Records from the Swedish poisons information centre as a means for surveillance of occupational accidents and incidents with chemicals

Safety Science 104 (2018) 269–275

Contents lists available at ScienceDirect

Safety Science journal homepage: www.elsevier.com/locate/safety

Records from the Swedish poisons information centre as a means for surveillance of occupational accidents and incidents with chemicals

T

⁎

Linda Schenka, , Karin Feychtingb, Anita Annasb, Mattias Öberga,c a b c

Unit of Work Environment Toxicology, Institute of Environmental Medicine, Karolinska Institutet, P.O. Box 210, 17177 Stockholm, Sweden The Swedish Poisons Information Centre, 171 76 Stockholm, Sweden The Swedish Toxicology Sciences Research Center (Swetox), Forskargatan 20, 151 36 Södertälje, Sweden

A B S T R A C T We present a retrospective analysis of records on occupational accidents from the Swedish Poisons Information Centre. The aim was to explore these data as a means for surveillance of accidents and incidents with chemicals at the workplace. We extracted data on all telephone consultations regarding occupational incidents (n = 8240) during 2010–2014. One third of the calls were made by health care staff (31%) and two thirds were made by the public (69%). For the latter group, about half (54%) received advice on how to manage on site. One out of five workplace incidents were assessed by the operating expert (pharmacists and physicians) as a major risk for severe symptoms. The three most commonly reported chemical groups were alkali (n = 1510, excluding ammonia), hydrocarbons (n = 1129, including halogenated hydrocarbons) and acids (n = 984). Eye exposure was the most common exposure route recorded (n = 3049), followed by inhalation (n = 2635) and skin (n = 1438). Data from the Swedish Poisons Information Centre offers insights about occupational accidents and incidents with chemical products and also include a higher number of accidents in absolute numbers as compared with the official injury statistics. With a clear focus on type of poisoning agent, treatment and health effects, poisons information data may serve as a means for surveillance on chemical incidents at the workplace.

1. Introduction Occupational use of chemical substance is the origin of a wide variety of occupational injuries, and statistics on occupational diseases and accidents are part of the foundation for identifying proper risk management efforts. In Sweden, the Swedish Work Environment Authority (SWEA) is responsible for occupational health and safety and also compiles the Swedish statistics on occupational diseases and accidents reported by the employers. However, despite obligation to report accidents and severe near-accidents (Swedish Parliament, 1977), it is highly likely that occupational disease and injury statistics underestimate the incidence of chemical incidents and accidents and injuries. Employers may lack awareness of reporting requirements, or may perceive other disincentives for instance that reporting would be incriminating or too time-consuming. Organisational factors such as safety incentive programs that penalise accidents and injuries may also disincentive employees from reporting injuries to their employers (e.g. US GAO, 2012). The incompleteness of occupational injury and disease statistics is a well-known issue also in other countries than Sweden (e.g. Gravseth et al., 2003; Leigh et al., 2004; Walters et al., 2011; Probst

⁎

et al., 2013). In addition, the Swedish reporting requirements are limited to comparatively severe accidents, i.e. accidents that led to sickleave or incidents that could have led to very severe consequences. This incomplete decision-basis may reduce efficiency of SWEA’s risk management measures. Another source of data on occupational related chemical accidents is the Poisons Information Centers (PICs). PICs provide information on risks, symptoms and treatments in case of acute poisonings with different agents, e.g. chemical products, pharmaceuticals and biological toxins. In Sweden there is only one PIC unit, established in 1960, that serves the whole country (population of 9.8 million inhabitants). The 24 h phone service is manned by pharmacists and physicians specialized in intensive care. The service is open to medical professionals as well as the general public and all telephone consultations are logged in the PIC database. The Swedish PIC is connected to the emergency response number. There is also a general Swedish health care information number, from which all calls concerning potential poisonings, including eye and dermal exposure to chemicals, are forwarded to the PIC. Furthermore, the Swedish PIC is appointed the Swedish responsible body for receiving information relating to emergency health response

Corresponding author. E-mail address: [email protected] (L. Schenk).

http://dx.doi.org/10.1016/j.ssci.2017.10.021 Received 13 January 2017; Accepted 29 October 2017 0925-7535/ © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/).

Safety Science 104 (2018) 269–275

L. Schenk et al.

3. Results

under Article 45 of the EU Regulation (EC) 1272/2008 on classification, labelling and packaging (CLP) and hence administers a database of chemical products on the Swedish market. Through the database the PIC can advise e.g. physicians on the nature of exposure based on product name. In addition, the PIC is the Swedish International Chemical Environment (ICE) Centre, i.e. the Swedish part of a European network of emergency response centers for chemical accidents, initiated and financed by the chemical industry. The Swedish PIC is thus a central component in several information chains concerning chemical agents and products. In their function as a source of information for the public, workplaces and health care personnel, PICs receive a substantial number of calls regarding workplace exposures (Litovitz et al., 1993) for which the overlap with other established occupational surveillance systems has been found low (Blanc and Olsen, 1986). There are several examples of the use of PIC data to investigate exposures to specific product groups such as pharmaceuticals (e.g. Smith et al., 2008) or pesticides (Olson et al., 1991; Meulenbelt and de Vries, 1997; Sudakin and Power, 2007) but also to specific population groups such as workers in small enterprises (Hinnen et al., 1994) or adolescents in the work-force (Woolf et al., 2001). In the present work we explore how the data from the Swedish PIC may provide knowledge about accidents and incidents with chemicals at the workplace.

From the 9266 telephone consultations concerning occupational exposures during 2010–2014, we extracted information on 8240 separate occupational incidents (Table 2). The number of occupational incidents for which PIC has been contacted each year has increased by 30% from 2010 (n = 1411) to 2014 (n = 1831), with the largest difference occurring between 2011 and 2012 (Table 2). Dividing these numbers by the total number of employed individuals in Sweden shows that the increase seen in PIC records is larger than the simultaneous increase in employment (Fig. 1). In 2010 the PIC records identify 3.2 accidents per 10,000 employed and year (n.b. Fig. 1 shows the corresponding numbers per month), for 2014 the corresponding number is 3.9 per 10,000, i.e. an increase of 21%. However, it is the number of calls from the public that have increased over the years (in total a 43% increase from 2010 to 2014). The number of calls from health care actors has been around 500 incidents per year (range 484–546, Fig. 2). During the same period, the number of reported accidents caused by hazardous materials (n.b. includes not only poisonings but also heat burns etc.) that resulted in sick leave in the official Swedish statistics has increased 80% from 2010 to 2014 (Table 3). A majority of first calls are made by the public (n = 5697), while about one-third of first calls are from a health care actor (Table 2). One out of five incidents were classified as potentially posing a major risk to the exposed individual (Table 2). Among calls from the public, i.e. exposed individuals not yet in health care 54% received advice on how to manage on site (Table 2). Generally, cases where the PIC expert judged the exposure to pose minor risk were manageable on site, e.g. diluted acid on small part of skin that was advised to rinse exposed skin with water for at least five minutes. Cases judged to pose a major risk were all advised to seek health care, e.g. alkaline products in the eye, for which immediate health care was advised in addition to continuous rinsing of the eye. The advice regarding health care varied for the category moderate risk depending on the nature of the exposure, health care was either advised immediately or in case symptoms worsen or persist within a certain time-frame. Among the 4381 cases judged to pose a moderate risk, 30% were already in health care (indicated by caller), another 31% received the advice to seek health care for examination and/or treatment. Some differences based on gender are discernible from the material presented in Fig. 1 and Tables 2 and 3. Of the recorded incidents 68% concern male workers (n = 5638) and 30% concern female workers (n = 2440). In 2% of cases (n = 162) the gender was not recorded. Participation on the labour market is comparatively equal between genders in Sweden, hence as with the absolute numbers presented in Table 2, the number of accidents per 100,000 employed women and month is around half that of men (on average 1.9 and 3.9, respectively). At first sight, the gender distribution of our material is similar to that of the number of reported accidents with hazardous materials that resulted in sick leave in the official Swedish statistics (on average 31% of injured are women, Table 3). However, it should be noted that almost all accidents leading to any sick leave probably would be judged as sufficiently severe to be categorized as a major risk by the PIC experts. In our material, 24% of incidents judged to pose a major risk concerned women. Calls concerning women are hence more likely to have been judged as minor risk incidents while cases involving male patients were more likely to be judged as major risks (chi-square statistic is 63.8, df = 3 and the P-value is < .00001, unknown gender and confirmed severe outcomes removed, n = 8071). Among the seven cases verified as severe six concerned men. Among calls concerning female patients not yet in health care (n = 1958), 58% received the recommendation to manage on site. For male patients not yet in health care (n = 3671) the corresponding number was 52%. Furthermore, calls concerning patients in health care more often concern male patients than female patients (77% of calls from health care actors

2. Methods We present a retrospective review of five years (2010–2014) of Swedish PIC records. Furthermore, official statistics on the number of employed men and women in Sweden per month during the investigated period were collected from Statistics Sweden (SCB, 2016). Official statistics on occupational accidents were compiled from the SWEA’s online database over occupational accidents (ISA, 2016). These data are reported by employers for accidents where the employee has been, or is expected to be, on sick-leave for at least one day. The data were analyzed using cross-tabulation and descriptive statistics. Ethical vetting was applied for and approved by the regional ethical board in Stockholm. 2.1. Compilation of PIC data All PIC database entries categorized as occupational accidents involving adults (due to PIC categories 18 and 19 year olds were also excluded) logged between January 1st 2010 and December 31st 2014 were extracted (n = 9266). The extracted information was imported to a spreadsheet format, cleaned and recoded when necessary for the purpose of the present study. The resulting database is outlined in Table 1. Repeated calls about the same incident, identified using the PIC cross-referencing, were consolidated into the same incident entry. Calls that are connected to incidents where the emergency care personnel asked precautionary questions about their or other patients’ safety (for instance need for decontamination of the patient before entering the emergency room) were also consolidated into the primary incident. The PIC recording of poisoning agents mainly relates to area of use (e.g. pool chemical, cleaning agent) and, if known, complemented with some specific chemical substances/groups (e.g. hypochlorite, alkali). Gases are generally logged depending on their health effects (e.g. irritant, systemic toxicity, inert). More details of the poisoning agent might also be specified in the free-text fields. Depending on the information provided by the callers and the PIC expert’s interpretation the poisoning agent in a specific case might be logged differently in a second or third call, making the recoding of the poisoning agents in the cases necessary. For the purpose of the present study we categorized the poisoning agents according to chemical groups. When available, detailed information about the chemical identity was collected from the free-text fields. 270

Safety Science 104 (2018) 269–275

L. Schenk et al.

Table 1 Description of PIC database content extracted for the purpose of the present study. Variable

Categories

Date and time

–

a

Caller

• Ambulance physician • Hospital, other • Hospital, care, physician • Primary care, other • Primary • Public • Eye • Dermal • Inhalation • Oral • Several • Other agent, alkali • Cleaning fume • Metal • Disinfectant, ethanol/isopropanol • Female • Male • Unknown b

Exposure routea

c

Poisoning agenta

Gender of exposeda

d

Caller’s question PIC response Risk judgment (based on described exposures and symptoms) or confirmed severitye

Free text field Free text field

risk: No or mild symptoms, possible to manage on site, health care recommended if symptoms persist • Minor or occur later. risk: Pronounced or prolonged symptoms might occur, immediate health care may be recommended. • Moderate risk: Risk for severe symptoms, immediate health care recommended. • Major possible to evaluate risk: e.g. due to limited information about symptoms and/or exposures. • Not Confirmed severe outcome : Only assigned to patients diagnosed by physician for a severe or life-threatening • outcome or in case a lethal outcome was confirmed. on site: advice on suitable first aid measures, seeking health care may be advised in case symptoms • Manage worsen or persist beyond a certain time point. health care: The exposed individual is advised to seek health care for further diagnosis and/or treatment. • Seek • Health care treatment advice: Advice to health care actors on how to diagnose and treat. f

Treatment advisea

a

Given examples do not include all possible categories included in original PIC data. Includes both occupationally exposed individuals that call the PIC themselves, or calls made on behalf of an occupationally exposed individual by a colleague or family member. Encompasses e.g. injection, high-pressure injection, nasal exposures. d Unknown gender also includes incidents where several persons were exposed and more than one gender was represented. e These category labels have been translated by the authors, they represent the estimated worst case degree of risk. f Corresponding to Poisoning Severity Score (PSS) Severe (3) and Fatal (4) (Persson et al., 1998). b c

potential corrosives were generally judged to be of minor or moderate risk and hence mainly manageable on site. Least likely to pose major risks were exposures to alcohols and glycols (1% of which were judged as major risk incidents), fibers, dusts and particles (2% of which major risk), pharmaceuticals (2% of which major risk), and surfactants (3% of which major risk) (Table 4). A comparably larger share of calls concerning exposure of the eyes or several routes were judged as major risk cases (32% and 27%, respectively) compared to the overall (21%). Only 8% of the inhalation exposures were judged as posing a major risk. There is a clear difference between chemical groups with respect to whether more than one call to the PIC was made for the same incident. Only 4% of incidents with pharmaceuticals are multiple call incidents (n = 11 of which 1 also was judged as a major risk incident), while 27% of incidents with hydrofluoric acid are multiple call incidents. Multiple calls regarding the same incident might indicate an increased risk posed by the exposure, but could also reflect exposure to chemicals that are less familiar to the health care personnel. Exposures whose main health hazards are systemic toxicity, especially with potential lethal outcome such as hydrogen sulphide, hydrogen fluoride and cyanogenic substances, are highly represented in the multiple calls column. Since these exposures are less common as eye exposures or injections (which are the main category of other routes) chemical identity would also explain a part of the higher number of multiple calls for inhalation, skin and several route exposures.

concern male patients and 19% female patients). Another indicator of complications is the number of calls made per incident, where male patients figured in three out of four cases for which more than one call was made to the PIC (Table 2). The nature of exposures (i.e. poisonous agent and route of exposure) recorded in our material is described in Table 4. Different forms of alkali are the most commonly implicated exposure agents, hydrocarbons and halogenated hydrocarbons also constitute a major group, followed by various acid substances and products (Table 4). The most common exposure route recorded in our material is exposure of the eyes (37%, n = 3049), followed by inhalation (32%, n = 2635) and skin (17%, n = 1438). In about 6% of cases the Swedish PIC records indicate several exposure routes (Table 4). In about one fifth of the cases the PIC expert judged the exposure to constitute a major risk at the time of the first call (n = 1736, Tables 2 and 4). In addition, our data includes seven incidents with a confirmed severe outcome (Table 2). More specifically these cases involved five exposures to alkali, one exposure to hydrogen sulphide, and one exposure to hydraulic oil As with the confirmed severe outcomes, most major risk incidents involve exposure to alkali. Also other potentially corrosive substances such as quaternary ammonium compounds and acids are frequently implicated in major risk incidents. More than one third (35%) of exposures to acids were judged to pose a major risk, and for alkali more than half (51%) of the exposures were judged to pose a major risk. Many of the commonly implicated substance groups besides 271

Safety Science 104 (2018) 269–275

L. Schenk et al.

Table 2 Overview of occupational incidents recorded by the Swedish PIC during the years 2010–2014. Numbers are presented for female and male patients, respectively, and in total.

Table 3 Number of accidents per year recorded by the official statistics on accidents with hazardous materialsa leading to sick leave (ISA, 2016). Female

Male

Total

63 125 135 137 156

194 204 315 309 313

257 329 450 446 469

Number Female

Male

Totala (%)

All valid cases

2440

5638

8240 (1 0 0)

Cases per year 2010 2011 2012 2013 2014

387 412 567 510 564

990 1068 1180 1165 1235

1411 1502 1788 1708 1831

Number of calls per case 1 2–3 4 or more (max 7)

2241 198 1

5026 585 27

7402 (90) 809 (10) 29 (0.3)

Caller (first call) Public of which manage on siteb Primary care, physician Primary care, other Hospital, physician Hospital, other Ambulance

1958 1133 65 112 198 81 26

3671 1904 255 321 945 312 134

5697 (69) 3077 329 (4) 450 (5) 1183 (14) 406 (5) 175 (2)

Risk/severity judgment Minor risk Moderate risk Major risk Undetermined Confirmed severe outcomec

676 1275 417 71 1

1161 3029 1278 164 6

1868 (23) 4381 (53) 1736 (21) 248 (3) 7 (0.08)

2010 2011 2012 2013 2014

(17) (18) (22) (21) (22)

a These are accidents caused by dusts, fumes, gases or liquids, and may also include non-poisoning outcomes (e.g. heat/cold burns).

4. Discussion We have presented an overview of telephone consultations performed by the Swedish PIC 2010–2014 concerning 8240 occupational incidents involving adults. These make up about 2% of all telephone consultations performed by the Swedish PIC in the same time period (GIC, 2011, 2012, 2013, 2014, 2015). Nevertheless, this database is the largest collection of data on current occupational incidents with chemicals in Sweden. During the investigated period the occupational incidents identifiable through PIC telephone consultations have increased from about 1400 to about 1800 per year. This increase (30%) is larger than the corresponding increase in number of employed during the same period (Fig. 1). A closer inspection shows that the increase is almost completely due to an increased number of calls from the public and that the number of calls from health care staff has been relatively stable. The increase in calls concerning occupational incidents may not necessarily be interpreted as solely due to an increase in the number of incidents. Recovery from the economic crisis in 2008 and the accompanying increase in employment probably plays some part. However, the Swedish work environment surveys for the period 2009–2013 indicate that there has been little to no increase of exposure to hazardous materials among the Swedish workforce, albeit only measured as self-reported exposure during more than 25% of the working time (SWEA, 2014). As the increase seen in our material is made up by calls from the public, rather than health care staff, increased awareness of the PIC among the public may play a larger role. This would also be supported by the overall increase in calls the Swedish PIC has experienced in the same period of time, the yearly number of telephone consultations increased with 12% from about 77,000 consultations to about 86,000 (GIC, 2011, 2015). The precautionary statements introduced by the CLP regulation may have promoted awareness of the PIC1. Notably, there are three statements that direct exposed individuals specifically towards a PIC: “Immediately call a POISON CENTER/doctor/…” (P310), “P311 Call a POISON CENTER/doctor/…” (P311) and, “Call a POISON CENTER/ doctor/… if you feel unwell” (P312). On the other hand, the increase in the SWEA data on accidents with hazardous materials leading to sick leave has increased 80% over the investigated time period (Fig. 2), which points towards a deteriorating work environment. However, this increase of accidents reported to SWEA is difficult to interpret. One factor may be the introduction of an online reporting form that facilitates the reporting procedure for employers and also coordinates with the reporting system to the social insurance, launched in November 2011. Assuming each incident identified in the Swedish PIC database concerns one separate individual, the numbers presented herein roughly translates to 3 person incidents per 100,000 employed and month. This number should be interpreted with caution as it does not consider that the differentiation of the workforce into subgroups more

a

Total includes also cases where gender has not been recorded (n = 162). The number of public callers that at time for first call were not judged to need immediate health care and were given advice on how to manage on site. c Among confirmed severe cases one was recorded as a fatality.

Incident rate

b

Number of incidents

Fig. 1. Number of incidents identified by PIC data divided per number of employed individuals in Sweden per month (collected from SCB, 2016).

2000 1800 1600 1400 1200 1000 800 600 400 200 0

Public Health care

Fig. 2. Number of total occupational incidents identified in the present study and whether first call came from the public or from actors within health care. For comparison the number of accidents recorded by the official statistics on accidents with hazardous materials (ISA hazmat accidents) are indicated as a line.

1 According to CLP, a substance or mixture classified as hazardous and contained in packaging shall bear a label including amongst other elements applicable precautionary statements.

272

Safety Science 104 (2018) 269–275

L. Schenk et al.

Table 4 Overview of exposures included in our database and the percentage of row and column totals that were judged as major risk incidents or were incidents with confirmed severe outcomes.

a

Exposure agents are often mixtures and are at time of call categorized according to the chemical ingredient judged to pose the largest risk. b Category totals (bold font) also include subcategories (normal font) not listed. c Other routes include injection, high-pressure injection, nasal and unknown. d Butane and propane are listed under fumes, gases and vapours and not hydrocarbons. e Two instances of liquid ammonia and one where dry ice was swallowed. f Includes also liquid ammonia (n = 67) for which 43% were judged as major risk incidents (for ammonia gas it is 10%). g Major toxic metals include arsenic, beryllium, cadmium, chromium (VI), lead, mercury and their compounds. h Pesticides in classes 1 and 2 may only be used for professional purposes.

corresponding exposure patterns as participation in the workforce is comparatively equal in Sweden (Fig. 1). For instance, the Swedish working conditions survey found that, with the exception of cleaning agents, men more often than women report that they are exposed to chemical agents at least 25% of their working time (SWEA, 2014). Specific examples from the survey performed in 2013 are cutting fluids (9% of men, 2% of women), air pollutants (26% of men, 17% of women) and chemicals (10% of men, 5% of women). For potentially corrosive acids and alkali the numbers were relatively equal (4% of men, 3% of women) and as mentioned for cleaning agents women more often reported exposure more than 25% of their working time (8% of men, 19% of women) (SWEA, 2014). However, there are also multiple behavioural components that may contribute to these noted gender differences. Both with regards to the exposed individuals’ safety behaviour, propensity to call for additional information, or how the exposed person’s need for health care is perceived (Loikas et al., 2015; Herlitz et al., 2009). However, the present study does not offer the sufficient resolution to investigate causal factors of the noted gender differences.

likely to be exposed or that for each incident more than one individual may have been exposed. However, this is a low number and, although significantly larger than corresponding numbers for the accidents identified by official statistics, most likely underestimation of chemical accidents and incidents that take place in Sweden. Not all exposed individuals, nor all treating health care staff, will call the PIC for assistance in case of chemical exposures as individuals may not be concerned about their exposure, lack of awareness of PIC or (believe they have) have sufficient knowledge on how to manage the exposure in question. Furthermore, when a call is made to the PIC, the incident will only be identified as of occupational origin if the caller reports it as such. An overrepresentation of men is seen in our material (70% of all cases with known gender), as well as in other sources of occupational accident statistics. Previous studies on PIC data from the US and Switzerland mostly report even larger proportions of men among the exposed (66–84%) (Blanc et al., 1989; Bresnitz, 1990; Litovitz et al., 1993; Hinnen et al., 1994). In this material the skewed gender distribution is most likely due to the structure of the labor market and 273

Safety Science 104 (2018) 269–275

L. Schenk et al.

protective equipment. Although personal identifying data would be most valuable from a surveillance perspective it may also be detrimental to PIC’s main task as some callers may wish to remain anonymous. The Swedish PIC does not routinely determine and record the medical outcome of a case, the judgment of the risk posed by an exposure is only based on the information presented at the call. Data recorded for each call may be incomplete, for instance due to interrupted calls, limited ability or willingness of the caller to provide information. However, these limitations apply also to other sources occupational injury data. Furthermore, as the PIC curates the database of chemical products PIC experts are generally able to characterize the chemical hazards if the caller provides the product’s name. In case of calls from health-care, we consider PIC data on severity of health-effects to be highly reliable.

A majority of accidents were considered as minor to moderate risk incidents; our results also show that callers from the public in 54% of cases received advice on how to manage on site (Table 2). In Litovitz et al. (1993) the medical outcome was known for approximately 78% of occupational cases, of which 85% (corresponding to 66% of all cases) resulted in no or minor health effects that did not need treatment. In the survey of cases with adolescents by Woolf et al. (2001) 50% were possible to manage on site. Hence, advice from a PIC may contribute significantly to reduce the health care burden of occupational accidents. Potential corrosives were implicated in the largest number of major risk incidents. In our material 30% of incidents involve some kind of acid or alkali. Products categorizations differ between studies, but previous works from the US and Switzerland point towards a lower share of acids and alkali (around 20% of reported exposures in Blanc and Olson, 1986; Litovitz et al., 1993; Hinnen et al., 1994; Woolf et al., 2001). The difference may be due to the national and/or time related differences in manufacturing, chemical product composition or occupational health and safety practices. One example of the latter could be improved routines and reduced use of certain product groups, e.g. solvents, while use and routines have stayed more similar for acids and alkali, leading to these kinds of products having increased in relative importance. Calls about incidents with alkali (e.g. present in cement and cleaning products) were more frequent than with acids, and also more frequently judged as major risk incidents, since alkalis penetrate tissue more rapidly and thereby are more prone to cause severe wounds than acids. The corrosive or irritative properties of acid and alkali products occur rapidly after exposure and hence it may be likely that a larger share of incidents involving these lead to patients contacting the PIC compared to other chemical products whose acute effects are not as distinctly noticeable. However, these findings also indicate that handling of potential corrosives often is performed without proper use of protective risk management measures, and the large number of eye exposures points specifically towards lacking use of protective goggles. This observation, as well as possible introduction of mitigation measures, needs to be further investigated. For instance, free text fields of PIC data and SWEA injury reporting could be mined for information on use of protective equipment. Follow up interviews with injured or exposed individuals would further illuminate in which respects risk management measures were lacking. The largest number of telephone consultations concerned eye exposures (37%) followed by inhalation (32%). It seems likely that exposed individuals are more prone to call about eye exposures compared to other routes. However, this exposure pattern was not found in previous studies reporting exposure routes. In these inhalation was the most commonly recorded exposure route (in 35–67% of cases), followed by skin (21–31% of cases) (Blanc et al., 1989; Bresnitz, 1990; Litovitz et al., 1993; Woolf et al., 2001). Eye exposures were reported in 11–27% of exposures investigated in these studies. Although comparatively small, the 6% oral exposures seen in our material attract interest in the occupational setting, especially as intentional poisoning is not coded as occupational accidents. Data routinely collected by the Swedish PIC is limited to meagre demographics (gender and age category) a general identification of exposure and acute medical management. Usefulness of Swedish PIC data for occupational monitoring purposes would be improved by systematic collection of additional demographic data. For instance, we suggest systematically recording the callers’ county (Sweden currently is divided into 21 counties). This would allow closer comparison to the SWEA database which includes detailed addresses of the concerned workplaces. Based on similar reasoning we would also like to propose that when possible (e.g. for single person incidents) age is noted in years rather than age category. Again this would allow closer comparisons to the SWEA database, but also provide a basis for analysis of age patterns of exposures, which possibly could in more detailed studies also be cross tabulated with free-text information such as use of

5. Conclusions and outlook Although the data compiled by the Swedish PIC have some limitations from the perspective of occupational surveillance regarding demographic data, we find that it may serve as a highly valuable source of data for occupational accidents and incidents with chemical products. For instance, as a means to identify chemical products with problematic uses, which could be indicated by the product being repeatedly involved in accidental exposures. A closer scrutiny of the official statistics on occupational accidents with hazardous materials would also allow us to investigate whether the pattern of chemical groups identified in the present work also is reflected in the official statistics. The PIC records also hold opportunities for experience feed-back from incidents that have not lead to severe health effects, but may well have. Such kinds of incidents are not well covered by official statistics, due to current reporting requirements and probably also poor compliance from employers. Systematic and recurring overviews of PIC data on occupational incidents may hence improve national risk management efforts. We have identified several issues that may be of interest for further study in order to more efficiently manage workplace risks. The increasing number of incidents recorded per year is not necessarily an indication of a deteriorating work environment; however, given the conflicting signals from our investigated materials the causes behind noted increases needs further investigation. The number of cases concerning potential corrosives is a more distinct indication that workplace risk management measures fail in many cases where acids and alkalis are used. In relation to these it would be of interest to more closely investigate the nature of products, areas of use and access to preventive risk management measures as well as to mitigation measures such as first aid equipment. Funding This work was supported by AFA Insurance [Project no AFA dnr 130288]. The parts performed at Swetox were supported by Stockholm County Council, Knut & Alice Wallenberg Foundation, and the Swedish Research Council FORMAS. References Blanc, P.D., Rempel, D., Maizlish, N., Hiatt, P., Olson, K.R., 1989. Occupational illness: case detection by poison control surveillance. Ann. Intern. Med. 111, 238–244. Blanc, P.D., Olsen, K.R., 1986. Occupationally related illness reported to a regional poison control center. Am. J. Public Health 76, 1303–1307. Bresnitz, E.A., 1990. Poison control center follow-up of occupational disease. Am. J. Public Health 80, 711–712. GIC – Giftinformationscentralen, 2011. Swedish Poisons Information Centre Annual Report 2010. GIC – Giftinformationscentralen, 2012. Swedish Poisons Information Centre Annual Report 2011. GIC – Giftinformationscentralen, 2013. Swedish Poisons Information Centre Annual

274

Safety Science 104 (2018) 269–275

L. Schenk et al.

opioid analgesic abuse: Application of a spatial mixed poisson regression mode using data from a network of poison control centers. Pharmacoepidemiol. Drug Saf. 17, 1050–1059. Sudakin, D.L., Power, L.E., 2007. Pyrethrin and pyrethroid exposures in the United States: a longitudinal analysis of incidents reported to poison centers. J. Med. Toxicol. 3, 94–99. SWEA – Swedish Work Environment Authority, 2014. The Work Environment 2013 Arbetsmiljöstatistik Rapport 2014:3 < https://www.av.se/globalassets/filer/ statistik/arbetsmiljostatistik-arbetsmiljon-2013-rapport-2014-03.pdf > (accessed 21. 11.16). Swedish Parliament, 1977. Swedish Work Environment Act. US GAO – United States Government Accountability Office, 2012. Workplace Safety and health: Better OSHA guidance needed on safety incentive programs. GAO-12-329. < http://www.gao.gov/products/GAO-12-329 > (accessed 21.11.16). Walters, D., Johnstone, R., Frick, K., 2011. Introduction: inspecting health and safety management and the consequences of restructuring in the modern world of work. In: Walters, D., Johnstone, R., Frick, K., Quinlan, M., Baril-Gingras, Thébaud-Mony, A. (Eds.), Regulating Workplace Risks: A Comparative Study of Inspection Regimes in Times of Change. Edward Elgar Pub, UK, pp. 3–20. Woolf, A., Alpert, H.R., Garg, A., Lesko, S., 2001. Adolecent occupational toxic exposures. Arch. Pediatr. Adolesc. Med. 155, 704–710.

Report 2012. GIC – Giftinformationscentralen, 2014. Swedish Poisons Information Centre Annual Report 2013. GIC – Giftinformationscentralen, 2015. Swedish Poisons Information Centre Annual Report, 2014. < http://www.giftinformation.se/globalassets/publikationer/ arsrapport-2014_engelska.pdf > (accessed 21.11.16). Gravseth, H.M., Wegeland, E., Lund, J., 2003. Underrapportering av arbeidsskader til Arbeidstilsynet [Under-reporting of occupational injuries to the labour inspectorate]. Tidskr Nor Lægefor 15:2057–2059 In Norwegian, English abstract available. Herlitz, J., Dellborg, M., Karlsson, T., Evander, M.H., Hartford, M., Perers, E., Caidahl, K., 2009. Treatment and outcome in acute myocardial infarction in a community in relation to gender. Int. J. Cardiol. 135, 315–322. Hinnen, U., Hotz, P., Gossweiler, B., Gutzwiller, F., Meier, P.J., 1994. Surveillance of occupational illness through a national poison control center: an approach to reach small-scale enterprises? Int. Arch. Occup. Environ. Health 66, 117–123. Leigh, J.P., Marcin, J.P., Miller, T.R., 2004. An estimate of the U.S. Government’s undercount of Nonfatal Occupational injuries. J. Occup. Environ. Med. 46 (1), 10–18. Litovitz, T., Oderda, G., White, J., Sheridan, M., 1993. Occupational and environmental exposures reported to poison centers. Am. J. Public Health 83, 739–743. Loikas, D., Karlsson, L., von Euler, M., Hallgren, K., Schenck-Gustafsson, K., Bastholm Rahmner, P., 2015. Does patient’s sex influence treatment in primary care? Experiences and expressed knowledge among physicians – a qualitative study. BMC Fam Pract 16, 137. Meulenbelt, J., de Vries, I., 1997. Acute work-related poisoning by pesticides in the Netherlands; a one year follow-up study. Med Rev (Przeglad Lekarski) 54, 665–670. Olson, D.K., Sax, L., Gunderson, P., Sioris, L., 1991. Pesticide poisoning surveillance through regional poison control centers. Am. J. Public Health 81, 750–753. Persson, H.E., Sjöberg, G.K., Haines, J.A., Pronczuk de Garbino, J., 1998. Poisoning severity score. Grading of acute poisoning. J. Toxicol. Clin. Tox. 36, 205–213. Probst, T.M., Barbaranelli, C., Petitta, L., 2013. The relationship between job insecurity and accident under-reporting: a test in two countries. Work Stress 27 (4), 383–402. Smith, M.Y., Irish, W., Wang, J., Haddox, J.D., Dart, R.C., 2008. Detecting signals of

Online databases ISA, 2016. Online database over Swedish occupational accidents and injuries. In Swedish. < http://webbstat.av.se/QvAJAXZfc/opendoc.htm?document=AccessPoint %2FArbetsskadestatistik.qvw&host=QVS%40vmextapp01-hk&anonymous= true > (accessed 21.11.16). SCB, 2016. Online databases for Swedish Labour market statistics. < http://www.scb.se/ en_/Finding-statistics/Statistics-by-subject-area/Labour-market/ > (accessed 21. 11.16).

275