On the evidence so far, working from home has no clear effect on health, good or bad

Working from home is one of the biggest shifts in how people live in a generation, which makes the evidence on its health effects unusually anticlimactic: so far, there isn't a clear one, in either direction. A Swiss longitudinal study that tracked employees switching to remote work found no measurable effect across 13 different health outcomes [s1]. And a systematic review of homeworking during the pandemic concluded there was no clear consensus on its relationship to mental health, because the effect varied so much from person to person and study to study [s2]. The confident claims on both sides — that remote work is quietly ruining or quietly saving people's health — run ahead of what the data show.

The strongest single study points to no effect

The most methodologically clean recent evidence comes from Switzerland. Researchers drew on the COVID-19 Social Monitor, a large, high-frequency longitudinal panel of the Swiss population, and analysed 3,381 tertiary-educated employees [s1]. Crucially, they used individual fixed-effects models — a design that compares each person against themselves before and after they started working from home, which strips out the confounding from the fact that remote workers differ from office workers in stable ways to begin with [s1].

Across 13 binary outcomes spanning general health, mental health, physical health, health behaviour, and social trust, the estimates suggested no evidence of an effect of transitioning to working from home on any of them [s1]. The authors frame the absence of adverse effects as meaningful for employers and policymakers weighing flexible arrangements: on this evidence, the switch did not harm the health of the workers who made it [s1].

The important limit is who was studied. The sample was tertiary-educated employees in a high-income country — precisely the group with the autonomy, space, and job type that make remote work manageable [s1]. It says little about a call-centre worker in a cramped flat, and the study itself positions its comprehensive-outcome, longitudinal design as a benchmark for future research rather than the last word [s1].

Why the wider literature is mixed, not null

The Swiss null result sits inside a broader literature that is genuinely messy, and the messiness is itself the finding. A systematic review of homeworking during COVID-19 screened 6,906 citations and included 27 studies, seeking the relationship between working from home and both mental health and productivity [s2]. Its conclusion was that the association varies considerably, pointing to a complex relationship shaped by many factors — demographics and occupation among them — with no clear consensus emerging [s2].

The review did surface one conditional signal worth separating out: there were indications that people homeworking for the first time during a pandemic were at risk of poor productivity, as were those experiencing poor mental health [s2]. That is a specific, stressful scenario — enforced, sudden, isolating homeworking under lockdown — and it is not the same as someone choosing hybrid work in ordinary conditions. Much of the alarming early evidence about remote work came from exactly that pandemic context, which is a poor guide to voluntary remote work now.

What the honest summary is

Put together, the two sources support a deflationary but useful conclusion. The best-designed recent study finds no health effect of switching to remote work in a population well suited to it, and the broader review finds too much heterogeneity to declare a general effect at all [s1] [s2]. The truthful answer to "is working from home good or bad for your health" is that it appears to depend heavily on who is doing it and under what conditions, and that the average effect measured so far is close to nothing — not the uniform harm or benefit the discourse often assumes [s1] [s2]. Where an individual falls within that variation, and what would make their own arrangement healthier, is not something a population study can decide for them.