How to Read a Clinical Trial

The 6 things to check on every trial

1. Primary endpoint. What does the trial measure to declare success? "Improvement in overall survival vs. placebo" is very different from "improvement in tumor response rate at 8 weeks." Endpoint quality determines whether positive data is approval-quality or just headline-quality.
2. Statistical significance threshold. Most trials require p < 0.05 on the primary endpoint. Some sponsors play games — registering multiple endpoints, splitting populations — which inflates false-positive risk.
3. Sample size and power. Underpowered trials (too few patients) miss real effects. Over-powered trials waste capital. Look for the power calculation in the protocol.
4. Patient population. Inclusion/exclusion criteria define who's in the trial. A trial in heavily pretreated late-stage patients is different from a trial in newly diagnosed patients — and the FDA-approval implications differ significantly.
5. Comparator arm. Placebo vs. standard-of-care vs. active comparator. Beating placebo in a disease with effective standard-of-care is much weaker than beating standard-of-care.
6. Interim analyses + stopping rules. Trials that allow early stopping for efficacy or futility have different risk profiles. Pay attention to whether the data we're waiting for is interim or final.

Where to find this information

• ClinicalTrials.gov — every US-relevant trial is here. Look for protocol summary, primary outcome measures, eligibility criteria, sponsor.
• Company SEC filings (10-K, 10-Q, 8-K). Material trial details are disclosed; "definitive material" 8-Ks announce major data.
• Investor presentations. Company-formatted summaries — useful but biased; cross-check against ClinicalTrials.gov.
• Clinical Investor's Trial Translator — paste any ClinicalTrials.gov URL or trial protocol, get a plain-English breakdown of design, endpoints, risks, and what to watch for.

Common red flags

• Open-label trials (no blinding) for subjective endpoints
• Single-arm trials in indications where placebo response is high
• "Composite" endpoints that obscure which sub-component drove the result
• Interim analyses that conveniently hit "significance" exactly when the company needed cash
• Sponsor-funded trials with no independent data monitoring committee

Common false alarms (often look bad but aren't)

• Higher-than-expected adverse events that are mechanism-of-action-related and were anticipated in the protocol
• Slow enrollment in rare-disease indications
• Modifications to the statistical analysis plan that improve power
• Switching primary endpoint mid-trial — sometimes legitimate based on FDA feedback, sometimes a red flag; check the timing and rationale

Frequently Asked Questions

Where do I find a trial's protocol?

ClinicalTrials.gov has a summary for every registered trial. Full protocols are sometimes published in journals (NEJM, Lancet) or as supplementary appendices. Many sponsors release the full protocol after data readout.

What's the difference between phase 1, 2, and 3?

Phase 1: safety and dosing in healthy volunteers or small patient groups. Phase 2: preliminary efficacy and dose-finding. Phase 3: large pivotal trials designed to support FDA approval. See our full phase guide.

How do I know if a trial result will be approval-quality?

Approval-quality data hits the pre-specified primary endpoint at the pre-specified statistical significance level in the pre-specified patient population. "Hitting" any one of those is good news; hitting all three is approval news.

Is open-label necessarily bad?

Not always. For oncology trials with overall-survival endpoints (an objective measurement), open-label is fine. For pain or depression trials with patient-reported outcomes (subjective), open-label introduces serious bias.

Related guides

• PDUFA Dates: Investor's Guide
• How to Invest in Biotech
• Biotech Catalyst Trading Strategy
• Clinical Trial Phases (1, 2, 3) Explained
• How to Evaluate a Biotech Pipeline
• Biotech IPO Investor Guide
• FDA Approval Process Explained