Related Experiment Videos
Multivariate models of self-reported health often neglected essential candidate determinants and methodological
Georgios D Mantzavinis1, Noula Pappas, Ioannis D K Dimoliatis
1Department of Hygiene and Epidemiology, University of Ioannina, School of Medicine, Ioannina 45110, Greece.
Journal of Clinical Epidemiology
|April 23, 2005
Summary
Self-reported health (SRH) models often lack appropriate determinants and exhibit methodological flaws. Future research should focus on comprehensive candidate lists and rigorous multivariate modeling for accurate SRH assessment.
Area of Science:
- Health outcomes research
- Biostatistics
- Epidemiology
Background:
- Self-reported health (SRH) is a key indicator of well-being.
- SRH is influenced by numerous, diverse parameters.
- Understanding these determinants requires robust modeling.
Purpose of the Study:
- To evaluate the range of determinants used in SRH models.
- To identify methodological issues in multivariate SRH studies.
- To improve the quality of SRH research.
Main Methods:
- Medline search for articles published in 2002.
- Included studies with SRH as an outcome and other variables as determinants.
- Excluded studies focused on specific diseases.
Main Results:
- 56 of 1,991 articles were eligible.
- Multivariate models were used in 91% of eligible studies.
- 133 unique determinants were identified, with a median of 7 per study.
- Significant methodological issues were found: overfitting (10%), non-linear gradients (29%), lack of interaction tests (63%), unspecified coding (49%), and variable selection (29%).
Conclusions:
- SRH models require comprehensive determinant lists.
- Methodological rigor in multivariate modeling is crucial.
- Addressing identified issues will enhance SRH model validity.