Development and Evaluation of the Firefighter Exposure to Carcinogens Scale

Firefighters face carcinogen exposure, necessitating effective decontamination practices. The Firefighter Exposure to Carcinogens Scale (FECS) provides a validated tool to assess these behaviors. Its strong reliability across English- and French-speaking populations supports its use in cancer prevention efforts, guiding interventions to enhance firefighter safety and long-term health.

Abstract

Objective

Firefighters are exposed to carcinogens from combustion, necessitating decontamination practices. This study developed and validated the Firefighter Exposure to Carcinogens Scale (FECS) to assess exposure-mitigating behaviors.

Methods

The sample included 179 volunteer firefighters from across Canada, comprising both English and French speakers, evaluated 20 items on exposure prevention across the following three dimensions: perceived importance, past behavior, and future intention. Principal axis factor analysis was conducted, and parallel analysis based on principal components determined the number of factors. McDonald’s Omega measured internal consistency, and item-total correlations were examined.

Results

A one-factor solution was acceptable for all scales, with high coefficient omega values indicating strong internal consistency. Small mean differences between language groups were nonsignificant, and no correlations were found with demographic variables.

Conclusions

The FECS is a reliable, valid one-factor model for both languages, supporting cancer prevention efforts.

LEARNING OUTCOMES

• Develop the Firefighter Exposure to Carcinogens Scale (FECS) to assess decontamination practices.

• Assess the internal consistency of the FECS in English- and French-speaking firefighters.

• Examine the factor structure of the FECS to determine its dimensionality.

The World Health Organization’s International Agency for Research on Cancer recently classified occupational exposure as a firefighter as a carcinogen.¹ This determination is further reinforced by several recent studies^2–4 and a recent overview of the systematic reviews of cancer prevalence in firefighters.⁵

Firefighters face a serious risk of exposure to carcinogens from combustion, which releases carcinogens including polycyclic aromatic hydrocarbons,⁶ asbestos,⁷ and diesel fumes, which may come either from the fire or trucks that can be transferred to firefighters via skin, inhalation and ingestion.² Although the direct link between firefighter behavior regarding exposure to carcinogens and cancer has been investigated in only a few studies,² the biological logic for a connection between behavior mediating exposure is widely accepted. Personal protective equipment (PPE) worn by firefighters, including pants, jackets, boots, gloves, helmets, facemasks, hoods, and self-contained breathing apparatus (SCBA), plays a critical role in protecting firefighters from exposure to carcinogens. While SCBA is designed to protect the respiratory tract during active exposure, other components of PPE can retain harmful contaminants if not properly cleaned, posing a risk of secondary exposure through dermal absorption and off gassing.⁸ Therefore, an important defense against carcinogen exposure is decontamination of the protective gear after exposure. Decontamination of equipment and the individual firefighter have been recommended to mitigate possible exposure to carcinogens.^9–11

Decontamination practices are evolving and need to be adapted to organizational contexts¹² because not all fire departments have the same resources. Practice varies widely and is influenced by a range of factors from attitudes and norms to logistical barriers and training.¹³ However, only a small percentage of firefighters adhere to all best decontamination practices on-scene, with urban firefighters more likely to follow these procedures compared to rural firefighters, who are predominantly volunteers.¹⁴

In Canada, there are about 89,000 volunteer firefighters, 71% of all Canadian firefighters, and 83.2% of the 3248 fire departments are fully volunteer.¹⁵ All-volunteer departments, which are located in smaller, more rural, and remote areas, typically have fewer resources than career or composite fire departments, potentially increasing the risk of exposure for these firefighters.

There is limited information about the decontamination and exposure-mitigating practices of firefighters.¹⁶ Moore et al conducted a texting-based survey of 250 Florida firefighters, asking them how frequently they cleaned their gear, their confidence in the effectiveness of the cleaning, where they cleaned their gear, their knowledge of how to properly dry bunker gear, and whether showered upon return to the station after attending a fire call.¹⁶ Harrison et al surveyed 485 firefighters about mixed attitudes, beliefs, perceived norms, barriers, and behaviors toward postfire decontamination.¹³ The survey included eight items focused on perceptions of cleaning gear, such as “Cleaning gear is a sign of professionalism.” There were three items on beliefs regarding the value of cleaning gear like, “Cleaning my gear will reduce my chance of getting cancer,” and three decontamination norms, for example “Most firefighters I work with clean their gear after structure fires.” Additionally, a series of questions addressed specific decontamination practices, such as, “How often do you use wipes to clean face hands and arms, change out dirty hoods.” Firefighters were asked to indicate their agreement or disagreement with statements on barriers to decontamination, such as, “Cleaning my gear after a fire or incident takes too much time.” Finally, five questions focused on issues related to wet gear from decontamination, including statements like, “Having wet gear from gross decontamination is a big problem for me.” The few studies on reducing firefighter exposure to carcinogens^13,17–19 have used single or just a few exposures to contaminant items. To advance research on preventing carcinogen exposure and cancer in firefighters, a comprehensive understanding of their behaviors and intentions related to decontamination is needed, ideally measured by a valid and reliable scale. Such a scale could help assess firefighter beliefs, behaviors, and intentions, as well as evaluate interventions aimed at reducing carcinogen exposure. Our objective was to develop and provide initial validation for the Firefighters Exposure to Carcinogens Scale (FECS). The scale would be a self-report scale using a standard set of behaviors to assess perceived importance in cancer prevention, past behaviors during the previous three fires, and future intentions to engage in these behaviors if resources were available.

Identification of Domain and Item Generation; Content Validity and Pretesting

The Canadian Association of Fire Chiefs (CAFC) was integral to this research project with four members of the research team coming from CAFC. The research was approved by the IWK Health Centre Research Ethics Board (IWK-REB) Halifax, NS, Canada (approval #1027461). Dr. Elena Laroche oversaw the French version of the FECS under the research ethics board of Laval University. The Behavior, Importance, and Intention items were initially generated in English and were then translated to French by a native French speaker who was an experienced translator. The translation was evaluated by independent back translation.

The 20 Behavior, Importance, and Intention items were generated from the current literature. One item focused on an annual cancer screening and the rest were on aspects of prevention of exposure and decontamination. Each item was evaluated by research participants on the following three dimensions: perceived importance in preventing cancer, past behavior in the last three fires attended, and future intention if resources were available.

Three members of the research team (an active-duty Deputy Fire Chief, an active-duty fire chief, and a retired firefighter who was a past president of the Canadian Volunteer Fire Services Association) reviewed the scale for content validity. They provided feedback on content relevance, representativeness, and technical quality of the items and responses.

Seven volunteer firefighters (five English- and two French-speaking firefighters) were interviewed by telephone. Participants provided feedback on comprehensiveness, clarity, and relevance of each scale applied to the 20 items. Following these evaluations modifications were made to grammar, word choice, and answer options. Table 1 contains the final FECS items.

TABLE 1.

Firefighter Exposure to Carcinogens Scale Items (English Version)

FECS Items
1. Ensure that you have checked your issued personal protective equipment (PPE) (eg, SCBA, turnout gear, helmet, gloves, balaclava) so that it is appropriately fitted, is in good working conditions and meets NFPA Standards.
2. Change fire hood and balaclava every time you exit and re-enter fire.
3. Wash your hands with soap and water or decontamination wipes before touching your face.
4. Wash your hands with soap and water or decontamination wipes before eating/drinking at fire scene.
5. Perform gross decontamination by having your boots, pants, jacket and hood brushed off or washed off before removal at the site of the fire.
6. If you are doing gross decontamination for other firefighters, wear PPE (eg, SCBA, turnout gear, helmet, gloves, balaclava).
7. Use full PPE including respiratory protection during salvage and overhaul phase of fire (eg, SCBA, turnout gear, helmet, radio, gloves, balaclava).
8. Use decontamination wipes or wash to remove soot from your skin (eg, face, head, jaw, neck and underarms) before clearing a fire scene.
9. Perform fine decontamination at the firehouse before putting your PPE back in service (eg, SCBA, turnout gear, helmet, radio, gloves, balaclava).
10. Transport dirty turnout gear in sealed containers or compartments away from the cab of the truck.
11. Shower within 1 hr of completing a fire call.
12. Entirely change to clean clothes within 1 hr of completing a fire call.
13. Wash and dry clothes worn under your turnout gear in a departmental (not your home) washing machine.
14. Wash and dry bunker gear in an industrial washer following NFPA Standards.
15. Avoid taking or wearing the clothes you wore under your PPE at the fire into your home.
16. Keep all equipment or fire hose used during the fire call away from the cab at all times.
17. Clean all used or contaminated tools, and hose at the fire station before putting the trucks back in service.
18. Use PPE (eg, N95 mask, nitrile gloves, protective eyewear, and coveralls) when handling or cleaning soiled bunker gear or equipment upon returning to the station.
19. Avoid the exposure of yourself or your turnout gear to diesel exhaust fumes from any source.
20. Have regular medical examinations including cancer screening.

To evaluate the three scales participants used a 5-point response (0, strongly disagree to 4, strongly agree with 2 labeled neutral). Participants rated the 20 items three times using the scales: (1) Belief: This plays an important role in cancer prevention; (2) Intention: I would do this in the future if the resources were available; and (3) Behavior: I did this when I performed firefighting duties at the last three working fires.

Our primary hypothesis was that the 20 items in each of the behavior, importance, and intention scales could each be treated as a general (one factor) scale. Second, based on the seminal work of Ajzen,²⁰ the theory of planned behavior, and the recent extensive literature on the attitude-intention-behavior gap,²¹ we hypothesized that the scores on the Behavior Scale would be lower than the scores on the Importance and Intentions Scale. In addition, we expected that the scores on the three scales would be positively correlated, but that the Importance and Intention Scales would be more highly correlated with each other than would their correlation with the Behavior Scale. Third, based on the shared experience of firefighting we hypothesized the French and English firefighters would have similar average scores on the three scales.

METHODS

Sampling of participants: The participants were recruited through their fire chiefs via email between June 2022 and November 2023. Participants came from 72 volunteer departments across Canada, including all provinces. A detailed description of the sample can be found in Table 2. All firefighters who were recruited for the study met the following criteria: 1. a volunteer firefighter, 2. working in an all-volunteer fire department, 3. living in Canada, and 4. able to read, write and understand English or French.

TABLE 2.

Sociodemographic Characteristics of the Study Participants

Demographic Characteristics	English (n = 101)		French (n = 78)
	n	%	n	%
Sex
Male	71	70.29	63	80.76
Female	15	14.85	14	17.94
Other	1	0.99	1	1.28
Missing	14	13.86	0	0
Age, median (range)	39 (21 to 68)		42 (169 to 66)
Education
High school	16	15.84	16	20.51
Technical training/vocational training	44	43.56	50	64.10
University degree	21	20.79	11	14.10
Missing	20	19.80	1	1.28
Employment
Full-time	73	72.27	51	65.38
Not full-time	11	10.89	20	25.64
Missing	17	16.83	7	8.97
Location
Urban setting	18	17.82	13	16.66
Suburban setting	10	9.90	9	11.53
Rural setting	58	57.42	43	55.12
Remote setting	1	0.99	13	16.66
Missing	14	13.86	0	0
Years working as a firefighter, median (range)	10 (0 to 43)		12 (0 to 45)
No. fires attended in the past 6 mos, median (range)	4 (0 to 50)		4 (0 to 75)

One hundred fifty-five English-speaking firefighters initially agreed to participate; however, between 57, 55, and 58 firefighters responded to fewer than half the items on the Behavior, Importance, and Intentions scales, respectively, and these were removed from the analysis. Missing item responses were imputed for firefighters who answered at least half of the 20 items on a given scale. Specifically, we used single imputation Predictive Mean Matching^22,23 to impute item responses for the 33, 23, and 25 firefighters who were missing between 1 and 10 item responses for the Behavior, Importance, and Intentions scales, respectively. The resulting sample size for the English-speaking firefighters was Behavior (98), Importance (100), and Intention (97).

One hundred eighteen French-speaking firefighters initially agreed to participate, however, 41, 40, and 45 firefighters responded to less than half the items on the Behavior, Importance, and Intentions scales, respectively. Missing item responses were imputed for between 18, 7, and 17 firefighters who were missing between 1 and 10 item responses for the Behavior, Importance, and Intentions scales, respectively. The resulting sample size for the French-speaking firefighters was Behavior (77), Importance (78), and Intention (74). We elected to use slightly different samples of firefighters for each measure as our primary hypothesis concerned the structure of each scale in each sample individually, and we wanted to maximize our sample size for each measure. Using only firefighters who completed all items would have reduced the sample size for the analyses by 20%.

Measures

The FECS (Table 1): The 20 items were all rated three times; once each for Behavior, Importance, and Intention was on a 5-point scale ranging from 0 to 4. Each of the three ratings was analyzed separately across the same 20 items.

Participants also completed a demographic questionnaire that included age, sex, education level, employment status, years of experience, and number of fires attended in the past 6 months.

Study Protocol

Volunteer departments were invited to participate through our partners, the Canadian Association of Fire Chiefs (CAFC) and the Canadian Volunteer Fire Services Association (CVFSA), as well as through personal networks of the study team members. Electronic posters containing a link to our study registration page were emailed to the relevant fire chiefs, who then forwarded them to their active volunteer firefighters. After reading and signing the informed consent form online, participants were given access to the survey, which was hosted electronically on Research Electronic Data Capture (REDCap, Vanderbilt University).²⁴ They could complete the questionnaires over multiple sittings, with a timeframe of a week to finish them. Three reminders were sent to the participants after a week, including a link that directed them to where they had left off on the surveys. The participants completed the demographic survey as well as the FECS Behavior, Importance, and Intention scales. All responses were anonymous.

Data Analysis

The data were analyzed with SPSS Statistics 29 (IBM Corp, USA), R 4.3 (R Development Core Team) and JASP 18.3 (University of Amsterdam). After calculating and inspecting the KMO statistic, the structural analyses of the primary hypothesis began with a principal axis factor analysis followed by an oblique (Promax) rotation. Parallel analysis based on principal components²⁵ was used to initially identify the number of factors. The factors were rotated to evaluate the interpretability of the number factors identified by parallel analysis and the correlations among the rotated factors were inspected. McDonalds Omega was computed as an index of the internal consistency of the 20 items, and the corrected item total correlations were inspected. Following these analyses, confirmatory factor analysis was used to further evaluate the model fit (using the CFI and root mean squared error of approximation) of the one and (if warranted) multifactor solutions. We used both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) as they provide different ways to evaluate structural models. In general, a one-factor solution for all the scales was determined to be acceptable (see results) and a total observed score was computed for each scale in each sample across the 20 items. To test the secondary hypotheses a repeated measures Analysis of Variance (ANOVA) was used to compare the mean scores of the three scales using the same items, within each sample and a between/within ANOVA was used to compare the scales between the English and French firefighters. The three scores were also correlated within each sample. In an exploratory analysis the scores were also related to the demographic variables in each sample.

RESULTS

Sample Description

Table 2 summarizes the demographic characteristics of the English and French firefighter samples.

Structural Analyses

Our primary hypothesis was that treating the Importance, Intention, and Behavior scales as one-dimensional and computing a total score for each would be justified by the structural analyses of these scales. Table 3 summarizes the results of both exploratory and confirmatory factor analyses of the three scales in the French and English samples. Except for the English Intentions scale, the results of the parallel analysis generally indicate more than one factor and only two of the six scales meet Lord’s (1959) criteria for unidimensionality (the ratio of the first to the second eigenvalues >4.0). However, the results in Table 3 appeared to support treating the 20 items as one single overall scale even if the item set is not strictly unidimensional. These results include the generally acceptable CFI’s for a one-factor model, the high correlations between the first and second factors following an oblique rotation, the fact that all 20 items load significantly on the first unrotated factor, and the high values of coefficient omega that indicate an internally consistent item set. In addition to these results, the rotated two- and sometimes three-factor solutions proved difficult to cleanly interpret. Moreover, the multifactor solutions were not consistent between the English and French samples. However, the one-factor solution was generally consistent between the French and English firefighters. Thus, we concluded that for the three scales treating the 20 items as a single scale as suggested by our first hypothesis was supported.

TABLE 3.

Summary of the Structural Analyses of the Behavior, Importance, and Intention Scales in the English and French Samples

Sample/Scale	Exploratory Factor Analysis				Confirmatory Factor Analysis			Internal Consistency
English	KMO	# factors	Ratio first/second	Correlation	CFI	RMSEA	# sig loadings	Coefficient	# ITC < 0.20
			Eigenvalues	first/second Factors				Omega
Behavior	0.79	2	3.2	0.60	0.94	0.13	20	0.89	0
Importance	0.88	2	4.5	0.67	0.99	0.06	20	0.92	0
Intention	0.81	1	4.5	—	0.98	0.10	20	0.90	1
French
Behavior	0.83	2	3.8	0.53	0.96	0.13	20	0.91	0
Importance	0.72	3	2.7	0.50	0.97	0.11	20	0.85	1
Intention	0.83	3	3.5	0.42	0.96	0.15	20	0.92	0

# factors is based on Horn’s parallel analysis using principal components, and the confirmatory factor analysis fit indices are for the one-factor solution across all 20 items.

CFI, Comparative Fit Index; KMO, Kaiser-Meyers-Olkin index of factorability; RMSEA, root mean squared error of approximation.

Description, Comparisons, and Correlations Among the Three Scale Scores

Table 4 shows the means and standard deviations of the Behavior, Importance, and Intentions scales for the English and French firefighters. The scores are based on the mean of the item responses so they can be interpreted based on the 5-point scale (0 = strongly disagree 5 = strongly agree). The mean score on the Behavior Scale indicates average slight agreement that the firefighters engage in the behavior, whereas the mean scores on the Importance and Intention scales indicate average fairly strong agreement that the firefighters view the behaviors as important and that they intend to engage in them if they had sufficient resources.

TABLE 4.

Means and Standard Deviations of the Scale Scores Based on the Mean Responses Across the 20 Items for the English and French Firefighters

Sample/Scales	Mean	Standard Deviation
English
Behavior	2.5	0.72
Importance	3.6	0.46
Intention	3.4	0.47
French
Behavior	2.4	0.86
Importance	3.5	0.67
Intention	3.2	0.67

The means were compared using a one-way repeated measures Multivariate Analysis of Variance (MANOVA). For the English firefighters the differences among the means were substantial (Wilks lambda = 0.308, F with 2 and 92 df = 103.1, P < 0.001, partial eta squared = 0.692). This was also the case for the French firefighters (Wilks lambda = 0.304, F with 2 and 71 df = 81.1, P < 0.001, partial eta squared = 0.696). These results support the second hypothesis that firefighters’ engagement in cancer prevention behaviors would be less than their attitudes and intentions about those behaviors.

Table 5 shows the correlations among the three scales for the English and French firefighters. For both groups, the correlation between the Intention and Importance Scale is larger than the correlation of either scale with the Behavior Scale. These results are also consistent with the second hypothesis that behavior is not highly correlated with intentions or perceived importance. We also correlated the three scales with the demographic variables shown in Table 2 in both samples. None of these correlations reached conventional levels of statistical significance.

TABLE 5.

Correlations Among the Behavior, Importance, and Intentions Scales for the English and French Firefighters

	Behavior	Importance	Intention
Behavior	—	0.33	0.16
Importance	0.35	—	0.48
Intention	0.28	0.81	—

Note. All correlations are unlikely to be 0.00 (P < 0.05, two-tailed test) except for the correlation between the Behavior and Intention scales in the French sample where P = 0.175.

Comparison of the Scores Between the French and English Firefighters

In a final analysis, we compared the mean scores on the three scales between the English and French firefighters using a mixed multivariate analysis of variance with the three scales as the repeated factor and group (English vs French) as the between subjects factor. Consistent with the analysis that compared the scales within each group, there was a large and significant effect across the three scales (Wilk’s lambda = 0.308, F with 2 and 164 df = 184.5, P < 0.001, partial eta squared = 0.692) with the Behavior Scale having a substantially lower mean than the other two scales in both groups. In contrast, there was a small and nonsignificant difference between the English and French firefighters in their scores across the three scales (F with 1 and 165 df = 2.1, P = 0.151, partial eta squared = 0.012). However, there was a modest, but significant, scale × group interaction (Wilk’s lambda = 0.896, F with 2 and 164 df = 9.5, P < 0.001, partial eta squared = 0.104). Inspection of the means indicates that this interaction reflects that the French firefighters exhibited a somewhat lower mean score on the Intention Scale compared to the Importance Scale, whereas for the English firefighters, the Importance and Intentions scores were essentially the same. Nonetheless, the overall pattern of scores was similar between the two groups

DISCUSSION

Research on cancer in firefighters has primarily focused on understanding the mechanisms of exposure. Few studies focused on firefighter attitudes, beliefs, or behaviors concerning cancer and risk reduction stemming from occupational practices.^13,26,27 In this study, we developed the first comprehensive scale on decontamination and exposure practices in firefighters, assessing 20 essential practices (scale items) based on their importance to cancer prevention, intention to implement, and behavior. This tool addresses a critical gap in firefighter safety and health research by providing a reliable and valid means to evaluate and improve cancer prevention practices. Providing a bilingual (English and French) format makes this tool accessible to a broader range of firefighters. The FECS may be suitable for translation in other languages. Unlike previous scales, which often focus on a very limited set of practices and mixed behavior with attitudes or are available only in one language,¹³ our approach ensures a more robust assessment.

There were only small differences in mean scores across the English- and French-speaking samples with an overall similar pattern. The one-factor model presenting a single scale for both English and French is accepted, which yields a reliable and valid scale with high internal consistency with ease of interpretation, making it suitable for use across both language groups.²⁸ Clearly though, further investigation of potential differences between different language versions of the scale is warranted.

The results from the Behavior, Importance, and Intentions scales showed firefighters’ engagement in cancer prevention behaviors is less than their attitudes and intentions about those behaviors. The consistent finding across various studies^29–32 is that while positive attitudes and intentions toward cancer prevention are prevalent, they do not always lead to the implementation of corresponding behaviors. We suspect that lack of resources may be key. Thus, for example, if there are no industrial washers at the firehall because the small rural municipality cannot fund them, washing gear could occur only if new resources were made available. Factors such as perceived behavioral control, self-efficacy, and societal influences play critical roles in bridging this gap.^33,34 Future research will address the possible barriers that may explain our results.

The absence of significant correlations with demographic variables indicates that the FECS subscales (Importance, Intention, and Behavior) likely measure constructs consistently across various demographic groups. This enhances the scales’ generality demonstrating that they are robust and not affected by external demographic factors.³⁵ However, subtle demographic effects could still be present and practically significant. Detecting these effects might require larger sample sizes.³⁶

This study serves as an example of using a completely electronic method to develop a scale tailored to a specific occupation. The Internet-based nature assisted in recruiting a more diverse population across Canada, enhancing the generalizability of the findings. However, it may have led to sampling bias, as individuals without Internet access were excluded and those who are less tech-savvy may have not participated.³⁷ Response rate was another challenge we faced, 58 English firefighters and 45 French firefighters were removed from the study as they completed less than half of the questions, which could be due to technical issues, such as website errors, slow internet connections, or compatibility problems with their devices, which can hinder their ability to complete the questionnaire.³⁸

Implications for Occupational Health Practice

It will be important to see if implementing this scale enhances firefighter safety by promoting best practices in cancer prevention behaviors. By regularly assessing and improving these practices, firefighters may reduce their exposure to hazardous substances and improve decontamination procedures, potentially decreasing the incidence of related health issues, such as cancer and respiratory diseases.

The scale can be integrated into training programs to educate firefighters on the importance of decontamination practices. It can also be used for routine assessments to ensure compliance and identify areas needing improvement. Furthermore, firefighting organizations can utilize the scale to develop and refine policies aimed at safeguarding firefighter health.

Limitations

One limitation of our study is the relatively small sample size, which may have affected the findings. This study was conducted in all volunteer departments, which could limit the generalizability of the results.

Further research is needed to explore the long-term effectiveness of cancer prevention practices assessed by our scale. Longitudinal studies could provide valuable insights into how these practices impact firefighter health over time. Additionally, adapting the scale for use in different firefighting environments and cultures could further its utility. As the science of exposure reduction advances, the scale will need to be updated accordingly, to remain aligned with emerging best practices.