An Analysis of Reported Dog Bites: Reporting Issues and the Impact of Unowned Animals.

Abstract

The authors investigated the epidemiology of dog bites in a large city by analyzing dog bite data from the period 1995 through 1997. The following data pertaining to victims were studied: gender, race, age, residing health district, exposure date, reporting date, duration between exposure and reporting (lag time), reporting agency, hospitalization, treatment, and location of bite. Data pertaining to the dogs--breed and ownership status--also were considered, Male-victim bite rates increased 21 percent over the three-year period, possibly because of an increase in stray dogs. Medical centers reported the majority of bites (58 to 68 percent). The mean lag time for the period ranged from 4.4 to 5.6 days. During 1997, the lag time was lower for owned dogs (4.0 days) than for stray dogs (5.2 days) (p [less than] .001). Lag time in 1997 also was lower for victims five to nine years of age (4.1 days) than for victims 15 to 19 years of age (63 days) (p = .006). Dog bites continue to increase dramatically. The invest igators identified several reporting characteristics to which attention should be paid so that dog bites can be reduced, along with the public health hazards they present.

Introduction

The epidemiology of dog bites has been studied intensively for the past 40 years. Previous studies have identified both victim and animal characteristics, leading to well-defined risk factors [1-10]. Nevertheless, dog bites continue to be a leading public health concern facing the United States. Each year, dog bites result in an average of 17 human deaths, 6,000 hospitalizations, and 330,000 emergency room visits [9,11-13]. An equal number of people do not seek treatment or report the incident, yet may suffer from psychological trauma, anxiety, and missed work or school [4,14]. In addition, dog bites remain a possible mode of rabies infection, requiring postexposure prophylaxis. The serious nature of this issue was demonstrated by a recent collaboration between the American Veterinary Medical Association and State Farm Insurance Company [15]. The goal of the partnership was to educate the public in order to decrease bites and the liability claims presented by victims. Annual direct medical care charges for do g bites in the United States have been estimated at $164.9 million [13].

The aim of this study was to investigate the epidemiology of dog bites in a large urban area--Philadelphia, Pennsylvania. The authors sought to validate the animal and victim characteristics previously defined, as well as to further characterize the reporting of dog bites. Specifically, they proposed to determine which agencies were responsible for reporting dog bites and the duration between the date of bite and reporting.

Methods

The authors conducted an epidemiological study of dog bites reported to the city of Philadelphia between 1995 and 1997. Data were collected from the Communicable Disease Registry (CDR) database compiled by the Division of Disease Control (DDC) at the Department of Public Health. All cases classified as dog bites reported between January 1, 1995, and December 31, 1997, were garnered and statistically analyzed.

Animal bites are reportable to the city under sections 6-201 and 6-202 of the Regulations Governing the Control of Communicable and Noncommunicable Diseases and Conditions, established under Title 6 (Health Code) of the Philadelphia Code. Notifiable disease case reports. (Form 55-A-116 A) are filed for each bite. Information for these reports is gathered from reporting agencies, such as hospitals, police departments, and health centers, as well as from victims. If the dog's owner is known, an investigator is sent to the residence 10 days following the incident to verify the health status of the charged animal. At this time, additional information may be gathered and filed (Form 55-A-2513) regarding the animal and circumstances of the bite. The information from both forms, if available, is entered into the CDR database maintained at the DDC office.

The following data pertaining to victims were analyzed: gender, race, age, residing health district, exposure date, reporting date, reporting agency, hospitalization, treatment, and location of bite. Data pertaining to the dogs--breed and ownership status--also were analyzed. With respect to ownership, animals were classified as either owned or unowned. They were designated as owned if an address was given. In virtually all cases, the animal investigator confirmed this address. All animals designated as strays by the reporting agency or for which ownership was unknown were classified as unowned. The reporting date was defined as the date when the case was filed with DDC. Lag time was defined as the number of days between the exposure date and reporting date. Reporting agencies were divided into seven categories:

1. medical centers (hospitals, private physicians, and medical health centers);

2. health departments (county, city, and state health departments);

3. schools;

4. police officers;

5. owners;

6. victims (or relatives of victims); and

7. other.

The category "other" includes veterinarians; employees from the U.S. Postal Service, the Society for the Protection of Animals, and insurance agencies; and individuals whose relationship to the victim was not noted. Dog breeds were classified either as pure breeds (when only one breed was listed) or as mixed. In addition, the authors independently studied specific breeds that were listed more than eight times in a single year. All data were sorted within each field, and the validity of outliers was confirmed with hard copies of the original reports.

To calculate bite rates, the authors used population estimates from vital-statistics reports, organized by census tract and health district, which were provided by the city's Division of Information and Research Systems. A t-test was used to compare quantitative data (a p-value less than .05 was considered significant). Descriptive analysis was performed with Excel 97, "Descriptive Statistics" (Microsoft Corporation).

Results

From January 1, 1995, through December 31, 1997, 6,293 animal attacks were reported to the city of Philadelphia. Of these cases, 5,390 (85.7 percent) were dog related, while the remaining 903 (14.3 percent) involved cats (633, or 10.1 percent); rats (39, or 0.6 percent); squirrels (35, or 0.6 percent); raccoons (three, or less than 0.1 percent); and ferrets, gerbils, groundhogs, guinea pigs, hamsters, horses, humans, mice, opossums, rabbits, skunks, snakes, and others (total of 193, or 3.1 percent). Of all the animal attacks involving dogs, bites (5,348, or 99.2 percent) were more prevalent than other types of exposure, notably saliva and scratch. During the three-year period, reported dog bites rose 13 percent from 1,659 in 1995 to 1,877 in 1996, and fell 9.3 percent to 1,702 in 1997. The bite rates per 100,000 people, according to population estimates, were 110.7 in 1995, 125.2 in 1996, and 120.8 in 1997--showing an increase of 9.1 percent over the three-year period.

The analyses of victim and animal characteristics gave results very similar to those outlined by previous authors [1-10]. The percentage of victims who were males ranged from 47 percent to 72 percent among the age groups studied. During the three years, males accounted for an increasing percentage of the victims--56 percent in 1995 and 1996, and 63 percent in 1997. Likewise, the bite rate for males rose over the three years, from 135.2 per 100,000 in 1995, to 151.5 in 1996, to 164.1 in 1997--an increase of 21 percent. An increase also was noted in the bite rate among females, from 77 per 100,000 to 100 per 100,000 (an increase of 30 percent)--although the rate returned to 81.3 per 100,000 in 1997 (giving an overall increase of 5.6 percent for the three-year period). The increase in bite rates appears to be attributable to an increase in bites by unowned dogs (Figure 1). While the bite rate from owned dogs dropped one percent among males (from 87.44 in 1995 to 86.57 in 1997), the rate for unowned dogs increas ed over 64 percent (from 47.2 in 1995 to 77.5 in 1997). An increase also was noted in the rate at which female victims were bitten by unowned dogs (from 24.8 to 35.4, or 42 percent), as was a decrease in those bitten by owned dogs (from 51.8 to 45.9, or -11.4 percent). Cases in which gender was unknown accounted for 6.0 percent in 1995 and 1.0 percent or less in 1996 and 1997.

As with previous studies, the bite rate was dramatically higher among children (less than 15 years of age) than among other age groups (Figure 2) [4,5,8,10]. Children between five and nine years of age experienced the most bites (224 to 288) and had the highest bite rates (200 to 257 per 100,000). The age of victims bitten varied significantly between 1995 and 1996 (t-test, p = .02). During 1997, there was a dramatic increase in the rate for the age groups zero to four and five to nine (49 percent, more than eight times that of the previous years). For all years, bite rates among victims older than 19 years of age showed a steep decrease from rates among younger victims, with a small peak during middle age (30 to 40 years of age). In roughly 20 percent of the cases, the victim's age was unknown. The authors also noted a seasonal variation in bite occurrence. More than 55 percent of the bites occurred between April and August when blank entries (18 percent) were discounted. This kind of variation has been not ed in several other studies [3,4,7,8]. In 1997, seasonal variation was apparent for all age groups, although the percentage varied with age (from 46 percent for victims 15 to 19 years of age, up to 64 percent for victims 40 to 44 years of age).

The distribution of dog bites on the victim's body was similar to that found by previous studies. Approximately 25 percent of bites occurred on the lower extremities, 35 percent occurred on the upper extremities, 12 percent occurred on the face or head, and relatively few occurred on the neck and trunk (three percent) [5,7,16]. More specifically, the hand and leg had the highest percentage of bites (total of 45 percent). In 16 percent of cases, the location of the bite was not recorded. Like previous researchers, the authors noted a dramatic increase in the percentage of bites to the face or head for children through 14 years of age (23.5 percent) [9,16]. In addition, those bitten by unowned dogs were less likely to be bitten in the face or head than those bitten by owned dogs (five percent for unowned and nine percent for owned), as previously reported by Wright [10]. Very few of the victims required hospitalization; eight in 1995, one in 1996, and two in 1997. These numbers are, however, lower than that re ported by Avner, who found that 12 Philadelphia children (seven percent) required hospitalization after dog bites in 1989 [1].

The dog's breed was not listed in over 74 percent of the cases. For cases in which breed was given, however, the authors' findings matched those of previous studies: Pure breeds were responsible for the majority of the bites (88 percent), with several specific breeds highly represented [1,5--7]. Pit bulls, German shepherds, and Rottweilers combined were responsible for over 59 percent of bites each year.

Unlike previous studies, this study found that ownership was confirmed in only 66 percent, 55 percent, and 54 percent of cases in 1995, 1996, and 1997, respectively These numbers are much lower than those typically cited, which tend to be greater than 75 percent [4,7,10,17]. As noted, the percentage of bites from unowned dogs increased 20 percent from 1995 to 1996.

Philadelphia is divided into 10 health districts. Each district comprises several contiguous neighborhoods. The authors found that eight of the districts had relatively similar bite rates (Figure 3). Two of the districts (District 6 and District 7) had rates two to three times those of the other districts. District 6 and District 10 were the only districts to show an increase in bites over the three-year period (42.2 percent and 9.7 percent, respectively). Medical centers were responsible for reporting the majority of the cases (58 to 68 percent), followed by victims or relatives of victims (12 to 14 percent), and police officers (six to nine percent). Schools, health departments of neighboring cities, and "other" each reported one to two percent of the cases (Figure 4).

Owners of dogs that were responsible for bites were a factor in 1995, reporting over six percent of the cases; however, this statistic became negligible (less than one percent) in 1996 and 1997. The authors did not note any trends between age groups and reporting agencies. The one exception was the obvious one--that schools play a larger part during childhood (up to six percent) than they do later in life (less than one percent).

Analysis of reporting-agency data on owner-ship status led to several interesting findings. First, the percentage of reports medical centers filed was much higher for unowned-dog cases than for owned-dog cases (77 percent versus 54 percent). This finding may indicate that victims bitten by unowned dogs are more likely to seek medical care than those bitten by owned dogs, contradicting Wright's results [10]. Second, the opposite was true for reports by victims or relatives of victims (unowned dogs accounted for five percent of the reports, while owned dogs accounted for over 18 percent). Approximately eight percent of the records for reported bites did not note the reporting agency In very few cases (less than 0.1 percent) did multiple agencies report the same incident, suggesting little overlap.

The mean lag times (the number of days between date of exposure and report to DDC) were 4.5 in 1997, 4.4 in 1996, and 5.6 in 1995. The larger number for 1995 was significant (t-test, p [less than] .001). During the three years, lag times ranged from zero to 186 days. The median was three days. During 1997, the mean lag time for owned dogs (4.0 days) was significantly lower than that for unowned dogs (5.2 days) (t-test, p [less than] .001). For the same year, the mean lag time varied from 4.1 (for victims five to nine years of age) to 6.3 (for victims 15 to 19 years of age). This difference also was significant (t-test, p = .006).

Discussion

It is clear that even with the increasing popularity of "alternative" pets, dogs accounted for the majority of animal bites in the years 1995--1997, and that dogs continue to present the leading public health concern among domestic animals. While the data are consistent with those found by previous studies in terms of bite locations, victims' gender, seasonal variations, ages at risk, and prevalence of biting among various dog breeds, the authors uncovered several issues that may be integral to the dog bite problem in urban environments.

Of particular concern is the dramatic increase in Philadelphia's reported bite rate over the three-year period (9.1 percent). Young males continue to significantly outnumber all other groups at risk of being bitten, and this study suggests that the trend is increasing in Philadelphia, as indicated by a 21 percent increase in the bite rate for male victims during the three years. The data attribute this dramatic increase to unowned animals. It is important to note the extraordinarily high percentage of unowned dogs seen in this study (60 percent) as compared with that found in previous studies (as low as 9.5 percent) [7]. The population of unowned dogs may be artificially elevated for several reasons. First, an animal was classified as unowned if its residence could not be verified. Second, although animal control officers do their best to obtain owner information, this procedure may not be as thorough as procedures used for previous studies, such as investigations performed by police officers or the scientis ts conducting the study Nevertheless, there was an undeniable increase in the bite rate for males bitten by unowned dogs. From 1996 to 1997 there was a 14 percent increase in this rate, while the percentages of bites from unowned dogs remained unchanged. The city of Philadelphia has recently instituted a mandatory dog-licensing program. Although currently in its infancy, this program may be a great asset in helping to determine dog ownership. The authors believe Philadelphia should aggressively pursue this program, especially the use of identification microchips, as a step toward correctly identifying the impact of ownership on dog bites in Philadelphia. This issue clearly warrants further investigation.

It was interesting to note that medical centers were responsible for reporting more than 58 percent of the dog bite cases. Since relatively few dog bites result in serious injury and the majority do not require medical attention, this finding suggests a very high degree of underreporting [3,7,14,18]. Of course, this result is not surprising in light of Chang's estimate that as many as 63 percent of bites are not reported [14].

Although the lag time between the bite and the filing of a report with DDC improved from 1995, it still remained quite large. The length of the lag crucially affects the provision of time-sensitive medical advice, the ability to obtain detailed information from the victim regarding the circumstances of the bite, the identification of the dog, and, if necessary, its capture. The numbers indicate insufficient public awareness that immediate reporting of dog bites is mandatory Although police, victims, and schools are involved in reporting bites, their participation is not adequate. Of notable concern is the decrease in reports from dog owners over the three years. As this group is relatively easy to address, through dog-licensing information and veterinarians, its members should be encouraged legally to take a larger role in the reporting process. The authors found that teenagers and victims bitten by unowned dogs waited the longest to report bites and that these groups should be targeted by additional educati onal efforts.

Conclusion

Overall, the risk factors for dog bite victims were consistent throughout the three-years covered by this study. Potential victims can be identified by gender, seasonal variations, age, and, possibly, breed of dog. It is important to note the dramatic increase (9.1 percent) in Philadelphia's reported bite rate over the three-year period. Of particular note is the 21 percent increase in the male-victim bite rate. This increase may be due to an increase in the city's dog population, particularly in the population of unowned dogs. The percentage of unowned dogs reported in bite cases (60 percent) is notably higher than percentages reported by earlier studies. If animal control inspectors are to be effective in reducing the incidence of dog bites in Philadelphia, perhaps they should be provided with additional resources that would help them better ascertain the ownership of dogs. An improved method of animal identification, combined with the city's animal licensing program, may prove to be an effective method for identifying more stray dogs and reducing the incidence of dog bites.

Acknowledgements: The authors wish to thank Robert Levenson and Alice Yang of Philadelphia's Division of Disease Control for their assistance with the collection of data from the Communicable Disease Registry (CDR) database. Work on this paper was supported by the Philadelphia Department of Public Health.

Corresponding Author: Robert R. Hodge, M.S., M.P.H., Ph.D., Department of Public Health, 500 S. Broad St., 2nd Floor, Philadelphia, PA 19146.

REFERENCES

(1.) Avner, J.R., and M.D. Baker (1991), "Dog Bites in Urban Children," Pediatrics, 88(1):55-57.

(2.) Beck, A.M., and B.A. Jones (1985), "Unreported Dog Bites in Children," Public Health Reports, 100(3):315-321.

(3.) Beck, A.M., R. Lockwood, and H. Loring (1975), "The Ecology of Dog Bite Injury in St. Louis, Missouri," Public Health Reports, 90(3):262-267.

(4.) Berzon, D.R., R.E. Farber, J. Gordon, and E.B. Kelley (1972), "Animal Bites in a Large City--A Report on Baltimore, Maryland," American Journal of Public Health, 62(3):422-426.

(5.) Gershman, K.A., J.J. Sacks, and J.C. Wright (1994), "Which Dogs Bite? A Case-Control Study of Risk Factors," Pediatrics, 93(6):913-917.

(6.) Greenhalgh, C., R.A. Cockington, and J. Raftos (1991), "An Epidemiological Survey of Dog Bites Presenting to the Emergency Department of a Children's Hospital," Journal of Pediatrics and Child Health, 27(3):171-174.

(7.) Hanna, T.L., and L.A. Selby (1981), "Characteristics of the Human and Pet Populations in Animal Bite Incidents Recorded at Two Air Force Bases," Public Health Reports, 96(6):580-584.

(8.) Ndon, J.A., G.J. Jach, and W.B. Wehrenberg (1996), "Incidence of Dog Bites in Milwaukee, Wisconsin," Wisconsin Medical Journal, 95(4):237-241.

(9.) Weiss, H.B., J.H. Coben, and D.I. Friedman (1998), "Incidence of Dog Bite Injuries Treated in Emergency Departments," Journal of the American Medical Association, 279(1):51-53.

(10.) Wright, J.C. (1990), "Reported Dog Bites: Are Owned and Stray Dogs Different?" Anthrozoos, 4(2):113-119.

(11.) Sacks, J.J., S.E. Bonzo, and R.W. Sattin (1989), "Dog Bite--Related Fatalities from 1979 through 1988," Journal of the American Medical Association, 262(11):1489-1492.

(12.) Sacks, J.J., J. Hornreich, R. Lockwood, and R.W. Sattin (1996), "Fatal Dog Attacks, 1989-1994," Pediatrics, 97:891-895.

(13.) Quinlan, K.P., and J.J. Sacks (1999), "Hospitalizations for Dog Bite Injuries," Journal of the American Medical Association, 281(3):232-233.

(14.) Chang, Y., J.H. Cohen, D.L. Hennon, R.E. LaPorte, and J.E. McMahon (1997), "Dog Bite Incidence in the City of Pittsburgh: A Capture-Recapture Approach," American Journal of Public Health, 87(10): 1703-1705.

(15.) Monti, D.J. (1998), "Dog Bite Prevention Campaign: Nipping a Problem in the Bud," Journal of the American Veterinary Medical Association, 212(9):1345.

(16.) Beck, A.M. (1991), "The Epidemiology and Prevention of Animal Bites," Seminars in Veterinary Medicine and Surgery (Small Animal), 6(3):186-191.

(17.) Beck, A.M. (1981), "The Epidemiology of Animal Bites," Compendium on Continuing Education for the Practicing Veterinarian, 3(3):254-257.

(18.) Lauer, E.A., B.A. Lauer, and W.C. White (1982), "Dog Bites: A Neglected Problem in Accident Prevention," American Journal of Diseases of Children, 136(3):202-204.

                     Sources of Dog Bite Reports, 1995
medical centers           58%
owners                     6%
victim/relative of victim 12%
schools                    2%
police officer             8%
health departments         2%
other                      1%
blank                     11%

Practical Stuff!

* Each year, dog bites result in an average of 17 human deaths, 6,000 hospitalizations, and 330,000 emergency room visits in the United States.

* An equal number of people do not seek treatment or report the incident, yet may suffer from psychological trauma, anxiety and missed work or school.

* Dog bites also remain a possible mode of rabies infection, requiring postexposure prophylaxis.

* The authors conducted an epidemiological study of dog bites reported to the city of Philadelphia between 1995 and 1997.

* Bites reported in Philadelphia increased dramatically (9.1 percent) over the three-year period.

* The bite rate was highest among children.

* The study also found a 21 percent increase in the bite rate for male victims.

* This increase may be due to an increase in the city's dog population, particularly in the population of unowned dogs.

* An extraordinarily high percentage of unowned dogs was found in this study (60 percent).

* Another significant finding was the length of the reporting lag time (the number of days between date of bite and report to Division of Disease Control).

* Teenagers and victims bitten by unowned dogs waited the longest to report bites.

* These groups should be targeted by additional educational efforts.

* Lag time crucially affects the provision of time-sensitive medical advice, the ability to obtain detailed information from the victim, the identification of the dog, and, if necessary, its capture.

* The authors also believe Philadelphia should aggressively pursue its new mandatory dog-licensing program, especially the use of identification microchips.