Use of reproducible research practices in public health: A survey of public health analysts.

Abstract

Use of reproducible research practices improves the quality of science and the speed of scientific development. We sought to understand use of reproducible research practices in public health and associated barriers and facilitators. In late 2017, we surveyed members of the American Public Health Association Applied Public Health Statistics section and others; 247 of 278 who screened eligible answered the survey, and 209 answered every applicable question. The survey included questions about file management, code annotation and documentation, reproducibility of analyses, and facilitators and barriers of using reproducible practices. Just 14.4% of participants had shared code, data, or both. Many participants reported their data (33%) and code (43.2%) would be difficult for colleagues to find if they left their institution. Top reported barriers to using reproducible practices were data privacy (49.8%) and lack of time (41.7%). Participants suggested training (50.9%) and requirements by journals (44.4%) and funders (40.2%) to increase use of reproducible research practices. Increasing use of reproducible research practices is important for public health and requires action from researchers, training programs, funders, and journals.