Guidance on data quality and dealing with data gaps

• Author: Clemm, Christan; Löw, Clara; Baron, Yifaat; Moch, Katja; Möller, Martin; Köhler, Andreas R; Gensch, Carl-Otto; Deubzer, Otmar
• Pages: 100-102

Methodology for Identification and Assessment

of substances for inclusion in Annex II under RoHS

100

A.7 Guidance on data quality and dealing with data gaps

This guidance is based on discussions of the Commission expert group accompanying future sub-

stance reviews under Directive 2011/65/EU and a proposal prepared as guidance on data quality

and dealing with data gaps, in which some revisions have been performed.

When do the recommended data quality requirements apply?

The methodology described in the manual consists of three parts. The first two parts (see Chapter

1 on Identification of substances and Chapter 2 on Prioritisation of substances) are aimed at the

prioritisation of substances which will be assessed in the last part (see Chapter 3 on Detailed as-

sessment of substances). The issue of data quality and data gaps is mainly relevant for the im-

plementation of Part three. Therefore, the assessment in stage three is dealt with in this section.

Is there any additional guidance available?

Article 6(1) further specifies that the review shall use publicly available knowledge obtained from

the application of chemical legislation such as REACH. Though this is not understood to mean

that other sources should not be used, it suggests that the review is to be based on publicly avail-

able data. Additional guidance is available in Recital 10 of the RoHS directive, i.e. that measures

should be based on an assessment of available scientific and technical information.

What is the main purpose to define data quality?

The most important reason is to avoid that poor-quality data are used to show that a restriction is

justified or is not justified. The assessment should collect and review all available data and

•only base decisions on results that are non-controversial within the research community; and

•assess thoroughly research that gives unusual and inconsistent data compared to the non-

controversial data and document such uncertainties within the assessment dossier.

Inconsistent data may be correct and usable, but it may also be wrong due to incorrect/unrealistic

testing conditions. If certain data is controversial, but it cannot be proved wrong, it may be used to

indicate the need for further research to allow the closing of a certain gap needed for coming to a

decision.

How can “data quality” be defined?

Data quality for a certain parameter can be described by a set of meta-data (data about data) that

can for instance be related to the data source (literature reference, date, place/region, experi-

mental procedure, test method, standards, reproducibility, uncertainties, owner, author, etc.).

One fundamental requirement for data is the need for a clear and traceable source. Data should

be used and documented in a transparent and reproducible way.

Documented use of meta-data includes an assessment as to whether the data are:

•adequate (useful, certain and accurate);

•relevant (fit for purpose);

•reliable (related to standardised methodology, experimental procedure or test method);