Risk Factors of Typhoid Infection in the Indonesian Archipelago

Abstract

Background: Knowledge of risk factors and their relative importance in different settings is essential to develop effective health education material for the prevention of typhoid. In this study, we examine the effect of household level and individual behavioural risk factors on the risk of typhoid in three Indonesian islands (Sulawesi, Kalimantan and Papua) in the Eastern Indonesian archipelago encompassing rural, peri-urban and urban areas.

Methods: We enrolled 933 patients above 10 years of age in a health facility-based case-control study between June 2010 and June 2011. Individuals suspected of typhoid were tested using the typhoid IgM lateral flow assay for the serodiagnosis of typhoid fever followed by blood culture testing. Cases and controls were defined post-recruitment: cases were individuals with a culture or serology positive result (n = 449); controls were individuals negative to both serology and culture, with or without a diagnosis other than typhoid (n = 484). Logistic regression was used to examine the effect of household level and individual level behavioural risk factors and we calculated the population attributable fraction (PAF) of removing each risk significant independent behavioural risk factor.

Results: Washing hands at critical moments of the day and washing hands with soap were strong independent protective factors for typhoid (OR = 0.38 95% CI 0.25 to 0.58 for each unit increase in hand washing frequency score with values between 0 = Never and 3 = Always; OR = 3.16 95% CI = 2.09 to 4.79 comparing washing hands with soap sometimes/never vs. often). These effects were independent of levels of access to water and sanitation. Up to two thirds of cases could be prevented by compliance to these practices (hand washing PAF = 66.8 95% CI 61.4 to 71.5; use of soap PAF = 61.9 95%CI 56.7 to 66.5). Eating food out in food stalls or restaurant was an important risk factor (OR = 6.9 95%CI 4.41 to 10.8 for every unit increase in frequency score).

Conclusions: Major gains could potentially be achieved in reducing the incidence of typhoid by ensuring adherence to adequate hand-washing practices alone. This confirms that there is a pivotal role for 'software' related interventions to encourage behavior change and create demand for goods and services, alongside development of water and sanitation infrastructure.

Fig 1. Location of study sites and participating health facilities.

Fig 2. Diagnostic algorithm.

¹ Fever on the day of consultation (body temperature measured axillary >37.5°C); fever duration ≥3 days; headache since the start of the fever. ² IgM lateral flow assay

Fig 3. Distribution of individual level risk factors by case definition [1].

¹ Type I controls were individuals negative to both serology and culture (n = 305); Type II controls were individuals negative to serology and with a diagnosis other than typhoid (n = 144); Type I cases were individuals with a culture positive result regardless of serological result (n = 174); Type II cases were individuals who had a culture negative but seropositive results (n = 310); ² Continuous score with values between 0 = Never and 3 = Always. ³ Categorical variable 0 = Sometimes/Never, 1 = Often, 2 = Always. ⁴ Categorical variable 1 = Less than once a week, 2 = Once a week, 3 = Few times a week, 4 = Every day