Cannabis in Postoperative Pain Management

Study Design This retrospective cohort study analyzes anonymized medical records of surgical patients who underwent surgery between January 2017 and January 2025 at the Assuta hospitals network.

Data Source Electronic health records from a hospital database, including postoperative pain scores, analgesic administration, and patient demographics. Pain levels will be assessed during hospitalization for up to one-week post-surgery. In the cannabis use research group, participants will be asked to report their daily use for at least the past six months. The study will utilize MDClone, a healthcare data analytics platform, to extract and analyze anonymized electronic health records. MDClone enables the generation of synthetic, privacy-preserving patient data, ensuring compliance with ethical and regulatory standards while allowing for robust statistical analysis.

Variables for Analysis

• Demographics: Age, sex, BMI, Hospital stay, Operation duration, type of anesthesia, region of residence, marital status.
• Medical History: Comorbidities, history of trauma, psychiatric conditions, prior surgeries.
• Surgical Data: Type of procedure, intraoperative factors, postoperative complications.
• Pain Management: Pain scores (e.g., VAS), opioid and non-opioid analgesic doses, use of regional anesthesia.
• Psychosocial Factors: psychiatric medication use (e.g., antidepressants).
• Hospital Course: Length of stay, ICU admissions

The study population The expected number of participants is 70,000 participants from the five medical canters in the Assuta network.

Statistical analysis include:

• Descriptive Analysis - Baseline characteristics will be summarized using means, medians, and proportions.
• Comparative Analysis - Pain levels and analgesic use across different surgical types, comorbidities and between cannabis users vs. non-users will be compared using t-tests, chi-square tests, or non-parametric equivalents.
• Machine Learning Models - Supervised learning algorithms (e.g., logistic regression, random forests, gradient boosting) will be employed to predict high analgesic requirements based on preoperative and intraoperative variables.
• Validation & Model Performance - ROC-AUC, sensitivity, and specificity will be used to assess model accuracy.

Inclusion criteria

• Patients aged 18 and over.
• Patients who underwent surgery under general anesthesia.

Exclusion criteria

• Minimally painful surgical procedures, including wrist and ankle tendon surgeries, minor rectal surgeries (e.g., fistula repair, rectal polyp removal), and minor gynecological procedures (e.g., vaginal procedures, transvaginal tape [TVT] insertion and transurethral procedures).
• Surgeries associated with potential neurological complications, such as craniotomy.
• Procedures involving percutaneous stent placement, including ureteral stent insertion.
• Incomplete pain assessment records
• Patients with severe cognitive impairments, affecting their ability to accurately report pain levels.
• Patients unable to express VAS scale.