Table 5.

Protective effectiveness of Vi polysaccharide and better household water, sanitation, and hygiene prevalence stratified by water, sanitation, and hygiene prevalence tertile and vaccine status

WASH prevalence	ViPS cluster			Hep A cluster			ViPS PE
	n	PY	Typhoid (IR per 1,000 PY)	n	PY	Typhoid (IR per 1,000 PY)	Crude PE, % (95% CI)	Adjusted PE, % (95% CI)*
Lower tertile†	7,673	14,612	9 (0.62)	12,813	24,452	72 (2.94)	79‡ (62–89)	67§ (27–85)‖
Middle tertile†	9,615	18,394	19 (1.03)	11,549	21,912	38 (1.73)	40 (–19 to 70)	48¶ (6–71)#
Upper tertile†	13,787	26,391	22 (0.83)	7,319	14,023	18 (1.28)	35 (–27 to 67	50¶ (10–73)#

Hep A = hepatitis A; IR = incidence rate; PE = protective effectiveness; PY = person-years; ViPS = Vi polysaccharide; WASH = water, sanitation, and hygiene. In models that expressed better WASH prevalence on a continuous scale, in the ViPS clusters, each 1% increase in better WASH prevalence was associated with a preventive impact against typhoid of –0.06% (P = 0.94; 95% CI, –1.59 to 1.45; crude) and –0.58% (P = 0.55; 95% CI, –2.50 to 1.30; adjusted).* In Hep A clusters, each 1% increase in better WASH prevalence was associated with a preventive impact against typhoid of 2.52% (P < 0.01; 95% CI, 0.93–408; crude) and 0.96% (P = 0.3; 95% CI, –1.00 to 2.88; adjusted).†

^*All models were adjusted for cluster stratification variables (ward [wards 29 and 30], number of residents ≤ 18 years [< 200 persons, ≥ 200 persons]; number of residents >18 years [< 500 persons, ≥ 500 persons]), design effect, and selected covariates.

^†Lower, < 6% prevalence; middle, 6% to 26% prevalence; upper, ≥ 26% prevalence.

^‡P < 0.001.

^§P < 0.01.

^‖ Model adjusted for the covariates age, Hindu religion, household population size, and longer distance to the nearest treatment center than median.

^¶P < 0.05.

# Model adjusted for the covariates age, household population size, and monthly household expenditures.