Interactive network calculator and method for predicting metatarsal bone fracture postoperative functional recovery condition

Abstract

The invention relates to an interactive network calculator and method for predicting the functional recovery condition after metatarsal bone fracture operation, and the computer comprises a data collection module which is used for obtaining the prediction factor data of a subject, and the prediction factor data comprises not limited to age and BMI; the model construction module is used for constructing a model based on the data acquired by the data acquisition module and performing analysis and calculation so as to calculate the metatarsal bone fracture postoperative functional recovery probability and 95% CI of the subject; the interactive page calculator application program development module is used for developing an interactive network calculator based on a result of the model building module; and the model evaluation module is used for evaluating the distinction degree, the calibration degree and the clinical net income of the model. The method can provide a rapid and accurate prediction tool for clinicians, assists decision making, is helpful for identifying high-risk patients, formulates targeted intervention measures, and improves treatment compliance. Installation is not needed, and cross-platform use can be achieved; the model can be updated according to new data, and prediction accuracy is kept.

Description

Technical Field

[0001] The invention relates to the field of rehabilitation assessment, and in particular to an interactive network calculator and method for predicting the functional recovery of a subject after metatarsal fracture surgery. Background Art

[0002] Metatarsal fractures are a common type of fracture, accounting for 3.2% to 6.8% of all fractures, with an annual incidence of 67-75.4 cases per 100,000 people. This type of fracture not only affects the patient's walking ability, but may also cause long-term pain and dysfunction, seriously affecting the quality of life. However, there is currently a lack of effective prediction tools to evaluate the postoperative functional recovery of patients with metatarsal fractures. Therefore, the development of an interactive network prediction model and the use of it to develop an interactive network calculator for predicting the postoperative functional recovery of subjects with metatarsal fractures are of great significance for formulating personalized treatment plans and optimizing patient management. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a fast and accurate prediction tool, an interactive network calculator and method that can be used to predict the functional recovery of a subject after metatarsal fracture surgery.

[0004] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0005] The interactive network calculator for predicting the postoperative functional recovery of metatarsal fractures of the present invention comprises a data acquisition module, a model building module, an interactive page calculator application development module and a model evaluation module;

[0006] The data acquisition module is used to obtain data including but not limited to age and BMI of the subject, and use the data to build a model and perform analysis and calculation, so as to calculate the probability of functional recovery after metatarsal fracture surgery of the subject and 95% CI;

[0007] The model building module uses LASSO regression to screen key predictive factors and build a clinical prediction model for functional recovery after metatarsal fracture surgery;

[0008] The interactive page calculator application development module uses R Shiny to build the user interface of the web page. Users can enter data through controls such as drop-down menus. The server is written in R language to process user input and perform calculations to finally generate results.

[0009] The model evaluation module uses C-index to evaluate model discrimination, adopts Hosmer-Lemeshow test and calibration chart to evaluate model calibration, and evaluates clinical net benefit through decision curve analysis (DCA).

[0010] The interactive network calculator for predicting postoperative functional recovery of metatarsal fractures of the present invention, wherein the data collection module includes: telephone follow-up and medical record query function modules, which are used to collect preoperative factors, intraoperative factors and postoperative factors. The prediction model also includes converting the subject's age, education, BMI, smoking, topography, fracture classification, injury type, postoperative infection, rehabilitation training and other predictive variables into categorical variables, and using the categorical variables as predictive variables to calculate the probability of postoperative functional recovery of the subject's metatarsal fracture.

[0011] The preferred model construction module uses LASSO regression to screen key predictive factors and incorporates them into a binary logistic regression model to derive independent risk factors associated with postoperative recovery of metatarsal fractures. The interactive page calculator application development module uses R Shiny to build the user interface of the web page and uses R language to write server-side logic. The model evaluation module uses C-index to evaluate model discrimination, uses the Hosmer-Lemeshow test and calibration chart to evaluate model calibration, and evaluates clinical net benefit through decision curve analysis (DCA). The prediction model is obtained by collecting preoperative, intraoperative and postoperative factor data of a large number of patients with metatarsal fractures, using LASSO regression to screen key predictive factors, and incorporating significant predictive factors into a binary logistic regression model for construction.

[0012] The calculator also includes an interactive calculator application development method, which includes the steps of building a dynamic prediction model, designing a user interface, and writing server-side logic to achieve real-time processing of user input data and instant feedback of prediction results.

[0013] The present invention also provides a method for using the above-mentioned interactive network calculator to predict the functional recovery after metatarsal fracture surgery, which comprises the following steps:

[0014] (a) Input the subject’s personal data;

[0015] (b) Calculate the probability and 95% CI of functional recovery after metatarsal fracture surgery for the subjects based on the input data and the constructed model;

[0016] (c) Display calculation results and provide immediate feedback.

[0017] The predictive factors screened out using LASSO regression in step (a) include, but are not limited to, gender, age, ethnicity, occupation, marital status, education, body mass index (BMI), blood type, season, terrain, smoking, drinking, injury cause, fracture classification, and common diseases of patients before surgery. The binary logistic regression model in step (b) is used to evaluate the independent risk factors associated with postoperative recovery of metatarsal fractures and calculate the weight of each factor.

[0018] Preferably, the clinical prediction model constructed in step (c) includes but is not limited to the following steps:

[0019] AO / OTA classification was performed based on the patient's imaging data to classify the fracture type;

[0020] Patients were divided into different groups according to age, such as 0--20 years, 21-40 years, 41-60 years, and 60 years;

[0021] The key predictors were converted into categorical variables for model calculation.

[0022] Preferably, the internal verification and external verification in step (d) include but are not limited to the following steps:

[0023] Use the training set for internal validation;

[0024] Use a validation set for external validation;

[0025] ROC curve, calibration plot and DCA curve were used to evaluate the discrimination, calibration and clinical effectiveness of the model.

[0026] Compared with the prior art, the interactive network calculator for predicting postoperative functional recovery of metatarsal fractures of the present invention has at least the following advantages:

[0027] 1. It can provide clinicians with a fast and accurate prediction tool, while assisting decision making to help identify high-risk patients, develop targeted intervention measures, and improve treatment compliance.

[0028] 2. Users can enter personal data (such as age, BMI, place of residence, whether there is infection after surgery, etc.), and the calculator can provide personalized risk prediction of functional recovery after metatarsal fracture surgery based on a statistical model of national data.

[0029] 3. Shiny apps provide instant feedback, allowing users to quickly understand their health status and adjust their lifestyle or seek medical advice based on the calculated results, thereby improving the public's awareness of diseases and self-management capabilities.

[0030] 4. Regular use of this calculator can help track changes in health status, identify people with potential poor recovery, and take preventive measures or adjust treatment strategies in a timely manner.

[0031] In addition, it has a user-friendly interface design and is easy to operate. It has an instant feedback mechanism to provide intuitive prediction results. It can be a web-based application that does not require installation and can be used across platforms. The model can be updated based on new data to maintain prediction accuracy.

[0032] The interactive network calculator for predicting postoperative functional recovery of metatarsal fractures of the present invention will be further described below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 To develop a nomogram for predicting functional recovery after metatarsal fracture surgery;

[0034] Figures 2A and 2B are the ROC curves of the training cohort A and validation cohort B of the prediction model for functional recovery after metatarsal fracture surgery, respectively;

[0035] Figure 3A , Figure 3B , Figure 3C These are the calibration plots of the modeling group (A), Bootstrap group (B), and external validation group (C);

[0036] Figure 4A , Figure 4B They are the training cohort A and the validation cohort B of the DCA curve of the functional recovery prediction model after metatarsal fracture surgery;

[0037] Figure 5 Establishing a flow chart for the interactive web calculator of the present invention for predicting functional recovery after metatarsal fracture surgery in a subject;

[0038] Figure 6 Example of a web-based interactive calculator application using R Shiny (where Fig. 6A :For web page icons, Figure 6B , Figure 6C and Fig.6D are the probabilities of functional recovery under different conditions). DETAILED DESCRIPTION

[0039] The interactive network calculator for predicting the functional recovery of subjects after metatarsal fracture surgery of the present invention mainly includes a data acquisition module, a model building module, an interactive page calculator application development module and a model evaluation module. The data acquisition module is used to obtain data such as the age and BMI of the subject, and then the above information obtained in the data acquisition module is used to build a model and perform analysis and calculation, so as to calculate the probability of functional recovery of the subject after metatarsal fracture surgery and 95% CI. The specific implementation example is as follows:

[0040] 1. Data collection and processing:

[0041] First, through telephone follow-up and medical record inquiry, the following research contents were collected:

[0042] Preoperative factors: gender, age, ethnicity, occupation, marital status, education, body mass index (BMI), blood type, season, terrain, smoking, drinking, injury cause, fracture classification, common diseases of patients before surgery (diabetes, hypertension, coronary heart disease, anemia, old cerebral infarction, respiratory system diseases, hepatobiliary system diseases, hypoproteinemia, hyperlipidemia, hyperuricemia, hypocalcemia, hypokalemia);

[0043] Intraoperative factors: surgical fixation method, surgical operation level and anesthesia method;

[0044] Postoperative factors: length of hospital stay, postoperative infection, and rehabilitation training.

[0045] Then, the patients were divided into 4 groups according to age: 0-20 years, 21-40 years, 41-60 years, and 60 years. Fractures were classified according to the AO / OTA classification of metatarsal fractures and judged based on the patients' imaging data.

[0046] 2. Model construction (R v.4.3.0 R Foundation for Statistical Computing, Vienna, Austria, SPSS v.23.0 IBM, Armonk, NY, USA) was performed as follows:

[0047] (1) LASSO regression was used to screen key predictive factors, and a total of 17 significant predictive factors were screened out;

[0048] (2) Significant predictors were included in the binary logistic regression model to obtain independent risk factors associated with postoperative recovery of metatarsal fractures

[0049] (3) Constructing a clinical prediction model for functional recovery after metatarsal fracture surgery

[0050] (4) Perform internal validation (using the training set) and external validation (using the validation set)

[0051] 3. Interactive page calculator application development (R shiny):

[0052] (1) After building the dynamic prediction model, obtain the web page key and the right to run the web page

[0053] (2) Design the user interface (UI), including input controls and result display area

[0054] (3) Write server-side logic to process user input and generate prediction results

[0055] 4. Model evaluation (R v.4.3.0 R Foundation for Statistical Computing, Vienna, Austria):

[0056] (1) Use C-index to evaluate model discrimination

[0057] (2) Use the Hosmer-Lemeshow test and calibration plot to assess model calibration

[0058] (3) Evaluating clinical net benefit through decision curve analysis (DCA)

[0059] Multivariate analysis revealed that factors such as living in mountainous areas, smoking, obesity, and rehabilitation training were associated with better postoperative recovery, while low education level, advanced age, high-energy injury, living in mountainous areas, infection, and anemia were identified as detrimental to postoperative functional recovery. A nomogram was drawn based on the results of multivariate logistic regression, see Figure 1 Shown is a nomogram for the prediction model of functional recovery after metatarsal fracture surgery.

[0060] As shown in Figures 2 and 3 , Figures 2A and 2B are the ROC curves of the training cohort A and validation cohort B of the functional recovery prediction model after metatarsal fracture surgery, respectively; Figure 3A , Figure 3B , Figure 3C The calibration graphs are for the modeling group (A), Bootstrap group (B), and external validation group (C). Through comprehensive evaluation, internal validation, and external validation, the discrimination, calibration, and clinical effectiveness of the model were evaluated using ROC curves, calibration graphs, and DCA curves. According to the ROC curve (Figure 2), the C index of the modeling group was 0.832 (95% CI 0.7798-0.8835), and the C index of the external validation group was 0.838 (95% CI 0.7472-0.929). In the calibration graphs of the modeling group, internal validation group, and external validation group (Figure 3), the prediction curve fits well with the reference line, indicating that the risk of poor prognosis predicted by the model is consistent with the actual risk of poor prognosis. See Figure 4, where Figure 4A , Figure 4B They are the training cohort A and the validation cohort B of the DCA curve of the prediction model for functional recovery after metatarsal fracture surgery. According to the DCA curves of the modeling group and the external validation group, when the threshold probability is in the range of 0.72-1, the net benefit of the model is higher than that of all treated populations and all untreated populations, indicating that the model has the best clinical effectiveness within this range, and taking treatment measures has a higher net benefit.

[0061] Combined with the above analysis results, please see Figure 5, is a flowchart of establishing an interactive web calculator of the present invention, which is based on R shiny to establish an interactive web calculator for predicting the probability and 95% CI of functional recovery after metatarsal fracture surgery in subjects. The interactive web calculator application development (R shiny) for predicting functional recovery after metatarsal fracture surgery in the present invention is specifically as follows:

[0062] It uses R Shiny's fluidPage function to build a web page user interface, where users can enter data through controls such as drop-down menus. Define how to handle user input and calculation logic. The server is written in R language, which processes user input and performs calculations to generate results. The calculator uses Shiny to automatically respond to changes in user input using reactive expressions and observers. When the user enters data, the relevant reactive expressions are automatically recalculated and the output is updated. Shiny creates data binding through the rsconnect function, so that user input and calculation results are synchronized in real time. For example, changes in the value of the input box will trigger related calculations and update the result display area. The calculation results are displayed on the user interface through the render function. These functions bind the calculation results to the user interface to ensure dynamic updating of the results.

[0063] The process of establishing an interactive web calculator for predicting the functional recovery of a subject after metatarsal fracture surgery of the present invention includes:

[0064] 1. Obtaining the right to run

[0065] After building the dynamic prediction model, obtain the web page key and the authority to run the web page.

[0066] 2. Design the User Interface

[0067] Includes input controls (such as drop-down menus, etc., for entering various data such as age, education, BMI, etc.) and result display areas

[0068] 3. Write server-side logic

[0069] Written in R, when the user enters data, the relevant reactive expressions are automatically recalculated and the output is updated. Shiny uses reactive expressions and observers to automatically respond to changes in user input, and creates data bindings through the rsconnect function, so that user input and calculation results are synchronized in real time.

[0070] In this system, the receiver operating characteristic (ROC) curve is used to detect age, education level, BMI, smoking, topography, fracture classification, injury type, postoperative infection, and rehabilitation training, and the predictive variables are converted into categorical variables, which are then used as predictive variables to calculate the probability of functional recovery after metatarsal fracture surgery in the subjects.

[0071] In the module for calculating the probability of functional recovery after metatarsal fracture surgery, the present invention converts the age of the subject into four categorical variables, namely, the age of the subject is 0-20 years old, the age of the subject is 21-40 years old, the age of the subject is 41-60 years old, and the age of the subject is 60 years old and above.

[0072] In the module for calculating the probability of functional recovery after metatarsal fracture surgery, the subject's educational level was converted into a 5-category variable, with the categories being the subject's educational level of primary school or below, the subject's educational level of junior high school, the subject's educational level of high school, and the subject's educational level of college or above.

[0073] In the module for calculating the probability of functional recovery after surgery for metatarsal fractures, the subject's BMI was converted into a 4-category variable, namely, the subject's BMI was less than 18.5, the subject's BMI was 18.5-23.9, the subject's BMI was 24-27.9, and the subject's BMI was 28 and above.

[0074] In the module for calculating the probability of functional recovery after metatarsal fracture surgery, the subjects' smoking, postoperative infection, and rehabilitation training were converted into binary variables, namely, yes or no.

[0075] In the module for calculating the probability of functional recovery after metatarsal fracture surgery, the terrain where the subject lives was converted into a 3-category variable, namely, the subject lives in a plain, the subject lives in a hilly area, and the subject lives in a mountainous area.

[0076] In the module for calculating the probability of functional recovery after surgery for metatarsal fractures, the fracture type of the subjects was converted into a 3-category variable, namely, the subjects were type A, the subjects were type B, and the subjects were type C.

[0077] In the module for calculating the probability of functional recovery after metatarsal fracture surgery, the subject's injury type was converted into a binary variable, namely, the subject had a high-energy injury and the subject had a low-energy injury.

[0078] As shown in the figure, an example of a web-based interactive calculator application using R Shiny of the present invention is shown (where Fig. 6A For web page icons, Figure 6B , Figure 6C and Fig.6DThe probability of functional recovery under different conditions is respectively): It can calculate the probability and 95% CI of postoperative functional recovery of metatarsal fracture in different genders by inputting clinical characteristics, see Figure 6 (A). For example, for individuals aged 21-40 years old, college degree or above, BMI of 18.5-23.9, non-smokers, living in the plains, type A fracture, low energy injury, no postoperative infection, no rehabilitation training, and no anemia, selecting "prediction" calculates the probability of good postoperative recovery to be 0.862 (95% CI 0.709-0.941), see Figure 6 (B). For individuals aged 41-60 years old, with less than primary school education, BMI of 18.5-23.9, smokers, living in the plains, type B fracture, low energy injury, postoperative infection, no rehabilitation training, and no anemia, the probability is 0.706 (95% CI 0.370-0.908), see Figure 6 (C). For individuals aged 60 years or older, with a junior high school education or less, a BMI below 18.5, smoking, living in mountainous areas, type B fractures, low-energy injuries, postoperative infections, no rehabilitation training, and anemia, the calculated probability of a good recovery was 0.1464 (95% CI 0.0088-0.768), see Figure 6 (D).

[0079] In summary, the present invention combines LASSO regression with R Shiny for the first time, integrates multiple predictive factors, develops a prediction model for postoperative functional recovery of metatarsal fractures, realizes the visualization and interactive application of complex statistical models, and provides personalized prediction results. In addition, the interactive web calculator can provide instant prediction results without waiting for a doctor's appointment or professional evaluation. The evaluation and prediction are based on the specific health data and environmental data entered by the user, which is more in line with personal circumstances than general tools. The complex risk assessment process is simplified, user-friendly, and simple to operate. It can help users pay attention to and manage their postoperative health care more conveniently.

[0080] The embodiments described above are merely descriptions of preferred implementation modes of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. An interactive network calculator for predicting postoperative functional recovery of metatarsal fractures, characterized in that: include: A data collection module, used to obtain predictive factor data of the subject, wherein the predictive factor data includes but is not limited to age and BMI; A model building module is used to build a model based on the data acquired by the data acquisition module and perform analysis and calculation to calculate the probability and 95% CI of functional recovery after metatarsal fracture surgery of the subject; An interactive web calculator application development module for developing interactive web calculators based on the results of the model building module; The model evaluation module is used to evaluate the model's discrimination, calibration, and clinical net benefit.

2. The interactive network calculator for predicting postoperative functional recovery of metatarsal fractures according to claim 1, characterized in that: The data acquisition module comprises: Telephone follow-up and medical record inquiry function modules are used to collect preoperative factors, intraoperative factors, and postoperative factors.

3. The interactive network calculator for predicting postoperative functional recovery of metatarsal fractures according to claim 1, characterized in that: The model building module uses LASSO regression to screen key predictive factors and incorporates them into the binary logistic regression model to obtain independent risk factors associated with postoperative recovery after metatarsal fracture.

4. The interactive network calculator for predicting postoperative functional recovery of metatarsal fractures according to claim 1, characterized in that: The interactive web calculator application development module uses R Shiny to build the user interface of the web page and uses R language to write the server-side logic.

5. The interactive network calculator for predicting postoperative functional recovery of metatarsal fractures according to claim 1, characterized in that: The model evaluation module uses C-index to evaluate model discrimination, adopts Hosmer-Lemeshow test and calibration chart to evaluate model calibration, and evaluates clinical net benefit through decision curve analysis (DCA).

6. A method for predicting postoperative functional recovery of metatarsal fractures using an interactive network calculator as claimed in any one of claims 1 to 5, characterized in that: The following steps are involved: (a) Input the personal data of the subject; (b) Calculate the probability and 95% CI of functional recovery after metatarsal fracture surgery for the subjects based on the input data and the constructed model; (c) Display calculation results and provide immediate feedback.

7. The method according to claim 6, characterized in that The predictive factors screened out using LASSO regression in step (a) include, but are not limited to, gender, age, ethnicity, occupation, marital status, education level, body mass index (BMI), blood type, season, terrain, smoking, drinking, injury cause, fracture classification, and common diseases of patients before surgery.

8. The method according to claim 6, characterized in that The binary logistic regression model in step (b) is used to evaluate the independent risk factors associated with postoperative recovery of metatarsal fractures and calculate the weight of each factor.

9. The method according to claim 6, characterized in that The clinical prediction model constructed in step (c) includes but is not limited to the following steps: AO / OTA classification was performed based on the patient's imaging data to classify the fracture type; Patients were divided into different groups according to age, such as 0--20 years, 21-40 years, 41-60 years, and 60 years; The key predictors were converted into categorical variables for model calculation.

10. The method according to claim 6, characterized in that The internal verification and external verification in step (d) include but are not limited to the following steps: Use the training set for internal validation; Use a validation set for external validation; ROC curve, calibration plot and DCA curve were used to evaluate the discrimination, calibration and clinical effectiveness of the model.

Images