It was with interest we read the results of the recent paper by Killackey et al evaluating the efficacy of individual placement and support (IPS) in first-episode psychosis (FEP) measured towards vocational outcomes and were excited to see positive results at earlier outcome points.^{Reference Killackey, Allott, Jackson, Scutella, Tseng and Borland1} The authors kindly offer multiple explanations of a loss of significance of such interventions at later end points, which we would like to contrast and expand upon.

The authors concluded that the research demonstrated a case for IPS in helping those with FEP return to work, despite this not being maintained at later time points. However, this may be optimistic given the key demographics examined could be potentially an overestimation.

First, the method of selection and randomisation denotes generous selection criteria but fails to fully account for an unequal distribution of affective presentations between group strata and self-selection of willing participants, suggesting a lower baseline functional impairment.

The evidence provided suggests short-term efficacy of IPS in motivated individuals of native language, with less severe symptoms, within a ‘schizonormative’ group and within a well-funded system. This is in keeping with previous research, but provides a relatively narrow margin for intervention, which does not clearly explain how and why it would be effective in those failing to make vocational recovery during usual treatment.

We would consider that those failing in the usual treatment arm may lie outside of the actionable group demonstrated, attributable to the coexistence of affective subtypes, more severe disease (omitted from the study), reduced language efficacy or other factors. Although some of this has evened out at later end-points, this may be because of other factors (remission of anergia, efficacy of pharmacological intervention). We would please request the authors address the above assertion.

Although the study does demonstrate some efficacy with those with a mood-affective component, as well as comorbid anxiety and post-traumatic stress disorder, unless further information is made available, it does not extrapolate the relevance of these factors into short- or long-term outcomes in a comprehensive way. We note that the control group included more affective presentations, which we hope have been accounted for.

The authors also comment that as a result of previous work, those operating in clinics may have been previously upskilled in IPS, which may also explain the high vocational rates ascribed to the treatment as usual arm and could have nullified significance at later end points. We would proffer an additional explanation, where some observer biases inherent in those clinics sharing both intervention and test cases would inspire more supportive treatment.

Finally, the measure of a successful outcome used may overestimate the success of the project, and the attenuation of employment in the intervention arm may further explain the loss of significance at later end points. We posit this could potentially empower the government into seeking more zero-hour contracts in the UK as evidence of employment. However, far from being a negative, it does suggest that IPS, and work around it, does favour positive outcomes and this has been reflected in long-term follow-up in a similar study in Switzerland^{Reference Hoffmann, Jäckel, Glauser, Mueser and Kupper2} and shown to improve outcomes. This suggests moreover that a team approach to support, regardless of vocational expert inclusion, may be the really successful intervention.

It may be that these programmes would present a viable intervention if tailored to adjust for affective symptoms, be maintained beyond 6 months, be 1:1,^{Reference Holt, Gossage-Worrall, Hind, Bradburn, McCrone and Morris3} account for heterogeneity in comorbid substance use, account for patient aspirations, allow for differences between Australia and UK service structures and funding prospects, and include a team-based approach to service delivery.