Page 123 - Freshwater-Biology-and-Ecology-Handbook

The audit also enabled an acceptable standard of analysis to the number of families recorded in standard analyses was
be identified for standard RIVPACS samples – Figure 2.20. increased, which enabled them to be included in the current
Based on what was achieved by some laboratories in the 1990 WHPT indices. In 2013, the Environment Agency adopted
audit, an average of no more than 2 gains (families found by mixed taxon analysis. Despite the fact that many of the
the auditor that the original analyst missed) was considered additional taxa were more difficult to recognise and caused
to be a realistic standard for samples analysed to the level more errors, the data standard remained at an average of no
required for BMWP indices, and it was adopted by the then more than 2 gains at family-level.
regulatory agencies in Great Britain (the NRA (National Rivers
Authority) in England and Wales, the River Purification Boards Performance rapidly improves when a laboratory is
in Scotland and the Department of the Environment for audited, as analysts have errors brought to their attention
Northern Ireland), and that standard is still used. Laboratories and are therefore able to act. Over the years, the quality
and individual analysts that do not meet this standard have to of Environment Agency laboratory analysis became more
take action to restore their quality to this standard. In 2000, stable, so after 2011 it stopped auditing to reduce costs.

Audit Quality
4
3.5
3
2.5
Gains 1.5 2

1
0.5
0
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

lcl ucl Average
Figure 2.20
Results of the audits of the Environment Agency laboratory analyses of invertebrate samples.
Gains are families identified by the auditor that the original analyst failed to record.

Although the Environment Agency has stopped the
independent auditing of its laboratories every year,
each of its laboratories continue to undertake internal
checks for both sorting and identification. One in
ten samples analysed are selected by a true random
method and are re-analysed within the laboratory
by another experienced analyst. In the past, internal
checking was linked to a statistical quality control
procedure to ensure that the laboratory achieved
the analytical target of an annual average of no
more than two gains. In more recent times the
Environment Agency laboratories have not analysed
a sufficient number of samples for the procedure
to work effectively. To overcome this, invertebrate
identification, which the Environment Agency
undertakes to mixed taxon level, is checked by a ring
test in which pre-identified samples are posted to the
analysts to check their identification skills.
Figure 2.21
Methods for these quality assurance schemes are Front cover of the Environment Agency’s
described in the Environment Agency’s Operational Operational Instruction on quality assurance for
Instruction for Quality Assurance. (37) laboratory analysis of invertebrate samples (37)

Freshwater Biology and Ecology Handbook | 123
–