The labor calculator's baseline model describes a first pregnancy with no adjusting factors. In that model, the largest share of spontaneous labor onset falls in the 40th week, at about 28% of all pregnancies. The 39th and 41st weeks hold about 19% and 24%. Weeks 37, 38, and 42 together account for about 22%, and labor before 37 weeks for 5%.
These are unconditional shares: the proportion of all pregnancies in which labor begins during each week, as seen from early in pregnancy. The figure most people want is different, and larger.
Two ways to state a weekly chance
The first is the share of all pregnancies that begin labor in a given week. The second is the chance that labor begins in the coming week given that it has not begun yet. The second number is always at least as large as the first, because the pregnancies that ended earlier are no longer in the denominator.
The table gives both for the baseline model. The "share" column is the unconditional figure. The "if still pregnant" column is the probability of labor within the next seven days for someone who reaches the start of that week without labor having begun.
| Week | Share of all pregnancies | If still pregnant at the start of the week |
|---|---|---|
| 37 | 2% | 2% |
| 38 | 8% | 9% |
| 39 | 19% | 23% |
| 40 | 28% | 43% |
| 41 | 24% | 64% |
| 42 | 12% | 88% |
The values for a later pregnancy are shifted earlier by about two days. Week 39 rises to 23% of all pregnancies and week 41 falls to 20%, and the "if still pregnant" chance at the start of week 40 is 48% rather than 43%.
Where the figures come from
The shares are read from a probability distribution built in two parts. The share before 37 weeks is 5%, rounded from the 4.5% spontaneous preterm rate among U.S. singleton births that Gyamfi-Bannerman and Ananth (2014) reported for 2012. The remaining 95% is a normal distribution over the term period. It is positioned so that the median of the complete distribution is day 284 for a first birth and day 282 for a later birth. Those are the values Smith (2001) reported for non-elective deliveries.
The width of the term distribution was set to match a late-pregnancy benchmark. Marquette and colleagues (2014) reported that among 15,253 uncomplicated pregnancies still ongoing at 41 weeks 0 days, 67.6% began labor spontaneously by 41 weeks 6 days. The baseline model's figure at the start of week 41 is 64% for a first pregnancy and 68% for a later one.
What the weekly figures leave out
The table describes spontaneous onset only. A pregnancy that ends by scheduled induction or planned cesarean is not represented in it. In the United States, more than a third of singleton births involve induction. Many care teams offer or recommend induction at 41 weeks. The week-42 figures therefore describe a situation that most pregnancies do not reach.
The figures also use no information beyond the questionnaire. Cervical examination findings and current contractions are outside the model. So are the reasons a care team might anticipate early or late labor. For a single pregnancy, those factors can move the chance further than any adjustment the calculator makes.
Adjustments
The answers on the form change the figures by small amounts. Smoking raises the before-37-weeks share from 5% to about 6%, and IVF or ICSI raises it to about 7%. A previous spontaneous birth at 35 or 36 weeks raises it to about 10%, and a previous spontaneous birth before 35 weeks to about 14%. Age and BMI move the term distribution by fractions of a day. The methodology page lists each adjustment with its source.
Using the table
The "if still pregnant" column answers the question most people are asking at a given appointment. At 39 weeks 0 days without labor, the baseline model gives about a one-in-four chance of labor in the coming week and about a three-in-four chance of reaching 40 weeks. At 40 weeks 0 days, the chance of labor in the coming week is slightly under one half.
The calculator reports these conditional figures directly for the current date, together with a 50% window and a 10% date. The weekly table on the result page uses the same distribution as this article, recalculated for the answers entered.