Risk of epilepsy after traumatic brain injury: a nationwide Norwegian matched cohort study.

Background: Post-traumatic epilepsy (PTE) is a well-known complication of traumatic brain injury (TBI). Although several risk factors have been identified, prediction of PTE is difficult. Changing demographics and advances in TBI treatment may affect the risk of PTE. Our aim was to provide an up-to-date estimate of the incidence of PTE by linking multiple nationwide registers. Methods: Patients with TBI admitted to hospital 2015-2018 were identified in the Norwegian Trauma Registry and matched to trauma-free controls on sex and birth year according to a matched cohort design. They were followed up for epilepsy in nationwide registers 2015-2020. Cumulative incidence of epilepsy in TBI patients and controls was estimated taking competing risks into account. Analyses stratified by the Abbreviated Injury Scale (AIS) severity score, Glasgow Coma Scale score and age were conducted for the TBI group. Occurrence of PTE in different injury types was visualized using UpSet plots. Results: In total, 8,660 patients and 84,024 controls were included in the study. Of the patients, 3,029 (35%) had moderate to severe TBI. The cumulative incidence of epilepsy in the TBI group was 3.1% (95% Confidence Interval [CI] 2.8-3.5%) after 2 years and 4.0% (3.6-4.5%) after 5 years. Corresponding cumulative incidences in the control group were 0.2% (95% CI 0.2-0.3%) and 0.5% (0.5-0.6%). The highest incidence was observed in patients with severe TBI according to AIS (11.8% [95% CI 9.7-14.4%] after 2 years and 13.2% [10.8-16.0%] after 5 years) and in patients >40 years of age. Conclusion: Patients with TBI have significantly higher risk of developing epilepsy compared to population controls. However, PTE incidence following moderate-severe TBI was notably lower than what has been reported in several previously published studies.

Prenatal exposure to benzodiazepine and z-hypnotics and fifth-grade scholastic skills - emulating target trials using data from the Norwegian Mother, Father and Child Cohort Study.

Evidence is limited regarding the effect of prenatal benzodiazepine and z-hypnotic exposure and long-term neurodevelopment in childhood. The objective of this study was to investigate the effects of initiating benzodiazepine or z-hypnotic treatment in early, mid and late pregnancy on fifth-grade numeracy and literacy scholastic skills in children, by emulating three target trials. The trials are identical except for the timing of enrollment and the number of eligible individuals. Eligibility to the trials required a history of anxiety and/or depression prior to pregnancy. We used data from the Norwegian Mother, Father and Child Cohort Study, linked to the Medical Birth Registry of Norway, to emulate the trials. We adjusted for baseline covariates that were available at time 0 for each trial by inverse probability of treatment weighting using the propensity score. The findings of this study did not show any effect of mothers' initiation of treatment with benzodiazepines or z-hypnotics in early, mid or late pregnancy on the children's 5th grade test scores in numeracy and literacy. The study results provide reassurance for patients in need of benzodiazepines and z-hypnotics during pregnancy; however, these findings need to be interpreted with caution due to low study power in some of the analyses.

Gender and educational differences in work participation and working years lost in Norway.

OBJECTIVES: This study aimed to quantify the duration of work participation and reasons for working years lost, according to gender and educational attainment, among a Norwegian population. METHODS: Register data on labor market attachment between 2000-2015 were obtained from Statistics Norway. We included five cohorts: individuals turning 20 (N=323 333), 30 (N=386 006), 40 (N=388 962), 50 (N=358 745), and 60 years (N=284 425) between 1 January 2000 and 31 December 2005. Individuals were followed for ten years. Data completeness allowed calculation of the average time spent in work and years lost to health-related absences and non-employment states per cohort. Changes in state probabilities over time were also depicted. Mean differences between genders and educational levels, and corresponding 95% confidence intervals were based on 1000 bootstrap samples. RESULTS: Both genders spent most time in work; however, per cohort, women worked approximately one year less than men. As cohorts aged, main reasons for working years lost changed from education and economic inactivity to sickness absence and disability pensioning; this trend was stronger for women than men. Individuals with a low education spent fewer years in work and more years in sickness absence and disability pensioning than highly educated peers. This difference tended to be larger for women and older cohorts. CONCLUSIONS: Per cohort, women participated one year less in work than men and, depending on age, spent more time in education, economic inactivity, sickness absence, and disability pensioning. Stronger educational gradients were seen for work and health-related absences for older cohorts and women.

The effects of a national, voluntary agreement for a more inclusive working life on work participation following long-term sickness absence: a Norwegian cohort study.

OBJECTIVES: This study aimed to estimate the average individual effect of the company-level Norwegian Agreement on a More Inclusive Working Life (IA Agreement) on individuals' (i) sustained return to work after a sickness absence (SA) episode, and (ii) recurrent SA. METHODS: Using register data, 79 253 men and 94 914 women born in Norway 1967-1976 were followed for one year between 2005 and 2010 after returning to work from an SA episode (>16 days). Weighted Cox proportional hazard models analysed time to first exit from work by companies' IA status (IA/non-IA). Weighted cumulative incidence differences between IA and non-IA groups with 95% bootstrapped confidence intervals (CI) were calculated for the competing events of full SA, graded (<100%) SA, unemployment/economic inactivity, education, disability pension, and death/emigration. Stabilised inverse probability of treatment weights balanced IA/non-IA groups according to nine covariates. Analyses were stratified by gender, and separately for two initial SA diagnoses (musculoskeletal and psychological). RESULTS: Both men [adjusted hazard ratio (HR) 0.96, 95% CI 0.93-0.99] and women (adjusted HR 0.97, 95% CI 0.94-0.99) in IA companies were less likely to exit work in the year following SA. Similar findings were seen among individuals with musculoskeletal diagnoses and women with psychological diagnoses. Men with psychological diagnoses were more likely to exit work. Recurrent full and graded SA were more likely, and unemployment/economic inactivity less likely, in IA companies. However, the estimated effects were small and the CI often included the null. CONCLUSIONS: Individuals working in IA companies were more likely to remain in work. This was mainly due to reduced unemployment/economic inactivity, suggesting the IA Agreement may have influenced work participation through other means than reduced SA.

Association of physical activity with overall mortality among long-term testicular cancer survivors: A longitudinal study.

Physical activity (PA) has been associated with reduced mortality among cancer survivors, but no study has focused on testicular cancer survivors (TCSs). We aimed to investigate the association of PA measured twice during survivorship with overall mortality in TCSs. TCSs treated during 1980 to 1994 participated in a nationwide longitudinal survey between 1998 to 2002 (S1: n = 1392) and 2007 to 2009 (S2: n = 1011). PA was self-reported by asking for the average hours per week of leisure-time PA in the past year. Responses were converted into metabolic equivalent task hours/week (MET-h/wk) and participants were categorized into: Inactives (0 MET-h/wk), Low-Actives (2-6 MET-h/wk), Actives (10-18 MET-h/wk) and High-Actives (20-48 MET-h/wk). Mortality from S1 and S2, respectively, was analyzed using the Kaplan-Meier estimator and Cox proportional hazards models until the End of Study (December 31, 2020). Mean age at S1 was 45 years (SD 10.2). Nineteen percent (n = 268) of TCSs died between S1 and EoS, with 138 dying after S2. Compared to Inactives at S1, the mortality risk among Actives was 51% lower (HR 0.49, 95% CI: 0.29-0.84) with no further mortality reduction among High-Actives. At S2, the mortality risk was at least 60% lower among the Actives, High-Actives and even the Low-Actives compared to the Inactives. Persistent Actives (≥10 MET-h/wk at S1 and S2) had a 51% lower mortality risk compared to Persistent Inactives (<10 MET-h/wk at S1 and S2; HR 0.49, 95% CI: 0.30-0.82). During long-term survivorship after TC treatment, regular and maintained PA were associated with an overall mortality risk reduction of at least 50%.

Third dose mRNA vaccination against SARS-CoV-2 reduces medical complaints seen in primary care: a matched cohort study.

BACKGROUND: SARS-CoV-2 mRNA vaccination has been associated with both side effects and a reduction in COVID-related complaints due to the decrease in COVID-19 incidence. We aimed to investigate if individuals who received three doses of SARS-CoV-2 mRNA vaccines had a lower incidence of (a) medical complaints and (b) COVID-19-related medical complaints, both as seen in primary care, when compared to individuals who received two doses. METHODS: We conducted a daily longitudinal exact one-to-one matching study based on a set of covariates. We obtained a matched sample of 315,650 individuals aged 18-70 years who received the 3rd dose at 20-30 weeks after the 2nd dose and an equally large control group who did not. Outcome variables were diagnostic codes as reported by general practitioners or emergency wards, both alone and in combination with diagnostic codes of confirmed COVID-19. For each outcome, we estimated cumulative incidence functions with hospitalization and death as competing events. RESULTS: We found that the number of medical complaints was lower in individuals aged 18-44 years who received three doses compared to those who received two doses. The differences in estimates per 100,000 vaccinated were as follows: fatigue 458 less (95% confidence interval: 355-539), musculoskeletal pain 171 less (48-292), cough 118 less (65-173), heart palpitations 57 less (22-98), shortness of breath 118 less (81-149), and brain fog 31 less (8-55). We also found a lower number of COVID-19-related medical complaints: per 100,000 individuals aged 18-44 years vaccinated with three doses, there were 102 (76-125) fewer individuals with fatigue, 32 (18-45) fewer with musculoskeletal pain, 30 (14-45) fewer with cough, and 36 (22-48) fewer with shortness of breath. There were no or fewer differences in heart palpitations (8 (1-16)) or brain fog (0 (- 1-8)). We observed similar results, though more uncertain, for individuals aged 45-70 years, both for medical complaints and for medical complaints that were COVID-19 related. CONCLUSIONS: Our findings suggest that a 3rd dose of SARS-CoV-2 mRNA vaccine administered 20-30 weeks after the 2nd dose may reduce the incidence of medical complaints. It may also reduce the COVID-19-related burden on primary healthcare services.

Causal inference in survival analysis using longitudinal observational data: Sequential trials and marginal structural models.

Longitudinal observational data on patients can be used to investigate causal effects of time-varying treatments on time-to-event outcomes. Several methods have been developed for estimating such effects by controlling for the time-dependent confounding that typically occurs. The most commonly used is marginal structural models (MSM) estimated using inverse probability of treatment weights (IPTW) (MSM-IPTW). An alternative, the sequential trials approach, is increasingly popular, and involves creating a sequence of "trials" from new time origins and comparing treatment initiators and non-initiators. Individuals are censored when they deviate from their treatment assignment at the start of each "trial" (initiator or noninitiator), which is accounted for using inverse probability of censoring weights. The analysis uses data combined across trials. We show that the sequential trials approach can estimate the parameters of a particular MSM. The causal estimand that we focus on is the marginal risk difference between the sustained treatment strategies of "always treat" vs "never treat." We compare how the sequential trials approach and MSM-IPTW estimate this estimand, and discuss their assumptions and how data are used differently. The performance of the two approaches is compared in a simulation study. The sequential trials approach, which tends to involve less extreme weights than MSM-IPTW, results in greater efficiency for estimating the marginal risk difference at most follow-up times, but this can, in certain scenarios, be reversed at later time points and relies on modelling assumptions. We apply the methods to longitudinal observational data from the UK Cystic Fibrosis Registry to estimate the effect of dornase alfa on survival.

Association of Prenatal Exposure to Benzodiazepines and Z-Hypnotics With Risk of Attention-Deficit/Hyperactivity Disorder in Childhood.

Importance: Evidence is limited regarding the safety of prenatal benzodiazepine and z-hypnotic exposure and its association with long-term neurodevelopment in childhood. Objective: To quantify the associations of the timing and number of intervals of prenatal exposure to benzodiazepines and/or z-hypnotics with the risk of attention-deficit/hyperactivity disorder (ADHD) in childhood. Design, Setting, and Participants: This cohort study used data from the 1999 to 2008 population-based Norwegian Mother, Father and Child Cohort Study, which are linked to the Medical Birth Registry of Norway, Norwegian Patient Registry, and Norwegian Prescription Database. Two populations of participants were created: a full sample and a mental health sample. The full sample included mothers and their live-born singletons, whereas the mental health sample was restricted to offspring of mothers who reported anxiety, depression, or sleeping problems during pregnancy or 6 months before pregnancy. Data were analyzed from September 2021 to February 2022. Exposures: Maternal self-report of benzodiazepine and/or z-hypnotic use during pregnancy was grouped into early pregnancy exposure and middle and/or late pregnancy exposure for analysis of the association with timing of exposure, and number of 4-week intervals of exposure was classified (single [1] vs multiple [≥2]) for analysis of the association with number of exposed intervals. Main Outcome and Measures: The outcome was ADHD, defined as time to ADHD diagnosis or filled prescription for ADHD medication. To control for confounding, inverse probability of treatment-weighted Cox proportional hazards regression models were used. Hazard ratios and 95% CIs were estimated. The weights were derived from propensity score modeling of the probability of benzodiazepine and/or z-hypnotic exposure as a function of potential confounders between the exposure and the outcome, including maternal symptoms of depression and anxiety. Results: The full sample comprised 82 201 pregnancies, and the mental health sample included 19 585 pregnancies. In total, 681 offspring (0.8%) in the full sample and 468 offspring (2.4%) in the mental health sample were prenatally exposed to benzodiazepines and/or z-hypnotics. After weighting, exposure in early (hazard ratio, 0.74; 95% CI, 0.39-1.94) and middle and/or late (hazard ratio, 0.76; 95% CI, 0.35-1.61) pregnancy was not associated with increased risk of childhood ADHD. There was no evidence of substantial association between the number of exposed intervals during pregnancy and childhood ADHD. Conclusions and Relevance: Results of this study suggest that there may be no increased risk of childhood ADHD associated with prenatal exposure to benzodiazepines and/or z-hypnotics, regardless of timing of exposure and number of exposed intervals. However, these findings should be interpreted with caution due to low study power.

Letter to the editor regarding "Covid-19 transmission in fitness centers in Norway - a randomized trial".

In a recently published paper in BMC Public Health we read about a randomized trial on Covid-19 transmission performed in five fitness centers in Oslo, Norway, during the spring of 2020. In our opinion, this study has major shortcomings in design and methodology, which have not been addressed by the authors.

An initiative for a more inclusive working life and its effect on return-to-work after sickness absence: a multistate longitudinal cohort study.

OBJECTIVES: To reduce sickness absence (SA) and increase work participation, the tripartite Agreement for a More Inclusive Working Life (IA) was established in Norway in 2001. IA companies have had access to several measures to prevent and reduce SA. Our aim in this paper was to estimate the average effect of having access to IA at the time of entering a first SA on later return-to-work (RTW) and on time spent in other work-related states. A secondary objective was to study how effects varied between women and men, and individuals with SA due to either musculoskeletal or psychological diagnoses. DESIGN: Population-based observational multistate longitudinal cohort study. SETTING: Individual characteristics and detailed longitudinal records of SA, work and education between 1997-2011 were obtained from population-wide registries. PARTICIPANTS: Each individual born in Norway 1967-1976 who entered full-time SA during 2004-2011, with limited earlier SA, was included (n=187 930). PRIMARY AND SECONDARY OUTCOME MEASURES: Individual multistate histories containing dated periods of work, graded SA, full-time SA, non-employment and education. METHODS: Data were analysed in a multistate model with 500 days of follow-up. The effect of IA was assessed by estimating differences in state probabilities over time, adjusted for confounders, using inverse probability weighting. RESULTS: IA increased the probability of work after SA, with the largest difference between groups after 29 days (3.4 percentage points higher (95% CI 2.5 to 4.3)). Differences in 1-year expected length of stay were 8.4 additional days (4.9 to 11.9) in work, 7.6 (4.8 to 10.3) fewer days in full-time SA and 1.6 (-0.2 to 3.4) fewer days in non-employment. Similar trends were found within subgroups by sex, musculoskeletal and psychological diagnoses. The robustness of the findings was studied in sensitivity analyses. CONCLUSION: Measures to prevent and reduce SA, as given through IA, were found to improve individuals' RTW after entering SA.

Gender Differences in Associations between Biomechanical and Psychosocial Work Exposures and Age of Withdrawal from Paid Employment among Older Workers.

BACKGROUND: Work exposures are known predictors of withdrawal from employment, but the associations between work exposures and withdrawal may vary with gender. This study evaluated gender differences in associations between biomechanical and psychosocial work exposures and age of withdrawal from paid employment among older workers in Norway. METHODS: 77,558 men and 67,773 women (born 1949-1953) were followed from age 62 until withdrawal from paid employment or end of follow-up in 2016 (up to five years follow-up). Information about eight biomechanical and seven psychosocial exposures was obtained from a gender-specific job exposure matrix. Using Cox regression, the difference in mean estimated time until withdrawal between non-exposed and exposed was calculated for each gender and work exposure separately. RESULTS: The largest gender difference was found for high psychological demands. Among men, the non-exposed withdrew earlier than the exposed (-3.66 months (95% CI: -4.04--3.25 months)), and contrary among women (0.71 (0.28-1.10)), resulting in a gender difference of 4.37 (3.81-4.97) months. Gender differences were also found for monotonous work (4.12 (3.51-4.69) months), hands above shoulder height (2.41 (1.76-3.10) months), and high iso-strain (2.14 (1.38-2.95) months). CONCLUSIONS: There were observed gender differences in the associations between some biomechanical and psychosocial work exposures and mean age of withdrawal from paid employment among older workers. However, the results are likely affected by the selection of who remains in the workforce at age 62 and should be interpreted accordingly.

Strategies to manage auxiliary pain medications in chronic pain trials: a topical review.

Chronic pain trials commonly allow auxiliary pain medications such as rescue and concomitant analgesics in addition to the randomized treatment. Changes in auxiliary pain medications after randomization represent intercurrent events that may affect either the interpretation or the existence of the measurements associated with the clinical question of interest, complicating the assessment of treatment efficacy. In chronic pain trials, pain intensity typically varies and patients may take the auxiliary medications 1 day but not the next or increase and decrease the dosages temporarily while continuing their randomized study medication. This distinctive feature of auxiliary pain medications as an intercurrent event has received little attention in the literature. Further clarifications on how to manage these issues are therefore pressing. Here we provide perspectives on issues related to auxiliary pain medication-related intercurrent events in randomized controlled chronic pain trials considering the strategies suggested in the E9(R1) addendum to the ICH guideline on statistical principles for clinical trials.

Injuries after violence and accidents - the forgotten pandemic? / Skader etter vold og ulykker ­ den glemte pandemi?

All information on injuries should be collected in a national injury registry devised for research. This will pave the way for a new strategy for injury prevention.

Impact of the Norwegian Agreement for a More Inclusive Working Life on diagnosis-specific sickness absence in young adults: a difference-in-difference analysis.

BACKGROUND: The Norwegian Agreement for a More Inclusive Working Life (the IA Agreement) aims to reduce sickness absence (SA) and increase work participation. Potential impacts of the IA Agreement have not been thoroughly evaluated. The study aimed to estimate the impact of the IA Agreement on musculoskeletal and psychological SA prevalence and duration among young adult men and women, and to identify whether the impact was modified by economic activity or SA grade. METHODS: Data from national registries were combined for 372,199 individuals born in Norway 1967-1976. ICPC-2 codes identified musculoskeletal (L) and psychological (P) diagnoses. A difference-in-difference method compared prevalence and mean duration of first SA > 16 days between 2000 and 2005 separately for men and women working in IA companies relative to non-IA companies. Analyses were adjusted for mean company size and stratified by economic activity and SA grade (full/graded). Average marginal change was calculated with 95% confidence intervals (CI). RESULTS: The impacts of the IA Agreement on SA prevalence were mixed as the direction and size of marginal changes varied according to diagnosis, gender, and economic activity. However, there was a general tendency towards reduced mean SA duration for both diagnosis groups, and in particular men with musculoskeletal SA (- 16.6 days, 95% CI -25.3, - 7.9). Individuals with full SA in IA companies had greater reductions in mean SA duration. Only the wholesale and retail economic activity indicated a beneficial contribution of the IA Agreement for both SA prevalence and duration, in both diagnoses and genders. CONCLUSIONS: Potential impacts of the IA Agreement on SA in young men and women varied according to diagnosis and economic activity. However, results indicated that the IA Agreement could reduce SA duration. Further research should identify reasons for gender and economic activity differences.

A hybrid landmark Aalen-Johansen estimator for transition probabilities in partially non-Markov multi-state models.

Multi-state models are increasingly being used to model complex epidemiological and clinical outcomes over time. It is common to assume that the models are Markov, but the assumption can often be unrealistic. The Markov assumption is seldomly checked and violations can lead to biased estimation of many parameters of interest. This is a well known problem for the standard Aalen-Johansen estimator of transition probabilities and several alternative estimators, not relying on the Markov assumption, have been suggested. A particularly simple approach known as landmarking have resulted in the Landmark-Aalen-Johansen estimator. Since landmarking is a stratification method a disadvantage of landmarking is data reduction, leading to a loss of power. This is problematic for "less traveled" transitions, and undesirable when such transitions indeed exhibit Markov behaviour. Introducing the concept of partially non-Markov multi-state models, we suggest a hybrid landmark Aalen-Johansen estimator for transition probabilities. We also show how non-Markov transitions can be identified using a testing procedure. The proposed estimator is a compromise between regular Aalen-Johansen and landmark estimation, using transition specific landmarking, and can drastically improve statistical power. We show that the proposed estimator is consistent, but that the traditional variance estimator can underestimate the variance of both the hybrid and landmark estimator. Bootstrapping is therefore recommended. The methods are compared in a simulation study and in a real data application using registry data to model individual transitions for a birth cohort of 184 951 Norwegian men between states of sick leave, disability, education, work and unemployment.

Pre-injury dispensing of psychoactive prescription drugs in a ten years trauma population: a retrospective registry analysis.

BACKGROUND: The use of psychoactive prescription drugs is associated with increased risk of traumatic injury, and has negative impact on clinical outcome in trauma patients. Previous studies have focused on specific drugs or subgroups of patients. Our aim was to examine the extent of psychoactive drug dispensing prior to injury in a comprehensive population of trauma patients. METHODS: The Oslo University Hospital Trauma Registry provided data on all trauma patients admitted to the trauma centre between 2005 and 2014. We linked the data to Norwegian Prescription Database data from 2004. Opioids, benzodiazepines, z-hypnotics, gabapentinoids, and centrally acting sympathomimetics dispensed during the year before trauma of each patient were identified. We determined the pre-trauma annual prevalence of dispensing and mean annual cumulative defined daily doses (DDD) for each drug class, and compared results with corresponding figures in the general population, using standardised ratios. For each drug class, dispensing 14 days preceding trauma was analysed in patients sustaining severe injury and compared with patients sustaining non-severe injury. RESULTS: 12,713 patients (71% male) were included. Median age was 36 years. 4891 patients (38%) presented with severe injury (Injury Severity Score > 15). The ratio between annual prevalence of dispensed prescriptions for trauma patients and the general population, adjusted for age and sex, was 1.5 (95% confidence interval 1.4-1.6) for opioids, 2.1 (2.0-2.2) for benzodiazepines, 1.7 (1.6-1.8) for z-hypnotics, 1.9 (1.6-2.2) for gabapentinoids, and 1.9 (1.6-2.2) for centrally acting sympathomimetics. Compared with the general population, mean annual cumulative DDD of opioids and benzodiazepines dispensed to trauma patients were more than two and three times as high, respectively, in several age groups below 70 years. The prevalence of dispensing 14 days pre-trauma was higher in severely injured patients for opioids, benzodiazepines, and z-hypnotics compared with patients without severe injury. CONCLUSIONS: Our results support previous findings that the prevalence of psychoactive drug use is high among trauma patients. In terms of both frequency and amounts, the pre-injury dispensing of psychoactive drugs to trauma patients supersedes that of the general population, especially in younger patients.

Effect of bariatric surgery on employment status-a 7 years controlled nationwide registry study.

BACKGROUND: Severe obesity is associated with a reduced ability to work. Bariatric surgery is the most effective method to achieve a sustained weight loss. Previous studies have reported conflicting results regarding the effect of bariatric surgery on employment status. To address this, we investigated the effect of bariatric surgery on employment status in the Danish population. METHODS: In this nationwide study, we identified 5450 subjects who underwent bariatric surgery and 10 900 control subjects matched for age, sex and municipality. From accessible registries, we extracted data regarding employment, absenteeism, sick leave and pension. Using a multistate model, we compared time in occupational states and transitions between these states to determine the effect of bariatric surgery on employment status. FINDINGS: Before surgery, cases had an absolute risk increase (95% CI)(ARI (CI)) and a relative risk (RR (CI)) of being in full-time employment of -0.12 (-0.14 to -0.10) and 0.84 (0.82 to 0.86) and were more often unemployed or in a subsidised job than the background population. Taking into account the employment status before surgery, the bariatric surgery group increased their probability of being in full-time employment 1-3 years after bariatric surgery. However, this positive effect was not present with a longer duration of follow-up. Being male, above 50 years of age, or employed as a craftsman or office worker were associated with a sustained positive effect of being in full-time employment (ARI (CI) and RR (CI) 0.05 (0.04 to 0.05) and 1.05 (1.04 to 1.06), 0.06 (0.06 to 0.07) and 1.08 (1.07 to 1.09) and 0.05 (0.05 to 0.06) and 1.05 (1.05 to 1.06), respectively). INTERPRETATION: Compared with a matched control group, those undergoing bariatric surgery did not improve their employment status in the long term. Certain subgroups had a more sustained positive effect.