Wednesday, March 2, 2011
Basic Knowledge About Study Design
Define:
Exposure: The exposure means the subjects were exposed to certain thing that we are interested in the given study. For example, if we want to investigate the reason of the food poisoning event occurred after a big party, we will collect the information about the food type people consumed during that party. In other words, the exposure in this case is eating or drinking the food.
Causality: In epidemiology, the causality is the association between exposure and disease. If the exposure results in the disease, they have the causality relationship.
Bias: Bias is an inclination to present or hold a partial perspective at the expense of (possibly equally valid) alternatives. Bias can come in many forms.
(This is the quotation from http://en.wikipedia.org/wiki/Bias.)
Cross-sectional study:
The cross-sectional study is used to determine the association between exposure and the disease or health-related characteristics among a interested population in a specific time. The hallmark of cross-sectional study is the snapshot, which means researchers collecting the data of exposure and disease simultaneously.
The cross-sectional study is best for quantifying the prevalence of disease (or health risk) and for quantifying the accuracy of a diagnostic test.
Strengths: The cross-sectional study is not expensive, easy to conduct and ethically safe.
Weaknesses: Only examine the association between variables, not causality.
The recall bias exist in this type of study.
The group size may vary.
Confounders may be unequally distributed.
Case-Control study:
The case-control study starts with identifying the case group and the control group. The case group are the individuals who have the certain disease or health outcome. By contrast, the control group are individuals who do not have the certain disease or health outcome. The control group should be appropriate to make sure the research accuracy. Afterword, we examine the exposure and non-exposure in case group and control group respectively. Based on these data, we can calculate the proportion of exposure and non-exposure in case and control group, thus, the conclusion about association between exposure and disease can be drawn.
Strengths: Quick and cheap, fewer subjects than cross-sectional study, only feasible method for rare disease and for the subjects who have long time interval between exposure and outcome.
Weaknesses: Recall bias, selection bias, difficulties in selection of control group, and the effect of confounders.
Cohort Study:
Compared with the case-control study, the cohort study begins with the exposure and non-exposure. Data were collect from the people who are exposed and not exposed, followed by obtaining the data about disease and non-disease in each group. The corhort study if best for the effect of predictive factors on disease.
Strengths: We can matching the subjects, can establish timing and directionality of events, ethically safe, we can standardise the eligibility criteria and outcome assessment.
Weaknesses: Difficulties in defining control group, binding is difficult, exposure may related to confounders that we do not know.
If we want to study the rare disease, the corhort study needs long time to follow up and large sample size.
References:
1.http://www.cebm.net/index.aspx?o=1039
Tuesday, March 1, 2011
Let us understand the health news!
I found a piece of interesting news on the website of New York Times. [1]This news reveals the potential association between childhood obesity and school lunch. The research was conducted among the children who study in Southeastern Michigan schools. The result indicates that, the children who consumed school lunch on daily basis have 29% more risk of developing obese problem than the children who ate home-made lunch. The research also explore the link between watching TV or video games and obesity issue. The long time spending on watching TV and video games could also contribute to the increase in risk of obesity.
Afterward, I searched the original article related to this piece of news. [2] I think asking questions is a good way to learn the health articles, so, I will pose some questions below to understand this article.
What is the exposure and disease?
Answer: The exposure in this research is the school lunch consumption, and watching TV or video games. The disease is obesity.
What are the measured and unmeasured confounders?
Answer: The measured confounders: The age, gender, ethnics/racial factors, height, weight, blood pressure, serum cholesterol, glucose, total cholesterol, high-density lipoprotein, physiologic characteristics, dietary characteristics, physical activity, sedentary activities, lipid profile and heart rate.
The unmeasured confounders: Genetic information,the BMI and other obesity indicators of parents, stress level, emotional status, disease history, medicine and treatment in the recent period and family income. In addition, besides lunch, we also need to know what type of food these students consumed for supper and breakfast, and what they often eat as the snack. If a student ate home-made food for lunch, but had dinner at KFC regularly, he/she may have the obesity problem. We need to collect the comprehensive information about subjectives' eating habit, not just for the lunch.
These unmeasured confounders may result in the outcome bias. The confounders could influence on both exposure and disease. If a study has too many unmeasured confounders, we would cast doubt on the causality exists between exposure and disease.
What is the study design and how do we know?
Answer: This research is a case- control study because the researchers firstly determined the group of obese (case) and the group of non obese ( control), and then collect the data regarding to the exposure ( eating habits) in each group, to calculate the proportion of the cases (and controls) were exposed and not exposed. Since this study is a case-control study, the researcher calculate the odds ratio to interpret the data.
What are the strengths and weaknesses of the study design?
Answer:
Strengths : This study collect the information of many variables, thus, it decreases the outcome bias by considering many confounders. In addition, the two groups were similar with respect to the age, gender. The similarity may diminish the bias come from the effect of gender and age. Also, this study is cheap and quick.It is not time-consuming and money-consuming.
Weaknesses: Firstly, as mentioned above, this study has many unmeasured confounders. These confounders may lead to bias. Secondly, since it is a case-control study, this research may has the "recall bias". It is possible that the children cannot recall their diet items and amount very clearly, thus causes bias in data. Third, the genetic factors could also lead to obesity. However, in this study, researchers did not collect any data about genetic information. Fourth, the objectives only come from the Midwest communities, and all the subjectives are participants in the health program. This process may cause the "selection bias". The restriction of sample selection might result in the restriction of conclusion.The conclusion can not be generalized to the broad population. Finally, the researchers used the questionnaire to get the data. Because the objectives are children, they may do not answer the questions very seriously. Also, they may lack the health literacy of understanding the question. All of these could lead to the data bias.
Do you agree with the author's interpretation of the findings?
Answer: In general, the outcome of this finding is reliable. So, I agree with the interpretation that the school lunch and watching TV or video games are related to the childhood obesity. However, I think the investigation of school lunch is also very important. If the school lunch is high-calories and low-nutrition, the interpretation of this research may be more persuasive.
Was this responsibly reported in the Internet article?
Answer: Yes, I think the article should be reported on the Internet. Because the research is basically valid, the conclusion of this article should be learned by the general public. In addition, this article will not lead to negative effect to life. After reading this article, more children may bring lunch from home, it would not be a bad thing for the health of children.
What is your impression of this article?
Answer: In general, I think the finding of this article is convictive. Before reading this article, I have known that the school lunch may result in obesity. According to my experience, I gained more pounds when I ate the school lunch. As a contrast, I became slimmer if I ate the home-made food. So, I believe the conclusion of this article. However, it would not be a surprise for me.
Would this change your behavior?
Answer: Yes, this article will remind me the healthy eating behavior. I will try to eat at home and eat in a health way, in order to control my weight.
Should you be alarmed?
Answer: Since the objectives of this research are children, the conclusion may can not be generalized to me. I will tell this finding to my brothers.
What is the value of this article?
Answer: This article may stimulate more researchers to explore the association of obesity and school lunch in other groups, such as the college students. Such researches may help us to deal with the obesity issue.
References:
1.http://www.nytimes.com/2011/02/08/health/research/08childhood.html
2. http://proxy.mul.missouri.edu:2159/science?_ob=ArticleURL&_udi=B6W9H-51NNFHD-12&_user=3419478&_coverDate=12%2F31%2F2010&_rdoc=1&_fmt=full&_orig=gateway&_origin=gateway&_cdi=6683&_sort=d&_docanchor=&view=c&_acct=C000049994&_version=1&_urlVersion=0&_userid=3419478&md5=937b9ae14f722ec7e1ee03d2f288b7dc&searchtype=a&artImgPref=F
3. Gordis, Leon, Case-Control Studies and Other Study Designs, Epidemiology, fourth edition, 2008, 169-177.
Afterward, I searched the original article related to this piece of news. [2] I think asking questions is a good way to learn the health articles, so, I will pose some questions below to understand this article.
What is the exposure and disease?
Answer: The exposure in this research is the school lunch consumption, and watching TV or video games. The disease is obesity.
What are the measured and unmeasured confounders?
Answer: The measured confounders: The age, gender, ethnics/racial factors, height, weight, blood pressure, serum cholesterol, glucose, total cholesterol, high-density lipoprotein, physiologic characteristics, dietary characteristics, physical activity, sedentary activities, lipid profile and heart rate.
The unmeasured confounders: Genetic information,the BMI and other obesity indicators of parents, stress level, emotional status, disease history, medicine and treatment in the recent period and family income. In addition, besides lunch, we also need to know what type of food these students consumed for supper and breakfast, and what they often eat as the snack. If a student ate home-made food for lunch, but had dinner at KFC regularly, he/she may have the obesity problem. We need to collect the comprehensive information about subjectives' eating habit, not just for the lunch.
These unmeasured confounders may result in the outcome bias. The confounders could influence on both exposure and disease. If a study has too many unmeasured confounders, we would cast doubt on the causality exists between exposure and disease.
What is the study design and how do we know?
Answer: This research is a case- control study because the researchers firstly determined the group of obese (case) and the group of non obese ( control), and then collect the data regarding to the exposure ( eating habits) in each group, to calculate the proportion of the cases (and controls) were exposed and not exposed. Since this study is a case-control study, the researcher calculate the odds ratio to interpret the data.
What are the strengths and weaknesses of the study design?
Answer:
Strengths : This study collect the information of many variables, thus, it decreases the outcome bias by considering many confounders. In addition, the two groups were similar with respect to the age, gender. The similarity may diminish the bias come from the effect of gender and age. Also, this study is cheap and quick.It is not time-consuming and money-consuming.
Weaknesses: Firstly, as mentioned above, this study has many unmeasured confounders. These confounders may lead to bias. Secondly, since it is a case-control study, this research may has the "recall bias". It is possible that the children cannot recall their diet items and amount very clearly, thus causes bias in data. Third, the genetic factors could also lead to obesity. However, in this study, researchers did not collect any data about genetic information. Fourth, the objectives only come from the Midwest communities, and all the subjectives are participants in the health program. This process may cause the "selection bias". The restriction of sample selection might result in the restriction of conclusion.The conclusion can not be generalized to the broad population. Finally, the researchers used the questionnaire to get the data. Because the objectives are children, they may do not answer the questions very seriously. Also, they may lack the health literacy of understanding the question. All of these could lead to the data bias.
Do you agree with the author's interpretation of the findings?
Answer: In general, the outcome of this finding is reliable. So, I agree with the interpretation that the school lunch and watching TV or video games are related to the childhood obesity. However, I think the investigation of school lunch is also very important. If the school lunch is high-calories and low-nutrition, the interpretation of this research may be more persuasive.
Was this responsibly reported in the Internet article?
Answer: Yes, I think the article should be reported on the Internet. Because the research is basically valid, the conclusion of this article should be learned by the general public. In addition, this article will not lead to negative effect to life. After reading this article, more children may bring lunch from home, it would not be a bad thing for the health of children.
What is your impression of this article?
Answer: In general, I think the finding of this article is convictive. Before reading this article, I have known that the school lunch may result in obesity. According to my experience, I gained more pounds when I ate the school lunch. As a contrast, I became slimmer if I ate the home-made food. So, I believe the conclusion of this article. However, it would not be a surprise for me.
Would this change your behavior?
Answer: Yes, this article will remind me the healthy eating behavior. I will try to eat at home and eat in a health way, in order to control my weight.
Should you be alarmed?
Answer: Since the objectives of this research are children, the conclusion may can not be generalized to me. I will tell this finding to my brothers.
What is the value of this article?
Answer: This article may stimulate more researchers to explore the association of obesity and school lunch in other groups, such as the college students. Such researches may help us to deal with the obesity issue.
References:
1.http://www.nytimes.com/2011/02/08/health/research/08childhood.html
2. http://proxy.mul.missouri.edu:2159/science?_ob=ArticleURL&_udi=B6W9H-51NNFHD-12&_user=3419478&_coverDate=12%2F31%2F2010&_rdoc=1&_fmt=full&_orig=gateway&_origin=gateway&_cdi=6683&_sort=d&_docanchor=&view=c&_acct=C000049994&_version=1&_urlVersion=0&_userid=3419478&md5=937b9ae14f722ec7e1ee03d2f288b7dc&searchtype=a&artImgPref=F
3. Gordis, Leon, Case-Control Studies and Other Study Designs, Epidemiology, fourth edition, 2008, 169-177.
Subscribe to:
Posts (Atom)