Performance Specifications for EQAS Graham Jones (Australia) - - PowerPoint PPT Presentation

Performance Specifications for EQAS

Graham Jones (Australia)

Membership

• Chair: Graham Jones (AU) Graham.Jones@svha.org.au
• Stéphanie Albarède (FR) steph.albarede@gmail.com
• Gabriela Gutiérrez (SP) gguti@clinic.ub.es
• Mauro Panteghini (IT) EFLMpresident@efcclm.eu
• Marc Thelen (NL) mthelen@amphia.nl
• Anne Vegard Stavelin (NO) anne.stavelin@noklus.no
• Annette Thomas (UK) annette.thomas2@wales.nhs.uk
• Pat Twomey (UK) taptwomey@aol.com
• Emma Ventura (SP) eventura@seqc.es

External Quality Assessment

• A “Pillar” of Traceability
• Confirms manufacturer’s / laboratory claims
• Assess

– Within-laboratory performance – Between-laboratory performance

• Assess:

– Precision – Bias – Analytical Specificity

EQA Assessment (single Result)

• Assessment: Distance from target

– Target – Distance from Target

• Performance Specification is used to assess

distance from Target

• “Defining Analytical Performance”

(Milan 2014)

Aims

• Apply Milan Criteria to describe EQA

performance specifications

• Develop common performance specifications

based on Milan Criteria

EQA Performance Specifications

• Currently wide variation in practice
• Range between very “tight” and very “loose”
• May be based on:

– Statistics (State of the Art, Milan 3) – Clinical effect (Outcome or BV, Milan 1 or 2) – Other (eg regulatory, combination)

• Can these be harmonised?

• 1. Description of Performance

Specification

• Minimum specification, which all reasonable

laboratories would be expected to pass;

• Expected specification, which most laboratories

should pass but with the aim to improve those that do not meet the specification

• Aspirational specification where some or many

laboratories will not meet until better methods are developed.

• Across this continuum, the specifications would be

expected to move from looser to tighter.

Performance Specifications

(looser) Regulatory (3)SAE ++ All labs/tests pass (tighter) Aspirational (2) Optimal BV Some tests fail

EQA Performance Specifications

• 2. Expected Response
• Regulatory. Laboratories should not fail. An example is

the CLIA regulations in the USA.

• Mandated Response. Failures require mandatory

investigation involving time and effort with compliance.

• Laboratory determined response. Results outside limits

should be followed up, with the amount of effort depending on the nature and severity of the failure.

• Industry response. If limits are too tight for current

methods, the response must come from industry rather than individual laboratories.

• 3. Clinical Implications
• Assays meet a tight specification: indicates that

patients can be appropriately monitored with sequential results from different laboratories where all their assays meet this standard.

• Assays meet a looser specification: monitoring

may be sub-optimal, but sharing of reference intervals is supported.

• Assays do not meet a loose specification: Use of

common decision points not supported, separate reference intervals are required.

Assessment of current EQAS performance specifications

• Surveys / other tools to:

– Compare the types of EQAS performance specifications in use – Assess the information provided by EQAS providers to their customers – Identify best practice

Recommending Specifications

• Consider criteria for “Common Specifications”
• Consider assessment process
• This should focus on “Type 1” EQA programs

(Commutable material, reference measurement for target, repeated samples).

Actions

• Develop terminology to describe EQA

Performance Specifications

• Use terminology to describe current limits
• Support EQA using descriptions to

communicate specifications (and meaning of specification) to clients

• Consider best specifications to meet goals

(May be different for different EQA programs)

Action Plan

• 1. Initial draft proposal circulated ~ 2 weeks ago
• 2. Meeting of most members tomorrow night
• 3. Aim for further meeting at EQALM (Bergen,

October 2015)

Comments / Questions (from me)

• 1. We need a good relationship with EQALM
• 2. Should we engage other non-European EQA

providers (eg CAP, CEQAL)?

Comments / Questions