year 16, Issue 2 (March - April 2022)
Iran J Med Microbiol 2022, 16(2): 173-176 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Alimohamadi Y, Sepandi M. Sample Size in Animal Studies (The number of laboratory animals in a Research study). Iran J Med Microbiol 2022; 16 (2) :173-176
URL: http://ijmm.ir/article-1-1566-en.html
URL: http://ijmm.ir/article-1-1566-en.html
1- Health Research Center, Life Style Institute, Baqiyatallah University of Medical Sciences, Tehran, Iran
2- Exercise Physiology Research Center, Life Style Institute, Baqiyatallah University of Medical Sciences, Tehran, Iran ,msepandi@gmail.com
2- Exercise Physiology Research Center, Life Style Institute, Baqiyatallah University of Medical Sciences, Tehran, Iran ,
Full-Text [PDF 542 kb]
(1314 Downloads)
| Abstract (HTML) (2528 Views)
Full-Text: (2928 Views)
Dear Editor
Prior to initiating interventional clinical studies in humans, researchers usually conduct in-silico (computer simulation) in vitro and in vivo studies in order to evaluate the initial efficacy, dose determination and toxicity, pharmacokinetics, and pharmacodynamics of the interventions. Accurate determination of sample size in animal studies is essential both in terms of the validity of scientific inference and animal rights. Insufficient sample size reduces the probability of detecting the relationship between variables. Furthermore, excessive sample size imposes unnecessary costs and kills more animals (1). According to paragraph 12 of the Helsinki Declaration, the rights of laboratory animals in research need to be respected (2).
The main approach recommended for sample size determination is a method based on statistical power analysis. However, calculating the sample size with this approach, on the one hand, requires the researcher familiarity with the basic concepts of statistical knowledge such as statistical power, statistical errors, and the concept of effect size, and on the other hand, this approach requires the use of data either from previous studies or pilot study. Therefore, in cases where the use of this method is not feasible, the resource equation method introduced by Mead in 1988 (3) could be used. The resource equation method is based on analysis of variance (ANOVA) and is applicable to all studies that use animals. In this method, it is assumed that if the sample size is calculated correctly, the degree of freedom (DF) would be between 10 and 20. If the DF is less than 10, as the sample size increases the probability of obtaining significant results would be increased; and when the DF is greater than 20, it means that an increase in the sample size may not increase the probability of detecting a significant result. Therefore, the general formula for the sample size determination in a study that aims to compare a quantitative variable in two (or more) groups would be as follows, where k represents the number of groups and n represents the sample size required:
Prior to initiating interventional clinical studies in humans, researchers usually conduct in-silico (computer simulation) in vitro and in vivo studies in order to evaluate the initial efficacy, dose determination and toxicity, pharmacokinetics, and pharmacodynamics of the interventions. Accurate determination of sample size in animal studies is essential both in terms of the validity of scientific inference and animal rights. Insufficient sample size reduces the probability of detecting the relationship between variables. Furthermore, excessive sample size imposes unnecessary costs and kills more animals (1). According to paragraph 12 of the Helsinki Declaration, the rights of laboratory animals in research need to be respected (2).
The main approach recommended for sample size determination is a method based on statistical power analysis. However, calculating the sample size with this approach, on the one hand, requires the researcher familiarity with the basic concepts of statistical knowledge such as statistical power, statistical errors, and the concept of effect size, and on the other hand, this approach requires the use of data either from previous studies or pilot study. Therefore, in cases where the use of this method is not feasible, the resource equation method introduced by Mead in 1988 (3) could be used. The resource equation method is based on analysis of variance (ANOVA) and is applicable to all studies that use animals. In this method, it is assumed that if the sample size is calculated correctly, the degree of freedom (DF) would be between 10 and 20. If the DF is less than 10, as the sample size increases the probability of obtaining significant results would be increased; and when the DF is greater than 20, it means that an increase in the sample size may not increase the probability of detecting a significant result. Therefore, the general formula for the sample size determination in a study that aims to compare a quantitative variable in two (or more) groups would be as follows, where k represents the number of groups and n represents the sample size required:
n = (DF / k) + 1
For example, suppose in a study with the aim of comparing 4 groups of rats:
n = (10/4) + 1 = 3.5 = 4 assuming DF = 10
n = (20/4) + 1 = 6 assuming DF = 20
n = (20/4) + 1 = 6 assuming DF = 20
Therefore, in this study, we need at least 4 and at most 6 rats in each group. It is worth noting to say if there is a possibility of attrition in the study, the percentage of attrition in the final sample size must be considered. Although the resource equation method is a simple method, it is not as accurate as the power analysis method. Therefore, it is recommended to use special software to calculate the sample size. One of the best options in this regard (for both clinical and animal studies) is the G-Power software (4). Using this software, the required sample size could be determined for studies in which various statistical methods such as correlation, mean difference test, regression, and other statistical methods, based on different effect sizes (large, medium, and small) are used (4). Other advantages include a short manual, user-friendliness, and availability of the software (5).
Type of Study: Letter to the Editor |
Subject:
Deep learning
Received: 2021/11/27 | Accepted: 2022/01/14 | ePublished: 2022/02/10
Received: 2021/11/27 | Accepted: 2022/01/14 | ePublished: 2022/02/10
References
1. Fitts DA. Ethics and animal numbers: informal analyses, uncertain sample sizes, inefficient replications, and type I errors. J Am Assoc Lab Anim Sci. 2011;50(4):445-53.
2. Organization WMAJBotWH. World Medical Association Declaration of Helsinki. Ethical principles for medical research involving human subjects. 2001;79(4):373.
3. Festing MF, Altman DGJIj. Guidelines for the design and statistical analysis of experiments using laboratory animals. 2002;43(4):244-58. [DOI:10.1093/ilar.43.4.244] [PMID]
4. Faul F, Erdfelder E, Lang AG, Buchner A. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods. 2007;39(2):175-91. [DOI:10.3758/BF03193146] [PMID]
5. Kang HJJoeefhp. Sample size determination and power analysis using the G* Power software. 2021;18. [DOI:10.3352/jeehp.2021.18.17] [PMID] [PMCID]
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
Copyright Policy
Iranian Journal of Medical Microbiology by Farname is licensed under CC BY-NC 4.0
