Adolescent Occupational Toxic Exposures: A National Study

Background  While many previous studies describe workplace-associated injuries in adolescents, few focus on toxic exposures. Such incidents are unlikely to be reported to either federal or state agencies. However, poison control centers often get called about these poisonings and might serve as a resource for monitoring their occurrence.

Objective  To describe the frequency and severity of job-related toxic exposures involving adolescents, the specific toxic agents involved, and trends over time.

Methods  Occupational toxic exposures occurring in the United States between 1993 and 1997 were analyzed using the Toxic Exposure Surveillance System database compiled by the American Association of Poison Control Centers. Contingency tables with the χ² statistic were used to test bivariate associations. Logistic regression was performed to investigate trends over time.

Results  Of 301 228 workplace toxic exposures reported over 5 years, 8779 (3%) involved adolescents younger than 18 years. The most common agents involved were alkaline corrosives (13.2%), gases and fumes (12.0%), cleaning agents (9.7%), bleaches (8.3%), drugs (7.4%), acids (7.2%), and hydrocarbons (6.9%). The injuries were rated as severe in 14.2% of exposures, life-threatening in 0.3%, and there were 2 deaths. The proportionate frequency of occupational exposures occurring among adolescents vs adults increased over time (odds ratio, 1.003; P<.001).

Conclusions  Adolescent occupational toxic exposures are an underrecognized hazard in the United States. Poison control center experience can be used to fill a gap in the surveillance of such injuries.

MORE THAN 5 million American children and adolescents are legally employed and another 1 to 2 million are employed in violation of provisions of the Fair Labor Standards Act (FLSA).¹ Child labor is resurgent in the United States because of a variety of factors, including increased immigration, the increased need for family income, a robust economy, employers' increased needs for unskilled help, the social acceptability of work among adolescents, and their desire for disposable income.^2-4 Working adolescents risk being injured while on the job, but previous studies do not provide much information about the working teenager's risk of toxic exposures.^5-9 Many adolescent jobs are not covered by workers' compensation and many poisonings occurring in small, unmonitored businesses go unreported to the Occupational Safety and Health Administration. Blanc et al¹⁰ have estimated that the incidence of occupational illness in the United States may be 3 to 5 times higher than that captured from these incomplete sources. Thus, the importance of adolescent occupational toxic exposures may go unappreciated because of the lack of a source of data.

Poison control centers (PCCs) have spread nationwide over the past 45 years, and have been aggregating their data on poisoning annually through the American Association of Poison Control Centers (AAPCC) since the 1980s.¹¹ They collect data on product safety and perform surveillance activities with relevance to occupational toxic exposures.^12-15 A pilot study in Massachusetts revealed that 269 adolescent occupational toxic exposures were reported to a single PCC over a 6-year period.¹⁶ The objective of the current study was to examine whether national PCC data could be used to describe the types, severity, and trends over time of adolescent workplace toxic exposures.

A secondary analysis of 1993 to 1997 Toxic Exposure Surveillance System (TESS) data maintained by the AAPCC was performed. Some calls previously reported in the pilot study were included here as well.¹⁶ The number of other PCCs participating in the TESS database during the period 1993 to 1997 ranged from 64 to 67 (covering 70.0%-93.5% of the US population).^17-21 Of 75 PCCs who submitted data in at least 1 of the 5 years under study, 58 centers (77%) reported data from all 5 years, with year-to-year consistency in the number of cases reported.

Occupational toxic exposures in the United States were identified 2 ways: by the site of the incident (the workplace) and by the circumstances given by the caller as a reason for the exposure (occupationally related). Calls to a PCC to obtain only information, without evidence of a human toxic exposure, were excluded. The analysis reported here was limited to records concerning children younger than 17 years, although the coincident validation study included cases outside of this age range.

A guide of rules and protocols for coding decisions is uniformly used by PCCs to reduce variation in the coding of data.²² The database included the following variables: date and time of call, PCC identity, locations of the exposure incident and the caller, age and sex of the victim, circumstances, toxin(s) involved, route of exposure, triage, disposition, and outcome. For certain analyses, toxins implicated in the exposure were classified into 12 broad descriptive categories: acids, alkali, bleaches, cleaners (not including bleach), drugs, gases, hydrocarbons, miscellaneous chemicals, pesticides, plants, soaps and detergents, and others. In this determination, only the first agent entered into the record (if there was more than 1 agent involved) was counted per incident. For analysis of severity, outcomes coded as moderate or major injurious effects were collapsed into 1 group ("major severity"), whereas those coded as no or minimal effect were collapsed into another ("minor severity").

Data from the original medical records were compared with those previously entered into the TESS database. A sample of 900 medical records was ascertained by a computer-generated random algorithm. Original paper medical records solicited from participating PCC were redacted according to written AAPCC guidelines²² by a trained, blinded research assistant (A.G.). Eight relevant variables were compared between the electronic database and the redacted paper record: agent type, age, exposure site, caller site, reason for the call, route of exposure, duration of effect, and outcome. The extent of correlation was assessed using the κ statistic. Thirty-five (45%) of 72 PCCs sent 453 records (50%) for analysis. Of these, 31 had missing data; 422 cases (93.2%) were analyzed. The κ was greater than 0.80 in 14 of 15 separate comparisons. Only the variable "outcome" showed agreement of less than 80% and a κ of less than 0.67.

Contingency table analyses with the χ² statistic were used to test bivariate associations. A 2-tailed α was set at .05 to establish significance, but corrected to .01 for analyses with multiple comparisons. Logistic regression was performed to investigate binomial trends of proportions over time and expressed as an odds ratio (OR) of change per month. Multiple logistic regression analysis was used to compare the severity of outcome by month, day, time of day, age, sex, and route. STATA version 6 for microcomputers (College Station, Tex) was used for data analysis.

This study was approved by the committee on clinical investigation at Children's Hospital, Boston, Mass. It was also authorized by the board of directors of the AAPCC.

Figure 1 shows the flow of the data assessed in this study. There were 301 460 occupational exposure cases reported to TESS in 1993 through 1997, of which 232 cases were invalid or duplicated, leaving 301 228 cases. There were 17 354 cases excluded from the study because of incomplete age information. There were 275 095 adult cases outside the inclusive age criteria; this left 8779 poisonings (3.1% of cases in which the age was known) among children aged 12 to 17 years, which formed the basis for further analysis. The percentage of exposure reports attributable to adolescents by month is shown in Figure 2. This percentage increased over time (OR, 1.003; P<.001) and was greatest in the summer months.

Figure 3 shows the number of cases reported to poison control centers, by sex and year, for 8758 cases (in 21 cases, sex was not specified). Males consistently predominated, accounting for 63.9% of the exposure incidents. The distribution of cases by age and sex are shown in Figure 4. The median age of the sample was 16 years; however, 2093 cases (23.8%) involved adolescents younger than 16 years and 287 cases (3.3%) involved 12-year-olds.

Adolescents were most likely to suffer toxic exposures during the summer months (38.2% of the total occurred in June, July, and August). Exposures were reported during weekends as frequently as during weekdays; 25.0% occurred between 6 and 9 PM; another 20.7% occurred between 3 and 6 PM. The initial call to the PCC was made from the home in 41.7% of cases, from the workplace in 28.3%, and from a health care facility in 23.3%.

Table 1 gives the frequencies of the 12 toxic agent categories most often involved in these exposures, by sex. The most common specific categories implicated in adolescent workplace exposures were to alkaline corrosives (13.2%), gases and fumes (12.0%), cleaning agents (9.7%), bleaches (8.3%), drugs (7.4%), acids (7.2%), and hydrocarbons (6.9%). A second toxic agent was implicated in the exposure in 862 (9.8%) cases. Females were more likely to suffer toxic exposures involving medications, gases, bleach, and cleaners, and were less likely to suffer toxic occurrences involving hydrocarbons, acids, pesticides, and plants.

Of the 8779 cases, 34.9% involved toxic inhalations, 26.8% ocular exposures, and 23.6% skin exposures. There were 19.3% ingestions, and 2.4% exposures involved bites or stings. In 1.7% of incidents, the toxin was injected, aspirated, or involved an unclear route. More than 1 route of exposure (eg, a chemical splashed both in the eyes and on the skin) was reported in 7.7% of cases. As shown in Figure 5, ocular splashes and dermal exposures were a common route of exposure for many toxins encountered by adolescents in the workplace. More than 71% of the exposures to alkaline agents, 71% of exposures to cleaners, 71% of exposures to soaps and detergents, and 55% of exposures to acids involved dermal or ocular splashes. Exposures to bleach included a number of inhalations (39%) and ocular contacts (41%), with a smaller number of dermal exposures (11.5%) and ingestions (15.3%).

In 7176 of the incidents (81.7%), adolescents reported that they developed some symptoms as a result of the exposure incident. Of those cases in which clinical symptoms were reported, the duration of the effects lasted less than 2 hours in 34.3% of cases and 1 day or less in 80.9%. In 5.6% of cases the symptoms reported were judged to be unrelated to a toxic exposure.

Ocular or dermal irritations were the most commonly reported symptoms (25.9% and 12.5% of all reported symptoms, respectively). Other dermal symptoms were also common, including erythema (7.2%), burns (7.1%), and swelling or edema (2.6%). Respiratory or throat symptoms were also common, with 9.5% of adolescents reporting coughing or choking, 6.7% reporting throat irritation, 4.7% reporting dyspnea, and 2.6% reporting chest pain. Gastrointestinal complaints included nausea (11.6%), vomiting (7.3%), and abdominal pain (2.8%). Headache (8.8%) and dizziness (6.0%) were also frequently reported.

Therapies were offered in 81.1% of all cases. These included dilution or irrigation (57.8%), fresh air (19.8%), oxygen (3.2%), antihistamines (1.2%), bronchodilators (1.3%), intravenous fluids (1.2%), and other therapies (19.2%).

As shown in Figure 1, 49.9% of these toxic exposures were treated on site (either at the workplace or at home). Another 36.6% of patients were treated and released from the emergency department of a health care facility; less than 2.2% required admission to the hospital. Another 11.4% of cases were classified as other, which includes those who left the clinic or emergency department without being seen or against medical advice and those who were lost to follow-up.

The outcomes in 8.4% of the cases were coded as resulting in no toxic effect or were deemed to be nontoxic incidents by PCC staff. In 65.8% of cases, only minor injurious effects were the reported or expected outcome of the incident. However, in 13.5% of cases, the toxic exposures resulted in injuries coded as of major severity, with another 6.6% coded as having potential for toxicity, but the poison center staff were unable to contact the patient for follow-up information.

Severity varied by the toxic agent involved (Figure 6). The highest percentage of cases with moderate or severe injury outcomes involved alkaline products (27.6% of the total), acids (20.7%), glues and pastes (19.0%), herbicides (17.6%), gases and fumes (17.4%), cleaners (16.8%), pesticides (15.6%), miscellaneous chemicals (15.4%), soaps and detergents (15.4%), and bleach (13.6%). Many of the more severe injuries from workplace exposures among adolescents involved dermal or ocular splashes, or inhalation-related injuries.

Proportionate severity of injury from adolescent workplace toxic exposures did not vary significantly by either season or year. However 16.4% of the males with toxic exposures suffered severe injuries vs 13.9% of females (χ² = 8.87; P = .003). A multivariate analysis of associations between selected variables and severity revealed no association between severity and the month, day, or year of occurrence. However, a dermal route of exposure (OR, 1.17; 95% confidence interval, 1.01-1.34); the time of day (between midnight and noon) that the incident took place (OR, 1.35; 95% confidence interval, 1.17-1.55); male sex (OR, 1.19; 95% confidence interval, 1.04-1.36); and age (OR, 1.08; 95% confidence interval, 1.03-1.14) all predicted severe outcomes.

There were 2 reported deaths, both of which involved exposures to caustic agents. A 13-year-old boy was taken to the emergency department after inhaling sulfuric acid fumes while cleaning out a pipe at work. He experienced respiratory arrest, coma, and cardiac arrest, and did not respond to cardiopulmonary resuscitation. This case was reported previously in abstract form.²⁰ The other involved a 16-year-old boy who swallowed an alkaline corrosive agent that he thought was a beverage while he was working at a construction site. The potassium hydroxide mixture was in a milk jug that he obtained from the back of a truck. He vomited, aspirated, and developed respiratory arrest. He also suffered oral and esophageal burns, chest and abdominal pain, pneumonia, and suffered a cardiac arrest. Despite intensive medical support, including artificial ventilation and vasopressors, the patient died.

This study supports a role for PCC in conducting surveillance of occupational toxic exposures among adolescents. The percentage of workplace toxic exposures occurring among adolescents reported here (3.1%) is similar to that reported previously (3.8%) in the pilot study conducted by the Massachusetts PCC.¹⁶ The seasonality we found is easily understood, with more adolescents working, and thus more likely to be injured from a toxic exposure, during the summer months than during the school year. This study found an increasing trend nationally in the relative frequency of such toxic exposures among adolescents between 1993 and 1997. There could be many reasons for this: it could be that more adolescents are gainfully employed in a tight labor market, giving them more opportunities to become victims of a toxic exposure. It could also be that the same tight market is causing employers to ask teenagers to perform different tasks on the job, involving chemicals more frequently, because they do not have enough experienced adults who otherwise would do such work. An increase in frequency could also be explained by improved familiarity of the population with the services offered by PCCs, such that families and health care professionals are using them more and reporting toxic exposures more often than in past years. It is also possible that the use of PCCs for reporting adult occupational exposures has, for unknown reasons, decreased over time.

The risk factors associated with agents of injury and both the physical and social environments to which adolescents are exposed in the workplace are incompletely understood.^23,24 According to Baker et al,²⁵ the first step in ascertaining determinants of risk should be to implement surveillance (eg, using poison control center data) so that we can then actively intervene to prevent these conditions. For example, the male predominance in the current study could be explained by their greater participation in part-time employment, although estimates of adolescent employment probably undercount participation by both males and females, making it difficult to speculate about employment rates between the sexes. The observation that older males working late at night or early in the morning were overrepresented in those toxic exposures resulting in severe injuries also suggests differences in the categories and types of jobs in which they are employed. Since circumstances and identification of the workplaces involved in these reports were not specified in the data, such differences are speculative at best. However, if males are more likely to be employed in construction, automotive, agriculture and landscaping work, and painting activities, then this could account for their exposure to more potent toxins such as caustics, hydrocarbon solvents, pesticides, herbicides, heavy metals, paints, and thinners.

While many of the 8779 exposures recorded were medically trivial, more than a third required medical attention in a health care facility and 2% of patients with such exposures were hospitalized. There were only 2 deaths reported in this study; however, there are other anecdotal reports^26-28 and systematic studies^29,30 of adolescent workplace injury deaths that include worksite toxic exposures. Previous studies have documented the incompleteness of PCC fatality reporting³¹ and the deaths reported here are likely to underestimate their true incidence. Our findings suggest that an important expenditure of scarce health care resources is directed toward these injuries; there are also the human costs of suffering or even death. Our results confirm the need for efforts directed to the prevention of adolescent occupational toxic exposures as a public health issue. The most frequently involved toxic agents were also in many cases those accounting for the highest percentage of severe injuries. Incidents involving caustics (acids and alkalis), gases, cleaners, bleaches, pesticides, and soaps and detergents deserve the particular attention of public health officials, because of their frequent use by adolescents, and because precautions can be taken to prevent such mishaps involving such chemicals. For example, many of the poisoning incidents described in this report involved dermal and/or ocular splashes, preventable by the appropriate use of barrier clothing, gloves, and goggles.

There are limitations to this study, which dictate that the results should be interpreted with caution. The number of PCCs contributing data increased slightly during 1993 through 1997 (64-67 reporting centers) although the population covered increased from 70% in 1993 to more than 90% of the United States by 1997. This may account for some of the increase in the absolute number of adolescent exposures over time, but does not explain their increasing percentage relative to those of adults.

As mentioned previously, while cases are labeled by exposure site at the time of the original telephone call, currently, no element in TESS specifies the occupation involved or the type of work effort undertaken at the time of the injury. Bresnitz et al^13,14 have previously pointed out the inadequacies of documentation and the questionable quality of recommendations of PCC staff regarding their management of occupational toxicity cases. We can only speculate here on the work activities that resulted in the toxic exposure; further research into the type of work establishment and the type of activity leading to an adolescent's toxic injury is necessary.

Coding errors could affect the accuracy of the data; however, our validity study found good concordance between data recorded in the original PCC record and that in the TESS file. There was a lower correlation only for the outcome variable; much of the discrepancy here seemed to be between labeling a case as either no effect or a minor effect on the patient vs an unknown effect. These mismatches point out the ambiguity in coding choices at this level of precision. While clear rules are given in the coding handbook furnished by the AAPCC, how they are applied may differ from center to center and specialist to specialist. However, marginal choices—the difference between the outcome of "minor effect" and "unknown effect, probably nontoxic"—are of negligible clinical importance. This analysis only considered the PCC record; no review of the emergency department's medical record of the poisoning was attempted. Thus, the study could not take into account those incidents in which erroneous information is given to and recorded by PCC staff.

Calls to PCC are acceptable as evidence of the standard of medical care administered to a poisoned patient. Nevertheless, such calls are entirely voluntary; TESS reporting is a passive surveillance system and there is no requirement to report toxic exposures to the AAPCC, so that results reported here almost certainly underestimate the frequency of these injuries. Health care providers may not call a PCC for many reasons: if they know how to treat the patient, if they fail to diagnose or misdiagnose the patient's problems, or if the injury from the toxic exposure is judged by them to be relatively minor. Poison control centers are more likely to be consulted for perceived emergencies and acute poisonings than for subacute or chronic exposures presenting with more indolent symptoms. Toxic exposures to carcinogens, teratogens, fertility-lowering chemicals, or other agents whose adverse effects are only apparent after long latency periods will likely never come to the attention of a PCC. Adolescents who are exposed to toxic chemicals may not attribute their symptoms to the exposure. They, their families, or their physicians may misdiagnose the poisoning as some other ailment or simply may not view the PCC as a resource to consult after a workplace toxic exposure. For all these reasons, PCC calls undercount the true incidence of workplace toxic exposures.

The Institute of Medicine has called for the investigation of "ambulatory services" databases to improve the monitoring of adolescent occupational injuries.³² This study suggests that occupational exposure surveillance activities of PCC can be useful to policymakers by providing information on sentinel occupational toxic events involving adolescents. By determining the effectiveness of PCC operations as a mechanism for surveillance of adolescent occupational exposures, this study informs the national debate concerning the vulnerability of children in the workplace to toxic injury and gives evidence-based arguments being developed in the pursuit of national consensus-building. Adolescent occupational exposures to toxins are an important but underrecognized injury category; PCC data can be used to fill in gaps in surveillance for this type of worksite-associated injury.

Accepted for publication December 8, 2000.

Supported in part by a grant from the National Institute of Occupational Safety and Health, Washington, DC, and grant R03 OH03796-01 from the National Occupational Research Agenda (Dr Woolf). Dr Woolf's work is supported in part by funds through a cooperative agreement (U50/ATU300014) with the Agency for Toxic Substances and Disease Registry, Public Health Service, US Department of Health and Human Services, Washington, DC.

Presented as a poster at the meeting of the National Congress of the American Association of Clinical Toxicologists, La Jolla, Calif, October 1-5, 1999, and in part at the annual meeting of the American Academy of Pediatrics, Chicago, Ill, October 31, 2000.

We thank poison control centers for their participation in providing original medical records used to validate the Toxic Exposure Surveillance System dataset.

Reprints: Alan Woolf, MD, MPH, IC Smith Building, Children's Hospital, Regional Poison Control and Prevention Center, 300 Longwood Ave, Boston, MA 02115 (e-mail: woolf@a1.tch.harvard.edu).