Table 1 summary of findings from included studies

Phillips-Howard et al. [15]

British travellers

326

Cohort study

48%

Complex information and receiving conflicting advice when they contacted other advisory services were recorded as reasons for non-adherence, with only 257 travellers felt to have fully understood the information. Forgetting, considering prophylaxis to be unnecessary, advised to stop (by local people or members of their peer group), side-effects and illness, and travelling at short notice so no tablets available

Low

Lobel et al. [16]

US citizens travelling to Haiti or Africa

4042

Cross-sectional survey

42.4%

Prophylaxis use was associated with receiving pretravel advice (p < 0.001) (forewarning risk of malaria, chemoprophylaxis recommendation, medical consultation). The source that provided the information was not an influence on chemoprophylaxis use

High

Hilton et al. [17]

US travellers

214

Retrospective telephone interview

81% in patients older than 40, 59% of patients younger than 40

A greater number of travellers older than 40 were fully adherent compared with those under 40 (significance not reported)

Low

Lobel et al. [18]

European and North American travellers to Kenya

5489

Longitudinal study

52% (of 3469 travellers who used chemoprophylaxis)

Business travellers (compared with tourists) were significantly more likely to be non-adherent (26%, p < 0.001), as were those visiting friends and relatives (32%, p < 0.001) compared with tourists. Logistic regression analysis showed that adherence was particularly poor among people who visited friends and relatives (37%), travelled for more than three weeks (39%), experienced adverse reactions (40%), used proguanil (31%), and among young travellers from the UK (43%) (P < 0.001)

High

Steffen et al. [19]

European travellers

42,202

Cross-sectional survey

55.4% (mean)

Weekly medication (p < 0.001) was significantly associated with increased adherence when compared with daily, twice weekly and thrice weekly medications. Those who stayed abroad over 3 months (p < 0.001) compared with those travelling for less than 3 months were significantly more likely to be non-adherent as were those who had had several previous journeys compared to those on their first journey to the tropics (p < 0.001). Severe adverse drug reactions were experienced by 1.5–4.4% of travellers and caused at least 200 of them to prematurely stop their chemoprophylaxis

High

Held et al. [20]

European or North American travellers returning to Berlin

507

Retrospective study

38% (195/507)

Significantly higher adherence was noted amongst: patients using only one information source compared to those who used no information source (p = 0.0026); shorter travel duration (37.2 ± 38.5 days (mean ± SD) in contrast to 69.8 ± 93.5 days in the group of patients with no adherence

p = 0.00001); older patients compared to those aged under 55 (20/27 adherent > 54 compared with 175/476 < 55, p = 0.0001); travel destination with Southern and East African destinations showing highest levels of adherence (p = 0.0054); package tours (p = 0.0001) compared with those who had organised the travel themselves

Medium

Cobelens et al. [21]

Dutch travellers

547

Cross-sectional survey

60%

Adherence was significantly affected by geographical areas travelled to (p < 0.001)—for example adherence was 45% in South America compared with 78% in East Africa, those aged under 29 (p = 0.027) were significantly more likely to be non-adherent compared with those aged over 29, those who had previous travel experience (p = 0.031) were significantly more likely to be non-adherent and those with adventurous travel style (p < 0.001)—such as backpacking—were also more likely to be non-adherent compared with their non-adventurous counterparts

Gender, education and travel purpose were not significantly associated with adherence in a logistic regression model

Self-reported reasons for early discontinuation included believing it unnecessary to continue prophylaxis during part of the journey due to the perceived low risk (43.9%), stopped on advice by others (12.2%), experiencing adverse reactions (11.6%), negligence (17.1%). Other reasons mentioned (all less than 5%) included: lack of awareness of the need to continue, loss of tablets, bad taste of the tablets, insufficient amounts prescribed, fear of developing adverse effects and (possible) pregnancy

Medium

Chatterjee [22]

Travellers to India

452

Cross-sectional survey

71% (320/452)

Females appeared more likely to be non-adherent than males (significance not reported), along with travellers under 30 years old (significance not reported). Travellers on visits lasting longer than 3 weeks tended to be less adherent (significance not reported)

Reasons for poor adherence included: inadequate dose or incorrect drug (21%); pretravel information deficit (45%); active decision (33%); side-effects (25%)

Low

Banerjee et al. [23]

UK GPs travelling to South Asia

145

Telephone survey

46%

Some self-reported reasons for poor adherence included: thinking the area was free from malaria (34%); not wishing to take prophylaxis (18%); experiencing previous side-effects (10%); believing they had long-term immunity (10%); had no time to obtain prophylaxis (4%); costs (2%) and thinking it was easier to cure than to take the medication (2%); travelled for a short period and took the risk (2%)

Low

Lobel et al. [24]

North American and European travellers to East Africa

6633

Cross-sectional study

61.7%

Adherence was lowest in those who used a daily drug as opposed to a weekly schedule (OR = 4.03; 95% CI 3.32–4.89), attributed an adverse event (OR = 2.23; 95% CI 1.80–2.76) to the prophylaxis, stayed more than a month (OR = 3.32; 95% CI 2.64–4.18), those who were non-tourism travellers (OR = 3.04; 95% CI 2.42–3.82), those aged under 40 (OR = 2.19; 95% CI 1.76–2.71)

High

Farquharson et al. [25]

Travellers attending a travel medicine clinic

130

Prospective study using regression analysis

62% full adherence and 25% partial adherence

There were no significant differences across the adherence groups for age, gender, ethnicity, nationality, education, previous travel or previous experience of anti-malarial medication

Multiple logistic regression showed that poor adherence (compared to full adherence) was associated with greater amounts of health professional discussion about adherence in the medical consultation (OR = 0.7, 95% CI 0.6–0.9). Increased likelihood of full adherence (compared to partial adherence) was associated with perceived benefits of taking prophylaxis (OR = 1.4, 95% CI 1.1–1.9), going for a longer trip (OR = 3.6, 95% CI 1.5–8.7), and greater amounts of traveller information and questions in the consultation (OR = 1.0, 95% CI 0.0–1.1). Poor adherence (compared to partial adherence) was associated with going for a longer trip (OR = 0.2, 95% CI 0.1–0.6) and greater amounts of traveller information and questions (OR = 1.0, 95% CI 0.9–1.0)

Medium

Jute et al. [26]

Expatriates working on a Mali mine

90

Cross-sectional survey

72%

Some self-reported reasons for poor adherence included concerns over adverse side-effects, presumed immunity from long-term residence in Africa and a high standard of on-site care

Low

Hamer et al. [27]

Expatriate corporate workers in Ghana

42

Cross-sectional survey

0% amongst those based over a year

81% of those based three months or less

Duration of stay (p < 0.01) was significantly associated with lowered adherence – for example none of those based over a year were still taking their chemoprophylaxis compared with 81% of those based three months or less

Common reasons for discontinuing malaria prophylaxis include medication side-effects, low perceived malaria risk and suggestions from colleagues on the job site

Medium

Ropers et al. [28]

German travellers to Kenya, Senegal and Thailand

1001

Cross-sectional survey

69% in Kenya

53% in Senegal

6% Thailand

Travel to Kenya compared to Senegal resulted in a significantly higher adherence rate to chemoprophylaxis (p = 0.021), Receiving advice (from either a medical or non-medical professional) significantly increased adherence when compared with those who received no advice (p < 0.001), Correct risk perception (p < 0.001) was associated with a significant increase in prophylaxis adherence. Increased length of travel (for example comparing those travelling less than 14 days to those travelling 15–21 days) was associated with increased adherence (p < 0.001)

Reasons for poor adherence included absence of mosquitoes (53%) and ‘adverse effects’ with their medication (22%)

High

Roukens et al. [6]

Non-immune expatriate business travellers

2350

Cross-sectional, web-based study

45%

Malaria awareness and CMK training (RR = 2.2; 95% CI 1.6–3.2); long-term travellers less likely to be adherent compared to rotators or visitors (p < 0.001)

High

Baggett et al. [29]

US residents travelling to India

1302

Cross-sectional study

VFRs (visiting friends and relatives) 16.3%; non-VFRs 39.4%

Factors significantly associated with lower adherence to chemoprophylaxis included travelling to India in the previous 5 years (POR = 0.46; 95% CI 0.31–0.67) and travelling with the purpose of VFRs (visiting friends and relatives) (p < 0.001). Taking chemoprophylaxis was also more common among US citizens (POR = 2.71; 95% CI 1.91–3.85)

High

Alon et al. [30]

Israeli travel clinic

394

Telephone interview

60.7% in over 60 age group

33.8% in 20–30 age group

Elderly travellers (p < 0.01)—those aged 60 and over—were significantly more likely to be adherent than those in the 20–30 age group

Medium

Depetrillo et al. [31]

Travellers from the United States

104

Prospective, non-blinded study

89%

Factors associated with increased adherence included travel destination, with those travelling to regions such as Sub-Saharan Africa having significantly higher adherence levels (p = 0.0063) compared with those travelling for example to Central America. Other significant predictors of non-adherence included previous travel to a malarious region (p = 0.0411) compared with those who had never travelled to a malarious region before

Travellers’ self-reported perception of need was felt to be a key influencer in adherence. 7/12 felt it was not necessary, 2/12 were told by their tour guides they did not need to take it and 3/12 reported adverse side-effects

Medium

Dia et al. [32]

French travellers to Senegal

358

Prospective cohort study

71.8%

Factors significantly associated with non-adherence included reporting at least one gastrointestinal symptom (p = 0.07) and non-reporting arthropod bite (p = 0.04)

The main reasons for not taking medications were: finding it useless (47.1%) and fearing side effects (44.1%)

Medium

Joshi et al. [33]

UK South Asians

400

Cross-sectional survey

49% (1994) and 32% (2004)

Factors associated with an increased adherence with prophylaxis included a basic knowledge of malaria (p = 0.003), perceiving malaria as a critical illness (p = 0.004) and defining trip as a holiday (as opposed, for example, to a ‘visit to friends and family’) (p = 0.043)

Age, gender and occupational status did not relate to adherence in either year; years of post-16 education did not relate to adherence in the 2004 sample (not asked in the 1994 survey)

Adherence was not related to experience of malaria or having been born in a malarial zone

Reasons given for non-adherence given by partial and zero adherers included: belief in personal immunity (47% in 1994, 43% in 2004); perceived low risk of getting malaria (42% in 1994, 26% in 2004); never heard of tablets (25% in 1994, 27% in 2004; forgot to take/get tablets (21% in 1994, 15% in 2004); dislike taking tablets (14% in 1994, 29% in 2004); believing malaria is easily treatable (9% in 1994, 22% in 2004); local norms (5% in 1994, 17% in 2004)

Medium

Pistone et al. [34]

French adult travellers

13,017 (3066 travellers to malaria-endemic countries)

Retrospective questionnaire study

47.6% in high-risk areas

9.5% in low-risk areas

Factors significantly associated with increased adherence with malaria chemoprophylaxis included awareness malaria was serious (OR = 2.03, p = 0.033) and receiving information from a physician (OR = 3.01, p = 0.042). When the analysis was reiterated for travellers to low and high-risk areas separately, older travellers were less likely to be adherent for the high-risk travellers only (OR = 0.95 for each incremental year of age p = 0.018)

High

Belderok et al. [35]

Dutch short-term travellers

620

Prospective cohort study

75% (466/620) took 100% of recommended tablets

Significant factors associated with adherence included: travelling to Africa (OR = 3.5; 95% CI 1.9–6.5) instead of Asia or Latin America; taking mefloquine (OR = 5.3; 95% CI 1.2–23.1) compared to atovaquone–proguanil or proguanil; spending 14–29 days in endemic areas (OR = 2.2; 95% CI 1.2–3.8) instead of ≤ 13 days or ≥ 29 days in endemic areas; concurrent use of DEET for more than 50% of days in high-endemic areas (OR = 2.6; 95% CI 1.4–4.8)

Medium

Caillet-Gossot et al. [36]

Children under 16 visiting travel medicine centre in Marseille, France

167

Prospective study

66%

Adherence was significantly higher in those visiting African destinations (p < 0.02) compared with those taking a trip to Asia or Indian Ocean

Being aged < 5 (p < 0.03) was also found to be a predictor of non-adherence as was being from a mono-parental family (p < 0.04)

Adherence was identical between VFR and tourist children, irrespective of trip duration

Medium

Muller et al. [37]

Adults consulting at a Medical Department for International Travellers’

287

Cross-sectional survey and telephone interview

76.3%

Travelling to areas of mass tourism (such as Kenya and Senegal) (p = 0.005) was found to be a predictor of adherence—it was noted that these travellers were also less likely to be seasoned travellers (compared with, for example, long-stay business travellers); trips shorter than 15 days were associated with better adherence (p = 0.001)

Side-effects (20.6%), forgetting (17.6%), too many pills—because of other treatments (17.6%), no mosquitoes seen (13.3%), tiredness (11.8%), did not like taking medication (10.3%), price (2.9%), lack of pills (1.5%)

Medium

Wieten et al. [38]

Travellers to Ghana from the Netherlands

154

Questionnaire survey

53.9% (had started chemoprophylaxis)

Attending pretravel clinic and receiving pre-travel advice was related to a greater likelihood of starting chemoprophylaxis (p < 0.01); if a participant incorrectly thought they had been vaccinated (p = 0.009) they were also more likely to use chemoprophylaxis

Higher age (p = 0.004) and travelling for family purposes (p = 0.022) rather than business were positively associated with starting chemoprophylaxis. Having had malaria (p = 0.028) and spending more than 6 weeks in West Africa (p = 0.001) were negatively associated with starting and buying chemoprophylaxis

Those who thought curing malaria was easier than taking preventative tablets (p = 0.046) were more likely to be non-adherent—it was felt that subjectively held information is more important than accurate information

Previous use of chemoprophylaxis was not found to influence current preventive behaviour

Medium

Cunningham et al. [12]

Foreign and Commonwealth Office employees on long-term placement in endemic areas

327

Questionnaire survey

25.1% had adherence > 95% of prescribed pills

54.4% had adherence < 25% of prescribed pills

Increasing age was shown to be significantly associated with improved adherence (Chi squared p < 0.00), living in an endemic area for more than a year was significantly correlated with adherence less than 95%, pregnancy was associated with lower adherence (87.5% of pregnant women took no prophylaxis)

Significant side-effects were reported by 39.5% of respondents and there was a trend between reported side-effects and self-reported adherence < 95% (p = 0.087)

Concerns with long term safety was cited by more than half of individuals with adherence < 25%

Medium

Goldstein et al. [39]

Israelis attending Haifa travel clinics

307

Questionnaire survey

34.7%

Shorter travel (p < 0.001), with those who adhered having a travel duration on average 2.6 times shorter than those who did not; travel to urban areas (p < 0.01) showed higher adherence; travellers older than 23 (p = 0.021) showed higher adherence, backpackers showed lower levels of adherence (p < 0.01) compared to other travellers

Medium

Landman et al. [40]

Peace Corps volunteers in the Africa region in 2013

781

Questionnaire survey

73%

Factors significantly associated with non-adherence included: being prescribed mefloquine (OR = 5.4; 95% CI 3.2–9.0) as opposed to doxycycline or atovaquone–proguanil; if they were in the peace corps for over a year (OR = 1.8; 95% CI 1.2–2.8); being under 26 years old (OR = 1.7; 95% CI 1.1–2.6); not worrying about malaria (n = 214; OR = 2.6; 95% CI 1.6–4.1); fears long-term adverse effects (OR = 1.6; 95% CI 1.1–2.4)

The most common reasons for non-adherence included: forgetting (n = 530, 90%), fear of long-term adverse effects (n = 316, 54%) and experiencing adverse events that volunteers attributed to prophylaxis (n = 297, 51%)

Medium

Shady [41]

Visitors to traveller’s health clinic to obtain malaria prophylaxis

928

Prospective comparative study

81.6% with mefloquine and 79.5% with doxycycline

University education (p = 0.005) was a predictor of non-adherence, travel organized through an agent showed increased adherence (p = 0.0001) whereas independently organized travel (p = 0.0001) was a predictor of non-adherence, blue-collar workers (p = 0.0001) showed higher non-adherence compared to white-collar workers

Predictors of good adherence for mefloquine group included travel to an African destination (p < 0.001), education above a secondary level (p < 0.001), organized travel (p < 0.05), travelling for leisure (p < 0.05) and Kuwaiti nationality (p < 0.001)

Predictors of good adherence in the doxycycline group included higher than a secondary level of education (p < 0.001), organized travel (p < 0.001), travel for leisure (p < 0.05), travel to an African destination (p = 0.05) and Kuwaiti nationality (p < 0.001)

Medium

Stoney et al. [42]

US travellers

370

Cohort study

71.6%

No significant difference for sex (p = 0.74), location of birth (p = 0.49), endemicity of country of birth (p > 0.99), daily vs weekly chemoprophylaxis (p = 0.19), visiting friends or relatives as a reason for travel (p = 0.44), destination as partially or entirely endemic (p = 0.89), or travelling for more or less than 2 weeks (p = 0.19)

Reasons for declining entirely: advised by peers not to take chemoprophylaxis (32%), low perceived risk in area (28%), no mosquitoes present during trip (16%), fear of side effects (16%), cost (8%), had a side effect (4%), unable to fill prescription before trip (4%), other (8%)

Reasons for not taking full course: Forgetting (cited by 50% of participants nonadherent during travel), side-effects (31%), not seeing mosquitoes (11%), low perceived risk in area (8%), lost medication (6%), other reason (6%). Data from a post-travel survey, completed by a smaller proportion of participants, are not reported here

Medium

Rolling et al. [43]

German travellers

928

Questionnaire survey

19% carried anti-malarial medication

Neither duration of travel or previous travel experience significantly differed between those carrying anti-malarial medication and those who did not

A medical consultation prior to travelling was associated with significantly higher odds of carrying anti-malarial medication after adjusting for age, with the highest odds in those having had their consultation at a travel medicine specialist (OR 7.83 compared to no consultation)

Medium

Pagès et al. [44]

Malaria cases in Réunion Island (a previously malaria-endemic island; the last indigenous cases were reported in 1967, but international travel has reintroduced the illness)

89

Epidemiological surveillance data; data from Regional Health Agency investigations

29 patients were prescribed chemoprophylaxis: 10 did not buy it, 13 stopped taking it early, 3 took it irregularly, and 5 reported proper adherence. Of the 56 patients not prescribed anti-malarial medication, 24 were not aware they should have consulted a doctor, 21 chose not to, and 11 were not prescribed a medication after their consultation

An absence of chemoprophylaxis or poor adherence was found in the majority of malaria cases (96%) regardless of the reason for travel (visiting friends and family vs. other reasons)

Low