between patients' characteristics and systemic treatments prescribed for psoriasis in a large group of patients observed in real-life clinical practice.

METHODS

This was a retrospective analysis of baseline data collected within the Swiss Dermatology Network for Targeted Therapies registry in Switzerland between March 2011 and December 2017. Semantic map analysis was used in order to capture the best associations between variables taking into account other covariates in the system.

RESULTS

A total of 549 patients (mean age 46.7 ± 14.7 years) were included in the analysis. Conventional therapies such as retinoids and methotrexate were associated with no previous systemic therapies for psoriasis, a moderate quality of life (QoL) at therapy onset and older age (≥60 years). Fumaric acid derivatives were associated with mild psoriasis (psoriasis area severity index < 10) and long disease duration (≥20 years). On the other side, cyclosporine and psoralen and ultraviolet A/ultraviolet B treatments were linked to a more severe condition, including impaired QoL, hospitalization and inability to work. Regarding biological therapies, both infliximab and adalimumab were connected to the presence of psoriatic arthritis, severe disease condition and other comorbidities, including chronic liver or kidney diseases and tuberculosis. Etanercept, ustekinumab and secukinumab were all connected to a complex history of previous systemic treatments for psoriasis, moderate disease condition, overweight and university education.

CONCLUSIONS

The analysis shows multifaceted associations between patients' characteristics, comorbidities, disease severity and systemic treatments prescribed for psoriasis. In particular, our semantic map indicates that comorbidities play a central role in decision-making of systemic treatments usage for psoriasis. Future studies should further investigate specific connections emerging from our data.