ABSTRAK Nama : Fitri Annisa Ahlul Jannah Program Studi : Kesehatan Masyarakat Judul : Analisis Efektivitas Asesmen Risiko dan Sistem Rekomendasi Kanker Payudara Berbasis Artificial Intelligence Pembimbing : R. Sutiawan, S.Kom., MSi Angka kejadian kanker payudara meningkat setiap tahun, namun angka cakupan deteksi dini sebagai program penanggulangannya masih sangat rendah. Salah satu penyebab hal tersebut adalah tingginya beban kerja tenaga kesehatan di puskesmas. Tujuan penelitian ini adalah untuk mengembangkan chatbot asesmen risiko dan sistem rekomendasi, serta menilai efektivitasnya untuk membantu puskesmas dalam meningkatkan cakupan deteksi dini secara efektif dan efisien melalui pendekatan selektif. Desain penelitian menggunakan desain kualitatif studi kasus di Puskesmas Jayagiri Lembang untuk mengetahui efektivitas chatbot dalam menilai risiko kanker payudara dengan menerapkan pendekatan model waterfall dalam membangun model aplikasi chatbot. Chatbot—yang dikembangkan menggunakan teknologi natural language understanding dan conditional statements sehingga membuatnya lebih dinamis dalam berinteraksi dan mengurangi error—teruji sangat efektif untuk melakukan asesmen awal risiko kanker payudara. Namun, sistem ini perlu dilakukan upaya revalidasi dan pengembangan lebih lanjut sebelum dapat digunakan oleh masyarakat secara masif. Kata kunci: Kecerdasan buatan, asesmen risiko, kanker payudara, deteksi dini, informatika kesehatan

ABSTRACT Name : Fitri Annisa Ahlul Jannah Study Program : Bachelor of Public Health Title : Analysis of the Effectiveness of the Artificial Intelligence-Based Breast Cancer Risk Assessment and Recommendation System Counsellor : R. Sutiawan, S.Kom., MSi The incidence of breast cancer increases every year, but the coverage rate for early detection as a prevention program is still low. One of the reasons is the high workload of health workers at the public health center. The purpose of this study was to develop a risk assessment and recommendation system and assess its effectiveness to assist public health center in increasing the coverage of early detection effectively and efficiently through a selective approach. A qualitative case study design at the Jayagiri Lembang Public Health Center to determine the effectiveness of chatbots in assessing breast cancer risk by applying the waterfall model approach in building a chatbot application model was used for the research design. Natural language understanding and conditional statements were used by the developed chatbot hence made it more dynamic in interacting and preventing errors. It was also proven very effective in doing early risk breast cancer assessment. However, this system needs to be revalidated and further developed before it can be used by the community on a massive scale. Keywords: artificial intelligence, risk assessment, breast cancer, early detection, health informatics

The tuberculosis treatment success rate in Indonesia in 2023 did not reach the 90% target. Treatment success impacts the reduction of infection spread and drug resistance cases, making early prediction of treatment success crucial. This study aims to develop a machine-learning model to predict treatment success. Data from Indonesia's Tuberculosis Information System (SITB) cohort was used. The study included productive-age patients (15-64 years) diagnosed with drug-sensitive tuberculosis who received treatment from January 1, 2020, to December 31, 2023. Data was randomly split into training (80%) and testing (20%) sets for model validation, with cross-validation performed. The algorithms used include decision tree, random forest, multilayer perception, extreme gradient boosting, and logistic regression. A consensus was reached for decision-making variables required in performing machine learning-based modeling of SITB data to predict treatment success using modeling of SITB data to predict treatment success using the Delphi method. The results of the study show that the random forest machine learning algorithm had the best performance and highest accuracy in predicting treatment success. This machine learning–based prediction tool can provide early predictions with SHAP (SHapley Additive ExPlanations) interpretation, helping healthcare workers make informed decisions more easily.

Penggunaan Rekam Medis Elektronik (RME) mempunyai tantangan tersendiri dalam praktiknya. Meski tak dipungkiri memiliki manfaat, namun masih ada inefisiensi dan inefektivitas dalam implementasinya dalam praktik klinis. Salah satunya gangguan komunikasi dan interaksi dokter-pasien sebagai akibat dari teralihkannya fokus dokter pada layar komputer acap kali terjadi terutama pada pasien rawat jalan baru. Dalam sistem RME penilaian efisiensi dan efektivitas ini dilakukan menggunakan SUS (System Usability Scale) dan UEQ (User Experience Questionnaire) yang berisikan sejumlah pertanyaan yang mewakili efisiensi dan efektivitas. Solusi dari masalah tersebut mungkin bisa terjawab dengan fitur speech-to-text sebagai bagian kecerdasan buatan serta pengintegrasian RME dalam kecerdasan buatan. Penelitian ini merupakan penelitian kuantitatif dengan desain potong lintang. Dokter-dokter spesialis selaku responden melakukan penilaian terhadap dua sistem, yaitu RME dan prototipe sistem RME dengan fitur speech-to-text berbasis kecerdasan buatan, yaitu pengubahan percakapan menjadi tulisan dalam komputer RME yang kemudian diintegrasikan kecerdasan buatan yang sudah ada dalam dunia kecerdasan buatan (misal: gemini AI). Usability diukur menggunakan System Usability Scale (SUS), sedangkan user experience diukur menggunakan User Experience Questionnaire (UEQ). Hasil penelitian menunjukkan bahwa RME yang digunakan saat ini mendapat median skor SUS 57,5 (5-70) dengan kategori usability “marginal” (mendekati “not acceptable”) dan user experience mendapat skor Daya tarik: 0,401; Kejelasan: 0,648; Efisiensi: 0,234; Ketepatan: 0,508; Stimulasi: 0,352; Kebaruan: -0,07 dengan kategori "bad" pada seluruh komponennya. Sementara itu, prototipe sistem RME dengan fitur speech-to-text berbasis kecerdasan buatan memiliki rerata skor SUS 69 ± 14.3 dengan kategori “marginal” (mendekati “acceptable”) dan user experience dengan skor Daya tarik: 1,880; Kejelasan: 1,813; Efisiensi: 1,852; Ketepatan: 1,172; Stimulasi: 1,938; Kebaruan: 2,078 dan termasuk "excellent" pada kategori daya tarik, stimulasi, dan kebaruan, "good" pada kategori kejelasan dan efisiensi, atau "above average" pada kategori ketepatan. Terdapat perbedaan yang bermakna antara kedua sistem RME, baik pada usability maupun user experience. Kesimpulannya skor SUS dan UEQ prototipe yang dibuat ini lebih dapat diterima daripada RME sebelumnya.

The use of Electronic Medical Records (EMRs) in clinical practice still faces several challenges. Even with numerous benefits, there are multiple inefficiencies and ineffectiveness in implementing through clinical practice. Doctor–patient communication is often disrupted and interaction may decline because physicians’ attention is frequently diverted to computer screens, especially towards newly admitted outpatients. In this EMR system, the evaluation of efficiency and effectiveness is conducted using System Usability Scale (SUS) and the User Experience Questionnaire (UEQ), which consists of a series of questions representing both efficiency and effectiveness. The proposed solution to these issues may be addressed through the implementation of speech-to-text features as part of artificial intelligence, as well as the integration of EMR systems with artificial intelligence. This study was a quantitative study with a cross-sectional design. Specialist physicians, as respondents, evaluated two systems: the existing EMR and an EMR prototype with an artificial intelligence–based speech-to-text feature, which converts spoken conversations into written text within the EMR system and is further integrated with existing artificial intelligence technologies (i.e., gemini artificial intelligence). Usability was measured using the System Usability Scale (SUS), while user experience was measured using the User Experience Questionnaire (UEQ). The results showed that the current EMR had a median SUS score of 57.5 (5–70), categorized as “marginal” usability and close to “not acceptable,” while its user experience scores were Attractiveness: 0.401; Perspicuity: 0.648; Efficiency: 0.234; Dependability: 0.508; Stimulation: 0.352; and Novelty: -0.07, all of which were categorized as “bad” in all of its components. Meanwhile, the EMR prototype with an artificial intelligence–based speech-to-text feature had a mean SUS score of 69 ± 14.3, categorized as “marginal” usability and close to “acceptable.” Its user experience scores were Attractiveness: 1.880; Perspicuity: 1.813; Efficiency: 1.852; Dependability: 1.172; Stimulation: 1.938; and Novelty: 2.078, which were categorized as “excellent” for attractiveness, stimulation, and novelty; “good” for perspicuity and efficiency; and “above average” for dependability. There were statistically significant differences between the two EMR systems in both usability and user experience. The conclusion is that the SUQ and UEQ scores in this prototype is more acceptable than the old EMR.