Sample size calculation for prevalence studies using Scalex and ScalaR calculators.

Naing L, Nordin RB, Abdul Rahman H, Naing YT. · 2022

Read Sample size calculation for prevalence studies using Scalex and ScalaR calculators. by Naing L, Nordin RB, Abdul Rahman H, Naing YT. (2022) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc9338613.txt.

Introduction

Naing and co-authors explain how to size a prevalence study using two free calculators. Picking the right sample size is vital, yet errors in formula and reporting are common. Their tools run in spreadsheet software and in RStudio. That keeps them open to researchers with small budgets.

Four inputs drive the math: confidence level, precision, data variability, and expected loss of participants. The guide shows how each choice changes the final number. The calculators walk users step by step and guard against common slips. They also draft a publication-style report of the result. The authors warn that no single formula fits every study design. Clear reporting, they argue, is part of doing the science right.

Catalog record

CREATORNaing L, Nordin RB, Abdul Rahman H, Naing YT.
YEAR2022
COLLECTIONPAPERS
FORMATTXT
FILE SIZE22 KB
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC9338613

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Sample size calculation for prevalence studies using Scalex and ScalaR calculators By Lin Naing, Rusli Bin Nordin, Hanif Abdul Rahman, Yuwadi Thein Naing Published in: BMC Medical Research Methodology DOI: 10.1186/s12874-022-01694-7 ABSTRACT Background Although books and articles guiding the methods of sample size calculation for prevalence studies are available, we aim to guide, assist and report sample size calculation using the present calculators. Results We present and discuss four parameters (namely level of confidence, precision, variability of the data, and anticipated loss) required for sample size calculation for prevalence studies. Choosing correct parameters with proper understanding, and reporting issues are mainly discussed. We demonstrate the use of a purposely-designed calculators that assist users to make proper informed-decision and prepare appropriate report. Conclusion Two calculators can be used with free software (Spreadsheet and RStudio) that benefit researchers with limited resources. It will, hopefully, minimize the errors in parameter selection, calculation, and reporting. The calculators are available at: (https://sites.google.com/view/sr-ln/ssc). BACKGROUND In quantitative research, when we take a sample from a study population or eligible population in order to save our resources, there are two important statistical processes namely using a probability sampling method (commonly known as “random sampling”) [1], and calculating an appropriate sample size [2]. Both are equally important to ensure a good representative sample for the study population. As we need a specific statistical analysis for a specific research objective, we also need a specific sample size calculation method for a specific research objective. Even if two research objectives may require a similar statistical analysis, the sample size might be different depending on the parameters that we use for the calculation. In this paper, we focus on the objective that estimates a prevalence or proportion, for example, to estimate the prevalence of obesity, the prevalence of smoking, the prevalence of heart disease, diabetes mellitus or any other diseases of a study population. The method in this paper will not be suitable for other type of objectives such as estimating mean, comparing means, comparing proportions or regression analyses. Books [3, 4] and published articles [5, 6] guiding the methods of sample size calculation for prevalence studies are available. Nevertheless, we observed that several parts of the sample size calculation process can be guided by a software or calculator and it can prevent incorrect calculation, incorrect use of formula, incorrect parameters, and incomplete sample size reporting. Sample size softwares and calculators are extremely helpful that are available through commercial licenses such as Power Analysis & Sample Size (PASS) [7], or via freely available softwares such as Epitools [8] and the “presize” package in R [9]. However, there are a lot of confusion that still exists, that resulted in users incorrectly calculating sample size of their studies [10, 11] especially the erroneous notion that one blanket formula can be used for all study designs [6]. In addition, users are expected to have some statistical knowledge to calculate and report the sample size calculation. Incorrect sample size calculation could introduce statistical errors that give rise to inaccurate results, which could be serious, particularly in medical research where evidences from these research studies are cornerstones of medical practices [12, 13]. Many reasons could be attributed to these confusion, inaccuracy, and misunderstanding, in particular, the complexity of available softwares and corresponding guidelines [13]. Therefore, in this paper, we are addressing these issues by introducing a user-friendly Excel calculator that guides users to use the correct method and parameters step-by-step. This calculator also generates a publication-style report of adequate sample size for users’ study. We believe that, this will improve sample size calculation in future prevalence studies in medical and health sciences. IMPLEMENTATION Method to calculate sample size For an objective that estimates a prevalence, the sample size calculation formula is fairly simple and available in a number of books. The following formula [2] shall be used:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$n=\frac{{Z}^{2}P\left(1-P\right)}{{d}^{2}}$$\end{document}n=Z2P1-Pd2 where n = Sample size, Z = Z statistic for a level of confidence (1.96 for 95% confidence level), P = Expected prevalence or proportion, and. d = Precision. However, we do not encourage researchers to use formula as it could have human error in manual calculation. We can use available softwares, and concentrate on carefully choosing appropriate parameters for the calculation. Appropriately choosing parameters The above formula indicates three parameters to be determined. Parameter 1: level of confidence When we take a sample but wish to know about the population (such as prevalence of smoking) from where the sample is taken, we will not know the exact prevalence of the population as we do not study all members of the population. However, the sample study gives us an estimation which has lower and upper limits (informally ‘a range’, but we call ‘interval’ in Statistics) for the population prevalence. We normally calculate these lower and upper limits or an interval with a certain level of confidence. Commonly used or almost always used “level of confidence” for these intervals or estimates, is 95% (which we called 95% confidence interval, CI) in medical and health fields. In addition, most data analysis softwares give the results with 95% CIs by default. For these reasons, and also to minimize users’ error by non-statisticians, we have fixed the level of confidence as 95% without giving users’ choice in these presented calculators. Parameter 2: precision As mentioned above, we will not know the exact prevalence of the population as we do not study all members of the population. Therefore, the prevalence we calculate from the sample could deviate (error) from the population prevalence. We call this deviation as sampling error. We also know that, the larger the sample size, the smaller the errors in estimation. The errors are calculated as precision or also known as ‘margin of error’. Practically, the precision reflects the width of 95% confidence interval. If we decide to choose an absolute precision of ± 2% in estimating a prevalence, we should expect, in the result, the width of 95% CI as 4% (example: 95% CI: 23%, 27%). If the absolute precision is ± 5% in estimating a prevalence, we should expect, in the result, the width of 95% CI as 10% (example: 95% CI: 20%, 30%). The width of the CI is twice that of the precision. Details are presented in Table 1.Table 1Relationship between Precision and width of Confidence Interval (CI)Prevalence in SampleAbsolute Precision95% CI for PopulationCI WidthRequired Sample size25% ± 2%(23%, 27%)4%180125% ± 5%(20%, 30%)10%28925% ± 10%(15%, 35%)20%7330% ± 2%(28%, 32%)4%201730% ± 5%(25%, 35%)10%32330% ± 10%(20%, 40%)20%81 It is an opportunity for researchers to decide the precision (margin of error) and the width of the CI that they wish to see in the results. Normally, researchers wish to have narrower width of CI but the narrower it is, the more expensive (bigger sample size) it is going to be. Even if researchers decide to go for a smaller sample size, the researchers can also foresee or appreciate how poor CI width is going to be in their results. Therefore, this is an informed decision to be made by researchers. Practically, we give some recommendations for choosing a precision value (Table 2). In general, well-funded studies or large scale studies, aiming to

Read the complete text