Closed-loop pumps moved the children whose HbA1c would not move

Design

retrospective multicentre observational cohort across 13 NHS paediatric diabetes units, before-and-after comparison

Population

220 children and young people aged 19 or under with type 1 diabetes for over a year and HbA1c 69 mmol/mol or above at closed-loop initiation

Primary outcome

change in HbA1c after starting hybrid closed-loop therapy

Effect

mean HbA1c 82.5 mmol/mol at baseline, 72.8 at 3 months, 69.3 at 12 months (p<0.001); largest falls above 100 mmol/mol and in those switching from injections

Trials of hybrid closed-loop systems generally enrol engaged families with reasonable baseline control, which leaves open the question that matters most: does the technology help the child whose HbA1c has been stuck above 69 mmol/mol? This retrospective cohort across 13 NHS paediatric diabetes units followed 220 children and young people aged 19 or under with exactly that profile.

Mean HbA1c fell from 82.5 mmol/mol at baseline to 72.8 within three months and 69.3 at 12 months, holding stable after the initial drop (p<0.001). The largest falls were in those starting above 100 mmol/mol and in those moving from multiple daily injections rather than from an existing pump. Results were similar across pump systems, and diabetes-related hospital admissions fell.

What the design cannot tell you is how much of this is the pump. A retrospective before-and-after comparison in a group selected for high HbA1c will show improvement from regression to the mean alone, and starting closed-loop therapy brings education, contact and attention that a control group would not have had. The effect is almost certainly real and almost certainly smaller than 13 mmol/mol. Taken as an argument for who to offer the technology to, though, it points the right way — away from rationing it to the families already doing well.

• High HbA1c is not a reason to withhold closed-loop therapy; in this cohort it predicted the largest gain
• Expect most of the fall within three months — if it has not happened by then, review engagement rather than waiting
• Pump system choice did not matter; availability and family preference reasonably decide it
• Count hospital admissions as an outcome alongside HbA1c when making the funding case
• Access in India is almost entirely self-funded; be explicit about consumable costs before raising the option

Why it matters

It undercuts the rationing logic that reserves closed-loop technology for families who are already managing well.

Don't overread it

Retrospective, uncontrolled, and selected on a high baseline value — this describes what happened, not what the pump caused.

The statistics, in plain English

This is a before-and-after comparison with no control group, so part of the 13 mmol/mol fall is regression to the mean: children are selected at a high HbA1c, and high values tend to be followed by lower ones regardless of treatment. That the fall persisted to 12 months argues against it being entirely artefact, but the true effect is smaller than the headline number.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.