Is chronic obstructive pulmonary disease, diabetes, and life-style variables reco(r)ded similarly in CPRD GOLD and Aurum?

Date of Approval: 
2020-08-21 00:00:00
Lay Summary: 
In England, many primary care practices changed the software which general practitioner (GP) uses to record patient‘s data. In 2017, Clinical Practice Research Datalink (CPRD) launched Aurum, a dataset which incorporates some of these migrating practices and some more new practices using the new software. Studies using Aurum data are ongoing and variables definitions in these studies need to be adapted from the existing ones which were used for the former GOLD dataset, including practices using old software. Our first objective is to evaluate migration of historical data from GOLD to Aurum by comparing the prevalence of two common diseases treated by the GP (chronic obstructive pulmonary disease (COPD) and diabetes), the use of medications to treat them, as well as clinical markers of these diseases. Additionally, we will assess patient’s lifestyle through body mass index (BMI), smoking status, and alcohol consumption. Our second objective is to assess the recording of new data in Aurum 1 year before and 1 year after the practice switched the software by comparing during those time periods the prevalence of COPD and diabetes diagnosis, the use of medications to treat those conditions, and clinical markers of these diseases.
Technical Summary: 
In England, many primary care practices migrated from VISION to EMIS software. In 2017, Clinical Practice Research Datalink (CPRD) launched Aurum, incorporating some of these migrating practices from GOLD and new practices using EMIS. Studies using Aurum data are ongoing and code lists as well as algorithms to define variables in the studies are adapted from the GOLD experience. Our objectives are to evaluate: - Data migration of historical data from GOLD to Aurum by assessing: o The prevalence of chronic obstructive pulmonary disease (COPD) diagnosis, medications, and distribution of FEV1 predicted. o The prevalence of Diabetes diagnosis, medications, and distribution of Hba1c. o The distribution of BMI, smoking status, and alcohol consumption. o Number of hospitalizations or emergency department visits - New recording of COPD data in Aurum by comparing data within 1 year before the migration date and 1 year after the first collection date on: o The prevalence of chronic obstructive pulmonary disease (COPD) diagnosis, medications, and distribution of FEV1 predicted. o The prevalence of Diabetes diagnosis, medications, and distribution of Hba1c.
Health Outcomes to be Measured: 
COPD and diabetes will be defined using READ codes and SNOMED codes in GOLD and Aurum, respectively, using data any time before date of migration from Vision to EMIS (migration date), and in Aurum both 1 year before the migration date and 1 year after the first collection date. Three life-style variables will be evaluated: - BMI within 3 years prior and closest to the migration date will be reported in categories: Underweight, Normal, Overweight, Obese, Missing. In GOLD, BMI (kg/m2) as a value will be derived from entity types for BMI, height, and weight. In Aurum, values corresponding to BMI, height or weight codes will be used to derive the BMI value. Once calculated, the BMI value will be categorized as <20 (Underweight), 20-25 (Normal), 25-30 (Overweight), and >30 (Obese). Both in GOLD and Aurum, clinical codes will be used to build the variable BMI in categories. - Smoking status within 10 years prior and closest to the migration date will be reported in categories: Never smoker, former smoker, current smoker, and missing. In GOLD, entity types will be used to define the patient’s smoking status. Both in GOLD and Aurum, clinical codes will be used to build this variable. - Alcohol consumption within 10 years prior and closest to the migration date will be reported in categories: Never drinker, former drinker, current drinker (divided in low or moderate (≤ 6 units/day), heavy (≥ 7 units/day), or unknown amount consumption), and missing. For current drinkers, amount of alcohol consumption will be derived from entity type in GOLD and values and units associated with alcohol codes in Aurum. Both in GOLD and Aurum, clinical codes will be used to categorize this variable. - Number of hospitalizations or emergency department visits recorded in both GOLD and Aurum
Application Number: 
20_183
Collaborators: 

Estel Plana Hortoneda - Chief Investigator - RTI Health Solutions
Estel Plana Hortoneda - Corresponding Applicant - RTI Health Solutions
CRISTINA REBORDOSA GARCIA - Collaborator - RTI Health Solutions
David Martinez - Collaborator - RTI Health Solutions
Jaume Aguado - Collaborator - RTI Health Solutions
Ryan Ziemiecki - Collaborator - RTI Health Solutions