| Objective:Smoking is an important social behavior factor influencing various diseases and health problems. We aimed to explore the application of meta-analysis of observational studies in research about the social factors and health, through analyzing the influence of smoking on health, using the methods of meta-analysis and overview of systematic reviews.Methods:There were three main methods:1. Literature research:We systematically retrieved international literature about social factors and health, smoking and health, overview of systematic reviews, to see the research progress.2. Methods of quantitative review:We judged heterogeneity between studies by index I2, synthesized data quantitatively by fixed effects model of Mantel Haenszel method or random effect model, analyzed sources of heterogeneity by subgroup analysis and meta-regression analysis, and conducted dose-response analysis based on the generalized linear regression model and fixed effect model of meta-analysis.3. Overview of systematic reviews:we conducted this overview of systematic reviews of of meta-analysis of observational studies according to the ideas and methods of overview of systematic reviews of meta-analysis of randomized controlled trials, and the key steps are:literature searching, formulation of inclusion and exclusion standards, systematic summarization of the research evidence, the necessary statistical analysis. Stata12.0software was used for data analysis.Results:1. Relationship between smoking/smoking cessation and the risk of erectile dysfunction(ED) Meta analysis of observational studies related to smoking and ED showed that the combined relative risk for prospective cohort studies is1.51(95%confidence interval (CI):1.34-1.71) and the combined odds ratio (OR) is2.14(95%CI:1.35-1.35). Meta-analysis of observational studies reporting dose-response data suggested that, with the increase of daily quantity and duration of smoking, the risk of ED increasing gradually:An increase of ten cigarettes smoked per day and an increment of ten years of smoking increased respectively the risk of ED by14%and15%statistically significantly. As the extension of time of smoking cessation, the risk of ED decreased:every ten years of smoking cessation decreased a9%lower risk of ED, OR of95%CI is0.85to0.99.2. The relationship between smoking and healthOverview of reviews based on94meta-analyses of observational studies suggested:(1)Active smoking and health:of the more than70kinds of diseases studied, the increased risk of11diseases was not confirmed to be related to smoking: multiple myeloma, meningioma, prostate cancer, non-hodgkin’s lymphoma, glioma, colorectal cancer, open-angle glaucoma, toxic nodular goiter, amyotrophic lateral sclerosis, carpal tunnel syndrome, and osteoarthritis. Smoking could reduce the risk of three diseases:endometrial cancer (OR:0.81;95%CI:0.74-0.88), skin melanoma (OR:0.83;95%CI:0.74-0.93), and Parkinson’s disease (OR:0.55;95%CI:0.51-0.59). The increased risk of the remaining nearly60kinds of disease was associated with smoking. Of them, diseases with the the greatest risk (increased more than four times) related to smoking were:lung cancer (OR:5.50;95%CI:5.07-5.96), oral cancer (OR:4.65;4.65-3.17), emphysema (OR:4.51;95%CI:3.38-6.02), ectopic pregnancy (OR:15.69;95%CI:2.87-85.76) initial psychosis (OR:6.04;95%CI:3.03-12.02), schizophrenia (OR:5.90;95%CI:4.90-6.70).(2) Passive smoking and health:Although the danger of it was less than active smoking, passive smoking still can increase the risk of many illnesses. Among them, the diseases which had higher risk (over50%) were:child invasive meningococcal disease (OR:2.18;95%CI:1.63-2.92), asthma (OR:1.85;95%CI:1.35-2.53), cervical cancer (OR:1.73;95%CI:1.35-2.21), child pharynx neisseria meningitidis infection (OR:1.68;95%CI:1.19-2.36),child pharyngeal streptococcus pneumoniae infection (OR:1.66;95%CI:1.33-2.07), and baby lower respiratory infection (OR:1.54;95%CI:1.40-1.69).(3) Effect of maternal prenatal smoking on health of offspring:Smoking mainly increased the risk of the following diseases:sudden infant death syndrome (OR:2.25;95%CI:2.03-2.50), cleft lip (OR:1.64;95%CI:1.33-2.02), overweight (OR:1.50;95%CI:1.36-1.65), obesity (OR:1.52;95%CI:1.36-1.70), cleft palate (OR:1.42;95%CI:1.06-1.90), and lower respiratory tract infection (OR:1.24;95%CI:1.111.38).3. The role of meta-analysis of obersevational studies in research about social factors and health and the existing main problemsThe role of meta-analysis of obersevational studies in research about social factors and health was:to quantitatively summarize the influence of a specific social factor on a specific disease worldwide, to point out the difference of this influence in different regions, and to pr ovide more comprehensive, more concentrated, and more level-higher research evidence. The existing main problems were:incomplete literature search, Lack or non-strictness of quality assessment of the original research, data analysis and interpretation of the results without considering quality assessment results, insufficiency of analysis and interpretation of heterogeneity, and lack of systematic summary of multiple meta-analyses of observational studies about a specific social factors and a specific disease.Conclusions:Although it had nothing to do with the increased risk of some diseased, and even could reduce the risk of a small amount of diseases such as endometrial cancer, skin melanoma, and Parkinson’s disease risk, active smoking could lead to a significantly increased risk of a large number of diseases. The diseases, of which the risk increased by more than four times (OR>4.0and having statistical significance), were:lung cancer, oral cancer, emphysema, ectopic pregnancy, psychosis, schizophrenia. Although the harm of it was small than active smoking, passive smoking still could increase the risk of many illnesses. Among these, the diseases with higher increased risk (over50%) were child invasive meningococcal disease, asthma, cervical cancer, child pharyngeal neisseria meningitidis infection, lower respiratory tract in infants, and child pharyngeal streptococcus pneumoniae infection. The influence of prenatal maternal smoking on offspring health was mainly inflected in the increased risk of sudden infant death syndrome, cleft lip, overweight, obesity, cleft palate, and lower respiratory tract infection. Meta-analysis of observational studies played an important role in research of social factors and health, but some issues should be of concern, such as non-standard reporting and inadequate methodology. The number of meta-analyses of observational studies about the same social factors and the same diseases was increasing. These meta-analyses provided more information and there were internal link and difference among the results of them. It is necessary to introduce the idea and method of overview of systematic reviews, to provide more comprehensive, more intuitive, available evidence-based medical evidence for research of social factors and health.Innovations and limitations:The main innovation were about the methodology of evidence-based medicine:1. To analyze the relationship of exposure and disease, we not only considered exposure itself, but the exposures in unit time, the duration of exposure, the cumulative effect of exposure, and the effect of eliminating exposure as well.2. We applied overview of systematic reviews in meta-analysis of observational studies for the first time to summarize the health effects of specific social factors. There were two major limitations. Firstly, literature retrieval is not very comprehensive. The main reasons were following:difficult to consider all the same or similar retrieval words, not to include the grey literature not published officially and non-English literature. Secondly, the reporting of overview of systematic reviews was not very standard. Since there were no unified international norms of overview of reviews, we referred to the cochrane manual and published overviews of systematic reviews of randomized controlled trials to report the present overview. |
PDF Full Text Request