This Article Free Full Text (Free PDF) Free References Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Saved Citations Download to citation manager Request Permissions Request Reprints Add to My Marked Citations Citing Articles Citing Articles via Google Scholar Citing Articles via Scopus Google Scholar Articles by Jenkins, P.L. Articles by Wall, S. Search for Related Content PubMed PubMed Citation Articles by Jenkins, P.L. Articles by Wall, S. Pubmed/NCBI databases Medline Plus Health Information Alcohol Social Bookmarking                         What's this?

The Norsjö-Cooperstown healthy heart project: A case study combining data from different studies without the use of meta-analysis

The Mary Imogene Bassett Research Institute, Cooperstown, New York, USA, paul.jenkins{at}bassett.org

L. Weinehall

Epidemiology, Department of Public Health and Clinical Medicine, UmeÅ University, UmeÅ, Sweden

T.A. Erb

The Mary Imogene Bassett Research Institute, Cooperstown, New York, USA

C. Lewis

The Mary Imogene Bassett Research Institute, Cooperstown, New York, USA

A.N. Nafziger

The Mary Imogene Bassett Research Institute, Cooperstown, New York, USA, Clinical Pharmacology Research Center & Department of Medicine, Bassett Healthcare, Cooperstown, New York, USA

T.A. Pearson

Department of Community & Preventive Medicine, University of Rochester Medical Center, Rochester, New York, USA

S. Wall

Epidemiology, Department of Public Health and Clinical Medicine, UmeÅ University, UmeÅ, Sweden

Objectives: This paper aims to develop and describe a method for combining, comparing, and maximizing the statistical power of two longitudinal studies of risk factors for cardiovascular disease that did not have identical data collection methodologies.

Methods: Subjects from a 1986 cross-sectional study (n= 180) were pair-matched with subjects of corresponding gender and age (+ 5 years) from a 1990 cross-sectional study. The methodology is described and results are calculated for various measures of cardiovascular risk or risk factors (e.g. cholesterol, Finnish Risk Score).

Results: Box's test of equality and symmetry of covariance matrices gave chi-square values of 223.8 and 710.0 for two cardiovascular risk factors (cholesterol and cardiac risk score, respectively); these values were highly significant (p=0.0001). For the North Karelia Risk Score, repeated measures ANOVA revealed a borderline significant interaction for treatment by time (p=0.054) and a significant interaction for treatment by time by country (p=0.035). These probabilities compared favorably with a randomized blocks model.

Conclusions: Creation of a synthetic longitudinal control group resulted in a statistically valid ANOVA model that increased the statistical power of the study.

Key Words: statistical • meta-analysis • epidemiology • evaluation.

Scandinavian Journal of Public Health, Vol. 29, No. 56 suppl, 40-45 (2001)
DOI: 10.1177/14034948010290021701


CiteULike    Complore    Connotea    Del.icio.us    Digg    Reddit    Technorati    Twitter    What's this?