A cross-sectional, quantitative survey-based design was used to evaluate the knowledge and attitudes about breast cancer risk factors among female patients.

The study was conducted at Allied Hospital, Faisalabad—one of the largest tertiary care hospitals in Punjab, Pakistan. Data collection was carried out within the oncology ward between March and August 2024.

The target population was comprised of female patients receiving care in the oncology department of Allied Hospital, Faisalabad. A convenience sampling technique was employed to recruit 150 female participants. This sample size was chosen to allow for potential incomplete or non-responses while still ensuring a statistically valid dataset for analysis. Patients were identified through hospital records, and consent was sought via their provided contact details. The study population consisted specifically of female oncology patients who were either receiving active treatment or attending follow-up consultations at the oncology ward of Allied Hospital, Faisalabad. These participants were selected to ensure that respondents had some level of interaction with breast cancer-related care and were more likely to have encountered relevant information through social media or healthcare professionals. Inclusion criteria required participants to be adult females willing to provide informed consent. Anonymity and the right to withdraw from the study at any time were assured. While convenience sampling may limit generalizability, it was deemed appropriate given time and resource constraints.

Sample size justification: The sample size of 150 participants was determined based on practical constraints including time, resource availability, and accessibility to the target population within the oncology ward of Allied Hospital, Faisalabad. Although a formal power calculation was not performed due to the exploratory nature of the study and the use of a convenience sampling method, the chosen sample size was considered adequate to observe meaningful patterns in social media usage, knowledge levels, and attitudes toward breast cancer risk factors. The sample size also accounted for potential non-responses or incomplete questionnaires, ensuring a sufficient number of complete cases for valid statistical analysis.

A structured questionnaire was used to collect data, which included multiple sections: 1) socio-demographic characteristics, 2) sources of breast cancer information, 3) knowledge of risk factors, 4) awareness and practice of breast self-examination (BSE), 5) signs and symptoms, 6) screening methods, and 7) perceptions regarding treatment. The questionnaire was pre-tested through a pilot study involving 15 participants, whose responses were not included in the final dataset. To ensure the reliability and clarity of the questionnaire, a pilot study was conducted involving 15 female participants from the same hospital setting who met the inclusion criteria. Their responses were not included in the final analysis. Based on the pilot data, the internal consistency of the questionnaire was assessed using Cronbach’s alpha, which yielded a value of 0.81, indicating good reliability. No modifications were necessary, as the items were found to be well-understood and appropriately structured. This preliminary validation helped confirm that the instrument was suitable for evaluating knowledge, attitudes, and social media usage related to breast cancer risk factors in the target population. Based on the pilot, no modifications were required. Correct answers were assigned a score of “1,” while incorrect or “don’t know” responses were given a “0.” Participants’ knowledge levels were categorized as Weak (25% - 49%), Average (50% - 74%), and Good (75% and above) based on their total scores.

The study included female patients aged 18 years or older who were receiving care—either active treatment or follow-up consultations—in the oncology department of Allied Hospital, Faisalabad. Participants were required to provide informed consent and possess the ability to understand and complete the questionnaire independently. Only those who reported using at least one social media platform in the past six months were eligible to participate, as the study specifically aimed to assess the impact of digital media on breast cancer awareness.

Patients were excluded if they had any cognitive impairments or communication difficulties that could interfere with their ability to respond accurately. Additionally, critically ill or hospitalized patients unable to participate due to their medical condition, and those with no history of social media usage, were excluded to ensure consistency and relevance in data collection. These criteria were implemented to ensure that participants were representative of the target population for this study (i.e., female oncology patients with social media exposure), and to strengthen the interpretability of findings related to digital health communication.

Due to practical limitations, convenience sampling was selected as the most feasible approach. The researcher requested a list of female oncology patients along with their contact information to facilitate recruitment. Survey forms were distributed in person, and informed consent was obtained before participation. Participants were assured of confidentiality, anonymity, and voluntary involvement, with the option to withdraw at any stage.

Data was analyzed using Minitab Version 18. Descriptive statistics (frequencies and percentages) were used to summarize socio-demographic characteristics such as age, marital status, education, income, and religion. Cross-tabulation analyses were conducted to examine relationships between independent variables (e.g., socio-demographics) and dependent variables (e.g., knowledge, screening practices, perceived barriers). Each correct response was given one point, and total scores were calculated for each participant. Knowledge and practice levels were then categorized accordingly. This approach allowed identification of key factors influencing breast cancer awareness and screening intentions.

Ethical approval for this study was obtained from the Institutional Review Board of the respective institution. Informed consent was obtained from all participants prior to data collection, and all ethical protocols regarding participant rights, confidentiality, and voluntary participation were strictly followed.

Table 1 presents detailed demographic characteristics of the respondents, including age distribution, marital status, area of residence, socio-economic class, and education level. All respondents were female, as the study focused on breast cancer awareness among women. The age range of participants was between 18 and 60 years, with the highest concentration in the 40 - 50 years age group. The mean and median ages were approximately 42 and 45 years, respectively. A significant majority were married (85.4%), while others were unmarried (12.6%) or divorced/separated (0.7%).

In terms of residence, 56.4% lived in urban areas, while 43.6% were from rural backgrounds. Most respondents belonged to the middle socio-economic class (70.2%), followed by lower (21.9%) and upper (6.6%) classes. Education levels varied; the majority had completed high school or had some college education, while fewer held a college or postgraduate degree.

Table 2 outlines the social media usage patterns and primary sources of health-related information among the respondents. Facebook and YouTube were the most used platforms, while Instagram, Twitter, and TikTok saw lower engagement. A majority of respondents reported that they did not use social media

regularly, and most spent between 0 - 2 hours on these platforms daily.

When asked about information sources regarding breast cancer, doctors were the most trusted and frequently used (72.5%), followed by social media, healthcare websites, and family or friends.

Respondents were asked about their knowledge of various risk factors for breast cancer. As presented in Table 3 , a large number of participants (79.2%) were not knowledgeable about age as a risk factor. Knowledge regarding family history, genetics, lifestyle, and dense breast tissue was moderate, with most respondents describing themselves as “not knowledgeable”.

Table 4 highlights respondents’ perceptions of social media’s role in raising breast cancer awareness. While many participants had encountered breast cancer-related content on social media, 65.8% viewed this information as not reliable, with only 3.4% considering it very reliable. Despite skepticism about credibility, most reported a slight increase in awareness due to exposure.

Furthermore, many respondents indicated that social media positively influenced their likelihood of undergoing breast cancer screening. However, healthcare professionals remained the most trusted source of breast cancer-related information.

The chi-square tests were conducted to examine the association between the type of social media platform used and knowledge about various breast cancer risk factors. The results varied across the factors analyzed, as shown in Table 5 .

The chi-square test showed a significant association between social media usage and knowledge of age-related risk factors for breast cancer. The Pearson

chi-square value was 4.642 with a p-value = 0.031, indicating that social media users are more likely to be aware of age as a risk factor compared to non-users. No significant association was found between social media use and awareness of family history as a breast cancer risk factor. The Pearson chi-square value was 3.231 with a p-value = 0.072. A significant association was found between social media platform and knowledge of genetic risk factors. The Pearson chi-square was 6.001, p-value = 0.014. The chi-square value was 6.594 (p = 0.010), indicating a statistically significant association. Social media users demonstrated greater awareness of lifestyle factors such as diet and exercise affecting breast cancer risk. A statistically significant relationship was also found for awareness of dense breast tissue as a risk factor, with chi-square = 5.899, p = 0.015. There was a strong association between social media platform and whether the respondent had seen information related to breast cancer. Chi-square = 7.673, p = 0.006. Significant differences were observed in perceived reliability across platforms. Chi-square = 11.909, p = 0.003. No significant association was found between platform used and perceived impact on knowledge. Chi-square = 0.651, p = 0.420. No significant relationship was observed. Chi-square = 0.302, p = 0.583. No significant difference in trust across platforms. Chi-square = 3.762, p = 0.152.

Multiple linear regression analysis was conducted to identify significant predictors of breast cancer knowledge and attitudes toward screening. Independent variables included education level, socio-economic status, residential area, family history, and perceived social media reliability. The models explained between 4.33% and 13.21% of the variance in knowledge-related outcomes (R²), with adjusted R² values ranging from 0.64% to 7.63%, indicating modest model fit. Significant predictors varied across outcomes; for example, education level significantly predicted knowledge about age-related and genetic risk factors (p < 0.05) with 95% confidence interval level, while socio-economic status influenced awareness of lifestyle and breast density-related risks. Social media reliability and information sources significantly impacted overall knowledge improvement and screening attitudes (p = 0.036 and p = 0.028, respectively). All regression assumptions were checked and met, and the analysis was performed using Minitab version.

The overall regression models were not statistically significant for most dependent variables, except for Knowledge: Dense Breast Tissue (p = 0.016). This suggests that the independent variables, as a group, did not significantly predict knowledge of breast cancer risk factors in most cases. For Knowledge: Dense Breast Tissue, the F-value was 2.37 (p = 0.016), indicating that the model was statistically significant.

Age: Age was not a significant predictor of knowledge for any of the dependent variables (p-values ranged from 0.191 to 0.843). Marital Status: Marital status did not significantly influence knowledge of breast cancer risk factors (p-values ranged from 0.356 to 0.921). Residential Area: Residential area was a significant predictor only for Knowledge: Dense Breast Tissue (p = 0.018). Participants from urban areas had slightly lower knowledge compared to those from rural areas. Socio-Economic Status: Socio-economic status was a significant predictor for Knowledge: Lifestyle Factors (p = 0.013) and Knowledge: Dense Breast Tissue (p = 0.008). Participants from lower socio-economic backgrounds had lower knowledge compared to those from upper socio-economic backgrounds. Education Level: Education level was a significant predictor for Knowledge: Age (p = 0.016), Knowledge: Family History (p = 0.011), and Knowledge: Genetics (p = 0.025). Participants with higher education levels (college degree or postgraduate degree) had significantly higher knowledge compared to those with only a high school diploma. Several observations had large residuals (marked as “R”), indicating that the model did not fit these data points well. Observation 111 was flagged as an unusual X-value (marked as “X”), suggesting it may be an outlier or influential point.

The multiple linear regression analysis was performed to assess the impact of various independent variables on the dependent variable, “Impact on Knowledge”. The results indicate that the overall regression model is statistically significant (F = 1.76, p = 0.036), suggesting that the included independent variables collectively explain a significant proportion of the variance in knowledge impact. However, the adjusted R-squared value is 9.65%, indicating that only a small percentage of the variance in the dependent variable is accounted for by the model.

Among the independent variables, the reliability of social media information (p = 0.025) and sources of information (p = 0.038) were found to have a statistically significant impact on knowledge. The coefficient for the reliability of social media information (category 2) was 0.363, indicating a positive association with knowledge impact. Similarly, sources of information category 5 had a coefficient of 0.750, suggesting that respondents relying on certain sources experienced a greater impact on knowledge. Other variables, including age, daily hours spent on social media, marital status, residential area, socio-economic status, and education level, did not show statistically significant effects (p > 0.05).

The model summary statistics reveal that the standard error of the regression (S = 0.7503) is relatively high, indicating some degree of dispersion in the residuals. The lack-of-fit test (p = 0.586) suggests that the model does not suffer from a significant misfit. However, certain observations exhibited large residuals, as indicated in the diagnostics for unusual observations. These residuals highlight the presence of individual responses that deviated considerably from the model’s predictions.

The multiple linear regression analysis was conducted to examine the impact of various predictors on the “Impact on Screening Attitude”. The model included 19 predictors, and the results indicate that the overall regression model is statistically significant (F-value = 1.82, p-value = 0.028), as shown in Table 6 . However, the adjusted R-squared value is 10.33%, suggesting that only 10.33% of the variance in the dependent variable is explained by the predictors in the model. This indicates that while the model is significant, it has limited explanatory power. The ANOVA in 6 shows that the regression model is significant (p-value = 0.028), but most individual predictors are not statistically significant. The only significant predictor is “Sources of Information” (p-value = 0.003), which has a strong impact on the screening attitude. Specifically, the coefficients for “Sources of Information” YouTube and others are significant (p-values = 0.002 and 0.004, respectively), indicating that these sources have a positive and significant influence on screening attitudes. Other predictors, such as age, daily hours on social media, marital status, residential area, socio-economic status, education level, social media platforms, and reliability of social media information, do not show. The coefficients table in Table 6 provides insights into the direction and magnitude

of the relationship between each predictor and the dependent variable. For example: Sources of Information” YouTube has a coefficient of 0.557, indicating a strong positive impact on screening attitude. Sources of Information” Others has a coefficient of 0.378, also showing a positive but slightly weaker impact. Other predictors, such as age and daily hours on social media, have very small coefficients and are not statistically significant. The model summary shows that the standard error of the estimate (S) is 0.288816, indicating the average distance that the observed values fall from the regression line. The R-squared value of 22.95% suggests that the model explains 22.95% of the variance in the dependent variable, but the adjusted R-squared value of 10.33% indicates that the model’s explanatory power is limited after adjusting for the number of predictors.