Chapter 24 — Further Reading

Almost everything worth reading on this subject is free, and most of it is written by the people who investigate the deaths. That is not true of many topics in this book, and you should take advantage of it.

Two rules before you start. First, read the primary source rather than a summary of it — a slide deck about a standard and the standard itself produce measurably different managers. Second, verify currency. Standards are revised, penalty schedules are inflation-indexed, reporting thresholds have changed over the life of the rule, and roughly half of U.S. states run their own approved plans with their own requirements. Nothing below is a substitute for the edition your contract adopts and the rules your jurisdiction enforces.

1. OSHA's construction standards — 29 CFR Part 1926 — read by subpart, in the actual regulatory text. Free on the agency's website. Do not try to read it front to back; read the subparts that own your hazards. M (fall protection), L (scaffolds), P (excavations), R (steel erection), K (electrical), CC (cranes and derricks), AA (confined spaces), Q (concrete and masonry), X (stairways and ladders), and Z (toxic and hazardous substances, including silica). Then read Part 1904 — the recordkeeping rule — because it defines what a recordable case actually is, and half the arguments you will have about a TRIR are arguments about a definition somebody has never read. Finally, read the General Duty Clause in the Occupational Safety and Health Act itself; it is one sentence and it is the rule that catches what the rules miss. For: everyone. This is the vocabulary every other party on your project uses.

2. OSHA's free guidance, training materials, campaigns, and enforcement directives. The agency publishes plain-language guidance, eTools, hazard alerts, and campaign material — including the Focus Four training resources and the long-running national fall prevention campaign run jointly with NIOSH and CPWR — aimed at contractors and workers rather than at lawyers. It also publishes its own compliance directives, which tell you how the agency instructs its inspectors to apply a standard; the multi-employer citation policy, describing the creating, exposing, correcting, and controlling employer roles in §24.2.6, lives there. Read the directive rather than somebody's paraphrase of it, and note that its application has been litigated and has varied across federal circuits and over time, and that state plans may take a different approach. OSHA's Recommended Practices for Safety and Health Programs, including a construction-focused version, is the closest thing to a free blueprint for the program this chapter describes. For: safety orientation, toolbox talks, training your own competent persons, and understanding your controlling-employer exposure.

3. OSHA's On-Site Consultation Program. Free, confidential, delivered through state agencies and universities, and — this is the part contractors do not believe — separate from enforcement. A consultant walks your site, identifies hazards, and helps you build a program, and findings do not generate citations, though serious hazards must be corrected. There is also a recognition program for small employers who achieve exemplary programs. If you run or subcontract to a small company that cannot afford a safety director, this is the single highest-value free resource in American construction and it is chronically underused. For: small and mid-size employers, and any GC trying to lift a weak subcontractor rather than replace them.

4. The Bureau of Labor Statistics — the Census of Fatal Occupational Injuries (CFOI) and the Survey of Occupational Injuries and Illnesses (SOII). This is where the numbers in §24.1 come from, and it is where you should get them yourself rather than repeating mine. CFOI counts fatal work injuries; SOII produces nonfatal injury and illness incidence rates by industry, which is what you need if you want to compare your TRIR to anything honestly. Both are annual and free. Knowing where a number comes from is worth more than knowing the number, because it lets you say "here is the source and here is the year" in a room where somebody else is quoting a statistic they heard in a class. For: anyone who reports safety performance to an owner, a surety, or an executive team.

5. NIOSH — the National Institute for Occupational Safety and Health — and especially the Fatality Assessment and Control Evaluation (FACE) program. NIOSH is the research agency; OSHA is the enforcement agency. FACE investigators reconstruct individual workplace fatalities and publish the reports, and they are the most useful safety documents in existence for a working manager, because they describe the ordinary, reasonable decisions that preceded the event rather than assigning blame. They are hard reading. Use them for toolbox talks and for exercise E2, where you build the five-whys chain yourself from the facts and then apply the substitution test to the report's own recommendations. NIOSH's construction program and the National Occupational Research Agenda's construction sector council publish the research agenda behind much of what OSHA later regulates. For: anyone who wants safety training that lands rather than complies.

6. CPWR — The Center for Construction Research and Training. The construction industry's research institute, NIOSH-funded, affiliated with North America's Building Trades Unions. Two things to know it for: The Construction Chart Book, a periodically updated compilation of construction demographics, injury, illness, and fatality data presented as charts with sources; and its data center and hazard-specific research on falls, silica, opioids in the trades, small-employer safety, and the effectiveness of specific interventions. When you need a defensible construction-specific number — not a general-industry number, and not one from a vendor's white paper — start here. For: anyone building a business case, writing a safety plan, or arguing a point with data.

7. ASSP — the American Society of Safety Professionals — and the consensus standards it stewards. ASSP is the professional body; it also serves as secretariat for a family of American National Standards that matter enormously in construction and that most project managers have never opened. The Z359 family covers fall protection and fall restraint — equipment, anchorages, program requirements, rescue — and is the technical backbone behind the arithmetic in §24.4.1. The A10 series covers construction and demolition operations by topic. Z10 covers occupational health and safety management systems, which is the systems-level framework this chapter argues for. These are consensus standards, not regulations: they are voluntary until your specification, your owner, or your insurer incorporates them by reference, which happens more often than you expect. They also tend to define what "state of the practice" means when somebody is arguing about whether you were reasonable. The international counterpart is ISO 45001, the management-system standard, worth knowing if you work for global owners. For: safety directors, and project managers who write or administer safety specifications.

8. NFPA 70E — Standard for Electrical Safety in the Workplace. The governing consensus standard for energized electrical work: shock and arc-flash boundaries, incident energy analysis, arc-rated clothing, and the written energized-work permit. It matters on a construction site mostly during commissioning, when the permanent system is partially energized and the trades are least sure which panels are live (§24.4.4). Read it alongside NFPA 70, the National Electrical Code, which governs the permanent installation, and NFPA's standard on fire prevention during welding, cutting, and other hot work, which is what your hot-work permit system should be built on. For: superintendents running commissioning, and anyone writing a hot-work or LOTO procedure.

9. The National Safety Council — Injury Facts and the Campbell Institute. Injury Facts is the NSC's long-running compilation of injury statistics across work, home, and transportation, useful for putting occupational risk in context for an audience that does not work in construction. More directly useful for this chapter: the Campbell Institute, NSC's center for environmental, health, and safety leadership, has published practitioner-facing work on leading indicators and on safety management systems — the research foundation under §24.9's argument, in a form you can hand to an executive who wants to know why you are asking for a dashboard instead of a rate. For: making the leading-indicator case to people who control budgets.

10. The systems view of accidents — the literature behind §24.11. This chapter's threshold concept is not original to me. It comes out of a body of work that is worth reading in the order below, and reading even the first two will permanently change how you run an investigation.

• James Reason, Human Error and Managing the Risks of Organizational Accidents — the origin of the "Swiss cheese" model and the distinction between active failures (the plank) and latent conditions (the acceleration). This is §24.11 stated academically, twenty years before I learned it in a trailer.
• Sidney Dekker, The Field Guide to Understanding 'Human Error' and Just Culture — the best available treatment of why "human error" is where an investigation starts rather than where it ends, and of how to hold people accountable without destroying your reporting system. The substitution test in §24.10 comes from this tradition.
• Diane Vaughan, The Challenger Launch Decision — the definitive account of normalization of deviance: how a group of competent, careful people come to treat an abnormal condition as normal, one small step at a time. Read it and you will understand exactly how a piece of loose plywood becomes furniture.
• Erik Hollnagel, on Safety-I and Safety-II — the argument that studying what goes right is more informative than studying the rare occasions when something goes wrong. Useful when you are trying to explain why you are measuring pre-task plans instead of injuries.
• Todd Conklin, Pre-Accident Investigations — the practitioner's version, short and aimed squarely at operations people rather than academics. The best first book if you are going to read exactly one.
• Charles Perrow, Normal Accidents, and Nancy Leveson, Engineering a Safer World — heavier going, for readers who want the structural and systems-engineering arguments in full.

An honest note on the oldest source. You will still encounter H. W. Heinrich's Industrial Accident Prevention everywhere, usually in the form of the accident triangle and the claim that a large majority of accidents are caused by unsafe acts. The triangle's basic shape — many minor events for each serious one — is intuitive and broadly useful. The specific ratios and the attribution to worker behavior rest on data from the 1930s that has been extensively criticized and is not reproducible, and that attribution is precisely the framing this chapter argues against. Know it because it is quoted; do not build a program on it. For: anyone who has been handed a triangle in a training class.

11. The rest of the world, because the hazards are identical and the legal architecture is not. The United Kingdom's Health and Safety Executive administers the Construction (Design and Management) Regulations, which push a meaningful share of safety responsibility onto designers and onto a formal pre-construction coordination role — a structurally different and, in my opinion, smarter allocation than ours, and worth reading even if you never build there. Safe Work Australia publishes the model Work Health and Safety laws, built around a broad primary duty of care and the concept of the "person conducting a business or undertaking." Canada regulates occupational health and safety mostly provincially, and the European Union sets framework directives that member states implement nationally. All four publish their guidance free and in plain language. For: anyone working internationally, and for anyone who wants to see that the multi-employer problem has been solved several different ways.

12. Your own documents — and this is not a filler entry. Before you buy a book, read these, this week:

• Your state's plan — is it a state-plan state, does it cover private construction, and what does it require that federal OSHA does not? Cal/OSHA's written injury and illness prevention program and heat illness rules are the famous example, but they are not the only one.
• Your site-specific safety plan, scored against the fourteen sections in §24.5, and against the only test that matters: could a new foreman run tomorrow's work from it?
• Your subcontractors' plans and competent-person designations, with the authority question asked out loud: can this person stop the work and spend money to fix it, without calling you?
• Your carrier's loss runs and your EMR worksheet. Ask your broker for the worksheet, not the number. It shows you which claims are driving your mod, how the primary and excess portions were split, and which year drops off next — which is the difference between managing an EMR and being told one.
• Your own OSHA 300 log and your last twelve weekly inspection reports. If the findings column is empty, you have just learned something about your program that no book will tell you.

A note on currency and jurisdiction. Standards, penalty schedules, reporting thresholds, recordkeeping requirements, and the treatment of incentive programs under the anti-retaliation provisions have all changed over the life of the rules described in this chapter, and several are actively contested. Requirements differ between federal jurisdiction and the states that run their own approved plans, and differ far more between countries. Workers' compensation is a state system with its own rules everywhere it exists. Where this chapter gives you a topic, a mechanism, or a question to ask, it should travel. Where it gives you a number, go verify it at the source before you put it in a document with your name on it.