Beyond Learning Styles: Rethinking How Birth Information Is Shared
26 May 2026
New population-level data show that approximately one in eight mothers in England now has a recorded diagnosis of gestational diabetes. Alongside this significant increase, existing evidence concerning diagnostic thresholds, ethnicity and individual metabolic risk raises important questions for maternity practice.
Gestational diabetes mellitus (GDM) is now one of the most commonly diagnosed complications of pregnancy.
A large population-based observational study led by the University of Edinburgh, with researchers from King's College London, analysed routinely collected NHS maternity data from more than 2.3 million mothers and approximately 2.8 million births across 184 hospitals in England between 2018 and 2022.
The study found that recorded GDM diagnoses increased from approximately 8% in 2018 to more than 12% in 2022 – an increase of around 60% over five years.
This equates to approximately one in eight mothers in England having a recorded diagnosis of GDM.
The increase warrants attention, particularly given the established associations between maternal hyperglycaemia and adverse maternal and neonatal outcomes.
However, when interpreting these findings, an important distinction should be maintained.
The study describes the prevalence of recorded GDM diagnoses within routinely collected NHS data. It does not represent universal prospective screening of the entire study population using a single standardised diagnostic protocol.
This is particularly relevant because the definition, screening and diagnosis of GDM remain areas of international clinical debate.
Diagnostic prevalence is not necessarily equivalent to biological prevalence
In England and Wales, NICE recommends a 75 g two-hour oral glucose tolerance test (OGTT) for women identified as being at increased risk of GDM.
NICE defines GDM as either:
- fasting plasma glucose ≥5.6 mmol/L; or
- two-hour plasma glucose ≥7.8 mmol/L.
These thresholds provide an essential framework for consistent clinical decision-making.
However, they should also be understood for what they are: defined clinical thresholds applied to a continuous biological variable.
The association between maternal glycaemia and adverse pregnancy outcomes does not appear to begin at one discrete glucose concentration.
The HAPO study demonstrated continuous associations between maternal glucose concentrations below those diagnostic of overt diabetes and outcomes including increased birthweight, cord-blood serum C-peptide and neonatal adiposity, without an obvious threshold at which risk suddenly appeared.
Consequently, any diagnostic definition of GDM necessarily establishes a point on a continuum at which additional clinical surveillance or intervention is considered appropriate. This distinction becomes important when interpreting prevalence.
Diagnostic criteria can substantially alter reported prevalence
There remains variation internationally in the thresholds used to diagnose GDM. This is more than a methodological consideration: the diagnostic criteria applied can materially alter the proportion of a maternity population classified as having the condition.
The Born in Bradford cohort provides an important UK example. Farrar and colleagues examined associations between maternal glucose and adverse perinatal outcomes among more than 10,000 White British and South Asian women.
When six different existing diagnostic criteria were retrospectively applied to the same population, estimated GDM prevalence varied substantially. Among White British women, prevalence ranged from approximately 1.2% to 8.7% depending upon the diagnostic definition applied.
Among South Asian women, the range was approximately 4% to 24%. These findings should not be interpreted as demonstrating that one set of criteria is inherently correct and another incorrect. They do, however, illustrate an important principle when interpreting epidemiological data: the prevalence of diagnosed GDM is partly dependent upon how GDM is defined.
Does a glucose threshold represent equivalent risk for every woman?
The Born in Bradford analysis also raised a second clinically important issue.
The relationship between maternal glucose concentrations and adverse outcomes was not identical between White British and South Asian populations. Researchers identified different glucose concentrations associated with a 75% increase in the relative risk of large-for-gestational-age birthweight and/or elevated infant adiposity.
For the composite outcome, the calculated fasting glucose thresholds were approximately 5.4 mmol/L for White British women and 5.2 mmol/L for South Asian women. A two-hour threshold of approximately 7.2 mmol/L was identified for South Asian women, while equivalent thresholds could not be identified for some outcomes among White British women.
This does not provide a basis for individual clinicians or maternity services to deviate from current NICE recommendations.
It does, however, contribute to the evidence that a single glucose measurement may not necessarily correspond to an identical level of absolute or relative clinical risk across all populations. This is particularly relevant within a diverse NHS maternity population.
Risk exists on a continuum
The clinical difficulty is readily illustrated around any diagnostic cut-off.
Under current NICE guidance, a fasting glucose of 5.6 mmol/L meets the diagnostic criterion for GDM, whereas a result of 5.5 mmol/L does not.
A threshold is required for clinical practice.
However, it would be biologically implausible to suggest that pregnancy risk changes abruptly as glucose increases by 0.1 mmol/L across that threshold.
Rather, the threshold identifies a point at which the evidence and guideline development process have determined that classification and intervention are appropriate.
For clinicians, this distinction between diagnosis and individual risk is important.
Diagnosis provides a pathway for care.
Risk describes the probability of an outcome.
The two are related, but they are not synonymous.
Can additional metabolic information improve risk stratification?
There is also increasing interest in whether additional biomarkers could improve the identification or stratification of metabolic risk during pregnancy.
Research involving Epsom & St Helier University Hospitals NHS Trust has examined first-trimester HbA1c as a predictor of subsequent GDM among women with moderate to severe obesity.
Balani and colleagues retrospectively examined women who subsequently underwent OGTT at 24–28 weeks.
The study identified an association between increasing first-trimester HbA1c and subsequent GDM and suggested that very low HbA1c values may identify a group at particularly low subsequent risk.
The study population was relatively small and selected, and the findings should therefore not be extrapolated to the wider maternity population.
Nevertheless, it contributes to a broader area of research examining whether risk stratification could eventually incorporate information beyond a single glucose threshold.
Other researchers have similarly explored whether the full glucose response during an OGTT can provide greater information about risk than categorisation based on individual cut-offs alone.
These approaches remain investigational, but they raise an important question for future maternity research:
Can we move from identifying the presence or absence of a diagnostic label towards more accurately quantifying individual metabolic risk?
Interpreting the new 'one in eight' finding
Against this background, the findings from the new Edinburgh-led study require careful interpretation.
The increase in recorded GDM is significant and may reflect a number of interacting factors.
These include genuine changes in population metabolic risk, including increasing maternal age and obesity; changes in ascertainment and screening; improved recording and data capture; and the diagnostic frameworks used within clinical practice.
The authors themselves acknowledge improvements in data capture as one possible contributor to the observed increase.
The study should therefore not be interpreted as evidence of either overdiagnosis or underdiagnosis.
Nor does it establish that one in eight pregnant women would necessarily meet GDM criteria if the entire maternity population underwent universal, prospective testing using an identical diagnostic protocol.
What it demonstrates clearly is that GDM now represents a substantial and increasing component of maternity care in England.
That has important implications.
Implications for maternity services
Increasing GDM prevalence affects both women and maternity services.
A diagnosis can lead to blood glucose monitoring, dietary intervention, additional antenatal appointments, pharmacological treatment where required, fetal growth surveillance, obstetric review and discussions regarding timing and mode of birth.
For women at increased risk of adverse outcomes, these interventions can be clinically important.
At service level, however, increasing numbers of women entering GDM pathways also create additional demand across midwifery, obstetric, diabetes, dietetic, ultrasound and laboratory services.
The objective should therefore not simply be to maximise or minimise the number of women receiving a diagnosis.
The clinical objective is greater precision:
identifying women and babies at increased risk, as early as practicable, and providing intervention proportionate to that risk.
Achieving this requires us to continue examining both the sensitivity of our diagnostic pathways and their specificity.
Why this matters as maternity becomes increasingly data-driven
There is a further implication for maternity research and the development of predictive analytics and artificial intelligence.
Healthcare datasets contain clinical diagnoses.
But a recorded diagnosis should not automatically be assumed to represent an objective biological ground truth.
A diagnosis is produced through a pathway involving eligibility for screening, attendance, the test performed, the diagnostic criteria applied and clinical recording.
This is particularly important when historical clinical datasets are used to develop predictive models.
If GDM diagnosis is used as an outcome variable, researchers need to understand:
- how consistently GDM was defined across the dataset;
- whether diagnostic criteria changed over time;
- which women were offered testing;
- whether access to and completion of testing differed between populations;
- whether missing diagnoses or differences in coding exist; and
- whether prediction of the diagnostic label is actually the clinically relevant outcome.
Without this consideration, there is a risk that predictive models become highly effective at reproducing existing clinical classifications without necessarily improving prediction of the maternal or neonatal outcomes we ultimately wish to prevent.
This distinction is particularly important in maternity care, where most women enter pregnancy without disease and where the challenge is to identify emerging risk without unnecessarily medicalising physiological pregnancy.
What does this mean for personalisation?
This evidence reinforces why personalisation in maternity care needs to extend beyond applying population-level thresholds.
At RealBirth, personalisation is about helping women and pregnant people understand evidence in the context of their own pregnancy, circumstances and choices. Through accessible, multilingual education and personalised learning, women can be better prepared to understand a diagnosis, discuss risk and participate meaningfully in decisions about their care.
This does not replace clinical assessment or established guidance. Rather, it supports an important principle of personalised maternity care: a diagnosis may determine a clinical pathway, but it should not define the individual.
As maternity becomes increasingly data-driven, personalisation should mean not only improving how we identify risk, but improving how we communicate and respond to it.
Towards more individualised assessment of risk
Current NICE guidance provides the appropriate framework for the diagnosis and management of GDM in NHS maternity services and should continue to underpin clinical practice.
At the same time, research should continue to examine whether our ability to stratify risk can become more sophisticated.
The new finding that approximately one in eight mothers in England now has a recorded diagnosis of GDM should therefore prompt more than a discussion about prevalence.
It should encourage us to consider what sits behind that figure.
How much represents changing population metabolic health?
How much reflects improved identification and recording?
How consistently are women being screened and diagnosed?
Do current thresholds identify equivalent levels of risk across different populations?
And can combinations of clinical, metabolic and demographic information eventually enable us to identify individual risk more precisely?
These questions do not undermine the importance of diagnosing and treating gestational diabetes.
They represent the next stage of understanding it.
As maternity care becomes increasingly data-driven, the opportunity is not simply to generate more diagnoses or more risk scores.
It is to become better at identifying which women and babies are most likely to benefit from additional surveillance and intervention – while continuing to protect physiological pregnancy wherever it is safe to do so.
References
- University of Edinburgh / King's College London (2026). Population-based observational analysis of routinely collected NHS maternity data examining gestational diabetes diagnoses and pregnancy outcomes in England, 2018–2022. BMJ Medicine. Study population included more than 2.3 million mothers and approximately 2.8 million births across 184 hospitals.
- National Institute for Health and Care Excellence (NICE). Diabetes in pregnancy: management from preconception to the postnatal period (NG3). Published 2015; last updated 2020; reviewed April 2025, with subsequent minor amendments.
- HAPO Study Cooperative Research Group (2008). Hyperglycemia and adverse pregnancy outcomes. New England Journal of Medicine, 358:1991–2002.
- Farrar D, Fairley L, Santorelli G, et al. (2015). Association between hyperglycaemia and adverse perinatal outcomes in South Asian and White British women: analysis of data from the Born in Bradford cohort. The Lancet Diabetes & Endocrinology, 3(10):795–804.
- Balani J, Hyer S, Johnson A, Shehata H. (2023). Predicting gestational diabetes mellitus by first trimester HbA1c: a retrospective study in women with moderate to severe obesity. Practical Diabetes. DOI: 10.1002/pdi.2466.
This article is intended for professional discussion and education. It does not recommend changes to current GDM screening, diagnostic or management pathways. Clinical practice should follow current NICE guidance and relevant local maternity and diabetes pathways.
