Background: Breast cancer is a leading cancer among women in Ethiopia. It accounts for onethirdrnrnof all newly diagnosed female cancers. Most women with breast cancer in Ethiopia arerndiagnosed with late-stage disease, do not receive high-quality care, and face a poor prognosis.rnLocally relevant information on the extent of delayed diagnosis, reasons for late diagnosis, care,rnand determinants of survival among women with breast cancer is essential to guide clinicalrnpractices and public health policy. However, little is known about the extent and reasons forrnpatient interval (from date of symptom recognition to the first consultation of health carernproviders), diagnosis interval (from consultation to diagnosis), and treatment initiation intervalrn(from diagnosis to treatment initiation). Moreover, evidence on the relationship between patientrndelay (> 90 days)/diagnosis delay (> 30 days) and stage at diagnosis, and its effect on survivalrnamong women with breast cancer in Ethiopia is limited. rnObjectives: To determine the magnitude of delays (patient and diagnosis delay) as well as stagernat diagnosis and its effects on the survival of women with breast cancer in Addis Ababa. rnMethods: The study employed mixed-methods (a cross-sectional study, a qualitative study, and arnprospective cohort study). A cohort of 441 women newly diagnosed with breast cancer betweenrn1rnstrn of January 2017 to 30rnthrn of June 2018 in Addis Ababa were recruited for the quantitative phasernof the study, and followed prospectively for two years. Data were collected at different points inrntime, as a cross-sectional study (Paper I and III) and prospective cohort study design (PaperrnIV) to address the quantitative study objectives.rnDuring recruitment, data on the participants' socio-demographic characteristics, date of firstrnsymptom recognition, and medical consultation after recognizing symptoms were collected usingrna structured interviewer-administered questionnaire. The date of diagnosis was taken from thernpatient's pathology report. One year after the recruitment, medical data, such as stage atrndiagnosis, date of diagnosis, histologic tumor type, date of receipt of treatment, and type wererncaptured using a data extraction tool from the study participantsâ€™ medical charts. Finally, at aboutrntwo years following diagnosis, data related to survival status were obtained using both face-tofacernrninterviews and telephone interviews. Also, medical charts were reviewed to update therntreatment status of the study participants. We have conducted a univariable descriptive analysis torndescribe each variable. Multivariable Poisson regression with a robust variance model was usedrnto determine the factors associated with patient/diagnosis delay and stage at diagnosis. Also, Kaplan-Meier and multivariable Cox regressions were used to determine the overall survival raternand factors contributing to the overall survival of women with breast cancer, respectively. Allrnstatistical tests were assessed for significance at p-value < 0.05.rnThe qualitative study (Paper II) was conducted to explore the patientsâ€™, family membersâ€™, andrnhealth care providersâ€™ perspective on late diagnosis of breast cancer. It was employed amongrnpurposively selected 23 in-depth interviewees. Each of the audio recordings was transcribedrnverbatim, coded, and analyzed using thematic analysis. rnResults: The magnitude of patient (>90 days) and diagnostic delays (>30 days) was 35.7%, 95%rnCI (31.1%, 40.3%) and 69.1%, 95% CI (64.6%, 73.3%), respectively. Patient delay wasrnsignificantly higher among women who used traditional medicine before consultation (adjustedrnprevalence ratio [aPR] =2.13, 95% CI (1.68, 2.71). Diagnosis delay was significantly higherrnamong women whose first consultation was at health centers (aPR=1.19, 95% CI [1.02, 1.39])rnand those visited â‰¥ 4 facilities before confirmation (aPR=1.24, 95% CI [1.10, 1.40]) but lowerrnamong women who recognized progression of symptoms before consultation (aPR=0.73, 95% CIrn(0.60, 0.90). rnThe qualitative study revealed that pre-diagnostic awareness about breast cancer risk, causes,rninitial symptoms, early detection methods, and treatment was low. Disregarding the clinicalrnimportance of the first symptom or seeking care from traditional healers were noted as commonrnpractices among women with breast cancer that contributed to late diagnoses. Also, lack ofrnawareness, and misperception about breast cancer treatment and its outcomes, competingrnpriorities, financial insecurity, fear of diagnosis of cancer, and weak health systems (e.g., delay inrnreferral and long waiting period for consultation) were identified as important causes of laterndiagnosis of women with breast cancer. rnThe median (interquartile range [IQR]) tumor size at diagnosis was 4 (3 to 6) centimeters. Sixtyfourrnrnpercent of the women (95% CI [59.5%, 68.8%]) were diagnosed at advanced-stages (44%rnstage III and 20% stage IV) of their disease. The prevalence of advanced-stage disease wasrnsignificantly higher among women who used traditional medicine before diagnostic confirmationrn(aPR=1.29, 95% CI [1.10, 1.52]), and in those who waited for > 6 months before diagnosisrn(aPR=1.35, 95% CI [1.12, 1.63]). On the contrary, it was lower among women who had everrnpracticed breast self-examination before symptom recognition (aPR=0.77, 95% CI [0.63, 0.96]). The median total interval (symptom recognition to first treatment initiation) was 7 (IQR: 2.7 torn15.7) months. One-fifth of the women started first treatment after one year of first symptom(s)rnrecognition. Adjuvant chemotherapy initiation was delayed (>90 days) in 30% of patients. Onlyrn31.4% (n=137) of the women had received radiotherapy, 64.2% (n=88) of which was adjuvantrnradiation. Adjuvant radiation initiation was delayed (>90 days) in 56.1% of the women. rnThe overall survival rate at year one was 88.3% (95% CI [84.9%, 91.0%]), and 75.2% (95% CIrn[70.7%, 79.0%]) at year two. Women diagnosed at stage I had a two-year survival of 100% inrncontrast to 26.7% at stage IV. The risk of death was significantly higher among women who hadrna symptom interval of >3 months (adjusted hazard ratio [aHR] = 1.87, 95% CI [1.15, 3.03]) andrndiagnosed with advanced-stages (aHR=3.32, 95% CI [1.81, 6.10]) but lower among those whornhad surgical (aHR=0.23, 95% CI [0.15, 0.35]) and hormonal therapy (aHR=0.26, 95% CI [0.17,rn0.40]). rnConclusions: Substantial proportions of women with breast cancer in Addis Ababa havernexperienced patient and diagnostic delays that contribute to the high proportion of advancedstagernbreastrncancer,rnandrnlowrnbreastrncancerrnsurvivalrnrates.rnThernfactorsrnidentifiedrnthatrncontributerntornrndelayedrnrndiagnosis and advanced-stage diagnosis are modifiable. These include poor awarenessrnabout breast cancer, using traditional and spiritual remedies, downplaying the clinical importancernof the first breast cancer symptoms, health care providers' limited provision of clinical breastrnexamination and delayed referral of women for diagnosis with suggestive of breast cancerrnsymptoms, and longer navigation process to get diagnosis. Once diagnosed, significant number ofrnwomen experienced delay to adjuvant chemo-and radiotherapy initiation. rnRecommendations: Breast health awareness campaigns that mitigate misconceptions andrnimprove awareness about breast cancer both in the community and frontline health care providersrnare essential. Interventions to enhance early detection and prompt referral following consultation,rnand decrease waiting time between symptom recognition and breast cancer diagnosis are neededrnto improve early-stage diagnosis and survival rate of women with breast cancer. Also, thernexpansion of cancer diagnostic and treatment centers is necessary to shorten the diagnosis andrntreatment delays.