An epidemiological investigation into the use of human health databases for surveillance of zoonotic enteric disease in Alberta.

摘要

Surveillance of enteric disease in Canada has traditionally relied upon data collected in passive surveillance systems. Information captured by passive surveillance systems, particularly for non-life threatening illnesses, suffers from under-reporting bias. An ecological study of the incidence and distribution of zoonotic enteric disease in Alberta was performed.; We examined the geographic and temporal distribution of zoonotic, enteric disease cases extracted from four human health datasets between 1995--1998. The databases included the Alberta Notifiable Disease Registry and three administrative datasets. Little correlation was seen among any of the datasets. However, when the geographic patterns of enteric cases were compared with those of selected control diseases, the correlation was high and significant within datasets. Results from further analyses showed that availability of health care services was having an important influence on which and how many enteric cases, and controls, were entered into the four datasets in Alberta.; The incidence of zoonotic enteric disease was calculated first for counties and then for communities within counties. The pattern of enteric disease changed depending on which geographic scale was used. Investigation by community provided a more complete picture and enabled us to identify where enteric disease incidence was highest in Alberta. However, lack of complete address information provided with the datasets limited our ability to correctly geolocate cases, particularly in rural areas. Population estimates taken from the Postal Code Conversion File (Statistics Canada) provided the most appropriate information for rate calculations by community.; We concluded that the Notifiable Disease Registry was the most appropriate dataset for surveillance of zoonotic enteric disease in Alberta using the methods employed in this study. Accordingly, the incidence of enteric disease was highest in Southern Alberta and where county cattle density was higher. However, neither the quality nor the specificity of the data collected in any of the four datasets was good enough to establish causal associations between risk factors and disease incidence. Until the quality and specificity of the data collected improve, these human health databases should only be used to calculate descriptive statistics and develop hypotheses for future testing.