H G Peach1, N E Bath. 1. University of Melbourne, Ballarat Health Services Base Hospital, Victoria, Australia.
Abstract
BACKGROUND: Raised plasma ferritin concentrations occur unexpectedly during iron studies done by primary care physicians. Plasma ferritin concentration has been positively associated with alcohol use among men. AIM: To determine the post-test probability that men in the community with raised plasma ferritin concentrations are hazardous drinkers. METHODS: The subjects were 152 men, randomly selected from a city's electoral roll. Nineteen (12.5 (2.7)%, mean (SEM)) admitted to drinking hazardously. The pretest probability of a man being a hazardous drinker was 0.125. This was converted to pretest odds of 0.14. The likelihood ratio (the ratio of the probability of obtaining a raised plasma ferritin concentration in a hazardous drinker (sensitivity) to the probability of obtaining a raised plasma ferritin concentration in a non-hazardous drinker (1-specificity)) was calculated for different plasma ferritin cut off points. RESULTS: A plasma ferritin level of > 652 micrograms/l gave the largest likelihood ratio, 4.16. Post-test odds were obtained by multiplying the pretest odds (0.14) by the likelihood ratio (4.16). A plasma ferritin level of > 652 micrograms/l had a post-test odds for a man being a hazardous drinker of 0.58. This was converted to a post-test probability of 0.37. CONCLUSIONS: Inquiries could usefully be made into the alcohol consumption of men with a plasma ferritin concentration > 652 micrograms/l, as approximately one in three would admit to drinking hazardously.
BACKGROUND: Raised plasma ferritin concentrations occur unexpectedly during iron studies done by primary care physicians. Plasma ferritin concentration has been positively associated with alcohol use among men. AIM: To determine the post-test probability that men in the community with raised plasma ferritin concentrations are hazardous drinkers. METHODS: The subjects were 152 men, randomly selected from a city's electoral roll. Nineteen (12.5 (2.7)%, mean (SEM)) admitted to drinking hazardously. The pretest probability of a man being a hazardous drinker was 0.125. This was converted to pretest odds of 0.14. The likelihood ratio (the ratio of the probability of obtaining a raised plasma ferritin concentration in a hazardous drinker (sensitivity) to the probability of obtaining a raised plasma ferritin concentration in a non-hazardous drinker (1-specificity)) was calculated for different plasma ferritin cut off points. RESULTS: A plasma ferritin level of > 652 micrograms/l gave the largest likelihood ratio, 4.16. Post-test odds were obtained by multiplying the pretest odds (0.14) by the likelihood ratio (4.16). A plasma ferritin level of > 652 micrograms/l had a post-test odds for a man being a hazardous drinker of 0.58. This was converted to a post-test probability of 0.37. CONCLUSIONS: Inquiries could usefully be made into the alcohol consumption of men with a plasma ferritin concentration > 652 micrograms/l, as approximately one in three would admit to drinking hazardously.