Decentralised trial

A conventional trial happens at a site: an academic centre where the patient travels for consent, dosing, scans, blood tests and questionnaires. A decentralised trial moves as many of those steps as possible to where the patient lives, using electronic consent, telehealth visits, home nursing, courier delivery of oral drugs, local laboratories and imaging, and electronic patient-reported outcomes. Fully virtual trials exist for simple oral drugs; most oncology trials that use these tools are hybrid, keeping infusions and protocol scans at the site while decentralising consent, follow-up visits and symptom capture. The pandemic forced the change, and regulators in the United States and Europe have since issued guidance accepting decentralised elements provided oversight, data integrity and safety reporting are preserved.

The case for the design is who gets into trials. A few percent of adults with cancer join one, and they are younger, whiter, wealthier and closer to academic centres than the population with the disease. Distance is a leading reason eligible patients decline, and it falls hardest on rural, older and poorer patients. Decentralised trials also collect more of the outcomes patients care about, because a weekly symptom questionnaire on a phone captures what a three-monthly clinic visit forgets. In the corpus, BWEL delivered its two-year weight-loss intervention to more than 3,000 women through 42 coaching calls by telephone, and CHALLENGE ran a three-year exercise programme through community physical activity consultants; neither is labelled decentralised, but both moved the intervention out of the hospital.

The limits are the parts of a cancer trial that cannot travel. Infusions of investigational drugs, protocol biopsies and central imaging review still need sites; local scans read by different radiologists add noise to progression endpoints; adverse events noticed at home must reach the investigator quickly; and the digital tools themselves exclude patients without connectivity or confidence with them. The design is a set of options to be chosen per protocol step, not a switch, and its value shows up in enrolment speed and diversity rather than in the headline result.