Six questions nobody has answered

Each of these was searched for deliberately, across more than one strand, and found not to exist. They are absences in the literature rather than gaps in this review, and each is a fundable question. A corpus that can show what is missing is more useful to a research partner than one that can only show what is there.

In preparation For each gap, the study design that would close it, the data assets it would need, the likely partners and an order-of-magnitude cost. That is professional judgement rather than a finding of the review, so it is being written by Stan Gilmour and Huw Williams and will appear here with the report.

No United Kingdom study links health records to police, court, prison or probation records to examine a justice outcome after brain injury

Gap 1

The linkage infrastructure exists. Scottish linkage has used the prison census as a population frame for hospitalised head injury prevalence, but not to follow offending, reoffending or any other justice outcome, and the Welsh and English assets have not been turned on this question. Ontario ICES and the New Zealand Integrated Data Infrastructure have each produced the study the United Kingdom has not.

Why it matters. Everything downstream of prevalence depends on it: who is affected, where they present first, and what happens to them afterwards.

No economic evaluation, and no formal feasibility study, of brain injury screening in any justice setting

Gap 2

The figures in circulation, including a United Kingdom cost of £43 billion a year and a 16:1 return on rehabilitation, are modelled assumptions from grey literature rather than measured savings. No screening programme in a justice setting has been costed against what it finds, and published accounts of screening programmes are descriptive implementation reports.

Why it matters. It is the first question a commissioner asks, and the field cannot currently answer it.

No United Kingdom study follows mortality in justice-involved people with brain injury

Gap 3

Both halves of the study exist separately in Scotland. Graham and colleagues linked 76,627 first-time prisoners to death records and found an age-standardised mortality rate ratio of 3.3 in men and 7.6 in women. McMillan and colleagues linked hospitalised head injury to mortality in homeless people and found a standardised mortality ratio more than twice that of those without. Neither crosses into the other population, and no study has linked brain injury to mortality in a justice cohort here. Mortality evidence for that group comes from United States trauma registries and a Western Australian linked cohort.

Why it matters. Premature death is the outcome that settles arguments about whether this is a health problem.

No United Kingdom sibling or twin fixed-effects study

Gap 4

The design that has moved the causal argument everywhere else has not been run here. The Danish cohort of 343,027 that collapsed the association, and the Swedish and Western Australian sibling comparisons that did not, are all from other jurisdictions.

Why it matters. The causal question is the one on which policy either rests or does not, and the United Kingdom has no domestic evidence on it.

No qualitative metasynthesis, and no ethnography with brain injury as its lens

Gap 5

Single interview studies exist and recur on the same themes. Nobody has synthesised them, and no observational study of a custody suite or a courtroom has taken cognitive impairment as its explicit frame.

Why it matters. The quantitative strands establish that identification fails. Only qualitative work explains how it fails in the room.

No system model of the justice pathway, and no needs assessment method that accounts for brain injury

Gap 6

Both tools exist and neither has been pointed at this problem. System dynamics modelling has been applied to traumatic brain injury, but to clinical care delivery rather than to a justice pathway, and its authors note it depends on data adequate to build it. Public Health England issued a health needs assessment toolkit for prisons, young offender institutions, immigration removal centres and police custody, and a text search of it found no mention of brain injury, head injury or acquired brain injury. No small-area estimation method exists for a commissioner wanting a local prevalence figure.

Why it matters. Without it, every local service is planned from a national average with a confidence interval sixteen points wide.

Working with us on any of these Three of the six sit inside existing infrastructure at the University of Exeter. Two need a data controller willing to link health and justice records. If you hold one of those assets, or fund work of this kind, get in touch.