IPS is not a conventional job-search scheme. Developed originally for people with severe mental illness, it follows a place-then-train philosophy: rather than screening and pre-training candidates before any contact with employers, employment specialists embedded within clinical teams move clients rapidly into competitive jobs matched to their preferences, then provide ongoing practical and health-related support to keep them there. Decades of randomised trial evidence show IPS consistently outperforms alternative supported employment approaches, and England has been at the forefront of extending the model beyond severe mental illness. Recent large-scale trials, most notably the Department for Work and Pensions’ Health-Led Employment Trial (HLET), tested IPS in people with low-to-moderate mental health (MH), musculoskeletal (MSK), or combined MH and MSK conditions, populations that had previously been largely excluded from the evidence base.

Despite that clinical evidence, policymakers have lacked a rigorous way to judge whether the economics stack up. A recent systematic review found that 79 percent of 56 economic studies reported a positive case for IPS, but few used long-term modelling, disaggregated results by population subgroup, or tested how sensitive returns were to service costs and performance. The new study addresses these gaps with IPS-MOD, an open-source Excel-based cost-benefit and outcomes model built to HM Treasury Green Book standards. It simulates a cohort of non-employed adults with MH, MSK or combined conditions over five years in quarterly steps, comparing a scenario in which they receive IPS against standard UK employment support such as Jobcentre Plus provision.

The model’s architecture is deliberately grounded in real-world data at every node. Labour market transition probabilities between employment and non-employment were estimated from 27 five-wave longitudinal datasets of the UK Labour Force Survey covering 2012 to 2018, then calibrated so the model reproduces the year-one employment outcomes actually observed in the HLET for both intervention and control arms. Population sizes draw on Office for National Statistics mid-year estimates combined with local authority prevalence data and the 2019 Health Survey for England. Health gains are captured in quality-adjusted life years (QALYs), combining length and quality of life, estimated from EQ-5D survey data and adjusted for the utility gain associated with moving into work using Understanding Society regression analyses. Each QALY is valued at £79,468 in 2024 prices, in line with Green Book guidance, while earnings, taxes and benefits are computed with the UKMOD tax-benefit microsimulation model applied to the Family Resources Survey.

In the base case, set in a typical local authority of 180,000 working-age adults, the model assumes 30 percent of the non-employed population with each condition voluntarily joins IPS. For the MSK group that yields 3,749 participants at a service cost of £2,326 per person. Trial-calibrated effectiveness figures show 28 percent of IPS participants in work at some point in the first 12 months compared with 24 percent receiving usual support; for the MH group the figures are 57 percent versus 53 percent, and 28 versus 23 percent for those with both conditions. The gap looks modest, but across thousands of people and compounded over five years it produces substantial economic and health dividends.

Those dividends accumulate quickly. By the end of year one, employment among MSK participants on IPS stood at 21.55 percent against 16.73 percent without the service, an absolute gain of 4.82 percentage points, equivalent to a 28.83 percent relative improvement. By year five the gap narrows in proportional terms but remains a 3.01 percentage point advantage, alongside an extra 8.65 weeks of paid work per person. Health utility improved by an average of 0.028 QALYs per participant in year one and 0.137 QALYs by year five, worth £2,201 and £10,881 per person respectively, or £8.25 million and £40.80 million across the population. Participants themselves accumulated average gains of £2,352 in year one and £11,718 over five years, while the exchequer collected £6.63 million in total benefits by year five through higher tax and national insurance receipts, £3.79 million in reduced social security payments and £467,105 in avoided NHS spending.

The headline result is the benefit-cost ratio. For the MSK population, every £1 spent on IPS returned £1.13 by the end of year one, rising steadily to £5.71 by the end of year five as wages, health gains and avoided benefit payments compounded. The MH group returned a similar £5.67 over five years, and the combined MH and MSK group £4.13, reflecting that group’s lower baseline rates of movement into work. Notably, the model’s returns are driven primarily by monetised health gains rather than fiscal effects alone, a departure from many economic evaluations that count only government savings. Sensitivity analysis indicates the five-year positive return would survive even if a QALY were valued at roughly £10,000, well below the official £79,468 figure.

The team stress-tested the model extensively. Two-way sensitivity analyses varied IPS effectiveness between 18 and 45 percent employment entry and service costs between £1,500 and £3,300 per participant. Most scenarios produced a positive return on investment within the first year, and by year five every modelled combination was cost-beneficial, even though the lowest-effectiveness, highest-cost corners delivered short-term negative returns. Because published trial evidence only calibrates the model for 12 months, the default scenario conservatively assumes treatment effects fade to 25 percent of their year-one value by year five, reflecting IPS’s continued in-work support, while direct effects on health utility are assumed permanent. Alternative persistence assumptions are built in as user-selectable options.

The limitations are acknowledged candidly. The model cannot run full probabilistic sensitivity analysis because the HLET published no distributions for key parameters. It assumes trial results from South Yorkshire and the West Midlands generalise nationally, and that the relative effect of IPS applies equally across health conditions and disability statuses. As a two-state Markov model it also cannot distinguish good jobs from insecure ones, even though the health benefits of work likely depend on job quality. Nevertheless, the timing of the analysis is significant: as the £1 billion Connect to Work programme scales IPS across England and Wales for jobseekers with low-to-moderate MH and MSK conditions, IPS-MOD gives commissioners a transparent, adaptable tool to forecast returns before committing budgets, and its framework could be exported to any country with a national health survey and longitudinal labour data.

What emerges is a rare case in public policy where the clinical evidence, the health economics and the fiscal arithmetic all point in the same direction. Unemployment is both cause and consequence of poor health, with people out of work two to three times more likely to experience mental health problems than those employed, while good work is linked to reduced suicide risk, relapse and better physical health. By quantifying those feedbacks in monetary terms and showing that returns grow rather than shrink over time, the study makes the case that helping people with common health conditions back into work is not merely a welfare expense to be managed but an investment with measurable, multi-stakeholder payoffs, from the individual’s payslip and quality of life to the Treasury’s balance sheet.

Subject of Research: Cost-benefit modelling of Individual Placement and Support supported employment for people with mental health and musculoskeletal conditions in England

Article Title: Estimating the health and economic impacts of Individual Placement and Support (IPS) for people with mental and physical health conditions: A cost-benefit modelling study

Article References: Clarke, H., Morris, D., James, M., Brennan, A., & Whitworth, A. (2026). Estimating the health and economic impacts of Individual Placement and Support (IPS) for people with mental and physical health conditions: A cost-benefit modelling study. Public Health in Practice, 12, Article 100867. https://doi.org/10.1016/j.puhip.2026.100867

Image Credits: AI Generated

DOI: Not provided

Keywords: Individual Placement and Support, IPS-MOD, supported employment, mental health, musculoskeletal conditions, cost-benefit analysis, QALYs, economic inactivity, Health-Led Employment Trial, health economics, UK labour market, Connect to Work

An umbrella review sits at the very top of the evidence hierarchy. Rather than analyzing individual clinical trials, it systematically collects and critically appraises all of the existing systematic reviews on a topic, effectively auditing the audits. The research team, led by Adrian Kużdżał of the University of Rzeszów in Poland and including Edzard Ernst of the University of Exeter, a longtime and often pointed critic of unsubstantiated complementary therapies, searched eight major databases including PubMed, MEDLINE, Embase, the Cochrane Database of Systematic Reviews, Web of Science, Scopus, PEDro and Epistemonikos. Their search window stretched from January 2000 to December 2025, capturing a quarter century of accumulated review literature on dry needling delivered with or without usual care by a range of healthcare professionals for musculoskeletal or non-musculoskeletal pain and dysfunction in human subjects. Narrative and non-systematic reviews were excluded, and the team strictly followed the PRIOR reporting guidelines for overviews of reviews.

The scale of the enterprise is striking. The researchers identified 75 systematic reviews that met their eligibility criteria, collectively covering 45,032 individual participants. That is an enormous evidence base by any standard, and it reflects just how much clinical and academic energy has been poured into dry needling over the past two decades. Yet the headline numbers tell a story of deep division within that literature. Fifty of the 75 reviews, or 66.6 percent, reported conclusions favoring dry needling. Seventeen reviews, or 22.6 percent, came down on the fence with equivocal conclusions, while eight reviews, or 10.6 percent, reported outright negative conclusions. On the surface, two thirds of the review literature appears to endorse the technique, and it is easy to see how advocates might cite that figure as proof of effectiveness.

But the authors dug one level deeper, and this is where the picture changes dramatically. Using the ROBIS tool, a validated instrument for assessing the risk of bias in systematic reviews, they found that 57 of the 75 reviews, a full 76 percent, demonstrated a high risk of bias across multiple domains. The problems clustered in the most consequential stages of the review process: study selection, where reviewers may cherry-pick which trials to include; data appraisal, where the quality of the underlying trials is judged; and synthesis, where results are pooled and interpreted. In other words, the machinery that is supposed to make systematic reviews trustworthy was itself frequently compromised. A review with high risk of bias in these domains can produce optimistic conclusions that simply do not survive rigorous scrutiny.

The most telling analysis concerns the minority of reviews that passed the quality test. Among the 18 systematic reviews judged to be at low risk of bias, the picture was far less rosy than the raw numbers suggested. Twelve of those 18, or 66.6 percent, still reported favorable conclusions, but five, or 27.7 percent, were equivocal and one, or 5.5 percent, was negative. Even among the best-conducted reviews, a third failed to find clear benefit, and the authors emphasize that the certainty of the evidence underpinning the favorable conclusions remained limited because of high risk of bias in the primary trials and substantial statistical heterogeneity between studies. Heterogeneity, in the language of meta-analysis, means that the pooled effect sizes extracted from the reviews varied widely across conditions, outcome measures and comparison groups, making it difficult to state with confidence what a typical patient should expect.

The umbrella review also examined the degree to which the 75 reviews were recycling the same primary trials, a problem known in evidence synthesis as overlap. When many reviews analyze overlapping pools of studies, the apparent volume of evidence is inflated and the reviews are not independent tests of the hypothesis. The researchers quantified this using the Corrected Covered Area, or CCA, a standard metric for overlap assessment. The result was a CCA of 1.09, which the authors classify as slight overlap. That finding is genuinely reassuring in one respect: the 75 reviews are largely built on different combinations of primary studies rather than repeatedly reanalyzing the same small trial set. The problem, then, is not redundancy but quality, since the sheer breadth of the literature has not translated into methodological rigor.

Safety is where the review delivers perhaps its most practical warning. Dry needling involves penetrating the skin with needles, and like any invasive procedure it carries potential for adverse effects, ranging from common transient soreness and bleeding at the needle site to rarer but more serious events such as pneumothorax when needles are placed near the chest wall. Yet the review found that only 36 of the 75 systematic reviews, or 48 percent, reported any adverse-effect data at all. The remaining 39 reviews, more than half, failed to provide any information on harms. This pattern of underreporting is a well-recognized blind spot in the complementary and manual therapy literature, where enthusiasm for benefit often outpaces systematic attention to risk. Patients and clinicians cannot make genuinely informed decisions when half of the evidence synthesis literature is silent on safety.

Putting all of this together, the authors conclude that no definitive judgments about the clinical effectiveness of dry needling in reducing pain and disability can currently be drawn, and that the existing findings should be interpreted with caution. That is not the same as declaring the technique useless; it is a statement about the state of the evidence. The favorable conclusions reported by many reviews may reflect real effects, but they may equally reflect biased trial design, small samples, unblinded outcome assessment, publication bias and the methodological weaknesses that pervade three quarters of the review literature itself. The authors call for high-quality, condition-specific systematic reviews and randomized clinical trials to resolve the uncertainty, and their work is registered in PROSPERO under identifier CRD420251058215, underscoring the transparency of their own protocol.

For clinicians, the implications are nuanced. Dry needling is often used for conditions such as myofascial pain syndrome, low back pain, tension-type headache, cervicogenic headache and fibromyalgia, and outcome measures in the reviewed literature included familiar instruments such as the Numeric Pain Rating Scale, the Visual Analog Scale, the Neck Disability Index, the Oswestry Disability Index, the Roland-Morris Disability Questionnaire and the WOMAC osteoarthritis index. The umbrella review does not show that the technique harms patients in the conditions studied, and the slight overlap analysis suggests the field has genuinely explored a wide range of clinical questions. But a therapy whose evidence base is dominated by high-risk-of-bias reviews cannot be presented to patients as proven, and the inconsistent safety reporting means that even the risk profile is incompletely mapped.

For the broader public, the study is a masterclass in why the phrase systematic review alone should not end an argument. Seventy-five reviews and forty-five thousand participants sound like overwhelming evidence, yet the umbrella review demonstrates that the architecture of that evidence, not just its quantity, determines what can be believed. The work, funded by the Statutory Research Fund of the Collegium Medicum at the University of Rzeszów with support from the Polish Ministry of Science and Higher Education, and published open access under a Creative Commons license, offers a comprehensive map of where dry needling research stands and where it must go. Until better trials and better reviews arrive, the needles may keep going in, but the verdict on dry needling remains, scientifically speaking, an open question.

Subject of Research: Effectiveness and safety of dry needling for pain and dysfunction, evaluated through an umbrella review of systematic reviews

Article Title: How effective is dry needling? an umbrella review

Article References: Kużdżał, A., Ernst, E., Trybulski, R., & Posadzki, P. (2026). How effective is dry needling? an umbrella review. BMC Complementary Medicine and Therapies. https://doi.org/10.1186/s12906-026-05607-9

Image Credits: AI Generated

DOI: 10.1186/s12906-026-05607-9

Keywords: dry needling, umbrella review, systematic review, myofascial trigger points, pain, disability, risk of bias, ROBIS, adverse effects, physical therapy, evidence-based medicine, musculoskeletal conditions