Validation or Not?

1. Validation or Not?

2. Introduction The past, the present and the future

3. A Project Client Phase 1 RA CONCLUSIONS Phase 2 SI RI ERA QUALITY ASSURANCE PROJECT PLAN DATA

4. QAPP • Spells the Regulations or Requirements for the Project • Sampling Scope and Sampling Procedures (Sampling Plan) • Analytical Methodology • Acceptance Criteria for the Data • Acceptance Criteria for Project Completion Two types of QAPPS: -The useless kind = same as not having one. -The Practical one = Project Bible http://www.deq.state.mi.us/documents/deq-wd-water-nps-QAPPguidance.pdf http://www.epa.gov/glnpo/fund/qareqs.html http://www.epa.gov/quality/qapps.html

5. Major Data Players in QAPP The Sampling Team and the Sampling Plan The Laboratory and the analytical Plan

6. The Sampling Plan -Seems to be easy to understand -Its execution requires a hands on approach -Its design requires a balance between budget and sampling confidence (Too very few samples or too many samples) -Its design requires knowledge of hydro/geology, statistics and a deep understanding of the chemistry and fate of pollutants

7. The Analysis Plan -Needs an experiences chemist to understand the laboratory, its methods capability, legal requirements, certification, internal SOPs, variability in methodologies, etc. -Have a working relationship with the QA officer and/or department leaders.

8. The Data No worries All performed to perfection If: -Phase 1 -Phase 2 -QAPP -Samplers -Laboratory Conclusions are flawless

9. Reality Check -Not everything gets to be planned to perfection -Not all plans are executed to perfection Need a system to evaluate the project and how close to perfection it is

10. Data Validation and Data Usability

11. What do end user’s use data for? • Site assessment • Remedial activities • Risk assessment • Pollution monitoring • Enforcement • Litigation

12. What do end user’s of data want? • “Reliable” data • Data of known quality • Low detection limits • Legally defensible data • Organized reports • Easy to understand reports • Supporting QA/QC data for all final results • ???????

13. Goals and Objectives of Data Validation • Why data are validated?The main objective of data validation is to provide data of known quality. Data of known quality is essential for the decision making process, supporting regulatory enforcement, litigation, and policy decisions regarding human health and safety. • How do you obtain data of known quality?To obtain data of known quality requires upfront planning, communication, and a rigorous QA/QC program prior to sample collection and analysis. These program requirements are considered in the final data review process which involves data validation by well trained and experienced personnel.

14. Levels of Validation • Level 2 (accuracy/precision) • Level 3 (cursory/reduced) • Level 4 (full/complete)

15. Tiered Validation Strategy • 90% Level 3/10% Level 4 • 80% Level 3/20% Level 4

16. Areas of Validation • Holding Times • Method Blanks • Lab Control Samples • Matrix Spikes • Surrogates • Internal Standards • Field QC (Trip blanks, Equip Blanks) • Calibration (Initial, Continuing) • Compound Identification • Compound Quantitation • Method Compliance

17. Areas of Validation • Holding Times • Method Blanks • Lab Control Samples • Matrix Spikes • Surrogates • Internal Standards • Field QC (Trip blanks, Equip Blanks) • Calibration (Initial, Continuing) • Compound Identification • Compound Quantitation • Method Compliance

18. Reference Documents • EPA National Functional Guidelines, Oct 1999 (Organics) • EPA National Functional Guidelines, Oct 2004 (Inorganics) • EPA National Functional Guidelines for Chlorinated Dioxin/Furan Data Review, August 2002”

19. Data Validation Process Log in Packages Log in EDD Review QAAP Perform Validation EPA Level 3/4 2nd Validation Review Report Generation Project Management Review EDD Qualification Final Report

20. Understanding Data Validation Qualifiers • Why are data qualified?Data are qualified to bring to the user’s attention the increased uncertainty in a measurement. This means that qualified data will have uncertainty or bias in excess of our specified limits. The qualifier is a flag (typically an alpha character) which denotes the level of uncertainty.

21. Reasons for Qualifying Data • Calibration error • Results near the detection limit • QC samples with high or low spike recoveries • Misidentification of compounds • Results associated with method blank contamination • Improper quantitation of concentration

22. Common reasons for “J” qualified data • Results between the IDL and CRQL • Instrument calibration does not meet criteria • Samples exceed holding time • Low or high MS or LCS spike recoveries • Low or high surrogate recoveries

23. Usable or Unusable Data The Big “R”, rejected

24. Understanding Data Validation • Data validators can be very conservative. • Data validation requires professional judgement. • Data validation is not data usability. • DQOs should dictate the validation guidelines. • Data validation does not make bad data good • Data validation is not well understood by industry and regulators.

25. Summary • Data validation is a tool to be integrated into data usability. • All data validation flags are not created equal. • DQOs should be used to develop validation criteria.. • Data validation should be performed and managed by professional chemists with hands-on laboratory experience. • Data validation is proportionally a small cost relative to overall project costs.

26. Contact Info Julio Paredes Laboratory Data consultants FL, Inc 112 Kings Way Royal Palm Beach, FL 33411 E-mail: jparedes@ldcfl.com Phone: (561) 753-0483 ftp://www.ldcfl.com