A total of 500 questionnaires were returned, of which 17 were excluded due to invalid or incomplete responses. The final number of valid questionnaires was 483, yielding a response rate of 96.6%. However, a number of questionnaires were excluded from the analyses based on specific criteria: including, 2 questionnaires with abnormal age values, 1 questionnaire with unreasonable response times, 5 questionnaires with unreasonable responses to the questions (smoking, medical insurance type, and left ventricular ejection fraction), and 9 questionnaires with incomplete or incorrectly entered data. Consequently, the final dataset consisted of 483 valid questionnaires. Among them, there were 254 (52.59%) males, while 246 (50.93%) participants fell within the 60–80 age range, 372 (77.02%) from urban areas, with 325 participants (67.29%) being retired. Moreover, the majority of participants had junior high school or lower education, accounting for 216 (44.72%) individuals, 97 (40.58%) with a monthly per capita income > 10,000 yuan, and marital status predominantly consisted of 446 married participants (92.34%). Smoking habits showed 370 (76.6%) never smoked. Finally, left ventricular ejection fraction status revealed 197 (40.79%) with HF with preserved ejection fraction (HFpEF) (Table 1).
Participants demonstrated a mean knowledge score of 10.75 ± 3.44 (on a 0–20 scale), indicating limited understanding of HF. The mean attitude score was 22.93 ± 2.43, reflecting moderately negative perceptions (scale: 8–40). The mean practice score was 32.21 ± 4.34 (scale: 8–40), suggesting a relatively proactive approach to disease management despite knowledge and attitude gaps. Notably, the knowledge score varied from heart failure patients with different age (p = 0.028), education (p = 0.003), and left ventricular ejection fraction status (p < 0.001). Additionally, as for the attitude score, there were difference among them with different monthly per capita income (p = 0.029). Finally, the difference of practice score were found among them with different left ventricular ejection fraction status (p = 0.019) (Table 1).
The distribution of knowledge dimensions revealed that the three questions with the highest number of participants choosing the “Well Understood” option were “Heart failure (HF) is a severe manifestation or advanced stage of various heart diseases,” (K1) with 53.42%, “HF is mainly characterized by difficulties in breathing, fatigue, and fluid retention (pulmonary congestion, systemic congestion, and peripheral edema)” (K2) with 34.16%, and “Diuretics can eliminate sodium and water retention, improve congestion, and relieve breathing difficulties and edema” (K3) with 25.67%. Conversely, the three questions with the highest number of participants choosing the “Uncertain” option were “Avoiding triggers like respiratory infections, overexertion, and emotional fluctuations in daily life is important.” (K7) with 26.5%, “If drug therapy is ineffective and HF worsens repeatedly, comprehensive treatment measures and surgical interventions need to be considered.” (K4) with 24.43%, and “During HF exacerbations, patients should rest in bed; performing passive leg exercises in bed can prevent deep vein thrombosis. After clinical improvement, moderate activity is recommended, avoiding muscle disuse atrophy.” (K8) with 23.81% (Table S1).
Regarding attitudes, what seems to give cause for optimism is a significant majority of the patients (76.19% totally) agreed or strongly agreed that they are willing to actively follow the treatment plan (A3). Similarly, a high percentage (69.36% totally) claimed that they hope to receiving more health education from doctors to assist them in managing the disease (A4). Additionally, an overwhelming (61.9% totally) of the patients realised that the disease is caused by poor lifestyle or unhealthy habits, and they plan to be more attentive and proactively intervene (A5). However, it is worth noting that a considerable portion (50.32% totally) of the patients admitted to being very afraid of exercise and reluctant to engage, even if it may be beneficial (A8). Meanwhile, what’s worrying is that a substantial number of patients acknowledged feelings of depression, anxiety, and unease, with 19.25% strongly agreeing and 56.94% agreeing that they often experience such emotions (A1). Furthermore, 16.77% strongly agreed, and 42.65% agreed that they feel a loss of control and autonomy, helpless and powerless against the disease (A2). These findings illuminate the complex attitudes and emotional states of patients in response to their disease and treatment (Table S1).
In practice dimension, 21.12% of participants consistently adhere to prescribed medications (P3.1), regular follow-up (P2) is a common practice among 43.06% of participants. On the other hand, lifestyle changes, including quitting bad habits (P1), show varying levels of commitment, 44.31% often engage in such changes, 25.67% do so only sometimes, and 14.08% occasionally adopt these practices. Furthermore, medication-related practices (P3.2) show that 39.54% often use methods to ensure timely medication, 29.61% do so sometimes, whereas 15.53% occasionally or never follow these methods for the time control of medication administration (Table S1).
Pearson correlation analysis showed that knowledge was negatively correlated with attitude (r = −0.200, p < 0.001), indicating that greater knowledge was unexpectedly associated with more negative attitudes. Similarly, attitude and practice were negatively correlated (r = −0.156, p = 0.001), suggesting that negative perceptions may hinder proactive behavior. Conversely, knowledge showed a positive correlation with practice (r = 0.198, p < 0.001), highlighting the importance of educational interventions to improve self-management behaviors (Table 2).
Multivariate logistic regression identified knowledge (OR = 1.078, 95% CI: 1.02–1.14, p = 0.008) and heart failure with mid-range ejection fraction (HFmrEF) (OR = 1.781, 95% CI: 1.097–2.891, p = 0.020) as significant independent predictors of proactive self-care practices. This suggests that improving patient knowledge—through targeted education—and paying closer attention to patients with HFmrEF may be critical for promoting better disease management behaviors (Table 3).
The SEM analysis showed that education exhibited a positive effect on knowledge (β = 0.366, p = 0.001), while medical insurance showed a notable negative effect on knowledge (β = −0.649, p = 0.010). Furthermore, left ventricular ejection fraction score demonstrated a positive effect on knowledge (β = 0.591, p = 0.003), and marital status had a significant positive effect on knowledge (β = 0.876, p = 0.032). Additionally, kidney disease emerged as a positive effect on practice (β = 0.486, p = 0.023), while attitude was found to have a negative effect on practice (β = −0.192, p = 0.015) (Fig. 1 and Table 4). The SEM model yielded favorable model fit indices, as indicated in Table S2. Specifically, the final model showed a chi-square/df (CMIN/DF) of 1.782, root mean square error of approximation (RMSEA) of 0.040, incremental fit index (IFI) of 0.841, Tucker–Lewis index (TLI) of 0.757, and comparative fit index (CFI) of 0.827, indicating an acceptable model fit. Furthermore, Table 4 presents the total, direct, and indirect effects of influencing variables on each dependent variable (Knowledge, Attitude, and Practice), including corresponding standardized coefficients, 95% confidence intervals, and p-values. These results highlight that education, medical insurance status, left ventricular ejection fraction, marital status, and kidney disease significantly influenced KAP dimensions. Additionally, R² values for the dependent variables were calculated: 0.192 for attitude and 0.224 for practice, indicating the explained variance by the included predictors.