Stool test kits and a trained guide got more people screened for bowel cancer

The review

1. 01

   Who took part

   The 73 trials covered 366,766 people, all in US clinics. They were adults at average risk of bowel cancer.

2. 02

   What they did

   Each trial tried a way to get people screened and compared it with usual care. Ways included giving stool test kits, a trained guide to help people past barriers, reminders, and teaching about screening. Some trials combined several parts.

3. 03

   What happened

   Stool test kits got about 22 more people in every 100 screened than usual care did. A trained guide got about 18 more in every 100, and reminders about 3 more, but the paper's guide counts don't match. A guide did not clearly help more people get a colonoscopy after an abnormal test.

What the review found

1. Helped

   Across 17 studies, more people who were given a stool test kit got screened than usual care did.

   Pooled analysis of 17 studies with 45,836 people · vs people who got usual care · behaviour was recorded · the authors rated their certainty high · ⚠ signs that small negative studies went unpublished; the studies disagreed a lot

2. Helped

   Across 16 studies, people who had a trained guide to help them got screened more than with usual care.

   Pooled analysis of 16 studies · vs people who got usual care · behaviour was recorded · the authors rated their certainty high

3. Helped

   Across 8 studies, reminders helped more people get screened than usual care did, though the gain was small.

   • Rule: Get a remindertested

   Pooled analysis of 8 studies with 33,339 people · vs people who got usual care · behaviour was recorded · the authors rated their certainty moderate · ⚠ signs that small negative studies went unpublished

4. Helped

   Across 48 studies, programmes with more than one part got more people screened than programmes with a single part.

   Pooled analysis of 48 studies with 163,570 people · vs programmes with a single part · behaviour was recorded · the authors rated their certainty moderate · ⚠ signs that small negative studies went unpublished

Source