Oral nutritional supplement compliance among hospitalized cancer patients: a cross-sectional study based on the KAP model.

We conducted a cross-sectional study at Taizhou Hospital in Zhejiang Province, China, from May 1, 2024, to August 31, 2024, involving cancer patients in the nonsurgical treatment phase. Convenience sampling was used to recruit study participants. Well-trained nurses provided guidance on the use of ONS, explained the rationale for the prescription, and emphasized the importance of proper intake. This study was approved by the Hospital Ethics Committee KL20240437), and all participants provided informed consent and participated voluntarily. The inclusion criteria were as follows: (1) patients diagnosed with tumors through pathological examination (including head and neck, digestive system, respiratory system, urinary system, and hematological system tumors) who were in the nonsurgical treatment phase (receiving radiotherapy, chemotherapy, or palliative care); (2) patients aged ≥ 18 years; (3) patients who scored 3 or higher on the Nutritional Risk Screening (NRS 2002) and had a Patient-Generated Subjective Global Assessment (PG-SGA) score of 4 or higher were prescribed ONS by oncologists; (4) patients on ONS powder during hospitalization. The exclusion criteria were as follows: (1) patients with a history of mental illness, mental disorders, severe hearing impairment, or current psychological problems; (2) individuals allergic to milk or protein; (3) patients who were required to fast due to conditions such as gastrointestinal obstruction, intestinal paralysis, severe short bowel syndrome, intestinal ischemia, or massive bleeding; (4) patients with severe nausea and vomiting exceeding the Common Terminology Criteria for Adverse Events version 4.0 (CTCAE v4.0) level III; (5) patients who did not provide consent to participate in the study. In this study, four scales were employed, including two three-dimensional scales, one five-dimensional scale, and one one-dimensional scale. Additionally, 11 items of general demographic data were incorporated, as shown in Fig.  1 . Sample size calculation using G*Power 3.1 software indicated that a minimum of 115 participants would be necessary for a multiple linear regression model with 14 predictors, assuming α  = 0.05, power = 0.85, and effect size f 2  = 0.2. The study included 127 participants, surpassing the required sample size for adequate statistical power. Fig. 1 Framework of independent and outcome variables Framework of independent and outcome variables Patient information was generally collected from the hospital’s medical record system. We conducted a cross-sectional, population-based online survey using the WeChat-incorporated Wen-Juan-Xing platform (Changsha Ranxing Information Technology Co., Ltd., Hunan, China), which is widely used in China. The detailed study design has been described previously [ 35 ].To ensure consistency throughout the survey process and maintain high data quality, all research staff underwent standardized training before data collection. The training program covered questionnaire administration, sampling methods, standardized interviewing techniques, data entry protocols, and procedures for resolving common issues. The 4-h training included both a theoretical examination and practical simulations (Table S1 ). Only staff members who successfully passed all assessments were permitted to participate in data collection. Participants voluntarily completed the questionnaire by scanning a QR code with their mobile phones. For those unable to complete the survey independently, such as certain patients or elderly individuals unfamiliar with smartphones, trained research staff assisted in conducting interviews and recording their responses. To ensure that the data was not duplicated, each WeChat account or IP address was only allowed to complete the questionnaire once. All questions were mandatory; incomplete questionnaires could not be submitted, effectively preventing duplicate or invalid entries. Upon completion of data collection, the research team exported the raw data from the Wenjuanxing platform and conducted systematic data cleaning. Through logical screening and eliminating questionnaires with suspected untrue answers and those with excessively short response times, two researchers independently reviewed all data cleaning procedures to ensure the accuracy and reliability of subsequent data analyses. Collectively, these steps guaranteed the integrity and quality of the final dataset. The KAP assessment scale was developed by Liu Manli in China [ 36 ]. All items were scored on a Likert-type scale and divided into three dimensions comprising 33 questions: knowledge (13 items), attitude (7 items), and practice (13 items). These were rated on a 5-point Likert scale (1 = lowest, 5 = highest). Higher scores indicated higher levels in each dimension. The results of the reliability test showed that the Cronbach’s alpha coefficients for KAP were 0.80, 0.78, and 0.66, respectively. This demonstrates acceptable internal consistency and reliability. Permission to use all the questionnaires was obtained from the authors. The General Self-Efficacy Scale (GSES) was used in this study to evaluate the self-efficacy of the research subjects. Developed by Schwarzer, this scale reliably measures general self-efficacy [ 37 ]. The Chinese version of the scale includes 10 items [ 38 ]. The items are scored on a 4-point Likert-type scale, with scores ranging from 10 to 40 points. Higher scores indicate a higher level of patient self-efficacy. Reliability tests revealed a Cronbach’s alpha coefficient of 0.96. A Perceived Social Support Scale (PSSS) was developed by Zimet [ 39 ]. This study used the Chinese version of the PSSS [ 40 ]. The scale comprises three dimensions and 12 items: family support, which assesses the degree of subjective support from family (items 3, 4, 8, and 11); friend support, which evaluates the subjective feeling of support from friends (items 6, 7, 9, and 12); and other support, which measures the degree of subjective support from individuals other than family and friends (items 1, 2, 5, and 10). A 7-point scoring system was used (1 point = "completely inconsistent,” 7 points = “completely consistent”). A higher total score indicates a greater level of perceived support from others in society. Specifically, scores were categorized as follows: 12–36 points indicate a low level of perceived support, 37–60 points indicate an intermediate level, and 61–84 points indicate a high level of perceived support. The scale has a Cronbach’s alpha coefficient of 0.93. The EuroQol Five- Dimensional Scale (EQ-5D) was developed by the European Quality of Life Group [ 41 ]. The Chinese version of this scale incorporates the EQ-5D health description system [ 42 ]. This study exclusively utilized the objectively quantified EQ-5D health description system to assess patients’ health-related quality of life (HRQoL). The EQ-5D includes five dimensions: mobility, self-care ability, daily activity ability, pain/discomfort, and anxiety/depression. Each dimension is rated on three levels: no difficulty, some difficulty, and extreme difficulty. The EQ-5D health utility value was calculated by subtracting the weighted values associated with each dimension from 1, followed by subtracting a constant term of 0.04. If any dimension was rated as level 3, an additional deduction of 0.022 was applied to the total score. The health effect value ranged from − 0.149 to 1, with higher scores indicating better quality of life. The Chinese version of the EQ-5D has demonstrated strong reliability and validity, with a Cronbach’s alpha coefficient of 0.88. Data analysis was conducted using SPSS version 25.0 (IBM Corporation, Armonk, USA). Categorical variables and sociodemographic characteristics were presented as frequencies and percentages. Shapiro–Wilk tests were performed to determine if continuous variables were normally distributed. Normally distributed continuous variables are reported as mean ± standard deviation and non-normally distributed continuous variables are reported as median with interquartile ranges. The two groups were compared using an independent samples t -test (continuous variables with normal distribution and homogeneity of variance) or Mann–Whitney U test (continuous variables with a non-normal distribution or heterogeneity of variance). A one-way ANOVA and Kruskal–Wallis test were used to compare the differences among the three groups. Pearson’s correlation analysis was used to investigate correlations between ONS compliance, KAP, social support, quality of life, and self-efficacy. A correlation coefficient ( r ) ≥ 0.7 was considered a significant correlation, 0.5–0.7 a strong correlation, 0.3–0.5 a moderate correlation, and r  ≤ 0.3 a weak correlation [ 43 ]. Multivariable linear regression was implemented, and the estimated model coefficients with 95% confidence intervals were reported. Model diagnostics included reporting of the coefficient of determination ( R 2 ), checks for multicollinearity (variance inflation factor, VIF > 10), and assessment of residuals for normality and homoscedasticity. Statistical significance was set at P  < 0.05.

The workflow used in the present study is illustrated in Fig.  2 . After applying the inclusion criteria, 132 eligible participants were selected from the initial cohort of 1605 participants for inclusion in the study. Including other questionnaires with short response times and logical errors, five invalid questionnaires were excluded. Our final sample, hence, comprised 127 valid cases, indicating an overall response rate of 96.21%. In this study, 79 (62.2%) were over 65 years old. All participants (100%) used powdered nutritional supplements. Among these patients, 89.76% resided in rural areas, and 66.14% were male. Chemotherapy accounted for 22.83%, radiotherapy accounted for 22.83%, and palliative care accounted for 54.34%. Detailed demographic information is presented in Table  1 . Fig. 2 Study sample selection process Table 1 Demographic characteristics of patients ( n  = 127) Variables Frequency Percentage (%) Age (year)  < 65 48 37.80  ≥ 65 79 62.20 Gender Male 84 66.14 Female 43 33.86 Education Illiterate 45 35.43 Primary school or above 82 64.57 Treatment stage Chemotherapy 29 22.83 Radiotherapy 29 22.83 Palliative care 69 54.34 Residence Rural 114 89.76 Non-rural 13 10.24 Healthcare payment Rural insurance 115 90.55 Urban insurance 12 9.45 Adverse reaction Yes 22 17.32 No 105 82.68 Family companionship during hospitalization Yes 107 84.25 No 20 15.75 Satisfaction Dissatisfied 67 52.76 Satisfied 60 47.24 Financial burden Yes 17 13.39 No 110 86.61 Attitude toward ONS Not helpful 26 20.47 Helpful 101 79.53 Total 127 100.0 Study sample selection process Demographic characteristics of patients ( n  = 127) The average score for the knowledge dimension was 38.61 ± 7.21 points, while the average score for the attitude dimension was 25.36 ± 3.48 points. The average score for the practice dimension was 38.39 ± 6.22 points, as detailed in Table 2 . Notably, within the knowledge dimension, the items “ONS preparation temperature” (3.78 ± 1.19) and “Preparation method” (3.76 ± 0.94) received higher scores. Patients showed higher scores on the items “single dose” (3.49 ± 1.14) and “total daily dose” (3.44 ± 1.17), but lower scores were observed in items related to managing complications and the importance of the dining environment (1.89 ± 0.74). For the attitude dimension, the total possible score was 35 points. The highest score recorded was 3.92 ± 0.60 points, while the lowest score was 3.29 ± 0.97 points. In the practice dimension, the highest score was obtained for the item “Stop taking supplements due to financial reasons” (4.49 ± 0.88), while the lowest score was observed for the item “I use a smartphone or note card to remind me to take supplements” (1.24 ± 0.61) (Table S2 ). Independent sample t -tests were used to analyze general data. The results indicated that age, educational background, financial burden, satisfaction, residence, and attitudes towards nutritional powder were significantly associated with compliance-related knowledge ( P  < 0.05). Furthermore, place of residence, family companionship, satisfaction, and payment method were significantly associated with attitudes toward compliance ( P  < 0.05). Additionally, family accompaniment during hospitalization, place of residence, satisfaction, and adverse reactions were significantly associated with compliance-related practices ( P  < 0.05) (Table  2 ). Table 2 Relationships between participants’ demographic characteristics and KAP scores ( n  = 127) Variable Knowledge Attitude Practice Mean ± SD or median (IQR) t / z / F / H P Mean ± SD or median (IQR) t / z / F / H P • Mean ± SD or median (IQR) t / z / F / H P Total 38.61 ± 7.21 25.36 ± 3.48 38.39 ± 6.22 Age (year)  < 65 42.00(37.3,44.8) 2.65 0.01** 25.00(22.0,27.0) 1.58 0.11 38.50(33.0,41.8) 1.22 0.22  ≥ 65 36.00(31.0,43.0) 25.00(24.0,28.0) 40.00(35.0,43.0) Gender Male 39.61 ± 6.94 1.45 0.15 26.00(24.0,28.0) 1.62 0.10 39.50(35.0,43.0) 1.12 0.26 Female 37.33 ± 7.65 25.00(22.0,27.0) 39.00(32.0,42.0) Education Illiterate 36.49 ± 7.41 2.51 0.01* 25.00(23.5,27.0) 0.58 0.56 39.00(34.5,43.0) 0.63 0.53 Primary school or above 39.78 ± 6.87 25.50(23.0,28.0) 39.00(34.0,42.0) Residence Rural 40.00(34.0,44.0) 2.33 0.02* 25.00(24.0,28.0) 2.52 0.01* 40.00(35.0,42.0) 2.42 0.02* Non-rural 33.00(28.0,38.0) 23.00(20.5,26.0) 35.00(29.5,37.5) Family companionship during Hospitalization Yes 39.04 ± 7.38 1.54 0.12 26.00(24.0,28.0) 2.70 0.01** 40.00(35.0,43.0) 2.02 0.04* No 36.35 ± 5.91 23.50(22.0,25.0) 38.50(33.3,39.8) Financial burden Yes 34.12 ± 7.94 2.84 0.01** 25.00(22.5,27.5) 0.21 0.83 38.00(31.0,41.0) 0.96 0.33 No 39.31 ± 6.87 25.00(23.0,27.0) 39.50(35.0,42.0) Satisfaction Dissatisfied 36.58 ± 6.72 3.50 0.01** 25.00(22.0,27.0) 2.62 0.01** 38.0(34.0,41.0) 2.03 0.04* Satisfied 40.88 ± 7.12 26.00(24.0,28.0) 40.0(35.0,44.0) Healthcare payment method Rural insurance 39.00(34.0,44.0) 0.03 0.97 25.00(23.0,27.0) 2.11 0.03* 39.0(35.0,42.0) 1.09 0.27 Urban insurance 38.00(33.8,44.5) 23.00(20.0,26.5) 38.0(33.0,40.8) Adverse reaction Yes 36.68 ± 6.14 1.39 0.17 25.00(22.0,27.3) 0.91 0.36 35.50(32.5,38.5) 2.84 0.01** No 39.02 ± 7.38 25.00(23.0,27.0) 40.00(36.0,43.0) Attitude towards ONS Not helpful 35.15 ± 6.92 2.81 0.01* 25.00(23.0,27.0) 0.70 0.49 37.50(31.5,41.3) 1.57 0.12 Helpful 39.50 ± 7.05 25.00(23.0,28.0) 39.00(35.0,42.5) Treatment stage Chemotherapy 39.34 ± 6.01 2.01 0.14 26.00(23.0,27.0) 0.05 0.98 38.62 ± 4.44 1.05 0.35 Radiotherapy 36.28 ± 8.87 25.00(24.0,27.0) 36.93 ± 6.49 Palliative care 39.29 ± 6.79 25.00(23.0,28.0) 38.90 ± 6.71 Data are presented as mean ± SD (compared using t -test) for normally distributed variables, or as median (IQR) [compared using Mann–Whitney U test] for skewed variables. For comparisons among more than two groups, one-way ANOVA was used for normally distributed data ( F -test), and the Kruskal–Wallis test for non-normally distributed data. Abbreviations: ONS , oral nutritional supplements. * P  < 0.05; ** P  < 0.01 Relationships between participants’ demographic characteristics and KAP scores ( n  = 127) Data are presented as mean ± SD (compared using t -test) for normally distributed variables, or as median (IQR) [compared using Mann–Whitney U test] for skewed variables. For comparisons among more than two groups, one-way ANOVA was used for normally distributed data ( F -test), and the Kruskal–Wallis test for non-normally distributed data. Abbreviations: ONS , oral nutritional supplements. * P  < 0.05; ** P  < 0.01 Pearson’s correlation analysis revealed that the practice scores of participants were positively correlated with their knowledge and attitude scores ( r  = 0.46, r  = 0.37, respectively, P  < 0.01). The results also indicated that the nutritional compliance KAP practice scores were positively correlated with the PSSS, EQ-5D, and GSES scores. They were weakly correlated with EQ-5D and GSES ( r  = 0.18 for both, P  = 0.04) and moderately correlated with PSSS ( r  = 0.33, P  < 0.01) (Table  3 ). Table 3 Analysis of ONS compliance KAP and correlation factors in hospitalized patients with cancer ( n  = 127) Knowledge Attitude Practice PSSS GSES EQ-5D Knowledge 1 Attitude 0.49** 1 Practice 0.46** 0.37** 1 PSSS 0.29** 0.32** 0.33** 1 GSES 0.13 0.30** 0.18* 0.39** 1 EQ-5D 0.18* 0.13 0.18* 0.05 0.57** 1 PSSS , Perceived Social Support Scale; GSES , General Self-Efficacy Scale; EQ-5D, the EuroQol  Five-Dimensional Scale * P  < 0.05, ** P  < 0.01 Analysis of ONS compliance KAP and correlation factors in hospitalized patients with cancer ( n  = 127) PSSS , Perceived Social Support Scale; GSES , General Self-Efficacy Scale; EQ-5D, the EuroQol  Five-Dimensional Scale * P  < 0.05, ** P  < 0.01 Multiple linear regression analysis was conducted to assess compliance as a function of KAP, using independent variables that demonstrated statistically significant differences in the univariate and correlational analyses. The numerical values assigned to the independent variables are listed in Table S3 . The results of the multiple linear regression models are summarized in Table  4 . The findings indicated that educational level, satisfaction, financial burden, social support, and attitudes toward ONS were independent factors associated with compliance-related knowledge. Specifically, the analysis revealed significant differences between satisfied and unsatisfied individuals ( β  = 3.00; P  = 0.01), between individuals with primary education or higher and those with less than primary education ( β  = 3.26; P  = 0.01), and between those experiencing financial burdens and those  not experiencing financial burdens ( β  = -3.55; P  = 0.04). Furthermore, attitudes towards ONS (helpful vs. not helpful) were significant ( β  = 3.09; P  = 0.03), as was PSSS score ( β  = 0.10; P  = 0.02). EQ-5D score was included in the model but was not a significant factor ( β  = 2.07, P  = 0.37). Table 4 Multivariate linear regression models for oral nutritional supplement compliance KAP ( n  = 127) Variable Knowledge Attitude Practice β SE 95%CI P VIF β SE 95%CI P VIF β SE 95%CI P VIF PSSS 0.10 0.04 0.02 ~ 0.19 0.02* 1.09 0.02 0.02  − 0.02 ~ 0.07 0.32 1.39 0.11 0.04 0.02 ~ 0.19 0.01** 1.44 GSES –- –- –- –- 0.10 0.04 0.02 ~ 0.17 0.02* 1.33  − 0.06 0.10  − 0.25 ~ 0.13 0.53 2.12 EQ-5D 2.07 2.32 −2.52 ~ 6.67 0.37 1.21 –- –- –- –- 3.25 2.51  − 1.74 ~ 8.24 0.20 1.88 Knowledge –- –- –- –- 0.18 0.04 0.11 ~ 0.26 0.01** 1.24 0.24 0.08 0.08 ~ 0.40 0.01** 1.51 Attitude –- –- –- –- –- –- –- –- 0.25 0.17  − 0.09 ~ 0.58 0.16 1.60 Satisfaction Satisfied vs. dissatisfied 3.00 1.14 0.74 ~ 5.26 0.01** 1.09 0.75 0.54  − 0.32 ~ 1.81 0.17 1.15  − 0.09 1.02  − 2.13 ~ 1.94 0.93 1.17 Residence Rural vs. non-rural 3.38 2.04  − 0.67 ~ 7.42 0.10 1.29 0.36 0.96  − 1.53 ~ 2.25 0.70 1.32 0.70 1.71  − 2.69 ~ 4.10 0.68 1.22 Age  ≥ 65 vs. < 65  − 1.71 1.28  − 4.25 ~ 0.81 0.18 1.29 –- –- –- –- –- –- –- –- –- Education Primary school or above vs. illiterate 3.26 1.30 0.70 ~ 5.83 0.01** 1.29 –- –- –- –- 0.61 –- –- –- –- Financial burden Yes vs. no  − 3.55 1.68  − 6.87 ~ − 0.23 0.04* 1.10 –- –- –- –- –- –- –- –- –- Healthcare payment method Rural vs. urban insurance –- –- –– –- –- 1.55 1.01  − 0.45 ~ 3.55 0.13 1.39 –- –- –- –- –- Family accompaniment during hospitalization 0.61 1.41 −2.18 ~ 3.39 0.43 1.16 Yes vs. no –- –- –– –- –- 1.66 0.75 0.18 ~ 3.14 0.03* 1.18 –- –- –- –- –- Adverse reaction No vs. yes –- –- –– –- –- –- –- –- –- –- 2.85 1.28 0.30 ~ 5.40 0.03* –- Attitude towards ONS Helpful vs. unhelpful 3.09 1.42 0.28 ~ 5.90 0.03* 1.11 –- –- –- –- –- –- –- –- –- –- R 2 0.32 0.38 0.31 PSSS , Perceived Social Support Scale; GSES , General Self-Efficacy Scale; EQ-5D , the EuroQol Five-Dimensional Scale * P  < 0.05; ** P  < 0.01 Multivariate linear regression models for oral nutritional supplement compliance KAP ( n  = 127) PSSS , Perceived Social Support Scale; GSES , General Self-Efficacy Scale; EQ-5D , the EuroQol Five-Dimensional Scale * P  < 0.05; ** P  < 0.01 Self-efficacy, knowledge, and family accompaniment during hospitalization were significant factors associated with attitudes towards compliance with ONS, with GSES ( β  = 0.10; P  = 0.02) and knowledge ( β  = 0.18; P  = 0.01) scores being particularly noteworthy. Additionally, the presence of family versus non-family companionship during hospitalization was associated with attitudes towards compliance with ONS ( β  = 1.66; P  = 0.03). Adverse reactions, knowledge, and social support were also identified as key factors were associated with compliance with ONS, with significant differences observed between individuals who did not experience adverse reactions and those who did ( β  = 2.85; P  = 0.03), as well as between PSSS scores ( β  = 0.11; P  = 0.01) and knowledge scores ( β  = 0.24; P  = 0.01). EQ-5D score was also entered into the final model but did not reach statistical significance ( β  = 3.25, P  = 0.20).

Our study demonstrated that adverse effects, compliance-related knowledge, and social support are significantly associated with ONS compliance practices. These findings have important clinical implications and offer guidance for the development of future intervention programs aimed at improving the KAP related to ONS compliance among this specific sub-population.

This study employed the KAP framework to deepen the understanding of the behavioral and psychosocial factors associated with ONS compliance among hospitalized cancer patients. The study is both timely and relevant, addressing an important clinical and public health issue: nutritional compliance in cancer patients. Unlike previous studies, this study included patients receiving palliative care, chemotherapy, or radiotherapy. It not only examined compliance with ONS practices but also comprehensively assessed compliance across the KAP dimensions. The findings indicated that educational level, patient satisfaction, family financial burden, social support, and attitudes towards ONS were significantly associated with knowledge related to compliance. Furthermore, self-efficacy, nutritional knowledge, and social support were linked to attitudes towards compliance, while adverse effects, nutritional knowledge, and social support were significantly associated with compliance-related practices. Social support was measured as the combination of received support (i.e., the quantity and quality of assistance from family, friends, and significant others) and perceived support (i.e., an individual’s subjective assessment of the availability and adequacy of these social resources) [ 39 , 44 ]. While a previous study on breast cancer established a connection between social support and general nutritional knowledge, this study specifically demonstrated an association between social and nutritional support for patients with cancer using a validated ONS scale [ 45 ]. The survey findings reveal a significant association between higher education levels among cancer patients and their superior performance in nutrition knowledge assessments, compared to those with lower educational attainment. This highlights the crucial role of education not only in enhancing patients’ understanding of nutrition but also in facilitating access to nutrition information and promoting the appropriate use of ONS. These observations are consistent with previous research, which indicates that individuals with limited education face greater difficulties in comprehending and obtaining nutrition-related information, particularly regarding the relationship between nutrition and health [ 46 ]. A significant positive correlation was observed between patient satisfaction and the knowledge dimension of compliance with ONS among hospitalized cancer patients. This finding is consistent with the previous results which emphasized that patient satisfaction serves as an important mediator between service quality and medication compliance, highlighting the importance of satisfaction as a leverage point for enhancing knowledge and promoting subsequent compliance practices [ 47 ]. Low socioeconomic status not only increases treatment non-compliance due to cost-related issues but also hinders patients’ ability to acquire disease-related knowledge. This phenomenon is evident in chronic conditions such as kidney disease, cancer, and other long-term illnesses [ 48 , 49 ]. This study highlights that patients with higher self-efficacy demonstrate more positive attitudes toward ONS compliance. This result is consistent with previous study [ 23 ].This finding suggests that patients with higher self-efficacy are more likely to have a positive attitude toward compliance with ONS, possibly due to greater confidence in their abilities and a greater willingness to take proactive steps in managing their health. This study found that patients with family support exhibited a more positive attitude towards compliance with ONS. This could be attributed to the emotional and practical support provided by family members, which enables patients to better understand and comply with the use of ONS [ 50 ]. The survey also revealed that cancer patients’ attitudes towards ONS compliance were closely related to their level of knowledge. Patients with a deeper understanding of the role, importance, and correct use of ONS were more likely to develop positive attitudes. This finding is consistent with a study on patients with endometriosis, which demonstrated that an increase in knowledge significantly improved patients’ attitudes towards treatment [ 51 ]. Further research suggests that social support was positively associated with cancer patients’ compliance with ONS, consistent with the findings of a previous study on postoperative gastrointestinal surgery [ 23 ].In addition, this study also showed that the level of knowledge about ONS among hospitalized patients with cancer was positively correlated with their compliance. This result is consistent with findings from a study on discharged patients [ 52 ], which suggested that the degree of patients’ understanding of ONS is closely related to their compliance with medical recommendations, whether during hospitalization or after discharge. This consistency further validates the fundamental role of knowledge in shaping patient behavior. In addition, ONS may elicit certain adverse reactions associated with the preparation method, including nausea, bloating, and diarrhea [ 10 , 53 ]. In this study, adverse reactions were significantly associated with patient compliance. These adverse effects may be associated with patient compliance with ONS, potentially leading patients to reduce their intake or discontinue use entirely [ 13 ]. The findings of this study indicate that knowledge significantly impacts both attitudes and practices. However, the mechanisms underlying the pathways of behavioral formation remain to be further explored. Future studies should aim to validate the mediating role of attitudes between knowledge and practice, as well as investigate the intrinsic mechanisms of practice formation alongside external determinants [ 10 ]. The EQ-5D is a widely used preference-based generic HRQoL instrument [ 41 ]. One study reported a positive correlation between improved medication compliance in hypertensive patients and enhanced HRQoL [ 54 ]. Another study found that poor compliance was associated with increased pain and anxiety levels in diabetic patients [ 55 ]. However, our study did not find a statistically significant relationship between HRQoL and the KAP dimensions of ONS. The possible reasons of this discrepancy may due to different study areas, different assessment tools, and study duration. Future research should focus on specific EQ-5D dimensions to further accurately elucidate the impact of compliance. Previous research suggested that improving compliance with ONS in hospitalized patients with cancer requires multidimensional interventions: enhancing patient knowledge through personalized education [ 52 ] boosting self-efficacy, encouraging family involvement, and promoting active communication to improve patient attitudes. Furthermore, optimizing preparation methods for ONS and providing social support are essential to improving practice behaviors. In clinical practice, patient compliance should be assessed comprehensively across three dimensions: knowledge, attitudes, and practices. Strengthening interdisciplinary collaboration [ 56 ], conducting regular follow-ups, and adjusting intervention plans accordingly are key strategies. Future research should investigate the mediating role of attitudes [ 10 ], explore the underlying mechanisms of behavioral formation, develop personalized intervention strategies, and assess long-term outcomes to enhance patient compliance, rehabilitation, and overall quality of life. This study has several strengths. First, the research addresses a relevant and timely topic, given the high prevalence of malnutrition among patients with cancer, and focuses on ONS compliance, which is both clinically and socially significant. Second, the application of the KAP model as the theoretical framework is appropriate, as it allows for a structured investigation of behavioral factors. Third, the study used multiple validated instruments (e.g., GSES, PSSS, and EQ-5D) and considered a broad range of demographic and clinical variables, enhancing the comprehensiveness of the data collected. Several limitations should be considered when interpreting the results of this survey. First, the study subjects were selected through convenience sampling, where participants were voluntarily recruited based on recommendations from cancer patients. This approach may introduce selection bias and trigger the Hawthorne effect, as individuals who choose to participate do so willingly in the selected hospital. Voluntary bias occurs when a study sample comprises only those participants who are inclined to participate in the study, particularly those who find the topic intriguing and are more likely to volunteer, similar to those anticipating favorable outcomes [ 57 ]. Second, all data were self-reported; some individuals may have been reluctant to disclose their true KAP, and social expectations bias may have led to over reporting. Third, because data were collected through a WeChat mini-program, our sample inevitably omitted rural and low-educated seniors who were digitally excluded. Future studies should adopt a mixed mode (paper-and-pencil plus face-to-face interviews) to mitigate this bias. The internal consistency of the practice dimension in the KAP scale was suboptimal (Cronbach’s α  = 0.66), falling below the conventional psychometric threshold of 0.70. Low internal consistency can result in unreliable scores and may bias study findings due to increased measurement error [ 58 ]. This suggests a restricted reliability that could potentially compromise the precision of evaluating participants’ compliance behavior [ 59 ]. Factors contributing to this inadequacy may stem from the restricted project sample size, ambiguities in definitions, and the intricate nature of compliance. Consequently, findings about the practice dimension warrant careful interpretation. Subsequent research endeavors should contemplate broadening project inclusivity, enhancing conceptual clarity, and integrating objective compliance metrics to bolster reliability. Finally, the cross-sectional design of this study limits the ability to establish causal relationships between KAP. Longitudinal studies are needed to explore the causal pathways and temporal dynamics of these factors, thereby providing a deeper understanding of how they are associated with each other over time.

Nearly 20 million new cancer cases occurred worldwide in 2022, and one in five individuals will develop cancer during their lifetime [ 1 ]. Cancer and the associated treatments cause considerable changes in patient’s nutritional status [ 2 , 3 ]. Nutrition is one of critical factors in cancer, with 40–80% of patients experiencing malnutrition during the disease [ 4 ]. Recent studies reveal that the prevalence of malnutrition among hospitalized cancer patients can be as high as 84.7% [ 5 ], mainly due to the tumor itself, treatment-related side effects, reduced food intake, and increased metabolic demands [ 6 ]. This condition is associated with an average extension of hospital stay by 5.20 days, an increase in healthcare costs by $15,722 [ 7 ], a 35% higher mortality risk [ 8 ], and a 42% increased likelihood of complications within 1 year after chemotherapy [ 8 ]. Severe malnutrition also correlates with poorer functional status, greater fatigue, and substantially reduced quality of life compared to patients with adequate or mild to moderate nutrition [ 9 ]. Therefore, malnutrition screening among hospitalized cancer patients is crucial for improving treatment outcomes and optimizing healthcare resource utilization. Oral nutritional supplements (ONS) are dietary foods consumed in addition to normal food to address nutritional needs [ 10 ]. ONS are the preferred nutritional therapy for malnourished or nutritionally at-risk patients, provided no contraindications to enteral nutrition exist [ 11 ]. Previous study showed that ONS improve nutritional status and clinical outcomes of patients [ 12 ]. They positively affect skeletal muscle mass, help maintain or increase weight, alleviate fatigue, improve tolerance to radiotherapy and chemotherapy, address malnutrition-related issues, and enhance overall quality of life [ 13 – 15 ]. However, effective nutritional improvement requires sustained, adequate, and regular ONS intake [ 16 ]. In addition, a compliance with ONS is defined as the ratio of actual intake to the recommended intake [ 12 ]. Despite the recognized importance of ONS, patient compliance remains suboptimal and often results in intake levels lower than the recommended doses [ 15 , 17 ]. Low compliance can reduce the effects of, or even render, nutritional interventions ineffective, leading to high medical costs and decreased quality of life [ 18 , 19 ]. Therefore, understanding the factors that are associated with patient compliance is critical for optimizing nutritional intervention strategies. Previous study has demonstrated a wide variation in compliance with ONS among cancer patients, with reported rates ranging from 24.7% to 98.0% [ 12 ]. The factors contributing to this substantial variability in compliance are multifaceted and encompass various dimensions. Prior research has identified various factors affecting cancer patients’ compliance with ONS, such as health beliefs and social support [ 20 ], economic status [ 21 ], and knowledge and attitudes towards ONS [ 10 , 22 ]. Strong social support and high self-efficacy also promote ONS compliance, which in turn improves patients’ nutritional status and overall quality of life [ 10 , 11 , 23 , 24 ]. Compliance is associated with multiple factors, including those related to the formulation, environmental factors, and educational level [ 17 , 23 ]. Existing studies on ONS compliance have primarily focused on investigating one or more factors [ 17 , 20 , 23 , 25 ]. Few studies have explored ONS compliance among patients undergoing radiotherapy, chemotherapy, or palliative care. Tumor types and treatment stage are crucial factors significantly associated with patients’ compliance with ONS [ 26 , 27 ]. These factors may vary greatly across diverse cultural backgrounds and healthcare settings [ 13 , 18 ]. Understanding these variables is critical for improving compliance among hospitalized cancer patients. The knowledge, attitudes, and practices (KAP) model is a cornerstone in public health research, providing a structured framework for comprehending health behaviors [ 28 ]. This model facilitates the development of impactful interventions and the assessment of their efficacy [ 29 ].This model emphasizes the importance of education and attitude changes in improving practices, which can promote more positive health behaviors by improving patient knowledge and being associated with patients’ attitudes [ 30 , 31 ]. In the context of managing nutrition for cancer patients, the KAP model is instrumental in identifying obstacles to effective dietary regimens, pinpointing educational deficiencies, and guiding the development of tailored interventions to enhance well-being. For instance, investigations on patients with gastrointestinal cancer revealed a spectrum of nutritional knowledge among participants, yet predominantly negative attitudes and limited engagement in beneficial dietary practices [ 32 , 33 ]. Moreover, a study on dietary supplementation in Saudi Arabia demonstrated that heightened knowledge levels and positive health attitudes facilitated the adoption of supplements [ 34 ]. These findings highlight the importance of targeted educational initiatives to address knowledge gaps and promote compliance to nutritional recommendations. Therefore, applying the KAP model to ONS compliance research among cancer patients is both timely and necessary. In this study, we utilized a KAP-based questionnaire to analyze factors associated with ONS compliance in hospitalized cancer patients, aiming to provide evidence that supports the development of targeted interventions to improve patient compliance and treatment outcomes.

Below is the link to the electronic supplementary material. ESM 1 (DOCX 25.8 KB) (DOCX 25.8 KB)