手法を比較

	メタアナリシス症例対照研究	前向き症例対照研究
分野	疫学	疫学
系統	Process / pipeline	Process / pipeline
提唱年	1980s–2000 (formalized with MOOSE reporting guidelines in 2000)	1970s–1990s (formalized alongside nested case-control methods)
提唱者	Systematic development attributed to multiple epidemiologists; MOOSE guidelines formalized by Stroup et al.	Evolved from classical retrospective case-control methodology; prospective embedding attributed to modern epidemiological practice (Rothman, Greenland, and others, late 20th century)
種類	Observational study synthesis	Observational analytic study design
原典	Shapiro, S. (1994). Meta-analysis/Shmeta-analysis. American Journal of Epidemiology, 140(9), 771-778.	Rothman, K. J., Greenland, S., & Lash, T. L. (2008). Modern Epidemiology (3rd ed.). Lippincott Williams & Wilkins. ISBN: 978-0781755641
別名	pooled case-control analysis, case-control meta-analysis, meta-analytic case-control design, systematic pooled case-control	prospective case-control design, ambidirectional case-control, bidirectional case-control, nested case-control (prospective variant)
関連	4	6
概要	A meta-analytic case-control study systematically identifies, critically appraises, and quantitatively synthesizes data from multiple independent case-control studies examining the same exposure-disease relationship. By pooling odds ratios across studies, it yields a more precise and generalizable estimate of association than any single study can provide, while formally quantifying heterogeneity across populations, settings, and study periods.	A prospective case-control study embeds the case-control logic within a defined cohort followed forward in time. Cases are identified as they occur, rather than looked up in records after the fact, and controls are sampled from the same prospectively monitored base population. This forward-looking approach allows collection of exposure data before outcome ascertainment, reducing recall bias — the principal weakness of the classic retrospective case-control design — while retaining the efficiency gains of sampling controls rather than enrolling a full cohort.