The data sources are presented in a sequence that mirrors the drug development process:
1. Pre-clinical Studies (Animal Studies)
- Purpose: These are the very first tests, conducted before a drug is given to humans. The main goals are to establish safe dosage levels and to identify which organs might be harmed by the drug (e.g., liver, heart).
- What they study: Major organ toxicity, long-term (chronic) toxicity, potential to cause cancer (carcinogenicity), genetic mutations (mutagenicity), and birth defects (teratogenicity).
- Limitations: Their ability to predict human reactions is only “moderate.” Some serious side effects are specific to humans and won’t show up in animals (as was the case with the drug practolol). A lack of major toxicity in animals allows testing in humans to proceed, but it doesn’t guarantee safety.
2. Human Volunteer Studies (Phase I)
- Purpose: The first stage of human testing, usually in a small number of healthy volunteers (except for very toxic drugs, like those for cancer). The goals are to determine a safe dosage range and see how the body processes the drug.
- Context: Participants are monitored extremely closely in a controlled setting.
- Limitations: While generally safe, unexpected, severe reactions can occur, as shown by the TGN1412 case in 2006.
3. Clinical Trials (Phases II and III)
- Purpose: These are the formal studies in patients with the target disease. They are designed to prove both efficacy (does the drug work?) and safety.
- Key Features: They use methods like randomization (assigning patients to treatment groups by chance) and blinding (patients and doctors don’t know who gets which treatment) to minimize bias.
- Strengths: They are the “gold standard” for establishing that a drug causes a particular effect because the comparison groups are designed to be similar in all ways except the treatment they receive.
- Major Limitations:
- Size: They include thousands, not millions, of patients, so they cannot detect rare side effects.
- Duration: They are often too short to detect side effects that develop after long-term use.
- Population: Patients are highly selected (often excluding the elderly, pregnant women, or those with other illnesses), so the results may not reflect the “real world.”
- Conditions: Close monitoring in trials may not reflect how the drug is used in normal practice.
4. Post-Marketing Surveillance (Phase IV)
This is the most critical phase for pharmacovigilance. Once a drug is on the market and used by a large, diverse population under real-world conditions, new safety issues can be identified. This phase relies on two main methods:
A. Spontaneous ADR Reporting Systems
- Purpose: This is the classic “early warning system.” Its main job is to generate signals of potential new, previously unrecognized hazards.
- How it works: Healthcare professionals (and, increasingly, patients) voluntarily report their suspicion that a drug caused a harmful reaction. These reports are collected in large national and international databases.
- Strengths:
- Simple and universal (covers all drugs, all the time).
- Can capture suspicions that would otherwise be lost.
- Can rapidly identify potential problems.
- Major Limitations:
- Under-reporting: Only a fraction of all ADRs are ever reported.
- Uncertainty: A report is just a suspicion; it doesn’t prove the drug was the cause.
- Bias: Reporting is influenced by factors like how serious the reaction is, how new the drug is, and media publicity.
- No denominator: It’s often hard to know how many people used the drug without having the reaction, making it difficult to calculate the true risk.
B. Pharmacoepidemiological Studies
- Purpose: When a signal is raised from spontaneous reports, these formal studies are used to test the hypothesis (i.e., to investigate and quantify the potential risk).
- Nature: They are observational, meaning researchers observe what happens in normal practice without intervening, unlike in clinical trials.
- Key Study Designs:
- Cohort Study: Follows a group of people taking a drug and compares their outcomes to a group not taking it. Good for calculating absolute risk.
- Case-Control Study: Starts with people who have already experienced the health problem (the “cases”) and compares their past drug use to people who didn’t have the problem (the “controls”). Good for studying rare outcomes.
- Tools: These studies often use large, computerized databases that link prescription records to health outcomes (e.g., the UK’s General Practice Research Database).
- Other Methods:
- Prescription-Event Monitoring (PEM): A specific system in the UK where all events (not just suspected reactions) are recorded for a large cohort of patients using a new drug.
- Registries: Databases that track patients with a specific disease, exposure to a certain drug, or a specific outcome (e.g., a pregnancy registry for women taking a particular medication).
5. Systematic Reviews and Meta-Analysis
- Purpose: These are methods for combining and analyzing data from multiple existing studies to get a more precise and reliable overall answer.
- Systematic Review: A rigorous, structured process of finding and evaluating all relevant studies on a specific question.
- Meta-Analysis: A statistical technique that combines the numerical results of these studies to produce a single overall estimate of risk or benefit. This is increasingly important for evaluating drug safety by pooling data to find signals that might be missed in individual, smaller studies.
