The DUP dilemma

In his accompanying commentary, Warner argues that, essentially, the ‘central tenet’ upon which EIS are based (namely, that DUP is the important driver of outcome and that it should be reduced) is fundamentally flawed. ^{Reference Warner2} Rather, he argues, it is mode of onset (acute v. insidious) that is the important issue (and which in turn influences DUP). Certainly, we know that a more insidious onset of illness is associated with poorer outcome and intuitively, it makes sense that someone with a more acute, florid presentation is likely to come to the attention of services (mental health or criminal justice system) more quickly (and thus have a shorter DUP) than those with a slower, less florid initial onset. Indeed, the ÆSOP study demonstrated that a more insidious onset was associated with a longer DUP. ^{Reference Morgan, Abdul-AI, Lappin, Jones, Fearon and Leese3} However, this does not necessarily mean that attempts to reduce DUP are in vain. Some would argue that, at the very least, any attempt to reduce a period of suffering (especially one as distressing as a first episode of psychosis) can only be a good thing, regardless of whether it changes prognosis. The more difficult question is whether, in our increasingly resource-constrained services, this is the best use of limited resources and staff. The truth is that we still do not know with sufficient certainty whether length of DUP centrally affects illness outcome or whether it is a mere indicator of poor prognosis illness. We do know that certain aspects of early illness course are ameliorated by EIS ^{Reference Craig, Garety, Power, Rahaman, Colbert and Fornells-Ambrojo4} compared with generic services, but equally, it is unclear whether, once people migrate to generic services after EIS, their progress is maintained. ^{Reference Marshall and Rathbone5} One interpretation of this is that it may not be any particular novel intervention that makes this early difference, but rather the quality and intensity of the service and support delivered. Indeed, EIS may be merely a blueprint for what all mental health services (given sufficient funding) should strive for.

Pathways

Ghali et al's findings in terms of differences in DUP and pathways to care in EIS largely mirror the findings of the original ÆSOP study (performed in parts of the same geographical location in the previous decade, prior to the advent of EIS). ^{Reference Morgan, Fearon, Hutchinson, McKenzie, Lappin and Abdul-AI6,Reference Morgan, Mallett, Hutchinson, Bagalkote, Morgan and Fearon7} In brief, they found shorter DUP for certain Black and minority ethnic (BME) groups and also higher rates of referral via primary care for White British people and higher rates via the criminal justice system for some BME groups. On this latter point, two issues are worth noting. First, that over a 10-year period, and one which has seen the initiation of EIS, little appears to have yet changed in terms of patterns of referral. And second, higher rates of referral via the criminal justice system persist for some groups. It is worth noting that EIS are still relatively young and are developing rapidly, and that changes in practice can take some time to be reflected in changes in outcome. That notwithstanding, these findings do pose a question as to whether EIS need to further develop, or even develop in a broader direction, in order to be more effective. For example, in some countries (most notably Australia but increasingly others such as Ireland), much closer links are being forged between statutory services and voluntary agencies that cater for younger adults with a broad range of mental health distress. Further, given that adolescents and younger adults may present with non-specific ‘symptoms’ that do not accurately predict one specific disorder, should EIS broaden their remit beyond psychosis? Put plainly, should EIS be less diagnosis specific? Should we be thinking more in terms of duration of untreated illness or distress rather than duration of untreated psychosis?

Strikingly, Ghali et al's paper shows continuing high rates of referral via the criminal justice system. About one-quarter of the overall sample was identified via this route, a proportion that was even higher for some BME groups. At first glance, this appears troubling. Indeed, many use the somewhat pejorative terms ‘coercive’ or ‘adverse’ as a synonym for the criminal justice system pathway. Ideally, everyone with early signs of illness would see their primary care physician who would then promptly refer them to their local EIS, who would then promptly assess them and treat them as appropriate. I would argue, however, that it is somewhat reassuring to see credible rates of referral via the criminal justice system (be they the police, the probation or the prison services). At least this suggests that these services appear able to detect and identify certain behaviours as indicators of illness and make the appropriate referrals. I would be more concerned if these rates were low, possibly suggesting that services were failing to detect genuinely unwell individuals.

Like ÆSOP, this study found that certain BME groups had higher rates of referral via the criminal justice system. According to Warner's thesis, this may be due to more acute, florid illnesses perhaps being more common in these groups leading to their being identified earlier and on the basis of behaviour that attracts the attention of the police and emergency services. ^{Reference Warner2} Intriguingly (and Ghali et al refer to this in their paper), the causes for this phenomenon may be more complex that at first glance. For example, in the ÆSOP study, ^{Reference Morgan, Mallett, Hutchinson, Bagalkote, Morgan and Fearon7} it was found that Black Caribbean family members were more likely than other groups to call the police rather than mental health services if they noticed unusual or disturbing behaviour in another family member.

Perhaps the most intriguing finding is that referral via primary care was associated with a longer DUP than other routes, a finding particular true for the White British group. Ghali et al speculate as to why this may be the case. Are individuals who are White British more likely to have a more insidious, less florid illness onset that is less likely to lead to a quick referral? Are general practitioners more likely to treat and/or observe such cases for longer prior to ultimately referring them on? Or are they not yet sufficiently quick at detecting the early signs of illness? The data in this paper do not allow us to arrive at a solid conclusion, but certainly this is an area that will need to be examined further, since the EIS model would argue that the earlier a referral is made to them, the better.