Generations of junior medical staff had internally rolled their eyes when I voiced my theory about conjunctivitis meaning that the patient would die soon, then been astounded by my prognostic skills when the patient deteriorated into the terminal phase. It was time to put my reputation where my mouth was and do a prospective audit. Did diagnosis and treatment of conjunctivitis in the inpatient palliative care setting mean that the patient had a very poor prognosis?

I carried a prospective observational audit over 15 months.

Accurately predicting when a patient will die is important, as it has implications for the patient and their family and for decision making.

The Wantirna Health palliative care unit is a 30 bed stand-alone facility in Melbourne, Australia; part of Eastern Health. It admits around 800 patients with life-limiting illnesses a year, mainly for specialist symptom management and end of life care.

Complete records were found in 38 patients who developed conjunctivitis – 6 additional patients were excluded for incomplete data.

Thirty-four of the 38 inpatients (90%) died after developing conjunctivitis and 4 were discharged. The mortality rate of 90% was statistically significantly different from the average mortality rate of 65% for the unit (p=0.002)

Of those who died, 24 (71%) developed conjunctivitis before being commenced on the LCP. Ten patients (29%) developed conjunctivitis on the same day or after being started on the LCP.

The length-of-stay of the patients who died ranged between 1-53 days, with an average of 15.1 days, compared with the average for all patients during the study period of 13 days.

The clinical phase of the patients at midnight on the day they developed conjunctivitis was: stable for 3 patients, unstable in 4, deteriorating in 20, and terminal in 7. Therefore on the day conjunctivitis was first observed, only 7 of 34 (21%) of the patients who died were in the terminal phase.

So in more than two thirds of cases, the conjunctivitis occurred before the staff perceived the patient was actively dying, as marked by the event of the patient being cared for on the LCP or being recorded as being in the “terminal” phase.

The average length of treatment for conjunctivitis was 5.1 days and the median was 3 days; time on LCP was average 2.3 days and median 2 days.

This prospective observational audit of palliative inpatients being treated for conjunctivitis showed that conjunctivitis is a poor prognostic factor and in two thirds of cases, preceded other indicators to staff that the patient was actively dying.

Ninety percent of palliative inpatients with conjunctivitis died, compared to an average annual death rate of 65% (p=0.002).

Further study of the reliability of this sign is warranted, as this is interesting and potentially useful for staff in prognostication and decision-making.

Have you observed this too? What do you think of “Sonia’s sign”? Let us know!

For those getting started in research and publication, this little project was great because it was an area of particular interest to me personally, and I think to others as prognosticating is such an important issue. It was easy to carry out and did not require any funding or equipment. It was published in the Journal of Palliative Medicine as a letter – I submitted it as a brief report but they asked me to revise and resubmit.

Dr Sonia Fullerton
Peter MacCallum Director of Pain and Palliative Care
MBBS FRACP FAChPM
Sonia.fullerton@petermac.org

@sonialf