Sec.2-Ch.1-Subsec.7：Single-Rate Meta-Analysis in R
From Data Preparation to Forest Plots—A Practical Case Study on the Prevalence of IBS in Patients with Dyspepsia
In today’s medical and clinical research, meta-analysis, as a method for synthesizing results from multiple studies, is receiving increasing attention. It not only improves statistical significance but also provides a more comprehensive perspective, helping decision-makers understand the effects of specific treatments or interventions.
Single-rate meta-analysis focuses specifically on the proportion of a particular event, such as the cure rate in a treatment group or the incidence of adverse events. A single rate refers to the proportion of a specific event occurring within a given sample. This analytical approach is widely used in epidemiology, clinical trials, and public health research because it effectively integrates event data from different studies and provides reliable statistical inference for broader populations.
I. Understanding Single-Rate Meta-Analysis
A single rate usually refers to the probability of a specific event (e.g., disease occurrence, treatment response) within a particular sample. For example, if 25 out of 100 patients develop a certain disease, the incidence rate is 25%.
The main purpose of single-rate meta-analysis is to summarize the incidence of the same event across multiple studies. It is widely used in epidemiology and clinical research. For instance, analyzing disease incidence across different regions can help inform public health policy and improve disease prevention strategies.
Advantages:
Data integration: Effectively combines data from different studies to provide more accurate estimates.
Increased sample size: Pooling multiple studies significantly increases the total sample size and statistical power.
Heterogeneity analysis: Differences may exist between studies; heterogeneity analysis helps explore the sources of variation.
II. Conducting Single-Rate Meta-Analysis in R
This article uses data from Ford et al.’s paper:
Systematic Review and Meta-analysis of the Prevalence of Irritable Bowel Syndrome in Individuals With Dyspepsia.
The study background indicates that dyspepsia and irritable bowel syndrome (IBS) are common and may coexist. Through systematic review and meta-analysis, the authors estimated the prevalence of IBS among patients with dyspepsia.
Relevant literature was retrieved from MEDLINE and EMBASE. Ultimately, 19 studies were included from 239 screened papers, covering 18,173 participants.
Results:
Prevalence of dyspepsia: 27%
Prevalence of IBS among dyspepsia patients: 37%
Prevalence of IBS in patients without dyspepsia: 7%
This indicates that IBS occurs eight times more frequently in dyspepsia patients than in the general population, suggesting a possible shared pathogenic mechanism. The study therefore recommends routine IBS assessment in patients with dyspepsia.
1. Dataset Construction
When performing a single-rate meta-analysis, the first step is to construct the dataset. In this study, the dataset ford.ibs is structured as follows:
study <- c("Talley", "Holtmann", "Holtmann", "Talley", "Agreus", "Schlemper", "Schlemper", "Kennedy", "Talley", "Caballero-Plasencia", "Shah", "Curioso", "Hu", "Locke", "Lu", "Papatheoridis", "Perona", "Minocha", "Bolling-Sternevald")
years <- c("1992", "1994", "1994", "1994", "1995", "1995", "1995", "1998", "1998", "1999", "2001", "2002", "2002", "2005", "2005", "2005", "2005", "2006", "2008")
event.e <- as.numeric(c(213, 44, 124, 200, 372, 46, 80, 833, 92, 63, 774, 87, 304, 90, 561, 339, 47, 247, 397))
event.n <- as.numeric(c(835, 180, 423, 919, 1154, 175, 473, 3169, 730, 264, 2549, 231, 1649, 643, 2018, 700, 70, 990, 1001))
ford.ibs <- data.frame(study, years, event.e, event.n)
ford.ibs