Search

The dietary exposure of Finnish children to benzo[a]pyrene (BaP) and the sum of four EU marker polycyclic aromatic hydrocarbons (PAH4) was estimated. In total, 154 analyzed samples and the literatu...

BIKE is a Bayesian dietary exposure assessment model for microbiological and chemical hazards. A graphical user interface was developed for running the model and inspecting the results. It is based on connected Bayesian hierarchical models, utilizing OpenBUGS and R in tandem. According to occurrence and consumption data given as inputs, a specific BUGS code is automatically written for running the Bayesian model in the background. The user interface is based on shiny app. Chronic and acute exposures are estimated for chemical and microbiological hazards, respectively. Uncertainty and variability in exposures are visualized, and a few optional model structures can be used. Simulated synthetic data are provided with BIKE for an example, resembling real occurrence and consumption data. BIKE is open source and available from github.

For the non-smoking and non-occupationally exposed population in Europe, food is the main source of heavy metal exposure. The aim of the study was to estimate the dietary exposure of the Finnish adult population to cadmium, lead, inorganic arsenic, inorganic mercury and methyl mercury as well as nickel using governmental as well as industry data on heavy metal occurrence in foodstuffs and the data from two national food consumption surveys conducted in 2007 and 2012. The sources of heavy metal exposure were estimated for the working-age population (25 to 64 years) and for the elderly (65 to 74 years). Exposure differences between years and between population groups were compared statistically. The mean exposure of women aged 25 to 45 years to cadmium and lead was statistically significantly (p &lt, 0.001) higher, and the methyl mercury exposure lower (p = 0.001) than that of women aged 46 to 64 years. For nickel and inorganic arsenic the differences were lower but still statistically significant (p &lt, 0.05). Between genders, significant difference (p &lt, 0.05) was only seen for lead and nickel. Mean cadmium exposure was significantly higher in 2012 than in 2007. For at least 95% of the adult population, the risk of health damage from mercury or nickel exposure is negligible, but the margin of exposure for lead and inorganic arsenic is small and shows a possible risk of cancer or neurotoxic effects.

Listeria monocytogenes causes severe consequences especially for persons belonging to risk groups. Finland is among the countries with highest number of listeriosis cases in the European Union. Although most reported cases appear to be sporadic and the maximum bacterial concentration of 100 cfu/g is not usually exceeded at retail, cold smoked and salt-cured fish products have been noted as those products with great risk especially for the elderly. In order to investigate the listeriosis risk more carefully, an exposure assessment was developed, and laboratory results for cold smoked and salt-cured salmon products were exploited. L. monocytogenes exposure was modeled for consumers in two age groups, the elderly population as a risk group and the working-age population as a reference. Incidence was assessed by estimating bacterial growth in the food products at three temperatures. Bayesian estimation of the risk was based on bacterial occurrence and product consumption data and epidemiological population data. The model builds on a two-state Markov chain describing repeated consumption on consecutive days. The cumulative exposure is probabilistically governed by the daily decreasing likelihood of continued consumption and the increasing bacterial concentrations due to growth. The population risk was then predicted with a Poisson distribution accounting for the daily probabilities of purchasing a contaminated product and the cumulative total probability of infection from its use. According to the model presented in this article, elderly Finns are at a greater risk of acquiring listeriosis than healthy adults. The risk for the elderly does not fully diminish even if the products have been stored at the recommended temperature (between 0 and 3 °C). It can be concluded that the stage after retail, i.e. food handling and storage by consumer or professional kitchens, is essential to protection against listeriosis. The estimation model provides means for assessing the joint impacts of these effects.

We investigated connections between antimicrobial use (AMU), biosecurity, and the numbers of pigs and staff in ten Finnish farrow-to-finish herds. Data on AMU in each herd were collected for 12 months. AMU was quantified as treatment incidences per 1000 days at risk (TI) using the consensus defined daily dose calculation. Biosecurity was scored using the Biocheck.UGent T system. We also examined antimicrobial resistance patterns of indicator E. coli isolated from faeces of selected pigs. In each herd, two groups of five pigs were formed: 1) antimicrobial treatment group (ANT: at least one pig in the litter was identified as sick and treated with antimicrobials) and 2) non-antimicrobial treatment group (NON: the litter was not medicated). Faecal samples were taken from these pigs at 5 and 22 weeks of age, cultured, and indicator E. coli isolates were tested for antimicrobial susceptibilities. The AMU varied considerably between the herds. Altogether, most of the antimicrobial treatment courses were assigned to weaned piglets. When AMU was quantified as TIs, suckling piglets had the highest TI (mean 46.6), which was significantly higher (P < 0.05) than TIs in fatteners and breeders (9.3 and 7.3, respectively). The difference between TI in suckling and TI in weaned piglets (19.1) was not statistically significant. There was a tendency for a negative correlation between the TI in breeders and the number of sows (r = 0.56, P = 0.09). Larger herds had higher external biosecurity scores than smaller herds (LS-means; 72 vs. 66, P < 0.05). The proportions of E. coli isolates resistant to at least one antimicrobial were higher in pigs at 5 weeks than in pigs at 22 weeks of age (Binomial proportion means; 40.5 % vs. 15.5 %, P < 0.05); as well as proportions of isolates resistant to at least three antimicrobial classes (23.0 % vs. 3.7 %, P < 0.01). These proportions did not differ between the ANT and NON groups at either 5 or 22 weeks of age (P> 0.05). We found few connections: enhanced external biosecurity levels found in the large herds co-occurred with lower use of antimicrobials and herds with low biosecurity scores - especially in the internal subcategories - appeared to have higher proportions of resistant isolates. Conclusively, we suggest that enhancing internal biosecurity might contribute to a reduction in the spreading of antimicrobial resistance in pig herds.

Small mammals are known to carryCampylobacterspp.; however, little is known about the genotypes and their role in human infections. We studied intestinal content from small wild mammals collected in their natural habitats in Finland in 2010–2017, and in close proximity to 40 pig or cattle farms in 2017. The animals were trapped using traditional Finnish metal snap traps.Campylobacterspp. were isolated from the intestinal content using direct plating on mCCDA. A total of 19% of the captured wild animals (n= 577) and 41% of the pooled farm samples (n= 227) were positive forC. jejuni, which was the onlyCampylobacterspecies identified. The highest prevalence occurred in yellow-necked mice (Apodemus flavicollis) and bank voles (Myodes glareolus) which carriedCampylobacterspp. in 66.3 and 63.9% of the farm samples and 41.5 and 24.4% of individual animals trapped from natural habitats, respectively. Interestingly, all house mouse (Mus musculus) and shrew (Sorexspp.) samples were negative forCampylobacterspp.C. jejuniisolates (n= 145) were further characterized by whole-genome sequencing. Core genome multilocus sequence typing (cgMLST) clustering showed that mouse and vole strains were separated from the rest of theC. jejunipopulation (636 and 671 allelic differences, 94 and 99% of core loci, respectively). Very little or no alleles were shared withC. jejunigenomes described earlier from livestock or human isolates. FastANI results further indicated thatC. jejunistrains from voles are likely to represent a new previously undescribed species or subspecies ofCampylobacter. Core-genome phylogeny showed that there was no difference between isolates originating from the farm and wild captured animals. Instead, the phylogeny followed the host species-association. There was some evidence (one strain each) of livestock-associatedC. jejunioccurring in a farm-caughtA. flavicollisand a brown rat (Rattus norvegicus), indicating that although small mammals may not be the original reservoir ofCampylobactercolonizing livestock, they may sporadically carryC. jejunistrains occurring mainly in livestock and be associated with disease in humans.

Riskinarvioinnissa tarkasteltiin leivästä, muroista ja mysleistä sekä savustetuista liha- ja kalatuotteista peräisin olevien PAH-yhdisteiden aiheuttamaa syövän riskiä sekä lainsäädännöllisten rajojen potentiaalista vaikutusta PAH-altistukseen. Ruoankäyttöaineistona käytettiin DAGIS-tutkimuksessa kerättyjä ruokapäiväkirjoja (n=815) Uudellamaalla ja Keski-Pohjanmaalla. Elintarvikkeiden PAH-pitoisuustiedot olivat peräisin tieteellisestä kirjallisuudesta sekä Evirassa vuosina 2012–2017 tehdyistä analyyseistä. Altistuksen arvioinnissa käytettiin bayesiläistä mallinnusta. Riskinluonnehdinnassa laskettiin turvamarginaali (MOE) EFSA:n julkaisemien BMDL10- annoskuvaajien (BMDL10(bentso[a]pyreeni)=0,07 mg/kg/päivä; BMDL10(PAH4)=0,34 mg/kg/päivä) ja altistuksen osamääränä. Hyväksyttävän turvamarginaalin alarajana oli 10 000. Keskiarvoaltistus oli bentso[a]pyreenille 1 400 pg/kg/päivä ja PAH4:lle 8 100 pg/kg/päivä, kun käytettiin elintarvikkeiden keskimääräistä pitoisuutta. Vastaavasti altistuksen 97,5. persentiilit olivat 2 700 pg/kg/päivä ja 14 000 pg/kg/päivä. Turvamarginaalien 2,5. persentiilit olivat bentso[a]pyreenille 263 000 ja PAH4:lle 24 000, joten PAH-yhdisteet eivät aiheuta tarkastelluista elintarvikkeista terveysriskiä suomalaisille 3–6-vuotialle lapsille.

A key feature of risk analysis is that risk assessment and risk management should be functionally separated. However, the usefulness of a risk assessment may be limited if the output is not designed to help with risk management decisions. The COMRISK project investigated the communication between risk assessors and risk managers. The overall goal of the project was to identify current practices and challenges in communication between risk assessors and risk managers during the risk analysis process, and thus increase and improve the understanding and the quality of the communication between them. Specific actions to achieve this aim included reviewing of historical food safety cases, analysing risk assessment requests, identifying communication guiding documents, including legislation and agreements, conducting semi‐structured interviews with risk assessors and risk managers, and identifying tools for facilitating the communication between risk assessors and risk managers. It was concluded that the usefulness of a risk assessment is strongly dependent on well‐defined and mutually recognised risk questions and that scarce or poor communication between risk assessors and risk managers is one of the major reasons when an output from risk assessment fails to support risk management. The communication between risk assessors and risk managers preceding the onset of the risk assessment, when the risk assessment requests with its risk questions are defined, is especially identified as one of the critical points to ensure a risk assessment that is fit for purpose. However, difficulties in understanding were also reported for the communication between risk assessors and risk managers during and after the risk assessment. Lack of communication is seldom a result of formal constraints or agreements nor can it be explained by a wish of the risk assessors or risk managers. Instead, perceived constraints or traditions appear to be possible underlying factors leading to scarce or poor communication between risk assessors and risk managers. It is essential that both risk assessors and risk managers acknowledge the crucial importance of communication between them while at the same time respect their different roles in a risk analysis.According to respondents, the best solution to facilitate the framing of the risk assessment questions isan open dialogue between risk assessors and risk managers to agree on the goal of the assessment and to build trust. Further, the interviewresults indicate that a formal systematic process may facilitate communication during the risk analysis. Where there is uncertainty, e.g due to data gaps or issues related to the methodology and models, it should be acknowledged and described properly by risk assessors to risk managers. Training of risk assessors and risk managers may improve the possibility of a timely and fit‐for‐purpose output.Such a training should give a deeper insight in the risk management process, give a better understanding of the risk managers role, and especially raise the awareness of the importance of the communication between risk assessors and risk managers.To improve the risk analysis process, it is also important that the risk assessor gets feedback regarding how risk assessments have met the needs of the risk managers.The present study also found that aspects of risk communication studied in this project are not extensively discussed in the guidance documents for risk analysis. More research is needed to identify the barriers for a fit for purpose communication.

The exposure of Finnish 1-year-olds to cadmium, lead and inorganic arsenic via food and drinking water was determined. The food consumption data consisted of 3-day records from 1010 children aged 12 months, collected during 2002 to 2005 in Southwest Finland. One fifth of these children were still breastfed when the consumption data were collected and their exposure was assessed separately from the non-breastfed children. The heavy metal concentration data in foodstuffs were mainly analysis results from national authorities and they were mostly from the years 2005 to 2012. Dietary exposure assessment was performed probabilistically using the online program MCRA. With middle bound estimates, 89% of the non-breastfed and 56% of the breastfed children exceeded the tolerable weekly intake of cadmium. The benchmark dose (BMDL01) for neurological damage caused by lead was exceeded by 60% of the non-breastfed and by 50% of the breastfed children, while the lowest BMDL01 for cancer risk increase caused by inorganic arsenic was exceeded by 77% of the non-breastfed and by 61% of the breastfed children. The assessment did not include the unknown heavy metal exposure from breast milk. Heavy metal exposure differences between the boys and the girls were also assessed. Breastfed girls had significantly higher heavy metal exposure relative to their bodyweight than the breastfed boys, while in the non-breastfed group there were no differences by sex.

The dietary exposure of Finnish 3-year-old and 6-year-old children to cadmium, lead, arsenic and mercury was determined using concentration data from Finland and individual food consumption data as well as individual weights of the children. Using middle bound estimates, 88% of the 3-year-olds and 64% of the 6-year-olds exceeded the tolerable weekly intake of cadmium. The benchmark dose for neurological damage caused by lead was exceeded by 14% and 1%, while the lowest benchmark dose of inorganic arsenic was exceeded by 43% and 29% for the 3-year-olds and 6-year-olds, respectively. The exposure of both age groups was below the tolerable weekly intake for inorganic mercury and methyl mercury. Although high, the exposures calculated with predominantly national concentration data were lower than previously estimated by EFSA, due to, for example, lower average concentrations in some much-consumed foods. The heavy metal exposure levels of the girls and the boys were also compared. Exposure to cadmium and lead was significantly higher for the boys than for the girls in both age groups, and exposure to inorganic arsenic was significantly higher for the 6-year-old boys than the girls of same age.

Pigs are an important source of human infections with Salmonella, one of the most common causes of sporadic gastrointestinal infections and foodborne outbreaks in the European region. Feed has been estimated to be a significant source of Salmonella in piggeries in countries of a low Salmonella prevalence. To estimate Salmonella risk to consumers via the pork production chain, including feed production, a quantitative risk assessment model was constructed. The Salmonella prevalence in feeds and in animals was estimated to be generally low in Finland, but the relative importance of feed as a source of Salmonella in pigs was estimated as potentially high. Discontinuation of the present strict Salmonella control could increase the risk of Salmonella in slaughter pigs and consequent infections in consumers. The increased use of low risk and controlled feed ingredients could result in a consistently lower residual contamination in pigs and help the tracing and control of the sources of infections.

Salmonella infections in pigs are in most cases asymptomatic, posing a risk of salmonellosis for pork consumers. Salmonella can transmit to pigs from various sources, including contaminated feed. We present an approach for quantifying the risk to pigs from contaminations in the feed chain, based on a Bayesian model. The model relies on Salmonella surveillance data and other information from surveys, reports, registries, statistics, legislation and literature regarding feed production and pig farming. Uncertainties were probabilistically quantified by synthesizing evidence from the available information over a categorically structured flow chain of ingredients mixed for feeds served to pigs. Model based probability for infection from feeds together with Salmonella subtyping data, were used to estimate the proportion of Salmonella infections in pigs attributable to feed. The results can be further used in assessments considering the human health risk linked to animal feed via livestock. The presented methods can be used to predict the effect of changes in the feed chain, and they are generally applicable to other animals and pathogens.

The Finnish Hygiene Passport System, a national legislative requirement, has been used to test the food safety knowledge of food handlers for almost 20 years, resulting in over one million approved hygiene passports. However, information on the relevance of the Hygiene Passport System is virtually nonexistent. In order to evaluate the relevance of the official hygiene passport test, we collected a sample of original official hygiene passport tests from test examiners. We also arranged a simulated hygiene passport test for volunteers without any professional background in the food sector to investigate whether the basic level of hygiene knowledge of Finns is sufficient to pass the test. Our study revealed that more than 80% of the participants in the official hygiene passport test passed. However, participants completing the test in a foreign language or with assistance had significantly more difficulties in passing the test. The results for the simulated test suggested that the food safety knowledge of most Finns would have enabled them to pass the official test without prior training, especially those older than 20 and with a higher educational level. The simulated test also revealed that preparation prior to the test, i.e. study or training, was effective in increasing the food safety knowledge of participants, especially when their knowledge level was initially low, as among young participants. Moreover, significant variation in the difficulty of the test itself was observed, which may affect the pass rate. The current study supports the relevance of the hygiene passport test, showing that without the national requirement for the hygiene passport test, the level of food hygiene knowledge of new employees in the food sector, especially young people aged under 20, would be lower. However, our results suggest that the relevance of the hygiene passport test should be improved by validation of the test and by using official translations when conducting the test in a foreign language. Digitalization of the test would enable the regular evaluation of the test based on accurate data collection.

A preliminary assessment of the cumulative exposure to heavy metals among Finnish preschool children is reported. Cadmium, lead, arsenic,and mercury affect many of the same organs in the human body. The effects are mostly caused by oxidative stress or disruption of enzyme function. The cumulative effects of the heavy metals on the central nervous system and on the kidneys are determined based on national concentration and consumption data, and comparison of the relative toxicity of the heavy metals is based on dose–response values found in the literature. The cumulative effects were assumed to be additive. The main contributors to kidney toxicity among the studied population groups were cadmium and lead, while lead was the main contributor to neurotoxic effects.

Several statistical models for salmonella source attribution have been presented in the literature. However, these models have often been found to be sensitive to the model parameterization, as well as the specifics of the data set used. The Bayesian salmonella source attribution model presented here was developed to be generally applicable with small and sparse annual data sets obtained over several years. The full Bayesian model was modularized into three parts (an exposure model, a subtype distribution model, and an epidemiological model) in order to separately estimate unknown parameters in each module. The proposed model takes advantage of the consumption and overall salmonella prevalence of the studied sources, as well as bacteria typing results from adjacent years. The latter were used for a smoothed estimation of the annual relative proportions of different salmonella subtypes in each of the sources. The source-specific effects and the salmonella subtype-specific effects were included in the epidemiological model to describe the differences between sources and between subtypes in their ability to infect humans. The estimation of these parameters was based on data from multiple years. Finally, the model combines the total evidence from different modules to proportion human salmonellosis cases according to their sources. The model was applied to allocate reported human salmonellosis cases from the years 2008–2015 to eight food sources.Several statistical models for salmonella source attribution have been presented in the literature. However, these models have often been found to be sensitive to the model parameterization, as well as the specifics of the data set used. The Bayesian salmonella source attribution model presented here was developed to be generally applicable with small and sparse annual data sets obtained over several years. The full Bayesian model was modularized into three parts (an exposure model, a subtype distribution model, and an epidemiological model) in order to separately estimate unknown parameters in each module. The proposed model takes advantage of the consumption and overall salmonella prevalence of the studied sources, as well as bacteria typing results from adjacent years. The latter were used for a smoothed estimation of the annual relative proportions of different salmonella subtypes in each of the sources. The source-specific effects and the salmonella subtype-specific effects were included in the epidemiological model to describe the differences between sources and between subtypes in their ability to infect humans. The estimation of these parameters was based on data from multiple years. Finally, the model combines the total evidence from different modules to proportion human salmonellosis cases according to their sources. The model was applied to allocate reported human salmonellosis cases from the years 2008–2015 to eight food sources.

The purpose of this study was to estimate the intake of selected bioactive compounds from fennel-containing plant food supplements (PFS) among Finnish consumers. The estimated average intake of estragole was 0.20mg/d, of trans-anethole 1.15mg/d, of rosmarinic acid 0.09mg/d, of p-coumaric acid 0.0068mg/d, of kaempferol 0.0034mg/d, of luteolin 0.0525μg/d, of quercetin 0.0246mg/d, of matairesinol 0.0066μg/d and of lignans 0.0412μg/d. The intakes of kaempferol, quercetin, luteolin, matairesinol and lignans from PFS were low in comparison with their dietary supply. The intake of estragole was usually moderate, but a heavy consumption of PFS may lead to a high intake of estragole. The intake of trans-anethole did not exceed the acceptable daily intake, but PFS should be taken into account when assessing the total exposure. To our knowledge, this study provided the first intake estimates of trans-anethole, p-coumaric acid and rosmarinic acid in human populations.

A Bayesian statistical temporal-prevalence-concentration model (TPCM) was built to assess the prevalence and concentration of pathogenic campylobacter species in batches of fresh chicken and turkey meat at retail. The data set was collected from Finnish grocery stores in all the seasons of the year. Observations at low concentration levels are often censored due to the limit of determination of the microbiological methods. This model utilized the potential of Bayesian methods to borrow strength from related samples in order to perform under heavy censoring. In this extreme case the majority of the observed batch-specific concentrations was below the limit of determination. The hierarchical structure was included in the model in order to take into account the within-batch and between-batch variability, which may have a significant impact on the sample outcome depending on the sampling plan. Temporal changes in the prevalence of campylobacter were modeled using a Markovian time series. The proposed model is adaptable for other pathogens if the same type of data set is available. The computation of the model was performed using OpenBUGS software.

Risk-based microbiological criteria can offer a tool to control Campylobacter in the broiler meat production chain. Recently two approaches have been applied to derive such criteria and to analyse their potential impact in terms of human health risk reduction: the risk-based version of the established microbiological criteria approach, that applies a microbiological limit (ML) for sample data, and the Danish “case-by- case” risk assessment approach, that applies a limit for the relative risk estimate (relative risk limit, RRL) based on sample data. In this study, data sets from Sweden and Denmark are used to compare the performance of the two approaches in terms of efficiency, i.e. the balance between the residual risk after implementation of the criterion and the percentage of non-complying batches, and the attending uncertainty. The analysis shows that the two approaches are equally efficient, and suggests that the RRL criterion is attended with less uncertainty. The two approaches are compared and their advantages and disadvantages are discussed. Given the uncertainties attending the results of the analysis, more research in terms of data collection, risk assessment and uncertainty analysis would be needed to develop these risk-based criteria further.

Human noroviruses (HuNoVs), a leading cause of food-borne gastroenteritis worldwide, are easily transferred via ready-to-eat (RTE) foods, often prepared by infected food handlers. In this study, the transmission of HuNoV and murine norovirus (MuNoV) from virus-contaminated hands to latex gloves during gloving, as well as from virus-contaminated donor surfaces to recipient surfaces after simulated preparation of cucumber sandwiches, was inspected. Virus transfer was investigated by swabbing with polyester swabs, followed by nucleic acid extraction from the swabs with a commercial kit and quantitative reverse transcription-PCR. During gloving, transfer of MuNoV dried on the hand was observed 10/12 times. HuNoV, dried on latex gloves, was disseminated to clean pairs of gloves 10/12 times, whereas HuNoV without drying was disseminated 11/12 times. In the sandwich-preparing simulation, both viruses were transferred repeatedly to the first recipient surface (left hand, cucumber, and knife) during the preparation. Both MuNoV and HuNoV were transferred more efficiently from latex gloves to cucumbers (1.2% ± 0.6% and 1.5% ± 1.9%) than vice versa (0.7% ± 0.5% and 0.5% ± 0.4%). We estimated that transfer of at least one infective HuNoV from contaminated hands to the sandwich prepared was likely to occur if the hands of the food handler contained 3 log 10 or more HuNoVs before gloving. Virus-contaminated gloves were estimated to transfer HuNoV to the food servings more efficiently than a single contaminated cucumber during handling. Our results indicate that virus-free food ingredients and good hand hygiene are needed to prevent HuNoV contamination of RTE foods.

Transmission of gastroenteritis-causing noroviruses may be significant via contaminated surfaces. Measures for control, e.g. disinfection with ultraviolet irradiation (UV), are therefore necessary for interrupting this transmission. Human norovirus (HuNoV) GII.4 and Murine norovirus (MuNoV) were used to study the efficacy of UV for virus inactivation on dry glass surfaces. MuNoV inactivation was measured using viability assay and the reduction in viral RNA levels for both viruses using reverse transcription quantitative PCR (RT-QPCR). For each UV dose, two parallel sample groups were detected using RT-QPCR: one group was enzymatically pre-PCR treated with Pronase and RNAse enzymes, while the other was not treated enzymatically. In the viability assay, loss of infectivity and a 4-log reduction of MuNoV were observed when the viruses on glass slides were treated with a UV dose of 60 mJ/cm2 or higher. In the RT-QPCR assay, a steady 2-log decline of MuNoV and HuNoV RNA levels was observed when UV doses were raised from 0 to 150 mJ/cm2. A distinct difference in RNA levels of pretreated and non-pretreated samples was observed with UV doses of 450–1.8 × 103 mJ/cm2: the RNA levels of untreated samples remained over 1.0 × 10³ PCR units (pcr-u), while the RNA levels of enzyme-treated samples declined below 100 pcr-u. However, the data show a prominent difference between the persistence of MuNoV observed with the infectivity assay and that of viral RNA detected using RT-QPCR. Methods based on genome detection may overestimate norovirus persistence even when samples are pretreated before genome detection.

Human norovirus (HuNoV), which causes gastroenteritis, can be transmitted to food and food contact surfaces via viruscontaminated hands. To investigate this transmission in food processing environments, we developed a swabbing protocol for environmental samples, evaluated the stability of HuNoV in the swabs, and applied the method in the food industry. Swabs made of polyester, flocked nylon, cotton wool, and microfiber were moistened in either phosphate-buffered saline (PBS) or glycine buffer (pH 9.5) and used to swab four surfaces (latex, plastic, stainless steel, and cucumber) inoculated with HuNoV. HuNoV waseluted with either PBS or glycine buffer and detected with quantitative reverse transcription PCR. HuNoV recoveries were generally higher with an inoculation dose of 100 PCR units than 1,000 PCR units. The highest recoveries were obtained when surfaces were swabbed with microfiber cloth moistened in and eluted with glycine buffer after a HuNoV inoculation dose of 100 PCR units: 66% ¡ 18% on latex, 89% ¡ 2% on plastic, and 79% ¡ 10% on stainless steel. The highest recovery for cucumber, 45% ¡ 5%, was obtained when swabbing the surface with microfiber cloth and PBS. The stability of HuNoV was tested in microfiber cloths moistened in PBS or glycine buffer. HuNoV RNA was detected from swabs after 3 days at 4 and 22uC, although the RNA levels decreased more rapidly in swabs moistened with glycine buffer than in those moistened with PBS at 22uC. In the field study, 172 microfiber and 45 cotton wool swab samples were taken from environmental surfaces at three food processing companies. Five (5.6%) of 90 swabs collected in 2010 and 7 (8.5%) of 82 swabs collected in 2012 were positive for HuNoV genogroup II; all positive samples were collected with microfiber swabs. Three positive results were obtained from the production line and nine were obtained from the food workers’ break room and restroom areas. Swabbing is a powerful tool for HuNoV RNA detection from environmental surfaces and enables investigation of virus transmission during food processing.

Microbial risk assessment provides a means of estimating consumer risks associated with food products. The methods can also be applied at the plant level. In this study results of microbiological analyses were used to develop a robust single plant level risk assessment. Furthermore, the prevalence and numbers of Listeria monocytogenes in marinated broiler legs in Finland were estimated. These estimates were based on information on the prevalence, numbers and genotypes of L. monocytogenes in 186 marinated broiler legs from 41 retail stores. The products were from three main Finnish producers, which produce 90% of all marinated broiler legs sold in Finland. The prevalence and numbers of L. monocytogenes were estimated by Monte Carlo simulation using WinBUGS®, but the model is applicable to any software featuring standard probability distributions. The estimated mean annual number of L. monocytogenes-positive broiler legs sold in Finland was 7.2 × 106 with a 95% credible interval (CI) 6.7 × 106–7.7 × 106. That would be 34% ± 1% of the marinated broiler legs sold in Finland. The mean number of L. monocytogenes in marinated broiler legs estimated at the sell-by-date was 2 CFU/g, with a 95% CI of 0–14 CFU/g. Producer-specific L. monocytogenes strains were recovered from the products throughout the year, which emphasizes the importance of characterizing the isolates and identifying strains that may cause problems as part of risk assessment studies. As the levels of L. monocytogenes were low, the risk of acquiring listeriosis from these products proved to be insignificant. Consequently there was no need for a thorough national level risk assessment. However, an approach using worst-case and average point estimates was applied to produce an example of single producer level risk assessment based on limited data. This assessment also indicated that the risk from these products was low. The risk-based approach presented in this work can provide estimation of public health risk on which control measures at the plant level can be based.

A Bayesian statistical temporal-prevalence-concentration model (TPCM) was built to assess the prevalence and concentration of pathogenic campylobacter species in batches of fresh chicken and turkey meat at retail. The data set was collected from Finnish grocery stores in all the seasons of the year. Observations at low concentration levels are often censored due to the limit of determination of the microbiological methods. This model utilized the potential of Bayesian methods to borrow strength from related samples in order to perform under heavy censoring. In this extreme case the majority of the observed batch-specific concentrations was below the limit of determination. The hierarchical structure was included in the model in order to take into account the within-batch and between-batch variability, which may have a significant impact on the sample outcome depending on the sampling plan. Temporal changes in the prevalence of campylobacter were modeled using a Markovian time series. The proposed model is adaptable for other pathogens if the same type of data set is available. The computation of the model was performed using OpenBUGS software.

Nitrite intake from the consumption of cured meat and tap water was estimated for Finnish children of 1, 3 and 6 years as well as Finnish adults of 25-74 years. Nitrite content in the foods was measured by capillary electrophoresis, and was then used together with individual food consumption data from the FINDIET 2007 and DIPP studies in a stochastic exposure assessment by a Monte Carlo Risk Assessment (MCRA) program. Nitrite intake from additive sources and tap water was assessed, and more than every 10th child between the ages 3 and 6 years was estimated to have a nitrite intake exceeding the acceptable daily intake (ADI) of nitrite. The high exposure levels were caused by frequent consumption of large portions of sausages, up to 350 g day(-1) or 750 g in 3 days, among the children. Median nitrite intake from cured meat was 0.016, 0.040, 0.033 and 0.005 mg kg(-1) body weight day(-1) for children of 1, 3 and 6 years and adults, respectively. Bayesian estimation was employed to determine safe consumption levels of sausages and cold cuts for children, and these results gave rise to new national food consumption advice.

Shifting from traditional hazard-based food safety management toward risk-based management requires statistical methods for evaluating intermediate targets in food production, such as microbiological criteria (MC), in terms of their effects on human risk of illness. A fully risk-based evaluation of MC involves several uncertainties that are related to both the underlying Quantitative Microbiological Risk Assessment (QMRA) model and the production-specific sample data on the prevalence and concentrations of microbes in production batches. We used Bayesian modeling for statistical inference and evidence synthesis of two sample data sets. Thus, parameter uncertainty was represented by a joint posterior distribution, which we then used to predict the risk and to evaluate the criteria for acceptance of production batches. We also applied the Bayesian model to compare alternative criteria, accounting for the statistical uncertainty of parameters, conditional on the data sets. Comparison of the posterior mean relative risk, $E(\mathit{RR}|\mathrm{data})=E(P(\mathrm{illness}|\mathrm{criterion is met})/P(\mathrm{illness})|\mathrm{data})$, and relative posterior risk, $\mathit{RPR}=P(\mathrm{illness}|\mathrm{data, criterion is met})/P(\mathrm{illness}|\mathrm{data})$, showed very similar results, but computing is more efficient for RPR. Based on the sample data, together with the QMRA model, one could achieve a relative risk of 0.4 by insisting that the default criterion be fulfilled for acceptance of each batch., Published at http://dx.doi.org/10.1214/15-AOAS845 in the Annals of Applied Statistics (http://www.imstat.org/aoas/) by the Institute of Mathematical Statistics (http://www.imstat.org)

Ten copies each of a mail survey was designed and distributed to 87 Finnish food manufacturing companies in order to be distributed to 870 employees representing both workers and managers. Respondents were asked about their attitudes towards food hygiene management strategies in their companies. The final response rates for companies and individual employees were 34.9% and 21.2%, respectively. Answers were stratified according to four job categories and four industry sectors (meat, dairy, fish and bakery). The employees’ attitudes towards various surveyed risk management practices were exclusively positive, regardless of job category or industry sector. All 30 companies that responded to the survey had a functioning own-checking plan (OCP), while other quality management programs were less prevalent. When asked what had caused most difficulties in devising the OCP/HACCP plan the most common answers were choosing the critical control points, committing the firm’s entire workforce and organizing the documentation of monitoring results. According to the respondents, the biggest benefits of the OCP/HACCP plan were product safety and quality.

The Finnish salmonella control program (FSCP) for beef production is based on both randomized and selective testing of herds and animals. Sampling of individual animals in abattoirs is randomized. Herds are selectively tested for salmonella on the basis of clinical symptoms and/or other factors. Risk assessment of FSCP is inherently difficult due to the complexity of the complete data set, especially if the detailed labeling of the testing types is lost. However, for a risk assessment of the whole production chain, methods for exploiting all available data should be considered. For this purpose, a hierarchical Bayesian model of true salmonella prevalence was constructed to combine information at different levels of aggregation: herds in geographical regions and individual animals arriving for slaughter. The conditional (municipality specific) probability of selection of a herd for testing was modeled given the underlying true infection status of the herd and information about the general sampling activity in the specific region. The model also accounted for the overall sensitivity of the sampling methods, both at the herd and at the animal level. In 1999, the 95% posterior probability intervals of true salmonella prevalence in the herd population, in individual cattle, and in slaughter animal populations were [0.54%, 1.4%] (mode 0.8%), [0.15%, 0.39%] (mode 0.2%), and [0.12%, 0.36%] (mode 0.2%), respectively. The results will be further exploited in other risk assessments focusing on the subsequent parts of the beef production chain, and in evaluation of the FSCP.

The requirements of implementing Hazard Analysis and Critical Control Points (HACCP) principles in food production are increasing. A practical risk quantification model, HYGRAM, was developed for small and medium-sized enterprises to meet this challenge. The model makes the user familiar with the HACCP principles by software-assisted guidance through the procedure, connecting special microbiological hazards, good hygiene practice, and other prerequisite programs to HACCP. HYGRAM is a tool to analyze and quantify risks of different processes, and to compare them. It is developed to relieve enterprises with limited resources in confirming the food safety of their production.

Enterohemorrhagic Escherichia coli (EHEC) has become a threat in the modern cattle sector because of its adverse impact on human health. Systems have been developed to reduce the risk of EHEC infection associated with the beef production chain. In Finland, the risk management of EHEC is mainly targeted at primary production, which is controlled by a national program. The prevalence of E. coli O157 in slaughter animals and herds appears to have remained relatively low over the years (0.2 to 1.2% and 0.3 to 1.5%, respectively). The effectiveness of the Finnish EHEC control program (FECP) was analyzed with a Bayesian statistical model based on the results from 2006 through 2010. According to the model, the estimated true prevalence of EHEC in slaughter animals was at its highest in 2007 (95% credible interval [CI], 0.94 to 1.85% of animals), and the estimated true prevalence in herds was its highest in 2007 (95% CI, 1.28 to 2.55% of herds). However, the estimated probability of the FECP detecting an EHEC-positive slaughter animal or herd was 0.52 to 0.58% and 4.74 to 6.49%, respectively. The inability to detect EHEC-positive animals was partly due to animal-based random sampling, which ignores herd-level testing and therefore emphasizes the testing of slaughter animals from herds that send more animals to slaughter. Some slaughterhouses collected samples incorrectly as a consequence of an incorrectly implemented FECP. Farmers may also have questionable reasons for choosing to send animals to be slaughtered in small abattoirs, in which testing is less likely, to avoid suspicion of EHEC or other zoonotic infections.

The Finnish Salmonella Control Programme and the special guarantees (SG) of import concerning Salmonella in the beef production chain were examined within the risk analysis framework. The appropriate level of protection (ALOP de facto since not referred to as ALOP in regulation), performance objectives (PO), and microbiological criteria (MC) were identified along the beef production chain. A quantitative microbiological risk assessment (QMRA) model using the Bayesian probabilistic method was developed for the beef chain to evaluate the capability of different POs to contribute to the ALOP. The influence of SGs was studied as an intervention protecting Finnish consumers. The QMRA made it possible to translate an ALOP without a stated food safety objective (FSO) to POs when implemented for both ready-to-eat (RTE) and non-RTE products. According to the results, the Finnish ALOP de facto for beef, beef preparations and products (10 human Salmonella cases/100,000) was reached in all of the years 1996–2004. However, if the prevalence at the slaughter, domestic cut beef, and retail levels would increase to the level of POs set (maximum 1%), the ALOP de facto would be exceeded by a factor of roughly two. On the other hand, the zero tolerance applied to MCs would keep the true Salmonella prevalence at production steps with POs clearly below 1%, and the ALOP would then be achievable. The influence of SGs on the total exposure was so small (average 0.1% added to the total prevalence of beef-derived foods at retail) that their relevance may be doubted with the current amount and Salmonella prevalence in beef-derived imports. On the other hand, a change in import profile could increase the protective effect of the SGs. Although practical follow-up has to be carried out as apparent prevalences, the objectives and criteria should be estimated as true prevalences and incidences with quantified uncertainties in order to achieve a sound, transparent scientific-based understanding of the risk. The QMRA model developed here provided a useful tool for studying the connection between the ALOP and other targets in the production chain.

Additional guarantees (AGs) for Salmonella in imported defined animal-derived foods were agreed on for Finland when it was admitted to the European Community. The aim of this project was to evaluate the impact of these AGs on the prevalence of Salmonella in the Finnish beef supply and the adequacy of their scope. According to the quantitative Bayesian model, the efficacy of AGs was mainly dependent on the proportions of different beef categories imported and the true prevalence in the countries of origin. According to the model, AGs were able to reach their target in the referred year 1999 and kept the true Salmonella prevalence of beef imports below 1% with quantified uncertainty. The extension of AGs to all imported fresh beef would have reduced the Salmonella prevalence of beef imports from three- to fourfold, whereas expanding the implementation of AGs to all imports of fresh beef, beef preparations, and beef products would have resulted in a sixfold decrease. If current AGs targeting fresh beef intended to be sold as fresh or to be processed by the Finnish industry with processes not achieving 70 degrees C were not implemented, the 95% credible interval of Salmonella prevalence in the Finnish beef supply would be 0.2 to 1.3% (mean, 0.6%) instead of 0.1 to 1.2% (mean, 0.5%). However, if the prevalence in the exporting countries were to rise or the main import countries and/or magnitudes were to change, AGs would be of greater importance.