TY - JOUR
T1 - Healthy Donor Effect and Satisfaction with Health
T2 - The role of selection effects related to blood donation behavior
AU - Shehu, Edlira
AU - Hofmann, Annette
AU - Clement, Michel
AU - Langmaack, Ann-Christin
PY - 2015/9
Y1 - 2015/9
N2 - The objective of this paper is to quantify selection effects related to blood donation behavior and their impact on donors’ perceived health status. We rely on data from the 2009 and 2010 survey waves of the German socio-economic panel (N = 12,000), including information on health-related, demographic and psychographic factors as well as monetary donation behavior and volunteer work. We propose a propensity score matching approach to control for the healthy donor effect related to the health requirements for active blood donations. We estimate two separate models and quantify selection biases between (1) active and inactive blood donors and (2) active donors and non-donors. Our results reveal that active donors are more satisfied with their health status; after controlling for selection effects, however, the differences become non-significant, revealing selection biases of up to 82 % compared with non-donors. These differences also exist between active and inactive donors, but the differences are less distinct. Our methodological approach reveals and quantifies selection biases attributable to the healthy donor effect. These biases are substantial enough to lead to erroneous statistical artifacts, implying that researchers should rigorously control for selection biases when comparing the health outcomes of different blood donor groups.
AB - The objective of this paper is to quantify selection effects related to blood donation behavior and their impact on donors’ perceived health status. We rely on data from the 2009 and 2010 survey waves of the German socio-economic panel (N = 12,000), including information on health-related, demographic and psychographic factors as well as monetary donation behavior and volunteer work. We propose a propensity score matching approach to control for the healthy donor effect related to the health requirements for active blood donations. We estimate two separate models and quantify selection biases between (1) active and inactive blood donors and (2) active donors and non-donors. Our results reveal that active donors are more satisfied with their health status; after controlling for selection effects, however, the differences become non-significant, revealing selection biases of up to 82 % compared with non-donors. These differences also exist between active and inactive donors, but the differences are less distinct. Our methodological approach reveals and quantifies selection biases attributable to the healthy donor effect. These biases are substantial enough to lead to erroneous statistical artifacts, implying that researchers should rigorously control for selection biases when comparing the health outcomes of different blood donor groups.
KW - Propensity score matching
KW - Blood donation services
KW - Socio-economic panel
KW - Donor management
KW - Propensity score matching
KW - Blood donation services
KW - Socio-economic panel
KW - Donor management
U2 - 10.1007/s10198-014-0625-1
DO - 10.1007/s10198-014-0625-1
M3 - Journal article
SN - 1618-7598
VL - 16
SP - 733
EP - 745
JO - European Journal of Health Economics
JF - European Journal of Health Economics
IS - 7
ER -