This multicenter cross-sectional KAP study was conducted on HNC patients who underwent radiotherapy between December 2023 and July 2024 in China and carried out in the radiotherapy and oncology departments of Tianjin Medical University Cancer Institute and Hospital and four other medical institutes—Peking University Cancer Hospital, Tianjin Medical University General Hospital, Tianjin People’s Hospital, and Tianjin Wuqing District People’s Hospital. The study has been approved by the Medical Ethics Committee of Tianjin Medical University Cancer Institute and Hospital (approval number: EK2023229), and informed consent was obtained from the study subjects.

The inclusion criteria were as follows: (1) patients diagnosed with HNC who were scheduled to undergo radiotherapy or had already received radiotherapy; (2) those aged 18 years and above; and (3) those able to understand the purpose of the study and who voluntarily signed informed consent. No specific exclusion criterion was applied in the study.

The design of the questionnaire referred to previous studies on oral nutritional supplements in HNC patients receiving radiotherapy (7, 8) and relevant guidelines, including the Radiation Therapy Guidelines for HNC in China (2021) and the Expert Consensus on Oral Nutritional Supplementation for Radiation Therapy Patients with Oncology (2017) (17, 18). After the completion of the first draft, the opinions of five experts in oncology and nutrition were taken into account to confirm the content validity, and the questionnaire was revised according to the experts’ suggestions. After the completion of the pilot investigation (31 cases), the reliability was tested, and the overall Cronbach’s α coefficient was 0.935. The Cronbach’s α coefficient was 0.971 for the knowledge dimension, 0.916 for the attitude dimension, and 0.762 for the practice dimension. The internal consistency of the questionnaire was therefore deemed acceptable. During the pilot study, participants were also asked about the clarity and relevance of the items. No issues of ambiguity or redundancy were reported, supporting the face validity of the questionnaire.

The final questionnaire was in Chinese, and the content of the questionnaire included four aspects: demographic information, knowledge dimension, attitude dimension, and practice dimension. The demographic data included age, gender, residence, education level, marital status, monthly household income per capita, height, weight, tumor location, time of diagnosis, and treatments. The body mass index (BMI) is determined by dividing weight in kilograms by the square of height in meters. The BMI thresholds were based on the Chinese classification system, which categorizes individuals as underweight (<18.5 kg/m2), normal weight (18.5 to <24 kg/m2), overweight (24 to <28 kg/m2), and obesity (≥28 kg/m2) (19).

The knowledge dimension consisted of 10 items. The score ranged from 0 to 20, with 2 points assigned for a “very familiar” response, 1 point for “heard about,” and 0 points for “unclear.” The attitude dimension consisted of 10 questions measured on a five-point Likert scale, with responses ranging from strongly agree to strongly disagree (5–1), resulting in total scores from 10 to 50. The practice dimension included four questions, with responses scored from 1 to 5, ranging from never to always, for a total score between 4 and 20 points. Questions 5 and 6 in the practice section, which assessed issues encountered during the use of oral nutritional supplements and sources of knowledge about oral nutrition supplementation, were not scored and were analyzed descriptively. The threshold of good knowledge, positive attitude, and active behavior was considered to be ≥70.0% (20).

The questionnaire was distributed to eligible participants in the departments of radiotherapy and oncology and in the outpatient ward. A convenience sampling method was used, and the questionnaire was distributed in both electronic (online) and paper (offline) versions. The electronic questionnaire was answered by scanning the code in the hospital. Dedicated researchers helped to interpret the contents of the questionnaire for the patients if they had questions. Special checks were performed for distribution and retrieval, data entry, and inspection.

The sample size was calculated based on the item-to-response ratio of 1:15 recommended for questionnaire research (21). With 24 KAP items in the questionnaire (excluding demographics and two descriptive questions), the required sample size was 360. Considering a 15% dropout rate, the minimal sample size was 414.

Statistical analysis was performed using SPSS 27.0 (IBM, Armonk, NY, USA) and AMOS 26.0 (IBM, Armonk, NY, USA). The distribution of the scores of each dimension was tested for data normality. If a skewed distribution was observed, the data were expressed as medians, 25% quantiles, and 75% quantiles. The Wilcoxon–Mann–Whitney U-test or the Kruskal–Wallis analysis of variance was used for comparison. Enumeration data for different demographic characteristics and responses to each question were expressed as N (%). Correlation analysis was performed using Spearman’s correlation coefficient. Combined with the KAP theoretical framework, the structural equation model (SEM) was used to verify whether attitude plays a mediating role between knowledge and practice, and the sizes of the indirect and direct effects were calculated and compared. The hypotheses tested were as follows: (H1) knowledge directly affects attitude, (H2) knowledge directly affects practice, and (H3) knowledge indirectly affects practice through attitude. The threshold criteria for the goodness-of-fit indices of the corresponding SEM model were as follows: root mean square error of approximation (RMSEA) < 0.08, standardized root mean squared residual (SRMR) < 0.08, comparative fit index (CFI) > 0.8, and Tucker–Lewis index (TLI) > 0.8. Two-sided p-values of < 0.05 were considered to indicate statistical significance for all analyses.

A total of 454 questionnaires were collected. Two questionnaires from participants under 18 years of age were excluded, along with one questionnaire that contained logical errors regarding the type of medical insurance. Four questionnaires with incomplete responses and 10 questionnaires with implausible BMI values were also excluded. Ultimately, 437 valid questionnaires were included in the analysis, resulting in an effective response rate of 96.26%. The majority of participants were men (70.71%), married (91.30%), urban residents (50.80%), and aged 41–60 (50.11%) years. The most common cancer type was nasopharyngeal carcinoma in 37.76% of participants, followed by oral cancer in 20.14% and laryngeal cancer in 11.90%. Approximately one-third (34.78%) of participants received surgical treatment, 62.02% underwent chemotherapy, 9.61% underwent targeted therapy, and 13.73% underwent immunotherapy. Only 8.92% of participants were underweight (BMI < 18.5). Moreover, only 27.0% of participants reported receiving educational information about oral nutritional supplements in the hospital (Table 1).

Knowledge scores were significantly lower among participants from rural areas (p = 0.016); those with lower education levels (p = 0.003); those with lower income (p = 0.020); those who underwent fewer radiotherapy sessions (p < 0.001); and those who had not received surgical treatment (p = 0.001), radiotherapy (p = 0.003), chemotherapy (p = 0.022), or targeted therapy (p = 0.023). Scores were most notably lower among participants who had never received any educational information about oral nutritional supplements in the hospital (p < 0.001).

Attitude scores were significantly lower among participants with lower education (p = 0.029), lower income (p = 0.003), fewer radiotherapy sessions (p = 0.004), and those who had not received surgical treatment (p = 0.023), radiotherapy (p = 0.017), or chemotherapy (p = 0.036). Scores were also significantly lower in participants who had never received any educational information about oral nutritional supplements in the hospital (p < 0.001). Practice scores were significantly lower among older participants (p = 0.035); those with lower education levels (p = 0.011), lower income (p = 0.008), or laryngeal/pharyngeal cancer (p = 0.003); and those who had undergone fewer radiotherapy sessions (p = 0.010). Scores were also lower among participants who had not received surgical treatment (p < 0.001), radiotherapy (p = 0.003), or chemotherapy (p = 0.035). Additionally, practice scores were significantly lower in participants who never received any educational information about oral nutritional supplements in the hospital (p < 0.001). Detailed information on participants and their KAP scores is shown in Table 1.

The median knowledge score in this study was 4 (0, 10), corresponding to a poor knowledge level. Questions unclear to the majority of respondents included the oral nutritional supplements lowering total medical costs (question K10, unclear to 70.02%, with 22.43% not sure), reducing complications (question K9, unclear to 61.56%, with 29.75% not sure), and the exact way of taking supplements (question K8, unclear to 62.93%, with 27.69% not sure) (Table 2).

The median attitude score was 34 (30, 39), corresponding to an unfavorable attitude. Although 45.54% of participants were willing to try oral nutritional supplements to alleviate the nutritional problems during radiotherapy for HNC (question A5), 45.31% remained neutral on the topic, 7.55% disagreed, and 1.6% strongly disagreed. However, more than half of the participants expressed concern about the high price of oral nutritional supplements, describing it as a heavy economic burden (question A10: 42.79% agreed and 15.33% strongly agreed) (Table 3).

The median practice score was 15 (13, 17), corresponding to proactive behavior. The majority of respondents (84.67%) reported always or often paying attention to adequate nutritional intake (question P2), but approximately 18.53% of participants rarely or never wanted to learn more about the topic (question P4) (Table 4).

Correlation analysis demonstrated that knowledge score was positively correlated with attitude score (r = 0.379, p < 0.001) and practice score (r = 0.395, p < 0.001); moreover, attitude and practice scores were also positively correlated (r = 0.363, p < 0.001) (Table 5).

SEM was used to further explore the influence of knowledge on practice, with model fit indices corresponding to the acceptable model fit (Supplementary Table S1). As demonstrated in Figure 1 and Supplementary Table S2, it was found that knowledge had a direct positive effect on attitude (β = 0.613, p < 0.001) and on practice (β = 0.807, p < 0.001), while attitude had a direct effect on practice (β = 0.614, p < 0.001).