Open AccessAge-Standardized Mortality Rates in the Caribbean: One Source, Three Different Interpretations
Author(s) -
Nicholas Wolf,
Camille Morgan,
John Flanigan
Publication year - 2018
Publication title -
journal of global oncology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.002
H-Index - 17
ISSN - 2378-9506
DOI - 10.1200/jgo.18.19000
Subject(s) - demography , medicine , jurisdiction , mortality rate , population , caribbean region , geography , latin americans , environmental health , surgery , linguistics , philosophy , sociology , political science , law
74 Purpose A recent publication in the Morbidity and Mortality Weekly Report (MMWR) provided the opportunity to calculate differences in published cancer mortality estimates for Caribbean jurisdictions from three organizations, MMWR, the International Agency for Research on Cancer, and the Institute for Health Metrics and Evaluation. This comparison study serves to educate end users of these data.Methods We downloaded the publicly available cancer mortality estimates for 15 jurisdictions and the United States from the three organizations. We compared reported age-standardized mortality rates for each jurisdiction, and calculated the range among the estimates for each jurisdiction. We repeated this analysis after applying the same world population standard to all estimates.Results For males, ranges of Caribbean estimates were between 49% (Grenada and Trinidad) and 201% (US Virgin Islands) of the MMWR value, with an average of 88%. For females, ranges were between 15% (Trinidad) and 171% (US Virgin Islands) of the MMWR value, with an average of 64%. After all estimates were compared using the same population standard, ranges of Caribbean estimates for males were between 6% (Grenada) and 111% (US Virgin Islands) of the MMWR value, with an average of 34%. For females, ranges were between 7% (Grenada) and 97% (US Virgin Islands), with an average of 28%.Conclusion The use of different standard populations complicates comparisons across organizations. Data modeling does not completely compensate for the quality of source data, as our analysis has demonstrated by the differences in mortality rates despite the good quality of the vital registration in the Caribbean.AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST No COIs from the authors.