Over-reporting significant figures-a significant problem?

Robert C. Hawkins*, Tony Badrick, Peter E. Hickman

*Corresponding author for this work

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

Background: Excessive use of significant figures in numerical data gives a spurious impression of laboratory imprecision to clinicians. We describe reporting practices in 24 Asia-Pacific laboratories, assess whether these reporting formats and those used in the literature can be justified based on actual laboratory performance and outline how to choose the appropriate number of significant places. Methods: Thirty-two laboratories in Asia-Pacific were surveyed as to their reporting practices for serum creatinine, ferritin, sodium and TSH. Imprecision data from the General Serum Chemistry program from the RCPA-AACB Quality Assurance Program (QAP) were used to assess whether the reporting unit magnitude implicitly suggested in Tietz, the RCPA Manual and the General Serum Chemistry program itself was justified. Results: There was a 75% response rate to the survey, with laboratories generally reporting data using unjustifiable deciles. Unit sizes from the RCPA manual, Tietz and the RCPA-AACB QAP were not justified by the majority of laboratories in the RCPA-AACB QAP. Conclusions: The reporting unit size used by many laboratories is not justified by present laboratory performance using a 95% probability level. A consensus on appropriate reporting unit size is needed to encourage laboratories to change their present reporting formats.

Original languageEnglish
Pages (from-to)158-161
Number of pages4
JournalClinica Chimica Acta
Volume375
Issue number1-2
DOIs
Publication statusPublished - Jan 2007
Externally publishedYes

Keywords

  • Imprecision
  • Quality assurance
  • Significant figures

Fingerprint Dive into the research topics of 'Over-reporting significant figures-a significant problem?'. Together they form a unique fingerprint.

Cite this