Associations between canine hookworm infection and dog owners' awareness, perception, and behaviour: A cross‐sectional study in Brisbane, Queensland, 2019–2020

Canine hookworms are found globally, and infections have been recorded in domestic dogs visiting dog parks in Queensland, Australia. Some canine hookworms also present a risk of zoonotic transmission to humans. Potential transmission of hookworms can occur in the household and at public places because of lack of owner awareness and poor coverage of canine deworming. Between April 2019 and March 2020, faecal samples from owned dogs were collected from 39 dog parks in metropolitan Brisbane, Queensland, and tested for the presence of hookworm eggs using faecal floatation technique. Dog owners who provided samples were requested to complete a survey on their awareness, perceptions and behaviour regarding the risks and consequences of canine parasitic infection. Associations between dog owner demographics and responses to the survey were measured using Goodman and Kruskal's gamma. Statistical associations between canine hookworm infection in dogs and their owner's responses were quantified using a two‐sample Wilcoxon rank‐sum test. A total of five canine faecal samples from 175 respondents were found positive for hookworm eggs, corresponding to a sample‐ and park‐level prevalence of 2.86% and 12.82%, respectively. Female dog owners were found to be more aware of the importance of regular deworming of their dogs (|γ| = 0.405). Our results indicate that hookworm infection was associated with a lower awareness of the importance of deworming (p = 0.007), less diligence in administration of deworming (p = 0.004), lower awareness of the risk of acquiring parasites from raw meat (p = 0.010), less likelihood of cooking meat before feeding it to their dogs (p = 0.028), and less likelihood to properly dispose their dog's faeces (p = 0.027). This study not only indicates a need for improving owner education towards the importance of deworming but also a need for changing owner's behaviours to reduce potential environmental contamination with infective hookworms. The latter in particular indicates a broader public health risk of disease transmission in public places due to improper disposal of dog faeces, especially in areas with higher canine hookworm infections.


| INTRODUC TI ON
Dogs are the most common pet in Australia, with an estimated pet dogs population of 4.2 million (Animal Health Alliance (Australia) Ltd, 2013). Companion animals such as dogs improve the mental health and physical well-being of humans; however, they also present a potential risk for human transmission of zoonotic parasites such as hookworms (Hawdon & Wise, 2021;Traversa, 2012). Hookworms are one of the most commonly encountered gastrointestinal parasites in canine companion animals in Australia (Palmer et al., 2007).
The species of canine hookworm most frequently identified in Australia include Ancylostoma caninum, A. ceylanicum, A. braziliense and Uncinaria stenocephala. Reports have found A. caninum clustering largely in tropical, sub-tropical and some temperate regions, U. stenocephala in temperate conditions, and A. braziliense mostly in tropical regions of Australia Palmer et al., 2007). With the exception of one study in South Africa that reported A. caninum completing its lifecycle in humans (Ngcamphalala et al., 2020), A. caninum and A. braziliense are generally not known to reach maturity in humans, however, their larval stages can cause CLM (Brenner & Patel, 2003). U. stenocephala has also been implicated in CLM (Hawdon & Wise, 2021). In comparison, A. ceylanicum can mature in human hosts, and readily infect humans without causing CLM. Of these species of hookworm, A. braziliense causes the most severe cases of CLM, whereas A. caninum, A. ceylanicum and U. stenocephala tend to cause more localized and short-lasting cases of CLM Hawdon & Wise, 2021). Under favourable environmental conditions, A. caninum larvae can survive and remain infective for up to 3-4 weeks in the environment (Hawdon & Wise, 2021). Infection in dogs and humans mostly occurs percutaneously or orally by infective L3 hookworm larvae contaminating soil, sand or grass in parks, playgrounds and yards. Dogs can also become infected by physical contact with contaminated faeces (Hawdon & Wise, 2021;Massetti et al., 2020).
Administration of anthelmintics is a principal means for treatment and prevention of hookworms in dogs (Lynn & Duquette, 2021) and dogs depend solely on their owners when it comes to control and treatment against parasites. Thus, awareness, perception and behaviours of dog owners towards canine parasites and potential zoonoses are essential factors with respect to regular and timely treatment of dogs against parasites (Harriott et al., 2019;Nguyen et al., 2021). Despite the availability of various dewormers off the shelf from chemist stores and pet shops, the application of these treatments has been hampered by lack of owner education regarding the correct usage of anthelmintic drugs and parasite transmission risk factors (Matos et al., 2015).
Furthermore, a study in Australia found that few pet owners considered regular deworming and proper disposal and cleaning up of pet's faeces as disease control measures (Steele & Mor, 2015). Proper disposal of dog faeces by owners in public areas can reduce environmental contamination with hookworm eggs and potential transmission of the parasite (Harriott et al., 2019), however a survey of Australian dog owners indicated that almost half of the respondents did not clean up their dog's faeces in public (Fawcett, 2013). A recent nationwide study of canine soil-transmitted helminths (cSTH) in dog faecal samples collected 1581 faecal samples from 190 dog parks across Australia and reported a nation-level true prevalence of hookworms of 9.1%, with A. caninum the most prevalent hookworm species (8.4%). The study also found an association between parks being contaminated with cSTH and lower social economic index of the postcode containing the park. Furthermore, A. caninum had the highest apparent prevalence of 44.5% in faecal samples collected from tropical areas, followed by 4.8% in samples from sub-tropical areas. Faecal samples contaminated with hookworms were detected in eight of 15 visited public parks in Brisbane (53.3%; 95% confidence interval: 30.1-75.2; Massetti et al., 2022). This reported prevalence of 9.1% is in spite of recent research having indicated that the majority of Australian dog owners are aware of the risks of hookworm and other parasitic infections in their pets, and administer anthelminthic products as prophylaxis Nguyen et al., 2021;Palmer et al., 2007).

K E Y W O R D S
Australia, canine hookworm, dog parks, owner's perceptions, zoonosis Impacts • Hookworm eggs were found in 2.86% of dog faecal samples, across 12.82% of surveyed dog parks in Brisbane.
• Female dog owners were found to be more aware of the importance of regular deworming of their dogs and younger dog owners and dog owners with a higher level of education better understood the need for veterinary advice on parasite control and prevention.
• Hookworm-infected dogs in metropolitan Brisbane dog parks were more likely to be owned by individuals with poor awareness of the importance of deworming and owners who do not properly dispose of their dog's faeces.
Studies targeting Australian pet owners have shown that awareness and perceptions of zoonotic risks vary demographically (Nguyen et al., 2021;Palmer et al., 2010). A recent study looking at the role of perceptions of dog owners towards canine gastrointestinal (GI) parasitism using a modified health belief model found that owners' perceptions were significantly associated with owners' sex and duration of dog ownership, with female owners and owners who had owned their dogs longer showing greater awareness of canine GI parasitism (Nguyen et al., 2021). Furthermore, the study reported a statistically significant and positive association between the frequency of visits to veterinary clinics and owner compliance towards deworming their dogs, suggesting that veterinarians may influence dog owners' perceptions and awareness towards canine parasites (Nguyen et al., 2021).
While previous studies have presented indirect evidence that owner awareness and perceptions about pet parasites does influence the likelihood of pet owners deworming their pets, these studies have not explored the possible associations between the presence of parasite infections in dogs and their owner's awareness and perception (Nguyen et al., 2021;Palmer et al., 2010). Such associations are important to measure in order to understand the role of owner education towards parasite control in pets and how it may reduce the potential risks of zoonoses.
This study aimed at quantifying associations between dog owners' demographics and their awareness, perception and behaviour regarding risks of parasite infection in their pet dogs in Metropolitan Brisbane, in the state of Queensland, Australia; and measuring associations between the presence of hookworm infections in pet dogs and their owners' awareness, perception and behaviour regarding risks of parasite infection in dogs.

| Dog faecal sample collection
Fresh faecal samples were collected from each dog after observing natural defecation and receiving consent from the dog owner. Faecal samples were collected by the volunteers from the ground using a wooden spatula and stored in leakproof 70 mL sterile plastic faecal collection pots (Livingstone International Pty. Limited). Each faecal collection pot was labelled with an individual identification number, which was matched with the identification number for the corresponding dog owner's response to the online survey. Faecal pots were stored and transported in containers that were kept at 4°C. Volunteers

| Diagnostic procedures for parasitic examination
Faecal samples were tested for the presence of hookworm eggs by faecal floatation technique using faecal diagnostic kits (Henry Schein®) and a saturated salt solution (SG 1.2). Faecal floatation technique was chosen over other diagnostic assays such as qPCR due to limited available funding for testing. To minimize the possibility of false-negative test results, three sections of each faecal sample were independently tested. Prepared faecal samples were examined under a compound microscope (Nikon® Eclipse E100) and hookworm eggs were identified using morphological keys (Zajac & Conboy, 2012). The test results were recorded as positive or negative for hookworm eggs.
The 95% confidence interval for the prevalence of hookworm eggs in tested faecal samples was calculated in R (R Core Team, 2022) by using a binomial regression with the following code: plogis(confint(glm(cbind(n _ infect, n _ uninfect) &#x007E; 1, data = data.frame(n _ infect = X, n _ uninfect = Y), family = 'binomial'))) where X represents the number of samples where hookworm eggs were identified and Y represents the number of samples where hookworm eggs were not identified.

| Dog owner survey
Dog owners who provided faecal samples of their dogs were invited to complete an online survey targeting their awareness, perception and behaviour towards canine parasitic infection. During the survey, brochures with details of accessing the survey were distributed by volunteers to dog owners within the parks visited. The survey questions asked for dog owners' demographics, their awareness and perception of the risks associated with zoonotic parasites and their behaviours for preventing parasitic infection in their dogs. The demographics recorded were age, sex, highest level of education or qualification achieved, and length of dog ownership (Appendix S1). Some survey responses were incomplete and therefore only complete and valid responses were included in the analysis. The analysis in the current study made use of dog owners' responses to 23 statements that captured different aspects of awareness, perception and behaviour towards parasite infection, with response options structured using a Likert-format five-level scale (1 = Strongly disagree, 2 = Disagree, 3 = No opinion/Unsure, 4 = Agree, 5 = Strongly agree). The statements were grouped into five categories, defined in the previous study: perceived susceptibility, perceived seriousness, perceived benefit, perceived barriers and actions (Nguyen et al., 2021). These were used for grouping the statements into larger conceptual categories for presenting results. The survey statements and their corresponding categories can be found in Table S1. Additional details of the survey design and its implementation are described in the study by Nguyen et al. (2021).

| Associations between dog owners' demographics and their survey responses
Pairwise statistical associations between the demographics of dog owners who participated in the survey and their responses to the survey were quantified using Goodman and Kruskal's Gamma (Goodman & Kruskal, 1954). Goodman and Kruskal's Gamma was used to test the null hypothesis that there is no difference in each survey response between each demographic (γ = 0). The measured γ value can range from −1 to 1, where −1 indicates a perfect negative association between the variables, 1 indicates a perfect positive association between the variables, and zero indicates no association between the variables. Goodman and Kruskal's Gamma results with a p-value of <0.05 were considered statistically significant. Additional details of this analysis are in Appendix S2.

| Associations between canine hookworm infection in dogs and their owners' survey responses
Statistical associations between the presence of canine hookworm eggs in dogs' faecal samples and their owner's responses were quantified using the two-sample Wilcoxon rank-sum test (Mangiafico, 2016). This test was run independently for the responses to all 23 Likert-format statements in the survey. It was performed to test the null hypothesis that the average survey scores would not significantly differ between owners of dogs whose dog's faecal sample tested positive for hookworm (hereon referred to as owners with hookworm-positive dogs) and owners whose dog faecal sample tested negative for hookworm (hereon referred to as owners with hookworm negative dogs). A greater average score represents stronger agreement with the statement in the survey. A p-value of p < 0.05 was considered to be statistically significant. The statistical effects of the responses to each statement were interpreted independently to identify which statements within each survey category produced statistically significant differences in owner's responses between the two groups of owners. Additional details of this analysis are in Appendix S3.

| Sample-and park-level prevalence of canine hookworm infection
A total of 175 dog owners provided both a faecal sample of their dog as well as completed the online survey. Of the 175 dog faecal samples provided, five samples tested positive for the presence of hookworm eggs (2.86%), with an estimated 95% confidence interval of 1.03%-6.04%. All five of these samples were detected in different dog parks, indicating environmental contamination at 12.82% of parks visited (5/39 parks), and were collected in different months (May and August 2019 and January, February and March 2020). There was no sign of spatial clustering in the locations of these five dog parks.

| Associations between dog owners' demographics and their survey responses
The most common age group of dog owners was 26-45 years (54.6%), followed by 46-60 years (24.1%). A majority of the dog owners were female (67.82%). More than two-thirds of dog owners had a university or graduate-level qualification (68.42%).
Approximately two out of five dog owners had owned their dog for more than 10 years (40.80%; Table 1). There were variations in the responses to all 23 survey statements among each dog owner demographic (Figures S1-S23).
In the perceived susceptibility category, a statistically significant and negative association was found between age and the statement that Access to wildlife or visiting places with wildlife increases a dog's risk of acquiring parasites (γ = −0.257; CI: −0.444, −0.071). A statistically significant and positive association was found between the highest level of education or qualification achieved and the statement Keeping pets indoors would help prevent them from acquiring parasites (γ = 0.164; CI: 0.023, 0.306; Table 2).
In the perceived seriousness category, statistically significant and negative associations were found between sex and five of the eight statements: I am aware of the parasites which affect dogs In the perceived benefits category, a statistically significant and negative association was found between age and the statement In the perceived barriers category, no statistically significant associations were found between any of the owner demographics and any of the four survey statements.
In the actions category, a statistically significant and positive association was found between highest level of education or qualification achieved and the statement I treat my dog(s) regularly for worms (γ = 0.267; CI: 0.080, 0.454). A statistically significant and negative association was found between the length of dog ownership and the statement I always make sure to cook meat before feeding it to dogs (γ = −0.223; CI: −0.377, −0.069).

| Associations between canine hookworm infection in dogs and their owners' survey responses
The responses to seven statements over four of the five survey statement categories were found to be significantly different between owners with hookworm-positive dogs and owners with hookwormnegative dogs ( Table 3). For all statements, there were differences in average response between these two groups ( Table 4 and In the perceived susceptibility category, average survey scores did not significantly differ between owners with hookworm-positive dogs and owners with hookworm-negative dogs for all four of the statements. In the perceived seriousness category, the average score given by owners with hookworm-negative dogs was significantly higher than those from owners with hookworm-positive dogs for two of the eight statements: I am aware of the parasites which affect dogs (p = 0.026) and I am aware that raw meat may contain parasites which can infect dogs (p = 0.010).
In the perceived benefits category, the average score given by owners with hookworm-negative dogs was significantly higher than those from owners with hookworm-positive dogs for the statement I am aware that without regular deworming, dogs would be at risk of acquiring parasite infection (p = 0.007).
In the perceived barriers category, the average score given by owners with hookworm-negative dogs was significantly lower than those from owners with hookworm-positive dogs for the statement It's hard to remember to give treatment to dogs at the recommended frequency (p = 0.011).
In the actions category, the average score given by owners with hookworm-negative dogs was significantly higher than those from owners with hookworm-positive dogs for all three statements: I always make sure dog poo is disposed properly (p = 0.027), I always make sure to cook the meat before feeding it to dogs (p = 0.028), and I treat my dog(s) regularly for worms (p = 0.004).

| DISCUSS ION
Previous studies in Australia have focused on either prevalence of canine hookworm infection in dogs (Blake & Overend, 1982;Cornack & O'Rourke, 1991;e Silva et al., 2006;Jenkins & Andrew, 1993;Palmer et al., 2007Palmer et al., , 2008Setasuban & Waddell, 1973;Traub et al., 2007) and humans (Burry, 1978;Croese, 1995;Croese et al., 1994;Hawdon & Wise, 2021;Koehler et al., 2013;Loukas et al., 1992;Prociv et al., 1994;Prociv & Croese, 1990) or dog owners' perceptions and awareness relating to risks of zoonotic parasitic infections (Nguyen et al., 2021;Palmer et al., 2010). A more recent study highlighted the current zoonotic risk of canine hookworms in Australian dog parks associated with the ecological characteristics of the geographical areas where sampled parks were located . In this study, hookworms (A. caninum as the dominant species) were detected in more than half of the parks sampled in Brisbane. However, none of  The average level of dog owner agreement was compared between dog owners whose dog faecal sample tested negative and dog owners whose dog faecal sample tested positive. Higher agreement means that the average response to the Likert statement of dog owners whose dog faecal sample tested negative was expected to be higher than that of dog owners whose dog faecal sample tested positive. Similarly, the lower agreement means that the average response to the Likert statement of dog owners whose dog faecal sample tested negative was expected to be lower than that of dog owners whose dog faecal sample tested positive. A recent study on prevalence of canine soil transmitted helminths also reported the presence of canine hookworms in Brisbane. The study reported a hookworm prevalence of 12.0%, with 14 of 117 faecal samples tested using qPCR or microscopy found to be positive for hookworms (Massetti, 2022). These hookworm positive faecal samples were collected in 8 of 15 parks (53.3%) visited within Brisbane . Compared to this study, the present study col-

| Differences in awareness and behaviour regarding risk of canine parasitic infections between different dog owner demographics
Our findings indicated that female dog owners showed greater awareness in the perceived seriousness category compared to male dog owners, supporting the findings of a previous study which applied a different analytical approach (Nguyen et al., 2021).

| Dog hookworm infection status and dog owners' perceived benefits, barriers and actions regarding deworming
Our findings indicate that dog owners whose dog's faecal sample tested positive for the presence of hookworm eggs have a lower awareness of the benefits of deworming and perceive the barriers to deworming to be greater than dog owners whose dog's faecal sample tested negative. These associations may explain the difference in their responses to the actions category statement regarding regular administration of deworming medication. Previous studies have also found an increased risk of hookworm infection in dogs that have not been regularly dewormed (Bowman, 2021;Massetti et al., 2020;Palmer et al., 2007Palmer et al., , 2008. This suggests that improved education of dog owners on the importance of deworming may positively influence dog owners' behaviours and reduce the risk of dog hookworm infections (Palmer et al., 2008). A study employing a modified health belief model on a superset of the data used for the current study reported a statistically significant association between increased frequency in veterinary clinic visitation by dog owners and regular deworming of their dogs, suggesting that these owners may be receiving additional education on prevention and control of canine parasitic infections from their veterinarians (Nguyen et al., 2021).

| Dog hookworm infection status and dog owners' perceived seriousness and actions regarding dog faeces disposal
This analysis demonstrated that dog owners whose dogs were positive for hookworms had a lower awareness of parasites which affect dogs than dog owners whose dogs tested negative. This indicates that poor perception of the seriousness of parasitic infections may be influencing owners' actions towards proper disposal of their dog's faeces. Previous studies have found an increased risk of hookworm infection in dogs when dog faeces have not been disposed timely and properly, helping development of infective hookworm larvae and environmental contamination (Bowman, 2021;Palmer et al., 2007Palmer et al., , 2008. This is particularly important in tropical climate areas such as Brisbane where environmental conditions are ideal for development and survival of hookworm during their infective stages (Bowman, 2021).
Environmental contamination with infective hookworm larvae is also a zoonotic risk to humans and can cause conditions such as cutaneous larva migrans (Burry, 1978;Gillespie & Bradbury, 2017;Lord, 1997) and one hookworm species, A. ceylanicum, can cause patent infections in humans .

| Dog hookworm infection status and dog owners' perceived seriousness and actions regarding the risk of parasites in raw meat
While hookworms are not known to be transmitted through the consumption of raw meat, there are other canine nematode parasites such as Toxocara spp. and Trichinella spp., and tapeworms such as Taenia spp. and Echinococcus spp., which can be present in raw meat of an infected animal (van Bree et al., 2018). Interestingly, the results of our analysis suggest that the dog owners whose dog's faecal sample tested negative for the presence of hookworm eggs have a greater awareness of the seriousness of the risk of acquiring parasites from raw meat than dog owners whose dog's faecal sample tested positive.
This difference is further illustrated in their actions, where dog owners with better awareness of parasites were also more likely to cook meat before feeding it to their dogs. This suggests that there may be a greater overall awareness of the risk factors of different parasites including hookworms among the dog owners whose dog's faecal sample tested negative for the presence of hookworm eggs.
A previous study, based on a superset of the data used in the current study, reported a statistically significant association between three or more veterinary clinic visitations per year by dog owners and regular cooking of meat before feeding it to their dogs (Nguyen et al., 2021). In contrast, a survey of dog owners in Italy reported that 9% of dog owners were recommended a raw meat-based diet for their dogs by veterinarians (Morelli et al., 2019). This suggests that while some dog owners may cook meat before feeding it to their dogs independent of veterinary advice, veterinary practices play an important role in providing additional advice to dog owners on the health risks associated with the consumption of raw meat by dogs.

| Limitations
The results of this study should be interpreted in light of the following While the two-sample Wilcoxon rank-sum test was able to identify statistically significant differences in perceptions between dog owners whose dog's faecal sample has tested positive for the presence of hookworm eggs and dog owners whose dog's faecal sample has tested negative for the presence of hookworm eggs using a small sample size, this approach is unable to identify causal effects because some of the survey responses are likely correlated. Furthermore, we were not able to fit a unified multivariable model to our two-sample Wilcoxon ranksum test results. In order to address any biases that may have been present in the results of our univariable two-sample Wilcoxon ranksum test results, future studies could make use of directed acyclic graphs (DAGs) to investigate the possible causal relations between variables and identify potential confounders (Greenland et al., 1999).
Appropriate multivariable models could then be created to address causal assumptions presented in the DAGs (Tennant et al., 2020).
Despite these limitations, the present study provides useful and novel information. The study indicates that prevalence of hookworm infections in dog faeces in Brisbane dog parks is low but not absent.
Moreover, the study showed that dog faecal samples collected from five different public dog parks in Brisbane were positive for hookworm infections, which is concerning considering the widespread availability of affordable preventatives. The study also found some variation in probability of canine hookworm infection when stratifying by owner demographics and survey responses. Whether these estimates can generalize to the broader pet owner population is unclear, but our work highlights this as an important question that warrants further study.

| CON CLUS IONS
Our study identified that persistent canine hookworm infections in Brisbane dog parks may be associated with dog owners' behaviours relating to deworming of their dogs and properly disposing of their dog's faeces, as well as with their risk perception. Combined with the previously reported zoonotic risk of exposure to A. caninum to humans that frequent these parks, these findings may in turn be used to provide better advice to dog owners about the risks of canine hookworm infection in dogs and humans.

ACK N O WLE D G E M ENTS
We thank all the volunteers involved in dog faecal sample collection and distribution of survey brochures. We would also like to thank the sur-

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy reasons.

E TH I C A L A PPROVA L
This study was carried out with the approval of the University of