Methodology

How the labour probabilities are calculated

Last updated 20 October 2026

Every probability on this site comes from one statistical model of when spontaneous labour starts. The model is a Johnson SU distribution fitted to NHS and CDC birth statistics, truncated at 42 weeks, and shifted by a few days for personal factors with published evidence. All displayed odds are conditional: they assume you are still pregnant today.

Statistical model

The model uses a Johnson SU distribution for the probability of spontaneous labour onset. Its heavy left tail captures preterm births that a normal or skewed-normal curve underestimates, while the peak and right tail match term and late-term births.

Heavy left tail: models early preterm births that simpler distributions miss.
Flexible shape: fits both the preterm tail and the term peak.
One distribution: simpler than a mixture model, with a better fit to the calibration targets.

Model against NHS data

Bars: all births in England 2024-25 by gestational week, including inductions and caesareans. Line: the model's distribution of spontaneous labour onset.

The NHS bars peak at 39 weeks because a large share of births are induced or scheduled. The model describes spontaneous onset, which is more spread out and peaks later, at 40 to 41 weeks. That is the quantity that matters when you are planning around natural labour.

Data sources

Model parameters were fitted to match official maternity statistics:

Model calibration

The fitted curve meets these targets for spontaneous labour onset:

10%

Preterm, before 37 weeks

52%

Born by the due date

80%

Born by 41 weeks

100%

Born by 42 weeks

The 42-week cut-off

The distribution is truncated at 42 weeks because almost no spontaneous labour is recorded beyond that point. Pregnancies that reach 42 weeks are, in practice, induced.

Conditional probability

All displayed probabilities are conditional on not having delivered yet. At 39 weeks, for example, only the part of the distribution from 39 weeks onward is considered, and the odds for each remaining day are rescaled to add up to 100%.

P(labour on day X | still pregnant) = P(labour on day X) / P(labour after today)

Personal factors

Each factor shifts the whole curve by a fixed number of days. The sizes come from these studies:

Factor	Effect on timing	Source
Parity	First pregnancies run about 2 days later	Bergsjø et al. 2001, BJOG
Fetal sex	Boys arrive about 1 day later (280.6 vs 279.8 days)	Divon et al. 2002, AJOG
BMI	Higher BMI is linked with later delivery, up to 5 days	Denison et al. 2008, BJOG
Interpregnancy interval	Gaps under 12 months are linked with earlier delivery	Shachar et al. 2017, Obstet Gynecol
Plurality	Twins about 3 weeks earlier (50% preterm); triplets about 7.5 weeks earlier	Murray et al. 2018, TOG
Maternal age	Under 25 or over 40: slightly earlier (small effect)	Population data

Limitations

The probabilities describe spontaneous labour onset in a population. They do not account for:

scheduled inductions or caesarean sections;
medical conditions that change timing;
individual variation beyond the modelled factors;
the limited accuracy of individual prediction in research (c-statistic about 0.56).