320 questions about employee benefits

The best evidence in employee benefits is for the least glamorous thing in the catalogue. When someone is off sick and their workplace gets involved in getting them back, total time off falls. The evidence for that is moderate, which in this field is about as good as it gets.

The benefits I see sold hardest are health cash plans, discounts and perks, and electric car salary sacrifice. For those we could find no published study of any outcome at all.

Some of this is penance. I spent years at a big consultancy being paid to sound as if I understood wellbeing, which I didn’t.

What we did

In early October we built an evidence base at Alltoogether. Thirty-two benefits, each against ten outcomes: sickness absence, long-term absence and getting back to work, productivity, staff turnover, recruitment, mental health, physical health, financial wellbeing, engagement, and whether anyone uses the thing. That’s 320 questions.

Each one was searched and graded with the logic medical guideline groups use. Randomised trials start high and observational studies start low, and both drop a level for bias, for testing something other than the thing you care about, for disagreeing with each other, or for being too small. Provider surveys were written down and never allowed to set a grade.

What came back

Nothing reached high certainty. Fourteen of the 320 reached moderate, 54 low and 64 very low. For 188 there was no usable published study.

Some of those gaps don’t matter. Nobody needs a trial of electric car salary sacrifice and back pain, and around 30 of the 188 are like that. The rest are questions someone could answer and hasn’t.

Health cash plans are empty on all ten. The number that does the rounds, 4.1 sick days saved, comes from a provider’s article that doesn’t say how it was worked out.¹ Discount platforms have one record of anything, from a fire and rescue service that found 6.1 per cent of its staff used its discount site in the scheme’s second year, back in 2010.² Electric car salary sacrifice has nothing. That’s the lot.

All it means is that nobody has checked, so nobody knows either way whether these benefits work. That should be said out loud more often, ideally by someone who doesn’t sell them for a living. Unfortunately, you’ve got me.

Evidence that something doesn’t work

Some pairs have decent evidence of no effect, which is a stronger finding than an empty box, and it’s where a few favourite claims live.

Two large US trials randomised workplace wellness programmes. Across 160 worksites of a warehouse retailer, the programme made no significant difference to absence, job tenure or job performance after 18 months.³ In Illinois, the people who signed up were already healthier, and already spent less on medical care, in the year before it started.⁴ Compare joiners with everyone else and the programme looks brilliant, because the joiners were healthier to begin with.

Faster physiotherapy is similar. In two UK trials, being seen sooner meant a shorter wait and the same recovery.⁵⁶ A shorter wait is worth something if you’re the one in pain. It’s a different thing to buy, and it should be sold as that.

General health checks have a Cochrane review of 17 trials whose authors concluded they’re unlikely to be beneficial.⁷

Where the moderate evidence sits

Return-to-work support that involves the workplace. Across 14 trials it got people back sooner and cut total time off sick.⁸ For depression, work-directed support combined with clinical care probably saved about 25 sick days a year, though a year or more on, no more people were in work.⁹ Coordination programmes on their own, which insurers, providers and employers all run, made no difference.¹⁰

Lower-grade studies point the same way. With physiotherapy plus an early conversation with the workplace, 85 per cent were working at a year against 74 per cent.¹¹ People at high risk of long absences who got an occupational health consultation had 19 days off on average, against 30. Phone advice for the medium-risk group did nothing, mostly because people didn’t use it.¹²

Digital mental health programmes and wellbeing apps help symptoms a bit. Across 81 trials, most of them at high risk of bias, they brought small reductions in stress, anxiety and depression.¹³ In the trials that measured sickness absence, it didn’t move.¹⁴¹⁵ So put them in the business case for symptoms and leave them out of the absence line.

Then there’s use. Staff given only an app used it for 45 of the 343 minutes they were asked to.¹⁶ That’s still more than I’d have managed. In Great Britain in 2020, half of employees said they had occupational health access, a fifth didn’t know, and 70 per cent of those with access hadn’t used it in the past year.¹⁷ A trial measures the effect on people who were offered something. In a real workplace, it also depends on people knowing it’s there.

What this can’t tell you

The grades are ours. We used the method properly, I think, but we’re not a guideline panel. I’m a broker with an MSc and a spreadsheet problem. Most sources were read at abstract level. A random fifth of the graded pairs was re-checked by reviewers who hadn’t done the grading, and about four in ten needed their wording fixed. The other four fifths haven’t been re-checked yet, so every grade is provisional.

Only 34 of the 132 graded pairs rest on UK evidence. Most trials tested a service, often in a clinic, which is a step removed from the benefit an employer buys, and they were marked down for it. An empty box isn’t a verdict. And a trial result is an average from somebody else’s workforce, which yours may not look like.

What would change my mind

For cash plans, a controlled study across UK employers using payroll absence records, or a staggered roll-out. Discounts need a job-advert experiment that varies whether they’re mentioned, plus real transaction data alongside a proper financial wellbeing measure. For insurers’ early intervention, a randomised trial against usual care for people off with common mental health problems. If any of those comes back, the grade changes and this note gets a dated correction.

Until then, the least glamorous thing in the catalogue has the best evidence. It still won’t make my brochure.

Footnotes

1. Hulme, J. (2025). Health cash plan myths: the hidden value employers can’t afford to ignore. REBA. reba.global. ↩

2. Nottinghamshire and City of Nottingham Fire and Rescue Authority (2011). Employee discount scheme: report to the Human Resources Committee, 28 January 2011, paragraph 2.7. committee.nottinghamcity.gov.uk. ↩

3. Song, Z. and Baicker, K. (2019). Effect of a workplace wellness program on employee health and economic outcomes: a randomized clinical trial. JAMA, 321(15), 1491. doi:10.1001/jama.2019.3307. ↩

4. Jones, D., Molitor, D. and Reif, J. (2019). What do workplace wellness programs do? Evidence from the Illinois Workplace Wellness Study. Quarterly Journal of Economics, 134(4), 1747-1791. doi:10.1093/qje/qjz023. ↩

5. Salisbury, C. et al. (2013). A pragmatic randomised controlled trial of the effectiveness and cost-effectiveness of PhysioDirect telephone assessment and advice services for physiotherapy. Health Technology Assessment, 17(2). doi:10.3310/hta17020. ↩

6. Jones, G.T. et al. (2019). Maintained physical activity and physiotherapy in the management of distal arm pain: a randomised controlled trial. RMD Open, 5(1), e000810. doi:10.1136/rmdopen-2018-000810. ↩

7. Krogsbøll, L.T., Jørgensen, K.J. and Gøtzsche, P.C. (2019). General health checks in adults for reducing morbidity and mortality from disease. Cochrane Database of Systematic Reviews, CD009009. doi:10.1002/14651858.CD009009.pub3. ↩

8. van Vilsteren, M. et al. (2015). Workplace interventions to prevent work disability in workers on sick leave. Cochrane Database of Systematic Reviews, CD006955. doi:10.1002/14651858.CD006955.pub3. ↩

9. Nieuwenhuijsen, K. et al. (2020). Interventions to improve return to work in depressed people. Cochrane Database of Systematic Reviews, CD006237. doi:10.1002/14651858.CD006237.pub4. ↩

10. Vogel, N. et al. (2017). Return-to-work coordination programmes for improving return to work in workers on sick leave. Cochrane Database of Systematic Reviews, CD011618. doi:10.1002/14651858.CD011618.pub2. ↩

11. Sennehed, C.P. et al. (2018). Early workplace dialogue in physiotherapy practice improved work ability at 1-year follow-up. Pain, 159(8), 1456-1464. doi:10.1097/j.pain.0000000000001216. ↩

12. Taimela, S. et al. (2008). The effectiveness of two occupational health intervention programmes in reducing sickness absence among employees at risk. Occupational and Environmental Medicine, 65(4), 236-241. doi:10.1136/oem.2007.032706. ↩

13. Stratton, E. et al. (2025). Optimizing workplace digital mental health interventions: systematic review and meta-analysis. Journal of Medical Internet Research, 27, e71253. doi:10.2196/71253. ↩

14. Taylor, H., Cavanagh, K., Field, A.P. and Strauss, C. (2022). Health care workers’ need for Headspace: findings from a multisite definitive randomized controlled trial. JMIR mHealth and uHealth, 10(8), e31744. doi:10.2196/31744. ↩

15. Udd-Granat, L. et al. (2023). Internet-delivered cognitive behavioral therapy (iCBT) for common mental disorders and subsequent sickness absence: a systematic review and meta-analysis. Scandinavian Journal of Public Health, 51(1), 137-147. doi:10.1177/14034948221075016. ↩

16. Bartlett, L. et al. (2022). Effects of a mindfulness app on employee stress in an Australian public sector workforce: randomized controlled trial. JMIR mHealth and uHealth, 10(2), e30272. doi:10.2196/30272. ↩

17. Moore, G. and Ezzard, O. (2023). Employee research phase 1: sickness absence, reasonable adjustments and occupational health. Department for Work and Pensions research report 1021, chapter 5. Fieldwork September to October 2020. assets.publishing.service.gov.uk. ↩