1. Introduction

Welcome to the online Help of the Exposure Assessment Expert System (SEEE). The software SEEE was developed by an interdisciplinary team from the Institute of Occupational Health Sciences, Lausanne, Switzerland, with the financial support of afsset. The main goal of SEEE is to assist industrial hygienists in the assessment of exposure to airborne contaminants in the workplace. SEEE is by no means a software that makes the decision by itself, but a functional tool that supplies some modern methods of exposure assessment and occupational risk management to practitioners working in the field of industrial hygiene. It focuses on the prospective assessment of chronic, long-term exposure to chemicals. We advocate its use only by competent, well trained persons. The user must recognize the constraints and limitations subject to the proper use of SEEE and bear the responsibility for such use. Finally, SEEE has not been developed for use as a legal standard.

Three different approaches of exposure assessment are at the roots of SEEE,and can be combined in a Bayesian approach:

• exposure assessment via measurements of airborne substances hazardous to health in the workplace atmosphere;
• exposure assessment via the use of a 2 compartment physical model, based on physical characteristics such as the emission, dispersion and geometry of the workplace;
• exposure assessment via the use of an empirical neural network model, which, in addition to physical characteristics of the workplace, uses general determinants of exposure, and expert judgments in particular.

The user can exercise either one approach separately, or any combination of these 3 approaches. The combination of all approaches results in the main output of SEEE, and is known to give a more reliable and precise evaluation of the exposure than any model used independently.

For more background information concerning the development and the specific characteristics of this tool, the user can consult the following references:

• Sottas PE, Bruzzi R, Vernez D, Mann S, Guillemin M, Droz PO (2005). A Bayesian framework for the combination of different occupational exposure assessment methods. IOHA 2005 Proceedings.
• Sottas PE, Lavoué J, Bruzzi R, Vernez D, Charrière N, Droz PO (2007), An empirical hierarchical Bayesian approach for unifying occupational exposure assessment methods. Statitics in Medicine (submitted).

1.1. General use

SEEE allows the user to assess the exposure at a workplace for an airborne contaminant for which an Occupational Exposure Limit (OEL) of some sort exists. A session of SEEE should be opened for exposure assessment at a single workplace. For the assessment of exposure in multiple workplaces, the user should open multiple sessions of SEEE simultaneously.

1.2. Version

Version 1.1. beta

• Development and validation : Pierre-Edouard Sottas, Pierre-Olivier Droz, David Vernez, Raffaella Bruzzi, Nicole Charrière, Jérome Lavoué, Institute of Occupational Health Sciences, Rue du Bugnon 21, 1005 Lausanne, Switzerland.
• Design:  Pierre-Edouard Sottas.

1.3. Copyright and responsibility

SEEE is a copyrighted computer software which is made available for use free of charge. The software SEEE, and/or one of its part, and/or the results obtained by SEEE, cannot be sold for commercial purposes without the written agreement of the proprietors of SEEE. The user recognizes the constraints and limitations of a proper use of SEEE. The tool is supplied as is, and the responsibility and/or liability of the owners of EASE cannot be searched and engaged in any case.

1.4. Ownership

SEEE is proprietary of the Institute of Occupational Health Sciences, Institut universitaire romand de Santé au Travail, Rue du Bugnon 21, 1005 Lausanne, Switzerland.

2. Datasheet

The process of data collection is separated into four distinct parts:

• general information about the workplace.
• exposure determinants.
• confidence levels in the different methods and measurements in the workplace.
• results.

3. Strategy

Three different strategies are proposed:

• descriptive statistics (predicted exposure distribution defined by its geometric mean GM and geometric standard deviation GSD).
• mean exposure strategy (estimation of the arithmetic mean of the day-to-day exposure distribution)
• exceedance strategy (estimation of the probability of exceedance of the chosen OEL.

4. General

General information about the workplace and the conditions of exposure. At the exception of the threshold limit value and its corresponding unit, this information has no effect on exposure assessment in this version. Filling all fields is therefore not mandatory.

4.1. General information

4.1.1. Date

Date of exposure assessment. Date as of today is proposed by default.

4.1.2. Occupation

Occupation of the worker. The user has the choice between:

• the International Standard Classification of Occupation, ISCO 1988 (default).
• the Swiss classification of the Federal Office of Statistics, “Répertoire des professions”, OFS 1990 (in French only).
• to load a database created by the user. The file should be an Microsoft Excel file whose first column contains the list of the occupations.
• to enter the occupation manually.

4.1.3. Industry

Industry class or type. The user has the choice between:

• the Swiss classification of the Federal Office of Statistics, NOGA 2002 (in French only)
• the French classification NAF 1.1 (in French only)
• the International Standard Industrial Classification, ISIC 3.1 (default).
• to load a database created by the user. The file should be an Microsoft Excel file whose first column contains the list of the industries.
• to enter the industry manually. 

4.1.4 Others

This field contain in free text all other useful information related to the workplace and the activity under study.

4.2. Contaminant

Information about the air contaminant. This includes:

• the name of the contaminant
• the long-term occupational exposure limit (OEL)
• the unit: either ppm or mg/m3
• the CAS registry number of the chemical, as given by the Chemical Abstracts Service (CAS): http://www.cas.org

4.2.1. VME

Following the Swiss SUVA 2007 list of air contaminants (in French only).

4.2.2. TLV

Default in the English version. Follows the 2006 edition of the Threshold Limit Values (TLV) published by ACGIH: http://www.acgih.org

4.2.3. Other database

The user can load its own database of air contaminants. The file should be a Microsoft Excel file whose columns have this information:

• first column: contaminant number (integer).
• second column: contaminant name (text).
• third column: CAS registry number of the contaminant (text).
• fourth column: unit of the contaminant, 0=ppm, 1=mg/m3 (binary).
• fifth column: long term OEL in ppm (mandatory if the unit is given in ppm).
• sixth column: long term OEL in mg/m3 (mandatory for all substances).

4.2.4. Manual entry

The OEL should be given in mg/m3 and is a mandatory field. If the OEL is unknown, the user should give any other reference value or recommendation.

5. Determinants

Exposure determinants give information on factors that influence exposure levels. They are either quantitative or semi-quantitative variables. They are integrated in SEEE in probabilistic figures, because a precise estimation may be difficult in practice. The user is asked to give the lower and higher bounds of the interval of possible values of the determinants. These bounds include both the incertaincy in the estimation and the temporal variability. Predefined categories are also proposed to assist in the introduction of the determinants. The category unknown determinant or impossible estimation may be chosen if the determinant cannot be estimated. 

5.1. Emission

Determinants characterizing the emission of the contaminant.

5.1.1. Average emission

Average emission of the source of the pollutant, or sources if they are multiple. The interval represents the lower and higher bounds of possible emission levels over a long period, typically one month or year. The unit is a mass [milligram, gram, kilogram] per unit of time [second, minute, hour, day]. In practice, it is often possible to estimate the emission via a mass balance. For example, a can of 1.5 litre of contaminant consumed in 30 days, with a density of 0.8 kilo per litre, results in an average emission of about 40 grams per day (40=1.5*0.8*1000/30), and can be introduced as 20-80 grams/day to take into account the uncertainty in the estimation.

5.1.2. Measured fraction

Fraction of the emitted contaminant which made airborne et which is measured by the given method or instrument.. This parameter takes into account that for certain situations only a fraction of the emitted material will become airborne (for example sedimentation of large aerosol particles), but also that the method used might not measure the total of what is airborne (for example when sampling the respirable fraction of aerosols). The unit is the percentage. For gaz and vapours this parameter is in principle 100%, for aerosols it might be smaller then 100% depending on losses in emission, on aerodynamic diameter and on toxicological considerations. The input is an interval of possible values of measured fraction (for example 50-70% for an aerosol with particles of medium size), or one of 6 predefined categories.

5.1.3. Initial momentum

Speed of the pollutant at the source or at the interface with the environment if the source is enclosed, even partially. The unit is the meter per second (m/s). The answer is either an interval of possible speeds (for example 0.8-1.2 m/s for a emission at high speed), or one of 4 predefined categories.

5.1.4. Local exhaust efficiency

Efficiency of the local exhaust or system used to reduce source emission. The unit is the percentage. The answer is either an interval of possible levels of efficiency of local exhaust, or one of 5 predefined classes.

5.2. Dispersion

Determinants characterizing the dispersion of the contaminant in the workplace.  

5.2.1. Air turbulence

Turbulence is defined here as the average omnidirectional air velocity in the zone between the source and the worker. The unit is the meter per second. The input is either an interval of possible values of velocities, or one of 4 categories.

5.2.3. Volume

Volume of the workplace. The unit is the cubic meter. The input is either an interval of possible values of volume, or one of 6 predefined categories.    

5.2.2. General ventilation

The total ventilation, given by the air flow in the workplace, from either a mechanical, dedicated installation, or from natural ventilation (caused by opened windows for example). The unit is the cubic meter per hour. The input is either an interval of possible levels of air flow, or one of 4 predefined categories.

5.2.4. Bias of dispersion

Bias on exposure due to the lack of homogeneity of the spatial distribution of the ventilation or of the geometry of the workplace. The unit is the percentage. Good ventilation conditions in the workplace can decrease exposure levels and are represented here by a value of less then 100%, a value of 100% is associated with homogenous conditions of ventilation, and a bad distribution of the ventilation to a value greater then 100%. The input is either an interval, or one of 3 predefined categories.

5.3. Others determinants

5.3.1. Expert judgment

General appreciation by the expert of the long term exposure of the worker. The expert judgment should be carried out independently of measurements of the concentration of the contaminant. It is the appreciation of the expert based on its experience and its unstructured observation of the workplace. The input is chosen among 5 categories.

5.3.2. Other factors

Other factors or determinants independent of those already mentioned that may have a significant impact on exposure. These factors describe some attributes or characteristics of the workplace. Overall working conditions, the behaviour of the worker with regard to exposure prevention, general conditions of hygiene in the workplace, industry type, age and state of repair of the facilities (new/archaic equipment may lead to low/high exposure), are some examples. The input is qualitative and chosen among 5 categories.

5.3.3. Source directivity

Directivity of the source with regard to the worker’s location. The input is one of the 4 categories.

5.3.4. Distance measurement-source

Distance between the worker (if personal sample) or the measurement apparatus (if area sample) and the source (in case of a unique source), or distribution of the distances between the worker and the sources (in case of multiple, disseminated sources). The unit is the meter. The input is either an interval of potential values (that take into account the possible positions of the worker), or one of the 3 predefined categories.

6. Measurements and Methods

6.1. Confidence levels

Level of confidence that the expert may have on the various approaches of exposure assessment. Evaluation is made with a relative scale of 0 to 100. A value of 0 indicates that the method is not adapted for the estimation of this particular workplace, a value of 100 that the method is well suited to this task.

6.1.1. Utility of the physical model

Exposure determinants that are related to the emission, dispersion and geometry of the workplace are the inputs of the physical model. A tricky or imprecise estimation of these parameters leads to a weak utility of a physical model (<50). on="" the="" contrary="" a="" precise="" estimation="" of="" emission="" in="" well="" controlled="" exposition="" suggests="" high="" confidence="" efficacy="" physical="" model="">50).

 6.1.2. Utility of the empirical model

The empirical model mines information from previous assessments of the exposure in other workplaces, through a relationship between determinants and the concentration of the pollutant. Thus, if the workplace has some characteristics that make it significantly different from other workplaces, or if the expert believe that this assessment has some particularities never encountered in the past, the utility of the empirical model is low (<50). on="" the="" contrary="" if="" conditions="" of="" exposure="" share="" strong="" similarities="" with="" often="" met="" in="" past="" utility="" is="" high="">50).

6.1.3. Utility of the measurements

If the general conditions of measurements are difficult, if a good sampling strategy is hard to implement, if the environmental variability can be considered as high at the workplace, if measurement uncertainty is high, the confidence in the results returned by the measurement device is low (<50). on="" the="" contrary="" easy="" conditions="" of="" measurement="" with="" use="" a="" standardized="" and="" well="" controlled="" procedure="" lead="" to="" high="" utility="">50). In the present version of the software, only confidence levels of the physical model and empirical model are used. These are combined to obtain a 100 confidence level. This result is then used as a prior which will be updated by the measurements, with an equivalent weight in the present version The utility of the measurements is reserved for a future version.

6.2. Measurements

Measurements are taken to represent average over time, typically time-weighted average single-shift concentrations. The unit is the same unit as for the threshold limit value. The number of measurements can be equal to zero. In that case, evaluation is only based on physical and empirical models.

6.2.1. Number

Sample or measurement number to trace back to files.

6.2.2 Date

Date of the measurement. The date of the exposure assessment as entered in 4.1.1 is proposed as default.

6.2.3  Value

Numerical value of the measurement. The unit is the unit chosen for the OEL.

6.2.4 Duration

Duration of the measurement. The duration is only indicative and is not taken into account for exposure assessment. The user is responsible to make sure that the measurement result represent the exposure over the same reference period as the chosen OEL.

6.2.5 Location

Location of the measurement device. The instrument can be either located on the worker (personal sample), or located in the workplace (area sample).

6.2.6 Method

Descriptive information on the measurement process in free text.

6.2.7 Number

Number of the sampler. Serves to identify the sampling device or the measurement.

6.2.8. Conditions 

Only informative, these parameters are not taken into account in the assessment.

6.2.8.1. Temperature

Temperature in the room during the measurement, in Celsius degrees (C°).

6.2.8.2. Pressure

Atmospheric pressure in hectopascals (hPa) during the measurement.

6.2.8.3. Humidity

Relative humidity in percents (%) during the measurement.

6.2.9. Others

Other information concerning the measurement in free text. 

7. Results

Different types of results as returned by SEEE depending on the strategy chosen. To update results, press the COMPUTE key.

For a better understanding of the results and their signification it is recommended to refer to the following documents:

Lavoué J, Deadman, JE. (2004). Enquête approfondie en hygiène du travail: Stratégies d'évaluation de l'exposition et d'interprétation des données. In: Manuel d'hygiène du travail - du diagnostic à la maîtrise des facteurs de risque, pp. 391-437. B. Roberge; J.E. Deadman; M. Legris; L. Menard; M. Baril, Eds. Modulo-Griffon, Mont-Royal (QC)

Bullock WH, Ignacio JS. (2006). A strategy for assessing and managing occupational exposures. AIHA Press, Farfay USA 

7.1. Descriptive

Posterior distributions of the geometric mean (GM), geometric standard deviation (GSD) of the worker’s exposures.

7.2. Mean exposure strategy

Statistics for the arithmetic mean concentration of the contaminant concentration, for 5 exposure assessment methods:

• physical model
• empirical model
• empirical and physical model combined according to their relative utilities
• measurements only
• Bayesian combination of all methods : SEEE result. Results are shown either numerically or graphically. The predicted distribution of the arithmetic mean and its comparison with the OEL is presented for the SEEE result (Bayesian approach).

7.3. Exceedance strategy

Statistics for the exceedance rate of the OEL (estimation and confidence interval) for the 5 exposure assessment methods:

• physical model
• empirical model
• empirical and physical model combined according to their relative utilities
• measurements only
• Bayesian combination of all methods : SEEE result

Results are shown either numerically or graphically. The predicted distribution of the arithmetic mean and its comparison with the OEL is presented for the SEEE result (Bayesian approach).

8. Installation

Recommended configuration for SEEE installation:

• Windows 2000 Professional or more recent, 256 Mb RAM, SXGA resolution 1280x1024 or more
• user must have administrator rights.

How to install SEEE:

• run SEEE_pkg.exe; installation can take a few minutes.
• chose language and follow instructions for installation of Matlab Component Runtime
• if the software requires the installation of Microsoft.NET framework, click OK
• run SEEE.exe. Some delay can occur at the first run of the software, due to installation of some software components.

SEEE is a self-installable software that can be downloaded upon request, free of charge.