Aguado J, Martinez D, Plana E, Ziemiecki R, Rebordosa C. Reco(r)ding of lifestyle variables in CPRD Aurum compared with GOLD. Poster presented at the 2020 36th ICPE International Virtual Conference on Pharmacoepidemiology & Therapeutic Risk Management; September 16, 2020.

BACKGROUND: In England, many primary care practices migrated from VISION to EMIS software. In 2017, Clinical Practice Research Datalink launched Aurum, incorporating some of these migrating practices from GOLD and new practices using EMIS. No publications to date exist evaluating data migration in Clinical Practice Research Datalink Aurum.

OBJECTIVES: To compare the recording of lifestyle variables in GOLD and Aurum before and after migration and check the validity of the new algorithm in Aurum.

METHODS: Among practices migrating from GOLD to Aurum, 7 were randomly selected, and adult patients registered in the practice at least 1 year before the last collection date for GOLD (migration) were included. Smoking status and alcohol consumption were defined in GOLD through entity types and codes, and the algorithm was adapted to Aurum using data up to 10 years before migration. Smoking status was defined as current, former, never, or missing. Alcohol consumption was defined as never, former, low to moderate (≤ 6 units/day), heavy (≥ 7 units/day) drinker, drinker unknown amount, or missing. Body mass index is under study.

RESULTS: A total of 40,196 adults in GOLD and 40,706 in Aurum were included in this study. Median age was 49 years, and approximately 50% were females in both data sources. Smoking status was recorded for 95.4% of patients in GOLD versus 95.6% in Aurum. The distribution of smoking status in GOLD versus Aurum was 19.3% versus 20.4% for current, 28.7% versus 28.1% for former, and 52.0% versus 51.5% for never-smokers. Alcohol consumption was recorded for 67.9% in GOLD versus 68.4% in Aurum. The distribution of alcohol consumption in GOLD versus Aurum was 19.8% versus 21.4% for never-drinkers, 3.2% versus 3.3% for former drinkers, 68.9% versus 66.9% low-to-moderate drinkers, 3.5% versus 3.6% heavy drinkers, and 4.6% versus 4.7% drinkers (unknown amount).

CONCLUSIONS: Using the GOLD and Aurum adapted algorithms, distribution of lifestyle variables in GOLD and Aurum were very similar and in line with population distribution in the United Kingdom.

Share on: