If this is you Your output at work has held steady for months, and you're running out of sleep, patience and reserve to keep it there.

Steady Output Can Hide a Person Running Out

Institutions track hours worked and patients seen, not the sleep, health and attention a person spends keeping those numbers steady. That part doesn't come back on schedule.

A case file: the rule, the person it landed on, and who could correct it.

The argument in 8 lines
If this is you
Your output at work has held steady for months, and you're running out of sleep, patience and reserve to keep it there.
The case
In Ann Rogers's 2004 study, 393 hospital nurses logged 5,317 shifts; about 40 percent ran past twelve hours, and error risk rose significantly past twelve hours and with overtime.
The claim
Institutions measure output, a flow, and miss the functional reserve that producing it drains, so steady output can hide a person running out.
The mechanism
Compensation holds the output rate steady while reserve drains, and the institution files that compensation as performance until the person decompensates into visible failure.
Who pays
People already near exhaustion, for whom ten more minutes can mean a missed dose, broken sleep, an error or leaving a job.
Who can correct it
Each institution sees only its own demand, so the sum the person receives stays invisible; measurement changes nothing until workers have authority to change the conditions.
What would change it
Instrument the workload: reserve rather than rate, recovery rather than only demand, compensation rather than only failure, plus enforceable recovery obligations.
The objection it answers
Average burden can look modest. Because harm climbs as reserve falls, the average is dominated by people with slack and hides a minority being ground down.
Reading settings

The question up top, the outline, the instrument and the what-next. Turn it off and you just get the essay.

An anchor. Its shadow falls across the person. A loop brings them back to where they began.

When 393 hospital staff nurses logged their shifts for a study published in 2004, about 40 percent of the 5,317 shifts ran past twelve hours. Past twelve hours, the risk of an error rose significantly.

Modern institutions are generally much better at measuring what people produce than at measuring what producing it does to their future ability to keep producing, caring, choosing, and living.

Most institutional reasoning models a person as a stable device that turns demand into output through effort. Demand arrives, effort is spent, output appears, and the cycle resets on demand. The model is convenient because it has one variable, and because a person whose output holds steady reads as a person who’s coping.

A human being carries the effects of one stretch of work into the next. A rough sketch puts that carryover on the page.

Rt+1=Rt+recovery−load−compensationR_{t+1} = R_t + \text{recovery} - \text{load} - \text{compensation}

This is an accounting sketch, not a validated physiological equation. It gives no units or method for measuring a person’s reserve, and the costs of load and extra effort can overlap. It pictures capacity carried forward, replenished by recovery and spent on work, including the extra effort that keeps output steady. The same demand can cost different amounts depending on the state it lands in. A count of completed work doesn’t record that cost.

The person who does this work usually gets drawn as a crumple zone, a figure taking a blow, and their repairs arrive on the dashboard as the system’s own performance. A figure taking a blow has no state variable, and the state is where the cost builds up.

Metrics Count Flows, Not The Reserve Beneath Them

Institutional metrics are almost entirely flows. Hours worked, patients seen, cases processed, forms completed, output per worker, response times, tickets closed. Every one of them is a rate, and rates are exactly the quantities that stay stable while a stock drains.

The variable that carries the biological weight is the stock being used up, which is functional reserve. Someone can hold output constant across months while the reserve falls the entire time, because the compensation term covers the difference.1

Maintained output gets mistaken for maintained capacity at exactly the point where the two come apart, which is whenever somebody is absorbing, because absorbing is what holds the rate constant. An institution reading its own throughput is reading the compensation term and filing it as performance.

Recovery is the other half of the stock and the half almost nobody accounts for. Capacity tomorrow depends on today’s load and tonight’s replenishment, as Theo Meijman and Gijsbertus Mulder argued in their effort-recovery model of workload.2 An organization that spends somebody’s evening, their sleep, a meal break, a weekend, or a stretch of attention has reached into the regeneration process itself. That makes the true cost larger than the hours appear to consume. Treating recovery as private leaves the production system depending on whether people slept, and its own accounting prices their sleep at zero. The same error shows up wherever the regeneration of a fishery or a wetland is treated as free because it happens off the balance sheet of the people drawing it down.

The Marginal Cost Of Burden Rises As Reserve Falls

Burden gets treated as additive almost everywhere in policy and operations. Ten extra minutes plus ten extra minutes equals twenty minutes of demand, and the arithmetic is the same whoever receives it.

Biology declines the arithmetic. Ten minutes imposed on a person with large reserve can be easy to absorb. Ten minutes imposed on a person already near exhaustion can land as a missed dose, a night of broken sleep, a childcare failure, an error, a withdrawal from a program or a job. The claim is that an extra demand can do more harm when someone is already depleted. The equation supplies no measured curve for that relationship.

Hospital nursing has measured a piece of this. The nurses in Ann Rogers’s 2004 study usually worked longer than they were scheduled to, and their risk of an error rose past twelve hours, with overtime, and past forty hours in a week.3 That unscheduled time is one possible form of the extra effort pictured above. The study links long hours to reported errors, but it doesn’t measure reserve or isolate the cost of that effort. A schedule that books the thirteenth hour of a shift as one more hour of coverage can’t see either.

That single property breaks the average-burden calculation more thoroughly than it breaks any individual one. A mean across people with very different reserves can conceal the cost to those who have little. A system can show a modest average cost, satisfy every threshold derived from that average, and be grinding down a minority badly enough to end their participation. Designing to that mean leaves out people who can’t absorb even the average demand.

Demands also arrive from several directions at once, and each institution sees only its own. A workplace sees the worker, a hospital the patient, a school the parent, a benefits office the claimant. Each of them is asking whether its own demand is reasonable, and by the standards of its own domain it usually is. The organism receives the sum. Ten individually defensible demands can add up to an environment that no single one of them would produce and that none of them can see. So the aggregate stays invisible for a structural reason: the accounting is partitioned and the person is not.

Visible Failure Is a Late Signal

Present systems are genuinely good at measuring visible failure. Absence shows up. Resignation shows up. Errors get logged, hospitalizations get recorded, missed deadlines and dropouts and complaints all leave marks. The measurement is competent and it’s consistently late.

A distinction from physiology does most of the work here. A system can be in compensated dysfunction for a long stretch, holding its outputs together by adjusting continuously. It reaches decompensated dysfunction only when adjustment stops being available. The metrics that record resignation and hospitalization and the complaint are recording the second state. Many of them are effectively asking when the person finally stopped successfully hiding the problem, which isn’t the same question as when the problem started. The interval between those two answers is the entire subject of compensation.

The interval has a name borrowed from the study of adaptive systems, and it’s hysteresis, here a model rather than a settled physiological law. After enough load, a system doesn’t return to where it began once the load is removed. Temporary burden produces consequences that aren’t temporary, because the person isn’t a bucket that tips and rights itself. The same relation shows up in the economy of a worker, where adaptive capacity gets spent so consistently that it runs down instead of recovering. A productivity model that treats adaptive capacity as a renewable input cannot represent this.

Survival is the bar most of this measurement clears, and the bar is very low. A person can keep functioning while their health, their sleep and their relationships go to shit. So treating continued function as the benchmark for acceptable conditions is the assumption doing most of the damage. Adaptation is evidence that adaptation happened, which is a different claim than adequacy.

Change the Model, Change What Counts as Evidence

Compressed, present institutional thinking assumes:

capacity is fixed, burden is additive, adaptation is free, recovery is private, and failure is visible\boxed{\text{capacity is fixed, burden is additive, adaptation is free, recovery is private, and failure is visible}}

A model that survives contact with a person in time assumes:

capacity is state-dependent, burden is nonlinear, adaptation has costs, recovery is constitutive, and compensation can hide failure\boxed{\text{capacity is state-dependent, burden is nonlinear, adaptation has costs, recovery is constitutive, and compensation can hide failure}}

Small modification is the wrong description. It changes what counts as evidence that a system is functioning, because the evidence that counts now is the evidence that stays stable while a person pays.

The practical consequence is an obligation to instrument the conditions the accounting leaves out: reserve rather than rate, recovery rather than only demand, compensation rather than only failure. What can be metered is the workload (hours, breaks, staffing, error rates, the interval between demand and recovery), not the biological capacity behind it, which no dashboard reads directly. Metering the workload is the difference between an institution that can discover it’s been spending people and one that can only notice when they stop showing up. But evidence doesn’t secure the remedy by itself. A measurement changes nothing until someone can act on it: workers need the authority to change the conditions the evidence names, and recovery obligations that are enforceable without first exhausting the person who invokes them. The error is hard to catch from inside for a plain reason: the compensation term exists to keep the rate steady, so it succeeds in exactly the case that matters most.

Notes

1

G. Robert J. Hockey, “Compensatory Control in the Regulation of Human Performance under Stress and High Workload: A Cognitive-Energetical Framework,” Biological Psychology 45, no. 1–3 (1997): 73–93, proposed that people can hold their performance steady under stress and heavy workload by recruiting extra effort, paying for it in strain and fatigue that the main task’s output doesn’t show.

2

Theo F. Meijman and Gijsbertus Mulder, “Psychological Aspects of Workload,” in Handbook of Work and Organizational Psychology, vol. 2, Work Psychology, 2nd ed., ed. Pieter J. D. Drenth, Henk Thierry and Charles J. de Wolff (Psychology Press, 1998), 5–33, set out the effort-recovery model: the strain of effort at work reverses when recovery is complete, and when it isn’t, the person starts the next stretch of work in a depleted state, has to spend extra effort to perform, and the strain accumulates toward chronic health problems. The reserve equation is this essay’s shorthand, not theirs.

3

Ann E. Rogers, Wei-Ting Hwang, Linda D. Scott, Linda H. Aiken and David F. Dinges, “The Working Hours of Hospital Staff Nurses and Patient Safety,” Health Affairs 23, no. 4 (2004): 202–212. The errors were self-reported in the nurses’ logbooks.

When they get you wrong or count on you

12
Who pays
People already near exhaustion, for whom ten more minutes can mean a missed dose, broken sleep, an error or leaving a job.
Who can change it
Each institution sees only its own demand, so the sum the person receives stays invisible; measurement changes nothing until workers have authority to change the conditions.

Continue reading

What to read next

This essay belongs to the Burden pattern; its toolkit, topic and guide are one step away.

Related essays

New essays by email

One situation a week, the essays that explain it, and one question to put to the institution.

New essays by email

One situation a week, the essays that explain it, and one question to put to the institution.