How Babyforecast calculates – and its limits

Which data are used?

These figures come from NICE NG207 (NHS England 2019–20) and describe only labours that eventually started spontaneously, not all pregnancies still ongoing today. The source reports weekly shares, which we spread equally across seven days for the chart – not measured daily values or a personal prediction.

How is Babyforecast Pro different?

The free calculator remains the unpersonalized NICE reference. Babyforecast Pro uses a separate model: age, early BMI, previous spontaneous labour onset and conception method can adjust the curve. A previous vaginal birth can contribute when earlier onset timings are unknown. The starting curve and the combination of study findings rely on model assumptions; they were not jointly estimated using individual trajectories or clinically validated. Greater accuracy has not been established.

From 37+0 weeks, Pro shows an outlook for up to seven days, including the selected day. A personal model estimate is still available only from 39+0 to 41+6 weeks and with confirmed prerequisites. Otherwise, a separate, explicitly nonpersonal historical NICE reference may appear. The assumed starting curve without personal factors is for comparison with the personal model only and is not the NICE distribution. Country and due-date dating do not change the percentage; clinical measurements such as cervical length or fFN are not used numerically in this model.

What do the three figures mean?

Daily and cumulative shares refer to all eventual spontaneous onsets. The conditional share divides the daily share by onsets still outstanding within that group. It is not the probability for a person still pregnant today, because induction and caesarean are missing as competing events.

Which assumption remains?

The source supplies weekly shares, not daily measurements. We divide the rounded shares by their sum of 99.9% and spread weeks 37 to 41 evenly across seven days each. Earlier shares are retained; from 42 weeks there is only an open-ended group without invented daily values. A displayed decimal place does not make this interpolation a precise prediction.

Which period does the result cover?

Free reference figures cover 37+0 to 41+6 weeks. Pro shows up to seven days, shortened at the end through 41+6 weeks. Every chart point uses the same period as the headline; the individual daily value is a separate detail. The personal research estimate remains limited to 39+0 to 41+6 weeks. Missing required details, planned interventions or certain reported conditions prevent a personal estimate; the reason is explained. Missing optional factors have no numerical effect. A historical reference is not a replacement personal prediction. A missing or low value says nothing about health or actual labour onset.

What applies to the account and email?

The separate Pro calculator does not automatically transfer inputs to your account. Saving is a separate, explicit step. In your account, the personal Pro model is used only after confirmation; email choices are separate. For the same day and saved details, your account and email use the same outlook and chart. Where a personal estimate is unavailable, any displayed historical reference remains explicitly nonpersonal; it does not activate the personal model.

What do tests establish?

Tests check the original weekly shares, calculation rules, date boundaries and agreement of each model’s values across calculator, preview, account and email. They do not replace external validation against individual trajectories. No confidence interval is calculated. The figures are not a basis for examination or treatment decisions.