Children's mouthing and food-handling activities were measured during a study of nondietary ingestion of pesticides in a south Texas community. Mouthing data on 52 children, ranging in age from 7 to 53 months, were collected using questionnaires and videotaping. Data on children's play and hand-washing habits were also collected. Children were grouped into four age categories: infants (7–12 months), 1-year-olds (13–24 months), 2-year-olds (25–36 months) and preschoolers (37–53 months). The frequency and type of events prompting hand washing did not vary by age category except for hand washing after using the bathroom; this increased with increasing age category. Reported contact with grass and dirt also increased with increasing age category. The median hourly hand-to-mouth frequency for the four age groups ranged from 9.9 to 19.4, with 2-year-olds having the lowest frequency and preschoolers having the highest. The median hourly object to mouth frequency ranged from 5.5 to 18.1 across the four age categories; the frequency decreased as age increased (adjusted R2=0.179; P=0.003). The median hourly hand-to-food frequency for the four age groups ranged from 10.0 to 16.1, with the highest frequency being observed in the 1-year-olds. Hand-to-mouth frequency was associated with food contact frequency, particularly for children over 12 months of age (adjusted R2=0.291; P=0.002). The frequency and duration of hand-to-mouth, object-to-mouth and food-handling behaviors were all greater indoors than outdoors. Infants were more likely to remain indoors than children in other age groups. The time children spent playing on the floor decreased with increasing age (adjusted R2=0.096; P=0.031). Parental assessment was correlated with hand-to-mouth activity but not with object-to-mouth activity. The highest combined (hand and object) mouthing rates were observed among infants, suggesting that this age group has the greatest potential for exposure to environmental toxins.
Studies have found that some children are exposed to pesticides at levels equal to, or greater than, adult levels (Loewenherz et al., 1997; Fenske et al., 2000; Lu et al., 2000, 2001; Shalat et al., 2003). Children's activities are considered to be important determinants of their actual dose of environmental contaminants (Cohen Hubal et al., 2000). Activities such as mouthing and food handling have been shown to impact children's blood lead levels (Charney et al., 1980; Bellinger et al., 1986; Freeman et al., 1997, 2001a). While algorithms have been developed to model children's exposure to environmental contaminants, children's microactivities such as mouthing and food handling behavior, are not well documented (Cohen Hubal et al., 2000).
Mouthing behavior generally includes all activities in which objects or hands are inserted into or touched by the mouth with the exception of eating or drinking (Groot et al., 1998). Mouthing has been studied using both direct observation and videotaping methodology (Zartarian et al., 1995, 1997a, 1997b, 1998; Groot et al., 1998; Reed et al., 1999; Juberg et al., 2001; Freeman et al., 2001b; Tulve et al., 2002). To date, only direct observation has been employed on children less than 1-year old (Groot et al., 1998; Juberg et al., 2001; Tulve et al., 2002). Direct observation may be done either by a parent or a trained observer in the child's home. Videotaping methodology has focused on older (≥2 years) children (Zartarian et al., 1995, 1997a, 1997b; Reed et al., 1999; Freeman et al., 2001b). This method employs a trained technician to record the child's behavior on videotape. The tape is analyzed at a later date and the results recorded either by manual translation or computer software.
It has been suggested that children's handling of foods (hand-to-food, food-to-surface and hand-to-surface contacts) may contribute 20–80% of their dietary exposure (Akland et al., 2000). Food-handling data have been collected using questionnaires (Freeman et al., 1997, 2001a, 2001b). Data typically show that a large percentage of children are reported to eat food with their fingers and to eat food that has dropped on the floor. In a study of children's eating habits (Akland et al., 2000), during each eating event, the number of contacts observed between food and hands ranged from 10 to 39 with a mean of 19.
Other behaviors (crawling on the floor, contact with dirt and grass) may be related to children's dermal and inhalation exposures (Fenske et al., 1990; Zartarian et al., 1997; Cohen Hubal et al., 2000; Lu et al., 2001). The frequency and timing of hand washing may impact the hand-to-mouth transfer of contaminants (Freeman et al., 1997, 2001a, 2001b). Children's use of objects in the home, particularly bottles, pacifiers and blankets, may also modify their exposures (Freeman et al., 2001b).
In this study, children's mouthing and food-handling behaviors were recorded on videotape as part of a 24-month longitudinal examination of pesticide exposure conducted in the mid-Rio Grande Valley (Shalat et al., 2003). The goal of the study was to evaluate children's exposure to pesticides in and around their homes and from drift from nearby farm fields. In order to determine the relationship between children's activities and pesticide exposure, environmental samples (housedust and soil) and personal samples (hand loadings and urine samples) as well as questionnaire and videography data were collected. Sampling was carried out over four growing seasons in the region. This report describes the behavioral characteristics of the children during the first round of the study as reported by parents and observed on videotape. This study applies a videotaping methodology, previously focused on the preschool child, to infants and toddlers. The use of a videotaping methodology to study infants and toddlers allows direct comparison of both mouthing duration and frequency in these age groups.
Subjects were recruited from a small colonia south of Laredo, Texas. Details of the study design have been described previously (Shalat et al., 2003). A total of 52 children from 29 homes participated in the first round (23 homes with two children and six homes with one child). The first round of videotaping was conducted between May and August 2000. Participation in the study was voluntary. All participants were informed of the nature of the study and signed consent forms approved by the UMDNJ-RWJMS Institutional Review Board (IRB no. 2708).
A detailed baseline questionnaire was administered to a parent or guardian in each household. The questionnaire collected information about household occupations, children's hygiene, mouthing and food-handling activities, household pesticide usage and household cleaning activities. The questionnaire was derived from those used in the RTI/EOHSI National Human Exposure Assessment Survey (NHEXAS, Pellizzari et al., 1995) and allows comparison with that population-based study. Additional questions were taken from a questionnaire used to characterize hand-to-mouth and personal hygiene activities from recent studies in Jersey City and Newark, New Jersey (Reed, 1997; Freeman et al., 1997, 2001a). Questions about pesticide usage were drawn from the Non-Occupational Pesticide Exposure Study (NOPES) and the National Home and Garden Pesticide Use Survey (Whitmore et al., 1994). A brief activity questionnaire about each participating child was also administered. This questionnaire asked about the child's contact with floors, hand washing, play outside and frequency of mouthing the previous day (usually the day the child was videotaped). The parental assessment of mouthing could then be compared to the measured mouthing from the videotapes.
Each child was videotaped for 4 h using a hand-held camcorder (Sharp Viewcam Model VL-H860U). If two children were living in the same household, videotaping was carried out on two separate days. The team of two videographers alternated videotaping approximately every 30 min. The videographers followed the child through the house and yard except in areas they were not permitted to enter (i.e., a room where a family member was sleeping). The children were not videotaped napping, changing or using the bathroom. Additionally, a record was kept of any significant interruptions during videotaping as well as any off-camera hand washing.
The videotapes were analyzed using VTD (Virtual Timing Device) software (© Colin Ong, SamaSama Consulting, 1998). The software is a refinement of the VideoTraq software described by Zartarian et al. (1997a) and used previously in children's mouthing and food-handling activity studies (Zartarian et al., 1997b, 1998; Akland et al., 2000; Freeman et al., 2001). The template was modified to record the activities of interest (Figure 1). The template has three grids: (1) location, (2) activity and (3) contact. The location grid was changed to include only inside, outside and in-other (when the child entered another home). The activity grid was modified to reflect both the child's contact with water (several children were in wading pools or tubs during taping) and the child's contact with the floor. The contact grid contained categories for all hand contacts as well as mouthing activity. Each tape was reviewed twice, first recording the right-hand contacts and then left-hand contacts. For comparison, the same criteria were employed in this study as in Reed et al. (1999). First, both hand-to-mouth and object-to-mouth contacts included contact with the lips as well as insertion in the mouth. Second, dietary (food utensil, cup and bottle) and pacifier contact with the mouth were not considered object-to-mouth contact. Third, when both hands were holding an object to the mouth, the activity was recorded as mouthing only under the right hand. Fourth, if hands also contacted the mouth during object mouthing, the transcriber scored whichever contact appeared to be greatest in terms of area of contact. Two individuals transcribed the tapes. Half hour segments of five randomly selected tapes were transcribed by both individuals. Reliability of both duration and frequency transcriptions were high (R2=0.96 for duration and R2=0.94 for frequency).
The children were grouped into four age categories: infants (7–12 months), 1-year-olds (13–24 months), 2-year-olds (25–36 months) and preschoolers (37–53 months). The same chronological age categories (1- and 2 year old) have been used to study the behaviors of children and impact on blood lead levels (Freeman et al., 1997) and have been associated with a number of developmental stages (Gallahue, 1989).
Microactivity data and questionnaire results were analyzed using SPSS (Statistical Package for the Social Sciences, v.11.5.1). Evaluation of the data using the four age categories was conducted using χ2-tests of homogeneity and Spearman's correlation. Linear regression was used to determine the effect of age (months) on the behaviors. The Wilcoxon signed-rank test was used to determine differences between indoor and outdoor behavior. Median values and nonparametric tests were used due to the lack of a normal distribution of the data.
A total of 52 children, 26 boys and 26 girls, ranging in age from 7 to 53 months participated in the first round (Table 1). Eight children regularly spent time away from home, ranging from 2 to 30 h a week (six were 10 h or less; one was 20 h and one was 30 h). Seven of these children were cared for in another home. The child who was away for 30 h a week was a 40-month-old boy who was in day care.
The majority (64.7%) of children were reported to wash their hands 3–5 times each day (69.2%. infants, 50% 1 year olds, 61.1% 2 year olds and 77.8% preschool). Parents were asked about certain events of the day that might cause children to wash their hands (Figure 2). Generally, infants were least likely to have their hands washed at these events and preschoolers were most likely to wash their hands. Several behaviors peak at the 1-year-old age category. Hand washing after using the bathroom was correlated with age (Rs=0.514, P<0.001) probably due to toilet training. In contrast to other events, coming indoors did not prompt hand washing. Less than half the children in the three youngest age groups washed hands after coming indoors. The highest rate of hand washing after coming indoors was 55% among the preschoolers.
Parents were asked questions concerning their children's behaviors (Figure 3). Most behaviors peaked before preschool age. Both types of mouthing behaviors (placing nonfood objects in the mouth and placing thumb or fingers in the mouth) are reported for more than 80% of infants and 1-year-olds. A significant decrease in reported behavior with increasing age was noted for both object mouthing (Rs=−0.329, P=0.017) and hand mouthing (Rs=−0.301, P=0.030). The remaining behaviors did not have a significant relationship with age category. Food-handling behaviors, eating food that had fallen on the floor and eating without utensils, peak in the 1 and 2-year-olds. Soil ingestion peaked in the 1-year-olds (42%), but was not reported for any of the preschool children. Other behaviors such as carrying around a blanket or soft toy and placing paint chips in the mouth are more often observed in the preschool children.
Activity questionnaires were completed by the parents for 49 of the 52 children. The questionnaire collected potential exposure data by asking if the child had skin contact with soil or grass or if the child played on the floor the previous day (usually the day of videotaping) (Table 2). For both contact with soil and contact with grass, the percent of positive responses increases with the age category (Rs=0.532; P<0.001 for dirt contact; Rs=0.540; P<0.001 for grass contact). In all, 30 children (61%) were reported to spend some time playing on the floor; the reported time on the floor ranged from 10 min to 12 h: 15 children reported to spend between 1 and 2 h on the floor. A linear regression analysis was carried out to determine the association of hours playing on the floor with age (in months) and gender (male=1; female=0). No significant correlation was found (the adjusted R2=−0.023; P=0.631). No consistent trend with age was found for time on the floor; infants spent 1.5 h, 1-year-olds spent 2 h, 2-year-olds spent 1 h and preschoolers spent 2 h (Rs=−0.003; P=0.985). Questions about hand washing the previous day and bathing were also asked. When asked about hand washing, the mean number of times hands were washed the previous day was similar across all the age categories with a mean of 4 and a range from 1 to 10. Of the 49 children with completed activity questionnaires, only two children, both 2-year-olds, did not take a bath or shower the previous day.
The mean tape time was 239 min; 51 of the tapes ranged from 200 to 247 min. Although the goal was to obtain 4 consecutive hours of activity, taping was interrupted in 16 cases due to children napping (eight infants; six 1-year-olds; one 2-year-old and one preschooler). In one case, only 114 min were videotaped when a 9-month-old child fell asleep and taping was ended. The total number of hand-to-mouth, object-to-mouth and food contacts for both the left and right hands was tabulated then divided by the total tape time to determine an hourly frequency. The total duration of hand-to-mouth, object-to-mouth and food contacts was divided by the total tape time to determine the percent of time for each type of contact. The median hourly frequency and percent of time for each age group are presented in Table 3. For comparisons to other studies, the mean and standard deviation of the mean are also presented in the table.
All but one child, a 30-month-old girl, had some hand-to-mouth activity during the videotaping. The median hourly hand-to-mouth frequency decreases from infants to 2-year-olds then increases among the preschool group. The median duration also decreases from infants to 2-year-olds but increases only marginally among the preschool group. The variation for both frequency and duration among the preschool group is high. One child in this group exhibited an hourly hand-to-mouth frequency of 76, largely due to nail biting. This was more than twice the rate of any other child in this age group and was the highest frequency among all the children. Object-to-mouth hourly frequency decreases between infants and 2-year-olds then increases to the preschoolers. Object-to-mouth duration also follows a similar pattern. Again, large variation is observed in both the infants and preschoolers. One preschooler, a 45-month-old boy, had an hourly frequency of 42, largely due to placing a food wrapper in his mouth. The next highest frequency in this group was 12. Food handling also varied with age category. The frequency peaked among the 1-year-olds. The duration of food contact increased from infants to 2-year-olds then remained the same for the preschoolers. At each age category, the median duration of food contact exceeded the duration of either hand-to-mouth or object-to-mouth activity.
Regression analysis was used to find any linear trends between age (in months) and mouthing (frequency and duration). Gender was also entered (1=male; 0=female) in the regression model. No significant linear trends in hand-to-mouth hourly frequency, hand-to-mouth duration, object-to-mouth duration, food contact frequency and food contact duration with respect to age or gender were observed. In contrast, hourly object-to-mouth frequency was found to decrease as age increased (adjusted R2=0.179; P=0.003). Regression analysis was also used to examine the relationship between food-handling and hand-to-mouth activity. For children over 12 months, hand-to-mouth frequency increased significantly with increasing food contact frequency (adjusted R2=0.291; P=0.002). This effect was marginal when all age groups were included (adjusted R2=0.090; P=0.058). No significant linear regression was observed for food contact duration and hand-to-mouth duration.
Almost all of the children (49) spent the majority of their time indoors; 13 of the 52 children did not go outside during the videotaping session. Playing outside differed by age category (χ2=8.1; P=0.04). In all, 54% of the infants remained indoors during taping compared to a range of 11–22% for the other age categories. For the remaining 39 children the mean time outdoors was 24.4% of the session (ranging from 0.3 to 187 min). In total, 13 of the children were in a pool or tub during part of the videotaping session; the time in the pool or tub ranged from 7 to 57 min. Time in water was not significantly correlated with age or gender. The time the children spent playing on the floor ranged from 0 to 57% of the tape time (mean=12.0; median=7.6; SD=12.6). The time on floor decreased with age (adjusted R2=0.096; P=0.031); gender was not significant. The variability in the time spent on the floor decreased as the age category increased (Figure 4).
Overall 39 children spent time both indoors and outdoors. Using the Wilcoxon signed-rank test, all three behaviors, hand to mouth, object to mouth and food handling, were significantly greater indoors, both by frequency and duration (Table 4). Two of the 39 children had no contact with food indoors or outdoors and 18 had some contact with food outdoors.
On the activity questionnaire parents were asked first, how frequently their child sucked on their thumb or fingers and second, how frequently did their child place nonfood items such as toys in their mouth the previous day (usually the day of videotaping). The children's mouthing activity on the videotape was compared to the parents' report of mouthing activity (Table 5). Linear regression analysis showed that both hand-to-mouth frequency (adjusted R2=0.057; P=0.060) and duration (adjusted R2=0.087; P=0.026) increased with reported increasing activity. However, no significant linear association was found between parental reports of object mouthing and object-to-mouth frequency or duration recorded on videotape.
The baseline and activity questionnaire results concerning children's hand washing are consistent with each other and similar to that found in other studies. Parents typically report that their children wash their hands between 3 and 5 times a day (Freeman et al., 1997, 2001a). Despite the consistency in the number of hand washings, differences between age categories in hand washing at certain events were observed. Washing before meals, snacks, after using the bathroom and before going to bed were all found to increase with age category. Most preschool children were reported to wash their hands at all the questioned events. Only hand washing after coming indoors was poor. A study of inner city children 13–36 months old found good reported compliance (65% for all children) with washing hands after coming indoors but poor compliance before and after snacks (30 and 33%, respectively; Freeman et al., 2001a). Social factors may account for the difference in washing hands after coming indoors. In this community, children went outside and re-entered the house frequently and without parental supervision. The inner city children may not have had the same opportunities to play outdoors.
The reported behavior data from the preschool age group can be compared to data from the youngest age group (3–4-year-olds) collected by the MNCPES telephone survey (Freeman et al., 2001b). The percentages of children reported to often place objects or hands in their mouths are similar. Carrying around a blanket or soft toy is higher in the MNCPES population than in our study, possibly due to climate differences. Placing paint chips in the mouth is lower in MNCPES, possibly due to the quality of the housing. In the colonia, many of the homes are still under construction. Eating soil appears to be more prevalent in MNCPES. When the 1 and 2-year-old groups in this study were compared to children in the Children's Lead Exposure and Reduction Study in Jersey City (Freeman et al., 1997), for both groups, the percentages of children to put objects in their mouth, eat food from the floor and eat with fingers were similar.
The mouthing data show great variability across and within the age categories. The median hand-to-mouth frequency and duration decreases from infants to 2-year-olds, supporting the findings in observational studies that younger children mouth more than older children (Groot et al., 1998; Juberg et al., 2001; Tulve et al., 2002). However, in all measures (hand-to-mouth frequency and duration, object-to-mouth frequency and duration), mouthing increases from the 2-year-olds to the preschool group, although to varying degrees. Children in all age categories were observed licking their fingers before and after eating. A significant increase of hand-to-mouth frequency was observed with increased food handling, particularly for children over 12 months old. Thus, the number and type of eating events (hand-held snacks vs. meals) included in the observation period may impact the mouthing data. This impact may be more pronounced among the preschoolers who were often observed to get snacks independently. Personal habits, such as nail biting, which emerge as children develop, may also impact mouthing behavior.
Comparisons with other studies are complicated by differences in methodology. The definition of mouthing (superficial contact, licking, biting or insertion), classification of objects, especially pacifiers, and the inclusion of meals or snacks in the observation period differ among studies. Using waking time (approximately 8 h) from Groot et al. (1998), daily hand-to-mouth activity can be estimated to range from 7 (2-year-olds) to 15 min (infants). This range of duration is higher than observed by Groot (6–8 min) for similar age groups. Object-to-mouth duration among infants can be estimated to be 15 min; only half of the 31 min of object mouthing (excluding toys meant for mouthing) in Groot et al. (1998). More overlap between the two studies was found in object-to-mouth duration among the older age groups. Tulve et al. (2002) reported a clear grouping of children in two distinct age categories. They found higher object-to-mouth activity for children under 24 months (a median of 44 events/h) compared to the infants and 1-year-olds in this study; however, no distinction is made for pacifier use. In this study, only two children (one 9 months old and one 31 months) were observed to use pacifiers during taping. In other populations, pacifier use may be more widespread. The object-to-mouth frequency (a median of 10 events/h) for older children (24–60 months) reported by Tulve et al. (2002) is more comparable to frequencies observed for 2-year-olds and preschoolers in this study. Reported hand-to-mouth frequencies (a median of 16 events/h for younger children and 10 events/h for older children) are similar to medians found in this study. The median frequency values in this study are higher than those found by Reed et al. (1999) in 2–6-year-old children (8.5 hand-to-mouth events/h and 3.6 hand-to-mouth events/h).
The Child Exposure Factor Handbook (USEPA, 2002) recommends using 49 total (hand- and object-to-mouth) contacts/h to represent mouthing behavior for children 10–72 months. This is higher than the combined medians observed in any age group in this study, but lower than five children in the infant category (combined frequencies 60–74 events/h) and one child in the preschool category (combined frequency of 88 events/h). Thus, the recommended contact value may not be protective of all children, particularly those 12 months and younger.
In conclusion, results of handwashing and mouthing activity reports from the baseline questionnaire show some differences in reported behavior between this community and others studied previously. Thus, behavior may vary by region, culture or socioeconomic group. No linear trend of decreasing hand-to-mouth activity with increasing age was observed. Object-to-mouth activity did decrease with age, largely due to the decrease from infants to 1-year-olds. The high rate of object-to-mouth activity among the infants suggests that object mouthing may be as important as hand mouthing in terms of exposure for this age group. This group also had the lowest rates for reported hand washing. Mouthing variability (both hand and object) was highest among the preschool age group. One explanation is that certain children at this age begin to develop habits that may affect mouthing activity. Food handling was found to affect mouthing activity. For children over 12 months of age, more contacts with food led to more hand-to-mouth behaviors. Meals and snacks will impact assessments of mouthing behaviors. All mouthing and food-handling activities were greater indoors than outdoors, but, in this population, hand washing after coming indoors was not widespread. Parental reports correlated decreasing hand and object mouthing with increasing age. However, parental reports are not strongly correlated with observed hand or object mouthing. Parents may be comparing the child's current mouthing to his/her earlier behavior rather than the mouthing of other children. One advantage of parental assessments is that parents observe their children over longer time periods and under more natural conditions than the videotaped observation. Future research may include alternate methods of asking about mouthing behavior to improve correlation of questionnaire data with videotaped observations or exposure metrics. Finally, the current recommendation of 49 combined hand- and object-to-mouth contacts/h may not be protective of children under 12 months of age; 38% of the infants in this study exceeded this value. This high mouthing rate, from both hand and object mouthing, suggests that infants have greater potential exposures than children in other age groups.
The results reported here are from the first phase of the 24-month longitudinal study. Of the 52 children, 44 (and 16 additional children) participated in the second phase, approximately 6 months later. In addition to baseline questionnaires, all children were once again videotaped for 4 h. Thus, changes in children's behavior and quantified mouthing activity can be assessed. The tapes have been transcribed and the results will be forthcoming. The study also includes collection of exposure data within the same time frame as behavioral data. The temporal linking of behavior to dose is essential in identifying the determinants (and, ultimately predictors) of exposure.
This project was supported by EPA STAR Grant No. R827440 and NIEHS grants to the Center for Environmental Health Sciences in Piscataway, NJ (P30-ES-05022) and the Center for Environmental and Rural Health at Texas A&M University (P30-ES-09106). This study would not have been possible without the cooperation and assistance of the Sisters of Mercy of Laredo, Texas and, in particular, Sister Rosemarie Welch.
About this article
Journal of Exposure Science & Environmental Epidemiology (2011)