Sample size calculation for prevalence study with examples

[@abubakarnur4271]

other wise very thankful for you clear explanations and examples you gave us

[@tarekegnbekele3451]

thank you dear doc for your nice explanation

[@PublicHealthResearchDrSandheep]

You are most welcome 🙏

[@hopeworship-7650]

Thanks sir for your help full video .pls help me this) A researcher wanted to conduct a research to determine the prevalence of ANC utilization of mothers in a district with infinite population. The prevalence of ANC utilization from previous study was found to be 45.7%. What is the final sample size to address the objective? If you don’t have any information about P, what are you going to do and what is the final sample size in this scenario?

[@PublicHealthResearchDrSandheep]

Please use 0.457 as P and 0.543 as 1 -P. Precision will be 0.05. Calculate the sample size substituting these values in the formula

[@PublicHealthResearchDrSandheep]

If you don’t have information about P from previous research - please take P as 50% - or 0.5

[@humphreylombe]

very helpful video.Thanks

[@PublicHealthResearchDrSandheep]

You are welcome

[@satyaprakashtripathi1279]

Sir i have a doubt suppose my prevalence is 1.4% and if i calculate it is like 22 sample which is not ideal

[@PublicHealthResearchDrSandheep]

Hi Satya. If prevalence is 1.4% ….. please look for some other articles also …. If you get a range of prevalence from 1 to 1.8% (for example)….then please take precision as 0.4…. Please try that way

[@satyaprakashtripathi1279]

@@PublicHealthResearchDrSandheep okay thank you so much sir for your kind response, i ll try that way and get back to you..

[@huzaifaimam7252]

Fantastic video! when do you need to use the sample size correction, how small should the population be? I have seen several suggestions calling for its use when the population is less than 10,000 or less than 5K etc. Which is correct Sir?

[@PublicHealthResearchDrSandheep]

correction formula for finite population can be applied when the total number in the sampling frame is known (or when the total number is already counted). I usually use when the number of individuals in the sampling frame is 2000 or less.

[@huzaifaimam7252]

@@PublicHealthResearchDrSandheep thank you prof

[@ranadam5568]

Can I use a prevalence calculated from a study conducted more than 5 years ago, for example 2016?

[@PublicHealthResearchDrSandheep]

Please use the latest prevalence percentage available from the published research articles

[@AdaRdor]

Do you know what the R code for the first formula is (sample size=(Zalpha)^2*P*(1-P)/precision^2)? Thanks in advance!

[@PublicHealthResearchDrSandheep]

Please use “stats” package. The command for estimation of single proportion is as follows sample_size

[@sushmitatamang1581]

Student wants to estimate the prevalence of antibiotic resistance in small state but there is no earlier research paper done for that state yet. How to calculate sample size?

[@PublicHealthResearchDrSandheep]

student can take the prevalence from other states in the same country / from nearby countries (with similar levels of health outcomes).

[@avinashtadvi5795]

My Objective:To asess Prevalence of cutaneous manifestation in internal malignancy patient in tertiary hospital cross sectioal study Prevalence from previous study -27% By chochran formula Z2Pq/d2 I got 303 as sample size -How to calculate for finite population if i dont know exact N as hospital data not available -Is it necessary to calculate for finite population -Can i assume N as 500 Can u help me to calculate sample size by above data sir I am affraid of not getting answers I wish sample around 150

[@PublicHealthResearchDrSandheep]

Please enquire about the average number of patients from the tertiary hospital. If you don’t know exact numbers, you can take approximate numbers

[@PublicHealthResearchDrSandheep]

sorry for the delay in replying

[@nathanaelsungte8326]

An investigator is intending to estimate the prevalence of Anaemia in a selected district. The expected prevalence is 28%. What is the required sample size to estimate the prevalence with a relative precision of 10% and alpha error of 5%? a) 514 b) 806 c) 988 d) 1526 What is the correct answer sir ? Kindly help

[@PublicHealthResearchDrSandheep]

987 is the sample size Proportion is 28% …. relative precision of 10% means absolute precision is 2.8% ….. alpha error of 5% means confidence level is 95%

[@Upasana-Sharma]

Can I use prevalence from previous study for sample size calculation for rct ?

[@PublicHealthResearchDrSandheep]

sample size calculation for rct is different. Outcome variable can be any one of the following - difference in proportions between experimental group and control group OR difference in mean values of continuous variable between experimental group and control group

[@PublicHealthResearchDrSandheep]

If you need more detailed explanations, I will make a video

[@Upasana-Sharma]

@@PublicHealthResearchDrSandheep thank you sir 🙏

[@anithasanthanam9787]

If we want to study prevalence in chennai is it appropriate to use prevalnce data from any other districts in tamilnadu

[@PublicHealthResearchDrSandheep]

if other research on prevalence in Chennai is not available, you can choose prevalence from another district with similar socio demographic characteristics - for calculation of sample size. Thank you

[@avinashtadvi5795]

How to consider N in finite population? When finite population is not known My study prevalence of cutaneous manifestation in internal maliganancy patients? Previous study P-27% Kindly help me sir waiting for your reply sir 😊❤

[@PublicHealthResearchDrSandheep]

you can define the sampling frame ( can include all the admitted patients with internal malignancy - from selected hospital or hospitals). Then define the duration of the time period. All those admitted with internal malignancy during that period in those selected hospitals - constitute N (it is a hypothetical scenario)

[@psychiatrysrmtrichy4736]

what if the prevalence is 0.3 percent sir? how to calculate precision then?

[@PublicHealthResearchDrSandheep]

If prevalence is so less, you can use relative precision. If prevalence is 0.3 % and relative precision is 5% of prevalence… 5% of 0.3% will be the relative precision. When prevalence is less, sample size calculated will be more

[@vivekthaware6507]

What if we don't know the percentage of prevalence..then what would be the prevalence?

[@PublicHealthResearchDrSandheep]

you can assume that the prevalence is 50% - that assumption gives you the maximum possible sample size

[@agriculturInfo]

if there is hav't previous research how to give ? p value?

[@PublicHealthResearchDrSandheep]

Please give anticipated prevalence as 50% and precision as 5%

[@deepaksingh-ut7lh]

How to calculate the sample size in case prevalence is given as a range .for example range of prevalence of diabetes is 10-14% ??

[@PublicHealthResearchDrSandheep]

you can take the middle value of the range as the anticipated frequency (my opinion…didn’t get any reference 🙏)

[@deepaksingh-ut7lh]

@@PublicHealthResearchDrSandheep thank u sir

[@drmanojlectures4233]

How to derive Absolute precision from Relative Precision

[@PublicHealthResearchDrSandheep]

To the best of my knowledge, absolute precision is used for calculating sample size of community based research. It can be based on the variation in prevalence in previous studies. Absolute precision can not be calculated from the relative precision value (according to my understanding). Thank you

[@husnat5420]

Study duration how much

[@PublicHealthResearchDrSandheep]

this formula is applicable for cross sectional study. Data will be collected only once.

[@abubakarnur4271]

incorrect formula I understood all things but which formula did you calculate the sample of 384?

[@PublicHealthResearchDrSandheep]

Sample size = [(z alpha) x (z alpha) x p x (1 - p) ] / (precision x precision)

[@abubakarnur4271]

@@PublicHealthResearchDrSandheep So where did you calculated the Prevalence (P) of those 600 students from the college Thank you for your time and helping me

[@PublicHealthResearchDrSandheep]

Prevalence can be taken from the previous research article - on the similar research topic - conducted in a similar population. If similar research article is not available. We can anticipate the prevalence as 50% - as it gives the highest sample size