Surveillance of travel-associated gastrointestinal infections in Norway , 2009 – 2010 : are they all actually imported ?

The Norwegian Surveillance System for Communicable Diseases (MSIS) includes variables related to travel for clinicians to fill when notifying travel-associated infections. We measured the completeness and validated the travel-history information for salmonellosis, campylobacteriosis, giardiasis and shigellosis reported in 2009–2010. Of all 8,978 selected infections in MSIS, 8,122 (91%) were reported with place of infection of which 5,236 (65%) were notified as acquired abroad, including 5,017 with symptoms. Of these, 2,972 (59%) notifications had information on both date of arrival in Norway and date of symptom onset, so time between travel and illness onset could be assessed. Taking in account the incubation period, of the 1,435 infections reported as travel-associated and for which symptom onset occurred after return to Norway, 1,404 (98%) would have indeed been acquired abroad. We found a high level of completeness for the variable ‘place of infection’. Our evaluation suggests that the validity of this information is high. However, incomplete data in the variables ‘return date to Norway’ and ‘date of symptoms onset’, only allowed assessment of the biological plausibility of being infected abroad for 59% of the cases. We encourage clinicians to report more complete travel information. High quality information on travel-associated gastrointestinal infections is crucial for understanding trends in domestic and imported cases and evaluating implemented control measures.

The Norwegian Surveillance System for Communicable Diseases (MSIS) includes variables related to travel for clinicians to fill when notifying travel-associated infections.We measured the completeness and validated the travel-history information for salmonellosis, campylobacteriosis, giardiasis and shigellosis reported in 2009-2010.Of all 8,978 selected infections in MSIS, 8,122 (91%) were reported with place of infection of which 5,236 (65%) were notified as acquired abroad, including 5,017 with symptoms.Of these, 2,972 (59%) notifications had information on both date of arrival in Norway and date of symptom onset, so time between travel and illness onset could be assessed.Taking in account the incubation period, of the 1,435 infections reported as travel-associated and for which symptom onset occurred after return to Norway, 1,404 (98%) would have indeed been acquired abroad.We found a high level of completeness for the variable 'place of infection'.Our evaluation suggests that the validity of this information is high.However, incomplete data in the variables 'return date to Norway' and 'date of symptoms onset', only allowed assessment of the biological plausibility of being infected abroad for 59% of the cases.We encourage clinicians to report more complete travel information.High quality information on travel-associated gastrointestinal infections is crucial for understanding trends in domestic and imported cases and evaluating implemented control measures.

Background
Increased harmonisation in preventing foodborne infections at the European level and an increase in international travel in recent years has led to the need for better knowledge on the epidemiology of travelassociated infections in Europe.To this effect, the European Centre for Disease Prevention and Control (ECDC) gathers surveillance data including data from the European Travel Medicine Network EuroTravNet [1,2] that is linked to the worldwide surveillance network on travel-associated morbidity GeoSentinel [3].
Already existing national surveillance systems may also provide information on travel-associated infectious diseases.The Norwegian Surveillance System for Communicable Diseases (MSIS) includes several variables on travel history that should be filled in by clinicians when reporting notifiable diseases.This information is notified to the European Surveillance System (TESSy), and is presented in yearly surveillance reports [4].).Of those with known travel history, the proportion that was travel-associated varied from 8% of campylobacteriosis cases to 62% of shigellosis cases [6,7].

Travel-associated gastrointestinal infections in Europe and Norway: current situation
In Norway, most cases of gastrointestinal infections notified to MSIS in 2010 were reported as travel associated, ranging from approximately 50% of campylobacteriosis cases to around 83% of shigellosis cases [8].

The challenge when notifying gastrointestinal travel-associated infections
Clinicians seeing patients returning from international trips play a critical role in recognising and notifying travel-associated public health risks [2].The importance of taking a travel history to establish the possibility of imported infection was highlighted almost 30 years ago by the classical publication 'Unde Venis?' ('Where do you come from?') [9].However, published studies conclude that there is still insufficient and inadequate travel history recording, which may directly have an impact on public health [10].
The lack of standardised case definitions for travelassociated infections makes the reporting and comparison of rates between countries difficult.In Europe, there is no general agreement on how to define a travelassociated infection.Information on travel-associated infections compiled at the international level, such as through TESSy, originates from multiple sources.Data from diverse European countries may therefore not be directly comparable due to differences in surveillance and national definitions.Additionally, guidelines for classifying 'travel-associated' cases versus 'domestic' cases may not exist at the country level.In Norway, for example, there are no strict criteria for defining a case with gastrointestinal infection as travel-associated.This is determined by the reporting clinician, based on time of symptom onset, place of travel, return date and incubation period of the disease.The endemic level of the infection in Norway is also taken into account when clinicians report a case as travel-associated or domestic.Since the endemic level of most notifiable gastrointestinal infections is very low in Norway, clinicians may report a case as travel associated without taking into account the incubation period for that specific infection in relation to when the patient returned to Norway from a high-endemic country.
In order to evaluate the quality of the travel-associated information on gastrointestinal infections available in MSIS, we measured the completeness of the travelassociated variables of gastrointestinal infections notified (specifically salmonellosis, campylobacteriosis, giardiasis and shigellosis) in MSIS during 2009-2010.We also validated the information about place of infection, Norway/abroad, by looking at the reported time between travel and time of symptom onset compared to the expected incubation period of the disease.

Methods
We considered 'travel-associated gastrointestinal infection cases' to be any registered cases of salmonellosis, campylobacteriosis, giardiasis and shigellosis in MSIS where the variable 'place of infection' was a country other than Norway.To be included in our study, a case's symptom onset date had to be between January 2009 and December 2010.
In Norway, clinicians report notifiable diseases using a paper form.When reporting, the clinician should state whether the patient acquired the disease in Norway or abroad.If the disease is reported as acquired abroad, the country visited and the date of arrival in Norway should be specified.Once the notification is filled-in, it is sent by post to the Norwegian Institute of Public Health (NIPH) where it is entered into MSIS.
Variables related to diagnosis, microorganism, place of infection, date of return to Norway after travel, and date of symptom onset were extracted from MSIS (Table 1).
To measure the completeness of the data on travelassociated infections we analysed the variables 'place of infection', 'return date to Norway' and 'date of onset of symptoms' for all cases of salmonellosis, campylobacteriosis, shigellosis and giardiasis notified to MSIS during the study period.For Salmonella infections we also did separate analyses for S. Typhimurium and S. Enteritidis as the endemic levels of these serovars are known to be very different within Norway.
Further analyses included only cases reported as having acquired infection abroad.
In order to validate the information, we used the variables 'date of onset of symptoms' and 'return date to Norway' to measure the time between travel and illness onset.We excluded all cases registered as asymptomatic (diagnosed, for example, as a result of a routine screening), since they would not have a recorded date of symptom onset.We then selected the cases where the variables 'date of onset of symptoms' and 'return date to Norway' were complete and studied the distribution of cases' dates of symptom onset around the return date to Norway after travel.
For those cases with symptom onset after arrival in Norway, we compared the number of days between the return date to Norway and the date of symptom onset to the incubation period described in the literature for each infection, to assess the plausibility of the case having been infected while travelling.As reference, we used two thresholds for each infection, the most common incubation period and the maximum incubation period (Table 2) [11,12].In addition to estimating infection abroad by incubation period, we also used seven days following return from travel as a cut off, as has been used in other studies [13].

Completeness of the variable 'place of infection'
Of all 8,978 gastrointestinal infections, 8,122 (90.5%) had information on place of infection, of which 5,236 (64.5%) were notified as contracted abroad.
Information on place of infection was most frequently missing for giardiasis (64 cases, 11.3%) and salmonellosis (272 cases, 10.5%).Among salmonellosis cases, S. Typhimurium (excluding the monophasic variant) was the sevovar for which place of infection was most frequently unknown (50 cases, 14.4% of all S. Typhimurium cases (n=348)) (Table 3).a According to [11,12].The percentage is calculated relative to the respective total reported symptomatic travel-associated cases.Fourteen campylobacteriosis cases (1.8%), eight salmonellosis cases (1.4%), five giardiasis cases (10.6%) and four of 44 shigellosis cases (9.1%) would have had symptom onset outside the maximun incubation period range defined in the literature, if the infections had been acquired abroad.

Discussion
The results of the evaluation of the travel information reported in MSIS indicate a high level of completeness with regards to the variable 'place of infection'.Approximately 90% of gastrointestinal infections notified to MSIS were reported with known origin.
Compared to studies published in other Scandinavian countries, such as Sweden, the level of completeness found in MSIS is higher for campylobacteriosis [14], but lower for shigellosis [15] and salmonellosis [16].
The reported level of completeness found in our study is also higher than the level of completeness reported by a Canadian study [17].
The validity of the travel information of those cases reported as travel associated is also high according to our assessment on time of illness onset related to the time of travel.The time of illness onset after travel for the majority of cases fell within the maximum incubation period reported in the literature and indicates good judgement among clinicians when reporting.This information supports the assertion that most gastrointestinal infections notified in Norway are travel associated.The low endemic level is considered to be caused by low endemic levels of the pathogens causing the diseases considered in this study in the food chain in Norway and historical low level of imports of animals and animal products in Norway [18].Of note is the high proportion of notified travel-associated infections with symptom onset on the day of return to Norway or the prior week.This might be explained by patients who got infected shortly before returning to Norway first sought medical attention at arrival because of persisting symptoms.
Shigellosis and giardiasis were the infections most frequently classified as travel associated although illness onset occurred after the common incubation period after return to Norway.11% of travel-associated reported giardiasis cases and 9% of shigellosis cases did not occur within a plausible incubation period.
The fact that the endemic level of these two diseases in Norway is low may lead clinicians to code them as travel-associated without taking into account the incubation-period of the disease.The endemic level of different serovars of salmonellosis may also play a role in clinicians' judgement when reporting.None of S. Typhimurium cases notified appeared to be misclassified.However, three S. Enteritidis cases reported as travel-associated had illness onset after the maximum incubation period after returning from abroad.Epidemiology of these two Salmonella serovars in Norway are very different: S. Typhimurium is the most common serovar that causes domestic salmonellosis in Norway and in addition to S. diarizonae the only serovar existing in Norwegian animals [19], while most S. Enteritidis cases are acquired abroad [8].Therefore, clinicians might classify a S. Enteritidis case as travelassociated with less consideration of the number of days since return to Norway.Three cases reported as travel-associated had a date of onset of symptoms more than one year after return from travel abroad.We think this is probably due to data entry errors of the dates rather than inaccuracy when reporting.It is also important to bear in mind that despite the biological plausibility of the calculated incubation period, it does not ensure that the infection occurred abroad as the source is not verified.
We only managed to assess the biological plausibility of the travel history in 59% of all cases reported as travel-associated as these were the only notifications with both variables 'return date to Norway' and 'date of onset of symptoms' completed.Giardiasis was by far the infection in which information regarding these variables was most frequently missing.The complex clinical course of giardiasis, which can be intermittent and chronic, might have contributed to the low completeness [20].Having been able to better assess the biological plausibility for this pathogen would have been useful, since the endemic level of this parasite in Norway is not very well described [21].
Surveillance of travel-associated gastrointestinal infections in Europe could benefit from improved   Mv: monophasic variant.
Interval of days of symptom onset before (-) or after (+) return date to Norway (0).

Table 1
Variables related to gastrointestinal infections extracted from the Norwegian Surveillance System for Communicable Diseases, Norway, 2009-2010 ranged from 51.2% for giardiasis (126/246) to 68.8% for salmonellosis caused by S. Typhimurium (108/157) (Table4).The variable 'date of onset of symptoms' was completed for 4,291 (85.5%) of all travelassociated symptomatic cases, ranging by disease from 61.8% for giardiasis (152/246) to 87.3% for salmonellosis due to S. Enteritidis (806/923) (Table4).The Figure illustrates the date of symptoms onset in relation to return date to Norway for cases reported as travel-associated.Overall, 1,435/2,972 (48.3%) reported onset of symptoms on the day of return or after return to Norway (Table5).
'return date to Norway' and 'date of onset of symptoms'Overall, 3,167 (63.1%) of all travel-associated symptomatic cases (n=5,017) had registered a 'return date to Norway' after a trip.The completeness of this variable by infection

Table 2
Reported incubation periods for gastrointestinal diseases considered in this study, surveillance of travel-associated gastrointestinal infections,Norway, 2009Norway,  -2010 Mv: monophasic variant; S.: Salmonella.

Table 4
Completeness of variables for selected travel-associated gastrointestinal infections registered in the Norwegian Surveillance System for Communicable Diseases, Norway, 2009-2010 Figure Distribution of travel-associated cases of gastrointestinal infection reported to the Norwegian Surveillance System for Communicable Diseases by date of symptom onset relative to travel-return date to Norway, 2009-2010 (n=2,972)

Table 5
Travel-associated cases with symptom onset before travelreturn to Norway among travel-associated cases with known travel-return and symptom onset dates registered in the Norwegian Surveillance System for Communicable Mv: monophasic variant; S.: Salmonella.Arrival in Norway (day=0).

Table 6
Days between arrival in Norway and symptom onset, travel-associated gastrointestinal infections registered in the Norwegian Surveillance System for Communicable Diseases,Norway, 2009Norway,  -2010 (n=2,972)    (n=2,972)

Table 7
Repartition, by incubation period, of travel-associated cases of gastrointestinal infection with symptom onset after travel return registered in the Norwegian Surveillance System for Communicable Diseases,Norway, 2009Norway,  -2010 (n=1,435)   (n=1,435)