To the Editor:
Syed et al1 recently published an observational study using real-world data to study patients with gastroparesis in a national population in the United States. We read with interest this paper and would like to congratulate the authors on publishing this very first epidemiologic study using a large-scale secondary database to estimate the prevalence of gastroparesis in the United States. The study used Explorys, an electronic medical records database with ICD-9 codes indexed diagnoses, to estimate the proportion of study population with gastroparesis. The unique strength of the study was that the database has a wide coverage of over 43 million patients from over 340 hospitals in 22 different health care systems in the United States and allowed an estimation of gastroparesis prevalence at the national level. The results showed that out of a total of 43,827,910 medical records, 69,950 (0.16%) have a diagnosis of gastroparesis recorded. The authors also observed concurrent gastroparesis in 4.59% medical records with type I diabetes mellitus, 1.31% medical records with type II diabetes mellitus, and 0.05% medical records with no history of any potential etiologies of gastroparesis.
As the authors correctly pointed out, previous literature on the prevalence of gastroparesis was limited to a single community and cannot be extrapolated to the wider US general population due to the lack of diversity in economic, geographic, and demographic characteristics.2,3 This new observational study mitigated the gap in disease occurrence knowledge using real-world evidence from a large representative sample of the United States general population. The traditional method of prevalence and incidence assessment involves complex sampling strategies in national surveys such as the National Health and Nutrition Examination Survey (NHANES) and has proven to be expensive, time-consuming, and unable to capture changing dynamics of a population.4 Meanwhile, as part of the medical practice for administrative purposes, secondary databases that are frequently updated over a wide population and readily accessible at a low cost have provided new data sources for epidemiologists and other researchers. It is also well-recognized that secondary databases are not collected for research purposes, often constrained by sociologic and economic limitations, and are censored.4 Therefore, all secondary database researches should implement a rigorous methodology, provide adequate documentation of methods and results, and adequately address the limitations to ensure quality and integrity. In this study, to the authors’ credit, they noted the retrospective nature of the study and addressed potential misclassification due to documentation errors.
On the basis of the findings of this study, we have a few suggestions for future research. First, the study of prevalence usually seeks to understand “the proportion of the population of the disease at the specified time.”5 This study reported the proportion of medical records which is associated with prevalence. The additional patient-level analysis would provide a clearer view of the burden of gastroparesis in the population. Second, in left and right-censored longitudinal databases, the length of lookback period will affect the ability to detect a disease and influence temporal trends.4 The impact of different lookback windows should be thoroughly assessed to ensure the robustness of the findings. Last, Syed and colleagues correctly pointed out that differences in study populations prevent meaningful direct comparisons with previous prevalence estimation of gastroparesis. This study has unique strengths, including a large covered population and diversity in patient demographics; however, more studies with rigorous methodology in additional populations are necessary to further understand the epidemiology of gastroparesis.
Yizhou Ye, PhD*
Dimitri Bennett, MD*†
*Department of Epidemiology, Takeda Pharmaceutical Company Limited Cambridge, MA
†Center for Clinical Epidemiology and Biostatistics, Perelman School of Medicine University of Pennsylvania, Philadelphia, PA
1. Syed AR, Wolfe MM, Calles-Escandon J. Epidemiology and diagnosis of gastroparesis in the United States: a population-based study. J Clin Gastroenterol. 2019. [Epub ahead of print].
2. Jung HK, Choung RS, Locke GR III, et al. The incidence, prevalence, and outcomes of patients with gastroparesis in Olmsted County, Minnesota, from 1996 to 2006. Gastroenterology. 2009;136:1225–1233.
3. Choung RS, Locke GR III, Schleck CD, et al. Risk of gastroparesis in subjects with type 1 and 2 diabetes in the general population. Am J Gastroenterol. 2012;107:82–88.
4. Rassen JA, Bartels DB, Schneeweiss S, et al. Measuring prevalence and incidence of chronic conditions in claims and electronic health record databases. Clin Epidemiol. 2019;11:1–15.
5. Rothman KJ, Greenland S, Lash TL. Modern Epidemiology. Philadelphia, PA: Wolters Kluwer Health/Lippincott Williams & Wilkins; 2008.