Measures for Bioaerosol Reduction in the Salmon Industry

The intervention trial is part of a broader study on assessment of exposure and health of workers in the salmon industry. Measurements of exposure, health examinations and questionnaire data serve as a basis for both intervention- and other substudies. The intervention trial itself is a parallel group, three-arm non-randomized superiority trial with a 1:1:1 allocation. Processing of salmon includes two main processes: 1) in the slaughtering department, the fish are slaughtered, sorted, automatically or manually gutted, and washed before further processing; and 2) in the filleting department, the salmon is filleted and skinned by machines before manual trimming of the fillets. The main source of exposure to fish proteins is by inhalation of wet bioaerosols containing a heterogeneous mixture of proteins, enzymes, endotoxins and microorganisms. It is hypothesized that a reduction in exposure will cause a decrease in the prevalence of symptoms from the airways and skin, as well as sensitization, in workers of factories that implement the exposure-reducing measures.Hence, the overall aim of identifying effective and feasible control measures (interventions) that reduce exposure to these bioaerosols is to lower the risk of, or preferably prevent, the development or chronification of respiratory diseases, eczema and allergy among salmon processing workers. The main hypothesis behind the interventions is that control measures that aim at using less water, at reduction of bioaerosol generation or at shielding workers from the production line have the potential to reduce bioaerosol exposure. The three arms of the trial are therefore 1) technical intervention targeting alteration of nozzles or nozzle function along the production line (NZ), 2) behavioral or technical intervention related to cleaning of surfaces (CS), or 3) no intervention (control). Each of the intervention arms represents a general intervention category in which details for the actual intervention are to be developed in collaboration with the factory staff to maximize the effect of the interventions. The intervention arms NZ and CS imply measures that have the potential to reduce bioaerosol exposure for the employees, while factories in the CTR arm will not receive any active intervention measures. NZ targets alteration of nozzles or nozzle function along the production line, e.g. by manipulation of nozzle dimensions, alteration of operating pressure or shielding of the nozzle's output stream. CS focuses on reduction of bioaerosol exposure for the employees while cleaning their personal operating areas during working operations or while cleaning floor areas . This can include increased use of swabbing instead of flushing with water hoses or alterations in use of water hoses for cleaning. Examples for the latter are a reduced flushing frequency, reduction of number of employees executing the work, alterations of hose dimension or nozzles on hoses. Suggestions from employees regarding details of the intervention will be taken into account for the final design of the intervention. This can include local specific characteristics, specific location of intervention or others. Randomization of intervention arms is not feasible due to different premises of factory infrastructure. The nine participating factories (three from each geographical region) will therefore be allocated to the study arms by the project group, reassuring that each region will cover all three study arms. Data assessment will be done by questionnaires, exposure measurements and health examinations before (T1) and after the intervention period (T2). Interventions will start 5-8 weeks after T1, permitting technical adjustments where necessary. Interventions are to be continued until follow-up measurements at T2, 12 months after T1. Two phone calls and one physical visit are to be carried out during the intervention period. Questionnaire data relevant for the intervention trial include background demographics, questions relevant for exposure assessment (work tasks, department, etc), as well as questions on symptoms from the eyes, the respiratory system and skin. Health examinations will include spirometry (lung function), skin prick testing (allergic sensitization) and blood sampling (sensitization, inflammation). Exposure assessment will be done by personal exposure measurements collected in the breathing zone of workers, as well as by stationary sampling. While all workers (approx. 1500-2000 subjects), will be eligible for answering the questionnaires, the number of invitees for health examination and exposure assessments will be much lower (ca 900 and 216, respectively), due to logistical and timely constraints. Sample size calculations was performed for the primary outcome; the effect of interventions in reducing the concentration of bioaerosols (total protein) in salmon processing plants. The inclusion of 24 workers with 2 repeated measurements in each of the three intervention arms (see section on inclusion criteria above) will give a total of 432 measurements (144 measurements in each of the intervention arms). According to the sample size calculations, 133 exposure measurements of total proteins in each of the intervention arms give a power of 80% of detecting a 20% reduction of exposure (significance level of 0.05). The corresponding number of measurements needed to detect a reduction in exposure of 40 and 30% were 34 and 59 measurements, respectively. Hence, 144 measurements as aimed at in the present study will ensure a buffer of about 8% in order to account for rejection of measurements due to sampling errors. The number of exposure measurements per work shift will also be limited by logistical constraints.