Latar belakang: Keberhasilan ASI eksklusif di Indonesia masih rendah. Self-efficacy ayah dan ibu berperan penting dalam praktik menyusui. Penelitian ini bertujuan mendapatkan model prediksi penghentian ASI eksklusif pada 0-12 postpartum menggunakan Paternal dan Maternal BSE framework.
Metode: Penelitian menggunakan metode kualitatif dan kuantitatif. Studi kualitatif menggunakan FGD terhadap 5 informan ayah secara purposive dengan tujuan menggali infomasi mengenai paternal BSE. Desain studi kuantitatif longitudinal bertujuan mengevaluasi model penelitian. Sebanyak 201 pasangan yang bersalin di rumah sakit ibu dan anak di Kota Tangerang Selatan berhasil diikuti sampai 12 minggu postpartum. Data kuantitatif dikumpulkan melalui kuesioner dan dianalisis menggunakan SEM-PLS untuk menguji pengaruh antar variabel dan mendapatkan nilai prediktif model.
Hasil: Instrumen Paternal BSE dan sumber self-efficacy ayah valid dan reliabel. Faktor signifikan yang berpengaruh terhadap penghentian ASI eksklusif meliputi paternal BSE, maternal BSE, pengalaman ayah, pengalaman ibu, persuasi verbal bagi ibu, kondisi fisik emosi ibu, usia ibu, frekuensi ANC, dan tipe keluarga. Secara tidak langsung, ada pengaruh pengalaman ayah terhadap penghentian ASI eksklusif melalui paternal dan maternal BSE.
Simpulan: Model paternal-maternal BSE dapat memprediksi penghentian ASI eksklusif dengan baik.
Saran: Penghentian ASI eksklusif dapat dicegah dengan melibatkan ayah secara aktif melalui edukasi, pendampingan, dan penguatan paternal BSE sejak masa kehamilan.

Background: The rate of exclusive breastfeeding in Indonesia remains low. Both paternal and maternal self-efficacy play a crucial role in supporting breastfeeding practices. This study aims to develop a predictive model for exclusive breastfeeding cessation during the first 0–12 months postpartum using the Paternal and Maternal Breastfeeding Self-Efficacy (BSE) framework.
Methods: This study employed both qualitative and quantitative methods. The qualitative phase involved focus group discussions (FGDs) with five purposively selected fathers to explore aspects of paternal BSE. The quantitative phase used a longitudinal design to evaluate the proposed model. A total of 201 couples who delivered at a maternal and child hospital in South Tangerang City were followed up to 12 weeks postpartum. Quantitative data were collected using questionnaires and analyzed with Partial Least Squares Structural Equation Modeling (PLS-SEM) to examine relationships between variables and assess model predictive.
Results: The Paternal BSE instrument and sources of paternal self-efficacy were found to be valid and reliable. Significant factors influencing exclusive breastfeeding cessation included paternal BSE, maternal BSE, fathers’ experience, mothers’ experience, verbal persuasion for mothers, mothers’ physical and emotional condition, maternal age, antenatal care (ANC) frequency, and family type. Indirectly, there was a relationship between fathers’ experience and exclusive breastfeeding cessation through paternal and maternal BSE.
Conclusion: The paternal-maternal BSE model effectively predicts exclusive breastfeeding cessation.
Recommendation: Exclusive breastfeeding cessation can be prevented by actively involving fathers through education, support, and strengthening paternal BSE starting from pregnancy.

ABSTRAK
Latar Belakang: Kanker payudara merupakan jenis kanker dengan insidensi tertinggi di Indonesia, dengan sebagian besar pasien terdiagnosis pada stadium lanjut salah satunya karena keterlambatan diagnosis di fasilitas kesehatan primer. Rendahnya kemampuan dokter umum dalam menegakkan diagnosis klinis kanker payudara serta alur rujukan yang panjang juga berkontribusi memperberat keterlambatan tersebut. Tujuan: Menghasilkan model prediksi diagnosis klinis kanker payudara di pelayanan kesehatan primer menggunakan skor malignansi “Probability of Breast Cancer (BOBAN)” Metode: Studi ini menggunakan mixed method, dengan pendekatan desain explanatory sequential, penelitian kuantitatif menggunakan desain potong lintang dilanjutkan penelitian kualitatif dengan desain studi kasus. Penelitian ini melibatkan 1.169 wanita usia ≥30 tahun yang melakukan deteksi dini di RS Kanker Dharmais (2020–2022). Variabel prediktor dianalisis dengan uji multivariat regresi logistik untuk penyusunan model skoring. Tahap kedua berupa uji akurasi (sensitivitas-spesifisitas) dan nilai probabilitas prediksi. Tahap ketiga berupa uji kualitatif melalui diskusi kelompok terfokus (FGD) dengan dokter umum di puskesmas. Hasil: Model prediksi terdiri dari tujuh variabel terpilih, yaitu usia, riwayat keluarga tingkat I, riwayat melahirkan, riwayat menyusui, benjolan payudara, kelenjar getah bening aksila, dan gejala lanjut kanker. Model ini memiliki nilai kalibrasi yang baik (p-value 0.826) dan nilai AUC pada ROC sebesar 0,920 (CI 95% 0,892 - 0,947; p-value 0,00) menunjukkan diskriminasi yang sangat baik. Total skor antara 0–177, dengan titik potong optimal pada skor 69 (sensitivitas 86,7%, spesifisitas 82,9%, dan nilai probabilitas 10,45%). Skor rendah (0-68) didiagnosis bukan kanker payudara dan skor tinggi (69-177) didiagnosis curiga kanker payudara. Evaluasi kualitatif menunjukkan bahwa skor malignansi BOBAN dapat diaplikasikan oleh dokter umum di fasilitas kesehatan pelayanan primer. Kesimpulan: Skor malignansi ini dapat memprediksi diagnosis klinis dan menghitung nilai probabilitas kanker payudara. Skor malignansi BOBAN direkomendasikan untuk digunakan sebagai instrumen deteksi dini kanker payudara di faskes primer dan dapat menjadi solusi bagi dokter umum untuk mempemudah skrining rujukan tatalaksana kanker payudara di Indonesia.

ABSTRACT
Background: Breast cancer is the most prevalent type of cancer in Indonesia, with the majority of patients diagnosed at an advanced stage, partly due to delayed diagnosis in primary healthcare settings. Limited diagnostic capabilities among general practitioners and lengthy referral processes contribute significantly to these delays. Objective: To develop a clinical prediction model for breast cancer diagnosis in primary healthcare using the "Probability of Breast Cancer (BOBAN)" malignancy score. Methods: This study employed a mixed-method approach, consisting of a quantitative cross-sectional study followed by a qualitative explanatory sequential design. A total of 1,169 women aged ≥30 years who underwent early detection at Dharmais Cancer Hospital (2020–2022) were included. Predictor variables were analyzed using multivariate logistic regression to construct the scoring model. The second phase involved evaluating the model’s diagnostic accuracy (sensitivity-specificity) and predictive probability values. The third phase included qualitative assessment through focus group discussions (FGDs) with general practitioners at community health centers (puskesmas). Results: The prediction model comprised seven selected variables: age, first-degree family history of breast cancer, childbirth history, breastfeeding history, presence of a breast lump, axillary lymph nodes, and advanced cancer symptoms. The model demonstrated good calibration (p-value = 0.826) and excellent discrimination with an AUC of 0.920 (95% CI: 0.892–0.947; p-value < 0.001). The total score ranged from 0–177, with an optimal cutoff score of 69 (sensitivity 86.7%, specificity 82.9%, predictive probability 10.45%). A low score (0–68) indicated a non-breast cancer diagnosis, while a high score (69–177) indicated suspected breast cancer. Qualitative evaluation indicated that the BOBAN malignancy score is feasible for implementation by general practitioners in primary care settings. Conclusion: The malignancy score is capable of predicting clinical diagnosis and estimating the probability of breast cancer. The BOBAN score is recommended as a screening tool for early detection in primary healthcare facilities and offers a practical solution for general practitioners to facilitate breast cancer management referrals in Indonesia.