From Member Design to Erection Risk: A Prevention-through-Design Framework for Safer Steel Buildings

Abstract

This article argues that safer steel buildings require structural engineers to treat construction safety as a bounded design-information responsibility. The position advanced is neither that engineers should assume control of erection means and methods nor that all site hazards can be foreseen from design documents. Rather, the article contends that engineers should eliminate or reduce foreseeable erection, access, fall, temporary-stability, and maintenance hazards where design can do so; encode safety-critical features in models, drawings, specifications, bids, and contracts; and transfer residual risks explicitly to the parties responsible for construction planning and control.

The article develops a conceptual framework from existing regulatory, professional, and scholarly evidence on prevention through design, steel erection hazards, hierarchy-of-controls reasoning, BIM-enabled design-for-safety review, and legal boundary management. It identifies a practical gap: the literature provides PtD principles, steel-specific interventions, model-based rule concepts, and cautions about liability, but not an integrated steel-building process that links design decisions, lifecycle hazards, information carriers, and residual-risk handoff. In response, the article proposes a four-layer framework of hazard recognition, design intervention, information representation, and boundary-governed transfer, supported by an implementation protocol for scope definition, element-phase reviews, hierarchy-of-controls evaluation, model and document encoding, residual-risk registers, and lifecycle tradeoff review. The framework is presented as a disciplined method for reasoning, documenting, and communicating safe constructability, not as experimental proof of injury reduction. Its significance lies in making steel erection and maintenance risk part of the design record while preserving the distinction between design influence and contractor control.

Introduction

Steel buildings are not made safe only after crews arrive on site. They are partly made safe—or made unnecessarily difficult—when members are sized, connections are detailed, holes are omitted or provided, roof edges are configured, anchor rods are specified, and erection assumptions are either communicated or left implicit. This article develops a conceptual framework for designing steel buildings that are easier and safer to build. Its central concern is not the finished structural adequacy of the building alone, but the sequence through which steel members are fabricated, transported, lifted, connected, stabilized, accessed, inspected, and maintained.

The motivation is empirical and professional. In the United States, fatal work injuries declined from 5,283 in 2023 to 5,070 in 2024, yet construction still accounted for the largest number of private-industry workplace deaths, with 1,034 fatalities in 2024 (U.S. Bureau of Labor Statistics 2026). Steel erection is a concentrated domain of construction risk. OSHA’s steel erection rule identifies hazards that are not merely matters of field supervision: working under loads, hoisting and placing decking, column stability, double connections, steel joists, and falls to lower levels are all affected by design and detailing decisions (Occupational Safety and Health Administration 2001). The same regulatory material records design-relevant requirements such as minimum anchor-rod provisions, double-connection details that maintain support during connecting, restrictions on shear connectors before erection, and perimeter-column devices for safety cables (Occupational Safety and Health Administration 2001).

Prior research has established the general premise that designers influence construction safety. Gambatese, Hinze, and Haas argued that design-for-safety should be treated as a design-support problem, because designers need usable checklists, examples, and tools linked to construction operations rather than general moral exhortation (Gambatese, Hinze, and Haas 1997). Behm’s review of 224 NIOSH fatality investigation reports found that 42 percent of the reviewed fatalities were linked to design-for-construction-safety opportunities, meaning the relevant risk could have been reduced or eliminated had such measures been used (Behm 2005). Gambatese, Behm, and Hinze subsequently situated designing for construction safety within the hierarchy of controls, where eliminating or avoiding hazards before exposure is preferred to relying on administrative controls or personal protective equipment (Gambatese, Behm, and Hinze 2005). NIOSH’s construction PtD toolkit makes the same principle explicit: designing out a hazard is presented as the most reliable and effective form of worker protection because an eliminated hazard does not require downstream inspection, training, supervision, or PPE to succeed (National Institute for Occupational Safety and Health 2024).

The literature also shows why adoption has remained uneven. U.S. designers have historically been kept at a distance from construction safety practice by limited safety education, limited construction-process knowledge, liability concerns, procurement arrangements, OSHA responsibility allocation, and legal precedent assigning site safety primarily to constructors (Gambatese, Hinze, and Haas 1997; Behm 2005). More recent work adds a cognitive and informational limit: Hallowell and Hansen found that approximately 25 percent of construction hazards in their study were latent in design and not identifiable from design documents alone, while designers recognized only 51 percent of identifiable hazards on average (Hallowell and Hansen 2016). Hardison and Hallowell’s review concludes that much of the CHPtD literature rests on logical argument and qualitative evidence, while objective validation of lifecycle risk reduction remains limited (Hardison and Hallowell 2019).

The gap, therefore, is not the absence of PtD ideals. It is the absence of a steel-building framework that simultaneously does four things: connects specific steel design decisions to erection and maintenance hazards; represents those decisions in drawings, models, bids, contracts, and risk registers; accommodates the fact that many hazards are not visible to designers working from drawings alone; and preserves a defensible boundary between design influence and contractor control of means and methods. Steel-specific educational materials and toolkits identify many useful interventions (Toole, Treppel, and Van Nosdall 2013; National Institute for Occupational Safety and Health 2024), while BIM-based DfS research shows how rules can be linked to model elements and residual-risk registers (Hossain et al. 2018). What remains unsettled is how structural engineers should organize these ideas into a bounded, repeatable, steel-focused design process.

The thesis advanced here is that structural engineers should treat construction safety in steel buildings as a bounded design-information responsibility: not as an assumption of site-safety control, but as a disciplined process of eliminating foreseeable steel erection and maintenance hazards where design can do so, embedding safety-critical features in the model and contract documents, and transferring residual risks explicitly to the parties who plan and control the work. This is a stronger claim than “designers should think about safety.” It asserts that safe constructability should become part of the design deliverable for steel buildings, subject to scope definition, interdisciplinary review, and residual-risk communication. The sections that follow develop this claim by first locating steel-building risk in design decisions, then proposing a model-linked PtD architecture, specifying steel design interventions, examining lifecycle cost and productivity implications, addressing counterarguments, and setting out an implementation protocol for professional practice.

Steel Building Design as a Construction-Safety Variable

From finished structure to construction process

Structural design conventionally privileges the completed facility: member strength, stiffness, stability, serviceability, durability, and code compliance. Those concerns remain indispensable. The argument here is that they are incomplete for steel buildings because the design also shapes temporary states. A steel frame exists for weeks or months as a partially completed structure: columns stand before the full lateral system is complete; beams are connected before composite action develops; joists, decking, stairs, and perimeter elements are installed under changing access and fall-exposure conditions. OSHA’s steel erection rule is significant because it treats several of these temporary-state hazards as matters affected by design and detailing, not only by field discipline (Occupational Safety and Health Administration 2001).

Column stability is a clear example. OSHA’s final rule requires a minimum of four anchor rods for columns, and the associated steel-erection safety material identifies inadequate anchor rod or bolt installation as a primary contributor to collapses; it also states that structural collapse is second only to falls as a cause of fatalities in steel erection (Occupational Safety and Health Administration 2001). The designer does not install anchor rods or plumb the column in the field. Yet the base plate, anchor-rod layout, erection assumptions, and tolerances create the conditions under which temporary stability is either robust or fragile. The boundary between design and field execution is real, but it does not make the design neutral.

Double connections provide another illustration. OSHA’s steel rule includes requirements for double-connection details that maintain support during connecting (Occupational Safety and Health Administration 2001). The risk arises because one member may be temporarily unsupported or destabilized while a second member is being connected at the same column or supporting element. A detail that appears efficient in the permanent frame can generate avoidable exposure during erection if it requires a connector to remove support or work in a precarious position. NIOSH’s steel-specific PtD manual accordingly recommends avoiding hanging connections, using beam seats, providing dummy holes for spud wrenches, minimizing bolt sizes where appropriate, and using at least two bolts per connection (Toole, Treppel, and Van Nosdall 2013).

Falls are similarly shaped by design. OSHA identifies falls to lower levels as a major steel-erection hazard (Occupational Safety and Health Administration 2001). NIOSH’s 2024 toolkit recommends designing 42-inch parapets or railings at roof edges, engineering fall-protection anchor points into structural drawings, showing lifeline attachment locations and capacities on contract drawings, and providing column holes at 21 inches and 42 inches for guardrails or lifelines (National Institute for Occupational Safety and Health 2024). These recommendations do not replace the contractor’s fall-protection plan. They make that plan less dependent on improvisation by providing permanent or designed-in features to which protection can be attached.

The hierarchy-of-controls argument for steel PtD

The hierarchy of controls supplies the technical rationale for PtD. Gambatese, Behm, and Hinze connect designing for construction safety to the hierarchy, arguing that eliminating or avoiding hazards before jobsite exposure is preferred to administrative controls and PPE (Gambatese, Behm, and Hinze 2005). NIOSH’s construction toolkit states the point more directly: PtD is “designing out a hazard,” and eliminated hazards are more reliable because they do not require downstream behavior, inspection, or supervision to work (National Institute for Occupational Safety and Health 2024).

In steel buildings, design-stage elimination rarely means eliminating all risk from an operation. More often, it means eliminating a particular exposure. Prefabricating assemblies or ground-assembling stairways can reduce the number of elevated connections; locating rooftop equipment away from roof edges and skylights can reduce maintenance exposure; installing permanent stairs early can reduce reliance on temporary access; providing beam-web holes and column holes can reduce ad hoc attachment practices (Toole and Gambatese 2008; National Institute for Occupational Safety and Health 2024). These are not abstract safety preferences. They change where work occurs, how often workers are exposed at height, and whether temporary protection depends on field improvisation.

This article therefore treats steel PtD as a form of lifecycle engineering. The lifecycle in question is not limited to energy use, material durability, or operational performance. It includes fabrication, transportation, erection, construction-phase inspection, maintenance access, and eventual interaction with permanent systems. The NIOSH steel module explicitly links structural design, detailing, fabrication, transportation, and erection as phases in which design choices affect worker risk (Toole, Treppel, and Van Nosdall 2013). The 2024 NIOSH toolkit extends the same logic to construction and maintenance, including roof-edge protection, mechanical-equipment location, permanent access, and drawing-based communication of fall-protection capacities (National Institute for Occupational Safety and Health 2024).

A conceptual risk transformation

The following expression is an author-generated formalization, not an empirical model. It is included to clarify the framework’s logic. Let H denote the set of foreseeable hazards associated with a steel element or assembly across lifecycle phases. Let D denote design decisions that eliminate or reduce hazards, C denote constructor-controlled means and methods, and R denote residual risks requiring communication. A bounded PtD process can be represented as:

Equation (1) does not calculate injury probability. Its purpose is conceptual: every design alternative should be examined not only for the hazards it reduces but also for hazards it introduces. Rajendran and Gambatese’s roof fall-protection case illustrates this point. A roof-anchor system might appear to provide fall protection, but the case study found that anchors required extensive temporary fall protection during construction and exposed more workers during temporary guardrail installation and leading-edge work; the parapet option was more expensive but safer, and project personnel estimated that the anchor system indirectly reduced worker productivity by at least 15 percent relative to the parapet system (Rajendran and Gambatese 2013). In the terms of Equation (1), the anchor alternative reduced one class of future fall-protection need while introducing construction-phase exposures that had to be considered.

A Model-Linked Framework for Safer Steel Buildings

Why checklists are necessary but insufficient

The earliest design-for-construction-safety work framed the problem as one of usable design support. Gambatese, Hinze, and Haas reported that a Construction Industry Institute study compiled design best practices that could be incorporated into project design to reduce or eliminate site hazards, and they emphasized that designers need tools tied to construction operations (Gambatese, Hinze, and Haas 1997). That insight remains sound. A structural engineer cannot rely on general concern for worker welfare to remember that perimeter columns may need holes or devices for safety cables, that beam seats can reduce connecting hazards, or that top-of-beam tripping hazards should be avoided (Occupational Safety and Health Administration 2001; Toole, Treppel, and Van Nosdall 2013).

Yet checklists alone have two weaknesses. First, they can become disconnected from the modeled element where the hazard arises. A generic reminder to “consider fall protection” is less useful than a model-linked prompt at a roof edge, perimeter column, beam line, stair opening, skylight, or rooftop mechanical platform. Second, checklists do not automatically preserve residual-risk information through design development, bidding, and construction. Hossain et al. developed a BIM-integrated DfS knowledge library that organizes rules as atomic rules, meta-element rules, and meta-rules, and their illustrative five-story BIM case showed how rule-based review can flag design-element risks early and maintain a risk register for residual risks (Hossain et al. 2018). Their work supports the proposition that a steel-building PtD framework should be tied to model elements rather than maintained only as a separate checklist.

The author’s proposal is to treat each safety-relevant steel element as carrying a small set of PtD attributes. These attributes need not make the structural model a construction plan. They should identify the lifecycle phase in which the hazard arises, the design feature intended to eliminate or reduce it, the residual risk if not eliminated, the party expected to address that residual risk during planning, and the drawing or specification location where the information is communicated. This proposal is consistent with NIOSH’s recommendation that PtD controls be written into bids, contracts, plans, and drawings so responsibilities and scope are explicit (National Institute for Occupational Safety and Health 2024), and with Toole and Erger’s argument that PtD can be pursued without implying designer control of site safety if roles, communications, contractual language, and residual-risk transfer are carefully managed (Toole and Erger 2019).

Four layers of the proposed framework

The proposed framework has four layers: hazard recognition, design intervention, information representation, and boundary-governed transfer. Each layer responds to a limitation identified in the literature.

Layer 1: hazard recognition. The framework begins with structured recognition of hazards associated with steel elements and construction phases. Hallowell and Hansen’s findings caution against assuming that designers will recognize hazards unaided: in their experiment, approximately 25 percent of construction hazards were latent in design and not identifiable from design documents alone; for identifiable hazards, average designer recognition was 51 percent; designers with field construction experience performed 45 percent better than those without; and a brief Haddon energy mnemonic intervention improved recognition by 27 percent (Hallowell and Hansen 2016). Steel PtD reviews should therefore include prompts, field-experienced participants where possible, and constructor input when procurement permits. This is not a courtesy add-on; it follows from evidence that recognition skill varies and that some hazards require construction-process knowledge.

Layer 2: design intervention. Once a hazard is recognized, the design team evaluates whether it can be eliminated, reduced, relocated, or made explicit as a residual risk. Steel-specific interventions include prefabricating assemblies to reduce elevated connections, locating column splices approximately four feet above the working floor, providing column holes at 21 inches and 42 inches for guardrails or lifelines, specifying at least four anchor rods in base plates, designing beams with at least six inches of width where workers may walk, designing beams to support lifelines, avoiding hanging connections, using beam seats, preferring shop welds over field welds, and avoiding top-of-beam tripping hazards (Toole, Treppel, and Van Nosdall 2013). The 2024 NIOSH toolkit adds roof-edge parapets or railings, engineered anchor points, rooftop-equipment placement away from edges and skylights, early permanent stairs, prefabricated or ground-assembled stairways, beam-web holes for support or lifeline attachment, and drawing-based lifeline capacities (National Institute for Occupational Safety and Health 2024).

Layer 3: information representation. Design choices must be represented in media that survive handoff. Hossain et al.’s BIM-integrated DfS system demonstrates one route: rule-based checks tied to model elements and a residual-risk register (Hossain et al. 2018). NIOSH’s toolkit emphasizes the contractual route: PtD controls should be written into bids, contracts, plans, and drawings (National Institute for Occupational Safety and Health 2024). The proposed framework combines these routes by requiring that each accepted steel PtD feature appear in at least one model property or coordination log entry and in the applicable contract document when it affects scope, capacity, or required installation.

Layer 4: boundary-governed transfer. The final layer distinguishes design influence from control of means and methods. Behm identifies the U.S. context in which traditional procurement, contract language, OSHA responsibility allocation, and legal precedent often place construction site safety on constructors, reducing designers’ incentives to adopt DfCS as standard practice (Behm 2005). Toole and Erger likewise warn that “pure design” firms face legal and professional-practice risks if PtD is allowed to blur responsibility for construction means and methods (Toole and Erger 2019). The proposed response is not for designers to retreat from PtD, but for them to document what is designed, what assumptions support the design, what residual risks are foreseeable, and who must resolve those risks during construction planning.

Designing the Steel Frame for Erection, Access, and Maintenance

Temporary stability as a design deliverable

Temporary stability is where the distinction between permanent structural design and construction safety becomes least defensible. A steel building that is stable in its completed state may pass through vulnerable configurations during erection. OSHA’s steel erection rule foregrounds column stability, double connections, steel joists, hoisting, placing decking, and falls as major hazards (Occupational Safety and Health Administration 2001). The rule’s design-relevant provisions show that temporary stability cannot be left wholly outside the design conversation.

The framework therefore requires the structural engineer to identify temporary-stability-sensitive elements during design development. This is an author’s proposed professional step, derived from OSHA’s identification of column stability and collapse hazards and NIOSH’s steel-specific interventions. The step does not require the engineer of record to prescribe crane locations, erection sequence, or temporary bracing means unless contracted to do so. It does require that base details, connection concepts, column splice locations, and member interfaces be reviewed for foreseeable erection hazards. NIOSH’s steel module recommends column splices approximately four feet above the working floor, at least four anchor rods in base plates, beam seats, dummy holes for spud wrenches, and avoidance of hanging connections (Toole, Treppel, and Van Nosdall 2013). These items belong in structural detailing because they affect how workers physically make the frame stable.

Steel joists and decking deserve parallel attention. OSHA identifies hoisting and placing decking and steel joists as major steel erection hazards (Occupational Safety and Health Administration 2001). The design team may not control the crew’s exact installation method, but it can reduce hazards by avoiding details that require awkward placement, excessive field welding, or unstable temporary conditions. NIOSH’s recommendation to prefer shop welds over field welds and to avoid top-of-beam tripping hazards is directly relevant here (Toole, Treppel, and Van Nosdall 2013). Where the structural system requires field operations that create residual risks, those risks should be documented for constructor planning rather than buried in assumptions.

Connections as worker interfaces

Connections are often discussed as force-transfer mechanisms. In construction safety terms, they are also worker interfaces. A connector must access the joint, align members, use tools, secure bolts or welds, and often do so at height. The NIOSH steel module’s recommendations—beam seats, dummy holes for spud wrenches, reduced bolt-size burdens where appropriate, at least two bolts per connection, and avoidance of hanging connections—are practical precisely because they address the work of connecting, not merely the final analytical model (Toole, Treppel, and Van Nosdall 2013).

This perspective changes the design review question. Instead of asking only whether the connection has adequate strength and stiffness, the PtD review asks: Can the first worker who reaches this connection make it stable without unsupported balancing, awkward tool use, or unnecessary exposure? Does the connection require field welding where a shop weld would be feasible? Does the geometry create a tripping hazard at the top of the beam? Does the connection support safe sequencing, or does it rely on temporary conditions that must be made explicit? These are author-generated review questions grounded in the steel-specific interventions recorded by NIOSH (Toole, Treppel, and Van Nosdall 2013) and OSHA’s treatment of double connections and erection hazards (Occupational Safety and Health Administration 2001).

Designers may object that connection design is often delegated to fabricators or specialty engineers. That practice does not eliminate PtD responsibility; it relocates part of it into delegated-design criteria and review. The author’s proposed framework therefore requires that safety-relevant connection assumptions be stated in connection-design criteria when they are known to affect erection risk. Examples include requirements for minimum temporary support at double connections, preferred use of shop welds, locations where beam seats are required, and details that avoid top-of-beam obstructions. These requirements are not instructions to the erector about means and methods; they are performance and detailing constraints intended to avoid foreseeable hazards.

Fall protection, access, and roof work

Fall hazards are both immediate during construction and persistent during maintenance. NIOSH’s 2024 toolkit recommends designing 42-inch parapets or railings at roof edges, engineering fall-protection anchor points into structural drawings, locating rooftop mechanical equipment away from edges and skylights, showing lifeline attachment locations and capacities on contract drawings, and providing beam-web holes for support or lifeline attachment (National Institute for Occupational Safety and Health 2024). These measures have different time horizons. Parapets and permanent railings protect construction and maintenance workers. Anchor points and lifeline capacity notes support temporary protection during construction and future work. Equipment placement changes maintenance exposure by spatial design rather than by PPE.

Rajendran and Gambatese’s case study demonstrates why roof fall-protection design should not be judged by first cost alone. In their comparison of roof anchors plus a parapet that did not meet OSHA guardrail height, a parapet meeting OSHA guardrail requirements, and no anchors or parapet, the compliant parapet was more expensive but safer than the roof-anchor system because anchors required extensive temporary fall protection during construction and exposed more workers during temporary guardrail installation and leading-edge work (Rajendran and Gambatese 2013). Project personnel estimated that the roof-anchor system indirectly reduced worker productivity by at least 15 percent compared with the parapet system (Rajendran and Gambatese 2013). The result is not a universal proof that parapets are always superior. It is evidence that lifecycle safety and productivity can reverse conclusions based on first cost.

Access design belongs in the same category. NIOSH recommends building permanent stairways early and using prefabricated or ground-assembled stairways where appropriate (National Institute for Occupational Safety and Health 2024). These choices affect the number of workers using temporary ladders or incomplete access systems. The structural engineer may share this decision with architects, contractors, and owners, but steel framing layouts, stair openings, supports, and sequencing assumptions influence whether early permanent access is feasible. Toole and Gambatese’s trajectories of PtD—greater prefabrication, safer material or system selection, more construction engineering by designers, and spatial design choices that reduce hazards—are directly applicable to stair assemblies, roof access, laydown areas, crane reach, and connection locations in steel buildings (Toole and Gambatese 2008).

Prefabrication and spatial planning

Prefabrication is not automatically safer, but it often changes the location and character of risk. Toole and Gambatese identify greater use of prefabrication as one likely trajectory of construction hazard prevention through design (Toole and Gambatese 2008). NIOSH recommends prefabricating assemblies or assembling multiple steel sections where practical to reduce work at height (National Institute for Occupational Safety and Health 2024). The safety mechanism is straightforward: tasks performed at ground level or in controlled shop conditions can reduce elevated connections, field welding, and repeated access exposures. The framework therefore treats prefabrication as a design alternative to be evaluated, not as a universal mandate.

Spatial planning is equally important. Locating rooftop equipment away from edges and skylights reduces fall exposure during maintenance; providing beam-web holes for lifeline support affects future attachment options; laying out framing so connections are reachable and stable affects erection; and early permanent stair access changes worker circulation (National Institute for Occupational Safety and Health 2024). These choices are made early enough that late-stage safety review may be too late. A PtD review conducted after structural geometry is fixed can still add anchor points or notes, but it may not be able to relocate equipment, alter bay spacing, or make prefabricated assemblies practical. That timing argument is the author’s reasoning, based on the nature of the design decisions identified in the NIOSH toolkit and steel module.

Evidence, Uncertainty, and the Case for a Bounded Framework

What the evidence establishes

The strongest evidence for steel-building PtD is cumulative rather than definitive. The occupational safety burden is real: construction remains the largest private-industry fatality category even in a year when overall U.S. fatal work injuries declined (U.S. Bureau of Labor Statistics 2026). Steel erection hazards are specifically recognized in regulation, and several OSHA requirements directly involve design and detailing (Occupational Safety and Health Administration 2001). NIOSH provides concrete steel-design interventions for structural design, detailing, fabrication, transportation, erection, roof edges, access, and lifeline attachment (Toole, Treppel, and Van Nosdall 2013; National Institute for Occupational Safety and Health 2024). Behm’s fatality-review finding that 42 percent of reviewed fatalities were linked to design-for-construction-safety opportunities gives the premise empirical weight beyond professional intuition (Behm 2005).

The evidence also establishes that implementation cannot be left to individual goodwill. Gambatese, Hinze, and Haas identify limited safety education, limited construction-process knowledge, and liability concerns as reasons designers historically were not directly included in construction safety practice (Gambatese, Hinze, and Haas 1997). Hallowell and Hansen show that hazard recognition is variable and improvable, with field experience and structured mnemonic support improving performance (Hallowell and Hansen 2016). Hossain et al. demonstrate that DfS knowledge can be structured into BIM-integrated rules and residual-risk registers (Hossain et al. 2018). Toole and Erger show that legal boundary management is necessary, especially for pure design firms in the United States (Toole and Erger 2019). Together, these findings support a framework that is procedural, model-linked, and contract-aware.

What the evidence does not establish

The evidence does not establish that every PtD intervention reduces total project risk in every steel building. Hardison and Hallowell’s review is important because it resists overclaiming: they conclude that much of the CHPtD literature is supported by logical argument and qualitative work, while objective validation of actual lifecycle risk reduction remains limited (Hardison and Hallowell 2019). That limitation matters. A parapet may reduce roof-edge fall exposure in one project, but it may interact with architectural constraints, wind loads, façade systems, cost limits, or maintenance requirements in another. Prefabrication may reduce elevated work, but it can introduce transportation, lifting, or load-control challenges. Beam-web holes can support attachment, but they must be structurally coordinated. These project-specific interactions are author reasoning consistent with the digest’s warning that objective validation remains limited.

Nor does the evidence establish that design documents reveal all construction hazards. Hallowell and Hansen’s finding that approximately 25 percent of hazards were latent in design and not identifiable from documents alone means that even a rigorous drawing-based review will miss some hazards (Hallowell and Hansen 2016). This finding undercuts a simplistic version of PtD in which the engineer is expected to foresee all site risk from the office. It also undercuts the opposite claim that designers should do nothing because they cannot see everything. The appropriate inference is bounded: designers should eliminate and communicate foreseeable hazards within their expertise and should seek construction input where hazards depend on means, methods, sequencing, or site conditions.

The evidence also leaves open how PtD should be evaluated quantitatively. Rajendran and Gambatese provide a useful case study with a productivity estimate, but it is not a universal cost model (Rajendran and Gambatese 2013). Behm’s fatality review links many fatalities to design opportunities, but it does not prove that a specific steel-detailing checklist would have prevented each event (Behm 2005). Hossain et al.’s BIM-integrated system shows how rule-based risk review can operate in an illustrative five-story case, but the digest does not report large-scale field validation of injury reduction from that system (Hossain et al. 2018). The framework proposed here is therefore conceptual and practice-oriented. It does not claim to report new experiments, collected data, executed code, or measured injury reductions.

Why the framework is still warranted

A framework is warranted because professional practice often must act before evidence is experimentally complete. The regulatory and technical sources identify specific design-sensitive hazards in steel erection (Occupational Safety and Health Administration 2001; Toole, Treppel, and Van Nosdall 2013). The empirical sources show that design is linked to a nontrivial share of construction fatalities, that hazard recognition is imperfect but improvable, and that design alternatives can change both risk and productivity (Behm 2005; Hallowell and Hansen 2016; Rajendran and Gambatese 2013). The implementation literature shows that knowledge can be formalized in tools and models, while the legal literature shows that role boundaries must be managed rather than ignored (Hossain et al. 2018; Toole and Erger 2019).

The gap is thus practical synthesis. A steel engineer needs a framework that says when to ask safety questions, what kinds of steel details to examine, how to encode decisions, and how to avoid assuming the constructor’s role. The proposed framework answers that need by treating PtD as a design-information process. It is deliberately more modest than a claim that designers control site safety, and more demanding than a claim that safety belongs wholly to contractors.

Counterarguments and Competing Interpretations

Counterargument 1: Construction safety belongs to constructors

The strongest objection is legal and organizational: construction means, methods, sequences, and site safety are traditionally controlled by constructors, not structural designers. Behm records this U.S. context, noting that procurement, contract language, OSHA responsibility allocation, and legal precedent often place construction site safety on constructors, leaving designers without strong legal, contractual, economic, or regulatory incentives to adopt DfCS as standard practice (Behm 2005). Toole and Erger sharpen the concern for pure design firms that do not contract to build or control means and methods; they identify legal and professional-practice risks if PtD is implemented in a way that blurs responsibility (Toole and Erger 2019).

This objection deserves to be taken seriously. A structural engineer who dictates erection sequence without contractual authority, reviews temporary works as though responsible for them, or gives field safety directions outside scope can create confusion and liability. The contractor’s site-specific knowledge, workforce control, equipment selection, and scheduling authority are indispensable. Many hazards depend on weather, site congestion, crane availability, subcontractor coordination, and worker behavior—conditions that are outside ordinary structural design services. Hallowell and Hansen’s finding that some hazards are not identifiable from design documents alone reinforces this point (Hallowell and Hansen 2016).

The response is that the objection defeats an overbroad PtD model, not a bounded one. OSHA’s steel rule itself contains design-relevant requirements for anchor rods, double connections, shear connectors, and perimeter safety-cable devices (Occupational Safety and Health Administration 2001). NIOSH’s steel module states that PtD is compatible with, but distinct from, contractor site-safety responsibility: designers retain an ethical duty to provide constructable drawings and to understand how early decisions influence construction and maintenance hazards (Toole, Treppel, and Van Nosdall 2013). Toole and Erger likewise argue that PtD can be pursued without implying designer control of site safety if roles, communications, contractual language, and residual-risk transfer are managed carefully (Toole and Erger 2019). The framework adopts that bounded view. It asks designers to shape what they design and communicate residual risk, not to supervise how the contractor builds.

Counterargument 2: PtD evidence is too limited for design mandates

A second objection is evidentiary. Hardison and Hallowell conclude that objective validation of lifecycle risk reduction remains limited and that much CHPtD support is logical or qualitative (Hardison and Hallowell 2019). A critic could argue that without stronger experimental or longitudinal evidence, PtD requirements risk becoming prescriptive burdens. They may increase design cost, complicate professional liability, and produce details that appear safer in theory but introduce other hazards during fabrication, transport, or erection.

This critique is partly correct. The framework should not be read as a mandate to adopt every listed intervention on every project. The evidence supports careful evaluation, not mechanical compliance. Rajendran and Gambatese’s roof case is instructive precisely because it compares alternatives and finds that the safer option was not the least expensive first-cost option (Rajendran and Gambatese 2013). The case supports lifecycle reasoning; it does not justify universal parapet selection regardless of context. Similarly, prefabrication can reduce work at height, but Toole and Gambatese present prefabrication as a trajectory of PtD, not as an unqualified rule (Toole and Gambatese 2008).

The response is that limited objective evidence calls for disciplined documentation, not inaction. The hierarchy of controls remains a well-established safety logic in the cited PtD literature, and designing out hazards is more reliable than relying on downstream controls when elimination is feasible (Gambatese, Behm, and Hinze 2005; National Institute for Occupational Safety and Health 2024). The framework therefore uses a review-and-record model: identify the hazard, evaluate alternatives, document the selected design feature, and record residual risks. This approach generates the kind of project-level information that future objective research could use, while allowing current projects to benefit from known steel-specific design measures.

Counterargument 3: Designers lack the construction expertise to recognize hazards

A third objection is cognitive rather than legal. Hallowell and Hansen show that designer hazard recognition is limited: average recognition of identifiable hazards was 51 percent, and field experience significantly improved performance (Hallowell and Hansen 2016). A critic could argue that asking structural engineers to perform PtD reviews will create false confidence. Designers may miss hazards, misunderstand erection practice, or select interventions that complicate field work.

This is not a reason to abandon PtD; it is a reason to structure it. Hallowell and Hansen also found that a brief Haddon energy mnemonic intervention improved hazard recognition by 27 percent, and that designers with field construction experience performed 45 percent better than those without (Hallowell and Hansen 2016). Those findings imply that recognition skill can be improved through prompts and experience. Hossain et al.’s BIM-integrated DfS knowledge library offers a complementary strategy: formalize knowledge into rules linked to model elements and maintain a residual-risk register (Hossain et al. 2018). The proposed framework therefore requires structured prompts, model-linked checks, and constructor or field-experienced review where feasible.

The deeper point is that PtD is not solitary prediction. It is a communication process. Designers should not pretend to know every construction method. They should identify foreseeable hazards created or affected by design, solicit input on hazards that depend on erection planning, and record residual risks in a form that constructors can use. This position aligns with NIOSH’s recommendation that PtD controls be written into bids, contracts, plans, and drawings (National Institute for Occupational Safety and Health 2024) and with Toole and Erger’s emphasis on communications and residual-risk transfer (Toole and Erger 2019).

Counterargument 4: Safer design will increase first cost and reduce competitiveness

A fourth objection is economic. Some PtD measures add material, detailing time, coordination effort, or design fees. A 42-inch parapet may cost more than a lower parapet with anchors; prefabrication may require earlier coordination; additional holes, seats, anchor points, or connection constraints may affect fabrication and design review. In competitive procurement, owners may resist visible first-cost increases unless safety benefits are contractually valued.

The strongest response is not that safety is always cheaper. The cited evidence does not support that claim. Rajendran and Gambatese found that the parapet option was more expensive but safer than the roof-anchor alternative in their case study (Rajendran and Gambatese 2013). The better argument is that first cost is the wrong boundary for evaluating design alternatives. In the same case, project personnel estimated that the roof-anchor system indirectly reduced worker productivity by at least 15 percent compared with the parapet system because of temporary fall-protection demands and leading-edge exposure (Rajendran and Gambatese 2013). A design that appears cheaper in the structural or architectural budget can shift cost and risk into construction operations.

Owners are central here. Toole and Erger note that PtD has been adopted in the United States especially by large safety-conscious owners and design-build or integrated firms, while Europe has required PtD for decades (Toole and Erger 2019). The procurement implication is clear: if owners want safer steel construction, they must request, price, and reward PtD services. NIOSH’s toolkit supports writing PtD controls into bids, contracts, plans, and drawings so responsibilities and scope are explicit (National Institute for Occupational Safety and Health 2024). Without that contractual recognition, designers may be asked to absorb additional coordination and liability without compensation or authority.

A Bounded Implementation Protocol for Steel Project Delivery

Stage 1: establish PtD scope and responsibility

The protocol begins before member sizes and connection concepts harden. The owner, structural engineer, architect, construction manager or contractor where available, and relevant specialty participants should define whether the project includes steel PtD services, what deliverables will be produced, and how residual risks will be communicated. This is an author’s proposed implementation step, grounded in NIOSH’s recommendation to include PtD controls in bids, contracts, plans, and drawings and Toole and Erger’s emphasis on roles, communications, contractual language, and residual-risk transfer (National Institute for Occupational Safety and Health 2024; Toole and Erger 2019).

The scope should state that the designer’s PtD role is to identify and reduce hazards affected by design, not to assume responsibility for contractor-controlled means and methods unless the contract expressly assigns construction engineering duties. This distinction responds to the U.S. barriers identified by Behm and the legal cautions identified by Toole and Erger (Behm 2005; Toole and Erger 2019). Where project delivery is integrated or design-build, the scope may allow closer coordination of erection engineering. Where the structural engineer is a pure design consultant, the scope should focus on design features, assumptions, and residual-risk communication.

Stage 2: conduct element-phase hazard reviews

The core review should be organized by steel element and lifecycle phase. For each major element category—columns, base plates, beams, connections, joists, decking, stairs, roof edges, anchor points, mechanical supports, and maintenance-access zones—the team asks how workers will fabricate, transport, lift, connect, stabilize, access, inspect, and maintain the element. This element-phase structure is an author’s synthesis of the NIOSH steel module’s linkage of structural design, detailing, fabrication, transportation, and erection, and the NIOSH toolkit’s construction and maintenance controls (Toole, Treppel, and Van Nosdall 2013; National Institute for Occupational Safety and Health 2024).

Because hazard recognition is imperfect, the review should not rely on unprompted designer intuition. Hallowell and Hansen’s findings support structured prompts, field-experienced participation, and mnemonic aids (Hallowell and Hansen 2016). A practical review prompt set for steel buildings should include at least the following author-proposed questions, each grounded in hazards or interventions identified in OSHA and NIOSH sources:

• Does the column-base design support stable erection, including anchor-rod quantity and layout? (Occupational Safety and Health Administration 2001; Toole, Treppel, and Van Nosdall 2013)
• Do double connections maintain support during connecting? (Occupational Safety and Health Administration 2001)
• Can connectors make initial attachments without hanging connections or avoidable unstable positioning? (Toole, Treppel, and Van Nosdall 2013)
• Can elevated work be reduced through prefabrication, ground assembly, or early permanent stairs? (Toole and Gambatese 2008; National Institute for Occupational Safety and Health 2024)
• Are fall-protection anchor points, lifeline capacities, column holes, beam-web holes, or parapets designed and documented where needed? (Toole, Treppel, and Van Nosdall 2013; National Institute for Occupational Safety and Health 2024)
• Are rooftop mechanical units located to reduce maintenance exposure near edges and skylights? (National Institute for Occupational Safety and Health 2024)
• Do details avoid top-of-beam tripping hazards and unnecessary field welding? (Toole, Treppel, and Van Nosdall 2013)

Stage 3: select design responses using a hierarchy-of-controls lens

After hazards are identified, design responses should be ranked by their position in the hierarchy of controls. Eliminating an exposure is preferred to reducing it; reducing it through permanent or designed features is preferred to relying on administrative controls or PPE; residual risks should be communicated when design cannot reasonably eliminate or reduce them. This ordering follows Gambatese, Behm, and Hinze’s hierarchy-of-controls framing and NIOSH’s definition of PtD as designing out hazards (Gambatese, Behm, and Hinze 2005; National Institute for Occupational Safety and Health 2024).

For steel buildings, the hierarchy can be translated into design alternatives. Eliminating elevated work may mean prefabricating an assembly or enabling ground assembly. Reducing fall exposure may mean a 42-inch parapet, permanent railing, designed anchor point, or relocated rooftop equipment. Reducing connection hazards may mean beam seats, connection geometry that supports stable initial bolting, shop welding instead of field welding, or avoiding hanging connections. Reducing temporary-stability risk may mean base-plate and anchor-rod details, column splice locations, and clearer connection criteria (Occupational Safety and Health Administration 2001; Toole, Treppel, and Van Nosdall 2013; National Institute for Occupational Safety and Health 2024).

Each alternative should also be checked for hazards introduced elsewhere. The parapet-versus-anchor case shows why: the roof-anchor system had safety value but introduced construction-phase exposure and productivity loss, while the compliant parapet had higher first cost but lower risk in the studied case (Rajendran and Gambatese 2013). Prefabrication likewise may reduce elevated connections but can affect transport and lifting. The framework therefore rejects one-directional safety claims. A design response is accepted only after the team records both the hazard it addresses and any residual or transferred hazards requiring construction planning.

Stage 4: encode PtD decisions in models, drawings, and contracts

The protocol’s computing and information-management component is intentionally modest but essential. Hossain et al.’s DfS knowledge library shows that safety rules can be structured and integrated with BIM to flag risks tied to model elements and maintain residual-risk registers (Hossain et al. 2018). The proposed steel-building implementation does not require every firm to build a custom software system. It does require that safety-relevant decisions be tied to identifiable elements and transmitted through project information channels.

At minimum, each PtD item should have five fields: element or location; hazard and lifecycle phase; design feature or rejected alternative; residual risk; and document or model location. This is an author’s proposed data structure, derived from BIM-linked risk-review work and NIOSH’s emphasis on contract-document communication (Hossain et al. 2018; National Institute for Occupational Safety and Health 2024). Examples include a perimeter column tagged as containing 21-inch and 42-inch holes for safety cables; a beam line tagged with lifeline attachment capacity shown on drawings; a roof edge tagged as protected by a 42-inch parapet; or a connection series tagged as requiring beam seats because of erection access and stability concerns.

The same information should appear in contract documents where it affects scope, fabrication, installation, or performance. NIOSH recommends that PtD controls be written into bids, contracts, plans, and drawings to make responsibilities and scope explicit (National Institute for Occupational Safety and Health 2024). Toole and Erger’s boundary-setting argument adds that communication should distinguish design features from means-and-methods instructions (Toole and Erger 2019). For example, a drawing may specify the capacity and location of a designed lifeline anchor point without dictating the contractor’s complete fall-protection plan. A connection criterion may require support during double connecting without prescribing the erector’s entire sequence.

Stage 5: maintain a residual-risk register through handoff

Residual-risk registers are the mechanism that prevents PtD from becoming either overreach or silence. Hossain et al.’s BIM-integrated case maintained a risk register for residual risks (Hossain et al. 2018). Toole and Erger emphasize residual-risk transfer as part of managing professional boundaries (Toole and Erger 2019). The author’s proposed register for steel buildings should include the following fields:

• Risk identifier: a unique label tied to a model element, drawing detail, specification section, or location.
• Steel element or assembly: column, base plate, beam, joist, deck, connection, stair, roof edge, anchor point, mechanical support, or other relevant element.
• Lifecycle phase: fabrication, transportation, lifting, temporary stability, connecting, decking, access, inspection, maintenance, or demolition planning if in scope.
• Foreseeable hazard: fall, collapse, struck-by load, unstable temporary condition, access constraint, tripping hazard, field-welding exposure, or unplanned attachment.
• Design action taken: eliminated, reduced, relocated, prefabricated, designed attachment provided, documented for contractor planning, or not adopted with reason.
• Residual risk owner for planning: contractor, construction manager, specialty engineer, owner maintenance team, or other party as defined by contract.
• Communication location: drawing sheet, model property, specification section, bid requirement, meeting record, or contract exhibit.

This register is not a substitute for the contractor’s safety program. It is a transfer instrument for design-relevant information. It records what the design team has done, what remains unresolved, and where the constructor or owner must plan further controls. That distinction is essential in the U.S. context described by Behm and Toole and Erger (Behm 2005; Toole and Erger 2019).

Stage 6: review lifecycle tradeoffs before design freeze

The final implementation stage is a lifecycle tradeoff review before design freeze. It should compare alternatives that materially affect construction or maintenance exposure, particularly roof-edge protection, access systems, connection strategies, prefabrication opportunities, and fall-protection anchorage. Rajendran and Gambatese’s case shows that a higher-first-cost parapet may be safer and less productivity-disruptive than a roof-anchor system in a specific project context (Rajendran and Gambatese 2013). Toole and Gambatese’s trajectories suggest that prefabrication, safer system selection, designer construction engineering, and spatial design can all shift construction risk (Toole and Gambatese 2008).

The review should explicitly record uncertainty. Hardison and Hallowell’s assessment of limited objective validation means that teams should avoid representing PtD choices as proven risk reductions unless project-specific evidence supports that claim (Hardison and Hallowell 2019). A more defensible record states the reasoning: the hazard identified, the design alternative selected, the cited PtD basis, the expected exposure reduction, any introduced risk, and the residual-risk owner. Such documentation makes the framework auditable without pretending that the design team has measured outcomes it has not measured.

The following author-proposed minimum specification can be inserted into a steel project’s design-management plan when contract scope allows: every major steel framing package shall undergo an element-phase PtD review; all accepted PtD features affecting fabrication, erection, access, fall protection, temporary stability, or maintenance shall be shown in the applicable model, drawing, or specification; all residual construction or maintenance risks recognized by the design team but not eliminated by design shall be entered into a residual-risk register; and no register entry shall direct contractor means and methods unless the contract expressly assigns that service. This specification operationalizes the article’s thesis as a bounded design-information obligation rather than an assumption of site-safety control.

Conclusion

This article has argued that safer steel buildings require a shift in what structural design is understood to deliver. The claim is not that structural engineers should control erection means and methods, nor that design can foresee every site hazard. It is that steel construction safety should be treated as a bounded design-information responsibility: engineers should eliminate or reduce foreseeable erection, access, fall, temporary-stability, and maintenance hazards where design can do so; encode safety-critical features in models, drawings, specifications, bids, and contracts; and transfer residual risks explicitly to the parties who plan and control the work. That position is deliberately narrower than assuming contractor site-safety responsibility, but more demanding than treating safety as wholly external to structural design.

The evidence reviewed supports this thesis through convergence rather than through a single definitive trial. Construction remains a major source of fatal work injuries, and steel erection contains hazards that OSHA itself links to design-relevant provisions, including anchor rods, double connections, perimeter safety-cable devices, restrictions on pre-erection shear connectors, column stability, decking, joists, and falls (U.S. Bureau of Labor Statistics 2026; Occupational Safety and Health Administration 2001). PtD scholarship supplies the broader hierarchy-of-controls rationale: hazards eliminated or reduced before exposure are more reliable than downstream dependence on training, supervision, administrative controls, or PPE (Gambatese, Behm, and Hinze 2005; National Institute for Occupational Safety and Health 2024). Steel-specific PtD materials then translate that rationale into concrete interventions, including beam seats, column and beam holes for fall protection, shop rather than field welding where feasible, early permanent stairs, prefabrication or ground assembly, designed lifeline anchorages, parapets, and rooftop-equipment placement away from edges and skylights (Toole, Treppel, and Van Nosdall 2013; National Institute for Occupational Safety and Health 2024).

The proposed framework responds to the article’s central gap: the literature offers PtD principles, steel safety examples, BIM-enabled rule concepts, and legal cautions, but not a bounded steel-building process that connects design decisions, lifecycle hazards, information carriers, and residual-risk transfer. The four-layer framework—hazard recognition, design intervention, information representation, and boundary-governed transfer—therefore organizes existing knowledge into a repeatable conceptual architecture. Its accompanying implementation protocol proposes scope definition, element-phase reviews, hierarchy-of-controls evaluation, model and document encoding, residual-risk registers, and lifecycle tradeoff review. These are proposed professional steps, not reported experimental results. Their purpose is to make safe constructability auditable and communicable without collapsing the distinction between design influence and contractor control.

The article has also emphasized limits. Hardison and Hallowell’s review cautions that objective validation of lifecycle risk reduction in CHPtD remains limited, and much of the field still rests on logical and qualitative evidence (Hardison and Hallowell 2019). Hallowell and Hansen’s findings show that many hazards are not visible from design documents alone and that designers’ hazard recognition is partial, variable, and improved by field experience and structured prompts (Hallowell and Hansen 2016). Case evidence, such as Rajendran and Gambatese’s comparison of parapet and roof-anchor alternatives, demonstrates the value of lifecycle reasoning but does not justify universal prescriptions (Rajendran and Gambatese 2013). Accordingly, the framework should be read as a disciplined method for reasoning, documenting, and communicating—not as proof that every listed intervention will reduce total project risk in every context.

The open problems follow directly from these limits. Future work should validate whether model-linked steel PtD reviews and residual-risk registers measurably improve hazard recognition, design coordination, and downstream safety planning. It should test which steel-specific prompts are most effective for designers with different levels of field experience; how BIM rules and contract documents can preserve safety information without directing means and methods; how owners price and procure PtD services; and how lifecycle tradeoffs among first cost, productivity, fabrication, transport, erection, access, and maintenance should be evaluated. Until that evidence matures, the article’s conclusion is pragmatic: structural engineers should neither overclaim control nor retreat into formal neutrality. They should make foreseeable steel erection and maintenance risk part of the design conversation, the design record, and the residual-risk handoff.

References