For a small proportion of the patients who attend an emergency department, no name can be elicited. Of the 51 628 attendances at the Royal Perth Hospital Emergency Department in the 2001–02 financial year, 84 (0.163%) were made by 78 patients who were recorded as having no name. I hypothesised that unknown people would have a worse outcome (in particular, a higher death rate) than those with a name. I present here a descriptive analysis of a study I conducted to test my hypothesis.

Of the 78 unknown patients who attended the emergency department, 56 (72%) were male (mean age, 32.9 years; range, 2–76 years; median, 31 years). Subsequently, patient names were able to be assigned to 36 (43%) of the 84 attendances.

There were six deaths (7.1%) in the 84 attendances without a name, compared with 840 deaths (1.6%) in the 51 544 attendances with a name. This represents an odds ratio for death in the unknown group of 4.6 (95% CI, 2.0–10.7). (This figure, calculated by Woolf’s method, is based on the number of attendances. An alternative approach is to use the number of patients. Preliminary data for the latter also indicate a statistically significant result.)

Of the six unknown patients who died, five were male (mean age, 38.3 years; range, 25–52 years). Four of the deaths were a result of trauma, and two were from medical causes. Of the six deaths in the unknown group, five occurred in the subgroup who were subsequently identified (effect size, 11.5%; 95% CI, 6.9%–16.1%).

Review of these findings has led me to consider a prospective trial wherein patients attending the emergency department without a name would be randomly allocated to either receive a name or remain nameless. As the data clearly suggest that having a name is protective, an ethics committee will need to consider whether a permanent name change by deed poll is required to maintain this protective effect.

On the other hand, the protective effect of having a name appears to be a “one off” benefit, as patients who are subsequently identified are more likely to die. I recommend that hospitals adopt the evidence-based approach of ensuring that patients remain unknown throughout their hospital stay in order to improve their outcome. Further study is required to determine whether the lives of these patients will be in danger once they resume their identity.