Clinical Biostatistics: Geometric mean length of hospital stay exercise

This website is for students following the M.Sc. in Evidence Based Practice at the University of York.

The following is the abstract of a research paper (Ng et al. 2002).

Objectives: To evaluate the impact of early abdominopelvic computed tomography in patients with acute abdominal pain of unknown cause on length of hospital stay and accuracy of diagnosis.

Design: Randomised, prospective controlled trial.

Setting: Teaching hospital in England.

Participants: 120 patients admitted with acute abdominal pain for which no immediate surgical intervention or computed tomography was indicated.

Intervention: 55 participants were prospectively randomised to early computed tomography (within 24 hours of admission) and 65 to standard practice (radiological investigations as indicated).

Main outcome measures: Length of hospital stay, accuracy of diagnosis, and, owing to a possible effect on inpatient mortality, deaths during the study.

Results: Early computed tomography reduced the length of hospital stay by 1.1 days (geometric mean 5.3 days (range 1 to 31) v 6.4 days (1 to 60)), but the difference was non-significant (95% confidence interval, 8% shorter stay to 56% longer stay, P=0.17). Early computed tomography missed significantly fewer serious diagnoses. Seven inpatients in the standard practice arm died. Only 50% (59 of 118) of diagnoses on admission were correct at follow up at 6 months, but this improved to 76% (90) of diagnoses after 24 hours.

Conclusions: Early abdominopelvic computed tomography for acute abdominal pain may reduce mortality and length of hospital stay. It can also identify unforeseen conditions and potentially serious complications.

Questions

1. What can we deduce about the shape of the distribution of hospital stay?

2. What is a geometric mean and how was it calculated?

3. What does the 95% confidence interval given as “8% shorter stay to 56% longer stay” mean? What is surprising about it and why might it be wrong?

4. How could the 95% confidence interval given as “8% shorter stay to 56% longer stay” have been calculated?

5. Under Results, the authors say that “Early computed tomography reduced the length of hospital stay by 1.1 days . . . but the difference was non-significant . . . ”. What do you think of this statement?

6. The authors first conclusion is that “Early abdominopelvic computed tomography for acute abdominal pain may reduce mortality and length of hospital stay.” Is this a useful conclusion to be drawn from a clinical trial?

7. In the body of the paper, we read that “seven inpatients in the standard practice arm died during the study (0% (0 of 55) early computed tomography v 11% (7 of 63) standard practice, P=0.014)”. In the discussion we read “Caution is needed in generalising our results as mortality was not an a priori end point of our study.” What would you conclude from this?

Reference

Ng CS, Watson CJE, Palmer CR, See TC, Beharry NA, Housden BA, Bradley JA, Dixon AK. (2002) Evaluation of early abdominopelvic computed tomography in patients with acute abdominal pain of unknown cause: prospective randomised study. British Medical Journal 325, 1387.

Back to Clinical Biostatistics index.