Travelers’ health problems and behavior: prospective study with post-travel follow-up

Volunteers to this prospective study cohort were enrolled at the Travel Clinic of Aava Medical Centre among travelers planning a journey outside the Nordic countries for a minimum of four days and a maximum of six months (Fig. 1). The only exclusion criterion was non-compliance in returning questionnaires. The subjects were recruited among consecutive clients at pre-travel appointments between December 2008 and February 2010. At the initial visit, they filled out a pre-travel questionnaire (Q1), on return home a post-travel questionnaire (Q2), and about three weeks later a follow-up questionnaire (Q3). Those who failed to return both Q1 and Q2 were excluded. In addition to the questionnaires, the subjects were asked to complete a structured diary on a voluntary basis. All volunteers were given pre-travel advice by a health care professional.

We have earlier reported risk factors for acquiring resistant intestinal microbes in the same population [28], and, data on diarrheal pathogens, first in our methodological investigation [29] and, recently, in an etiological study [30].

The questionnaires consisted of 134 multiple-choice or open-ended questions: Q1 comprised 47, Q2 60, and Q3 27 questions. These questionnaires were modified from a set of survey questions routinely used for more than 10 years for ill travelers admitted to the Clinic of Infectious Diseases at HUCH. All items covered are listed in Fig. 1. The structured diary contained more specific questions on duration and severity of symptoms, and use of antibiotics.

Subjects were classified as ill if they reported symptoms implying health problems which could be travel-related. Diarrhea was defined according to WHO criteria, i.e. passage of 3 or more loose or liquid stools per day, or more frequently than is normal for the individual (World Health Organization [31]). At follow-up, symptoms which had set in more than two days after a journey and could be travel-related were categorized as newly onset.

The countries visited were grouped into nine geographic regions (UN categorization, modified [32]): Southern Asia, South-Eastern Asia, Eastern Asia and Central Asia, Southern Africa, Eastern Africa, Western Africa and Middle Africa, Northern Africa and Western Asia, Latin America and the Caribbean, and Europe and Northern America (Table 1). The destination with the greatest health risk was considered primary for subjects traveling to several places. Here we used a rating based on the risk map drawn up by International SOS (Additional file 1: Figure S1 HealthMap 2010) which takes into account a range of factors: the standard of local medical and dental care, access to prescription drugs, the possible prevalence of serious infectious diseases, and known cultural, linguistic and administrative barriers.

Univariable, bivariable and multivariable models were used. The p-value of Pearson Chi-square tests and Fisher’s exact test < 0.05 was considered statistically significant. If bivariable p-value was less than 0.10, the factor was chosen to the multivariable logistic regression model, and its adjusted odds ratios and 95 % confidence intervals were calculated. Missing data were assumed to be missing at random (MAR) and missing values were imputed using multiple imputations in SPSS. The statistical analyzes were carried out with SPSS Statistics (version 22.0.0.2, IBM Corp., USA).

Of the 524 initially recruited travelers, 88 % completed both pre- and post-travel questionnaires. The final study population thus consisted of 460 volunteers (Fig. 1, Table 1). A total of 233 (51 %) returned follow-up questionnaires and 295 (64 %) filled in diaries. An underlying disease or condition was recorded for 192 (42 %)(Table 1), more frequently among the oldest subjects (56– years, 73 %; p < 0.001), and females (45 %; p = 0.046) than in other age groups (0–55 years, 23–38 %) or among males (36 %). Data on vaccinations are presented in Table 2 and on antimalarials in Table 3.

The 460 travelers visited 77 countries altogether (UN definition); India (64), Thailand (60), Gambia (50), Tanzania (48), and Kenya (40) ranking as their most popular destinations. Overall 662 country visits were made (average 1.45 per person; range 1–8). The 30 most favored countries accounted for 560 journeys (85 %); only three destinations were in developed regions (UN definition [33]): USA (8), Spain (7), and the Netherlands (5).

Uncooked meat/fish (Table 4) was eaten more frequently by young adults (18–35 years, 17 %) and the middle-aged (36–55 years, 13 %) than children (0–17 years, 6 %) and older travelers (56– years, 6 %) (p = 0.016). Young adults proved more likely than the others to have freshwater contact, neglect hand washing, and not to shun salads or eating without utensils. Men consumed more alcohol than women (p < 0.001).

Factors with p < 0.1 in the bivariable analyses (Table 5) were included in the multivariable analyses (Table 6). Destination associated strongly with health problems. Females were more predisposed to falling ill. Overall, the risk was greatest at 31.5 years of age. The longer the travel, the greater the risk of contracting an illness – it increased each day by 2.5 %. Eating raw meat/fish was associated with healthier travelers.

Geographic regions with demographics and proportions of sick travelers are presented in Fig. 2. Illness rates by geographic regions were found to accord with respective risk ratings by International SOS (Additional file 2: Figure S2).

Health care was sought by 10 % of all subjects and by 13 % of those fallen ill during travel; 10 % and 13 % visited a physician and 0.7 % and 0.9 %, respectively, were hospitalized (Table 7). The most common reason for consulting health care professionals while abroad was TD – 4 % of all travelers visited a doctor, and 0.4 % were subsequently admitted to hospital because of this illness.

During follow-up, about one-third reported new-onset health problems (Table 7). Their most common single symptom was fever, followed by respiratory tract infection, skin problems, and TD.

Investigations addressing travelers’ health have generally been retrospective and/or conducted on those seeking medical care after their journeys, whereas prospective studies are scarce. Furthermore, research tends to focus on the effects that one or just a few specific factors have on travelers’ behavior and illness: destination [4, 26, 27], length of journey [4, 5, 26, 27], purpose of travel [27], gender [4, 5, 26, 27], age [4, 5, 26, 27], risk behavior [5], and particular diseases, such as TD [34‐41]. We are not aware of any earlier investigations into all these factors with a single prospective cohort of travelers; only one [4] provides data on the same subjects during and after travel, and at follow-up.

A central finding of our work is high morbidity rate. Despite efficient prophylactic measures taken before travel – exemplified by our vaccination data – the amount of health problems proved striking: as many as 76 % of our subjects reported illness while abroad, and 25 % still had ongoing symptoms or new complaints within two days after returning home. The overall proportion of our subjects with any symptoms while overseas (76 %) correspond to that reported for American (64 %; [4]), U.K. (64 %; [25]), and Israeli travelers (70 %; [5]), but proved higher than percentages presented for German (10-43 %; [26, 27]), Swiss (38 %; [24]), and Swedish travelers (49 %; [23]). As all these investigations are based on questionnaires, the format of the various questions concerning symptoms may account for the differences.

As regards morbidity in different geographic regions, the percentage of illness proved, as expected, significantly smaller for travelers to advanced countries (20 %) than those visiting developing regions (81 %); the rates were highest for visitors to Southern Asia, South-Eastern Asia, and Eastern Africa. TD accounted for the majority of health problems while abroad / after return (69 %). This result accords with several previous studies [4‐6, 8, 10, 21‐24, 26, 27, 42‐44].

While the symptoms of illness mostly set in abroad, the figures proved unexpectedly high also during the follow-up: new health problems were reported by 32 % of our subjects. This percentage slightly exceeds the result of the only previous study exploring illness on return and at follow-up; the morbidity rate reported by Hill is 26 % [4]. The duration of follow-up may account for the small difference. In our study it was somewhat shorter (three weeks vs two months). But since the subjects were not asked to give exact dates for their symptoms, the onset could afterwards not be limited to 14 days in cases of TD, respiratory illness, and skin disorders, like in Hill’s report.

At follow-up, the four leading symptoms were fever, skin problems, respiratory illness, and TD, a finding according with Hill’s results [4]. In our study, fever was the most frequent single symptom – often associated with TD. Diarrhea remained the most common cause for seeking health care.

The data revealed regional differences which enabled profiling our subjects: travelers to Africa tended to be older and more cautious than those visiting Asia, whereas visitors to Asia were typically younger and favored longer trips. Likewise, the data show differences between travelers by age: the young traveled for longest and stayed in guest houses more frequently than the others, were the most likely to eat uncooked meat/fish and salads, not to use utensils, to follow a vegetarian diet, to neglect hand washing, and not to avoid freshwater contact. Not surprisingly, the longer the stay overseas, the lower the degree of travelers’ compliance with hygiene instructions. Such traveler profiles can be used as a tool to target advice at various groups according to their special characteristics.

Destination, gender, age, and duration of travel were shown by multivariable analysis to be factors predisposing to illness. Southern Asia proved the riskiest resort, as also reported in earlier studies [4, 26]. Female travelers proved to be at greater risk of acquiring symptoms (OR 1.7), a finding according with the results published by Hill [4]. In the present data, each day increased the risk of contracting illness by 2.5 %; the respective figure obtained by Hill was 3.1–3.7 % [4]. Young age has often been reported to be associated with illness [4, 5, 8, 23, 26, 27]; in our cohort the risk was highest at 31.5 years. Eating raw meat or fish proved to be a protective factor. The reasons for this are not obvious and can only be speculated on.

The present investigation has some limitations which deserve to be discussed. Firstly, the results are not representative of all travelers, but of volunteers with pre-travel appointments at a travel clinic. Thus visitors to Africa and other developing regions were overrepresented. Secondly, due to small sample size, statistical analyses of some subgroups were poorly powered. The third limitation concerns the follow-up: the reliability of the conclusions suffers from the fact that only 51 % of our volunteers completed the questionnaire, and many delivered it later than requested. Importantly, however, even if it could be assumed that none of the remaining 49 % had been ill, the number of newly onset symptoms during follow-up remains substantial. The data would have benefited from a comparison to non-traveling controls. Our data collection can also be regarded as a limitation, for questionnaire-based studies may distort the results in various ways. Due to the common use of questionnaires, this tends to be characteristic of research into travelers’ health problems [3, 4, 6, 22‐27, 42, 44‐48]; a few previous reports include data collected by telephone surveys [2, 4, 5, 8, 27]. In our study, travel diaries may have improved the accuracy of data to some degree. The exact number of clients declining to participate was not recorded. However, based on the number of clients seen by the recruiting doctor and number of those recruited, we estimate that at the maximum 10 % of the potentially eligible clients declined.

We recently reported incidences of illness and injury among more than 50 000 Finnish travelers visiting various regions [21] by relating cases recorded by an assistance organization to numbers of travelers to each region. The results were considered to cover the most severe cases. The present data complete that picture by showing even the mildest symptoms contracted. This puts the two studies into perspective. Since 90 % of our travelers did not contact health care, the proportion covered by our previous report may not amount to more than 10 % of the illness altogether – slightly less than estimated in that study [21]. The strengths of the present research include the prospective study design and the fact that we combined data of three time points (pre-travel, travel/return, and follow-up) for a single cohort. Interestingly, the rate of health problems per region presented in our study accords with risk estimates presented by International SOS (Additional file 2: Figure S2).

Advice on malaria prophylaxis is a cornerstone of pre-travel appointments. In our data, 289 of the 296 travelers (98 %) taking malaria prophylaxis reported compliance. Interestingly, however, four of them, all diagnosed in Africa, were also treated for malaria while abroad. One case was microbiologically verified, two reported a negative malaria test, and one had taken medication without laboratory diagnostics. Our earlier data collected by the assistance organization shows malaria to be rare in travelers (8/50000 cases; Siikamäki, personal communication). Indeed, the diagnoses of our cases may not be correct, as presumptive treatment is often given in Africa [49], and malaria diagnostics may not always be accurate [50].