From diabetes.co.uk/news
- A data-driven model suggested people at higher risk may benefit from shorter follow-up intervals when HbA1c rises, while stable lower-risk patients could safely have longer gaps.
- A separate study found the first appointments following a diabetes-range HbA1c were particularly important opportunities to document a diagnosis.
- Both studies used US healthcare data and neither proves that algorithm-based scheduling would improve NHS outcomes.
Seeing a GP more frequently is not automatically better for everyone with type 2 diabetes, according to research investigating how primary-care appointments could be scheduled according to individual risk.
Two studies from researchers at the University of Massachusetts Amherst examined different parts of the type 2 diabetes pathway: how frequently people already diagnosed with diabetes should be reviewed, and how quickly an abnormal HbA1c result leads to a documented diagnosis.
The findings suggest that the organisation of primary care may matter alongside individual patient characteristics.
But they do not support simply replacing existing appointment schedules with an algorithm.
Should higher-risk patients be seen more frequently?
The first study used electronic health records from 10 primary-care practices to develop a mathematical model for choosing follow-up intervals.
Rather than automatically reviewing everyone after one, three or six months, the model adjusted the recommended interval according to the person’s current health and diabetes trajectory.
People in the higher-risk group tended to be older and have more existing conditions such as chronic kidney disease and cardiovascular disease.
When HbA1c increased, the model favoured more frequent follow-up in these patients.
For lower-risk people whose glucose remained stable, it allowed longer intervals between appointments.
The model performed better on its mathematical optimisation measure than fixed appointment schedules, including a benchmark based on American Diabetes Association recommendations.
The basic idea is straightforward.
Seeing someone again too soon may provide little new information because HbA1c reflects blood glucose over approximately the previous two to three months.
Waiting too long, however, could allow worsening glucose control to go undetected.
A personalised schedule tries to find the point between those two extremes.
This was a modelling study – not a trial of patient outcomes
That distinction is important.
Researchers did not randomly assign thousands of people to algorithm-selected appointment intervals and then show fewer diabetes complications.
Instead, they used existing electronic health records to build and evaluate a Markov decision model.
Its better performance refers to the model’s predefined optimisation criteria.
It does not yet prove that the approach reduces hospital admissions, kidney disease, cardiovascular complications or mortality in real clinical practice.
A prospective trial would be needed before drawing those conclusions.
A second study examined delays in diagnosis
A separate study investigated 594 adults attending two US primary-care clinics between 2017 and 2023.
All had recorded an HbA1c in the diabetes range – at least 6.5%, equivalent to 48 mmol/mol – before receiving a documented diagnosis of type 2 diabetes.
The median time from the first diabetes-range HbA1c to diagnosis differed between the two clinics, at approximately 1.6 months and two months.
The investigators found that the first few appointments following the abnormal HbA1c appeared to be particularly important opportunities for diagnosis.
However, this finding needs nuance.
After adjustment for other factors, the study did not find statistically significant associations between primary-care location, continuity of care or visit regularity and time to diagnosis.
It would therefore be too strong to conclude that staffing levels or the clinic someone attends have been proven to cause diagnostic delays.
The study was exploratory and involved only two practices.
The researchers argue that differences in staffing, appointment availability and local practice processes deserve further investigation.
Why might diagnosis be delayed?
An HbA1c result does not operate in isolation.
Clinicians may need to repeat tests, review symptoms, rule out factors that can distort HbA1c and consider whether another form of diabetes is possible.
Diabetes.co.uk has guides explaining how diabetes is diagnosed and how the HbA1c test is used in type 2 diabetes.
The healthcare system also has practical constraints.
Appointments may be missed or delayed, laboratory results need reviewing and primary-care teams differ in workload and staffing.
That makes the period immediately following an abnormal result an obvious place to examine whether patients are being lost between testing and treatment.
What could this mean for the NHS?
The studies were conducted using US healthcare data, so the results cannot simply be transferred to NHS general practice.
The UK has different diagnostic pathways, staffing structures, diabetes reviews and commissioning arrangements.
But the underlying question is relevant.
People with type 2 diabetes are highly heterogeneous.
Someone with stable glucose, few complications and a consistent treatment regimen may not need exactly the same follow-up frequency as someone whose HbA1c is rising rapidly alongside kidney or cardiovascular disease.
More appointments are not automatically better if they add workload without changing care.
Equally, leaving a higher-risk person too long between reviews may delay medication changes or recognition of complications.
The potential value of the research is therefore risk-based follow-up, rather than simply increasing or reducing appointment numbers for everyone.
Before an algorithm like this could be used in the NHS, it would need prospective UK validation and evidence that it improves clinically meaningful outcomes without widening inequalities or making access harder for people who need support.
The studies make a sensible case for personalising follow-up.
They do not yet prove the best schedule for doing it.


