Table 3:

Comparison of quantitative survey responses in intervention versus control groups

	Intervention: mean scores			Control: mean scores			Difference in score changes: intervention vs control (95% CI)	p value*
	Baseline	Year 1	Change (95% CI)	Baseline	Year 1	Change (95% CI)
How familiar are you with IPT?†	3·72	4·73	+0·52 (0·03 to 1·0)	4·05	4·09	+0·05 (−0·46 to 0·55)	+0·47 (0·44 to 0·80)	0·0034
How strong is the evidence that isoniazid prevents active tuberculosis in HIV-infected patients?‡	3·75	4·38	+0·63 (−0·1 to 1·36)	4·14	4·18	+0·05 (−0·63 to 0·72)	+0·59 (0·06 to 1·12)	0.015
How difficult is it for providers in this district to add isoniazid to standard care for HIV-positive people in order to prevent tuberculosis?§	2·38	2·23	−0·15 (−0·99 to 0·69)	2·52	2·15	−0·36 (−1·31 to 0·58)	+0·21 (−0·26 to 0·69)	0·183
How hard is it to influence changes in practice among frontline providers around tuberculosis management?§	2·78	2·42	−0·37 (−1·0 to 0·27)	2·68	2·45	−0·23 (−0·86 to 0·41)	−0·14 (−0·62 to 0·35)	0·282

The left column shows the survey questions, which were scored on a Likert scale with a range of 1–5. Responses to 1 and 5 scores are listed in the footnotes. IPT=isoniazid preventive therapy.

^*One-sided p value.

^†1=no knowledge of IPT, 5=high knowledge of IPT.

^‡1=very weak, 5=very strong.

^§1=very easy, 5=very difficult; declining score (negative change) indicates decreasing difficulty (ie, increasing ease) for these questions.