Session 3
Missing Data
Incomplete Data
Sally C. Morton
Department of Biostatistics
Member
PCORI Methodology Committee
Co-Investigator
“Doing research differently”
Car analogy:
Patients moving into the front
seat of research
• Define terms
• Consider cultures
• Standard statistical approaches may
not work…but it’s worth thinking
about them
• Incentives, feasibility and impact
• Standards?
PCORI Methodology Report: 47 Standards
• Formulating questions (6) • Patient centeredness (4)
• Data integrity and Rigorous Analyses (6) • Causal inference (6)
• Missing data (5)
• Heterogeneity of treatment effects (4) • Data networks (2); Data registries (3) • Adaptive and Bayesian trial design (5) • Studies of diagnostic tests (5)
• Systematic reviews (1)
• “Regardless of study design, proposals must adhere to all
relevant PCORI methodology standards”
Nov. 19, 2013 –
PCORI today issued its revised PCORI Methodology Report,
which offers minimal requirements for following best practices in the conduct of scientifically valid patient-centered outcomes research.
PCORI Methodology Standards
• Minimum requirements
• Reflect best practices
• Supported by scientific evidence
• Are used to assess PCORI proposals
• Not meant to be comprehensive
• Means to educate and influence?
• Potential to stifle innovation?
“Doing methods differently”
Reframe how we conceptualize
and tackle issues of missing,
incomplete, and inconsistent data
