nobody counts

the whole day

Problem statement

Fatigue is regulated in transportation more tightly than almost any other occupational hazard: hours-of-service rules cap driving and duty time, and electronic logging devices enforce them. But those rules count duty hours recorded by one employer, and the NORA Transportation, Warehousing and Utilities Council points out that this is not the quantity that determines whether someone is dangerously tired. "Rail, maritime, air, and passenger transit sub-sectors all have workers whose time on-call is not well tracked and who may also have long commutes or second jobs," the agenda states, and concludes: "We need a complete picture of time at work in order to determine impacts on sleep, fatigue, and injury." No instrument produces that picture. A driver can be fully compliant on their log while having spent the preceding fourteen hours on unpaid, untracked on-call status, commuting, and working a second job for a different employer who has no idea the first one exists.

Why this matters

Transportation and warehousing is the sector where this gap is widest and the consequences are shared with the public. The agenda reports that 31.5% of transportation and warehousing workers are on "alternate shifts" — shifts falling at least partly outside 6 a.m. to 6 p.m. — against 17.7% for U.S. workers overall, and that the sector has "a much higher prevalence of on-call workers (11.3% in 2015) than any other sector." The health consequences are already visible in the same population: obstructive sleep apnea at 8% versus 4% in the general population, and the council notes that poor sleep and other health conditions are associated with increased risk of occupational safety incidents. Because the regulated quantity is the recorded one, every fatigue countermeasure — rest rules, scheduling standards, screening programs, the accident investigator's question of whether fatigue was a factor — operates on a number that systematically understates exposure by an unknown and worker-specific amount. And an unmeasurable quantity cannot support a dose-response study, so the research needed to set better rules is blocked by the same gap the rules are.

What’s been tried and why it hasn’t worked

Regulation went first and defined the measurement: hours-of-service caps, and later electronic logging devices that made duty hours tamper-resistant. That solved the falsification problem and left the boundary problem untouched, because the device logs the vehicle, not the human — nothing outside the cab is in scope, and the employer has neither the authority nor the appetite to record a second employer's schedule. Self-report was the fallback and is unreliable in exactly the direction that matters: workers whose earnings depend on being available under-report the hours that would disqualify them, and the agenda documents the surrounding pressures — misclassification as independent contractors that strips protections, and lease-purchase arrangements in trucking that bind drivers "to their employer with debt they cannot sustain." Even the vocabulary is not settled. The council identifies a prerequisite most fields would consider embarrassing: there is "a need to establish standard meanings for terms such as temporary, on-call, shift work, etc., so that study results can be understood and compared." So studies that do attempt to measure the sector's real work time cannot be pooled. The agenda's summary of the evidence base is that "[t]here is so far, little health and safety research on how non-standard work arrangements are affecting the safety and health of workers, including transportation and warehousing workers" — and it notes that scheduling algorithms and load-brokering apps "might be making the problem worse."

What would unlock progress

The measurement has to move from the employer's record to the worker's own device, with the worker controlling disclosure — a worker-held time-at-work ledger that captures on-call windows, commute, second jobs and sleep opportunity, and reports an aggregate exposure metric rather than raw whereabouts. The privacy structure is the hard and interesting part, and it has a precedent: personal exposure dosimetry in industrial hygiene, where a worker wears the sampler and the aggregate goes into the exposure record without the employer seeing minute-by-minute detail. A second, cheaper unlock is definitional and would help immediately: a published, tested taxonomy of work-arrangement terms with operational definitions, which is what the council asks for and what would let existing datasets be compared. The technical ingredients for the first — passive sleep and activity estimation from consumer wearables, geofenced commute detection — already exist; nobody has assembled them into an instrument designed for a worker's interest rather than an employer's.

Entry points for student teams

A team could build and pilot the worker-held ledger with a small cohort (student drivers, delivery workers, hospital staff on call), measuring the gap between logged duty hours and total time-at-work plus sleep opportunity, and reporting the distribution of that gap — a number nobody currently has. A privacy-focused team could design the disclosure architecture: what aggregate the worker can share with a regulator or clinician that supports a fatigue judgment without exposing location history, and test whether workers find it acceptable. A third team could do the definitional work as a genuine research contribution — draft operational definitions for on-call, standby, split-shift and second-job time, then apply them to two existing public datasets and show how much the estimated prevalence of long work weeks changes. Relevant skills: mobile sensing and app development, sleep and fatigue science, survey methodology, privacy engineering, transportation policy.