Questionnaire DesignThe questionnai

Questionnaire Design
The questionnaire was developed and assessed as previously reported.[10] Items were constructed based on information obtained through in-depth interviews with a variety of individuals identified through travel clinics; those who had previously traveled to malaria-endemic areas, individuals planning to travel to malaria-endemic areas for the first time, and those who had no experience of or planning to travel to malaria-endemic areas. From the themes that emerged from the interviews, a questionnaire was constructed using closed question and Likert-type scales. Face validity of the questionnaire was assessed and subsequently confirmed through review of the questionnaire by practitioners (a nurse, pharmacist, and physician) in travel medicine. The test–retest method with a 2-week interval was employed to examine the questionnaire reliability. Usability was further investigated through pilot studies. In total the questionnaire consisted of:

7 Questions relating to traveler details, eg, age, gender, and itinerary
8 Questions on specific use of bite-avoidance measures, eg, repellents and whether they were frequently, sometimes, or never used
3 Questions relating to experience while away, eg, whether many, few, or no mosquito bites were experienced
17 Attitude statements exploring attitudes toward different methods of bite avoidance on a 1 to 5 Likert scale indicating level of agreement to the statement. Examples of items include, “Repellents are unpleasant to use,” and “Using insecticide sprays on clothes reduces mosquito bites.”
Further four statements relate to attitudes toward malaria measured on a 1 to 5 Likert scale. An example includes, “Malaria is a severe condition.”
The total score of the attitude scale was calculated by reversing selected negatively worded items and summing them.

Analysis
Data were analyzed using ibm spss version 20 for Windows. Prior to detailed analysis of the data, the total score for the attitude scale was obtained by first reverse-scoring the relevant items such that higher scores indicate a positive attitude and then summing all scores for the attitude scale items. In addition, various subscales were constructed based on categories of bite-avoidance measures. Reliability was measured by calculating the Cronbach's alpha.[11] Categorical relationships were examined using Chi-square test of independence with Pearson Chi-square being used where expected counts were less than five in more than three cells. Other relationships were explored by linear regression. Multinomial logistic regression was employed to examine the predictive relationship between the attitude scale scores and the behavioral outcome of bite-avoidance measures undertaken.