Introduction
The relationship between healthcare and statistical approaches is not new, particularly in medical research. However, the amount of available information that can help inform healthcare decisions and the application of science in the health industry has recently increased. For this reason, biostatisticians play a vital role in the medical landscape by giving room to reinforce decisions concerning patient care. They also help ensure a better focus of the medical research, hence providing meaning to the available data. Biostatistics is an area of study that links biology and statistics through the traditional findings approaches applied in clinical trials and public health. Biostatistics is an essential tool for clinical decisions, medical research, and health management as an applied science branch. Therefore, it is necessary to understand statistical principles and the scientific decision-making in research as influenced by biostatistics alongside calculation of the sample sizes in medical studies.
Biostatistics Role in Research
Using biostatistics, researchers can make inferences from the available information, particularly in the clinical field. In many cases, biostatisticians are typically called at the very beginning of research studies to help develop group objectives, data analysis techniques, and general study design to help in enhancing study outcomes (Vallat, 2018). With the evolution of the biostatistics methods in the past centuries, they are now capable of assisting in experimental design through perspective technique. The use of the p-value has made the use of biostatistics to attain desirable outcomes (Vallat, 2018). P-value helps the researcher in deciding on whether the available data could be realized by chance and is commonly helpful in clinical studies involving:
- Evaluating suggested clinical treatments
- Determining relative gains on various competing therapies
- Accessing possible optimal cure combinations
Biostatisticians help in the development of data management plans through the identification of potential areas of weakness during the process of data collection. This also helps support the high-level validity of data both at the collection and analysis stage. In addition, biostatisticians consider the information collected as a significant part of clinical research and use statistical approaches to summarize the info alongside reporting any abnormal data trends.
Moreover, medical trials evaluating patient results based on new treatment methods are usually drafted with proper considerations of biostatistics. In this case, biostatisticians operate behind the sights in designing fair clinical trials that give room for reliable data collection techniques (Vallat, 2018). Medical researchers are known to spearhead the development of new drugs in the industry and the evaluation of existing medical involvements affecting public wellbeing.
Advances in epidemiology as a critical aspect of community wellbeing policies and a significant cornerstone of preventive care are significantly reinforced by data from biostatisticians. In this regard, biostatistics focuses on the relationship between the effect and cause of diseases alongside other factors that might have influenced the general widespread of infection (Amrhein, Trafimow and Greenland, 2019). On the other hand, biostatistics can also identify the lack of connection between the theoretical bases of a disease, hence making researchers redirect their focus toward eliminating other possible risk factors.
Biostatistics research can also be applied locally to ensure the improvement of desired patient care. Data garnered from existing studies can be synthesized and then utilized to make more effective health systems or general individualized plans that will help to improve patient outcomes. The use of evidence-based assessments in biostatistics helps in eliminating bias in healthcare systems (Wainwright, 2019). In addition, the practice helps in managing the risk associated with ineffective patient care and treatment.
Moreover, it is essential to note that biostatistics does not always involve the analysis of newly collected data. Instead, it also implies a meta-analysis of various sources’ already available secondary data. Evaluation of clinical research around a specific disease, for instance, can focus on statistical trends revolving around known risk factors, anticipated lifespans, and the probability of any genetic factors. This kind of trend and statistical material can be used to identify the possible treatment approaches to aid in improved patient outcomes (Wainwright, 2019). In clinical research, biostatistics helps prevent fraud or any unintentional error that might occur in medical studies. Fraud occurs due to data fabrication or in the event of falsification, which tends to change data values. In reported cases, falsification involves cheating on inclusion measures where an ineligible individual is allowed to participate in a trial.
Sample Size Calculation
The optimal sample size is an important aspect worth considering for any research in various fields. The main aim of tester size calculation is to identify the number of samples required to arrive at desirable changes in treatment effects, clinical parameters, or relations after data collection (Van Smeden et al., 2019). In many instances, studies tend to be underpowered, thereby failing to determine the impact of the existing treatment due to insufficient model size.
In designing medical research, calculating sample sizes is one of the first significant steps to be considered. In this regard, sample size refers to a portion of the population that needs to be incorporated in research to respond to the available study hypothesis. The main focus of this calculation is to determine the sufficient units required to detect the unknown parameters (Van Smeden et al., 2019). In the events where the sample size is too small, the researcher may be unable to respond to the existing research question. Therefore, it is required that the investigator first determines the optimal sizes of samples before gathering data, as this will help avoid the mistakes associated with too small pieces.
In making the sample size calculations, various assumptions must be made, including type I and II errors, variability, and the lowest effect of interest. The outcome variability is usually estimated using the expected standard deviation approach, where researchers use available estimates from previous studies (Smeden et al., 2019). The type I fault occurs where a true null hypothesis is overruled, and the type II mistake is committed through failure to decline a false premise. The two appear when an investigator deals with a minimum sample to determine its effect on a particular size. The effect size refers to the difference between estimation and the anonymous factor that a researcher intends to evaluate.
Sample Size Calculations in Various Studies
When dealing with cross-section research, the main focus is to approximate the prevalence level of the unknown parameter from the object population through a random sample. Therefore, enough samples are required to determine the degree of population prevalence with proper precision (Lakens, 2022). As a result, the following formula can be used in a prevalence study to arrive at an appropriate sample size:
- n= Z2 P (1-P)
- d2
From the formula, n stands for the expected sample scope, Z represents confidence level, P denotes the prevalence expected, and d is for the precision. The confidence interval ranges from 95%-99%, depending on the researcher’s need (Braun and Clarke, 2021, p. 209). To use the above formula, the researcher must have the assumed value of P as the precision is typically selected depending on the amount of P. Since there are no specific criteria for choosing d, a conservative selection is known to be one-fifth or fourth of prevalence as the size of precision in the event of small P (Braun and Clarke, 2021). The table below presents calculations using three distinct values of P and d. From the table below, it is clear that precision changes depending on the assumed value of P, and an incorrect precision leads to the wrong sample scope (large or small).
Calculation of Sample Size in Case-Control Research
This involves an observational study where risk factors associated with a disease are compared using subjects affected (cases) and those not affected (controls). The assumptions that need to be made in this case include; the estimated size of the controls and subjects from the pilot and similar study, the confidence levels (95%), and suggested study power ( from 80%) (Braun and Clarke, 2021, p. 209). In the event of a significant perplexing factor, investigators must have a large sample size to attain the desired significance. Good research may fail to respond to the hypothesis in medical trials if the sample is small (Blaikie, 2018). For a randomized trial, availability of significance level, power, population event rate, and the scope of sought treatment effect are required to calculate the sample size.
Conclusion
Biostatistics plays a crucial role in decision-making by ensuring enhanced efficiency and efficacy in delivering health services and patient care. In other words, the part of biostatistics is seen in turning clinical trials and public wellbeing data into more practical knowledge. The optimal sample size is essential to any sample size as it helps detect the treatment effects. Therefore, the calculation of enough scope of the sample is a critical practice in any clinical research. Researchers need all material information for the sample size to be estimated using correct assumptions.
Reference List
Amrhein, V., Trafimow, D. and Greenland, S. (2019) ‘Inferential statistics as descriptive statistics: There is no replication crisis if we don’t expect replication,’ The American Statistician, 73 (sup1), pp.262-270.
Blaikie, N. (2018) ‘Confounding issues related to determining sample size in qualitative research,’ International Journal of Social Research Methodology, 21(5), pp.635-641.
Braun, V. and Clarke, V. (2021) ‘To saturate or not to saturate? Questioning data saturation as a useful concept for thematic analysis and sample-size rationales.’ Qualitative Research in Sport, Exercise and Health, 13(2), pp.201-216.
Lakens, D. (2022) ‘Sample size justification,’ Collabra: Psychology, 8(1), p.33267.
Vallat, R. (2018) ‘Pingouin: statistics in Python,’ J. Open Source Softw., 3(31), p.1026.
Van Smeden, M., et al. (2019) ‘Sample size for binary logistic prediction models: beyond events per variable criteria,’ Statistical Methods in Medical Research, 28(8), pp.2455-2474.
Wainwright, M.J. (2019) High-dimensional statistics: A non-asymptotic viewpoint (Vol. 48). Cambridge University Press.