You can run and ANOVA with the mean and standard deviation.

In R:

https://epidemiologist.blog/2019/03/17/one-way-anova-analyses-by-hand-and-in-r/

https://stackoverflow.com/questions/26170002/anova-in-r-using-summary-data

https://brilliant.org/?utm_medium=cpc&utm_source=youtube&utm_term=VA_Topic_Math&utm_content=Whats_Like_Learn_Math_030&utm_campaign=Acquisition_VAC_BestTargeting&gclid=Cj0KCQiA1sucBhDgARIsAFoytUsCIjhdwm5Drq1rZCD9XA02IZnqhO21WPGmxB23AMmbBQ4RlmY1QEYaAgoEEALw_wcB&overwritten_utm_medium=cpc&overwritten_utm_source=youtube&overwritten_utm_term=VA_Topic_Math&overwritten_utm_content=Whats_Like_Learn_Math_030&overwritten_utm_campaign=Acquisition_VAC_BestTargeting

Am i performing statistics correctly?

Presumably not. You did not explain in sufficient detail what the response variable is and how it is measured. As you write that this is a "biomarker level" I assume is't a kind of concentration of a biomolecule - something for which a normal distribution should not be assumed. It's typically more reasonable to assume a log-normal distribution instead. Further, you follow a sample of patients over time, and such measurements are not statsitically independent. This must be considered in the analysis. And finally, you write that you "selected 5 patients". It's not clear how they were selected. Is it a proper random sample of some targeted population? And what is with the controls?

Should not i worry about non significant results?

Given your sampling strategy was ok and the statistical model was appropriate, a non-significant result tells you that your data are not sufficient to make any conclusion about the direction of the change in the population. Taken in isolation, such a study is actually worthless. Being part of a collaborative activity of more centers/scientists, it adds data that may be conclusive when analysed all together.

What are the statistical tests I should use?

Hard to say without better understanding your response variable. It should be some kind of reapeated-measures model addressing the interaction between time and group.

How will I represent my data for publication purposes?

I don't think that data from only 5 patients is worth being published, unless the findings and implications are plausible and really really drastic.

I strongly suggest discussing your study with a local statistician. This may not rescue your current data but may help you to better plan future studies.

Jochen Wilhelm Thankyou for your reply. I am studying changes in biochemical parameters in blood sample of patients. we are comparing changes in venous blood and tissue blood at inflammation sites. So I have selected female patients of same age who come with such disease with inflammation.Then we are trying to evaluate quantitative changes in biochemicals or biomarkers from venous blood and tissue blood and then we give them therapy and recollect the blood from the tissue area.

So this brief summery of my research. As we have planning this study and try to plan future studies with large sample size.

I am thankful for your response.

May you please help me to understand with this data set

patient Biomarker 1 from VB. Biomarker from cb post tpy

Pt 1 184 215 253

pt 2 692 675 849

pt 3 117 107 112

pt 4 279 282 310

pt 5 297 282 310

normal value of biomarker in body is 40-160

How to compare the results and how to apply statistics

I applied anova for them

DF Sum of Squares Mean Square F Value Prob>F

Model 2 8807.98533 4403.99267 0.07538 0.92782

Error 12 701062.792 58421.89933

Total 14 709870.77733

descriptive statistics

Sample Size Mean Standard Deviation SE of Mean

vb bio 5 314.24 223.53627 99.96846

cb1 bio 5 329.7 215.54712 96.3956

CB II 5 371.6 280.77869 125.56805

Tukey test

MeanDiff SEM q Value Prob Alpha Sig LCL UCL

cb1 vb 15.46 152.86844 0.14302 0.99438 0.05 0 -392.37021 423.29021

cb2 vb 57.36 152.86844 0.53065 0.9258 0.05 0 -350.47021 465.19021

cb2 cb1 41.9 152.86844 0.38762 0.95957 0.05 0 -365.93021 449.73021

Please help now.

Dear Shahnawaz Ahmad Wani , I'd really encourage you to discuss this with a local statistician. Your medical institute should have a division for bio/medical statistics to help you.

This is possibly an address where you could ask for further help:

http://www.skims.ac.in/index.php/departments/professional/others/4997-clinical-research

At least they write: "Leveraging on strong scientific, clinical, biostatistical and regulatory acumen of our proposed team we shall work collaboratively with global and local institutions to plan, execute, and document research studies with accuracy, quality, expediency, and cost-effectiveness."

You wrote:

So as i apply anova results show significant difference in their mean values but due larger difference in their standard deviations i am getting non significant results.

I think your statement is self-contradicting...

Provide your data and I will analyse them for you.

Chears

Massimo Sivo very nice of you that you want to try to help your colleague, however, as was mentioned earlier, you should understand experimental design very well before you have sufficient information to construct an appropriate statistical model (that can than provide meaningful insights). To understand experimental design of a clinical trial it is not enough to understand some statistics, you should also understand some medicine. E.g. From what Shahnawaz Ahmad Wani described one can by no means understand what was actually measured, how, and why. It is super challenging to sample 5 patients with appropriate controls to make any kind of biologically meaningful inference about the population of patients. Tons of confounders and no power to take them into account. From how the problem was presented I can image such (confounder) data was not collected. Using such data and shoving them into e.g. mixed model to account for repeated measures (and dependence of some parameters) will most likely give meaningless results. Even if everything was done right but you did not account for a biologically meaningful confounder... it would still be meaningless. For example, Shahnawaz Ahmad Wani told us he is measuring some kind of biochemical parameter from the blood from female patients. Many parameters in the blood change with the part of the menstrual cycle, and it is not clear whether the authors analysed this. Without accounting for such a serious confounder the study makes no sense. Furthermore, how would you construct a model around the dependence of venous blood and "tissue blood". The variables are most definitely related, but we cannot imagine the nature of their dependence without having much more information. So, before asking for data, you should really ask for the explanation of the design to make sure all the confounders were ruled out the the level of design as this is the only scenario in which several lines of code or several clicks in some statistical software will help you in any way to make some kind of inference.

Please be mindful about this kind of stuff, uncertain and misleading conclusions can be very dangerous in medicine and it is healthier for the community to talk about them than to just generate some numbers from the data.

Best,

jan

Thankyou JAn Homolak,

Brother I have selected

5 patients I have selected have age of 60-65 years, We want to check the effect of therapy on the patients. If you want me to share a small part of data you will see that there is difference between means.

PLease check the data below

Can you provide me some solution to the problem.

I am thankful for your response.

May you please help me to understand with this data set

patient Biomarker 1 from VB. Biomarker from cb post tpy

VB, TB1, TB2 Venous blood= VB

Pt 1, 184, 215, 253 Tissue blood = TB , TB2

pt 2, 692, 675, 849

pt 3, 117, 107, 112

pt 4, 279, 282, 310

pt 5, 297, 282, 310

normal value of biomarker in body is 40-160

How to compare the results and how to apply statistics

I applied anova for them

DF Sum of Squares Mean Square F Value Prob>F

Model 2 8807.98533 4403.99267 0.07538 0.92782

Error 12 701062.792 58421.89933

Total 14 709870.77733

descriptive statistics

Sample Size Mean Standard Deviation SE of Mean

vb bio 5 314.24 223.53627 99.96846

cb1 bio 5 329.7 215.54712 96.3956

CB II 5 371.6 280.77869 125.56805

Tukey test

MeanDiff SEM q Value Prob Alpha Sig LCL UCL

cb1 vb 15.46 152.86844 0.14302 0.99438 0.05 0 -392.37021 423.29021

cb2 vb 57.36 152.86844 0.53065 0.9258 0.05 0 -350.47021 465.19021

cb2 cb1 41.9 152.86844 0.38762 0.95957 0.05 0 -365.93021 449.73021

Please help now.

Jan Homolak and your recommendersù

I understand your idea and I agree:

• Massimo Sivo very nice of you that you want to try to help your colleague, however, as was mentioned earlier, you should understand experimental design very well before you have sufficient information to construct an appropriate statistical model (that can than provide meaningful insights). To understand experimental design of a clinical trial it is not enough to understand some statistics, you should also understand some medicine.

I wanted to provide my analysis, leaving the interpretation to the EXPERTS!!!

Good analysis ... depends on good knowledge ...

I know cases where EXPERTS on the subject thought to know the response to the experiments in advance...

If just 5 samples better to make it as case series rather than case control study.

Yes, it is case series?

Even though I am not an "expert in the field", here are my findings:

thanks we can use friedman anova.

Shahnawaz Ahmad Wani

What did you get with Friedman ANOVA?

I found NONSENSE values!

Thanks

M. Sivo

Although I got insignificant P values but correct way to apply statistics.