Sample size and outcomes

Sample size

Samples size calculations are essential in randomized controlled trials to ensure the study’s statistical power is high enough to detect a true effect, if there is one, whilst also minimizing the margin of error.

In studies involving both pregnant and non-pregnant populations, power calculations tend to be based on the broader population and not specifically to pregnant women. From the outset, sub-group analysis on pregnant populations should be planned and powered accordingly to ensure conclusions on efficacy can be drawn specifically to pregnant and lactating women. It should be noted that even if a study is not powered to demonstrate efficacy in pregnancy and during the breastfeeding period, an assessment of safety may be demonstrated.

Key resources

ICH E9 Statistical principles for clinical trials

Example power/sample size calculation tool, provided in the context of conducting research on antiretrovirals in pregnancy

See also the toolkit on vaccines.

 

SPECIAL CONSIDERATIONS: LACTATION

In lactation pharmacology, there are many situations with no prior data upon which to base some of the necessary assumptions. In this scenario, it may be helpful to incorporate an interim analysis into the design plan to enable assumptions to be checked and to inform any necessary amendments to the protocol.

Outcomes

Standardized comprehensive core outcomes with harmonized definitions that consider both mothers and infants and include a mix of safety and efficacy outcomes that should be used by researchers. This helps to facilitate data sharing and meta-analysis to enable comparability across diverse populations, including those with co-infections such as HIV or malaria.

A WHO interdepartmental task force co-produced a core data set of essential indicators for surveillance and health monitoring of maternal and newborn health. This core data set is called the WHO minimum maternal and newborn health data set and comprises 15 indicators to be used for clinical trials as well as during active and routine surveillance in all studies or projects that pertain to maternal and newborn health. Specialized core outcomes sets exist that are specific to different conditions and time points in pregnancy. Notable resources for specialized core outcome sets include the International Consortium for Health Outcomes Measurement (ICHOM).

Careful consideration must be given to the method of evaluating an outcome (e.g., blood test, imaging) to ensure that those used do not pose risks during pregnancy. Use of universal pregnancy-specific normal lab ranges and harmonized and clear definitions (e.g., WHO’s International statistical classification of diseases and related health problems. Tenth revision [ICD-10]) enable accurate outcome assessment.

Pregnancy-related outcomes should be evaluated both individually and in the context of the disease’s known effects on pregnancy (e.g., congenital malformations linked to diabetes). Understanding how a treatment affects both the disease and pregnancy outcomes is essential. Extended follow up of the woman and any baby should occur beyond the end of pregnancy to ensure capture of any safety events.

Consideration of how post-trial data is protected through storage and access is essential in order to protect women. This is especially important in settings where access to reproductive health measures is limited, criticized, or even criminalized.

Key resources

The WHO minimum maternal and newborn health data set (pending)

International Consortium for Health Outcomes Measurement (ICHOM)

International statistical classification of diseases and related health problems. Tenth revision [ICD-10]

See also the toolkit on vaccines for more on the harmonization of outcome assessment in immunization.