Background. Of all the diseases an individual can encounter in the world, malaria is one of the most destructive. Simple measures like sleeping under a bednet would greatly reduce the burden (Abeku, 2007). When people estimate their risk relative to others, they are most often unrealistically optimistic, which may explain why those at risk often fail to perform behaviors, such as using a bednet that will reduce their risk. However, one study showed that people at high risk for malaria held pessimistic perceptions of their risk for the disease, but the reasons for this finding are unclear (Morrison, Ager, & Willock, 1999).

Purpose. In this study I examined risk perceptions about malaria, specifically absolute and comparative risk for the disease, and the role such perceptions play in encouraging or discouraging preventive behavior in areas with high (>50 cases/1,000) and low (<1 case/1,000) rates of endemic malaria. In addition, I tested the accuracy hypothesis (Brewer, Cuite, Herrington, & Weinstein, 2004) between perceived risk of malaria and engaging in the preventive behavior of using a bednet.

Method. This study used a cross-sectional, non-equivalent group comparison observational design and was conducted in Belize, Central America. The data were collected using both self-reported surveys and personal interviews from residents of 20 selected villages across the country with high (n = 10 villages) and low (n = 10 villages) endemic malaria. A minimum of 15 people per village were recruited, for a total sample size of 300 people, with approximately equal numbers of males and females. The survey assessed absolute and comparative risk perceptions, based on the recommendations by Brewer et al., as well as perceived control and risk behaviors related to malaria, and standard demographics.

Results. More people in the high risk area had ever had malaria compared to those in the low risk area (42.7% vs. 8.8%, p<.001). Average perceived risk of ever getting malaria was 48.4%, with no significant difference between the high (51%) and low (46%) areas; however, those in the high area who had ever had malaria reported a significantly higher risk of ever getting malaria in the future (mean=59% vs. 45% chance, p=.003), indicating a fairly accurate perception among those in the high risk area but pessimism among those in the low risk area. However, participants in both high and low risk areas were most likely to show an optimistic bias for comparative risk items. When compared to others in their town, 47.1% thought they were below average, 32.5% average, and 20.3% above average risk (with no differences between high and low areas). Bednet use varied, with 40.7% of those in the high, and only 2.0% of those in the low, risk area always using a bednet (p<.001); those who had a history of malaria in the high risk area were more likely to always use a bednet than those who have no history of malaria (50.0% vs. 33.7%, p<.05). There was no evidence for the accuracy hypothesis: greater perceived risk regarding malaria was not associated with regular bednet usage. However, perceived control was significantly correlated with preventive behavior (r=0.236, p<0.001).

Conclusions. These results show both pessimistic and optimistic biases regarding risk of malaria, depending on the type of risk assessed (absolute vs. comparative risk), actual risk of malaria (high vs. low risk areas), and perceptions of control. However, risk perceptions regarding malaria were not reliably associated with preventive behavior, unlike that found by other researchers, although it is unclear why this is so. Future studies should be longitudinal in nature, to determine if risk perceptions and perceived control are related to future behavior. These studies must use standardized measures of risk perceptions to adequately examine risk-behavior hypotheses.