Behavioural and environmental factors associated with the risk of human tick bites in the Grippenet/Covidnet cohort in France

Abstract

Background

Ticks are one of the most important vectors of infectious disease pathogens in Europe, posing threats to human health. However, human exposure to tick bites and associated factors remains poorly documented, despite its importance to better mitigate the risk of tick-borne diseases. Current participatory datasets report information from individuals being bitten by ticks, while information on exposed but non-bitten individuals is scarce. Here, by collecting data on exposed individuals from a nationwide participatory surveillance cohort in France, we assessed the risk of self-reporting tick bites, and explored associated factors.

Methods

We conducted a longitudinal study within the Grippenet/Covidnet cohort from November 2022 to February 2025, using online surveys sent every four months, and collected data on participants’ tick-bite experiences and related behaviours. We estimated the incidence of participants self-reporting at least one tick bite over a year. We then fitted a two-part Bayesian model using Integrated Nested Laplace Approximation (INLA) to look at factors associated with being bitten at least once and, among bitten participants, factors associated with an increased number of tick-bite episodes.

Results

The reported number of tick bites showed a seasonality with peaks occurring in June 2023 and 2024. We estimated the annual incidences of participants reporting at least one tick bite at 5,615 (95% CI [4,910-6,321]) and 5,402 (95% CI [4,670-6,133]) per 100,000 inhabitants in 2023 and 2024, respectively. Our model showed that factors positively associated with reporting a tick bite included being female, living in Grand Est region, having a green outdoor area at home, with an average temperature above 7 °C and higher percentage of forest and seminatural areas near the residence, having outdoor activity in green areas more than once a week, and visiting forest or meadow. The use of prevention measures was also associated with tick-bite reporting, which might reflect greater awareness of ticks or higher outdoor exposure. Factors associated with a higher number of self-reported tick bites were getting bitten in forest, meadow, or private garden, using prevention measures, and removing at least one tick after 24 h.

Conclusion

We present a nationwide and seasonal study providing baseline quantitative estimates of human behavioural and environmental factors associated with tick-bite reporting. Combined with other nationally collected data sources related to tick bites, for example providing data on tick species or at a higher resolution, this study could contribute to better inform prevention efforts.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12889-026-28382-1.

Background

Ticks are hematophagous arthropods that can transmit pathogens of public health importance, leading to emerging or re-emerging tick-borne diseases (TBD) such as Lyme borreliosis (LB), tick-borne encephalitis (TBE), or Crimean-Congo haemorrhagic fever [1, 2].

TBDs are a growing concern for human health. LB is the most common TBD in France with an incidence estimated at 53 cases per 100,000 inhabitants (95% CI [43–63]) in 2024, and could lead to chronic manifestations affecting the skin, joints, or nervous systems if not treated at an early stage [3, 4]. In addition, sporadic cases of TBE have been reported in the country for decades, but with mandatory reporting since 2021. Although TBE is not as common as LB, the number of reported TBE cases in 2024 has increased by 68% compared to 2023 [5]. Most TBDs have non-specific symptoms, making it difficult to be properly diagnosed. Moreover, the limited availability of disease-specific treatments and widely available vaccines make personal protection against tick bites (protective clothing, tick repellents) a major strategy to prevent TBD occurrences [6].

The most common tick vector in France is a hard tick, Ixodes ricinus, which is the vector for both LB and TBE [1]. Their presence depends on environmental factors such as host availability and abiotic conditions [6]. I. ricinus ticks are often found in forest areas as well as urban green spaces [1, 7, 8], and they are more active between March and October, although they can be found throughout the year [7, 9]. When humans are bitten by ticks, the ticks may remain attached to the skin for hours or several days without being noticed [7], offering the opportunity of pathogen transmission.

Humans are bitten by ticks mainly during outdoor activities, for occupational or recreational purposes, and in different kinds of outdoor spaces, including forests, gardens, meadows, and parks [10–12]. In addition, frequent outdoor activity has been associated with a higher risk of getting bitten, as illustrated by a study conducted in the United States reporting an association between the risk of finding an attached tick and time spent outdoors [13].

In Europe, tick-bite occurrence has been documented through several voluntary citizen science programmes. In the Netherlands, two platforms were developed to collect tick-bite reports, Natuurkalender (Nature’s calendar) and Tekenradar (Tick radar), with almost 10,000 reports in six years and 25,000 in a span of three years, respectively [14]. In Switzerland, human tick-bite reports were collected using a phone application (Tick Prevention), which allowed them to produce maps of tick attachment showing higher attachment rates in suburban areas with green spaces [15]. In France, CiTIQUE programme (https://www.citique.fr/) allows participants to report tick bites, and also send the corresponding ticks for pathogen screening. CiTIQUE have collected more than 86,000 tick-bite reports between January 2017 and May 2024, where over 72,000 concerned humans [16]. Another French data source, the “Sentick” indicator from Réseau Sentinelles, captured human tick bites through voluntary reports made by general practitioners (GP) following patient consultations [17]. However, although these data document tick bites, they do not capture individuals who were exposed to ticks but not bitten, making it difficult to estimate the risk of tick bites and associated factors.

The objective of this study was to assess self-reported human tick-bite occurrence in metropolitan France and to identify associated behavioural and environmental factors, using seasonal online surveys nested within a national online health cohort. By doing so, we first aimed to provide an estimation of the annual incidence of participants reporting at least one tick bite, and second, we identified factors that were associated with the risk of reporting tick bites.

Methods

Study design and data collection

We conducted a longitudinal online study among Grippenet/Covidnet cohort participants (https://www.grippenet.fr/), an online voluntary participatory cohort in metropolitan France launched in 2012 for influenza-like illness surveillance and also used to study other health-related topics [18]. Metropolitan France refers to the French territory located in Europe, including Corsica and excluding the overseas territories. Cohort participants were invited by email to participate to the survey. The survey was launched seven times covering the study period between November 2022 and February 2025. Each survey covered a 4-month period: November to February (Nov-Feb), March to June (Mar-Jun), or July to October (Jul-Oct), and was sent in the month that followed each study period. We selected these periods to approximate seasonal variation in both tick activity and human outdoor activity in green areas, assuming lower activity in November-February, increased activity in March-June, and decreasing activity in July-October [19]. The survey consisted of 3 sections. Section  1 contained general questions regarding prior knowledge and history of tick bites. Section  2 investigated tick-bite experiences in the previous 4-month period. Section  3 focused on outdoor activities in green areas for the same period as Sect.  2, for all participants, regardless of whether they reported a tick bite. Details on the questionnaire structure including the types of outdoor environments and outdoor activities investigated, the frequency of outdoor activity, and the types of prevention measures considered are available in Supplementary Materials (see SI section Study design). Finally, demographic data (age, sex, and municipality of residence) of the participants were collected from Grippenet/Covidnet database.

Environmental data

To characterise the environment around participants’ place of residence, we considered environmental variables within a circular buffer of 10 km radius around the centroid of each participant’s municipality. This environment was assumed to represent environmental conditions participants were likely to be exposed to, and also conditions related to tick questing activity in this area [19]. We used air temperature, relative humidity, and total precipitation. Monthly data were downloaded from the Copernicus Climate Data Store, ERA5 monthly averaged data on single levels from 1940 to present, DOI: https://doi.org/10.24381/cds.f17050d7 [20]. We used the air temperature at 2 m above the land surface (variable name: 2m temperature, t2m) measured in kelvin, which we converted to degrees Celsius (°C), and the dewpoint temperature at 2 m above the Earth’s surface (variable name: 2m dewpoint temperature, d2m). Then, we estimated the relative humidity using the August-Roche-Magnus approximation [21]. We also used the total precipitation (variable name: total precipitation, tp), which corresponds to the accumulated liquid and frozen water, including rain and snow, that falls to the Earth’s surface, measured in metres that we converted to millimetres.

We included land cover types to account for the environment surrounding place of residence, using CORINE Land Cover (CLC) 2018 [22]. In our analysis, we focused on 10 land cover types from 3 CLC main levels, for which tick presence was considered ecologically plausible: forest and seminatural areas, agricultural areas, and artificial surfaces (see SI section Environmental data) [22, 23].

Whilst CLC was a time-independent variable, monthly temperature, relative humidity, and precipitation were aggregated over 4-month periods to match the survey periods.

Incidence estimation

We estimated the incidence of participants reporting being bitten at least once within a one-year period. We considered one “epidemiological year” to correspond to three survey periods. Epidemiological year 2023 comprised the 2nd (Mar – Jun 2023), 3rd (Jul – Oct 2023), and 4th (Nov 2023 – Feb 2024) surveys, while the epidemiological year 2024 comprised the 5th (Mar – Jun 2024), 6th (Jul – Oct 2024), and 7th (Nov 2024 – Feb 2025) surveys. A case was defined as a participant living in region  r and bitten by ticks at least once during an epidemiological year  y, in their home region. The denominator included all survey participants who participated at least once in epidemiological year  y and lived in region  r.

The demographics (in terms of age and sex) of the surveyed participants did not reflect the French population as documented in the national statistics database (Supplementary Fig. 1) [18, 24]. Therefore, we adjusted the incidence estimation by assigning a post-stratification weight to each participant based on their age group and sex, therefore minimising the difference between the population surveyed and the general population. The weight used for a given combination of age-sex () was calculated as the ratio of the proportion of that age-sex group in the French general population to the proportion of the same group among survey participants (Eq. 1). The Grippenet/Covidnet cohort included adults aged above 18 years old, but since we collected only 17 questionnaires from 7 participants aged between 18 and 19 years old, we limited the incidence estimation to participants aged ≥ 20 years old.

In Eq. (1), is the number of individuals in the French population within the age group  a and sex  s, is the total French population aged 20 years and older, is the survey participants within the age group  a and sex  s, and  T referred to all survey participants. The age group  a in Eq. (1) was grouped as 20–34, 35–49, 50–64, 65–79, and above 80 years old. For population numbers, we used the French national statistics database from 2025 (INSEE) [24].

1

To estimate the incidence of participants being bitten at least once during an epidemiological year, we first calculated being the weighted number of bitten participants (at least once) in epidemiological year  y and in region  r (Eq. 2), with  j being the number of bitten participants in year  y and region  r. Then, we calculated , the weighted total number of unique participants per year  y and region  r (Eq. 3), with  i representing the number of unique participants in year  y and region  r.

2

3

From these, we estimated the adjusted proportion of bitten participants among the total participants () (Eq. 4). We then multiplied by the population size per region () to calculate , the number of bitten people in region  r and year  y (Eq. 5). Then, we produced annual regional incidence estimates () by dividing by the population size per region (), scaled per 100,000 inhabitants (Eq. 6):

4

5

6

Looking at Eqs. 4, 5 and 6, we could simplify the incidence calculation by taking the adjusted proportion of bitten participants among the total participants () and multiply by 100,000, assuming that reflected the adjusted proportion of bitten individuals in the wider population (Eq. 7).

7

To provide national annual estimates () (Eq. 8), we summed up the regional estimates weighted by each region’s proportion of the national population (). We also adjusted the variances based on regional population size to obtain the national variance, which was then used to compute the national 95% confidence intervals (CI) (see code in Data availability section).

8

Bayesian hierarchical statistical model

To investigate the behavioural and environmental factors associated with human tick bites, we used a Bayesian hierarchical two-part model with INLA (Integrated Nested Laplace Approximation) [25]. We modelled the risk of reporting at least one tick-bite occurrence over one survey period using a binomial likelihood, and conditional on being bitten, we modelled the number of tick-bite episodes using a zero-truncated Poisson likelihood (nzpoisson or non-zero Poisson). The outcome variables for both models were defined per participant per survey.

Participant ID was included as an independent and identically distributed (IID) random effect to account for repeated participation across surveys, while the survey period was modelled as a seasonal random effect to account for the survey seasonality [25]. Candidate variables potentially associated with the risk of reporting at least one tick bite were: age group, sex, whether participants’ home had a green outdoor space (garden and/or field), frequency of outdoor activity in green areas during the study period (everyday, two to three times a week, once a week, once or twice a month, less than once a month), types of outdoor environments visited during the study period (forest, meadow, private garden, mountain, public parks, agricultural land), types of outdoor activities (leisure, professional activity, hunting/fishing, sport), frequency of implementing prevention measures against tick bites during outdoor activities, and region of residence. Regarding the types of outdoor activities, leisure included slower motion activities such as hiking, walking, gardening, and picnic, whilst sports included running and collective sports in green areas. Hunting/fishing were also categorised separately as they assumed a higher risk of exposure due to potential contact with animals or long-standing activities in forest (for more details see SI section Study design). Four environmental variables tested: 4-month averaged temperature, 4-month averaged relative humidity, 4-month averaged total precipitation, and the percentage of each CORINE Land Cover (CLC) type (forest and seminatural areas, agricultural areas, artificial surfaces), within a 10 km buffer around the place of residence.

For the zero-truncated Poisson part, we tested the frequency of outdoor activity in green areas, the frequency of using prevention measures against tick bites during such activities, and also variables focusing on tick-bite experiences, such as the types of environments where participants reported getting bitten (forest, meadow, private garden, agricultural land, public park, indoor, farm, near waterbody), the types of outdoor activities when bitten (leisure, professional activity, hunting/fishing, sport), whether the participant was bitten outside their department of residence at least once, and post-tick-bite behaviours including whether participants consulted a GP at least for one episode, and the time elapsed between tick bite and tick removal (“time to tick removal “: within 24 h, after 24 h, other).

As participants could report up to four tick-bite episodes per survey, and each episode could involve multiple ticks with different removal times, the time to tick removal variable was defined at the participant-survey level. If at least one tick from one episode was removed after 24 h, the corresponding participant-survey was categorised as “after 24 h”. Only participant-surveys in which all reported ticks from all episodes were removed within 24 h were categorised as “within 24 h”. The group “other” corresponds to participant-surveys who did not report removing any ticks; either the ticks fell or the participants did not remember.

In the model, we removed participants without any outdoor activity in green areas reported during the study period, with the assumption that they were not at risk of exposure (15.1% overall, n = 3,853 participant-surveys, ranging from 9.4% to 19.6% depending on survey season) (Supplementary Table 1). For participant-surveys who reported outdoor activity with unknown frequency (0.8%, n = 169/21,607), we performed multiple imputations using the mice package in RStudio [26]. We set the frequency of activity as an ordinal outcome, which was modelled using proportional odds regression (polr), with participants age group, season (Nov-Feb, Mar-Jun, Jul-Oct), and whether participants have garden or field in their residence as predictors [27]. We performed the imputations five times, and we extracted the values from the first imputation to be used in data description and the Bayesian model.

Correlations between variables were checked using Spearman’s rho (ρ) for numeric variables (Supplementary Table 2) [28]. For the categorical variables, we measured the strength of association based on the Phi () coefficient for 2 × 2 contingency tables (binary variables) or Cramér’s V for other categorical variables with more than two levels (Supplementary Fig. 2) [29].

For model selection, we first run the full model with all candidate variables. Sociodemographic variables (age, sex, and region of residence) were forced in the Binomial model. Variable selection was done first for the Binomial model, then for the zero-truncated Poisson model. Within each model, we investigated the set of variables by group (meteorological, environmental, and behavioural) and selected the best combination, based on the Deviance Information Criterion (DIC) and the Watanabe-Akaike Information Criterion (WAIC), where lower values indicate a better model fit. We used Gaussian priors with mean 0 and precision 1. For the fixed effects, we exponentiated the posterior distributions, to obtain odds ratio (OR) for the Binomial model and incidence rate ratio (IRR) for the zero-truncated Poisson model, with their 95% credible intervals (CrI). All statistical analyses were performed in RStudio version 2025.05.1. The two-part Bayesian model was performed using R-INLA package version 24.12.11 [30].

Results

Participatory survey data

Overall, 4,840 people from Grippenet/Covidnet cohort participated in the study, and 47.3% participated in seven survey periods (Supplementary Fig. 3), resulting in a total of 25,460 completed surveys. The participants’ age during their first participation ranged from 18 to 98 years old, with a median age of 63 years old [IQR, 52–71]. There were more women (59.6%; n = 2,887) participants in the survey compared to men (40.4%; n = 1,953). Grippenet/Covidnet cohort showed some familiarity with the survey topic, with 97.7% of participants reporting prior knowledge of ticks or TBD. Moreover, based on the responses provided in the last survey participation of each participant, 2,295 out of 4,840 participants (47.4%) reported having been bitten by ticks at least once in their lifetime.

The second section of the survey focused on tick-bite episodes that occurred during the study period. A total of 1,348 tick-bite episodes were reported from 535 participants (11.1%). 260 participants (48.6%) reported one tick-bite episode, while 275 participants (51.4%) reported two or more episodes across all survey periods. We observed a seasonal pattern, with a higher number of tick bites reported in Mar-Jun and Jul-Oct surveys, peaking in June 2023 (n = 142 tick-bite episodes) and June 2024 (n = 138 tick-bite episodes) (Fig. 1, blue line). Within each survey period, we found that around 5–6% of participants were bitten by ticks over the Mar-Jun or Jul-Oct surveys, while less than 2% of participants reported being bitten in the Nov-Feb surveys (Fig. 1, red bars).

Fig. 1.

Temporal pattern of self-reported tick-bite episodes in the Grippenet/Covidnet cohort. The left y-axis (blue line) shows the number of reported tick-bite episodes per month, and the right y-axis (red bars) shows the percentage of participants bitten at least once among all participants per survey season. All tick-bite episodes are shown, regardless of location, within or outside the participant’s region of residence

We collected spatial information from 1,258 tick-bite episodes (Fig. 2). 18.7% of tick-bite episodes were reported in Nouvelle Aquitaine (n = 235), 18.2% in Auvergne-Rhône-Alpes (n = 229), 14.1% in Occitanie (n = 177), 12.7% in Grand Est (n = 160), 6.8% in Bretagne (n = 85), 5.3% in Bourgogne-Franche-Comté (n = 67), 5% in Ile-de-France (n = 63), 4.9% in Centre-Val de Loire (n = 62), 4.6% in Hauts-de-France (n = 58), 4.1% in Normandie (n = 52), 3.3% in Pays de la Loire (n = 41), and 2.3% in Provence-Alpes-Côte d’Azur (n = 29). No participant reported being bitten in Corsica. Overall, 83.2% tick-bite episodes (n = 1,047) occurred in participants’ region of residence, while 16.8% happened outside their region of residence (n = 211).

Fig. 2.

Total number of self-reported tick-bite episodes per region over the study period (November 2022 – February 2025). ARA: Auvergne-Rhône-Alpes, BFC: Bourgogne-Franche-Comté, BRE: Bretagne, CVL: Centre-Val de Loire, COR: Corsica, GE: Grand Est, HDF: Hauts-de-France, IDF: Ile-de-France, NOR: Normandie, NA: Nouvelle-Aquitaine, OCC: Occitanie, PDL: Pays de la Loire, PACA: Provence-Alpes-Côte d'Azur

Information regarding outdoor activity performed when the tick bites occurred was available for 1,298 tick-bite episodes (Fig. 3A). 82.1% of reported tick bites happened during leisure activities (n = 1,066), 10.1% during sport activity (n = 131), 2.5% during professional activity (n = 32), 0.6% during hunting/fishing (n = 8), and 4.7% during other unspecified activities (n = 61). The information on the environments of tick bites was available for 1,232 tick-bite episodes (Fig. 3B). More than half of tick-bite occurrences were reported from forest areas (59.2%, n = 729), followed by 40.6% in meadows (n = 500), and 27.3% in private gardens (n = 336). Meanwhile, the other environments such as public park (n = 19), vicinity to waterbody (n = 20), agricultural land (n = 52), farm (n = 9), and indoor/inside a dwelling (n = 28), were reported in less than 5% of tick-bite episodes.

Fig. 3.

Percentage of tick-bite episodes by (A) type of activity performed when the bite occurred and (B) environment during tick-bite episodes. Categories are ordered from highest to lowest percentage, except the “other” category in (A). Leisure activity included hiking, walking, picnic, and gardening; Sport included running or collective sports in green areas

In most tick-bite episodes, participants only reported getting bitten by one (84.7%, n = 1,061) or two ticks (12.2%, n = 153). Only a small fraction (3.1%) of tick-bite episodes were related to 3 to 10 ticks. Throughout the seven surveys, in 63.9% of tick-bite episodes (n = 830), the ticks were removed within 24 h, while in 34.7% (n = 451) episodes, the ticks were removed the following day (within 48 h), and under 2% (n = 24) reported that the tick had detached spontaneously. Participants were allowed to choose more than one removal time, as they could get bitten by more than one tick in one episode. The method of tick removal was also surveyed, and participants could choose more than one method (see SI section Study design). Among tick-bite episodes with information regarding general practitioner (GP) consultation (n = 1,306), 108 episodes were followed by a consultation with a GP (8.3%), reported by 79 out of 535 bitten participants (14.8%). The average percentages of tick-bite episodes followed by a GP consultation varied from 5.9% to 19.4% depending on the survey period (Fig. 4).

Fig. 4.

Percentage of tick-bite episodes followed by general practitioner (GP) consultation per survey period. Blue bars show the average percentage and black whiskers correspond to 95% CIs

The third section of the survey focused on participants’ outdoor activities in natural environments or green areas, regardless of tick-bite reporting. Overall, 84.9% of participant-surveys (n = 21,607) reported having outdoor activities during the study period, and this number varied between 80.4% and 90.6% across different survey seasons. Among participant-surveys who reported outdoor activities during a study period (n = 21,607), 15.6% reported outdoor activities everyday, 39.2% two to three times a week, 23.7% once a week, 16.3% reported once or twice a month, and 5.2% reported having outdoor activities less than once a month. These frequencies varied from one survey season to another (Supplementary Table 1).

Moreover, we also surveyed the types of outdoor activities and the types of green outdoor environments visited by the participants during each study period. Participants were allowed to choose more than one type of outdoor activity and environments. From 7 surveys (n = 21,607 participant-surveys), 91.6% participant-surveys reported outdoor leisure activities (including hiking, walking, gardening, picnic), 29.5% chose sport (running in green areas, playing collective sport on grass), 3.3% professional activity, and 1.5% hunting/fishing. For these outdoor activities, the most frequently visited environments were forest (60.5%) and private garden (59.0%), followed by public park (42.7%), meadow (40.0%), mountain (23.5%), and agricultural area (16.2%). The use of prevention measures (wearing long trousers, long sleeves, or using repellents) during outdoor activities was reported in 33.4% (n = 7,214) participant-surveys.

Incidence of bitten participants

The incidence of participants self-reporting being bitten by ticks in their home region at least once over an epidemiological year was estimated at 5,615 (95% CI [4,910-6,321]) and 5,402 (95% CI [4,670-6,133]) per 100,000 inhabitants, for the March 2023-February 2024, and March 2024-February 2025 periods, respectively. In both years, the highest incidences were found in Grand Est with annual estimates exceeding 10,000 per 100,000 inhabitants (Fig. 5). Regional incidences slightly varied between years (Fig. 5). In epidemiological year 2024, the incidences in Grand Est and Nouvelle-Aquitaine were lower, while the incidence in Provence-Alpes-Côte d’Azur was higher compared to 2023. However, the 95% CIs for both years overlapped. Details on the incidence of participants reporting tick bites per survey season, and stratified by the frequency of outdoor activity in green areas are available in Supplementary Materials Supplementary Fig. 4.

Fig. 5.

Incidence estimation of bitten participants per region and per epidemiological year with 95% confidence intervals. March 2023-February 2024 is referred to as epidemiological year 2023, and March 2024-February 2025 as epidemiological year 2024. Regions are listed in decreasing order according to the 2023 estimates

Behavioural and environmental factors associated with tick bites

The results of the best model selected are presented in Table 1 (binomial part) and Table 2 (zero-truncated Poisson part). For the binomial part of the model (Table 1), being female was positively associated with reporting at least one tick bite (OR = 1.30; 95% CrI [1.07–1.58]) compared to male; whilst older age groups (45–74 and over 75 years old) were associated with lower risk compared to participants aged 18–44 years. Having a green outdoor area at home was also associated with a higher risk of reporting a tick bite (OR = 1.46; 95% CrI [1.13–1.90]). Regarding the region of residence, living in Grand Est (OR = 1.94; 95% CrI [1.29–2.92]) almost doubled on average the risk of reporting a tick bite, while living in Provence-Alpes-Côte d’Azur or Corsica was associated with a lower risk compared with Ile-de-France region. Average temperature above 7 °C in the vicinity of residence was also associated with an increased risk of reporting a tick bite. In addition, for each 10% increase of forest and seminatural area cover within the 10 km buffer, the risk of reporting a tick bite increased by 12%, while the increase in artificial surface cover was associated with a lower risk. Regarding participants’ behaviours, visiting either forest (OR = 2.35; 95% CrI [1.88–2.92]) or meadow (OR = 1.92; 95% CrI [1.60–2.31]) was associated with an increased risk of self-reporting a tick bite, whilst having outdoor activity in private garden, agricultural land, public park, or mountain was associated with a lower risk. Regarding frequency, participants with outdoor activity in green areas two to three times a week (OR = 1.36; 95% CrI [1.10–1.70]) and everyday (OR = 1.97; 95% CrI [1.51–2.57]) were at higher risk of reporting a tick bite, compared to those reporting outdoor activity once a week. Finally, using prevention measures against tick bites during outdoor activity was associated with a higher risk of reporting a tick bite (sometimes: OR = 1.63, 95% CrI [1.34–1.98]; always: OR = 1.46, 95% Crl [1.14–1.88]), compared with no prevention measures.

Table 1.

Results of the binomial model. Odds ratio (OR) and 95% credible intervals (CrI) for the variables included in the final model

Variable	OR	95% CrI
Sociodemographic
Sex (Ref: Male)
Female	1.30	[1.07–1.58]	*
Age (Ref: 18–44)
45–74	0.63	[0.46–0.86]	*
≥ 75	0.35	[0.23–0.52]	*
Green outdoor area at home (Ref: No)
Have garden and/or field	1.46	[1.13–1.90]	*
Region (Ref: Ile-de-France)
Centre-Val de Loire	1.30	[0.79–2.15]
Bourgogne-Franche-Comté	0.83	[0.50–1.39]
Normandie	0.74	[0.42–1.31]
Hauts-de-France	0.86	[0.51–1.44]
Grand Est	1.94	[1.29–2.92]	*
Pays de la Loire	0.83	[0.50–1.39]
Bretagne	0.64	[0.38–1.06]
Nouvelle-Aquitaine	1.02	[0.69–1.52]
Occitanie	0.88	[0.58–1.32]
Auvergne-Rhône-Alpes	1.27	[0.88–1.84]
Provence-Alpes-Côte d’Azur	0.29	[0.17–0.51]	*
Corsica	0.21	[0.05–0.85]	*
Environmental conditions in the vicinity of residence
Temperature (Ref: ≤ 7 °C)
> 7 - ≤ 13 °C	1.68	[1.18–2.37]	*
> 13 - ≤ 17 °C	2.00	[1.33–3.00]	*
> 17 °C	1.56	[1.01–2.42]	*
Land cover types (per 10%)
Forest and seminatural areas	1.12	[1.06–1.18]	*
Artificial surfaces	0.60	[0.36–0.99]	*
Behaviours
Environment visited during study period (Ref: Did not visit)
Forest	2.35	[1.88–2.92]	*
Meadow	1.92	[1.60–2.31]	*
Private garden	0.60	[0.50–0.73]	*
Agricultural land	0.80	[0.64–0.99]	*
Public park	0.36	[0.29–0.44]	*
Mountain	0.74	[0.60–0.91]	*
Types of outdoor activities performed during study period (Ref: Did not perform)
Professional activity	1.18	[0.79–1.78]
Hunting/fishing	0.61	[0.30–1.21]
Sport	0.80	[0.65–0.97]	*
Leisure	0.77	[0.56–1.06]
Frequency of outdoor activity (Ref: Once a week)
Less than once a month	0.57	[0.33–0.98]	*
1–2 times a month	0.64	[0.47–0.89]	*
2–3 times a week	1.36	[1.10–1.70]	*
Everyday	1.97	[1.51–2.57]	*
Frequency of using prevention measures (Ref: No prevention)
Sometimes	1.63	[1.34–1.98]	*
Always	1.46	[1.14–1.88]	*

* Variables for which the 95% CrI excludes 1

Table 2.

Results of the zero-truncated Poisson model. Incidence Rate Ratio (IRR) and 95% credible intervals (CrI) for the variables included in the final model

Variable	IRR	95% CrI
Environments where participants were bitten
Forest	2.08	[1.59–2.71]	*
Meadow	1.37	[1.09–1.73]	*
Private garden	1.48	[1.13–1.93]	*
Agricultural land	1.09	[0.68–1.74]
Public park	1.07	[0.53–2.17]
Indoor	1.76	[0.75–4.12]
Farm	0.62	[0.16–2.39]
Near waterbody	0.86	[0.36–2.05]
Behaviours
Frequency of outdoor activity (Ref: Once a week)
Less than once a month	0.50	[0.17–1.45]
1–2 times a month	1.00	[0.58–1.73]
2–3 times a week	1.30	[0.94–1.80]
Everyday	1.39	[0.96–2.02]
1–Frequency of using prevention measures (Ref: No prevention)
Sometimes	1.32	[1.02–1.72]	*
Always	1.48	[1.05–2.10]	*
Post-tick bite behaviours
Consult GP for at least 1 bite episode (Ref: No)
Yes	0.69	[0.44–1.09]
When was the tick removed (Ref: Within 24 h)
After 24 h	1.75	[1.38–2.23]	*
Other	1.59	[0.90–2.83]

* Variables for which the 95% CrI excludes 1

The results of the zero-truncated Poisson part of the model (Table 2) showed that being bitten in either forest, meadow, or private garden was associated with a higher number of self-reported tick-bite episodes. The use of prevention measures during outdoor activity was also associated with a higher number of bite episodes (sometimes: IRR = 1.32, 95% CrI [1.02–1.72]; always: IRR = 1.48, 95% CrI [1.05–2.10]). Concerning post-bite behaviours, participants who removed at least one tick after 24 h (IRR = 1.75; 95% CrI [1.38–2.23]) were associated with a higher number of self-reported tick-bite episodes compared to those who always remove the ticks within 24 h.

From the zero-truncated Poisson model, we obtained posterior mean predictions of the expected number of self-reported tick bites, which increased with observed values (Fig. 6). The mean predicted values were 1.2 for participant-surveys reporting one tick-bite episode, 1.7 for two episodes, 2.3 for three episodes, and 3 for four or more episodes, indicating that the model captured an increasing trend in bite intensity across observed groups.

Fig. 6.

Distribution of posterior mean zero-truncated Poisson predictions stratified by observed tick-bite counts

Discussion

This study explored behavioural and environmental factors associated with self-reported tick bites in France, using a longitudinal survey among Grippenet/Covidnet cohort participants. 47.4% of our survey participants (n = 2,295) reported being bitten at least once in their lifetime, higher than Santé publique France barometer in 2019 which reported 30.2% of surveyed individuals had been bitten at least once [31]. This difference could partly be explained by the difference in population surveyed, where the Grippenet/Covidnet population is older than the French population [18], resulting in higher chance of getting bitten at least once in their life. Tick-bite reports showed a clear seasonality that is similar to tick questing activity, where tick abundance reached its peak in May-June and September [9]. Moreover, the decreasing trend after the month of June is consistent with a study by Perret et al. which showed that the proportion of questing (host-seeking) ticks decreased after the drought around June [32].

The estimated incidence of people getting bitten by ticks at least once over a year was 5,615 (95% CI [4,910-6,321]) per 100,000 inhabitants in 2023 and 5,402 (95% CI [4,670-6,133]) per 100,000 inhabitants in 2024. Whilst excluding tick bites occurring outside the home region could lead to an underestimation of the national incidence, having a health-focused cohort, more likely to have higher awareness of tick bites, might lead to an overestimation. However, our survey results corroborated with Santé publique France study conducted in 2024, reporting that 4.9% (95% CI [4.6%-5.1%]) of the adult population of metropolitan France had been bitten by a tick at least once in the previous 12 months [33]; with similar regional findings exhibiting higher proportions of bitten participants mainly in north-eastern and south-western France, and lower estimates in Provence Alpes-Côte d’Azur and Corsica. Whilst the lack of tick-bite reports in Corsica from our survey was probably due to the limited number of participants, our results support the quality of the data we collected. In another study (“Sentick”) conducted in France between May 2023 and April 2024, the incidence of tick bite seen in general practice was estimated at 324 per 100,000 inhabitants (95% CI [290–369]) [17]. Whilst this is more than 10 times lower than our average estimation of self-reported tick bites, it remains consistent with our findings which indicated that only 8.3% of self-reported tick bites were followed by a GP consultation [34].

Our Bayesian model showed that female participants were at a higher risk of reporting tick bites than male participants, whilst Santé publique France found higher percentage of bitten individuals among males compared to females (5.3% vs. 4.5%) [33]. However, Santé publique France also reported that females were more likely to look for ticks after potential exposure compared to males (68% vs. 62.1%) [33], suggesting that increased awareness among females might explain part of this contrasted result. Regarding age groups, our study found a lower risk of reporting tick bites among older individuals (45–74 and ≥ 75 years old), whilst the incidence of tick-bite consultation estimated by Réseau Sentinelles showed higher incidence in age group 65 years and older (418 per 100,000; 95% CI [354–498]) compared to 15–64 years old (246 per 100,000; 95% CI [211–286]) [17]. One possible explanation for this discrepancy could be that older individuals performed outdoor activities less frequently [35], or have different healthcare-seeking behaviours. We highlight that healthcare-seeking behaviour following a tick bite should be investigated further in the future.

Our model showed that participants living in the Grand Est region had a higher risk of reporting tick bites compared to Ile-de-France (Paris and surrounding municipalities). Meanwhile, Provence-Alpes-Côte d’Azur and Corsica were associated with a lower risk. These differences in regional trends align with the distribution of I. ricinus ticks, where northeast, central, and southwest France were considered as a suitable habitat, while Mediterranean region was considered less favourable for I. ricinus tick development [36]. This finding was also in line with LB incidence, for which Grand Est is considered at higher risk [3]. Although our survey did not collect information on tick species involved in self-reported tick bites, data from CiTIQUE showed that 96% of ticks collected from humans in France were identified as I. ricinus, the primary vector of LB in France [37]. This could suggest that the majority of tick bites in our study were likely caused by I. ricinus.

Having a green outdoor area (garden, field) at home was associated with a higher risk of reporting a tick bite, and private garden was an environment that was associated with a higher number of tick bites. Aligning with these findings, a citizen science study conducted in Nancy (north-eastern France) private gardens detected the presence of ticks in 32% of surveyed yards [38]. In contrast, our model showed that spending time in private garden, agricultural land, public park, or mountain was protective against reporting a tick bite. These results should be interpreted cautiously, as the former variable investigated tick-bite location, whilst the latter meant to measure potential exposure. In other words, those spending time in a garden may not be necessarily bitten there. Indeed, among participants without a green outdoor area at home, we found similar proportions of people reporting at least one tick bite regardless of whether they spent time in a private garden or not (~ 2%). However, among participants with green outdoor areas at home (garden and/or field), the proportion of participants reporting at least one tick bite was higher among those who did not spend time in a private garden during study period (6.2%) compared to those who spent time in a private garden (4.4%) (Supplementary Fig. 5). Over seven survey periods, private garden was the second most visited green outdoor area, with 27.3% of tick bites were reported there, which was in line with CiTIQUE data reporting 23% of human tick bites occurred in garden [39].

Average temperature above 7 °C in the vicinity of residence was also positively associated with the risk of reporting a tick bite. This can be explained by the fact that at cooler temperatures participants may spend less time outdoor or wear more clothing layers when outdoors, which provides additional protection against tick bites. An increase in forest cover within the buffer zone, as well as having outdoor activities in either forest or meadows, were associated with a higher risk of reporting tick bites, in line with the literature, which reports an increased presence of ticks in forests areas, especially shrubs, deciduous, and mixed forests [1]. Furthermore, forest or meadows were also environments associated with an increased number of self-reported tick bites.

Having outdoor activities in green areas more than once a week was associated with an increased risk of reporting at least one tick bite. A survey conducted in Sussex County, New Jersey, reported that the risk of finding ticks increased with weekly time spent outdoors [13]. However, in our study, frequency of outdoor activity was not associated with the number of reported tick-bite episodes, potentially reflecting lack of statistical power in the zero-truncated Poisson model due to the smaller number of observations among participant-surveys reporting at least one episode. Furthermore, participation in sport activities, such as running or team sport on grass, was associated with a lower risk of reporting at least one tick bite. This might reflect differences in exposure behaviour, as sport often involve movement and shorter contact duration with vegetation, thus potentially reducing opportunity for tick attachment.

We found that using prevention measures against tick bites was associated with a higher risk of reporting a tick bite and a higher number of reported tick-bite episodes over a study period, compared to not using prevention measures. Most studies have mixed results on the effectiveness of the use of personal protective measures [40]. Indeed, personal protective measures against tick bites are not 100% protective, and their effectiveness heavily depends on the way they are implemented. In the example of repellents, the effectiveness also depends on the type of molecule and dosage used. In addition, confounding factors such as frequency of exposure and awareness might explain those results. Jepsen et al. reported that prevention measures were more common among individuals with tick-bite experience [41], and similarly, barometer from Santé publique France showed that prevention measures were more frequent in regions with higher incidence of LB [31]. In our survey, a high number of participants (97%) reported to have heard about ticks or TBD, yet we did not check that knowledge using tick or TBD-related images, and around 33% participant-surveys reported the use of protection measures during outdoor activities in green areas. Furthermore, we did not ask participants to specify which measures they actually applied. Although some examples were mentioned in the survey question (long trousers, long sleeves, and repellents), participants might have considered other common measures, such as checking for ticks after outdoor activities [41]. Thus, our study unfortunately did not investigate well enough this variable.

Finally, concerning participants behaviour following tick bites, removing at least one tick after 24 h was associated with a higher number of self-reported tick-bite episodes compared to removing all ticks within 24 h. This might indicate that participants who reported more episodes were more likely to experience delayed detection of at least one tick, even though tick check was reported as the second most common prevention measure according to the survey by Santé publique France [33]. In addition, while tick bites mostly occurred on the legs, ticks could also attach to less visible areas such as behind the ears, the scalp, or the genital region, which may result in delayed detection [42].

This study has several limitations that need to be considered when interpreting the results. First, the use of a participatory survey could result in selection bias, especially as the cohort was not representative of the general French population. This was partially addressed in incidence estimation by using post-stratification weighting. In addition, this study may be subject to recall bias and exposure misclassification as we relied on self-reported tick-bite occurrences. Furthermore, we did not collect detailed information on specific prevention measures taken by participants during outdoor activities. This limits our ability to evaluate the effect of each measure on the risk of reporting tick bites. In particular, prevention measures like tick check, which occurred after outdoor activity, may increase the likelihood of detecting ticks and potentially introduce reporting bias across participants. Finally, the zero-truncated Poisson model was based on smaller sample compared to the binomial model, as only participant-surveys with at least one tick bite were included in the zero-truncated Poisson model, which may limit the ability to detect certain associations with the number of tick bites.

Conclusion

This study presents nationwide and seasonal estimates of self-reported tick-bite occurrence in a participatory cohort and identifies environmental and behavioural factors associated with reporting a tick bite. Being female, having a green outdoor area at home, having outdoor activity in green areas more than once a week, living in Grand Est region, visiting forests or meadows, living in areas with an average temperature above 7 °C and a high percentage of forest and seminatural land cover, and using prevention measures were positively associated with reporting a tick bite. Meanwhile, types of environments in which participants reported being bitten (forest, meadow, or private garden), use of prevention measures, and removing at least one tick after 24 h were associated with a higher number of reported tick-bite episodes. The associations with the use of prevention measures might reflect greater exposure or awareness of the risk of tick bite. Our study also identified regional differences in the risk of reporting tick bites.

This study may contribute to public health practice by identifying behavioural and environmental factors associated with self-reported tick bites. Our findings suggest that prevention efforts should not be limited to individuals frequently exposed to forest environments, but should also include those spending time in peri-domestic green areas where tick bites have also been reported. In addition, awareness of the risk of tick bites may influence the use of protective measures, highlighting the importance of improving public awareness that tick bite can occur in a wide range of green environments. Furthermore, future studies integrating multiple data sources, such as national participatory surveillance or GP consultation data, could help refine the identification of populations, regions, and periods at higher risk of tick bites in France at a finer resolution. Finally, studies incorporating tick species or detailed information on preventive behaviours would help improve understanding of tick-bite risk factors.

Abbreviations

CI

Confidence Intervals

CrI

Credible Intervals

CLC

CORINE Land Cover

DIC

Deviance Information Criterion

GP

General Practitioner

IID

Independent and Identically Distributed

INLA

Integrated Nested Laplace Approximation

IRR

Incidence Rate Ratio

LB

Lyme Borreliosis

OR

Odds Ratio

TBD

Tick-Borne Disease

TBE

Tick-Borne Encephalitis

WAIC

Watanabe-Akaike Information Criterion

Authors’ contributions

CB, JD, MD, and RM conceptualised the study and designed the questionnaire. MD coordinated the set-up of the survey in Grippenet/Covidnet cohort. IES, RM, and JG analysed the data. IES, CB, JD, JG, MD, and RM interpreted the results. IES and RM wrote the initial manuscript. IES, CB, JD, JG, MD, and RM commented and approved the final manuscript.

Funding

IES is funded by a Sorbonne Université PhD fellowship. JD is funded by the interdisciplinary program ARTEMIS of “Initiative d’Excellence Lorraine” (ANR-15-IDEX-04-LUE). This study was conducted under the Agence Nationale de la Recherche JCJC MoZArt project (www.mozartonehealth.com), grant number ANR-22-CE35-0003. Grippenet/Covidnet cohort is supported by public funds from Sorbonne Université, Inserm, and Santé publique France.

Data availability

Aggregated and post-processed data are available upon request. The code for incidence estimation and two-part Bayesian model can be found in https://github.com/ilona-suhanda/Tick-bite-paper.

Declarations

Ethics approval and consent to participate

Grippenet/Covidnet was reviewed and approved by the French Advisory Committee for research on information treatment in the field of health (i.e., CCTIRS, authorization 11.565), and by the French National Commission on Informatics and Liberty (i.e., CNIL, authorization DR-2012-024). Participant consent is informed and provided through registration in the cohort. This study was considered as non-interventional research according to article L1221–1.1 of the public health code in France and only required the non-opposition of the participant (per article L1211–2).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.