Analysis Considerations for Integrated Summaries of Safety & Efficacy

Statisticians and statistical programmers are crucial team members in planning, producing and presenting integrated summaries. Planning well in advance for ISS/ISE is the first step in saving time and money. Statisticians will need to outline the specific issues that the integrated summary will need to cover and should attend pre-submission meetings with regulatory authorities to discuss the statistical analysis aspects.

The statisticians will produce a Statistical Analysis Plan while the programmers need to focus on integrated datasets. As CROS NT has pointed out in the past, centralizing data can be the first step towards reducing costs and time as it allows for traceability, data quality and standardization. A thorough gathering and review of existing information is important – especially if outsourcing to a vendor. Information such as protocols, Statistical Analysis Plans (SAPs), annotated CRFs, coding dictionaries, define.xml etc. need to be made available at the outset to avoid unnecessary delay.

Proper statistical planning also includes similar structures for datasets, for example, CDISC standards. Standardization is also applicable to reporting and using standard forms for tables, figures and listings.

Therefore, CROS NT lists the top analysis considerations Sponsors should think about when preparing Integrated Summaries of Safety & Efficacy:

– Utilize both positive and negative trials
– Comparie trials of similar designs (weighting of sample size; common covariates and stratifications)
– Make use of graphical presentations such as Forest Plots
– Consider controls, durations, patient populations, endpoints, drop-outs and statistical analyses
– Consider inconsistencies in the data
– Indicate denominator over time
– Always show gender (or subgroup) specific denominators
– For laboratory data, apply conversions where necessary to ensure same unit of measure for each parameter if multiple laboratories within/between studies
– Make documentation relating to individual laboratory reference ranges available
– Finally, schedule additional time to review all the data, information and knowledge you are supplying in support of the label. Use this time to pre-empt the questions from the regulatory authority.

CROS NT has particular expertise in biometrics services including statistical analysis and reporting, clinical data management and medical writing. CROS NT has completed over 1,000 studies in a wide range of therapeutic areas, and several study programs have included the preparation of Integrated Summaries for both the EMA and FDA. We have an expert team with particular expertise in trial design, data analysis and implementing and mapping CDISC standards.