Uporaba umetne inteligence pri oskrbi ran : diplomsko delo

Popovič, Filip Gabrijel (author), Ficzko, Jelena (mentor), Birk Tot, Karin (comentor), Pajnič, Manca (reviewer)

Uvod: Kronične rane predstavljajo veliko fizično breme za posameznika in finančno breme za zdravstveni sistem. Zaradi teh bremen znanstveniki raziskujejo možnosti uporabe novih tehnologij, s katerimi bi lahko pridobili objektivne meritve o površini, globini in stanju ran ter izboljšali njihovo oskrbo in način dokumentiranja zdravstvene obravnave. Ena izmed tehnologij, ki je pritegnila njihovo pozornost, je umetna inteligenca. Namen: Namen diplomskega dela je predstaviti ugotovitve raziskav s področja uporabe umetne inteligence pri oskrbi ran. Cilj diplomskega dela je na osnovi raziskav prikazati uporabo umetne inteligence pri oskrbi ran in morebitna tveganja pri njeni uporabi. Metode dela: V diplomskem delu smo izvedli narativni pregled literature. Za vir informacij smo izbrali dve elektronski podatkovni zbirki, CINAHL in Medline. Rezultati: Iz zbranih člankov smo v nadaljnjo obravnavo vključili 15 člankov, ki so ustrezali našim kriterijem. Na osnovi teh smo identificirali štiri področja, kjer so avtorji ugotavljali uporabnost umetne inteligence za oskrbo ran: zaznavanje področja ran in izvajanje meritev na fotografijah ran, klasifikacija ran, predvidevanje nastanka ran ter vodenje in pomoč pri oskrbi ran. Razprava in zaključek: V vseh izpostavljenih kategorijah so modeli umetne inteligence dosegli obetavne rezultate in avtorji so bili optimistični glede uporabe umetne inteligence pri oskrbi ran. Njihova uporaba pa bi lahko bila tvegana, če bi neizkušeni zdravstveni delavci vnašali napačne podatke. S temi podatki bi model umetne inteligence lahko predlagal nepravilne odgovore oziroma nasvete. Pogosta omejitev zbranih raziskav je, da so za testiranje modelov umetne inteligence uporabili majhno število podatkov ali pa so uporabili le eno ali dve vrsti ran. Zaradi tega je večina avtorjev v svojih raziskavah predlagala oziroma napovedala dodatna testiranja njihovih modelov umetne inteligence. Ker vse zbrane raziskave izvirajo iz tujine, priporočamo, da se v slovenskih zdravstvenih ustanovah izvedejo podobne raziskave s fokusom na komercialno dostopnih pripomočkih in programski opremi za oskrbo ran, ki uporabljajo umetno inteligenco. Prav tako bi bilo koristno raziskati odnos slovenskih zdravstvenih delavcev do uporabe umetne inteligence pri oskrbi ran in njihovo poznavanje tehnologij, povezanih z umetno inteligenco, saj bo ta v zdravstvu čedalje bolj prisotna.

• Language: Slovenian
• Keywords: diplomska dela, zdravstvena nega, umetna inteligenca, strojno učenje, kompleksne rane
• Work type: Bachelor thesis/paper
• Typology: 2.11 - Undergraduate Thesis
• Organization: ZF - Faculty of Health Sciences
• Place of publishing: Ljubljana
• Publisher: [F. G. Popovič]
• Year: 2024
• Number of pages: 41 str.
• PID: 20.500.12556/RUL-160468
• UDC: 616-083
• COBISS.SI-ID: 205676035
• Publication date in RUL: 29.08.2024

Secondary language

• Language: English
• Title: The use of artificial intelligence in wound care : diploma work

Abstract

Introduction: Chronic wounds represent a significant physical burden on a person and a financial burden for the healthcare system. This has led scientists to research if new technologies could gain objective measurements of the surface, depth and state of wounds, improve wound care, and improve healthcare documentation. One of the technologies, which has been the focus of researchers, is artificial intelligence. Purpose: The purpose of our diploma work is to present the findings of studies regarding the use of artificial intelligence in wound care. Our goal is to present the use of artificial intelligence in wound care with the help of gathered studies, as well as to highlight the possible risks that could come with its use. Methods: We performed a narrative literature review. We searched CINAHL and Medline for articles regarding the use of artificial intelligence in wound care. Results: From the gathered studies we chose 15 studies that fit our criteria. From them we identified four categories where authors were researching the use of artificial intelligence in wound care. These categories are: precepting wound area and gathering information about wounds from photographs, wound classification, predicting wounds, and helping manage wound care and offering advice regarding wound care. Discussion and conclusion: In all four identified categories, the artificial intelligence models performed adequality and authors were optimistic regarding its use in wound care. Using artificial intelligence may pose a risk, if untrained healthcare workers enter incorrect data. The model would then give false answers or advice. The most common limitation of these gathered articles was, that authors often used only one or two types of wounds or small amounts of data. Because of this, most authors recommended additional testing for their models. Due to the fact that all gathered studies were carried out in foreign countries, we recommend that similar studies be conducted in Slovenian healthcare institutions with a focus on commercially available programs and accessories that use artificial intelligence for wound care. Additionally, we would recommend researching the Slovenian healthcare workers attitudes towards using artificial intelligence in wound care and their knowledge of the technology used in artificial intelligence, as its use is becoming more and more widespread.

• Keywords: diploma theses, nursing care, artificial intelligence, machine learning, complex wounds