Johannes C, McQuay L, Tennis P, Camargo CA, Schatz M, Midkiff K, Lanes S, DiSantostefano R, Davis K. Characteristics of initiators of asthma maintenance medications in 10 health care populations. Poster presented at the 29th ICPE International Conference on Pharmacoepidemiology & Therapeutic Risk Management; August 2013. Montreal. [abstract] Pharmacoepidemiol Drug Saf. 2013 Aug; 22(Suppl 1):273-4.

Background: Level of asthma control can be a source of confounding in observational studies of medication-related asthma mortality. As part of a collaborative feasibility study of medication-related asthma mortality, we enumerated use of asthma medications across a distributed network of health care databases and explored variability in asthma control indicators among treatment groups.

Objective: Describe characteristics of patients initiating different asthma maintenance medications and examine variation across 10 data sources.

Methods: In a cohort (N = 994,627) of persistent asthma patients aged 4-100 years enrolled at 10 United States health plans (2001-2010), use of asthma medications was categorized. New use was first use of any study-defined asthma maintenance treatment medication on or after cohort entry without any prior recorded exposure of interest. We tabulated the frequency of three indicators of asthma control in the 6 months before new exposure in each treatment group, in the pooled cohort and across data sources.

Results: A total of 144,574 persons (15%) had new use: 42% aged >= 40 years, 56% female. The three most common treatments were salmeterol+fluticasone propionate (ADVAIR), n = 51,840; inhaled corticosteroid monotherapy (ICS-m), n = 53,798; and ICS+leukotriene receptor agonist (LTRA), n = 26,994. ADVAIR users were older (52% >= 40 years vs. 33% ICS-m and 32% ICS+LTRA); this was consistent across the data sources. Overall, 13% had >= 1 asthma hospital stay or emergency room visit, ranging across data sources by exposure: ADVAIR, 8%-17%; ICS-m, 7%-16%; ICS+LTRA, 9%- 14%. Overall, 33% had &>= 1 oral corticosteroid dispensing—ADVAIR, 27%-42%; ICS-m, 26%-40%; ICS+LTRA, 27%-52%—and 44% had >= 2 short-acting beta-agonist dispensings—ADVAIR, 44%-55%; ICS-m, 49%-66%; ICS+LTRA, 31%-45%.

Conclusions: The observed variation in characteristics of exposure groups across data sources was beyond what would be expected from differences in size and demographics by health plan. Such information will be used in developing adjustment strategies for future pooled analyses of medication-related asthma mortality.