Discussion

This study reports on the duration of IPS support received and the reach, effectiveness and equity in access to, and outcomes from, IPS in routine clinical practice. The median time to gaining employment was 4.3 months, and most people did not require time-unlimited in-work support. They received in-work support for a median of 4.8 months. Those who remained unemployed spent a significantly shorter time on the IPS case-load. We also found that the IPS service reached almost 9% of the working-age population who are using adult community mental health teams in a year. Those who received IPS were younger, but representative compared with the adult community mental health teams case-load in terms of gender and diagnosis. Positive outcomes were achieved for a total of 46.2% of people who either gained or retained employment, or were supported to achieve their educational goals.

The IPS evidence-based principle of rapid job search can sometimes cause concern for mental health professionals, who may interpret this to mean people move rapidly into jobs; however, its actual meaning is that the job search process begins rapidly, within 30 days after first contact with IPS. This principle can also be an operational challenge – demand and capacity – as an IPS employment specialist's case-load is limited to a maximum of 20 clients in adherence to the highest possible score for IPS fidelity. Although the median time from first IPS contact to gaining employment was 132 days (4.3 months), by 3 months, 33.5% had gained employment, and by 9 months, most people (84.2%) had gained employment. Our findings are consistent with the first European IPS RCT, where employment outcomes were predominantly achieved within the first 9 months,^{Reference Burns, Catty, Becker, Drake, Fioritti and Knapp17} and then replicated in the ‘IPS-Lite’ RCT in England that restricted job-seeking support to 9 months without significantly altering the effectiveness of IPS.^{Reference Burns, Yeeles, Langford, Montes, Burgess and Anderson12}

Gaining and maintaining employment involves both social adaptation and psychological adjustment, and the evidence-based principle of time-unlimited in-work support is important to ensure people manage the transition into work and maintain employment.^{Reference Rinaldi and Perkins18} However, this principle also poses an operational challenge: how much support is needed and for how long? Although 84.7% of people who gained employment did not require time-unlimited in-work support, the majority of those (61%) required support for up to 6 months, with a further 20.9% requiring support for up to 9 months. Research suggests ongoing support promotes job tenure,^{Reference Bond and Kukla19} but there are complex interacting individual, employment specialist and job-related factors that predict job tenure.^{Reference Corbière, Villotti, Berbiche and Lecomte20} Even so, our findings suggest people need longer periods of in-work support than found in the ‘IPS-Lite’ RCT, which limited in-work support to 4 months.^{Reference Burns, Yeeles, Langford, Montes, Burgess and Anderson12}

Our findings align with the existing literature, in that the goal of gaining employment is both realistic and achievable for people using mental health services. Both the Royal College of Psychiatrists²¹ and the World Psychiatric Association^{Reference Heron, Kawakami, Thirumalai, Morganstein, Jetly and Glozier22} recommend psychiatrists routinely explore a person's current employment status and consider appropriate work as a treatment outcome for any care provided as this can positively contribute to an individual's recovery. We found when IPS employment specialists are integrated into community mental health teams and teams addressed people's recovery goal of gaining employment, then 44.8% of people diagnosed with schizophrenia/psychosis and half of people (49.9%) diagnosed with a personality disorder gained employment. These findings are consistent with the IPS RCT literature showing people can gain and maintain employment regardless of diagnostic, clinical, functional and personal characteristics.^{Reference De Winter, Couwenbergh, Van Weeghel, Sanches, Michon and Bond23}

Interestingly, this study found people from Black/Black British, Asian/Asian British, mixed and other ethnic group backgrounds were significantly overrepresented in accessing IPS, whereas people from White backgrounds were underrepresented. Although not significant, these findings followed through to employment outcomes, where 47.8% of people from mixed ethnic groups, 44.2% of Asian/Asian British people and 43.4% of Black/Black British people gained employment, whereas the lowest employment outcomes achieved, although very positive, were for people from a White ethnic background (40.2%). These findings are consistent with a previous study,^{Reference Perkins, Patel, Willett, Chisholm and Rinaldi24} and now provide evidence from across ten London boroughs that people from different ethnic groups appear not to be differentially disadvantaged in relation to either access to, or outcomes of, IPS services. These findings also align with the international IPS literature on ethnicity, and intersectional characteristics and outcomes.^{Reference Bond, Mascayano, Metcalfe, Riley and Drake25,Reference Witte, Strandberg, Granberg and Gustafsson26}

Although the primary outcome of IPS is helping people to gain employment, there are other important outcomes. Research shows people often come into contact with mental health services and are in work, but typically lose their jobs within a short period of time and do not return to work.^{Reference Rinaldi, Killackey, Smith, Shepherd, Singh and Craig27,Reference Hakulinen, Elovainio, Arffman, Lumme, Pirkola and Keskimäki28} Even after a short period of unemployment (3 months or more), people with mental health conditions are already less likely to return to work compared with people without mental health conditions, and they are particularly sensitive to the negative effects of unemployment and the loss of structure, purpose and identity.^{29,Reference Rinaldi and Perkins18} Additionally, enabling people to complete their education is a determinant of health, is associated with better employment prospects and people diagnosed with schizophrenia typically have poor educational outcomes.^{Reference Dalsgaard, McGrath, Østergaard, Wray, Pedersen and Mortensen30,Reference Crossley, Alliende, Czepielewski, Aceituno, Castañeda and Diaz31} It is therefore important for IPS services to monitor and report these outcomes as part of service performance, recognising the clinically significant outcome of providing employment support to facilitate job retention or educational support to enable a person to complete their education and make the transition into the labour market.³²

This study has several limitations. First, it is a naturalistic design and cannot attribute causality to the findings; however, we believe this limitation is also a strength, providing real-world evidence on the duration of IPS support needed and the effectiveness of IPS in routine clinical practice. Second, the findings for time-unlimited in-work support provide evidence that people who gained jobs were also maintaining them, but details of the types of jobs, hours worked, duration of employment and their experiences were unavailable. Although a previous evaluation found people worked in both a wide range of occupations and occupational levels,^{Reference Rinaldi, Perkins, Hardisty, Glynn and Souza33} such details could have provided greater insight into whether people in this study were engaged in meaningful work. Third, employment specialist turnover affects service performance, productivity and the continuity of employment support provided within IPS services.^{Reference Butenko, Rinaldi, Brinchmann, Killackey, Johnsen and Mykletun34,Reference Brichmann, Wittlund, Lorentzen, Moe, McDaid and Killackey35} During the study period, there was a level of employment specialist turnover, and we did not have access to the vacancy rate, which would have enabled us to reliably calculate the average employment specialist case-load size, and understand whether the reach of 8.7% of the overall clinical case-load was good or not. However, the reach found is greater than the 2% found in the USA.^{Reference Bruns, Kerns, Pullmann, Hensley, Lutterman and Hoagwood36, Reference Twamley, Baker, Norman, Pittman, Lohr and Resnick37}

Since 2018, IPS services have been scaled up to be a core intervention within the transformation of community mental health services in England.³⁸ The initial focus was to establish or expand existing IPS services to increase the capacity and quality (fidelity) of services and deliver against NHS England access targets. However, at an NHS Trust or Integrated Care Board level, questions arise in terms of the planning and performance of these services. While achieving equity in outcomes, the employment outcomes achieved in this study are consistent with the 44% outcomes found across European IPS RCTs,^{Reference Drake, Becker and Bond39} and suggest there is little to no research to practice gap. However, during the study period, the healthcare quality regulator, the Care Quality Commission,⁴⁰ found only a quarter of people (26%) using mental health services in England reported they had ‘definitely’ received help and advice with finding work, and half (50%) did not receive help or advice but would have liked it. This suggests an unmet need for employment support for people using mental health services. Although there may not be an IPS research to practice gap, there might be a gap between those who want and need employment support, and those who have access to it.