This study is a cross-sectional study. The convenience sampling method was used, and 316 patients with chronic diseases hospitalized in the geriatrics department of a tertiary hospital in Deyang City were selected from January to June 2025 as the survey subjects. The inclusion criteria for patients were as follows: (1) clinically diagnosed with at least one chronic disease; (2) age over 60 years; (3) informed consent and voluntary participation in this survey. The exclusion criteria were: (1) consciousness disorder or intellectual disability, resulting in inability to communicate normally; (2) patients in the acute stage of disease, not suitable for the survey. Sample size calculation: According to Kendall’s criterion, the sample size for regression analysis should be 10–20 times the number of independent variables. This study included 15 independent variables, resulting in a sample size of 150–300 (17). Considering a 20% non-response rate, the minimum sample size N = 15 × 10 + [(15 × 10) × 20%] = 180. A total of 330 questionnaires were collected in this study. After screening, 14 questionnaires with inconsistent responses or obvious logical errors (such as selecting the same option for all items) were completely excluded, yielding 316 valid questionnaires for final analysis. The effective questionnaire recovery rate was 95.8%. The study was approved by the Ethics Committee (No. 2022-04-010-K01).

Latent profile analysis (LPA) was conducted in Mplus 8.3. Model fit indices included: AIC, BIC, aBIC (lower = better), Entropy (0–1; higher = better classification,an Entropy value ≥ 0.80 suggests approximately 90% accuracy in class assignment), LMR-LRT and BLRT (p < 0.05 favored k-class over (k-1)-class models) (21). Analyses used SPSS 23.0. Mean ± SD (normally distributed) Median (P25, P75) (non-normal) n (%) (categorical). Group comparisons: Kruskal-Wallis (continuous), χ²/Fisher’s exact (categorical). Predictors of medication literacy profiles were analyzed via ordinal logistic regression. A p-value of < 0.05 indicated statistical significance.

Among the 316 included patients, 176 (55.7%) were male and 140 (44.3%) were female, with a mean age of 74.94 years (SD = 8.89). A total of 77.8% were married, and 74.1% had more than one chronic disease. Among the 14 excluded patients, 9 (64.3%) were male and 5 (35.7%) were female, with a mean age of 74.07 years (SD = 5.01). A total of 64.3% were married, and 64.3% had more than one chronic disease. Sensitivity analysis was conducted by comparing the demographic and clinical characteristics between the included participants (n = 316) and the excluded participants (n = 14). The results showed no significant differences between the two groups in terms of gender, age, education level, or number of chronic diseases (all p > 0.05), indicating that the exclusion of cases did not introduce significant bias. We further conducted a best-worst case scenario analysis to test the impact of extreme assumptions regarding the missing data. Under the worst-case scenario (assuming all 14 excluded participants belonged to the lowest medication literacy profile), and the best-case scenario (assuming all belonged to the highest profile), the results of the ordinal logistic regression model were re-examined. The significance and direction of the associations for the key predictors (age, education, monthly income, and self-efficacy) remained unchanged in both scenarios, confirming the robustness of our primary findings even under highly conservative assumptions. Additional characteristics are presented in Table 1.

The results of this study showed that the total score of Medication Literacy and Disease Self-Efficacy of Older adult patients with Chronic Diseases was 70.50 (50.00, 89.00) and the total score of Disease Self-Efficacy was 47.00 (38.00, 52.00), where the scores of each of the dimensions are presented in Table 2.

In this study, item scores of medication literacy among older adult chronic disease patients were used as manifest indicators. Initial models with 1 to 5 latent profiles were fitted, and the fit indices for each model are presented in Table 3. As the number of latent profiles increased, the AIC, BIC, and aBIC values demonstrated monotonic decreases. The entropy values for all five-class solutions exceeded 0.800, and the Bootstrap Likelihood Ratio Test (BLRT) results were statistically significant across all latent profiles. However, the Lo-Mendell-Rubin adjusted likelihood ratio test (LMR) for the five-class model was not statistically significant (p > 0.05), suggesting that adding more classes might alter model stability. Based on comprehensive evaluation, the four-class model was identified as the optimal solution, consequently retaining four distinct profiles (C1, C2, C3, C4).

To evaluate the accuracy of the latent profile classification, we examined the posterior classification metrics of the model. As shown in Table 4, the average posterior probabilities for the four latent profiles were 0.988, 0.990, 0.976, and 0.971, all well above the acceptable threshold of 0.90. Furthermore, the diagonal values of the classification probability matrix were all greater than 0.90, while the off-diagonal values were consistently low (all < 0.05). In addition, the average variance extracted (AVE) exceeded 0.6, and the composite reliability (CR) surpassed 0.9, collectively indicating high within-class homogeneity. These results demonstrate that the model’s classification of individuals is highly accurate and reliable, with a very low probability of misclassification.

Based on the conceptual model of medication literacy proposed by Pantuzza et al. (7), the four identified types of medication literacy among older adult patients were named according to their performance across measurement dimensions (see Figure 1): C1: Comprehensive-Deficit Type (n = 53, 16.8%): This profile demonstrated significantly lower scores across all dimensions, with particularly notable deficiencies in functional literacy (e.g., information acquisition and comprehension) and critical literacy (decision-making abilities). C2: Communication-Advantage Type (n = 91, 28.8%): This group showed low to moderate overall performance but exhibited relative strength in communicative literacy (e.g., patient-provider interaction and medication communication), while functional and critical literacy remained low. C3: Balanced-Development Type (n = 94, 29.7%): This profile displayed well-balanced capabilities across all dimensions, reflecting comprehensive literacy skills with an overall intermediate level of proficiency. C4: Knowledge-Proficient Type (n = 78, 24.7%): This group achieved the highest scores across all dimensions, with particularly outstanding performance in critical literacy (e.g., application of disease knowledge and decision-making), coupled with high functional and interactive literacy.

The results of univariate analysis showed that the differences in the distribution of medication literacy profiles among older adult patients with chronic diseases residing in different places, marital status, number of children, education, monthly income, type of medical insurance, age, Whether regular checkups are received, and self-efficacy in managing their diseases were statistically significant (p < 0.05). Details are provided in Table 5.

Using variables that showed statistical significance in the univariate analysis as independent variables (Place of residence: rural = 1, urban = 2; Marital status: married = 1, divorced or widowed = 2; Number of children: ≤1 = 1, 2 = 2, ≥3 = 3; Education level: primary school and below = 1, junior high school = 2, high school and above = 3; Personal monthly income (yuan): ≤1,000 = 1, 1,001–3,000 = 2, 3,001–5,000 = 3, ≥5,001 = 4; Health insurance type: employee insurance = 1, resident insurance = 2; Age (years): 60–64 = 1, 65–74 = 2, 75–89 = 3, ≥90 = 4; Regular follow-up: yes = 1, no = 2; Disease self-efficacy was entered as the actual score), and the latent profiles of medication literacy in older adult patients with chronic diseases as the dependent variable (Comprehensive Deficiency Type = 1, Communication Strength Type = 2, Balanced Development Type = 3, Knowledge Proficiency Type = 4), an ordered multinomial logistic regression analysis was performed using the Comprehensive Deficiency Type as the reference group. The mean scores of each dimension across the four profiles showed a sequentially increasing trend, and the test of parallel lines (χ² = 40.552, p = 0.095) indicated that the ordered multinomial logistic regression model could be adopted; the Pearson goodness-of-fit test (χ² = 611.576, p > 0.05) indicated good model fit, and the overall fit test showed that the regression model was statistically significant (p < 0.001). The results revealed that age, education level, personal monthly income, and disease self-efficacy significantly predicted the health literacy of older adult patients, as shown in Table 6.

The results of this study indicate that the total medication literacy score among older adult patients with chronic diseases was 70.50 (50.00, 89.00), with a scoring rate of 61.3%, which is at a moderate level and warrants further improvement. This finding is consistent with the interview results reported by Rahman et al. (22). Regarding patients with chronic conditions. Due to aging, the vision, hearing, and cognitive abilities of older adults decline, making it difficult for them to receive, understand, and remember Medication information. At the same time, incorrect medication attitudes and behaviors have led some older adults to not take standardized medication as prescribed, which further hinders the improvement of their medication literacy (11). The information acquisition ability dimension scored the lowest at 46.0%. This may be because, due to their limited knowledge and educational background, older adults have difficulties accurately interpreting and understanding complex Medication instructions and medical terminology, thus hindering their ability to effectively acquire medication-related information. In addition, older adults have relatively limited access to Medication information channels, mostly relying on guidance from medical staff. Their ability to apply information technology is relatively weak, which limits their effective use of the Internet and other platforms to obtain diversified Medication information (23).

Findings from this study reveal significant heterogeneity in medication literacy among older adult patients with chronic diseases, which can be categorized into four distinct profiles: comprehensive deficiency type (16.8%), communication strength type (28.8%), balanced development type (29.7%), and knowledge proficiency type (24.7%).

The comprehensive deficiency type scored the lowest across all dimensions, accounting for 16.8% of the sample. This profile is predominantly characterized by individuals residing in rural areas, having an elementary school education or below, a monthly income ≤ ¥1,000, and being enrolled in resident health insurance. Most patients in this category live in rural settings with relatively low educational attainment and weak information acquisition skills, making it difficult for them to understand complex medication-related information, thus resulting in lower medication literacy. For these patients, healthcare providers should adopt diversified medication guidance strategies, such as using simple, humorous, and easy-to-understand language to explain complex medication knowledge; organizing drug information into large-print, illustrated medication lists and pill organizers based on patient needs to facilitate understanding for both patients and their families; and fully leveraging the supervisory role of family members through repeated verification and assessment of the patient’s medication knowledge and skills to enhance their overall medication literacy.

The communication strength type had relatively low scores across most dimensions but scored highest in the communicative interaction dimension, comprising 28.8% of the sample. This group is primarily characterized by being married, having an elementary school education or below, having 2–4 chronic diseases, and no history of adverse drug reactions. Although their low educational level makes it challenging to grasp medication knowledge, resulting in overall lower medication literacy, their multiple chronic conditions and high marriage rate enable them to acquire knowledge with family support and discuss medication matters collectively, leading to better performance in communicative interaction. This suggests that healthcare providers should capitalize on these patients’ communication strengths by recommending reliable medication management platforms, teaching them how to search for and discern medication information, and strengthening their mastery of medication knowledge.

The balanced development type showed relatively balanced, moderate scores across all dimensions, accounting for 29.7% of participants. This profile is mainly composed of male urban residents who are married and covered by employee health insurance. These patients generally benefit from relatively better healthcare access and living conditions, which contribute to their medication literacy, though there remains room for further improvement. For this group, targeted health education initiatives—such as regular medication knowledge lectures and consulting sessions—should be implemented to provide personalized medication guidance. Additionally, they should be encouraged to set self-monitoring goals for medication use, integrating knowledge with action to enhance their medication knowledge, attitudes, and behaviors (32).

The knowledge proficiency type scored the highest across all dimensions, constituting 24.7% of the sample. This profile is characterized by urban residence, high school education or above, employee health insurance, and a monthly income ≥ ¥5,001. These patients not only possess relatively high educational levels, enabling them to effectively understand and master medication-related knowledge and information, but also have sufficient income to access high-quality healthcare services and medication resources, allowing them to adhere to medical advice and use medications appropriately. Healthcare providers should affirm and support the high level of medication literacy in these patients, encouraging them to maintain their good medication habits and behaviors. Furthermore, these individuals can be trained as “Medication Health Ambassadors” and encouraged to participate in peer education, sharing their medication management experiences to improve medication literacy among more older adult patients, thereby reducing the incidence of adverse health outcomes.