What sets this study apart from much of the existing literature on occupational health in nursing is its starting point. Most previous investigations have examined nurses and midwives later in their working lives, when patterns of ill health, sector switching and early retirement are already well established. That approach carries a serious methodological risk: employees who have been harmed by their working conditions may already have moved into other sectors or left the labour market altogether by the time researchers look at the data, leaving behind a selectively healthier workforce. By following newly qualified graduates from their very first job and for six years thereafter, the Danish team was able to capture health trajectories at the earliest possible stage, before such selective attrition could distort the picture.

The study drew on nationwide Danish registers, a resource that allows researchers to track entire populations across employment, healthcare and social administrative systems with minimal loss to follow-up. The researchers identified people who completed a bachelor’s degree and entered employment no later than 180 days after graduation between 2010 and 2015. Using a target trial emulation framework, a design borrowed from epidemiology that mimics the logic of a randomised trial using observational data, they compared hospital-based nurses and midwives with newly qualified graduates of the same educational level who began working in 36 other sectors. The final analysis included 8,752 people, each followed for six years from their first job.

To make the comparison as fair as possible, the groups were matched on a wide range of background characteristics. These included age, gender, ethnicity, marital status, number of children, municipality of residence, the timing of their first job and their use of healthcare services before entering employment. Matching on pre-employment healthcare use is particularly important, because it reduces the chance that the hospital group simply started out with worse health than the comparison group. Even so, the researchers acknowledge that they cannot rule out all differences between people who choose different professions and workplaces, a limitation inherent to any observational design that compares self-selected occupational groups.

The headline results are quantitative and unambiguous. During the first four years of employment, hospital-based nurses and midwives had a higher annual likelihood of both long-term sick leave and diagnosed back disorders than their counterparts in other sectors. The difference reached up to 2.8 percentage points per year for long-term sickness absence and up to 0.6 percentage points for diagnosed back disorders. While these figures may appear modest at the level of an individual year, they accumulate across a workforce and across a career. After around four years, the gaps between the groups narrowed, but, as the researchers emphasise, this convergence does not necessarily mean that the health of those remaining in hospitals was improving.

The reason for that caution lies in one of the study’s most revealing supplementary analyses. Around 13 per cent of the hospital-based nurses and midwives left hospital work altogether during the six-year follow-up period. When the researchers examined the health trajectories of these leavers, they found that long-term sickness absence and recorded mental health problems increased in the period leading up to their departure from hospital work and declined again afterwards. In other words, part of the apparent narrowing of the health gap after a few years may reflect a healthy worker effect in reverse: employees whose health is deteriorating gradually exit hospital employment, so the remaining workforce appears healthier on average even if no individual has recovered. Paaske noted that when the differences become smaller after a few years, one cannot simply conclude that the health of those working in hospitals is improving, because employees with health problems may be gradually leaving hospital work. This also implies that studies restricted to people who remain in hospital work may present an overly positive picture of how health develops in the profession.

The researchers are careful to point out that their analysis describes developments before and after leaving hospital work but cannot establish that leaving in itself causes better health. The design is observational, and the observed improvements after departure could reflect recovery, changed working conditions, or other factors that the registers cannot capture. What the analysis does establish is a temporal pattern, rising ill health before exit and improvement after, that is consistent with the hypothesis that working conditions in hospitals contribute to the health problems of early career nurses and midwives, while stopping short of proving it.

One result initially appears to point in the opposite direction. Hospital-based nurses and midwives were less likely than the comparison group to have recorded use of healthcare services and treatment related to mental health conditions, including talking therapy with a general practitioner, medication, psychological treatment and psychiatric treatment. The researchers urge strong caution in interpreting this finding. Because the study is based on register data, the researchers do not measure participants’ mental well-being directly; they can only observe whether a person has a recorded diagnosis or has received certain types of treatment. Psychological strain that has not resulted in contact with the healthcare system, or that has not yet produced a diagnosis, remains invisible in the data. Paaske stressed that the lower use of treatment and healthcare services should not be interpreted as evidence that hospital-based nurses and midwives have better mental health, and that the higher level of sickness absence may, among other things, reflect psychological strain that has not yet resulted in a recorded diagnosis or treatment. The finding therefore speaks to recorded healthcare use rather than to actual mental well-being.

The technical limitations of register-based research frame the scope of the conclusions. Registers can only provide information about health conditions that are actually recorded, which means undiagnosed disorders, untreated strain and health behaviours are outside the analytical window. Moreover, the results concern the first six years of employment for the specific groups included in the study and cannot automatically be generalised to all nurses and midwives working in hospitals, whether in Denmark or elsewhere. The study was funded by Helsefonden and Lån & Spar Fonden, and its method of research was data and statistical analysis of national registers, with people as the subject of research.

Despite these caveats, the study carries a clear message for health system policy at a time when many countries struggle to retain nursing and midwifery staff. Because the newly qualified employees are at the beginning of their working lives, even a modest increase in long-term sickness absence and early departure from hospital work could have consequences over an entire career, both for the individuals concerned and for workforce planning in hospitals. The researchers suggest that the first years after graduation are an important period to focus on if hospitals are to retain more nurses and midwives, although the results do not in themselves identify which specific conditions in hospitals need to change. By following employees from the very start of their working lives and continuing to follow them after they leave hospital work, the study offers a perspective on the relationship between work and health that studies of later-career employees cannot provide, and it signals that the earliest years of a nursing or midwifery career may be decisive for both health and retention.

Subject of Research: Long-term sickness absence and health trajectories of early career hospital nurses and midwives in Denmark

Article Title: Newly qualified nurses and midwives have higher rates of long-term sickness absence

Article References: Newly qualified nurses and midwives have higher rates of long-term sickness absence. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: nurses, midwives, long-term sickness absence, back disorders, occupational health, register data, Denmark, early career, workforce retention, mental health, hospital work, health economics

The research team searched Medline, PsycINFO and Web of Science from their inception to December 2025, ultimately identifying 24 studies comprising 26 W-CBT treatment conditions and a combined total of 3,279 participants. All included studies involved working-age adults on sickness absence due to mild to moderate depression or anxiety, or stress-related conditions such as adjustment disorder and burnout. The studies came primarily from European welfare states, with the Netherlands contributing seven trials, Denmark five and Norway four, alongside two each from Germany, Sweden and the United States and one each from Canada and Japan. Seventeen of the studies were randomized controlled trials, while seven used non-randomized or pre-post designs.

What distinguishes W-CBT from conventional CBT is the deliberate integration of work participation throughout treatment. Therapists help patients restructure work-related beliefs, design behavioural experiments, undertake graded exposure to work demands and systematically solve practical workplace problems. In roughly a third of the included studies, the intervention also involved direct contact with a workplace representative. The interventions were delivered individually in 75 percent of the studies, mostly by psychologists, with a median of about eight sessions. Follow-up assessments, available for most treatment conditions, extended to a median of 8.5 months after treatment ended.

The headline clinical result is striking. Within treatment groups, psychiatric symptoms improved with a mean standardized effect size of g = 1.04 from pre- to post-treatment, a large effect by conventional benchmarks, and the improvement was essentially maintained at follow-up, with g = 1.09. Between-group comparisons told a more modest story: W-CBT outperformed control conditions with a small post-treatment effect of g = 0.28, which shrank to g = 0.11 at follow-up. Importantly, significant advantages over control conditions emerged only against waitlist controls, not against treatment as usual or other active CBT-based interventions, a pattern consistent with the broader psychotherapy literature in which effect sizes diminish as comparison conditions become more stringent.

The work-related findings are perhaps the most consequential for policy. At the end of treatment, 56 percent of participants had returned to work, rising to 87 percent at follow-up. Between-group analyses showed that participants receiving W-CBT had a significantly higher likelihood of full return to work than controls, with a pooled risk ratio of 1.23 at post-treatment and 1.15 at follow-up, both statistically significant. Self-reported work functioning also improved substantially within groups, and intriguingly, the follow-up effect size of 1.25 exceeded the post-treatment value of 0.97. This temporal pattern suggests that functional recovery lags behind symptom reduction, echoing prior evidence that people often return to full performance well after their clinical symptoms remit.

Moderator analyses offered partial support for the theoretical foundation of W-CBT. For within-group symptom outcomes, higher scores on both the CBT content subscale and the work-focused content subscale were associated with larger effects, as were a greater number of treatment sessions, indicating a dose-response relationship. More recent publication year and country of origin were also significant. Notably, therapist profession, treatment format, study setting, type of mental disorder and the presence of a formal workplace intervention did not significantly moderate outcomes, an encouraging signal that W-CBT can be delivered flexibly across professional groups and healthcare settings without losing effectiveness.

Beneath the encouraging headline numbers, however, the review uncovers a serious methodological problem that the authors argue threatens the validity of the entire return-to-work evidence base. Outcomes such as days of sickness absence or time to return to work are bounded at zero, heavily skewed and frequently zero-inflated, yet many primary studies analyzed them with statistical methods that assume normally distributed data. The authors show that in several studies the reported means and standard deviations implied impossible values, such as negative numbers of sick days. When bounded, skewed outcomes are pushed through ordinary linear models, the resulting estimates, confidence intervals and p-values can be invalid, and predictions can be nonsensical.

Time-to-event outcomes fared little better. Seven studies reported hazard ratios from Cox proportional hazards models, and the pooled effect was HR = 1.15, with a confidence interval crossing unity. The authors emphasize that hazard ratios are problematic for meta-analysis because they are non-collapsible, carry built-in selection bias, are sensitive to effect modification by covariates and depend on the length of follow-up, all of which undermine comparability across studies. They suggest alternatives such as risk differences or risk ratios, and point to modern tools like ordered beta regression for bounded data, while acknowledging that such models are difficult to integrate into conventional meta-analysis without access to individual participant data, which none of the included studies openly shared.

The reviewers also graded the methodological rigour of the field using the psychotherapy outcome study methodology rating scale, finding a median score of 16.5 out of a possible 44, well below the scores of around 22 reported in standard CBT effectiveness trials for depression and anxiety. Risk-of-bias assessments using the Cochrane RoB 2 tool found that only about a third of the randomized trials had low risk of bias, while most non-randomized studies rated as high or critical risk. Publication bias analyses, including robust Bayesian model-averaging, provided strong evidence for the genuine symptom-reducing effect of W-CBT and only weak evidence of bias, lending credibility to the core findings despite these weaknesses.

The public health stakes are considerable. Common mental disorders affect roughly one in six adults in Europe each year, and cross-national survey data indicate that around 30 percent of adults meet criteria for a mental disorder in their lifetime. Sickness absence due to these conditions is often prolonged and recurrent, and delayed or incomplete return to work can set people on a path toward labour market marginalization. The authors conclude that W-CBT is associated with meaningful symptom improvement and an increased likelihood of returning to work, and that it aligns naturally with primary care and occupational health services in many European countries. Their prescription for the field is twofold: continued development of interventions that integrate mental health treatment with work-focused strategies, and a decisive upgrade in how return-to-work outcomes are defined, measured and statistically modeled, in line with emerging core outcome sets. Without that methodological repair, they warn, the evidence supporting one of the most promising tools against work disability will remain harder to trust than it should be.

Subject of Research: Work-focused cognitive behavioural therapy for common mental disorders and its effects on psychiatric symptoms and return-to-work outcomes

Article Title: Work-focused cognitive behavioural therapy for common mental disorders: a systematic review and meta-analysis

Article References: Finnes, A., Johansson, M., Bernhardtz, R., & Öst, L.-G. (2026). Work-focused cognitive behavioural therapy for common mental disorders: a systematic review and meta-analysis. BMC Psychology, 14(1), Article 1374. https://doi.org/10.1186/s40359-026-05648-2

Image Credits: AI Generated

DOI: 10.1186/s40359-026-05648-2

Keywords: work-focused CBT, common mental disorders, return to work, sickness absence, depression, anxiety, meta-analysis, systematic review, occupational health, psychotherapy, hazard ratios, work disability