Analysis of Port Accidents and Calibration of Heinrich’s Pyramid

Academics and industry have used the Heinrich Pyramid for decades to justify overall safety theory, risk assessments and accident prevention strategies. Most use Heinrich’s original severity ratios (1:29:300) for accident causation developed in a factory setting. However, to use the Pyramid effectively and mitigate risks/hazards, it must be calibrated to represent specific industry reality. This paper, for the first time, focuses on calibration of Heinrich’s Pyramid to maritime accident data, using databases from the Marine Accident Investigation Branch of the Department for Transport. This research clusters five years (2013-2017) of accident data using K-Means clustering on categorical variables and severity levels of accidents, similar logic to Heinrich’s analysis. This approach and descriptive statistics provide a new set of ratios between accident severity classifications for casualties with a ship (CS) and occupational accidents (OA) separately. Results show that the data does not appear to fall into Heinrich’s Pyramid shape and yields a vastly different and lower ratio to that of Heinrich’s. Especially concerning was that Very Serious and Serious accidents occurred at a 1:5 ratio for CS and 4:1 for OA, very different from Heinrich’s 1:29. While these results calculated a new ratio, it may not represent reality due to accident reporting/investigation requirements under UK law, a lack of an agreed taxonomy of risk and hazard definitions, and likely underreporting of less severe accidents. This is proven by the fact that, in 2017, CS data became pyramid-shaped, after a decrease in the number of accidents and a 17% increase in near-misses.

Language

  • English

Media Info

  • Media Type: Web
  • Features: Figures; References; Tables;
  • Pagination: 20p

Subject/Index Terms

Filing Info

  • Accession Number: 01763794
  • Record Type: Publication
  • Report/Paper Numbers: TRBAM-21-02560
  • Files: TRIS, TRB, ATRI
  • Created Date: Dec 23 2020 11:11AM