Materials and Methods

Data were collected to identify specific risk factors for work-related bite injuries among veterinary technicians certified in Minnesota. The study was approved by the University of Minnesota Institutional Review Board Human Subjects Committee.

Population and study design—Veterinary technicians who were certified in Minnesota through the end of October 2004 and who worked as CVTs within the 12 months prior to initial survey administration were eligible for inclusion in the study. A total of 1,465 CVTs were identified from a database maintained by the Minnesota Veterinary Medical Association, and surveys were sent to 1,427 who had valid mailing addresses; those who had not worked as CVTs during the 12 months prior to the survey were excluded from the study (although certified in Minnesota, participants could have worked in any state during the specified period). The initial survey^a sent through postal mail between April 30 and June 30, 2005, was used to identify individuals with and without work-related injury events during the previous 12 months, to collect information on the consequences of those injuries, and to identify potential risk factors associated with those injuries. Of 1,052 individuals who responded to the initial survey, 873 were eligible for inclusion in the analysis. Data obtained from that survey were described elsewhere.¹⁵

For the present study, a nested case-control design was used to identify specific risk factors and potential protective factors associated with animal-bite injury–related outcomes identified in the previous study.¹⁵ Questionnaires^a that consisted of 41 multiple-choice or open-ended questions were sent between August 31 and November 30, 2005, to all participating CVTs who had answered questions regarding bite-related injuries in the initial survey. Similar to the initial survey method, case-control questionnaires were mailed up to 3 additional times if individuals did not respond. All mailings included cover letters that provided information about the study and informed consent. To encourage participation, each participating CVT could choose to be included in a random drawing with a 1 in 30 chance to win a retail gift card. Follow-up was performed by mail and telephone to clarify missing or unclear written responses.

Bite injuries—To be defined as a work-related bite injury, an injury must have been associated with occupational activities, involved bruising or breaking of the skin, or resulted in ≥ 1 of the following: restriction of normal activities (for any length of time), loss of consciousness, or use of medical assistance, including self-treatment. In the initial survey,¹⁵ descriptive information was requested for up to 4 of the most severe injuries incurred by a respondent; for purposes of the present study, only those respondents with bite injuries that had this descriptive information provided were included in the case group.

Group assignment and data collection—Case CVTs (301 surveys sent) were defined as those who reported ≥ 1 qualifying work-related bite injury within the 12 months prior to the initial survey.¹⁵ If > 1 eligible bite injury event was reported by a CVT, a single event was selected randomly as the event of interest. This method was chosen over selecting the most remote event, the most recent event, or the most severe event because each of the alternate methods of case event selection could have contributed to potentially greater study bias owing to misclassification and recall bias. Exposure information was collected for the month prior to the month in which the selected bite injury occurred.

All initial survey respondents who did not report a qualifying work-related bite injury in the 12 months prior to the survey were eligible as controls. Of 572 possible control CVTs identified, questionnaires were sent to 567 (5 CVTs had withdrawn from the study). To obtain exposure information, each control was randomly assigned a month on the basis of the months he or she indicated working during the study period; data were then collected for the month prior to the selected month.

The CVTs included in the case and control groups were asked to provide personal information (including demographic data; time, place, and type of animal restraint training received; duration of experience working in a veterinary facility of any kind; and whether they had incurred work-related animal bite injuries that resulted in bruises or broken skin prior to the month covered by the questionnaire). Additionally, information regarding employer practices (eg, animal restraint and bite prevention practices, support and encouragement for use of proper animal restraint, training requirements and practices, and whether any staff member could stop a procedure if they felt staff safety may be compromised), facility environment (eg, type of clinic or practice; number of veterinarians, CVTs, assistants, and other staff that handled animals in the facility or department; relevant coworker training and experience; coworker support of safety practices; workload and scheduling; quality of staff communication; species and number of animals handled daily; and availability of adequate restraint equipment), and animal owners (eg, whether owners were allowed to restrain their own animals during examination or treatment, the CVT's perception of owners’ capability to restrain their animals, and whether presence of the owner affected restraint procedures) were collected. These data were then used to compare exposures between the case and control groups.

Data analysis—Risk factors for animal bite injuries were assessed through multivariate modeling with a causal model and specific directed acyclic graphs^16–18,a pertinent to exposures of interest. This method was used in a prior study¹⁹ of injuries acquired by veterinarians. To determine associations between exposures of interest and bite-related injuries, the directed acyclic graphs facilitated identification of potential confounders to be controlled for in each multivariate model. These methods identify parsimonious models and exclude covariates that should not be entered into the regression because of the potential to introduce bias. Multiple logistic regression was used to calculate ORs and 95% confidence intervals. Findings were considered significant if the 95% confidence interval did not include 1.0. Differences in the levels of exposure variables were considered significant if the confidence intervals did not overlap. In addition, the potential impact of an unmeasured confounder related to the magnitude and direction of potential bias was examined by means of sensitivity analyses.²⁰

Results

Of 868 CVTs that were sent questionnaires for the nested case-control study (301 and 567 recruited as potential cases and controls, respectively), 489 responded. These included 176 case CVTs who had incurred a work-related bite injury in the 12 months prior to the initial survey and 313 control CVTs who had not; not every respondent answered all questions. Characteristics for each group were summarized (Table 1). Both groups predominantly comprised females. Fifty-four of 176 (31%) respondents in the case group were ≥ 35 years of age, compared with 139 of 313 (44%) controls (data were not reported for 9 CVTs). One hundred twelve (64%) CVTs of the case group worked in small animal clinics, and 10 (6%) worked in a college or university setting, compared with 166 (53%) and 44 (14%), respectively, for the control group. Other apparent differences among groups included the proportions that reported working ≥ 8 h/d on average (case group, 153/176 [87%]; control group, 251/313 [80%]), having had a work-related bite injury prior to the period included in the questionnaire (case group, 155 [88%]; control group, 245 [78%]), and having worked < 5 years in any veterinary facility (case group, 80 [45%]; control group, 87 [28%]). Case CVTs more commonly handled ≥ 10 animals during a typical day (163 [93%]) and handled ≥ 3 species during a typical day (82 [47%]), compared with control CVTs (244 [78%] and 114 [36%], respectively).

One hundred one of 176 (57%) CVTs of the case group indicated that animal owners were always present or frequently present during examinations, compared with 149 of 313 (48%) in the control group. One hundred six (60%) and 151 (48%) case and control CVTs, respectively, indicated that animal owners were allowed to restrain their pets during examination or treatment.

On a scale ranging from 1 (always) to 5 (never), 85 of 176 (48%) and 170 of 313 (54%) CVTs of the case and control groups, respectively, indicated that they always had adequate support from their employer to use proper animal restraint. On the same scale, 54 (31%) case and 141 (45%) control CVTs indicated they always had adequate staff available for animal restraint, and 41 (23%) case and 109 (35%) control CVTs reported that sufficient time was always allowed in the schedule to perform adequate animal restraint. By use of a scale from 1 (strongly agree) to 4 (strongly disagree), 84 (48%) CVTs in the case group strongly agreed with the statement “I felt confident in handling animals when applying restraint techniques,” whereas 180 (58%) CVTs in the control group selected this response.

Given that most (337/489 [69%]) CVTs in the study worked in small or mostly small animal clinics (cases, 132/176 [75%]; controls, 205/313 [65%]), as identified in the previous study,¹⁵ multivariate analysis was focused on this more homogeneous population to investigate potential differences in types of exposures (Table 2).

Significant associations between risk of bite injury and age or duration of employment were identified; CVTs < 25 years of age were > 3 times as likely to incur a bite injury, compared with those ≥ 35 years of age. Similarly, CVTs who worked < 5 years in any veterinary facility were > 3 times as likely to receive a bite injury, compared with those who worked in such facilities for ≥ 10 years. Working with ≥ 5 species of animals on a typical day was associated with a risk of bite injury 2 times that of working with < 3 species. The risk for CVTs who reported they frequently had adequate staff available for animal restraint was 2 times that of CVTs who indicated they always had adequate staff for this purpose; however, risk did not differ significantly from that of the referent category for those who reported adequate staff were sometimes or infrequently (to never) available for this task. The risk of bite injury was significantly lower for CVTs who reported handling < 10 versus ≥ 20 animals on a typical day.

Sensitivity analyses were conducted to examine the impact of an unmeasured confounder related to the magnitude and direction of potential bias. Given that one of the results of this study suggested that CVTs’ perceptions that they frequently, but not always, had adequate staff available for proper animal restraint was associated with increased risk of bite injury, sensitivity analyses were conducted to consider the potential effect of the population density of the clinic location in which the CVTs worked during the month for which they had answered questions about their exposures. This factor could be associated with the pool of available trained veterinary staff within the community and the risk of bite injury; variability in characteristics of animals, such as animal species and behaviors, between rural and urban areas could also result in dissimilarity in risk of bite injuries to CVTs. With prevalence of the unmeasured confounder being defined as the proportion of CVTs who worked in a more densely populated work area, the examination of estimated ranges of this prevalence between staffing exposure rates was used to determine the effect of the unmeasured confounder on the risk associated with the exposure. To reverse the effect of frequently versus always having adequate staff available for restraint (eg, to change the OR from 2.02 to 0.93), the unmeasured confounder would have to be a relatively strong risk factor (OR, ≥ 3.0) with differences in prevalence across the exposed and unexposed groups (frequently having enough staff vs always having enough staff) of 0.8 and 0.1, respectively. The point estimate was considered, however, and not the precision of the estimate for this sensitivity analysis.

Discussion

This population-based case-control study enabled identification of potential risk factors for work-related bite injuries to veterinary technicians certified in Minnesota, with a focus on those working in small animal or mostly small animal clinics. A similarly designed study¹⁹ previously investigated work-related bite injuries in a population of veterinarians through multivariate logistic regression analyses.

In the present study, the number of hours worked (on average) per day as reported by CVTs was analyzed as a surrogate measure of fatigue, and multivariate analysis revealed no significant differences in the risk of bite injury among CVTs working ≥ 8 h/d, compared with those working < 8 h/d. A prospective study design may be more effective in identifying any associations between risk of bite injury and number of hours worked daily.

Significant associations were found between the risk of bite injury and younger age (when CVTs < 25 and ≥ 35 years of age were compared) as well as fewer years worked in any veterinary facility (when CVTs with < 5 years and ≥ 10 years of experience were compared). Similar findings were reported by Gabel and Gerberich.¹⁹ Further research may attempt to explain the effects of the type and quality of work experience on the risk of bite injury.

Results of the present study suggested that having no prior bite injuries (that resulted in bruises or broken skin) was associated with a decreased risk of bite injury during the study period, similar to findings reported for veterinarians.¹⁹ It is possible that CVTs reporting a bite injury during the study period may have had enhanced recall of more distant events that may, in part, have been related to severity of the events or that they worked in an environment where bite-related injury was more likely to occur.

Our results also suggested that the risk of bite injury was associated with the CVTs’ perceptions of availability of staff for animal restraint. The CVTs who reported frequently having adequate staff for animal restraint were more than twice as likely to experience a bite injury, compared to those who reported always having adequate staff available. However, interpretation of this result was difficult, as lower frequencies (sometimes, infrequently, or never) of having adequate staff for animal restraint were not statistically associated with increased risk of bite injury. The findings of reduced risks associated with handling lower numbers and fewer species of animals on a typical day are also important considerations. In particular, this suggests a need for veterinarians and practice managers to consider, carefully, the patient load and to provide adequate staffing. Further research may also elaborate on the effect of types and ratios of staff to patients and the availability of other resources on the outcome of bite injury.

Our questionnaire-based study relied on the CVTs’ perceptions and self-reported descriptions of relevant bite injury events and exposures. These methods may possibly have resulted in some biases, including those related to recall, information misclassification, selection, and confounding. In an attempt to reduce recall bias, the recall period for exposures was limited to 1 month before a bite injury that occurred in the previous 12 months. This method was used in a prior study¹⁹ of work-related injuries among veterinarians. To reduce information bias, follow-up was performed by mail and telephone to clarify missing or unclear written responses. All veterinary technicians with active certification in Minnesota were selected for inclusion in the initial study.¹⁵ No demographic information was available in the CVT certification database, so it was not possible to consider response bias or eligibility. Although there may have been bias attributable to the influence of an earlier event for some case CVTs who experienced multiple bite-injury events during the study period, random selection of 1 event for analysis was expected to result in less bias than was selection of an event on the basis of timing or severity of the injury. The selection of potential confounders for logistic regression analyses was guided by directed acyclic graphs^16–18,a to minimize the effect of confounding. The risk of bite injury may vary among different regions of the country and may also differ between certified and uncertified veterinary technicians.

This population-based study provided preliminary findings and analysis of case-control data pertinent to risk factors for work-related bite injuries among CVTs. The risk factors identified included younger age (< 25 years vs ≥ 35 years), shorter experience (< 5 years vs ≥ 10 years), perception of frequently versus always having adequate staff available for proper restraint, and greater number of animal species handled on a typical day (≥ 5 vs < 3 species). A decreased risk was associated with handling fewer animals on a typical day (< 10 vs ≥ 20). The results from this study add to our understanding of the relationships between various occupational exposures and work-related bite injuries among CVTs. Future studies might explore whether changes in various exposures over time have an impact on the frequency or severity of work-related bite injuries.