Characteristics of included studies evaluating screening followed by feedback, advice and/or referral
| Study characteristics | Population | Screening tool; delivery | Screening process | Sample size; gender (% male); mean age (years) | Study design; measurement time points; intervention; control group | Advice description; delivery; adherence (% of participants) | Outcomes of interest (scale) | Efficacy of screening*/intervention |
| Kant et al;† The Netherlands in 200337 | Employees of large banking company (formally employed, not specified if full time/part time, mean work hours/week≈34). | Balansmeter, 34 items assessing multiple health domains including mental health; Delivered by occupational health service from participating company using posted questionnaire. | N=9863 invited to complete screening (completed by n=4950). Those screening ‘at-risk’ of sickness absence (n=299) (and who were not absent from work, not pregnant, not receiving treatment from the OP at the time of screening) were randomised. | IG: n=132 73.50% 46.32 CG: n=131 68.70% 46.58 | RCT; Baseline, 1-year follow-up; Screening+consultation with OP; Screening+care as usual | Invited to OP consultation (1–1.5 hours) focusing on clarifying symptoms from screening, discussing benefits of treatment, and referral to targeted intervention); Delivery same as screening; 75%. | Sickness absence: duration, frequency, incidence of long-term absence (individual record linkage using company registers of certified sickness absence). | ITT analyses showed no significant intervention effect on sickness absence outcomes. Modified ITT (excluding those who had already sought treatment) revealed significant effect of intervention on sickness absence duration, but not frequency or incidence of long-term absence. |
| Klasen et al;†36 | – | – | – | – | 2-year, 3-year, 4-year and 5-year follow-up | – | Sickness absence: duration, frequency, incidence of long-term absence (individual record linkage using company registers of certified sickness absence). Work-related: termination of employment contract (company registers). | Kant et al 37 observed a marginally significant decrease in sickness absence duration and incidence of long-term absence at 5 years, no effect on frequency. No effect on termination of employment observed. |
| Gärtner et al;‡ The Netherlands in 201140 | Nurses/allied health professionals employed at an academic medical centre (formally employed, mix of temporary, part time and full time, mean work hours/week≈30). | WHS Mental Module screening for work functioning and mental health complaints; Delivered by the research team online. | Participants prerandomised according to cluster (ward). N=1152 invited, n=379 enrolled (completed baseline and were not expected to be on sick leave for >2 weeks at the study commencement). Once completed, intervention participants received automatically generated results and those screening positive (n=151) were invited to OP consult. Control participants did not receive screening results or any intervention. | IG: n=191 18.00% 43.00 CG: n=188 23.00% 42.00 | Cluster RCT; baseline, 3-month, 6-month follow-ups; Screening and feedback+OP consultation; wait-list control. | Voluntary face-to-face OP consultation followed 7-step protocol similar to employee-initiated care-as-usual consultation but with a preventive focus; delivered by in-company occupational health service; 34% | Help-seeking: formal (visiting 1 of 11 caregivers). Mental health symptoms: psychological distress (4DKL), depression and anxiety (BSI), PTSD (IES). Work-related: impaired work functioning (NWFQ), need for recovery after work (VBBA). | Marginally significant effect of intervention on help-seeking at 3-month follow-up (stable in intervention, decline in control) attenuated at 6-month follow-up. No effect of intervention on mental health outcomes observed. Significant effect found for impaired work functioning favouring intervention at both 3 and 6 months. No effect on need for recovery after work. |
| Ketelaar et al;§ The Netherlands in 201141 | – | – | – | – | 3-month and 6-month follow-ups. | – | User satisfaction (participant perspectives on screening including usefulness of advice/e-mental health programme provided, acceptability of feedback, perceived efficacy or the advice/e-mental health programme, and preferences regarding future screening). | 41% (51 from 125) of participants from Gartner et al attended preventive consult, 51% (18 from 35) reported receiving advice, and 80% (16 of 20) reported following such advice. 13% (16 from 126) of participants reported wanting feedback differently. 60% (9 from 15) participants felt advice helped improve their mental health/work functioning. 37% (41 from 112) would appreciate to be offered the screening intervention in the future. |
| Farzanfar et al; United States date not specified44 | Employed individuals from various industries/workplaces (type of employment not specified). | WHO-5 and 1-item from PHQ; delivered by the research team via telephone. | N=463 volunteered to participate. Those screening positive (n=164) for emotional distress (WHO-5) and functional impairment (PHQ), who were not currently undertaking mental health treatment, were randomised. Control participants were advised to consult their usual care provider if required. | IG: n=87 27.00% 39.00 CG: n=77 21.80% 39.20 | RCT; baseline, 3, 6 months; Screening+TLC-Detect (all modules); screening only+TLC-detect (screening only). | Automated phone intervention with three modules: (1) screening, (2) intervention, containing education and referral for self-help or professional care tailored to symptoms, and (3) follow-up, involving monthly calls to monitor progress and adherence; delivered via algorithm on the phone; 78% responded to follow-up call. | Mental health symptoms: depression (PHQ-9), stress (PSS). Positive mental health: well-being (WHO-5). QoL and functioning: Mental and physical well-being (SF-12). Work-related: productivity, absenteeism, presenteeism (Work Limitation Questionnaire). User satisfaction: usability (HTQ). | Significant intervention effect on depression at 3 months, attenuated at 6 months. No effect on stress or well-being observed. No significant effect on QoL at 3 months, but effect observed at 6 months. No effect of intervention on work-related outcomes, except for time/scheduling at 6 months and interpersonal scale at 3 months. 84% intervention participants reported ease of use and 76% said intervention was informative. Not assessed in control. |
| Addley et al (2014);¶ UK in 200943 | Administrative and executive grade employees within civil service (formally employed, 87% full time). | Health risk appraisal (‘LPAA’) assessing multiple health domains including stress; delivered by the company occupational health service. | N=1503 invited to complete baseline questionnaire. From the n=334 who completed, n=180 were randomised. Those allocated to intervention completed the LPAA, which provides a printout of an individual’s health risk profile and general advice. Those allocated to control received no LPAA or subsequent intervention. | IG—group B: n=43 54.00% 60% aged 25–44 CG: n=48 46.00% 50% aged 25–44 | RCT; Baseline and 12 months; IG-group B: screening; No intervention. | IG—group B received the 45 min HRA assessment only; Delivered by the company occupational health service and company sport organisation; Not reported. | Mental health symptoms: psychological distress (GHQ-12). Positive mental health: well-being (WHO-5 and MHC-SF). Work-related outcomes: work ability (WAI) and job satisfaction (single item). | No effect of either intervention on any outcome variables. |
| van Holland et al; The Netherlands in 201242 | Meat processing plant employees (formally employed for at minimum 12 hours/week). | POSE programme assessing multiple health domains including mental health; Delivered part online part face to face in the workplace by a registered vocational physiotherapist. | Participants prerandomised by cluster (workplace). N=986 invited. Intervention participants agreeing to participate completed screening tests to create a risk profile. | IG: n=305 89% 50.60 CG: same as IG (see trial design) | Cluster randomised stepped wedge trial; baseline, 1–3 years follow-up (depending on wedge); POSE (screening+counselling/referral); care as usual. | Built into screening procedures (as part of POSE) was a counselling session regarding participant results. Based on scores, participants received recommendations on future interventions (ie, consulting a GP, physio, etc); Not reported. | Positive mental health: vitality (RAND-36). QoL and functioning: self-rated health (EuroQol-5D). Work-related: sickness absence (company data), work ability (WAI), productivity (QQ), psychosocial work factors (COPSOQ II). | No effect of intervention found on any outcomes. Significantly higher odds of longer-term sickness absence and lower productivity in intervention vs control at follow-up. |
| Steel et al;** Belgium in 201946 | Hospital workers in roles requiring safety functions, heightened vigilance, work involving physical, biological, or chemical agents or tasks that are ergonomically or mentally burdensome (52.6% medical personnel). | Physical assessment and OP consultation or online health screening questionnaire (GHQ-12 and other measures of work-related illness); OP consultation conducted face to face, screening questionnaire conducted online. | Participants allocated to one of four groups (randomly allocated and non-random allocated IG and randomly allocated and non-randomly allocated CG). IG participants completed online health screening questionnaire and the top 20% of participants reporting symptoms most indicate of needing OP contact were targeted and referred for OP consultation to discuss results. The other 80% did not receive further care. CG participants were invited to attend health screening with an OP consultation. | IG: n=380 17.1% 45.10 CG: n=396 19.7% 45.63 | RCT trial with quasi- randomisation; baseline, 6-month and 18-month follow-ups; random and non-random IG and GC. | OP could refer participants on to other healthcare providers depending on their presenting issue; OP was employed by hospital site; All CG participants received OP consult but % taking up subsequent referral not stated, all IG participants received online health screening and N=126 were referred to OP but N/% taking up OP consult, or any subsequent referral not stated††. | Mental health symptoms: stress and burnout (COPSOQ-III), general mental health (GHQ-12) Work-related: absenteeism and presenteeism (iPCQ), need for recovery after work (NFR), role conflicts and job satisfaction (COPSOQ-III), turnover intention (VBBA) User satisfaction: trust in OP (Trust in Physician Scale) QoL and functioning: self-perceived health (EQ-5D) Help-seeking: spontaneous consultations with OP | No overall group differences between groups in GHQ-12 but reduced general mental health for IG at 18-month follow-up. Non inferiority between groups reported for GHQ-12 and stress. No overall difference in turnover intention, but IG had fewer absences not observed at follow-up. Superior post hoc differences favouring targeted screening in absenteeism. Non inferiority between groups reported for absenteeism, presenteeism, need for recovery, and job satisfaction. IG reported lower trust in OP overall. No overall difference between groups for EQ-5D. Non inferiority reported. IG had fewer spontaneous consultations with OP overall not sustained at 18-month follow-up. Superior post hoc differences favouring CG in spontaneous consultations. |
*A significant intervention effect refers to instances where a study reported a significant difference in a particular outcome between intervention and control groups comparing preintervention and postintervention scores on that outcome.
†Klasen et al 36 reports results of Kant et al 37 and Lexis et al 38 (see table 2) at 2–5 years follow-up (in 1-year increments).
‡Gärtner et al 40 and Boiler et al 39 (see table 2) report revaluations from the one, three-armed trial ‘Mental Vitality @ Work’ trial. As such, the control groups between these two studies are the same.
§Ketelaar et al 41 report additional data from the intervention groups reported in Gärtner et al 40 and Boiler et al.39
¶Three-arm trial comparing two interventions to control, both reported in the one paper. Results from other intervention arm are reported in table 2. Note the Cochrane recommended formula for calculating SD from SEs was applied to extracted quantitative data from this study.
**Note this article was identified via updated database searching and follow-up data collection occurred during the COVID-19 pandemic.
††Note number of referrals measured and means for whole sample at each survey wave reported, however, N/% of referrals not reported.
BSI, brief symptom inventory; CG, control group; COPSOQ II, Copenhagen Psychosocial Questionnaire; 4DKL, Four-Dimensional Symptoms Questionnaire; EQ-5D, EuroQol 5-dimension Health-Related Quality of Life Instrument; GHQ-12, 12-Item General Health Questionnaire; HTQ, Health Technology Questionnaire; IES, Impact of Events Scale; IG, intervention group; iPCQ, IMTA Productivity Cost Questionnaire; ITT, intention to treat; LPAA, Lifestyle and Physical Activity Assessment; MHC-SF, Mental Health Continuum Short Form; NFR, Need for Recovery Scale; NWFQ, Nurses Work Functioning Questionnaire; OP, occupational physician; PHQ, Patient Health Questionnaire; POSE, Promotion of Sustained Employability; PSS, Perceived Stress Scale; QQ, Quality and Quantity Questionnaire; RCT, randomised controlled trial; SF-12, 12 Item Short Form Survey; VBBA, Questionnaire on the Experience and Evaluation of Work; WAI, Work Ability Index; WHS, Work Health Surveillance.