Should young people be paid for getting tested? A national comparative study to evaluate patient financial incentives for chlamydia screening

Table 3 Linear regression model of the difference in average percentage point change in screening coverage between exposed and unexposed PCTs

	Number of PCT Pairs	Difference in average % change	95% Conf. interval	P value	P value (Effect modif.)
Any financial incentive	42	0.43%	(0.04; 0.82)	0.03
Males	42	0.14%	(-0.29%; 0.57%)	0.5
Females	42	0.73%	(0.30%; 1.17%)	0.001	0.002
Prize draw	36	0.16%	(-0.22%; 0.54%)	0.4
Vouchers	6	2.35%	(1.55%; 3.14%)	<0.0001	<0.0001
Prize draw in males	36	-0.05%	(-0.47%; 0.38%)	0.8
Prize draw in females	36	0.37%	(-0.06%; 0.80%)	0.09	0.03
Vouchers in males	6	1.55%	(0.63%; 2.46%)	0.001
Vouchers in females	6	3.18%	(2.25%; 4.11%)	<0.0001	0.0005
Prize draw via outreach	19	0.14%	(-0.40%; 0.68%)	0.606
Prize draw via post	11	0.16%	(-0.39%; 0.71%)	0.57
Prize draw via clinic	2	1.58%	(0.24%; 2.92%)	0.021
Prize draw via other	4	0.03%	(-0.99%; 1.05%)	0.95	0.15
Voucher via outreach	1	3.15%	(1.11%; 5.20%)	0.003
Voucher via post	2	3.66%	(2.24%; 5.08%)	<0.0001
Voucher via clinic	1	2.24%	(0.25%; 4.23%)	0.027
Voucher via other	2	1.10%	(-0.13%; 2.34%)	0.08	0.02
£5-9 Voucher	2	-0.66%	(-2.22%; 0.9%)	0.4
£10 Voucher	4	2.65%	(1.70%; 3.59%)	<0.0001	0.36

The above is from a number of different models of screening coverage used in the analysis. The results are adjusted for the IMD quintile of the PCT and the pairing of the analysis. P Values in the first column are a test of whether the effect of the PFI is different from zero. P Values in the second column (effect modification) are a test of whether the effect of the PFI varies by the type of PFI (or by gender). Both significance tests are Wald tests.
We explored but found no evidence for effect modification by age group or by the length of time an incentive was run in a quarter affected by the PFI (1, 2 or 3 months). There was also no significant effect of health promotion campaigns or "simultaneous mail-out" (whether or not a mail-out was undertaken to promote an incentive).
All differences are between the average of PCTs with a particular sort of incentive and the average of PCTs without an incentive.