Table 3.

 Details of analysis and results for between country comparison studies of legislation for presumed consent for organ donation that had a robust analysis

Type of analysis	Statistical significance of factors considered in regression analysis
	PC law (or practice)	CVA mortality	RTA mortality	GDP	Healthcare expenditure	Transplant capacity	Religion (Catholicism)	Education	Legislative system	Blood donation rate	Internet access
Abadie (2006)w6
Fixed regression with panel (longitudinal) data*	P≤0.05	P≤0.05	P≤0.05	P≤0.05	NS	—	NS	—	P≤0.05	NS	—
Neto (2007)w11
Quantile regression for panel (longitudinal) data†	P≤0.05	P≤0.05¶	P≤0.05	P≤0.05	P≤0.05	—	P≤0.05**	—	P≤0.05	—	P≤0.05††
Healy (2005)w9
Linear mixed-effects regression using time series data‡	NS	NS	P≤0.05	NS	NS	—	—	—	—	—	—
Gimbel (2003)w8
Linear ordinary least squares regression using single data point per country§	P≤0.05	—	—	—	—	P≤0.05	P≤0.05	P≤0.05	—	—	—

PC = presumed consent, CVA = cerebrovascular, RTA = road traffic accident, GDP = gross domestic product, NS = not significant.

*Different combinations of variables considered in a series of models.

†Analysis based on Koenker 2004.⁸ Two models were used—one with GDP and one with health expenditure (these were highly collinear). A generalised least squares regression was also performed for comparison.

‡The initial model did not fit the data, and the analysis was repeated excluding outliers (Spain and Italy).

§This study classified countries based on whether there was presumed consent in practice rather than whether presumed consent legislation was in place.

¶Significant in model using health expenditure per capita but not GDP per capita.

**Significant at 25th centile only on one model and 25th and 50th centiles but not the 75th.

††Significant for 25th and 75th centiles.