SDS Authoring – Man vs. Machine

How accurate are Safety Data Sheets authored using only software to determine classifications?

SDS Authoring Services - Man vs. MachineSome software companies tout that you can create and publish Safety Data Sheets in minutes instead of days or weeks, without regulatory or chemical expertise. While many of the SDS authoring processes can be automated, classification of substances and mixtures is still a complex art form.

SDS Authoring Services

The subtle differences between the regulations requires sophisticated programming to address those slight variations which may lead to a classification with one agency but not the others. In addition, there are distinct nuances within an agency’s regulation which requires a certain amount of expert judgement when determining a classification for physical, health or environmental hazards. There are many gray areas that require years of experience with regulations, manufacturing and use of various products in order to apply a proper classification that protects human health and the environment. Listed below are a few examples which exemplify this very concept.

Example 1

The following substance data has been published in a well-known database for Acute Toxicity via Inhalation route: LC50 = 4.5 mg/L. Without knowing which method of exposure was used for the study (vapor vs. dust/mist); a classification is not achievable. Therefore, plugging this number into a software program does not provide a useful calculation result. It could be a Category 3 under the vapor exposure or a Category 4 under the dust/mist exposure. Further research and expert human judgement is needed to make an appropriate determination.

Example 2

In a similar scenario, using a dust/mist exposure of LC50 > 4.5 mg/L, it would be reasonable to assume that this material would be a Category 4 classification. However, the numerical test data alone does not provide a complete reflection of the data. Expert judgement is needed to review the study to understand that the numerical test data is the maximum concentration achievable and that all animals in the study survived therefore a classification is not warranted.

Example 3

When calculating an Acute Toxicity Estimate (ATE) of a mixture, some assumptions and expert judgement need to be considered. This is especially relevant when the formula has ingredients with unknown data for specific routes of exposure and when the mixture formulation is complex. For example, ingredients that are common in foods may not have Acute Toxicity data for all routes of exposure. We use our expert judgement when determining if an ingredient has ‘unknown’ toxicity versus ‘classification criteria not met’ via a particular route. This determination has a significant impact on the outcome of the ATE result and overall mixture classification. Software is not easily programmed to account for all types of exceptions to these rules. Another common scenario encountered is for mixtures with ingredient ranges resulting in greater than 100%. A proper ATE calculation must account for no more than 100% of the formulation. In these complex formulations, common sense and good judgement must be used to be able to make a calculation that is defensible and repeatable. Computers can run millions of calculations a minute but how do you know what data was used in the calculation and if it is selecting the logical one for a hazard determination.

There is no magic wand when it comes to determining substance or mixture classifications. A certain amount of logic, common sense and professional judgement must be applied and that is not yet something that can be built into a software program.

The SiteHawk Authoring Services Department has over 60 years combined experience authoring Safety Data Sheets and classifying materials. SiteHawk uses the right blend of technology and human expertise to tailor solutions for each customer’s specific need and budget.  SiteHawk’s powerful authoring system includes a robust database of toxicological data for physical, health and environmental hazards from many sources such as RTECS, Toxnet, HPV Challenge Program, IUCLID data sets, OECD SIDS program just to name a few. Each piece of data provides a part to the classification puzzle for a material and is analyzed individually and collectively with the remaining data to provide a rational classification.

About Sheila Westerveld

Sheila Westerveld has a BS Chemistry, Stockton State College; MS Environmental Science, Pace University. She has worked in the Chemical Regulatory Compliance field for over 30 years. At SiteHawk she provides subject matter expertise for SDS authoring services and material classification. She has previously worked in the paint and lubricant industries, notifying new chemicals under the various US, EU and Asian regulations. She is experienced in Hazardous Material Transport for US DOT, Canada TMG, EU ADR/ADN, UN IMDG and IATA agencies. She is also HAZWOPER trained.

About SiteHawk

SiteHawk is the global leader in chemical data Intelligence and compliance solutions. Our cloud-based solutions and services deliver a complete approach to SDS and chemical data management, providing data, intelligence and reporting to support safety, compliance and risk management. Organizations in virtually every industry worldwide utilize SiteHawk solutions to manage their hazard communication programs, meet EHS compliance and product stewardship initiatives, manage chemical inventories and data, publish safety data sheets and promote workplace safety.