A risk prediction model for oral frailty in elderly patients with COPD was constructed based on the health ecology model Authors: Xuan Qiao, Shasha Gao, Haixia Zhao, Yiming Lu, Huijun Zhang Published: 2025 ABSTRACT BackgroundElderly patients with chronic obstructive pulmonary disease (COPD) often experience oral health problems, such as dry oral mucosa, tooth loss, and gum disease. Dry oral mucosa makes it difficult for food to slide and chew in the mouth. Tooth loss or gum disease can affect chewing and occlusal functions, causing difficulty in swallowing during meals for such patients and increasing the risk of difficulty in swallowing. Oral frailty is an independent risk factor for dysphagia in elderly patients with COPD. This study aims to explore the influencing factors of oral frailty in this group and construct a nomogram prediction model using R software.MethodsFrom July 2025 to August 2025, this study selected 320 elderly patients with chronic obstructive pulmonary disease (COPD) from a tertiary grade A hospital in Shandong Province, China as the research subjects by using the convenience sampling method. Patients were screened by the Oral Frailty Index − 8 (OFI-8), and those with a score of ≥ 4 were defined as having oral frailty. The research subjects were randomly divided into the modeling group (223 cases) and the validation group (97 cases) in a ratio of 7:3. Univariate and multivariate Logistic regression analyses were conducted on the data of the modeling group using SPSS software to identify the risk factors affecting oral frailty in participants. Based on the results of regression analysis, a risk prediction model was established using R software and visualized through Nomogram. To verify the predictive effect of the model, methods such as ROC curve, Hosmer-Lemeshow (H-L) test, correction curve and decision curve analysis (DCA) were adopted to evaluate the discrimination, calibration and clinical practicability of the model.ResultsThe research results show that the incidence of oral frailty among participants are as high as 92.5%. Analysis shows that the influencing factors of oral frailty mainly include the patient’s nutritional status, the degree of breathing difficulties and the type of chronic disease. In the modeling group and the validation group, the areas under the ROC curve were 0.97 (95% CI: 0.94-1.00) and 0.96 (95% CI: 0.91-1.00), respectively, indicating that the model had a relatively high predictive accuracy. In addition, the calibration curve fitting of the two groups was good (P = 0.994 in the modeling group and P = 0.540 in the verification group). The results of the decision curve analysis also show that this model has high clinical practicability.ConclusionsThe nomogram prediction model constructed in this study can effectively assess the risk of oral frailty in elderly patients with chronic obstructive pulmonary disease and has a good predictive ability. This model helps clinical staff identify oral frailty at an early stage and provides a basis for implementing targeted preventive interventions. Through this model, nursing staff can screen high-risk patients more accurately, thereby formulating personalized intervention measures and improving the quality of life and health management level of patients.Ethical committee approvalThis study was approved by the Medical Ethics Review Committee of Jinzhou Medical University (Approval No. JZMULL2025269) on 17 March 2025.Trial registrationNot applicable.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12903-025-07186-6. Background Chronic obstructive pulmonary disease (COPD) is a respiratory disease characterized by chronic airway inflammatory response and chronic structural destruction of lung tissue. Incomplete reversible airflow limitation is its main pathophysiological feature [1, 2]. According to the 2024 World Health Organization (WHO) data [3], COPD is the fourth leading cause of death in the world, causing 3.5 million deaths in 2021, accounting for about 5% of the total global deaths. There are about 100 million COPD patients in China, among whom the prevalence rate of elderly people over 60 years old is as high as 27%, which has become one of the main diseases causing disability and death in the elderly in China [4]. Oral Frailty [5–7] refers to the change of various oral conditions (number of teeth, oral hygiene, oral function, etc.) with the increase of age, accompanied by the decline of oral health interest, physical and mental reserve capacity, and eating disorders. The overall impact of oral frailty is the deterioration of physical and mental function. Studies have shown [8] that oral frailty is an independent risk factor for dysphagia in elderly COPD patients, similar to the research results of Lu Qian et al. [9]. The reason may be that elderly COPD patients often have oral health problems, such as dry oral mucosa, tooth loss, gum disease, etc. Dry oral mucosa makes the sliding and chewing process of food in the mouth difficult. Tooth loss or gum disease can affect chewing and occlusal functions, causing difficulty in swallowing during meals in elderly COPD patients and increasing the risk of dysphagia. Liang Yuanjun et al. found that the incidence of oral frailty in elderly hospitalized COPD patients with severe conditions was 96.1%, which is a predictor of frailty in elderly COPD patients. At present, there are relatively few studies on oral frailty in elderly COPD patients both at home and abroad, and there is no risk prediction model for oral frailty in COPD patients. The health ecology model holds that individual health is the result of complex interactions between personal traits and environmental factors. This model can be divided into the following levels from the inside out: First comes the individual’s innate traits, followed by psychological behavior and lifestyle. The outer layer includes the influence of family and social networks. The outer layer is the environmental conditions of life and work. The outermost layer is the macro policies and environmental factors. This model constructs a comprehensive analytical framework for the integrated analysis of the formation mechanism of health conditions [10, 11]. Unlike previous studies that only focused on biomedical or behavioral determinants, health ecology models can integrate the impact of environmental, social and policy aspects on oral frailty in chronic obstructive pulmonary disease. Recent post-COVID-19 observations have shown that fungal colonization (such as Candida and mucormycosis) in patients treated with oxygen dependence or steroids is associated with dry mouth, which may increase the risk of oral frailty. Especially COPD patients who have been using steroids for a long time may face a higher risk of oral frailty. To fill this research gap, this study, guided by the theory of health ecology models, conducted multi-dimensional classification and comprehensive summary of the influencing factors of oral frailty in elderly hospitalized COPD patients, and constructed a nomogram prediction model for the risk of oral frailty. This model is designed to assist clinical medical staff in early screening for oral frailty in such patients and timely intervention, thereby improving the quality of life of patients and providing a basis for reducing their adverse health consequences. Although existing studies have gradually revealed the impact of oral frailty on dysphagia and nutritional decline in elderly COPD patients, there is still no validated predictive model for oral frailty to date. Therefore, this study fills this important research gap by constructing an oral frailty risk prediction model based on a health ecology model, providing new tools and theoretical basis for early clinical screening and intervention. Materials and methods Research design This is a cross-sectional study, with the research subjects being patients in the respiratory ward of a tertiary general hospital in Shandong Province, China. This study was conducted solely through questionnaires and did not involve any invasive procedures, thus posing no risks or harm to patients. Subjects Using the convenient sampling method, a total of 320 patients with chronic obstructive pulmonary disease from a Class III Grade A hospital in Shandong Province were selected as the research objects from July to August 2025. Inclusion criteria: (1) meet the diagnostic criteria for chronic obstructive pulmonary disease in the “Guidelines for Primary diagnosis, treatment and Management of Chronic obstructive pulmonary Disease in China (2024)“ [12]; (2) age ≥ 60 years old; (3) the disease is in a stable stage; (4) have a clear consciousness and can cooperate to complete the investigation; (5) Informed consent of patients. Exclusion criteria:1) patients with severe organic diseases; 2) malignant tumors; 3) patients with mental illness; 4) unable to communicate normally; 5) Oral diseases due to trauma. Sample size Estimation According to the sample size estimation formula of quantitative research in nursing research [13], N = 4U2αS2/δ2, where U is the U value corresponding to the test level, S represents the standard deviation, and σ represents the allowable error. The standard deviation S and the allowable error δ can be derived from the data in the pilot experiment (or from other results in the literature) [14]. Usually, α = 0.05; this study takes a two-sided test, U = 1.96, the allowable error is [0.25 S,0.5 S] [15], and the standard deviation is S = 1.93 [16]. Based on the above formula and considering a 20% sample loss rate, the sample size range is only 4 to 15 cases. Meanwhile, referring to the calculation formula for estimating the overall rate in the sampling survey, N=[Z2α/2P(1-P)]/σ2, the allowable error σ is set to 0.05. Through reading the literature, it is found that the incidence of oral frailty in elderly hospitalized patients with COPD was approximately 96.1% [16], P = 0.961,α = 0.05, Z0.05/2=1.96. The sample size n was calculated to be 58. Considering a sample loss rate of 10%−20%, the final sample size range was determined to be 64–72 cases. By integrating two sample size calculation methods to ensure the goodness of model fit and sample representativeness, a total of 350 elderly COPD patients were investigated in this study. Excluding invalid questionnaires, a total of 320 elderly COPD patients who met the criteria were included in this study. Selection of variables Based on the health ecology model and literature review, this study selected the influencing factors that might lead to oral frailty in elderly COPD patients. (1) Personal innate characteristics: gender, age, occupation, place of birth, course of COPD, degree of dyspnea, type of chronic disease, type of oral medication, nutrition, frailty, intact teeth, and oral diseases; (2) psychological behavior: smoking, exercise, eating speed, eating habits, oral health related self-efficacy (3) family community network: marital status; (4) Working and living conditions: previous occupation, hukou location, and monthly income; (5) Policy environment: type of medical insurance. Instrument of survey General information questionnaire By reviewing the relevant literature, The questionnaire included gender, age, occupation, place of origin, marital status, monthly income, type of medical insurance, course of COPD, degree of dyspnea, types of chronic diseases, types of oral drugs, smoking, exercise, eating speed, eating habits, types of daily diet, intact teeth, and oral diseases. Oral frailty Index-8 (OFI-8) The scale was developed by Tanaka et al. [17]in 2021, and Chen Zongmei et al. [18]translated it into a Chinese version in 2023 to screen for oral frailty in the elderly. The prevalence OF oral frailty in elderly patients with chronic obstructive pulmonary disease screened by scholars such as Liang Yuanjun using F-8 was 96.1%, and the prevalence of oral frailty in elderly patients with chronic obstructive pulmonary disease was high [16]. The Cronbach’s α coefficient OF the Chinese version of F-8 was 0.692, indicating good reliability and validity. The scale included 5 dimensions (whether to use dentures, swallowing function, social participation, oral health-related behaviors, and chewing ability) and 8 items. The total score ranged from 0 to 11 points, and ≥ 4 points were defined as oral frailty. Higher scores indicated worse oral conditions. The Cronbach’s α coefficient of the scale in this study was 0.834. Fatigue resistance ambulation illness and loss of weight scale (FRAIL) The scale was proposed by the International Nutrition and Aging Society (IANA) in 2008 and is widely used in the assessment of frailty of the elderly in China. Wei Yin et al. [19] completed the localization revision in 2018, and its Cronbach’s α coefficient was 0.826, which had good reliability. The scale was composed of five dimensions, including fatigue degree, resistance, gait speed, number of chronic diseases, and weight change. Each dimension was scored using a dichotomous method, with 1 point assigned to a “yes” response and 0 points assigned to a “no” response. The total score was evaluated as follows: a score of 0 indicated no frailty, 1–2 indicated pre-frailty, and ≥ 3 was considered as frailty. The Cronbach’s α coefficient of the scale in this study was 0.830. Mini-Nutritional assessment short form (MNA-SF) This scale was developed by Rubenstein et al. [20]in 2001 for the screening of malnutrition in the elderly with high sensitivity and specificity. Zhang Yan et al. [21]analyzed the internal consistency of the scale, and its Cronbach’s α coefficient was 0.711, indicating that the reliability of the scale was good. The scale consists of six domains, including changes in diet, weight loss in the past 3 months, mobility, stress or acute illness, neuropsychiatric illness, and body mass index or calf circumference. The total score was 14. A score of 0–7 indicates that malnutrition is already present, 8–11 indicates that there is a risk of malnutrition, and ≥ 12 indicates normal nutritional status. The Cronbach’s α coefficient of the scale in this study was 0.828. Geriatric Self-Efficacy scale for oral health (GSEOH) The scale was developed by Ohara et al. [22]in 2017 to investigate the oral health-related self-efficacy of the elderly in Japan. Xu Yuxin et al. [23]conducted the Chinese version and reliability and validity test of the scale in 2021. The Cronbach’s α coefficient of the Chinese version of the scale was 0.913; the reliability of the scale was good, and the scale had good validity. The scale was composed of three dimensions, including the oral hygiene habits dimension, the oral function dimension, and the oral treatment habits dimension. The Likert 4-point scoring method was used, with 1 being not confident at all, and 4 being very confident. The total score ranged from 20 to 80; the higher the score, the higher the level of self-efficacy. The Cronbach’sα coefficient of the scale in this study was 0.932. Survey methods In this study, the on-site questionnaire method was used to complete the data collection, and the investigators were responsible for the distribution and recovery of the questionnaires. Then, the purpose and significance of the survey were explained to the subjects and their accompanying family members, and the information was promised to be confidential. The survey was carried out after informed consent. For subjects who could not complete the questionnaire independently, the researchers or their family members would assist them in filling in the questionnaire. In this study, a total of 350 questionnaires were distributed, and 320 valid questionnaires were collected, with an effective recovery rate of 91.4%. Statistical analysis SPSS 25.0 software was used for data analysis. For quantitative variables with skewed distribution, median and interquartile range M (P25, P75) were used for descriptive statistics. Categorical variables were presented as frequencies and percentages. Non-parametric tests such as the Mann-Whitney U test (two-group comparison) or Kruskal-Wallis H test (multi-group comparison) were used for quantitative variables comparison between groups, and the Chi-square test or Fisher’s exact test was used for categorical variables comparison between groups. In addition, the significant variables screened by univariate analysis were included in the multivariate Logistic regression model to determine the relevant influencing factors for the risk of oral frailty in elderly patients with chronic obstructive pulmonary disease. RStudio software was used to draw a nomogram according to the influencing factors, and calculate the confusion matrix of the prediction model according to the best Youden index to evaluate the performance of the model. The ROC curve took the true positive rate as the vertical axis and the false positive rate as the horizontal axis, and the area under the curve (AUC) measured the discrimination power. The closer the AUC was to 1, the stronger the discrimination power was. The calibration curve was compared between the predicted probability and the actual frequency; the closer to the 45-degree diagonal, the better the calibration. Hosmer-Lemeshow test P > 0.05 indicated a good fit. The decision analysis curve was based on the net benefit of different threshold probabilities. When comparing the model with “full intervention” and “no intervention” strategies, the clinical value of the model was higher when the net benefit was obvious. All tests α = 0.05. Results Characteristics of the participants A total of 320 elderly COPD patients who met the criteria were included in this study, of whom 296 had oral frailty and 24 had non-oral frailty (Fig. 1). There was no significant difference in the main baseline characteristics between the training set and the validation set (Table S1) (P > 0.05), indicating that the data of the two groups were balanced and comparable, which provided a reliable basis for subsequent model validation. Fig. 1Sample screening Flowchart Sample screening Flowchart Univariate analysis of oral frailty risk in elderly patients with COPD The results of the univariate analysis are shown (Table 1). There were statistically significant differences in the incidence of oral frailty among elderly COPD patients in terms of age, malnutrition, types of chronic diseases, degree of dyspnea, and intact teeth (P < 0.05).Table 1Univariate analysis of oral frailty(n = 320)VariablesTotal(n = 320)Non-oral frailty(n = 24)Oral frailty(n = 296)StatisticPMNA -SF sum, M (Q₁, Q₃)3.00 (2.00, 4.00)12.50 (4.00, 13.25)2.00 (2.00, 4.00)Z=−5.84< 0.001GSEOH sum, M (Q₁, Q₃)62.00 (59.00, 65.00)62.50 (59.75, 66.00)62.00 (59.00, 65.00)Z=−1.200.230FRAIL sum, M (Q₁, Q₃)1.00 (0.00, 2.00)1.00 (0.00, 1.00)1.00 (0.00, 2.00)Z=−0.600.546Sex, n(%)χ²=0.300.587Male216 (67.50)15 (62.50)201 (67.91)Female104 (32.50)9 (37.50)95 (32.09)Age, n(%)χ²=33.31< 0.00160–69129 (40.31)23 (95.83)106 (35.81)70–7999 (30.94)1 (4.17)98 (33.11)80–8969 (21.56)0 (0.00)69 (23.31)≥9023 (7.19)0 (0.00)23 (7.77)Occupation, n(%)-0.520Personnel of public institutions60 (18.75)4 (16.67)56 (18.92)Professional and technical personnel84 (26.25)6 (25.00)78 (26.35)Enterprise employees93 (29.06)5 (20.83)88 (29.73)Freelancer51 (15.94)7 (29.17)44 (14.86)Farmer32 (10.00)2 (8.33)30 (10.14)Registered permanent, n(%)χ²=1.850.174City213 (66.56)19 (79.17)194 (65.54)Rural107 (33.44)5 (20.83)102 (34.46)Marital status, n(%)-0.368Unmarried32 (10.00)1 (4.17)31 (10.47)Married237 (74.06)17 (70.83)220 (74.32)Divorced51 (15.94)6 (25.00)45 (15.20)Pension, n(%)-0.415027 (8.44)2 (8.33)25 (8.45)<100083 (25.94)5 (20.83)78 (26.35)1000-<3000139 (43.44)12 (50.00)127 (42.91)3000-<500053 (16.56)2 (8.33)51 (17.23)≥500018 (5.62)3 (12.50)15 (5.07)Insurance, n(%)-0.242Resident Medical Insurance193 (60.31)18 (75.00)175 (59.12)Employee Medical Insurance91 (28.44)3 (12.50)88 (29.73)Public expense21 (6.56)2 (8.33)19 (6.42)Own expense15 (4.69)1 (4.17)14 (4.73)The Course of COPD, n(%)-0.008≤1 (Year)99 (30.94)6 (25.00)93 (31.42)2–5158 (49.38)7 (29.17)151 (51.01)6–939 (12.19)8 (33.33)31 (10.47)≥1024 (7.50)3 (12.50)21 (7.09)Types of chronic diseases, n(%)χ²=18.13< 0.0011 (kind)102 (31.88)17 (70.83)85 (28.72)≥2218 (68.12)7 (29.17)211 (71.28)Types of oral medications, n(%)-0.3501 (kind)89 (27.81)7 (29.17)82 (27.70)2–3158 (49.38)10 (41.67)148 (50.00)4–549 (15.31)3 (12.50)46 (15.54)>524 (7.50)4 (16.67)20 (6.76)Degree of breathing difficulty, n(%)χ²=73.85< 0.0010 (Grade)39 (12.19)16 (66.67)23 (7.77)1107 (33.44)6 (25.00)101 (34.12)≥ 2174 (54.38)2 (8.33)172 (58.11)Smoke, n(%)χ²=2.070.150No129 (40.31)13 (54.17)116 (39.19)Yes191 (59.69)11 (45.83)180 (60.81)Sport, n(%)χ²=0.190.661No213 (66.56)15 (62.50)198 (66.89)Yes107 (33.44)9 (37.50)98 (33.11)The speed of eating, n(%)χ²=1.080.581Slow107 (33.44)10 (41.67)97 (32.77)Moderate179 (55.94)11 (45.83)168 (56.76)Fast34 (10.62)3 (12.50)31 (10.47)Healthy diet, n(%)χ²=2.630.105No85 (26.56)3 (12.50)82 (27.70)Yes235 (73.44)21 (87.50)214 (72.30)Oral disease, n(%)χ²=0.560.452No102 (31.88)6 (25.00)96 (32.43)Yes218 (68.12)18 (75.00)200 (67.57)Perfect teeth, n(%)χ²=14.72< 0.001No188 (58.75)23 (95.83)165 (55.74)Yes132 (41.25)1 (4.17)131 (44.26)Z Mann-Whitney test, χ²: Chi-square test, -: Fisher exact M: Median, Q₁: 1 st Quartile, Q₃: 3 st Quartile Table 1Univariate analysis of oral frailty(n = 320)VariablesTotal(n = 320)Non-oral frailty(n = 24)Oral frailty(n = 296)StatisticPMNA -SF sum, M (Q₁, Q₃)3.00 (2.00, 4.00)12.50 (4.00, 13.25)2.00 (2.00, 4.00)Z=−5.84< 0.001GSEOH sum, M (Q₁, Q₃)62.00 (59.00, 65.00)62.50 (59.75, 66.00)62.00 (59.00, 65.00)Z=−1.200.230FRAIL sum, M (Q₁, Q₃)1.00 (0.00, 2.00)1.00 (0.00, 1.00)1.00 (0.00, 2.00)Z=−0.600.546Sex, n(%)χ²=0.300.587Male216 (67.50)15 (62.50)201 (67.91)Female104 (32.50)9 (37.50)95 (32.09)Age, n(%)χ²=33.31< 0.00160–69129 (40.31)23 (95.83)106 (35.81)70–7999 (30.94)1 (4.17)98 (33.11)80–8969 (21.56)0 (0.00)69 (23.31)≥9023 (7.19)0 (0.00)23 (7.77)Occupation, n(%)-0.520Personnel of public institutions60 (18.75)4 (16.67)56 (18.92)Professional and technical personnel84 (26.25)6 (25.00)78 (26.35)Enterprise employees93 (29.06)5 (20.83)88 (29.73)Freelancer51 (15.94)7 (29.17)44 (14.86)Farmer32 (10.00)2 (8.33)30 (10.14)Registered permanent, n(%)χ²=1.850.174City213 (66.56)19 (79.17)194 (65.54)Rural107 (33.44)5 (20.83)102 (34.46)Marital status, n(%)-0.368Unmarried32 (10.00)1 (4.17)31 (10.47)Married237 (74.06)17 (70.83)220 (74.32)Divorced51 (15.94)6 (25.00)45 (15.20)Pension, n(%)-0.415027 (8.44)2 (8.33)25 (8.45)<100083 (25.94)5 (20.83)78 (26.35)1000-<3000139 (43.44)12 (50.00)127 (42.91)3000-<500053 (16.56)2 (8.33)51 (17.23)≥500018 (5.62)3 (12.50)15 (5.07)Insurance, n(%)-0.242Resident Medical Insurance193 (60.31)18 (75.00)175 (59.12)Employee Medical Insurance91 (28.44)3 (12.50)88 (29.73)Public expense21 (6.56)2 (8.33)19 (6.42)Own expense15 (4.69)1 (4.17)14 (4.73)The Course of COPD, n(%)-0.008≤1 (Year)99 (30.94)6 (25.00)93 (31.42)2–5158 (49.38)7 (29.17)151 (51.01)6–939 (12.19)8 (33.33)31 (10.47)≥1024 (7.50)3 (12.50)21 (7.09)Types of chronic diseases, n(%)χ²=18.13< 0.0011 (kind)102 (31.88)17 (70.83)85 (28.72)≥2218 (68.12)7 (29.17)211 (71.28)Types of oral medications, n(%)-0.3501 (kind)89 (27.81)7 (29.17)82 (27.70)2–3158 (49.38)10 (41.67)148 (50.00)4–549 (15.31)3 (12.50)46 (15.54)>524 (7.50)4 (16.67)20 (6.76)Degree of breathing difficulty, n(%)χ²=73.85< 0.0010 (Grade)39 (12.19)16 (66.67)23 (7.77)1107 (33.44)6 (25.00)101 (34.12)≥ 2174 (54.38)2 (8.33)172 (58.11)Smoke, n(%)χ²=2.070.150No129 (40.31)13 (54.17)116 (39.19)Yes191 (59.69)11 (45.83)180 (60.81)Sport, n(%)χ²=0.190.661No213 (66.56)15 (62.50)198 (66.89)Yes107 (33.44)9 (37.50)98 (33.11)The speed of eating, n(%)χ²=1.080.581Slow107 (33.44)10 (41.67)97 (32.77)Moderate179 (55.94)11 (45.83)168 (56.76)Fast34 (10.62)3 (12.50)31 (10.47)Healthy diet, n(%)χ²=2.630.105No85 (26.56)3 (12.50)82 (27.70)Yes235 (73.44)21 (87.50)214 (72.30)Oral disease, n(%)χ²=0.560.452No102 (31.88)6 (25.00)96 (32.43)Yes218 (68.12)18 (75.00)200 (67.57)Perfect teeth, n(%)χ²=14.72< 0.001No188 (58.75)23 (95.83)165 (55.74)Yes132 (41.25)1 (4.17)131 (44.26)Z Mann-Whitney test, χ²: Chi-square test, -: Fisher exact M: Median, Q₁: 1 st Quartile, Q₃: 3 st Quartile Univariate analysis of oral frailty(n = 320) Z Mann-Whitney test, χ²: Chi-square test, -: Fisher exact M: Median, Q₁: 1 st Quartile, Q₃: 3 st Quartile Multivariate logistic regression analysis of oral frailty in elderly COPD patients Taking the variables with P < 0.05 in the univariate analysis of the risk of oral frailty in elderly COPD patients as independent variables and oral frailty as the dependent variable, a multivariate Logistic regression analysis was conducted. The results (Table 2) showed that malnutrition, types of chronic diseases, and the degree of dyspnea were independent risk factors for oral frailty in elderly COPD patients (P < 0.05).Table 2The logistic regression analysis of oral frailty (n = 320)VariablesβS.EZPOR (95%CI)Intercept0.781.010.770.4382.18 (0.30 ~ 15.65)Types of chronic diseases1 (Kind)1.00 (Reference)≥23.391.123.030.00229.76 (3.32 ~ 266.79)MNA -SF sum−0.330.09−3.37< 0.0010.74 (0.62 ~ 0.88)Degree of breathing difficulty0 (Grade)1.00 (Reference)13.001.092.750.00619.99 (2.37 ~ 168.51)≥24.931.413.49< 0.001138.66 (8.71 ~ 2208.49)OR Odds Ratio, CI Confidence Interval Table 2The logistic regression analysis of oral frailty (n = 320)VariablesβS.EZPOR (95%CI)Intercept0.781.010.770.4382.18 (0.30 ~ 15.65)Types of chronic diseases1 (Kind)1.00 (Reference)≥23.391.123.030.00229.76 (3.32 ~ 266.79)MNA -SF sum−0.330.09−3.37< 0.0010.74 (0.62 ~ 0.88)Degree of breathing difficulty0 (Grade)1.00 (Reference)13.001.092.750.00619.99 (2.37 ~ 168.51)≥24.931.413.49< 0.001138.66 (8.71 ~ 2208.49)OR Odds Ratio, CI Confidence Interval The logistic regression analysis of oral frailty (n = 320) OR Odds Ratio, CI Confidence Interval Construction of a nomogram of oral frailty in elderly COPD patients and evaluation of its testing efficacy For the variables that were meaningful in the multivariate analysis, a nomogram was constructed using R Studio to visualize the risk factors (Fig. 2). Based on the total score of micronutritional assessment screening in elderly COPD patients, the degree of dyspnea, and the type of chronic disease, the total score could be calculated corresponding to the nomogram to predict the probability of oral frailty. The area under the receiver operating characteristic curve of the training set was 0.97(95%CI:0.94–1.00.94.00) (Fig. 3A), and that under the receiver operating characteristic curve of the validation set was 0.96(95%CI:0.91- 1.00) (Fig. 3B). The AUC values of both the training set and the validation set were greater than 0.8, indicating that the model has a strong ability to distinguish oral frailty. The calibration curve shows that the risk of oral frailsia predicted by the model is highly consistent with the actual situation (Fig. 4A and B), and the Hosmer-Lemeshow test (P = 0.994 for the training set and P = 0.540 for the validation set) indicates a high degree of model fit. The decision curves of the oral frailty prediction model show that the net benefits of the decision curves corresponding to the models in both the training set and the validation set are higher than those of the “total intervention” and “no intervention” strategies, indicating that the model has high clinical application value. (Figure 5A and B).Fig. 2Nomogram for predicting the risk of oral frailtyFig. 3A The ROC curve of the training dataset B The ROC curve of the verification datasetFig. 4 A Calibration curve of the training dataset B Calibration curve of the verification datasetFig. 5A Calibration curve of the verification dataset B The DCA curve of the verification dataset Nomogram for predicting the risk of oral frailty A The ROC curve of the training dataset B The ROC curve of the verification dataset A Calibration curve of the training dataset B Calibration curve of the verification dataset A Calibration curve of the verification dataset B The DCA curve of the verification dataset Discussion Based on the theoretical guidance of the health ecology model, this study comprehensively integrated multi-dimensional factors such as personal innate characteristics, psychological and behavioral patterns, family and community network, working and living conditions, and policy environment of elderly COPD patients, and constructed a nomogram model for oral frailty in them. The research results show that the incidence of oral frailty among participants are as high as 92.5%, which were consistent with the study of Liang Yuanjun et al. (96.1%) [10]. Compared with the community elderly (70.3%) [24], the rural elderly (44.7%) [25], the elderly in nursing homes (31.0) [26], the elderly patients with Parkinson’s disease (28.67%) [27], the elderly patients with stroke (47.8%) [28], and the prevalence of oral frailty in patients with diabetes mellitus (45.4%) [29]. This indicates that patients with advanced chronic obstructive pulmonary disease (COPD) are at high risk of oral health issues due to the severity of the disease and the intensity of treatment. This high-risk environment is associated with the cumulative impact of long-term use of multiple drugs by COPD patients on the oral microenvironment. Studies have shown that severe systemic diseases similar to COPD, such as COVID-19, have common oral manifestations in patients, mainly attributed to disease-related immune impairment and/or drug treatment, rather than the primary viral infection itself [30]. Therefore, among COPD patients, the high prevalence of oral diseases may be closely related to the combined impact of long-term drug use on the oral microenvironment. Some patients with COPD require long-term use of antibiotics, inhaled glucocorticoids, or long-acting β₂ receptor agonists. These drugs [31, 32] are associated with the inhibition of salivary gland function, reducing salivary gland secretion and causing changes in physical and chemical properties (such as increased viscosity and decreased buffering capacity). It can lead to a series of oral problems, such as dry mouth, dental caries, periodontal disease, and fungal infection, and eventually increase the risk of oral frailty. Previous studies have suggested [33] that for such patients, increasing the frequency of oral cleaning (such as increasing the frequency of brushing teeth and gargling) and gargling 2%−4% sodium bicarbonate solution regularly for oral cleaning can effectively reduce the incidence of oral fungal infection and improve oral health. The results of this study show that the types of chronic diseases are risk factors for oral frailty in elderly COPD patients. The more types of chronic diseases elderly COPD patients have, the higher the probability of oral frailty, which is consistent with the research of Fan Xiaoli et al. [34, 35]. When multiple chronic diseases (such as diabetes, cardiovascular diseases, osteoporosis, etc.) coexist, Multiple types of drugs, such as bronchodilators, glucocorticoids, antihypertensive drugs, and hypoglycemic drugs need to be used in combination. Drug interactions are associated with inhibiting salivary gland function, altering the balance of oral flora, leading to dry mouth, weakened mucosal barrier, and increased risk of infection [35]. The pathological mechanisms of different chronic diseases can produce synergistic effects. For instance, the hyperglycemic state of diabetes and the chronic inflammation of COPD jointly disrupt the oral immune microenvironment, accelerating the destruction of periodontal tissues and the decline in mucosal repair capacity [36]. In addition, the coexistence of multiple diseases is often accompanied by an unbalanced intake of nutrients (such as protein and vitamin deficiencies), which further weakens the metabolic function of oral tissues, exacerbates the decline of oral muscle strength, and causes abnormal sensations and functional degeneration [37]. Therefore, clinical medical staff need to optimize drug treatment plans, reduce adverse drug reactions, maintain the balance of oral flora, and lower the incidence of infections. When COPD patients have other diseases, medical staff should consider the treatment plans for different diseases, pay attention to the mutual influence among diseases, keep the patients’ oral immune microenvironment normal, and, at the same time, patients with multiple coexisting diseases should pay attention to supplementing protein and vitamins to meet the body’s nutritional needs. This study showed that the lower the MNA screening score, the more prone to oral frailty, and the risk of oral frailty increased in patients with malnutrition (MNA screening score 0–7). Consistent with the study by Iwasaki et al. [38], malnutrition determined using the MNA-SF score was directly related to oral weakness. Studies have shown that the direct manifestations of oral frailty, such as masticatory dysfunction and swallowing dysfunction, are important independent risk factors for malnutrition. Both of them affect the mechanical processing and transport of food, forming a vicious circle of “oral function degradation, nutrient intake limitation, and body metabolic imbalance“ [39]. Therefore, medical and nursing staff should pay close attention to the nutritional status of patients and supplement nutrients according to the patient’s constitution. For elderly patients with chewing inconvenience, they can ensure nutritional intake by adjusting the diet form to meet the body’s metabolic needs. This study also shows that elderly COPD patients with severe dyspnea are more likely to suffer from oral frailty, which is consistent with the research of Chen Zongmei et al.[10]. When COPD patients experience severe dyspnea due to airway obstruction or decreased lung function, they are more likely to compensate through mouth breathing. This abnormal breathing pattern may lead to accelerated evaporation of water from the oral mucosa (causing dry mouth) and persistent spasms of oral muscle groups (such as masticatory muscles and tongue muscles). Persistent dry mouth can weaken the function of the oral mucosal barrier and increase the risk of dental caries, periodontal disease, and fungal infections. Long-term spasms of the oral muscle groups may accelerate the degenerative changes of muscle fibers, leading to functional degeneration such as decreased tongue pressure and reduced chewing efficiency, and ultimately inducing oral weakness. Therefore, medical staff need to optimize drug treatment plans and guide COPD patients to undergo long-term oxygen therapy and pulmonary rehabilitation training measures to alleviate breathing difficulties. In addition, increasing the frequency of oral hygiene for COPD patients, applying 4℃ coconut water spray (to relieve dry mouth), acupoint massage (to relax oral muscle groups), and oral function training (such as tongue muscle strength training and chewing function rehabilitation) can directly improve the oral microenvironment and prevent further deterioration of oral function [40–42]. Conclusion Under the framework of the health ecology model, this study comprehensively analyzed multiple dimensions of elderly COPD patients, including their individual innate traits, psychological and behavioral factors, family and community relationships, work and living environments, and policy environments. Our research indicates that oral frailty exists in the majority of elderly COPD patients. The types of chronic diseases, MNA screening scores, and the degree of dyspnea in the individual’s innate characteristic layer in the health ecology model are factors influencing the occurrence of oral frailty. This study constructed a visual nomogram prediction model based on three core risk factors. By quantitatively calculating the cumulative scores of each indicator, personalized risk prediction probabilities for oral frailty were generated. The application of this tool helps accurately identify high-risk groups, providing a basis for early targeted intervention measures, thereby optimizing the allocation of public health resources and reducing the medical expenditure burden of the elderly population. It has significant health economic value. Limitations The oral frailty risk prediction model constructed in this study has a good screening efficiency. The visual nomogram tool has the advantages of convenient operation and intuitive results, and can effectively quantify the oral frailty risk of the elderly. However, this study has some limitations. Firstly, although the OF-8 questionnaire adopted in this study is convenient for large-scale screening oral frailty, the questionnaire data come from the self-reports of the subjects, which may affect the objectivity of the results. Future research should consider the combination of subjective and objective investigation tools to enhance the authenticity and reliability of the assessment results. Secondly, this study adopted a cross-sectional design, which could only reveal the correlation between predictive factors and oral frailty, but could not determine their causal relationship. Therefore, this study was unable to determine which factors occurred temporally before oral frailty or whether there was a bidirectional causal relationship. Future research suggests adopting a longitudinal design to track the research subjects through time series, thereby revealing the causal mechanism and its dynamic change process. In addition, this study adopted the convenience sampling method and selected patients from a tertiary grade A hospital in Shandong Province as samples. Shandong Province is a populous province in China, with a developed economy and abundant medical resources. Tertiary grade A hospitals usually admit patients with severe COPD, and most of them are in the stable stage, which is consistent with the overall disease characteristics of COPD patients in China. Although the hospitals selected in this study have certain regional representativeness, there are also COPD patients who cannot represent all regions and types of hospitals in our country. Future research should adopt multi-center, large-sample cohort studies to verify the universal applicability of the model. Finally, although the sample size of this study has certain statistical power overall and can provide reasonable support for multivariate regression analysis, the sample size of patients with oral asthenia is significantly higher than that of non-oral asthenia patients, with a specific ratio of 296:24. This may lead to an unstable data distribution in the oral asthenia group in multivariate regression analysis. This thereby affects the predictive ability of the model and the reliability of the results. Although the model demonstrated excellent ROC/AUC values (0.97) on the training set, there was a risk of overfitting considering the small validation sample size (n = 97). To verify the generalization ability of the model, future research should use a larger sample size for training to enhance the model’s generalization ability. At the same time, external validation should be conducted on the model to enhance its stability and accuracy in practical applications. Supplementary Information Supplementary material 1. Supplementary material 1.