In the United States, an estimated 300,000 Lyme disease cases occur annually (Nelson et al. 2015), and the disease is expanding geographically (Schwartz et al. 2017). Throughout eastern North America, the bacterium Borrelia burgdorferi, the causative agent of Lyme disease, is transmitted by the blacklegged tick Ixodes scapularis. Exposure to infected blacklegged ticks is thought to occur peri-domestically (Dister et al. 1997, Cromley et al. 1998, Connally et al. 2006, Eisen and Dolan 2016). In the United States, most Lyme disease cases occur in the Northeast and upper Midwest (Nelson et al. 2015). In one study of 70 people bitten by ticks in Westchester County, New York, 69% reported acquiring the tick in their yard (Falco and Fish 1988). On the basis of the expectation that exposure is primarily peri-domestic, there has been widespread interest in deploying methods for controlling ticks and avoiding exposures in residential yards (Ostfeld et al. 2006a).

Many residents of areas endemic for Lyme disease undertake one or more environmental controls against ticks. In a survey of residents in three Connecticut towns in 2004, 24%–44% used a chemical pesticide in their yards against ticks, 44%–52% removed brush or leaf litter as an environmental control against ticks, 17%–29% placed wood chip or gravel barriers intended to discourage tick incursions, and 12%–25% fenced their properties to keep out deer (Gould et al. 2008). Chemical and landscaping measures entail investments by homeowners; 30%–58% of Connecticut respondents reported being willing to pay over $100 ($134 in 2018 dollars) for yard tick control. In addition to these active controls, residents might affect tick abundance inadvertently by enhancing or diminishing habitat quality for ticks or their vertebrate hosts.

Public health officials recommend an array of property management measures to control abundance of ticks and risk of exposure to tick-borne infections. The advice includes measures expected to decrease habitat for hosts of ticks, decrease tick habitat, limit tick dispersal, and decrease human contact with ticks. Advice to reduce habitat for tick hosts includes recommendations about stone walls, wood piles, bird feeders, and fences. Stone walls are recommended to be sealed (Stafford 2004). P. leucopus use stone walls to orient and to take refuge (Barry and Francq 1980). In a study in Westchester County, New York, nymphal tick abundance on unsealed, remnant stone walls was similar to the abundance on nearby forest floor (Frank et al. 1998). Wood piles, which the U.S. Centers for Disease Control (CDC) advises to keep neat and in areas of the yard away from the home (Centers for Disease Control 2018b), have been associated with increased tick-borne disease risk in some studies (Smith et al. 2001, Jones et al. 2018) but not all (Connally et al. 2009). Bird feeders, which homeowners are advised to remove altogether or keep away from the house for small mammal and bird management (Stafford 2004), have been associated with increased risk of tick encounters (Mead et al. 2018) and increased tick-borne disease risk in some studies (Smith et al. 2001, Jones et al. 2018) but not all (Townsend et al. 2003). Deer-exclusion fencing reduced tick abundance in residential forest plots (Daniels et al. 1993, Stafford 1993). Fencing is a recommended action in yards (Centers for Disease Control 2018b), where it has been correlated with reduced Lyme disease risk in one study (Fritz et al. 1997, Connally et al. 2009) but not others (Ley et al. 1995, Perkins et al. 2006, Jones et al. 2018). Recommendations to reduce tick habitat include raking of leaf litter, removal of trash (Centers for Disease Control 2018b), and avoid planting Japanese barberry (Berberis thunbergii) (Stafford 2004). Removal of leaf litter reduced immature tick abundance (Schulze et al. 1995) but has not been found to correlate with a reduction in Lyme disease incidence (Klein et al. 1996, Connally et al. 2009). Higher abundance of I. scapularis on P. leucopus has been found in nonresidential areas with denser Japanese barberry (Williams et al. 2009). Removal of Japanese barberry, which buffered microclimatic conditions, reduced I. scapularis by 60% in nonresidential forested areas (Williams and Ward 2010).

To reduce tick dispersal, public health officials have advised people to establish mulch, gravel, or wood chip barriers between forest and lawn. Past studies have found that silt barriers reduced tick dispersal (Carroll and Schmidtmann 1996). Frequent mowing is also recommended. To reduce ticks being brought into the home by pets, residents have been encouraged to keep cats and dogs out of woods (Stafford 2004). The effects of pet ownership on human risk are mixed, with cat ownership associated with increased risk in some studies (Jones et al. 2018) but not all (Ley et al. 1995, Belongia et al. 1999), and dog ownership similarly increasing risk of tick encounters (Jones et al. 2018) or disease in some studies (Fritz et al. 1997) but not others (Finch et al. 2014). For some recommendations, such as those regarding pets, wood piles, and bird feeders, associations found between the feature and human risk could arise from effects on tick abundance, human exposure, or both.

Here we asked how larval tick abundance in residential yards relates to recommended nonchemical yard management measures as environmental tick controls. We catalogued the yard management measures recommended to reduce tick bites by the CDC and the State of Connecticut's Tick Management Handbook (Stafford 2004), which is cited by the CDC and other organizations across the United States (Harvard Health Letter 2010, UMaine Cooperative Extension 2018) and Canada (Public Health Ontario 2018). We assessed residential yards across Dutchess County, New York, for their use of strategies included in these recommendations. We do not know whether the features we documented were imposed by homeowners with the intention of controlling ticks; rather, we analyzed these features as consistent with recommended control measures. For some recommendations, such as those regarding pets, wood piles, and bird feeders, associations found between the feature and human risk could arise from effects on tick abundance, human exposure, or both. We evaluated whether the use of yard management recommendations predicted larval tick abundance. Abundance of larvae was assessed both by flagging for questing ticks and by visual counts of ticks on small mammals. Our results expand the evidence base for recommendations about how to reduce tick exposures and tick-borne disease risk in residential yards. We focused on larval ticks, because our sampling coincided with larval peak activity. While nymphal ticks are of chief concern for public health, assessing how yard features influence larval ticks is relevant to human exposure to nymphs, because larvae transition to nymphs the following year.

Materials and Methods

Results

Questing Larval Ticks

Trash (df = 328, t = 2.678, P = 0.008) and stone walls (df = 318, t = 2.237, P = 0.026) were each associated with increased abundance of questing larval ticks in the full model that included all property management features and random effects of neighborhood and property address (Table 1). We also observed higher abundance of questing larval ticks in forest samples (df = 195, t = 4.171, P < 0.0001). For other property management variables, we detected no significant effects on abundance of questing larvae ( Supp Figs. 1–3 [online only]). Summary statistics for each habitat (mean number of ticks, standard error, range, sample size, number of properties with nymphs) are available in  Supp Table 2 (online only) (questing ticks) and  Supp Table 3 (online only) (ticks on mice). Counts of properties containing each management feature are available in  Supp Table 4 (online only). Data are available, on questing ticks and ticks on P. leucopus, in the figshare repository (Fischhoff et al. 2018).

Table 1.

Effects of property management on questing tick abundance

In models that included fixed effects of one property management feature, forest area, and their interaction (with random effects of neighborhood and property address), there was again a positive effect on tick abundance of the presence of trash or stone walls (trash: df = 324, t = 2.932, P = 0.004; stone wall: df = 318, t = 3.327, P = 0.001). The amount of forest habitat in a yard was found to be an important moderating effect (interactive force) on the influence of property management features. Trash or stone walls were each a less positive predictor of larval tick abundance in yards with greater forest area (trash*forest: coefficient = –0.27, df = 323, t = –2.098, P = 0.037; stone wall*forest: coefficient = –0.43, df = 358, t = –2.719, P = 0.007; Fig. 1). In the forest*trash model, we did not find a significant effect of forest area: coefficient = 0.08, df = 173, t = 1.144, P = 0.25. Neither was there a significant effect of forest in the stone wall*forest model: coefficient = 0.07, df = 173, t = 1.060, P = 0.290.

Discussion

Residential yards are thought to be the primary location at which exposure to blacklegged ticks occurs (Dister et al. 1997, Cromley et al. 1998, Connally et al. 2006, Eisen and Dolan 2016). In some communities endemic for Lyme disease, majorities of residents surveyed used nonchemical measures to control ticks in their yards, and reported willingness to pay over $100 for tick control (Gould et al. 2008). Messages from public agencies influence residents' decisions about whether and how to manage their yards for tick control. We found that two widely used sources, the CDC and Connecticut Tick Management Handbook, together included 22 recommendations for nonchemical yard-level tick control. Given the costs of managing yards to reduce ticks, and the potential benefits of reducing ticks, it is important for public health authorities' messages and homeowners' decisions to be grounded in evidence about how yard management relates to tick abundance. Here we evaluated the relationship between adherence to a subset of yard management recommendations and the abundance of larval ticks, both questing and on small mammals, in residential yards. The yards are in neighborhoods in Dutchess County, New York, which ranks among the highest counties in the United States in Lyme disease incidence (Centers for Disease Control 2018c).

The presence of trash or stone walls each predicted higher abundance of questing larval ticks. Trash and stone walls may increase tick abundance by providing improved habitat for small mammals. These effects are consistent with prior research identifying positive associations between P. leucopus and stone walls and logs (Barry and Francq 1980). Natural debris provides P. leucopus with cover protection from predators in urban environments (Persons and Eason 2017), and trash benefits rodents in urban settings (Johnson et al. 2016). The effects of trash and stone walls on tick abundance were each reduced in yards with greater forest area. In yards with less forest, trash and stone walls may each substitute for the refugia that greater forest area would otherwise provide for P. leucopus and other small mammals. With increasing forest area, the benefits to small mammals of trash or stone walls may decline relative to the increasing benefits of greater forest area. The benefits to residents of removing trash or stone walls in reducing risk are greater in yards with little or no forest.

Despite the increased questing larval tick abundance in yards with stone walls and trash, we saw no effect of these features on larval abundance on white-footed mice captured in yards. One possible explanation is that abundance of larvae on mice, which may range across property boundaries, is subject to greater influence by features of neighboring yards. Fewer larvae were found per white-footed mouse in yards where at least 75% of the area is enclosed by fence, compared with yards with less or none of the property fenced. However, no effect was detected of the interaction between fencing and forest area. Thus, the effect of fencing appears to be independent of forest area.

Fig. 1.

Questing larval tick abundance in relation to forest area, and the presence (1, light gray triangles) or absence (0, dark gray circles) of A) stone walls and B) trash in residential yards. Each line is based on a linear model fitted to the data; the gray bounds are 95% confidence intervals for predictions from the fitted model. To plot on log scale for greater ease of interpretation, 0.00001 has been added to each questing larval tick value. The data have been horizontally and vertically ‘jittered’ to reduce overlap of points.

Table 2.

Effects of property management on total larval ticks observed on all mice (per habitat and property)

We conclude that there is partial support for public health agency recommendations regarding trash, stone walls, and fencing. The potential benefits of each of these recommended measures needs to be weighed against each measure's costs, both financial and in terms of the value of yards as wildlife habitat, value which may be reduced by fencing. We detected no effects of ground barriers (gravel, mulch, or woodchip), bird feeders, wood piles, or indoor/outdoor cats. In the case of barriers, only two properties had this feature, resulting in low power to detect an effect while also reflecting low adoption rates of this recommendation. Property management features evaluated had inconsistent effects between tick sampling measures. The different patterns for questing larvae versus larvae on mice, and for total versus per capita larvae on mice, indicate that other factors, in addition to the property management features measured here, affect variation in abundance of ticks and their interactions with the wildlife host community.

Experimental manipulations would be necessary to establish causal relationships between adherence to tick control recommendations and larval tick abundance. Nevertheless, our observational study accounted statistically for sources of variation that would be controlled for directly in an experimental design. During brief visits to yards we were unable to assess some recommended environmental tick control measures, such as mowing, that require longer-term monitoring of yards or surveys. Given the time available to make observations, it was necessary to define as present or absent some property features that were described by public health authorities as continuous aspects of yard management ( Supp Table 1 [online only]). For example, homeowners are advised to place bird feeders in the yard away from the home; we recorded bird feeders as present or absent rather than their distance from the home. Yard features that we enumerated as binary contain heterogeneity, for example in the size or neatness of a firewood pile, or the length of a stone wall, that may have contributed variation to the results. Variation in results may also have arisen due to chemical treatments in prior years, about which we do not have information. We do not expect variation in prior year pesticide use, or spot spraying in the studied years, to have directionally influenced results with respect to any yard features.

Our study focused on larval ticks, due to sampling at the time of larval peak activity. Public health agencies are chiefly concerned with reducing exposure to nymphal ticks; larvae have not yet acquired the most common pathogens. Assessing the effects of yard management on larval ticks is relevant to controlling risk from nymphs, given that larvae transition to nymphs the following year. Moreover, the larval ticks attached to white-footed mice are more likely to survive and to become infected with tick-borne pathogens than are larvae that attach to other hosts (Keesing et al. 2009). It is possible, given the potential for hosts to transport larvae, and the potential for yard management to affect survival to the nymphal stage, that the same set of yard management features examined here may have different effects on nymphal ticks.

Table 3.

Effects of property management on per capita larval tick abundance on white-footed mice Peromyscus leucopus

Fig. 2.

Bar plot (with standard errors) of abundance of larval ticks on white-footed mice in relation to habitat and fencing. Fencing was considered present in properties with at least 75% of the yard enclosed in fencing, absent in yards with <75% fenced.

In addition to tick abundance, an important source of variation in risk for residents is human behavior. Our study addressed yard management measures expected to affect tick abundance, but other recommended yard management actions (e.g., placement of yard equipment away from forest edge) are based on expectations about how human behavior affects exposure to the ticks that are present.

Our study presents results on eight public health agency recommendations across a county, in 143 yards subject to the management decisions of residents. In contrast, in previous studies on the property management features examined here, data were typically collected in a more limited number of residential (Stafford 1993) or nonresidential locations (Daniels et al. 1993), on fewer interventions (Frank et al. 1998, Townsend et al. 2003, Mead et al. 2018), or relied on self-reporting about property features (Smith et al. 2001, Connally et al. 2009, Jones et al. 2018, Mead et al. 2018). Prior research often involved experimental manipulation of a single intervention (Stafford 1993, Schulze et al. 1995, Carroll and Schmidtmann 1996). The effects of certain interventions, for example trash removal, do not appear to have been previously studied in the context of tick control. To inform the decisions of residents, we suggest that when making property management recommendations, public health officials provide information on the scientific support and uncertainty about each recommendation.

Ethics statement

The authors declare no conflicts of interest. The Institutional Review Board of Cary Institute of Ecosystem Studies (CIES) approved the survey that collected data on participating households, including presence of indoor/outdoor cats. The CIES Institutional Animal Care and Use Committee approved the protocols for live-capturing small mammals.

Supplementary Data

Supplementary data are available at Journal of Medical Entomology online.