About Prompt
- Prompt Type – Dynamic
- Prompt Platform – ChatGPT, Grok, Deepseek, Gemini, Copilot, Midjourney, Meta AI and more
- Niche – Drug Interaction Alert System
- Language – English
- Category – Pharmacovigilance
- Prompt Title – AI Prompt for Predicting Drug Interactions and Side Effects Based on Patient Data
Prompt Details
**Prompt Type:** Dynamic
**Target Audience:** All AI Platforms (e.g., large language models, machine learning models)
**Purpose:** Pharmacovigilance – predicting potential drug interactions and side effects based on individual patient data within a Drug Interaction Alert System.
**Prompt Structure:**
“`
## Drug Interaction and Side Effect Prediction
**Patient Data:**
* **Patient ID:** {patient_id}
* **Age:** {age}
* **Sex:** {sex}
* **Weight (kg):** {weight}
* **Height (cm):** {height}
* **Medical Conditions (ICD-10 Codes):** {list_of_icd10_codes}
* **Current Medications (RxNorm Codes with dosage and frequency):** {list_of_rxnorm_codes_with_details}
* **Allergies (RxNorm Codes):** {list_of_allergy_rxnorm_codes}
* **Lab Results (Name, Value, Unit):** {list_of_lab_results} *(Optional, include if available)*
* **Genetic Information (e.g., CYP2D6 genotype):** {genetic_information} *(Optional, include if available)*
**New Medication (RxNorm Code with proposed dosage and frequency):** {new_rxnorm_code_with_details}
**Tasks:**
1. **Drug Interaction Prediction:**
* Identify potential interactions between the new medication and the patient’s current medications.
* For each identified interaction, provide:
* **Interacting Drugs (Names and RxNorm Codes):** {drug1}, {drug2}
* **Interaction Type (e.g., pharmacodynamic, pharmacokinetic):** {interaction_type}
* **Severity (e.g., minor, moderate, major):** {severity_level}
* **Description of the interaction mechanism:** {interaction_mechanism}
* **Clinical Significance:** {clinical_significance} *(e.g., dose adjustment needed, monitoring required, contraindicated)*
* **Evidence Source (e.g., Micromedex, DrugBank, clinical trials):** {evidence_source}
* **Confidence Level (e.g., Low, Medium, High):** {confidence_level}
2. **Side Effect Prediction:**
* Predict potential side effects for the patient given their profile and the new medication.
* For each predicted side effect, provide:
* **Side Effect (MedDRA Preferred Term):** {side_effect_name}
* **Likelihood (e.g., very common, common, uncommon, rare, very rare):** {likelihood}
* **Severity (e.g., mild, moderate, severe):** {severity_level}
* **Onset (e.g., immediate, delayed):** {onset_time}
* **Actionability (e.g., monitoring required, intervention required):** {actionability}
* **Evidence Source:** {evidence_source}
* **Confidence Level (e.g., Low, Medium, High):** {confidence_level}
**Output Format:**
Return the results in a structured JSON format. Ensure the JSON structure is well-defined and easily parsable. Provide clear labels for each data field. Example Partial JSON Structure:
“`json
{
“patient_id”: “12345”,
“drug_interactions”: [
{
“drug1”: {
“name”: “Drug A”,
“rxnorm_code”: “1234”
},
// … (other interaction details)
}
],
“side_effects”: [
{
“side_effect_name”: “Nausea”,
// … (other side effect details)
}
]
}
“`
**Additional Instructions:**
* Prioritize accuracy and clinical relevance.
* If no interactions or side effects are predicted, clearly indicate “No interactions found” or “No side effects predicted.”
* Consider drug-drug-disease interactions.
* Consider the patient’s age, sex, weight, height, medical conditions, and other relevant factors when making predictions.
* If possible, quantify the uncertainty associated with predictions.
* Use standardized terminologies and coding systems (e.g., RxNorm, ICD-10, MedDRA) wherever possible.
* Clearly distinguish between known interactions/side effects based on established knowledge and predictions based on AI inference. Indicate the source of information (e.g., knowledge base, predictive model).
“`
This dynamic prompt provides a flexible framework for querying AI systems about drug interactions and side effects based on patient-specific data. By using standardized codes and a structured output format, it ensures interoperability and facilitates integration with Drug Interaction Alert Systems for effective pharmacovigilance. The inclusion of confidence levels and evidence sources allows clinicians to evaluate the reliability of the predictions and make informed decisions. The optional fields for lab results and genetic information allow for even greater personalization and accuracy as these data become more readily available. This detailed approach minimizes ambiguity and maximizes the clinical utility of the AI’s output.