(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .
Preventing Lyme disease through identification of key beliefs [1]
['Pierre Valois', 'Observatoire Québécois De L Adaptation Aux Changements Climatiques', 'Oqacc', 'Faculté Des Sciences De L Éducation', 'Secteur Mesure Et Évaluation', 'Université Laval', 'Québec', 'Catherine Bouchard', 'National Microbiology Laboratory', 'Public Health Risk Sciences Division']
Date: 2023-12
Lyme disease and other vector-borne diseases are on the rise because of climate change. In the province of Quebec, Canada, Lyme disease has become a public health problem deserving the attention of health authorities. Despite their recognized effectiveness at preventing tick-to-human transmission, rates of adoption of Lyme disease adaptive behaviours (LDAB) remain relatively low in the population. Using the Theory of Planned Behaviour (TPB), the aim of this study is to identify specific and actionable beliefs associated with the adoption of Lyme disease adaptive behaviours. Specifically, 2,011 people were surveyed to determine the decision-making process behind specific beliefs, which could be targeted for raising awareness. Statistically significant associations were found between the three determinants of the TPB (i.e., attitudes, perceived social pressure and perceived behavioral control) and the intention to adapt. In addition, the intention itself was significantly associated with adopting LDAB. Belief-based analyses indicated that 8 primary beliefs (4 behavioral beliefs, 2 normative beliefs, and 2 control beliefs) were associated with LDAB. Among these, control beliefs (barriers and facilitating factors) appeared to have the greatest impact on adaptation. These findings can be used to guide educational and awareness-raising campaigns to promote LDAB by changing or reinforcing these primary beliefs.
In addition, the TPB states that attitudes, perceived social norms, and perceived behavioral control are based on three kinds of beliefs: behavioral beliefs (i.e., beliefs about the potential benefits and risks of adopting a particular behaviour), normative beliefs (i.e., beliefs about the extent to which their important referents (e.g. friends, physicians, etc.) approve of them protecting themselves from Lyme disease), and control beliefs (i.e., their perception of the presence of factors that facilitate or hinder their adoption of LDAB) ( Fig 1 ). These beliefs underlie the formation of attitudes, perceived social norms, and perceived behavioral control. Understanding these specific beliefs leads to a more comprehensive understanding of the full psychosocial process that leads to the adoption of health behaviours aimed at preventing Lyme disease, an essential step toward designing effective health communication and interventions [ 59 ].
According to the TPB, people’s intention to adopt preventive behaviours against tick bites and Lyme disease should therefore be high if: (a) they have favourable attitudes toward adaptive behaviours (i.e., they positively value the adoption of LDAB); (b) they believe that significant others (e.g., relatives, physicians) have expectations about or support LDAB; and (c) they believe they have the ability to perform LDAB (i.e., believe they can overcome any obstacles to carrying out their intentions, and/or recognize facilitating factors such as prior experience with Lyme disease or support from friends to overcome obstacles).
Consistent with the concept of reasoned action, which is the basis of the TPB, people’s behavioural intentions towards preventing tick bites and Lyme disease are presumed to be guided by some measure of deliberation [ 76 ]. Using the TPB as a guide, the intention to adopt preventive behaviours against tick bites and Lyme disease precedes the adoption of LDAB ( Fig 1 ). Similarly, this intention is preceded by three factors: attitudes, social norms, and perceived behavioural control. Attitudes relate to the positive or negative values a person has towards the adoption of LDAB, social norms correspond to the social pressure a person perceives in engaging or not in LDAB, and perceived behavioral control refers to a person’s perception of their own capacity to perform LDAB [ 77 ]. In addition, an individual’s perception of control, to the extent that it faithfully reflects actual control, can moderate the effect of intentions on behavior [ 78 ]. In Fig 1 , this moderating effect is illustrated by the dotted arrow.
A variety of other theories are available to assess health and adaptation behaviors, such as the health beliefs model (HBM) [ 69 ]), the protection motivation theory (PMT) [ 70 ], the model of private proactive adaptation to climate change [ 71 ], or the Value-Belief-Norm model [ 72 ]. The TPB, however, offers a unique advantage by combining the main variables discussed in other models into a single theoretical framework [ 73 ]. For instance, the notions of efficacy and control included in the definition of adaptation beliefs, a variable proposed by Grothmann et al. [ 74 ] are well represented by the behavioral and control beliefs of the TPB. Similarly, the behavioral evaluation construct found in the HBM, which integrates beliefs about the advantages and the effectiveness of a behavior are reflected in the behavioral beliefs of the TPB. In addition, according to Fishbein & Ajzen [ 59 ], it is possible to add other predictors to the TPB. The perception of risk, for example, an important construct of the PMT (threat assessment) and the HBM (perception of the threat) has been included successfully as a construct of the TPB in other studies [ 75 ]. These characteristics have motivated us to retain the TPB as our theoretical framework.
Previous studies examining LDAB have focused mainly on evaluating their overall levels of adoption and understanding the relationships between adaptation and various psychosocial variables. These include perceived risk of contracting the disease, knowledge about the disease [ 44 – 47 ] and its mental and physical impacts [ 48 ], as well as attitudes and barriers to adaptive behaviour [ 49 ]. Others also explore the perceived efficacy of protective behaviours against the disease, perceived control over Lyme disease prevention measures [ 50 , 51 ], and the acceptability and effectiveness of Lyme disease prevention programs [ 52 , 53 ]. Past research has also evaluated the determinants of intention to adopt LDAB using psychosocial models such as the Attitude-Social-Influence-Efficacy (ASE) model [ 54 ] and the Theory of Planned Behaviour (TPB) [ 55 , 56 ]. Indeed, Omodior et al. [ 55 ] and Valois et al. [ 56 ] found that the constructs of the TPB (attitudes, social norms, and perceived behavioural control) were effective predictors of intention to engage in LDAB. Complementary to these previous results, the current research seeks to uncover key beliefs underlying these precursors of intention, which represents a gap in the literature on LDAB. As a well-established social-cognitive model, the TPB is an excellent tool to identify beliefs that can be addressed in public health messages aiming to limit Lyme disease by promoting adoption of LDAB [ 57 ].
Lyme disease is caused by the Borrelia burgdorferi sensu lato (s.l.) complex of spirochete bacteria, which is transmitted to humans during a bite from an infected tick. The first and most common symptom of Lyme disease is a red skin rash known as erythema migrans, which is present at the site of the tick bite in 70% to 80% of cases [ 35 ]. Other symptoms include fever, fatigue, headache, stiff neck, and muscle and joint pain. In the absence of early treatment, the disease can cause chronic multisystem morbidity including cardiac or neurological issues and arthritic joint deterioration [ 35 , 36 ]. With no available vaccine for Lyme disease, behaviours aimed at reducing skin exposure to ticks or detecting and removing ticks during a tick bite remain the most effective Lyme disease prevention strategies. These Lyme disease adaptation behaviours (LDAB) include wearing long, light-coloured clothes that cover the skin, applying insect repellent, and thoroughly checking one’s skin and clothes for ticks when returning from the outdoors, among others [ 37 – 44 ].
The first human case of Lyme disease in the province of Quebec was identified in 2008 [ 15 , 25 , 26 ]. The number of cases since then has markedly increased and is actively monitored by the Quebec Ministry of Health and Social Services [ 27 , 28 ]. The number of reported cases originating from within the province increased from five cases in 2011 to 650 cases in 2021 [ 29 ]. This increase can be partly explained by variations in temperature, rain, and humidity, which can have a direct impact on the reproduction rates and development of ticks, their biting activity, the density of their populations and their search for hosts [ 15 , 19 , 30 – 32 ]. The abundance and diversity of host species (e.g., deer, migratory birds, small rodents, and chipmunks) may also favor tick survival, as well as the increased suitability of habitats for ticks due to warmer winters associated with climate change [ 18 , 33 , 34 ].
Among the myriad risks to human health and wellbeing exacerbated by climate change [ 1 – 5 ] is an increase in vector-borne diseases [ 6 – 10 ]. In North America, Ixodes scapularis, a vector of Lyme disease’s causative agent, Borrelia burgdorferi, has spread northward from the United States into Canada due to climate and environmental changes [ 9 , 11 – 15 ]. As a result, the number of Lyme disease cases reported in the Canadian population has increased, from 144 cases in 2009 to 2,634 cases in 2019 [ 16 , 17 ]. Evidence suggests that in North America, the numbers, types, activity levels, and geographical distribution of I. scapularis are increasing due to the rising temperatures associated with climate change [ 18 – 20 ]. The magnitude of the impact of climate change on Lyme disease risk is likely to vary regionally [ 9 , 21 ]. Despite evidence of the increasing emergence and spread of Lyme disease, some studies show that relatively few Canadians understand the impacts of climate change on health and the expansion of zoonotic diseases [ 22 – 24 ].
The Tucker Lewis Index (TLI), the Root Mean Square Error of Approximation (RMSEA), and the Comparative Fit Index (CFI) were used in the SEM analysis to assess model fit. The TLI and the CFI indicate a good or excellent model fit when their value is greater than or equal to 0.90 or 0.95, respectively. A RMSEA value between 0.06 and 0.08 indicates an acceptable model fit, while a value lower than 0.06 indicates an excellent fit [ 89 , 90 ].
First, psychometric analyses were conducted to test the reliability of each scale. Item-total correlation was used to exclude items with low discrimination. The minimum threshold for inclusion based on item-total correlation was 0.2 [ 86 ]. Below this threshold, items were removed using a backward selection process; the item with the lowest value under 0.2 was removed first and the statistics were recalculated before removing any further items from the scale. Additionally, Cronbach’s alpha was calculated and reported for each resulting scale. An alpha value of 0.7 is widely considered an acceptable threshold for internal consistency of a scale and was therefore used here [ 87 ]. Second, structural equation modelling (SEM) with Mplus 8.3 [ 88 ] was performed to test the TPB model and identify which of the three constructs of the TPB (attitudes, social norms, and perceived behavioral control) significantly predicted the intention to adopt LDAB (Objective 1). Third, three multiple regression analyses were performed using Mplus 8.3 to identify behavioral, normative and control beliefs significantly related to attitudes, social norms, and perceived control, respectively (Objective 2).
Control beliefs. Participants’ beliefs about facilitating and impeding factors related to the adoption of LDAB were assessed by asking participants to rate each of five factors elicited in the pilot study in terms of its perceived expected occurrence, and the extent to which they would facilitate/hinder the adoption of the behaviour (see Table 3 ). These multiplicative items were then averaged to obtain an overall score for the control beliefs construct.
Behavioral beliefs. An expectancy-value index [ 58 ] was used to assess participants’ beliefs about Lyme disease prevention. Participants were asked to rate each of the five consequences elicited in the pilot study in terms of its perceived likelihood and the subjective value of each of these consequences (see Table 3 ). A behavioral belief score was obtained by multiplying the score for the likelihood of an outcome by the score of its desirability. These multiplicative items were then averaged to obtain an overall score of the behavioral beliefs construct.
Lyme disease adaptation behaviours. To measure this construct, we used a previously validated index of Lyme disease prevention behaviours in Québec [ 56 ]. Respondents indicated how often they adopt each of the 10 behaviours in the index (see Table 2 ). The first question was a “Yes” (1) or “No” (0) question, while the remainder (questions 2 to 10) were coded using a five-point scale with the following options: “Always”, “Often”, “Occasionally”, “Rarely”, and “Never”. The responses were then dichotomized. “Always” and “Often” were considered as LDAB adoption and assigned a value of (1), while “Occasionally”, “Rarely”, and “Never” were considered as a lack of LDAB adoption and assigned a value of (0). These values were then summed to obtain a measure of adaptive behaviour.
The data collection, from the preliminary focus groups to the main survey dissemination, was managed by the survey firm Léger. Participants were selected from the company’s pool of 225,000 panelists residing throughout the province of Quebec. To attain the desired regional quotas and sample size (2,000 participants), the survey firm sent invitations to a subset of 17,752 panelists. These were sent in waves rather than all at once to increase the likelihood of obtaining a more representative sample by soliciting responses throughout the data collection period. In total, 11 invitations waves were sent. First, invitations were sent to 8,945 people in 5 waves over 5 days (one by day). Second, 5,950 additional people were contacted over 4 days in 4 waves. Finally, 2,857 more people living in regions where quotas were not yet met were sent invitations in 2 waves. Before ending the data collection, a general reminder was sent to all those invited, but who had not yet accessed the survey, ensuring that each potential respondent had the same probability of being contacted during the survey period. The raw data ( S1 Data ) collected during the survey (and whose codes correspond with the questionnaire: S1 File ) can be accessed in the supporting information. The survey was accessible to participants on any day of the week, at any time, from any Internet-enabled device (computer, tablet, smart phone, etc.). The median completion time for the survey was 19 minutes and 43 seconds. Léger’s web panel point program, in which panelists can trade points earned by completing surveys for gift cards, was the main incentive for participation.
Table 1 shows that the proportion of men and women in our sample corresponds to that of the target population according to Statistics Canada census data [ 83 ]: χ 2 = 1.14, n.s., nil effect size. However, the sociodemographic characteristics differ from the target population in age, income, and education. The respondents were generally more educated (large effect size) and had a higher household income (large effect size). The proportions of respondents aged between 65 to 74 and older than 75 years old are lower than the proportion in the target population (small effect size) [ 84 , 85 ].
The target population for this study was individuals aged 18 and up living in the following 12 administrative regions of the Canadian province of Quebec (1,667,441 km 2 ; population 8.7 million): Chaudière-Appalaches, Capitale-Nationale, Centre-du-Québec, Estrie, Lanaudière, Laurentides, Laval, Mauricie, Montérégie, Montréal, Outaouais, and Saguenay-Lac-Saint-Jean. These regions were specifically targeted because they comprise municipalities where the public is at risk of acquiring Lyme disease [ 27 , 81 ]. The target sample size was 2,000 participants, with a quota of participants for each administrative region based on its proportion of the province’s total population, then corrected using Kish’s over-sampling method [ 82 ]. More information on the survey sample characteristics can be found in Table 1 .
Before conducting a TPB-based study, Ajzen [ 60 ] recommends performing a pilot study to determine the salient beliefs associated with a specific behaviour in a given population. It is crucial to conduct this preliminary study to avoid researchers making an arbitrary selection of these beliefs [ 60 ]. Thus, five focus groups of 4–5 persons (n = 21) recruited by the polling firm Léger were conducted in early 2021. Beliefs about LDAB were determined through an open-ended questionnaire. Specifically, participants were asked to discuss: (a) the benefits and disadvantages they would experience by adopting LDAB over the next year; (b) which people or groups of people would approve or disapprove of their adoption of LDAB in the next year; and (c) which factors might facilitate or impede their adoption of these behaviours in the next year. Quantitative measures of behavioral, normative, and control beliefs were subsequently elaborated and included in the survey questionnaire using the more frequently cited outcomes, referents, and control factors reported during the focus groups.
In addition, according to the coefficients of determination (R 2 ), the predictive variables of the TPB model account for 94.4% of people’s intentions to adopt LDAB ( Fig 2 ). While this exceeds the R 2 values commonly observed in other TPB studies, including transversal studies where intentions and behaviour are measured at the same timepoint, our large sample size should preclude issues related to overfitting [ 59 ]. In contrast, the predictive variables explain 28.3% of the variance in reported LDAB ( Fig 2 ). The model’s fit to the data was good to excellent: CFI = 0.93, TLI = 0.92 and RMSEA = 0.06.
The results in Table 4 also demonstrate significant correlations between all TPB variables. More precisely, intentions correlated better with predictive variables (attitudes, perceived social norms, and perceived control over behaviours) than with the LDAB index. Furthermore, behavioral, normative, and control beliefs are significantly related to their associated construct: attitudes (r = 0.47), perceived social norms (r = 0.63), and perceived behavioral control (r = 0.44). In other words, these results indicate that the behavioral, normative, and control beliefs can explain the three main determinants of behavioural intention according to the TPB, respectively [ 60 ].
Cronbach’s alphas, means, standard deviations, and item-total correlations (item discriminations) are presented in Table 3 . After examining the item-total correlation, three items were removed due to their failure to meet the 0.2 threshold: one behavioral belief (“Protecting yourself from ticks will be too unpleasant”; rho = 0.11), and two control beliefs (“You will experience discomfort while adopting measures and behaviours that can protect you from Lyme disease”; rho = -0.01; and “You will lack the items required to protect yourself from Lyme disease”; rho = 0.07). Following this step, the results of the psychometric analyses showed that all the alpha coefficients were adequate, albeit with two scales (perceived behavioral control and control beliefs) showing values slightly below 0.70 (see Table 3 ).
Discussion
The results showed that the population of the province of Quebec is still in the early stages of adopting preventive behaviours towards Lyme disease. Despite our respondents residing in high Lyme disease risk regions, they adopted on average 4.19 of the 10 behaviours included in the LDAB index in 2018 [56], while in 2021 the average was 4.24. These results are consistent with those from other studies in Canada [44, 52, 91] and in other countries where Lyme disease is endemic, including the United States [49], France [92] and the Netherlands [93]. Previous studies conducted in regions of the province of Quebec where Lyme disease incidence is the highest (Montérégie and Estrie regions) also showed that the adoption of preventive behaviours was low [52, 91]. These results suggest that more tailored messages and educational campaigns aiming to promote these preventive behaviours should be developed (e.g., public awareness campaigns, training for medical professionals, preparation of government documents). To design such efficient interventions, we need to better understand the psychosocial underpinnings of expert-recommended preventive behaviours, and the TPB’s behavioural, normative and control belief constructs provide a robust framework for this. Essentially, interventions should challenge people’s beliefs that are in opposition to the desired adaptive behaviours using information that engages their existing supportive beliefs, or that encourage them to form new ones in line with the desired behaviours [76, 94]. The current study sheds some light on this subject by identifying certain specific and actionable behavioral, normative, and control beliefs that underlie Quebecers’ more general attitudes, social norms and perceptions of behavioural control determining their intention to adopt LDAB.
The first objective of this study was to predict and explain individuals’ adaptive behaviour to Lyme disease using the full TPB model. According to the results, attitudes, perceived social norms, and perceived behavioral control accounted for a large proportion of the variance in intentions to adopt Lyme disease adaptive behaviours (LDAB). These intentions were also a good predictor of self-reported LDAB. Particularly noteworthy is the strong estimated effect of perceived behavioral control on intentions and presumably, behaviour (through intentions). This finding emphasizes the importance of removing potential barriers and creating conditions that will encourage the adoption of LDAB, which has also been highlighted by other studies [49, 91, 95]. To a lesser extent, perceived norms and attitudes are also linked to intentions and behaviour according to our model, and thus our results suggest that interventions must subjectively transmit accurate information that links adaptation to positive outcomes and social norms followed by certain referent individuals or groups who support adopting the adaptive behaviours or who have adopted them themselves. Though these findings on the importance of constructs such as perceived control on intentions to adapt to Lyme disease are significant in themselves, the theoretical and practical relevance of this study lie in our identification of the specific beliefs that should be targeted in interventions.
Examination of the specific control beliefs that predicted adaptation through overall perceptions of control revealed Quebec residents’ view that comprehending the measures and behaviours recommended by experts to protect themselves from Lyme disease and being familiar with high Lyme disease risk areas will facilitate adaptation. This finding is consistent with the results obtained by Beck et al. [49], who found that lack of awareness was a barrier to practicing prevention behaviours. Anecdotally, this finding is also possibly reflected in the beliefs and TPB constructs’ excellent ability to predict intentions to adopt LDAB in the future, but limited ability to explain variance in current LDAB patterns, as respondents simply may not have had prior knowledge of many of the behaviours at the time of the survey. Though one’s internal integration and understanding of communicated information is an important process in surmounting these knowledge barriers, there is an important external component to these control beliefs. Indeed, these barriers can largely be overcome with improved accessibility of information, both in terms of ease of availability and intelligibility. Therefore, campaigns aiming to favour LDAB adoption should emphasize educating the public on the behaviours and raising awareness on the locations of high-risk areas. Fortunately, quality resources on endemic areas and tick detection in Québec and Canada exist (e.g.,
https://www.hopkinslymetracker.org/explorer/); these should be further developed and awareness of them should be promoted.
The results also showed that four behavioural beliefs have significant associations with LDAB via attitudes and intentions. Three of these beliefs (i.e., adopting LDAB will: 1) avoid tick bites, 2) protect the health of one’s loves ones, and 3) reduce the risk of transmitting Lyme disease) were related to the same cognitive advantage: adapting to Lyme disease has a protective effect on health. This is not surprising, considering that Lyme disease can have severe and long-lasting consequences [35, 36]. This suggests that this health protection benefit is associated with respondents’ positive attitudes toward LDAB. The fourth behavioral belief, on the other hand, relates more to preserving one’s lifestyle and quality of life, as it cites the ability to continue to pursue outdoor activities as an advantage of LDAB. Thus, educational messages must also strengthen this last belief since the practice of outdoor activities is a lifetime objective for many people. This last result is consistent with the Theory of Reasoned Goal Pursuit (TRGP) [96], an extension of the TPB. This theory postulates that a person’s current set of goals reinforce their motivation to perform a specific behaviour [97]. In our case, a person could have a high intention to protect themselves against ticks and eventually adopt the proposed adaptive behaviours because they align with their current objective to practice outdoor activities. Our results show that the population already has a positive perception of outcomes following LDAB adoption (the behavioural beliefs’ average score is 3.31 on a scale ranging from -6.25 to +6.25) and a strong positive attitude toward the adoption of LDAB (mean attitude score of 5.26 out of 6), which bodes well for increasing adoption of LDAB if they become socially normalized and knowledge barriers to their adoption are removed.
The results also revealed that two normative beliefs are associated with LDAB via social norms and intentions. This finding demonstrates that the approval of groups of influence such as friends, family as well the municipal government is important for LDAB adoption. Therefore, to be effective, a norm-based intervention should focus on encouraging people to not only adopt LDAB themselves but also support their loved ones’ adaptation, and municipal authorities should be directly involved in LDAB advocacy and awareness raising in at-risk constituencies. This participation could both improve social normalization of LDAB and harness aforementioned levers to behaviour adoption (knowledge of behaviours and knowledge of at-risk areas). Some municipalities in Quebec have already taken steps in this direction. This is the case, for example, of the municipality of Bromont, which is located in a Lyme disease endemic region in Southern Quebec. This municipality launched a project in 2019 to raise awareness and engage citizens in adopting preventing behaviors in partnership with the Estrie Public Health Direction and the University of Montreal [98, 99]. On their website, users can find highly useful resources such as posters and videos about the behaviours to adopt, where on one’s body ticks tend to be located, and periods of the year they are active. Municipalities should also be involved in disseminating the aforementioned maps and interactive tools identifying areas at risk for ticks or developing their own.
The present research has significant strengths, such as a strong theoretical framework, a sample population that resembles the target population, a validated behaviour index, and the use of advanced statistical modelling techniques, but its limitations should also be considered when interpreting the findings. First, we describe LDAB as behaviours that can protect individuals from Lyme disease in our survey questions, which may introduce an element of social desirability and bias participants’ responses; this is a limitation common to TPB research on health behaviours [100] and has been deemed a minor issue [101]. Additionally, the observational cross-sectional design of this study precludes inferring causal relationships between variables. Though transversal TPB studies have repeatedly demonstrated success at predicting behaviour [59], a suggestion for future research would be to conduct a longitudinal study to test whether the beliefs significantly associated with LDAB in the present study can predict these behaviours over a period of one year, the target timeframe proposed in the survey questions. A second limitation involves the overall low response rate (11.3%, although typical for a large-scale study), which may have impacted the generalizability of results. Similarly, since the present study considered only adult respondents, inferences derived from its results are limited to that population. Future research should attempt to perform a similar study with adolescents and children, as well as their parents, as the latter are often making the decision for their children [54]. As a first step, we suggest that the LDAB index be adapted and validated for this young population. Although consistent with findings in other countries, another limitation is that the study’s sample of adaptive behaviours may not necessarily represent all possible adaptive behaviours. For instance, in contrast to Beck et al. [49], “Wear permethrin-treated clothing” was not an indicator in our LDAB index because it was not recommended as a behaviour to adopt in the province of Quebec at the time of data collection. Finally, as this study focused on identifying beliefs in Quebec’s high-risk regions only, we also suggest that future studies evaluate beliefs in regions at high as well as low Lyme disease risk areas in order to facilitate comparison and distinguish differences.
[END]
---
[1] Url:
https://journals.plos.org/climate/article?id=10.1371/journal.pclm.0000176
Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.
via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/
