TY - JOUR
T1 - Be careful with ecological associations
AU - Roumeliotis, Stefanos
AU - Abd ElHafeez, Samar
AU - Jager, Kitty J.
AU - Dekker, Friedo W.
AU - Stel, Vianda S.
AU - Pitino, Annalisa
AU - Zoccali, Carmine
AU - Tripepi, Giovanni
N1 - Publisher Copyright: © 2021 Asian Pacific Society of Nephrology
PY - 2021/6
Y1 - 2021/6
N2 - Ecological studies are observational studies commonly used in public health research. The main characteristic of this study design is that the statistical analysis is based on pooled (i.e., aggregated) rather than on individual data. Thus, patient-level information such as age, gender, income and disease condition are not considered as individual characteristics but as mean values or frequencies, calculated at country or community level. Ecological studies can be used to compare the aggregated prevalence and incidence data of a given condition across different geographical areas, to assess time-related trends of the frequency of a pre-defined disease/condition, to identify factors explaining changes in health indicators over time in specific populations, to discriminate genetic from environmental causes of geographical variation in disease, or to investigate the relationship between a population-level exposure and a specific disease or condition. The major pitfall in ecological studies is the ecological fallacy, a bias which occurs when conclusions about individuals are erroneously deduced from results about the group to which those individuals belong. In this paper, by using a series of examples, we provide a general explanation of the ecological studies and provide some useful elements to recognize or suspect ecological fallacy in this type of studies.
AB - Ecological studies are observational studies commonly used in public health research. The main characteristic of this study design is that the statistical analysis is based on pooled (i.e., aggregated) rather than on individual data. Thus, patient-level information such as age, gender, income and disease condition are not considered as individual characteristics but as mean values or frequencies, calculated at country or community level. Ecological studies can be used to compare the aggregated prevalence and incidence data of a given condition across different geographical areas, to assess time-related trends of the frequency of a pre-defined disease/condition, to identify factors explaining changes in health indicators over time in specific populations, to discriminate genetic from environmental causes of geographical variation in disease, or to investigate the relationship between a population-level exposure and a specific disease or condition. The major pitfall in ecological studies is the ecological fallacy, a bias which occurs when conclusions about individuals are erroneously deduced from results about the group to which those individuals belong. In this paper, by using a series of examples, we provide a general explanation of the ecological studies and provide some useful elements to recognize or suspect ecological fallacy in this type of studies.
KW - confounding
KW - ecological fallacy
KW - ecological studies
UR - http://www.scopus.com/inward/record.url?scp=85100878354&partnerID=8YFLogxK
U2 - https://doi.org/10.1111/nep.13861
DO - https://doi.org/10.1111/nep.13861
M3 - Review article
C2 - 33570780
SN - 1320-5358
VL - 26
SP - 501
EP - 505
JO - Nephrology (Carlton, Vic.)
JF - Nephrology (Carlton, Vic.)
IS - 6
ER -