Cohens d is usually calculated by mean difference (pooled) divided by standard deviation. Cohen (1988) defined effect sizes as "small, d = .2," "medium, d = .5," and "large, d = .8. (Borenstein, 2009)
Thus, it can be safely concluded that the effect size of the intervention is small and not clinically significant. Association and causality is not proven in this case. The data set given was imported to Stata and summary statistics were used and graphs and scatter plots on Stata to visualise and analyse the data.
3 A linear regression analysis was done with Stata to find the treatment effect to adjust for baseline co variates. (propensity score matching)
The results are as below- weight after treatment being the dependent variable (continuous …show more content…
Even though the study findings were not significant they were encouraging and a better designed study could have had better results.
Part C – Research study design.
“Does obesity in childhood increase the risk of developing Type I Diabetes?”
1.The question above is to test the association between childhood obesity and Type 1 diabetes. The best study design to answer this question is possibly a long-term case control study- i.e. cases with Type 1 diabetes (from the local GP database) and less than 14 years of age with age and sex matched controls without Type 1 diabetes.
The growth data of these children could be obtained from the GP and the school health clinics. Data as height, weight baseline BMI and BMI z scores. These are compared yearly till the time they develop or are diagnosed with type 1 diabetes.
This is because we are choosing an outcome(diabetes) and looking back to see if there was an exposure that caused it (childhood obesity).
A scoping search to see what is already known in the topic on Medline is useful to plan the …show more content…
The inclusion criteria would be –
Children with Type 1 diabetes. (on the diabetic register.)
Children with a local GP as in their local area.
The exclusion criteria are-
Children who have other co morbidities as hypertension, asthma, genetic causes of obesity, family history of obesity.
Children who have learning difficulties or mental health problems.
Children diagnosed with diabetes before 5 years of age. (It is anticipated that the numbers will be small.)
Family history of diabetes- as the association is to see if obesity causes Type 1 diabetes and not genetic causes.
3. A sample size calculation need to be done at the start to assess the number needed to gain statistical power and prove association.
The first set of data you would need to collect is baseline characters and see if any adjustments need to be made to avoid bias and confounders. These would normally be age, sex, year in school, baseline weight when they first were 5 years old, baseline height, educational status of parents, socio economic status of parents, baseline BMI, baseline waist measurements, baseline BMI z scores (adjusted for UK growth chart cut offs). Similar set of data need to be collected for the control groups that are equally
Thus, it can be safely concluded that the effect size of the intervention is small and not clinically significant. Association and causality is not proven in this case. The data set given was imported to Stata and summary statistics were used and graphs and scatter plots on Stata to visualise and analyse the data.
3 A linear regression analysis was done with Stata to find the treatment effect to adjust for baseline co variates. (propensity score matching)
The results are as below- weight after treatment being the dependent variable (continuous …show more content…
Even though the study findings were not significant they were encouraging and a better designed study could have had better results.
Part C – Research study design.
“Does obesity in childhood increase the risk of developing Type I Diabetes?”
1.The question above is to test the association between childhood obesity and Type 1 diabetes. The best study design to answer this question is possibly a long-term case control study- i.e. cases with Type 1 diabetes (from the local GP database) and less than 14 years of age with age and sex matched controls without Type 1 diabetes.
The growth data of these children could be obtained from the GP and the school health clinics. Data as height, weight baseline BMI and BMI z scores. These are compared yearly till the time they develop or are diagnosed with type 1 diabetes.
This is because we are choosing an outcome(diabetes) and looking back to see if there was an exposure that caused it (childhood obesity).
A scoping search to see what is already known in the topic on Medline is useful to plan the …show more content…
The inclusion criteria would be –
Children with Type 1 diabetes. (on the diabetic register.)
Children with a local GP as in their local area.
The exclusion criteria are-
Children who have other co morbidities as hypertension, asthma, genetic causes of obesity, family history of obesity.
Children who have learning difficulties or mental health problems.
Children diagnosed with diabetes before 5 years of age. (It is anticipated that the numbers will be small.)
Family history of diabetes- as the association is to see if obesity causes Type 1 diabetes and not genetic causes.
3. A sample size calculation need to be done at the start to assess the number needed to gain statistical power and prove association.
The first set of data you would need to collect is baseline characters and see if any adjustments need to be made to avoid bias and confounders. These would normally be age, sex, year in school, baseline weight when they first were 5 years old, baseline height, educational status of parents, socio economic status of parents, baseline BMI, baseline waist measurements, baseline BMI z scores (adjusted for UK growth chart cut offs). Similar set of data need to be collected for the control groups that are equally