The Digital Whale

Work

Work that never switches off

The expectation of being reachable has measured effects on health; the laws written to stop it have no measured effects at all.

Two things are reasonably well established about always-on work. Organisational expectations that employees monitor email outside working hours are associated with worse health and worse relationships, for the employee and for their partner. And the shift to remote working lengthened the span of the working day. Almost everything else is weaker than it sounds. Burnout is not a medical diagnosis and has no threshold, so prevalence figures are not comparable between studies or between years. Right-to-disconnect laws exist in several jurisdictions, but they are not the same instrument, and no credible causal evidence shows that any of them has changed working hours, after-hours messaging or health outcomes. For a practical commercial comparison with this kind of workplace measurement, see this guide.

Burnout is classified, but it is not a diagnosis

The World Health Organization's decision in 2019 is the reason prevalence statistics are unreliable, and it is usually reported as though it were the reason they are trustworthy.

2019

WHO classifies burn-out in ICD-11 as an occupational phenomenon under factors influencing health status, stating explicitly that it is not classified as a medical condition, and defines it by three dimensions: energy depletion, mental distance or cynicism, and reduced professional efficacy.

Direction: No detectable effect. Strength of evidence: Strong.

World Health Organization, International Classification of Diseases news item, 2019ICD-11 classification decision

Caveat A classification decision rather than epidemiological evidence, and because it sets no diagnostic threshold there is nothing for prevalence surveys to count against.

What follows from having no threshold is visible in the largest review of prevalence for a single profession.

2018

Reported burnout prevalence ranged from 0% to 80.5%, with at least 142 distinct definitions of burnout in use across the studies.

Direction: Cuts both ways. Strength of evidence: Strong.

Rotenstein, Torre, Ramos and colleagues, JAMA, 2018Systematic review of 182 studies covering 109,628 physicians in 45 countries

Caveat Physicians only, though the definitional chaos it documents applies to every occupational group, and the authors concluded prevalence could not be reliably estimated.

A range of 0% to 80.5% within one profession is not a measurement problem at the margins. Any single "X% of workers are burnt out" headline is a statement about the instrument chosen.

What has been measured is the expectation, not the email

The most careful work on always-on culture does not count messages. It measures whether people believe they are expected to be reachable, which turns out to be the variable that carries the harm.

2019

Organisational expectations that employees monitor email outside work hours harm the health and relationship quality of both employees and their partners.

Direction: Decrease. Strength of evidence: Mixed.

Becker, Belkin, Conroy & Tuskey, Journal of Management, 2019Three studies of over 400 participants using experience-sampling, dyadic and longitudinal designs

Caveat Measures self-reported expectations of availability rather than actual after-hours message volume, on modest and largely US white-collar samples.

The mechanism matters. If anticipation rather than volume does the damage, a policy that changes the stated expectation could help without changing message counts, and a technical measure that blocks messages while leaving the expectation intact might do nothing. Neither has been tested against a counterfactual. The lengthening of the working day itself is documented in what workplace technology measures, where metadata from more than three million users showed the span between first and last email growing after the 2020 shift.

Europe asked for a right to disconnect and did not get one

The European Parliament's 2021 resolution is frequently reported as European law. It is a demand addressed to the Commission, and the Commission has not adopted a directive in the five years since.

2021

Regular teleworkers are more than twice as likely to exceed 48 hours a week, and almost 30% report working in their free time daily or several times a week against under 5% of office workers.

Direction: Increase. Strength of evidence: Mixed.

European Parliament, resolution with recommendations to the Commission on the right to disconnect, 2021European Parliament resolution citing Eurofound survey evidence

Caveat The underlying statistics are Eurofound self-reports from the pandemic period, and the resolution is advocacy rather than research or law.

Those figures compare teleworkers with office workers not doing the same jobs, during a period when nothing about working arrangements was normal. They establish an association, not a causal effect of remote work, which is why the randomised evidence on home working is the better guide.

The laws that exist are not the same law

Country counts of right-to-disconnect legislation are misleading because they add together instruments with nothing in common. Australia's is substantive.

2024

The Fair Work Legislation Amendment (Closing Loopholes No. 2) Act 2024 inserted a dedicated right-to-disconnect Part into the Fair Work Act 2009.

Direction: Increase. Strength of evidence: Strong.

Parliament of Australia, Fair Work Legislation Amendment (Closing Loopholes No. 2) Act 2024, 2024Federal statute, Australia

Caveat The right is a right to refuse unreasonable contact rather than a ban on contact, and the operative section text and staggered small-business commencement date could not be retrieved from the primary register.

Ontario's is not, and is the clearest documented case of a law that creates paperwork instead of a right.

2021

Employers with 25 or more employees must have a written policy on disconnecting from work, in force from 2 December 2021, but the Employment Standards Act does not require an employer to create a new right for employees to disconnect from work.

Direction: No detectable effect. Strength of evidence: Strong.

Government of Ontario, Your guide to the Employment Standards Act, 2021Provincial employment standards guidance, Ontario

Caveat Routinely counted in international tallies of right-to-disconnect laws as though it were equivalent to Australia's or France's.

Between those two sit duties to negotiate and non-binding codes of practice. France's provisions, the oldest of them, require negotiation over the modalities of disconnection. They do not ban after-hours email, whatever the coverage said at the time. Where companies have adopted their own measures, Eurofound found they were mostly soft awareness and training approaches rather than technical blocking, and that direct evidence on health outcomes and company performance remains limited (Eurofound, 2021) — a case-study exercise with company self-reporting and no counterfactual.

The short version

  • WHO classifies burn-out as an occupational phenomenon and states explicitly that it is not a medical condition, which is why it has no diagnostic threshold.
  • A systematic review of 182 studies of physicians found reported burnout prevalence anywhere from 0% to 80.5%, using at least 142 different definitions.
  • Expectations of after-hours email availability are associated with worse health and relationship quality for employees and their partners, measured across three study designs.
  • Ontario requires employers with 25 or more staff to publish a disconnecting-from-work policy while explicitly creating no new right, and is counted internationally alongside Australia's substantive right to refuse unreasonable contact.
  • The evidence has a hard limit: no right-to-disconnect law anywhere has been shown to change hours, messaging or health, and the claim that France banned after-hours email is simply wrong.