Gestational diabetes — structured oral station (12 minutes)

Opening prompt — "Interpret her test for me."

Model response — lead with the classification, then the number that makes it: ^[1]

• "This is gestational diabetes. Her one-hour value of 11.2 mmol/L is at or above the ADIPS 2025 threshold of 10.6 mmol/L. One abnormal value is enough."
• "Her fasting of 5.1 and two-hour of 8.7 are both below their thresholds of 5.3 and 9.0, so the one-hour sample is the only reason we have a diagnosis — which is exactly why the one-hour sample is worth taking."
• "This is not overt diabetes in pregnancy: that needs fasting 7.0 or more, two hours 11.1 or more, or HbA1c 6.5 percent or more."^[1]

Examiner is listening for: the correct threshold set, the specific abnormal value, and confident exclusion of overt diabetes. ^[1]

Probe 1 — "Under the old criteria, would she have been diagnosed?"

• "Yes, and on two values. The superseded ADIPS 2014 and IADPSG criteria were fasting 5.1, one hour 10.0, two hours 8.5 — her fasting and one-hour values both cross those. ADIPS raised the thresholds in 2025 to the HAPO odds-ratio 2.0 tier."^[1]
• "The honest point is that no threshold is biologically correct. HAPO showed a continuous relationship between glucose and outcome with no inflection, so every cut-point is a policy trade-off between over-diagnosis and missed benefit."^[1]

Probe 2 — "Defend raising the thresholds. Isn't that just under-diagnosing?"

This is the discriminating probe. Argue both sides and land somewhere. ^[2][8]

• "For raising them: GEMS randomised 4061 New Zealand women to lower or higher criteria. Lower criteria diagnosed 15.3 percent versus 6.1 percent, yet large-for-gestational-age rates were 8.8 versus 8.9 percent — no difference — with more induction, more pharmacotherapy and more neonatal hypoglycaemia in the lower-criteria group."^[2]
• "ScreenR2GDM said the same thing about screening strategy: one-step doubled diagnoses to 16.5 percent versus 8.5 percent, with no difference in large-for-gestational-age infants, the perinatal composite, hypertensive disease or caesarean."^[8]
• "Against raising them: within GEMS, the women whose values fell between the two criteria and who were treated did better than those who were not. So the intermediate band is not a null group — it is a group we now identify differently, mainly through risk-based early testing and clinical surveillance."^[2]
• "My position: I follow ADIPS 2025 because RANZCOG endorses it and it is the ANZ standard, and I keep a low threshold to escalate surveillance in a woman like this one who has every risk factor."^[1]

Probe 3 — "What are your targets, and what will you actually do this week?"

• Targets, said as numbers: "Fasting 5.0 or less, one hour after meals 7.4 or less, two hours 6.7 or less."^[3]
• This week: "Diabetes educator and dietitian this week, meter and strips today, testing fasting plus one or two hours post-meal, and a written plan. Medical nutrition therapy with carbohydrate spread across three meals and two to three snacks, lower glycaemic index choices, and a walk after her largest meal."^[1]
• Escalation trigger: "If a meaningful proportion of readings are above target after one to two weeks, I start metformin per the local formulary (starting low, uptitrating to the local maximum), or bedtime isophane insulin if the fasting value is the problem."^[5]
• "I would tell her that in the MiG trial nearly half the women on metformin also needed insulin, so needing insulin later is not a failure — it is the expected path for many women."^[5]

Probe 4 — "Why bother treating at all? She had it last time and the baby was fine."

• "ACHOIS answered that. Treating gestational diabetes reduced serious perinatal complications from 4 percent to 1 percent, with an adjusted relative risk of 0.33."^[4]
• "It is not free: ACHOIS also increased induction from 29 to 39 percent and neonatal nursery admission from 61 to 71 percent, and TARGET showed that tighter targets increased serious maternal morbidity from 3.0 to 5.9 percent while reducing the serious infant composite from 2.6 to 1.3 percent."^[4][3]
• "So I treat, and I say out loud that treatment buys fewer big babies and fewer birth injuries at the price of more monitoring and more intervention."^[4][3]

Probe 5 — "When and how will she give birth?"

• "If she stays diet-controlled with normal fetal growth, I would aim for birth by around 40+6 weeks. If she needs metformin or insulin, I would plan birth at about 38 to 39 weeks."^[6]
• "Growth ultrasound every 3 to 4 weeks from 28 to 30 weeks, watching the abdominal circumference trajectory rather than a single estimated fetal weight."^[6]
• "In labour, hourly capillary glucose targeting 4 to 7 mmol/L. A woman on diet alone rarely needs an insulin infusion; I start a variable-rate insulin infusion with dextrose if she is persistently above 7 mmol/L or is fasting on substantial insulin doses."^[6]
• "Her previous 4.15 kg baby matters. I would discuss shoulder dystocia risk honestly and document the conversation, including the wide error margin on ultrasound estimated fetal weight."^[6]

Probe 6 — communication and cultural safety: "She says 'is it bad this time?' Talk to her."

Three of the eight domains are being scored here. Say the words out loud. ^[1]

• Sit at her level, use her name, check whether she would like a support person or an interpreter present, and ask what she already knows.
• "Your sugar test shows gestational diabetes again. That means your body needs help managing sugar while you are pregnant. It is common, it is treatable, and most women who do the plan have a healthy baby."
• "Two things matter most: eating in a way that keeps the sugar steady, and checking it four times a day so we can see if you need tablets or insulin. I am not going to leave you to work this out alone — you will see a diabetes educator and a dietitian this week."
• Acknowledge that Pasifika women carry a higher burden of gestational diabetes and worse outcomes, and that this is about access and services, not about her. Offer a community-based or telehealth model if travel is a barrier.^[1]
• Check understanding by asking her to say the plan back, not by asking "does that make sense?"

Probe 7 — "What happens after she goes home with the baby?"

• "Stop all diabetes medication at birth, check her glucose before discharge, and book a 75 g two-hour oral glucose tolerance test at 4 to 12 weeks postpartum, then screening every 1 to 3 years for life."^[6]
• "Her relative risk of later type 2 diabetes is about tenfold, and much of that risk lands in the first five years — so this test is not paperwork, it is the intervention."^[7]
• "I would also expect neonatal hypoglycaemia: early feeding, protocol glucose monitoring, buccal dextrose gel rather than a reflex nursery admission, and keeping them together."^[6]