Article type
Year
Abstract
Background: In 1904, the British Medical Journal (BMJ) published "Report on certain enteric fever inoculation statistics" by Karl Pearson1. This paper contains what is often regarded as the first meta-analysis using health care data.
Objectives: To assess the effects of enteric fever inoculation in terms of relative risk, absolute risk difference, and number needed to treat, using data published by Karl Pearson.
Methods: We obtained data from the James Lind Library2 as published in the BMJ1. These data came from soldiers in the British Army treated in five hospitals in South Africa, and one hospital in India. We estimated the relative risk (RR) for mortality as a dichotomous variable and assessed statistical heterogeneity using I23. We used a random effect model for meta-analysis.
Results: We included all six reports, which referred to a total of 13,635 soldiers. Inoculation was associated with a reduced risk of death in patients with enteric fever (relative risk: 0.61; 95% confidence interval (95% CI) 0.5 to 0.74, P < 0.01, I2 = 24.91; absolute risk difference: 7.69; 95% CI = 6.35 to 9.03). The Number Needed to Treat was 13 (95% CI = 11.07 to 15.75).
Conclusions: The data from Professor Pearson's original paper provided encouraging evidence that inoculation against enteric fever could be protective. However, because Pearson was concerned that the correlations he had observed might reflect selection bias, he recommended a controlled trial using alternation to obtain more reliable data. This extended analysis of a meta-analysis published more than a century ago is a little tribute to the memory of Professor Karl Pearson.
References:
1. Pearson K. Report on certain enteric fever inoculation statistics. British Medical Journal 1904;3:1243-1246.
2. www.jameslindlibrary.org/trial_records/20th_Century/1900_1920/pearson/
pearson_kp.html
3. Higgins JP, Thompson SG: Quantifying heterogeneity in a meta-analysis. Stat Med 2002, 21: 1539-1558.
Objectives: To assess the effects of enteric fever inoculation in terms of relative risk, absolute risk difference, and number needed to treat, using data published by Karl Pearson.
Methods: We obtained data from the James Lind Library2 as published in the BMJ1. These data came from soldiers in the British Army treated in five hospitals in South Africa, and one hospital in India. We estimated the relative risk (RR) for mortality as a dichotomous variable and assessed statistical heterogeneity using I23. We used a random effect model for meta-analysis.
Results: We included all six reports, which referred to a total of 13,635 soldiers. Inoculation was associated with a reduced risk of death in patients with enteric fever (relative risk: 0.61; 95% confidence interval (95% CI) 0.5 to 0.74, P < 0.01, I2 = 24.91; absolute risk difference: 7.69; 95% CI = 6.35 to 9.03). The Number Needed to Treat was 13 (95% CI = 11.07 to 15.75).
Conclusions: The data from Professor Pearson's original paper provided encouraging evidence that inoculation against enteric fever could be protective. However, because Pearson was concerned that the correlations he had observed might reflect selection bias, he recommended a controlled trial using alternation to obtain more reliable data. This extended analysis of a meta-analysis published more than a century ago is a little tribute to the memory of Professor Karl Pearson.
References:
1. Pearson K. Report on certain enteric fever inoculation statistics. British Medical Journal 1904;3:1243-1246.
2. www.jameslindlibrary.org/trial_records/20th_Century/1900_1920/pearson/
pearson_kp.html
3. Higgins JP, Thompson SG: Quantifying heterogeneity in a meta-analysis. Stat Med 2002, 21: 1539-1558.