I would like to conduct research about the effect of spleen/liver stiffness measured by new MR modality on detection of vascular hypertension in the liver in people with cirrhosis by dividing the participants into 3 groups (controls, moderate hypertension, and severe hypertension). How can I calculate the sample size? I used one and only article as reference to predefine the following assumptions: standard deviation of 3.24, a significance level (α) of 0.05, and a power (1-β) of 90% to have a sample size of 20 subjects to detect a difference in spleen stiffness of 2 kPa with as withdrawn from the animal study reported earlier.
It is however not clear for me if this when using 20 healthy / 20 severe or not? I am not sure how to distinguish between severe and not severe by means of calculating the moderate hypertension.
Appreciate your help.
First of all, I would suggest NOT to define groups. Measure the degree of hypertension and use this as a continuous predictor in your analysis.
You will also have to take care of many other influential variables you cannot control, like age, BMI, sex, comorbidities, diet, etc. You should record these variables and use them in the analysis, too.
The sample size might be estimated using the “resource equation method": the total number of patients minus the number of parameters estimated in the statistical model should be between 10 and 20. So when you model hypertension, age, and BMI as linear and sex as a two-level factor, the model will estimate 5 parameters and the total sample size should be between 15 and 25.
Firstly, You need to know the three groups population.
Then you may determine your sample size by using stratified sampling.
proportionally sampling i.e Stratified Random Sampling Method will be useful. For instance you use 10% of each group.
But if the population is unknown for each group, you may use Purposive sampling method.
First of all, you need to know the population of each group in order to determine the total population (summation of the three). Then, you apply the proportionate stratified random sampling technique formular.
As long as the population of the three group can be determined. It is best to use proportionate stratified random sampling. So that each group can be sampled according to the proportion of it's population.
- Badges
- Science topic