Evidence-Based Prospective Study of the Artificial Intelligence Risk Assessment Model for Complication Prevention in Plastic Surgery
Evidence-Based Prospective Study of the Artificial Intelligence Risk Assessment Model for Complication Prevention in Plastic Surgery
The goal of this observational study is to determine if an AI-based risk assessment model can help prevent complications in plastic surgery patients by improving decision-making, providing recommendations to address risk factors, and assisting doctors in choosing the optimal timing and setting for elective plastic surgery. The study aims to answer if the AI model can effectively identify high-risk patients and what specific risk factors predict complications.
Purpose:
Evaluate the clinical effectiveness of an AI-based risk assessment model in preventing complications in plastic surgery patients by analyzing clinical data and patient history, providing personalized recommendations to mitigate risk factors and enhance outcomes.
Hypothesis:
The AI model can more accurately identify high-risk patients and provide effective recommendations to reduce complications compared to traditional methods.
Participants:
Individuals undergoing elective plastic surgery. They will complete an online form collecting data on age, height, weight, smoking habits, and comorbidities. The system calculates risk scores, BMI, and Caprini scores.
Study Procedures:
Risk assessment using the AI model, which evaluates multiple factors and generates personalized recommendations, including weight management, smoking cessation, blood pressure control, Doppler ultrasound for DVT, nutritional consultations, and specialist referrals. Recommendations are reviewed and approved by plastic surgeons.
Follow-Up:
The follow-up period ranges from 2 to 41 months, with a mean of 15 months. Data on patient outcomes, including complication rates and satisfaction, will be collected and analyzed.
Outcomes Measured:
Incidence of complications, the accuracy of the AI model in predicting complications, and its impact on improving surgical outcomes.
Impact:
The study aims to provide insights into AI use in plastic surgery, leading to better risk assessment tools and protocols, enhancing preoperative planning, postoperative care, and patient safety and satisfaction.
Study Design and Procedures:
The goal of this observational study is to determine if an AI-based risk assessment model can help prevent complications in plastic surgery patients by improving decision-making, providing recommendations to address risk factors, and assisting doctors in choosing the right time and setting for elective plastic surgery.
Methodology:
The study was conducted from January 2021 to May 2024, involving 3,347 patients assessed using the AI risk assessment model in a solo practice setting. The model uses an algorithm to evaluate clinical data and patient history, calculate risk scores, highlight risk factors, and generate personalized recommendations.
Patient Assessments:
Data Collection: Participants complete an online form collecting data on various clinical factors, including BMI, age, Caprini score, smoking habits, and gender.
Risk Calculation: The algorithm calculates risk scores, flags abnormal values, and screens for Body Dysmorphic Disorder (BDD) using the Body Dysmorphic Disorder Questionnaire (BDDQ).
Risk Categorization: Patients are categorized into low, moderate, or high-risk groups based on their risk scores.
Personalized Recommendations: The model generates specific recommendations for each patient based on their risk assessment, including:
Weight Management: Target weight recommendations for patients with BMI ≥ 25.1. Smoking Cessation: Advice for patients who smoke to quit. Blood Pressure Control: Recommendations for hypertensive patients to monitor blood pressure daily and consult a cardiologist.
Deep Vein Thrombosis (DVT) and Varices: Suggestions for Doppler ultrasound of the lower limbs 24 to 72 hours before surgery for patients with DVT, coagulopathies, varices, or Caprini score ≥ 8.
Comorbidities: Recommendations for cardiac echocardiogram and stress tests for patients over 50 with hypertension or vascular pathology antecedents.
Nutritional Guidance: Advice for patients with BMI ≤ 25.1 to consult a nutritionist and undergo screening for eating disorders with psychiatric consultation.
Specialist Referrals: Suggestions for consultations with endocrinologists, bariatric surgeons, cardiologists, hematologists, or other specialists according to comorbidities findings, and psychiatric referrals according to BDD screening.
Data Management:
Data collected from participants are anonymized and stored securely to protect patient privacy and confidentiality. The data management process includes:
Anonymization of Sensitive Data: All personally identifiable information (PII) is removed or masked to ensure that participants' identities are protected. Anonymized data is used for analysis to maintain confidentiality.
Quality Assurance Plan: Regular data validation and registry procedures, including site monitoring and auditing, to ensure data integrity.
Data Checks: Consistency checks for data fields and predefined rules for range. Source Data Verification: Comparison of registry data with external sources, such as medical records, to assess accuracy and completeness.
Data Dictionary: Detailed descriptions of each variable, including source, coding, and normal ranges.
Standard Operating Procedures (SOPs): Procedures for patient recruitment, data collection, management, analysis, adverse event reporting, and change management.
Statistical Analysis:
Statistical and inferential analyses were performed using a Colab notebook to ensure robustness and reproducibility. The primary outcome measure was the incidence of complications in each risk group. Secondary outcome measures included the correlation between risk factors and complications.
Sample Size Assessment:
The sample size was determined to ensure sufficient power to detect differences in complication rates between the risk groups.
Plan for Missing Data:
Procedures were in place to address missing data, including imputation techniques and sensitivity analyses.
Statistical Analysis Plan:
The analysis plan included descriptive statistics, correlation analyses, and regression models to evaluate the relationship between risk factors and complications.
Inclusion Criteria:
Exclusion Criteria: