Method and system for evaluating hypertension risk of residents in residential area under influence of noise exposure

Abstract

The invention provides a method and system for evaluating the hypertension disease risk of residents in a residential area under the influence of noise exposure, and belongs to the technical field of disease risk evaluation. The method comprises the following steps: collecting personal basic information, dietary habits, noise exposure level data, building influence factors related to noise propagation and hypertension illness condition data, and carrying out preprocessing and data verification; the hypertension disease risk serves as a dependent variable, other data serves as an independent variable, the variables are assigned, and single-factor logistic regression analysis is conducted on the assigned variables; performing multi-factor logistic regression analysis on the variables with low significance corresponding to the variables; establishing a plurality of evaluation models for the hypertension risk of residents in the residential area under the influence of noise exposure based on a multi-factor logistic regression analysis result, and screening an optimal evaluation model according to the accuracy of a prediction result; and outputting the importance of each risk factor in the optimal evaluation model prediction result. And the hypertension disease risk aiming at the noise exposure influence can be quantitatively estimated.

Description

Technical Field

[0001] This invention relates to the field of disease risk assessment technology, and in particular to a method and system for assessing the risk of hypertension in residential residents affected by noise exposure. Background Technology

[0002] Numerous epidemiological studies have shown a link between road traffic noise and hypertension prevalence. Buildings, as the ultimate barrier protecting residents from noise, can alter the characteristics of noise entering indoor spaces due to their inherent characteristics. However, the exact impact of building factors on the relationship between road traffic noise and hypertension prevalence remains unclear, making it difficult to accurately estimate the prevalence of hypertension among residents who spend long periods indoors under noise exposure.

[0003] To date, existing models of "noise exposure-hypertension" do not fully consider the corrective effects of building-related variables. In Leon Bluhm G, Berglind N, Nordling E, et al. Road traffic noise and hypertension [J]. Occupational and Environmental Medicine, 2007, 64(2): 122-126, the "noise exposure-hypertension" model only includes three building-related variables besides basic personal information (gender, age, occupation, length of residence), dietary habits (smoking frequency), and noise exposure: window type, bedroom window orientation, and construction year. The model does not include any other building factors. In Babisch W, Swart W, Houthijs D, et al. Exposure modifiers of the relationships of transportation noise with high blood pressure and noise annoyance [J]. The Journal of the Acoustical Society of America, 2012, 132(6):3788-3808, a "noise exposure-hypertension" model was established. This model only added building-related variables such as room orientation, housing type, window opening status, and window visibility to the street to modify the noise exposure level variable. It did not include other variables related to residents' basic personal information, dietary habits, etc. In another model developed by Babisch W, Wolke G, Heinrich J, et al. Roadtraffic noise and hypertension – Accounting for the location of rooms [J]. Environmental Research, 2014, 133: 380-387, although it included variables such as basic personal information (gender, age, occupation, length of residence, body mass index), dietary habits (frequency of smoking, frequency of drinking, frequency of exercise, salt intake), shape of the road adjacent to the residence, and noise exposure, it did not add any building-related variables for correction.The "noise exposure-hypertension" model established in Banerjee D, Das PP, Fouzdar A. Urban residential road trafficnoise and hypertension: a cross-sectional study of adult population [J]. Journal of Urban Health, 2014, 91(6): 1144-1157, only includes two building-related variables—residential type and bedroom window orientation—in addition to basic personal information (gender, age, occupation, length of residence), dietary habits (smoking frequency), and noise exposure. It does not include other building-related variables. The model contains few building-related variables, lacks systematicity, and does not rank the degree of influence of various factors. Furthermore, the model is mainly applicable to populations in Europe and America, and lacks applicability to the Chinese population. Summary of the Invention

[0004] To address the technical problems existing in the prior art, this invention provides a method for assessing the risk of hypertension among residents in residential areas affected by noise exposure, aiming to improve the accuracy of hypertension risk assessment.

[0005] This invention provides a method for assessing the risk of hypertension in residential residents affected by noise exposure, comprising the following steps: Step S1: Collect personal basic information, dietary habits, noise exposure level data, building impact factors related to noise transmission, and hypertension prevalence data, and perform preprocessing. Step S2: Verify the preprocessed data; Step S3: Using the risk of hypertension in the validated data as the dependent variable and other data as independent variables, assign values ​​to the variables and perform univariate logistic regression analysis on the assigned variables. Step S4: Perform multi-factor logistic regression analysis on variables whose significance level is less than or equal to a preset threshold. Step S5: Based on the results of multivariate logistic regression analysis, establish multiple assessment models for the risk of hypertension among residents affected by noise exposure, and select the best assessment model based on the accuracy of the prediction results. Step S6: Output the importance of each risk factor in the prediction results of the best evaluation model.

[0006] Preferably, the basic personal information in step S1 includes gender, age, occupation, length of residence, and body mass index; dietary habits include smoking, drinking, and exercise; architectural factors include window type, bedroom window orientation, seasonal window opening, window visibility to the street or greenery, floor level, and strength of standing wave resonance effect in the bedroom; the above data and data on hypertension prevalence are obtained through questionnaires or epidemiological surveys; and noise exposure level data are obtained through noise simulation.

[0007] Preferably, the pretreatment includes screening and preliminary treatment.

[0008] Preferably, data from questionnaires that are incomplete, unclear, or contain obvious errors are primarily filtered out.

[0009] Preferably, the preliminary processing includes data entry (e.g., entering questionnaire information into Excel software), logical verification (e.g., observing whether the numerical range conforms to common sense and the set range, whether the related questions are consistent, etc.), and manual verification (e.g., for a randomly selected portion of the questionnaires, comparing the original questionnaires with the entered data to check for omissions or errors).

[0010] Preferably, the verification in step S2 includes the chi-square test, the Shapiro-Wilk test, and the rank-sum test.

[0011] Preferably, the chi-square test is performed on the categorical variables according to formulas (1)-(3): (1) in, It is the total observation frequency of the i-th row; It is the total observation frequency of the j-th column; N is the total number of samples, which is the sum of the observation frequencies of all cells; (2) in, It is the observation frequency in the i-th row and j-th column; It is the expected frequency of the i-th row and j-th column; It is the chi-square statistic; (3) in, It is the number of categories in the row. It is the number of categories in the column. It's about freedom.

[0012] Preferably, the Shapiro-Wilk test is performed on continuous variables using formula (4): (4) in, n is the number of data items; It is the i-th observation in the data; It is the mean of the data; It is a constant related to the normally distributed data after sorting, and includes the number of data points. n The theoretical parameters for both normal and normal distributions were calculated.

[0013] Preferably, the Wilcoxon rank-sum test is performed on variables that do not satisfy a normal distribution according to the following formula; (5) (6) in, and These are two sets of statistics, and ; The final verification statistic is: .

[0014] Preferably, in step S3, a single-factor logistic regression analysis is performed using the following formula: (7) (8) (9) (10) Preferably, in step S4, multivariate logistic regression analysis is performed on variables with a significance level ≤ 0.05 according to the following formula: (11) (12) (13) (14) in, It is the probability that the dependent variable Y takes the value 1; It is the log odds of the logistic regression model, defined as the natural logarithm of the ratio of the probability that the dependent variable Y is 1 to the probability that the dependent variable Y is 0. It is the intercept term; It is the independent variable The regression coefficients; It is an independent variable of multiple factors; It is a linear combination of all independent variables; It is the dominance ratio of each independent variable; X is a single-factor independent variable; e It is the base of the natural logarithm; It is a linear combination in the logistic regression model, representing an estimate of the influence of the independent variable X on the dependent variable Y on a log-odds scale; Logit is short for logarithmic probability, which is the natural logarithm of the ratio of the probability of an event occurring to its probability of not occurring. OR It is the odds ratio, which represents the factor by which the logarithmic probability of an event increases when the independent variable X increases by one unit.

[0015] Preferably, in step S5, the risk of developing hypertension is used as the dependent variable, and different combinations of other data are used as independent variables. Multiple factors are continuously modified through stepwise regression to establish multiple assessment models for the risk of developing hypertension. Based on the analysis results of the assessment models, the predictive ability of the models, the hypothesis testing of the models, the goodness of fit, and in combination with professional knowledge, the optimal model is selected from the multiple assessment models established.

[0016] The present invention provides an assessment system for the risk of hypertension among residents in residential areas affected by noise exposure, comprising a processor capable of executing a computer program that implements the aforementioned assessment method for the risk of hypertension among residents in residential areas affected by noise exposure.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Estimate the risk of developing hypertension in response to the effects of noise exposure. Because this invention incorporates more building factors related to noise propagation than existing "noise exposure-hypertension" models, it can more accurately output an assessment model of the risk of hypertension among residents affected by noise exposure. Based on this, the risk of hypertension caused by noise exposure can be quantitatively estimated.

[0018] 2. Quantitatively determine the extent to which building factors, noise, and other risk factors affect the prevalence of hypertension. The assessment model for the risk of hypertension among residential residents affected by noise exposure, established according to this invention, quantifies the impact of building factors, noise, and other risk factors on hypertension prevalence by ranking the regression coefficients of each variable. This identifies the most important risk factors among all factors, as well as the most important factors within each category, helping decision-makers propose more scientific solutions. Such technical effectiveness has not yet been achieved in existing "noise exposure-hypertension" models. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of a method for assessing the risk of hypertension among residents in a residential area affected by noise exposure, according to an embodiment of the present invention. Detailed Implementation

[0021] The specific embodiments of the present invention will be described in detail below.

[0022] This invention provides a method for assessing the risk of hypertension in residential residents affected by noise exposure, comprising the following steps: Step S1: Collect personal basic information, dietary habits, noise exposure level data, building impact factors related to noise transmission, and hypertension prevalence data, and perform preprocessing. Step S2: Verify the preprocessed data; Step S3: Using the risk of hypertension in the validated data as the dependent variable and other data as independent variables, assign values ​​to the variables and perform univariate logistic regression analysis on the assigned variables. Step S4: Perform multi-factor logistic regression analysis on variables whose significance level is less than or equal to a preset threshold. Step S5: Based on the results of multivariate logistic regression analysis, establish multiple assessment models for the risk of hypertension among residents affected by noise exposure, and select the best assessment model based on the accuracy of the prediction results. Step S6: Output the importance of each risk factor in the prediction results of the best evaluation model.

[0023] According to a specific embodiment of the present invention, the basic personal information in step S1 includes gender, age, occupation, length of residence, and body mass index; dietary habits include smoking, drinking, and exercise; architectural factors include window type, bedroom window orientation, seasonal window opening, window visibility to the street or greenery, floor level, and strength of standing wave resonance effect in the bedroom; the above data and data on hypertension prevalence are obtained through questionnaires or epidemiological surveys; and noise exposure level data are obtained through noise simulation.

[0024] According to one specific embodiment of the present invention, the preprocessing includes screening and preliminary processing.

[0025] According to a specific embodiment of the present invention, data from questionnaires that are incomplete, unclear, or contain obvious errors are mainly filtered out.

[0026] According to a specific embodiment of the present invention, the preliminary processing includes data entry (e.g., entering questionnaire information into Excel software), logical verification (e.g., observing whether the numerical range conforms to common sense and the set range, whether the related questions are consistent, etc.), and manual verification (e.g., for a randomly selected portion of the questionnaires, comparing the original questionnaires with the entered data to check for omissions or errors).

[0027] According to a specific embodiment of the present invention, the verification in step S2 includes the chi-square test, the Shapiro-Wilk test, and the rank-sum test.

[0028] According to a specific embodiment of the present invention, the chi-square test is performed on the categorical variables according to formulas (1)-(3): (1) in, It is the total observation frequency of the i-th row; It is the total observation frequency of the j-th column; N is the total number of samples, which is the sum of the observation frequencies of all cells; (2) in, It is the observation frequency in the i-th row and j-th column; It is the expected frequency of the i-th row and j-th column; It is the chi-square statistic; (3) in, It is the number of categories in the row. It is the number of categories in the column. It's about freedom.

[0029] According to a specific embodiment of the present invention, the Shapiro-Wilk test is performed on continuous variables using formula (4): (4) in, n is the number of data items; It is the i-th observation in the data; It is the mean of the data; It is a constant related to the normally distributed data after sorting, and includes the number of data points. n The theoretical parameters for both normal and normal distributions were calculated.

[0030] According to a specific embodiment of the present invention, the Wilcoxon rank-sum test is performed on variables that do not satisfy a normal distribution according to the following formula; (5) (6) in, and These are two sets of statistics, and ; The final verification statistic is: .

[0031] According to a specific embodiment of the present invention, in step S3, a single-factor logistic regression analysis is performed using the following formula: (7) (8) (9) (10) According to a specific embodiment of the present invention, in step S4, multifactor logistic regression analysis is performed on variables with a significance level ≤ 0.05 according to the following formula: (11) (12) (13) (14) in, It is the probability that the dependent variable Y takes the value 1; It is the log odds of the logistic regression model, defined as the natural logarithm of the ratio of the probability that the dependent variable Y is 1 to the probability that the dependent variable Y is 0. It is the intercept term; It is the independent variable The regression coefficients; It is an independent variable of multiple factors; It is a linear combination of all independent variables; It is the dominance ratio of each independent variable; X is a single-factor independent variable; e It is the base of the natural logarithm; It is a linear combination in the logistic regression model, representing an estimate of the influence of the independent variable X on the dependent variable Y on a log-odds scale; Logit is short for logarithmic probability, which is the natural logarithm of the ratio of the probability of an event occurring to its probability of not occurring. OR It is the odds ratio, which represents the factor by which the logarithmic probability of an event increases when the independent variable X increases by one unit.

[0032] According to a specific embodiment of the present invention, in step S5, the risk of hypertension is used as the dependent variable, and different combinations of other data are used as independent variables. Multiple factors are continuously modified through stepwise regression to establish multiple assessment models for the risk of hypertension. Based on the analysis results of the assessment models, the predictive ability of the models, the hypothesis testing of the models, the goodness of fit, and in combination with professional knowledge, the optimal model is selected from the multiple assessment models established.

[0033] The present invention provides an assessment system for the risk of hypertension among residents in residential areas affected by noise exposure, comprising a processor capable of executing a computer program that implements the aforementioned assessment method for the risk of hypertension among residents in residential areas affected by noise exposure.

[0034] Example 1 This invention provides a method for assessing the risk of hypertension in residential residents affected by noise exposure, comprising the following steps: Step S1: Collect personal basic information, dietary habits, noise exposure level data, building impact factors related to noise transmission, and hypertension prevalence data, and perform preprocessing. Step S2: Verify the preprocessed data; Step S3: Using the risk of hypertension in the validated data as the dependent variable and other data as independent variables, assign values ​​to the variables and perform univariate logistic regression analysis on the assigned variables. Step S4: Perform multi-factor logistic regression analysis on variables whose significance level is less than or equal to a preset threshold. Step S5: Based on the results of multivariate logistic regression analysis, establish multiple assessment models for the risk of hypertension among residents affected by noise exposure, and select the best assessment model based on the accuracy of the prediction results. Step S6: Output the importance of each risk factor in the prediction results of the best evaluation model.

[0035] Example 2 Existing research has demonstrated the role of building-level factors in the impact of road traffic noise on hypertension among residents. These factors specifically include room location, bedroom window orientation, and the number of panes of glass in exterior windows, primarily affecting the amount of noise exposure after it enters the building, thereby altering the prevalence of hypertension. Babisch et al.'s study, considering the room's location relative to the road, showed that daytime noise exposure in the living room was higher and lasted longer than nighttime noise exposure in the bedroom, resulting in a greater impact on the risk of hypertension. Studies by Banerjee et al. and Bluhm et al. focused on exterior window factors, finding that residents with bedroom windows facing the street had a significantly higher risk of hypertension than those with other window orientations. Furthermore, Bluhm et al. found that triple-pane windows corresponded to a lower risk of hypertension than double-pane windows.

[0036] The method of the present invention will be described in detail below according to a specific embodiment of the present invention. For details not covered herein, please refer to the above-mentioned relevant content. The present invention provides a method for assessing the risk of hypertension in residential residents affected by noise exposure, comprising the following steps: Step S1: Collect personal basic information, dietary habits, noise exposure level data, building impact factors related to noise transmission, and hypertension prevalence data, and perform preprocessing. Research design and target population for this invention: This invention screened the surveyed residential areas based on factors such as average housing price, construction year, building structure, community size, and surrounding road traffic characteristics to ensure comparability. The final selected residential areas were Dihuali and Changhuali, and traffic information such as road grade, traffic volume, vehicle type ratio, and vehicle speed is shown in Table 1. This invention used adult residents (≥18 years old) who had lived in the areas for more than one year as the research subjects, and ultimately recruited 566 volunteers, including 334 from Dihuali and 232 from Changhuali. The questionnaire survey began in November 2021, with 10 trained investigators distributing paper questionnaires in the two residential areas and collecting information related to hypertension and residential information. All survey participants signed informed consent forms during the survey.

[0037] Architectural Factor Screening: Based on the acoustic principles related to the propagation of road traffic noise and referring to the conclusions of studies on road traffic noise and the prevalence of hypertension, the architectural factors involved in the main paths of road traffic noise entering the room were screened. At the same time, their controllability during the design was also considered. In this embodiment, the architectural factors only include factors below the level of individual buildings, such as: building floor plan, sound insulation performance of the external envelope, floor level, bedroom size ratio (strength of standing wave resonance in the bedroom), etc.

[0038] (1) Bedroom window orientation Compared to daytime noise, nighttime noise has a greater impact on residents' health. Since bedrooms are where residents primarily spend their time at night, they are more severely affected by noise compared to other rooms. Furthermore, actual measurements and related studies have found that bedroom windows facing away from or away from the street have lower noise exposure levels than those facing the street in the same household. This may have varying degrees of impact on the prevalence of hypertension among residents; therefore, the orientation of bedroom windows warrants attention.

[0039] (2) Window type In the examples, most of the residences are brick-concrete structures with 370mm thick solid brick exterior walls, achieving a weighted sound insulation of 54dB(A). In comparison, windows are a weak point in sound insulation. The sound insulation performance of windows is related to their type; different window types vary in the number of glass layers, thickness, glass spacing, window construction, and sealing, resulting in differences in sound insulation. A review of relevant standard drawings shows that the difference in sound insulation for some exterior windows can reach as high as 24dB(A). Therefore, window type needs to be analyzed as one of the building's individual factors.

[0040] (3) Floor Several scholars have studied the relationship between building height and noise exposure through actual measurements or simulations, finding that the noise distribution level on the building facade changes with building height. In this embodiment, residential buildings are mostly multi-story buildings, and the noise exposure level on the facade is also related to building height. Different floors correspond to different building heights, therefore this invention requires analysis of floor factors. (4) Bedroom size proportions According to the principle of standing wave resonance in a room, when sound waves propagate in a room, they form standing waves, which generate simple harmonic vibrations (also known as normal modes) in different dimensions. Considering the auditory characteristics of the human ear, the distribution interval of normal modes should be between 2Hz and 20Hz. If the interval is less than 2Hz, the human ear will find it difficult to distinguish them, resulting in degeneracy of the normal modes. The distribution of normal modes depends on the length, width, and height ratio of the room. When the room's dimensions are inappropriate, frequency degeneracy will occur in the lower frequency bands, amplifying noise in some frequency bands and greatly increasing the noise level, causing sound distortion. As mentioned above, noise has a greater impact on bedrooms than on other rooms; therefore, the impact of standing wave resonance on bedrooms cannot be ignored. The strength of the standing wave resonance effect in a bedroom is related to its size ratio, which in turn causes changes in the sound pressure level and spectrum within the bedroom. Therefore, the size ratio of the bedroom needs to be included in the analysis of this invention.

[0041] Noise exposure data estimation: This invention estimates the daily noise exposure level for each respondent based on collected residential information and the planar distribution of noise exposure in the residential area. The planar distribution data of residential noise exposure is obtained through simulation using the RLS90 model of SoundPLAN (version 8.0, Shenzhen Municipal Design and Research Co., Ltd. CHINA, PR). This method is the primary noise estimation method and has high feasibility and reliability. The outdoor noise exposure evaluation index of this invention is L. dn It is China's current outdoor sound environment evaluation index, and it is consistent with the L in the international ISO standard. den In comparison, although the time periods differ, both indicators take into account the significant impact of nighttime noise on residents and have been weighted accordingly. The main input parameters for noise simulation include: building parameters, road attributes such as road width, road type, vehicle type ratio, vehicle speed, and average daily traffic volume, as well as ground structures such as sound barriers and topography. Building parameters were obtained from the as-built drawings provided by the Tianjin Municipal Urban Construction Archives, topography was obtained through GIS software, and other information was obtained based on on-site surveys.

[0042] When outputting the noise distribution map of a residential area, this invention sets the ground clearance as the height of the center of the building facade to reflect the average noise distribution. Based on simulation data, residents' noise exposure can be divided into five levels: L dn <45dB(A), 45dB(A)≤L dn <50dB(A), 50dB(A)≤L dn <55dB(A), 55dB(A)≤L dn <60dB(A), 60dB(A)≤L dn <65dB(A). Due to exposure to 60dB(A)≤L dn The number of respondents exposed to noise levels <65dB(A) was small, therefore they were compared with those exposed to levels ≤55dB(A) in the data analysis. dn Respondents with an A level <60 dB(A) were grouped together.

[0043] Standardized questionnaire design for cross-sectional surveys: The study employed a cross-sectional standardized questionnaire, which typically includes three aspects: the diagnosis of hypertension, basic information about the research subjects, and their lifestyle habits. The questionnaire of this invention aims to obtain information on individual building factors, potential confounding factors related to hypertension, and the residential information of the research subjects. To achieve this objective, the questionnaire ultimately includes four main parts: building factors, the diagnosis of hypertension, basic information about the research subjects, and their lifestyle habits.

[0044] (1) Architectural factors The architectural factors in the questionnaire mainly included: bedroom window orientation (facing the street, facing elsewhere), window type (double-glazed soundproof window, single-glazed soundproof window, ordinary window), bedroom size ratio (categorized according to the strength of the standing wave resonance effect: strong standing wave resonance effect, weak standing wave resonance effect), and floor level (1-5). In this invention, since the standing wave resonance effect is most pronounced in buildings facing the street, questions related to the strength of the standing wave resonance effect were limited to answers from residents facing the street.

[0045] (2) Determination of hypertension When a participant can answer affirmatively to the question, "Were you diagnosed with hypertension at your current residence in the past?", they are considered to have hypertension.

[0046] (3) Basic information of the research subjects The basic information of the study participants mainly included: gender (male, female), age (18-80 years), BMI, employment status (student, employed, unemployed, retired), and length of residence (1-22 years). BMI was measured in meters (m). 2The data were calculated from body weight (kg). Previous research suggests that length of residence may influence hypertension outcomes; therefore, this variable is included in this section.

[0047] (4) The life habits of the research subjects This invention ranked common influencing factors of hypertension (all related to the lifestyle habits of the study subjects) according to their importance. Considering the need to control the number of questionnaire items and the total number of items in the three sections mentioned above, the following variables were ultimately selected: sleep quality (no sleep disorder, suspected sleep disorder, sleep disorder), smoking status (never smoked, used to smoked, occasionally smoked, frequently smoked), alcohol consumption (never drank alcohol, used to drink alcohol, occasionally drank alcohol, dangerous drinking, harmful drinking), and exercise status (never exercised, occasionally exercised, frequently exercised irregularly, frequently exercised regularly). Among these, alcohol consumption and exercise status are medically recognized important factors that cause hypertension. Sleep quality was statistically analyzed using the Athens Insomnia Scale (AIS). This invention also provides detailed definitions for smoking and alcohol consumption, such as: frequent smoking refers to more than 3 times per week for 6 months or more; dangerous drinking refers to men consuming 41 grams or more but less than 61 grams of alcohol per day, and women consuming 21 grams or more but less than 41 grams of alcohol per day; harmful drinking refers to men consuming 61 grams or more of alcohol per day, and women consuming 41 grams or more of alcohol per day.

[0048] Research location and research population: (1) Research location This invention selects 10 typical residential areas based on the annual average daytime and nighttime noise exposure levels of Class 1 (residential) and Class 2 (mixed) noise functional zones in the main urban area of ​​Tianjin, according to the 2020 Tianjin Environmental Bulletin and the Tianjin Municipal "Noise Environmental Quality Standard" Applicable Area Classification Report. These residential areas are representative in terms of overall noise exposure levels and are consistent with the average situation of residential areas in Tianjin. The surrounding roads include expressways, arterial roads, and secondary arterial roads. Considering the correlation between hypertension prevalence and long-term road traffic noise, and that residential areas built before 2000 (including 2000) are worse in terms of building envelope and architectural layout than those built after 2000, making them more susceptible to noise, the 10 selected residential areas were all built before 2000. Furthermore, these residential areas all have more than 1,000 households, and the building structures, except for some street-facing buildings which are frame structures, are all brick-concrete structures, with 1-5 floors. The exterior walls are mostly 370 mm thick solid brick walls, and the window types include ordinary windows, single-pane soundproof windows, and double-pane soundproof windows. The layout of the buildings in the residential area mainly includes row-and-column, perimeter, and mixed layouts. Among these residential areas, Changhuali and Dihuali are quite similar in terms of construction year, building structure, construction scale, surrounding environment, and noise exposure level. They are both close to expressways, outer ring roads, or main roads, resulting in relatively higher noise exposure levels and making it easier to observe the impact of noise on the prevalence of hypertension among residents. Therefore, Dihuali and Changhuali were ultimately selected as the research sites.

[0049] (2) Study population The sample size of the study population depends on the expected prevalence of hypertension (p), the margin of error (d), and the significance level (α). The significance level α is generally set to 0.05, and the margin of error d can be 0.2 times the prevalence p. The expected prevalence p of hypertension is calculated based on the prevalence, new incidence, and total population of hypertension in the Tianjin Residents' Health Report [88-91]. According to the calculation formula n=100×(1-p) / p, the theoretical sample size was finally determined to be 164 people. This calculation formula and parameter values ​​are widely accepted in the field of preventive medicine. To ensure the high accuracy of this invention, the expected sample size is 650 people, which is about 4 times the estimated size. The sample size of each residential area is sampled according to the ratio of the total sample size to the total number of households in each residential area. In the actual survey, 84 people were excluded because they did not meet the inclusion criteria set by this invention for the research subjects. Therefore, the final actual sample size was 566 people, which is about 3.5 times the estimated size. It included 334 people from the Dihuali residential area and 232 people from the Changhuali residential area. The age range was 18-80 years old (mean ± standard deviation = 54.72 ± 16.47). Among them, there were 266 males and 300 females.

[0050] Data preprocessing: First, the collected questionnaires are screened for validity, mainly discarding those that are incomplete, unclear, or contain obvious errors. The selected questionnaire data is then entered into an Excel spreadsheet, followed by logical verification (such as observing whether the numerical range conforms to common sense and the set range, and whether the correlation questions are consistent). Finally, manual verification is performed (randomly selecting a portion of the questionnaires and comparing them with the entered data to check for omissions or errors) to ensure the validity and accuracy of the data.

[0051] Step S2: Verify the preprocessed data; Chi-square test and rank-sum test: This invention mainly uses the SAS 9.4 (SAS Institute Incorporation, Cary, North Carolina, USA) version installation package suitable for Windows, and uses chi-square test, rank-sum test, univariate, multivariate logistic regression and stratified analysis to analyze the data, with a significance level of 0.05 for all. In this step, continuous variables such as age, BMI, length of residence, and floor level were tested using formula (4) with Shapiro-Wilk test, and variables that do not meet the normal distribution were tested using formula (5)-formula (6) with Wilcoxon rank-sum test. At the same time, gender, work status, sleep quality, smoking status, drinking status, exercise status, and L were also analyzed. dn Categorical variables such as bedroom window orientation, window type, and strength of standing wave resonance effect were subjected to chi-square tests using formulas (1)-(3) to screen out variables that have a significant impact on the prevalence of hypertension. The results are shown in Tables 1-2.

[0052] Table 1. Distribution of potential influencing factors of hypertension in different hypertension scenarios

[0053] Note: Data are presented as mean. The standard deviation or frequency (composition ratio) is used. The superscript 'a' indicates a chi-square test, used to compare differences between groups of categorical variables.

[0054] The superscript b indicates the Wilcoxon rank-sum test, which is used to compare differences between groups of continuous variables.

[0055] BMI; Body Mass Index Table 2. Distribution of standing wave resonance effect intensity under different hypertension conditions

[0056] Note: Data are presented as mean. The standard deviation or frequency (composition ratio) is used. The superscript 'a' indicates a chi-square test, used to compare differences between groups of categorical variables.

[0057] Step S3: Using the risk of hypertension in the validated data as the dependent variable and other data as independent variables, assign values ​​to the variables and perform univariate logistic regression analysis on the assigned variables. The variables involved in the questionnaire were subjected to univariate logistic regression analysis using formulas (7) to (10). The odds ratio (OR) and 95% confidence interval (CI) of hypertension under the influence of a single factor were obtained, and the results are shown in Table 3. At the same time, the variables were assigned values, and the assignment results are shown in Table 4.

[0058] Table 3. Results of univariate logistic regression

[0059] Table 4. Detailed information on variable assignment

[0060] Step S4: Perform multivariate logistic regression analysis on variables whose significance level is less than or equal to a preset threshold. The variables that showed a significant association with the risk of hypertension in the univariate logistic regression results (gender, age, work status, length of residence, smoking status, drinking status, exercise status, noise exposure, bedroom window orientation, window type, floor) were subjected to multivariate logistic regression analysis using formulas (11)-(14).

[0061] Step S5: Based on the results of multivariate logistic regression analysis, establish multiple assessment models for the risk of hypertension among residents affected by noise exposure, and select the best assessment model based on the accuracy of the prediction results. By employing a stepwise regression approach and continuously refining the assessment of confounding factors such as basic personal information and dietary habits, multiple models have been established to evaluate the risk of hypertension among residents in residential areas affected by noise exposure. The currently established models include: Model 1: Dependent variable: risk of hypertension; Independent variables: gender, length of residence, noise exposure, window type, floor, and bedroom window orientation. Model 2: The dependent variable is the risk of developing hypertension, and the independent variables are: gender, length of residence, alcohol consumption, smoking, exercise, noise exposure, window type, floor, and bedroom window orientation. Model 3: The dependent variable is the risk of developing hypertension, and the independent variables are: gender, length of residence, sleep quality, alcohol consumption, smoking, exercise, noise exposure, window type, floor, and bedroom window orientation. Model 4: The dependent variable is the risk of developing hypertension, and the independent variables are age, gender, noise exposure, window type, floor, and bedroom window orientation. Model 5: The dependent variable is the risk of developing hypertension, and the independent variables are: age, gender, exercise status, noise exposure, window type, floor, and bedroom window orientation. Model 6: The dependent variable is the risk of developing hypertension, and the independent variables are: age, gender, work status, exercise, noise exposure, window type, floor, and bedroom window orientation. The analysis results are shown in Tables 5-10. Table 5. Results of Model 1

[0062] Table 6. Results of Model 2

[0063] Table 7. Results of Model 3

[0064] Table 8. Results of Model 4

[0065] Table 9. Results of Model 5

[0066] Table 10. Results of Model 6

[0067] Based on the analysis results, model predictive ability (Somers D and Gamma values), model hypothesis testing (likelihood ratio test, scoring test, and Wald test values), goodness of fit (AIC, SC, and -2 Log L value), and combined with professional knowledge, the optimal model was selected from the multiple established models: Models 1-3 had moderate predictive ability and good fit, and the overall models were significant. Models 4-6 had good predictive ability and good fit, and the overall models were significant, with Model 6 having the strongest predictive ability; see Tables 11-16.

[0068] Table 11. Fitting effect of Model 1

[0069] Table 12. Fitting effect of Model 2

[0070] Table 13. Fitting effect of Model 3

[0071] Table 14. Fitting effect of Model 4

[0072] Table 15. Fitting effect of Model 5

[0073] Table 16. Fitting Results of Model 6

[0074] Step S6: Output the importance of each risk factor in the prediction results of the best evaluation model: Importance assessment of influencing factors: Model 1: Sorting by noise exposure: Noise exposure > occupancy time > gender > floor level > window type > bedroom window orientation. Sorting by significance: Occupancy time (0.0031) > floor level (0.0134) > gender (0.0125) > noise exposure (0.0263) > window type (0.0704) > bedroom window orientation (0.6376). Model 2: Contribution ranking: Noise exposure > Length of stay > Alcohol consumption > Gender > Floor level > Window type > Exercise activity > Bedroom window orientation. Significance ranking: Length of stay (0.0031) > Floor level (0.0134) > Gender (0.0125) > Noise exposure (0.0263) > Alcohol consumption (0.0468) > Window type (0.0505) > Exercise activity (0.0863) > Bedroom window orientation (0.5918). Model 3: Contribution ranking: Noise exposure > Length of stay > Alcohol consumption > Gender > Floor level > Window type > Exercise activity > Sleep quality > Bedroom window orientation. Significance ranking: Length of stay (0.0031) > Floor level (0.0134) > Gender (0.0125) > Noise exposure (0.0263) > Alcohol consumption (0.0468) > Window type (0.0505) > Exercise activity (0.0863) > Sleep quality (0.1342) > Bedroom window orientation (0.5600). Model 4: Contribution ranking: Age > Window type > Floor > Bedroom window orientation > Gender > Noise exposure ranking: Age (<0.0001) > Window type (0.0365) > Floor (0.0295) > Bedroom window orientation (0.0575) > Gender (0.0998) > Noise exposure (0.3302) Model 5: Contribution ranking: Age > Window type > Floor > Bedroom window orientation > Gender > Noise exposure > Exercise status. Significance ranking: Age (<0.0001) > Window type (0.0365) > Floor (0.0295) > Bedroom window orientation (0.0575) > Gender (0.0998) > Noise exposure (0.4218) > Exercise status (0.6001). Model 6: Contribution ranking: Age > Window type > Floor level > Bedroom window orientation > Gender > Job status > Noise exposure > Exercise status. Significance ranking: Age (<0.0001) > Window type (0.0365) > Floor level (0.0295) > Bedroom window orientation (0.0575) > Gender (0.0998) > Job status (0.2125) > Noise exposure (0.4218) > Exercise status (0.6001). Logit (risk of hypertension) = Intercept + SUM(Intercept, Age, Gender, Job status, Exercise status, Noise exposure, Window type, Floor level, Bedroom window orientation). In the model, age has the largest contribution, followed by noise exposure, window type, and floor level. Gender, job status, exercise status, bedroom window orientation, and the strength of the bedroom standing wave resonance effect have smaller contributions. Age is the most important risk factor among all factors. Among personal basic information factors, the most important risk factor is age; among dietary behavior factors, the most important risk factor is exercise; and among building factors related to noise transmission, the most important risk factor is window type.

[0075] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A method for assessing the risk of hypertension among residential residents affected by noise exposure, characterized in that, The steps include: Step S1: Collect personal basic information, dietary habits, noise exposure level data, building impact factors related to noise transmission, and hypertension prevalence data, and perform preprocessing. Step S2: Verify the preprocessed data; Step S3: Using the risk of hypertension in the validated data as the dependent variable and other data as independent variables, assign values ​​to the variables and perform univariate logistic regression analysis on the assigned variables. Step S4: Perform multi-factor logistic regression analysis on variables whose significance level is less than or equal to a preset threshold. Step S5: Based on the results of multivariate logistic regression analysis, establish multiple assessment models for the risk of hypertension among residents affected by noise exposure, and select the best assessment model based on the accuracy of the prediction results. Step S6: Output the importance of each risk factor in the prediction results of the best evaluation model.

2. The method for assessing the risk of hypertension in residential residents affected by noise exposure according to claim 1, characterized in that, In step S1, basic personal information includes gender, age, occupation, length of residence, and body mass index; dietary habits include smoking, drinking, and exercise; architectural factors include window type, bedroom window orientation, seasonal window opening, window visibility to the street or greenery, floor level, and strength of standing wave resonance in the bedroom; the above data, as well as data on hypertension prevalence, were obtained through questionnaires or epidemiological surveys; noise exposure level data were obtained through noise simulation.

3. The method for assessing the risk of hypertension in residential residents affected by noise exposure according to claim 2, characterized in that, Preprocessing includes screening and preliminary treatment.

4. The method for assessing the risk of hypertension in residential residents affected by noise exposure according to claim 2, characterized in that, The verification in step S2 includes the chi-square test, Shapiro-Wilk test, and rank-sum test.

5. The method for assessing the risk of hypertension among residential residents affected by noise exposure according to claim 4, characterized in that, Chi-square tests were performed on the categorical variables according to formulas (1)-(3): （1） in, It is the total observation frequency of the i-th row; It is the total observation frequency of the j-th column; N is the total number of samples, which is the sum of the observation frequencies of all cells; (2) in, It is the observation frequency in the i-th row and j-th column; It is the expected frequency of the i-th row and j-th column; It is the chi-square statistic; （3） in, It is the number of row categories. It is the number of categories in the column. It's about freedom.

6. The method for assessing the risk of hypertension in residential residents affected by noise exposure according to claim 4, characterized in that, For continuous variables, the Shapiro-Wilk test is performed using formula (4): (4) in, n is the number of data items; It is the i-th observation in the data; It is the mean of the data; It is a constant related to the normally distributed data after sorting, and includes the number of data points. n The theoretical parameters for both normal and normal distributions were calculated.

7. The method for assessing the risk of hypertension in residential residents affected by noise exposure according to claim 6, characterized in that, For variables that do not satisfy a normal distribution, perform the Wilcoxon rank-sum test according to the following formula; （5） （6） in, and These are two sets of statistics, and ; The final verification statistic is: .

8. The method for assessing the risk of hypertension in residential residents affected by noise exposure according to claim 2, characterized in that, In step S3, a one-way logistic regression analysis is performed using the following formula: (7) (8) （9） （10） In step S4, multivariate logistic regression analysis is performed on variables with a significance level ≤ 0.05 according to the following formula: （11） （12） （13） （14） in, It is the probability that the dependent variable Y takes the value 1; It is the log odds of the logistic regression model, defined as the natural logarithm of the ratio of the probability that the dependent variable Y is 1 to the probability that the dependent variable Y is 0. It is the intercept term; It is the independent variable The regression coefficients; It is an independent variable of multiple factors; It is a linear combination of all independent variables; It is the dominance ratio of each independent variable; X is a single-factor independent variable; e It is the base of the natural logarithm; It is a linear combination in the logistic regression model, representing an estimate of the influence of the independent variable X on the dependent variable Y on a log-odds scale; Logit is short for logarithmic probability, which is the natural logarithm of the ratio of the probability of an event occurring to its probability of not occurring. OR It is the odds ratio, which represents the factor by which the logarithmic probability of an event increases when the independent variable X increases by one unit.

9. The method for assessing the risk of hypertension in residential residents affected by noise exposure according to claim 1, characterized in that, In step S5, the risk of developing hypertension is used as the dependent variable, and different combinations of other data are used as independent variables. Through stepwise regression, multiple factors are continuously modified to establish multiple assessment models for the risk of developing hypertension. Based on the analysis results of the assessment models, the predictive ability of the models, the hypothesis testing of the models, the goodness of fit, and combined with professional knowledge, the optimal model is selected from the multiple assessment models established.

10. A system for assessing the risk of hypertension among residents in a residential area due to noise exposure, characterized in that, The device includes a processor capable of executing a computer program that implements the method for assessing the risk of hypertension among residents affected by noise exposure, as described in any one of claims 1-9.

Images