Middle Earth is a dangerous place

The Lancet Neurology last week published Incidence of traumatic brain injury in New Zealand: a population-based study. Prof. Valery Feigin from AUT University, with Dr. Alice Theadom, Dr. Suzanne Barker-Collo, Dr. Nicola Starkey, Prof. Kathryn McPherson, and colleagues, reported on their impressive study that applied a fine sieve to an entire urban and rural catchment population of over 170,000 in the Waikato region of New Zealand for a one year period. This is the most thorough incidence study of traumatic brain injury that has been conducted—the first large scale population-based study covering both urban and rural areas. The study defines a new standard for future research in this area. (It metaphorically but also literally defines a standard—see the paper's Panel 2: Suggested criteria for population-based studies of traumatic brain injury incidence and outcomes, p. 10.)

The case identification methodology is impressive:
”We aimed to assure complete case ascertainment using multiple overlapping sources of information about all cases, admitted and not admitted to hospital, both fatal and non-fatal. This case ascertainment included the following: daily checks of all public hospitals and emergency departments (including surgery and neurosurgery departments) in the study region; monthly checks of CT and MRI records, hospital discharge registers for public and private hospitals in the wider Waikato region, family doctors, rehabilitation centres, and outpatient clinics; quarterly checks of coroner and autopsy records and rest homes; and a yearly check of ambulance services, the prison located within the study region, and the Accident Compensation Corporation (ACC) database. The ACC is a government-supported no- fault insurance agency that funds treatment and rehabilitation for all New Zealand residents with injuries. Cases were also identified from the national death register (we ascertained all death certificates with any mention of TBI). We made every effort to capture data for all individuals with mild TBI who were not admitted to hospital, by including those from family doctor practices providing direct referrals of new and suspected cases of TBI, and by doing checks of accident records of community health services, schools, and sport centres (within and just outside the catchment area), and through self-referrals (the study was widely advertised in the study area via television, newspaper articles, and newsletters and posters). Final checks for complete case ascertainment included reviewing computerised hospital separations data (deaths, discharge, and transfers) for public hospitals with ICD-10 S00-S09 codes for head injury (via the National Health Index number). All TBI cases were checked against existing cases in our TBI registry, to identify any duplicates. Remaining suspected cases (ie, cases for which the presence of TBI was not clear and needed to be verified) of TBI were cross checked with hospital discharge lists, hospital inpatient management records, lists of excluded cases (ie, TBI criteria not met, individuals who did not live in the study area at the time of injury), and lists from other sources (ie, schools, sports groups, rest homes).” (pp. 5-6).

By now it will not surprise you that their evaluation of case information was equally as thorough once participants were identified.

The study identified 1,369 traumatic brain injuries that occurred during the study year, including 71 moderate to severe injuries. This equates to an overall incidence of 790 cases per 100,000 person years. The authors note this was substantially higher than the incidence observed in other high income countries in Europe (47–453 cases), North America (51–618 cases) and is also higher than World Health Organisation estimates. It is possible that there is something different about New Zealand. However, given the rigorous methodology of this study the more likely outcome is that future research will confirm this incidence rate in other high income countries as well. The authors note that regrettably even higher incidence rates again are expected in lower income countries.

There are many ways in which good epidemiological data contributes to health service delivery. Good data guides injury prevention efforts. The kind of partnerships this study describes is an example to us all. Building and maintaining such networks for not just research but clinical purposes is a goal worthy of consideration in itself. Meanwhile, a key question that arises for rehabilitation services is: what is the outcome for the many, many mild (and moderate) injuries that are not being captured into the health system whatsoever, let alone receiving rehabilitation services? Do they spontaneously make a good recovery? How much worse off are they than those who receive services? Given the high numbers of people not accessing services, are there additional population-based interventions we could provide to mitigate at a distance some negative outcomes? On the whole, we don't have good answers to these questions. With this new high water mark set for how many injuries are occurring, the importance of these questions is further underscored.