Colorectal cancer (CRC) is an emerging public health concern in Ethiopia and across sub-Saharan Africa, driven by urbanization, changes in diet, physical inactivity, and improvements in life expectancy. In 2020, Ethiopia recorded an estimated 3,121 new CRC cases, making it the country's third most common cancer. Two-thirds of Ethiopian patients are diagnosed at stage III or IV, which drastically limits survival prospects.
Time to treatment initiation (TTI) refers to the interval between a confirmed cancer diagnosis and the start of treatment. A TTI of more than 60 days is considered late by international standards. Timely treatment initiation is a cornerstone of optimal cancer care, and delays have been linked to poorer survival outcomes in multiple cancer types including colorectal cancer.
Despite the known impact of treatment delays, there has been little systematic evidence on the extent and causes of late TTI specifically for colorectal cancer patients in Ethiopia. This study aimed to fill that gap by assessing the prevalence and determinants of delayed treatment initiation across three oncology centers in the Amhara Region.
This was a hospital-based cross-sectional study conducted from March 30 to April 20, 2024, at three oncology centers in the Amhara Region of Ethiopia: Felege Hiwot Referral Hospital, University of Gondar Specialized Hospital, and Dessie Specialized Hospital. These are among the few facilities offering specialized cancer care in the region.
Data were extracted from the medical records of colorectal cancer patients diagnosed and registered between July 1, 2018, and June 30, 2023. Patients were included if they had a confirmed CRC diagnosis and a documented treatment initiation date. Those with incomplete records or missing key dates were excluded. The final analysis included 464 patients with 96% data completeness.
The dependent variable was late TTI, defined as treatment initiation more than 60 days after the first confirmed cancer diagnosis. Potential predictors included demographic factors (age, sex, residence, education, insurance, distance from the treatment center), pathological factors (tumor location, stage, histological type, CEA level), and clinical factors (performance status, comorbidities, presenting symptoms).
Data were coded and entered into EpiData version 4.2, then exported to STATA version 14 for analysis. Both bivariable and multivariable logistic regression were performed. Variables that showed an association with delayed TTI at a significance threshold of p less than 0.25 in the bivariable analysis were carried forward into the multivariable model.
The multivariable analysis controlled for all included variables simultaneously, producing adjusted odds ratios (AOR) with 95% confidence intervals. Only associations with p less than 0.05 in the adjusted model were considered statistically significant. The Hosmer-Lemeshow goodness-of-fit test yielded a p-value of 0.2167, indicating that the final model fit the data well.
Study bias was minimized through standardized data abstraction protocols and training of data collectors. The sample size of 484 (slightly larger than the 464 included) was calculated using STATA 14 based on anticipated prevalence estimates for the key determinants of treatment delay. Ethical approval was granted by the Institutional Review Board of Woldia University, and all data were de-identified prior to analysis.
The 464 included patients had a mean age of 48 years. A slight majority were male (58.6%), and more than half lived in urban areas (54.7%). Most patients were married (80.4%) and identified as Ethiopian Orthodox Christian (74.8%). Regarding education, 43.1% were unable to read or write.
Nearly 43% of patients had no health insurance coverage. Adenocarcinoma was the dominant histological type at 79.5%, with mucinous or signet-ring cell carcinoma accounting for the remaining 20.5%. More than 62% of patients had carcinoembryonic antigen (CEA) levels above 5 ng/mL, suggesting high tumor burden at presentation. Approximately 62% of patients were diagnosed at late stage (III or IV).
Regarding treatment, the most common approach was chemotherapy alone, received by 66.4% of patients, followed by surgery combined with chemotherapy (20.5%) and surgery alone (10.8%). Among those receiving chemotherapy, FOLFOX was the most frequently used regimen (26.3%), followed by CAPOX (22.2%) and FOLFIRI (14.2%). Notably, 30.4% of patients did not receive chemotherapy at all.
Of the 464 colorectal cancer patients, 260 patients (56.0%; 95% CI: 51.4 to 60.0%) initiated treatment more than 60 days after diagnosis. The median time to treatment initiation was 117 days (95% CI: 95 to 150 days), far longer than reported in Poland (38 days), Italy (28 days), and the United States (26 days).
The strongest independent predictor of delayed treatment initiation was geographic distance. Patients living more than 81 km from the treatment center had 3.54 times higher odds of delay compared to those living closer (AOR = 3.54, 95% CI: 2.23 to 5.60). Patients without health insurance had 83% higher odds of delay (AOR = 1.83, 95% CI: 1.20 to 2.78).
Two clinical factors also independently predicted delay. Patients with good WHO performance status were 2.55 times more likely to initiate treatment late (AOR = 2.55, 95% CI: 1.53 to 4.24), and patients with non-metastatic disease had 67% higher odds of delay compared to those with metastatic disease (AOR = 1.67, 95% CI: 1.08 to 2.58). Two factors were associated with reduced odds of delay: rural residence (AOR = 0.56) and having more than two presenting symptoms (AOR = 0.53).
Geographic distance was the dominant predictor of delayed treatment initiation, consistent with findings from other studies in Ethiopia and low- and middle-income countries (LMICs). Patients living far from oncology centers face compounding barriers: transportation costs, prolonged travel time, and the logistical burden of making multiple trips for diagnostic workups, consultations, and treatment initiation. In rural Ethiopia, these challenges are amplified by poor road infrastructure and limited transport options.
Lack of health insurance independently doubled the odds of delay, consistent with international evidence linking absence of financial coverage with delayed diagnosis, suboptimal care, and poorer survival outcomes. Cancer care in Ethiopia requires significant out-of-pocket spending, which can cause patients to defer treatment until their condition worsens or financial support is obtained from family members or community networks.
Recommended policy responses include decentralization of oncology services to zonal and district hospitals closer to where patients live, provision of transport subsidies for those traveling long distances, expansion of health insurance to cover oncology services, and streamlined clinical pathways that cluster diagnostic procedures into fewer visits. These structural changes would address the most impactful and actionable barriers identified in this study.
Two findings in this study run counter to intuition. First, patients with good WHO performance status were more likely to experience treatment delays. This paradox may reflect informal triage practices in which patients with poor performance status are perceived as more urgent and fast-tracked for treatment, while those who appear healthier may face longer waits. Patient-driven delays may also contribute, as individuals feeling relatively well may be less motivated to rapidly navigate a complex healthcare system.
Second, patients with non-metastatic disease had higher odds of delay than those with metastatic cancer. This likely reflects the clinical management pathway: patients with metastatic disease are often initiated on systemic therapy quickly given the urgency of their condition, while non-metastatic patients may undergo additional staging procedures, surgical consultations, and pre-treatment optimization steps that extend the time before treatment begins.
These findings highlight the importance of establishing maximum acceptable wait times for all patients regardless of apparent clinical urgency. Patient navigation programs that actively guide patients through the diagnostic and pre-treatment process could help reduce delays for those who might otherwise fall through the cracks of the healthcare system. The study also underscores the need for prospective longitudinal designs to better establish causal pathways to delay.