Improving data sharing in research with context-free encoded missing data.

ABSTRACT: Lack of attention to missing data in research may result in biased results, loss of power and reduced generalizability. Registering reasons for missing values at the time of data collection, or-in the case of sharing existing data-before making data available to other teams, can save time and efforts, improve scientific value and help to prevent erroneous assumptions and biased results. To ensure that encoding of missing data is sufficient to understand the reason why data are missing, it should ideally be context-free. Therefore, 11 context-free codes of missing data were carefully designed based on three completed randomized controlled clinical trials and tested in a new randomized controlled clinical trial by an international team consisting of clinical researchers and epidemiologists w

SUBMITTER: Hoevenaar-Blom MP 

PROVIDER: S-EPMC5595279 | biostudies-literature | 2017

REPOSITORIES: biostudies-literature