From: Blinding of study statisticians in clinical trials: a qualitative study in UK clinical trials units
Overview of statistician roles | |||||
Role name | Description | ||||
Trial statistician | Typically responsible for day-to-day statistical input into the design, conduct and reporting of the trial. Roles include (but are not limited to): • Input into database and CRF design • Producing regular monitoring reports and liaising with trial committees (e.g. TMGs, DMCs and TSC) • Planning and conducting the interim (if applicable) and final statistical analysis | ||||
Lead statisticiana | Leads on and has overall responsibility for statistical aspects of the trial. Responsible for oversight of the TS including advice and quality control. Often a co-applicant involved in the grant funding application | ||||
Non-blinded Statistician | A statistician (not the trial or lead) statistician that is able to access data by allocation. The main purpose of an unblinded statistician is to enable trial and lead statisticians to remain blinded where possible | ||||
Statistical models | |||||
Role name | Blinding | Meeting attendancec | |||
TMG | DMC (o) | DMC (c) | TSC | ||
Model 1—Trial statistician can be unblindedb to maintain blind of lead statistician | |||||
Trial statistician | Blinded at outset but can be unblinded to preserve blinding of lead statistician | (✓) | ✓ | ✓ | (✓) |
Lead statistician | Remains blinded until database lock | ✓ | ✓ | ✓ | |
Model 2—TS remains blinded, lead statistician (LS) can be unblindedb | |||||
Trial statistician | Remains blinded until database lock. | ✓ | ✓ | ✓ | |
Lead statistician | May become unblinded (if required) to maintain the TS blind. If not previously unblinded, they are unblinded at database lock | (✓) | ✓ | ✓ | (✓) |
Model 3—TS and LS remain blinded by having a third unblinded statistician | |||||
Trial statistician | Remains blinded until database lock | ✓ | ✓ | ✓ | |
Lead statistician | Remains blinded until database lock | ✓ | ✓ | ✓ | |
Non-blinded Statistician | Unblinded | (✓) | |||
Model 4—TS and LS remain ‘pseudo-blinded’ by using coded groupsd | |||||
Trial statistician | Remains “blinded” to true treatment allocations until database lock | (✓) | ✓ | ✓ | (✓) |
Lead statistician | Remains “blinded” to true treatment allocations until database lock | (✓) | ✓ | ✓ | (✓) |
Model 5—Both statisticians can be unblindedb during the trial | |||||
Trial statistician | May become unblinded if required. If not during the trial, they are unblinded at database lock | (✓) | ✓ | ✓ | (✓) |
Lead statistician | May become unblinded if required. If not during the trial, they are unblinded at database lock | (✓) | ✓ | ✓ | (✓) |
Model 6—All statisticians are unblindedb | |||||
Trial statistician | Unblinded | (✓) | ✓ | ✓ | (✓) |
Lead statistician | Unblinded | (✓) | ✓ | ✓ | (✓) |