Item – Theses Canada

​
OCLC number
46496231
Author
DiCenso, Alba,1952-
Title
Systematic overviews of the prevention and predictors of adolescent pregnancy.
Degree
Ph. D. -- University of Waterloo, 1995
Publisher
Ottawa : National Library of Canada = Bibliothèque nationale du Canada, 1995.
Description
4 microfiches.
Notes
Includes bibliographical references.
Abstract
Adolescent pregnancy is a significant social and health problem. Although much has been published about the topic, there remains continuing uncertainty about effective approaches to its prevention. This project consisted of two systematic overviews, the first, of the available evidence regarding the effectiveness of primary prevention strategies in delaying sexual intercourse, in improving consistent birth control use, in improving responsible sexual behaviour, and in reducing the incidence of pregnancy in the adolescent population; the second, of the available evidence regarding the predictors of initiation of sexual intercourse, consistent birth control use, and pregnancy. Retrieval strategies included searches of personal files, computerized literature searches of 11 databases, review of reference lists of articles and published reviews, hand searches of key journals, and consultation with experts. This resulted in 1,034 papers that were rated for relevance using defined eligibility criteria. Twenty-five studies were identified for inclusion in the evaluation overview and 38 studies were identified for inclusion in the predictor overview. These studies were rated for methodologic quality and relevant data were abstracted from them. In the evaluation overview, the meta-analysis strategy consisted of calculating fixed effects summary odds ratios, using logistic regression to explain heterogeneity, calculating random effects summary odds ratios, and comparing results when analyses were restricted to randomized controlled trials only. Generally, the evaluation overview demonstrated that adolescent pregnancy prevention programs have not had a significant impact on the outcomes of interest in females or in males. For all outcomes for both females and males, the summary odds ratio for cohort designs were consistently more in favour of the intervention group. For the predictor meta-analysis, a preliminary analysis strategy is described that permits calculation of odds ratios for each predictor studied. This approach permits inclusion of non-significant and partially reported findings in the analyses. In total, 61 of 194 unique predictor variables were examined in more than one study for one or more outcomes and were meta-analyzed. Of the 95 meta-analyses, 42 (44%) produced summary odds ratios that were statistically significant different from 1.00. However, the strength of prediction was often weak.
ISBN
0612035735
9780612035737