Fatal Asthma: An Audit of 30 Million Patient-Years of Health Plan Membership from 2007 to 2015

Without accurate data on deaths directly caused by asthma, prevention will be difficult.

To determine how often asthma could be confirmed as a proximal cause of death in a large well-defined population with active health plan membership and no acute barriers to medical care.

All deaths occurring in active Kaiser Permanente Southern California health plan members between 2007 and 2015 were identified. Asthma-coded deaths were manually audited for cause. Health care and asthma medication use in the 6 months before death were determined.

There were 248 (0.80 per 100,000 patient-years) unaudited asthma-coded deaths. There were only 63 (26.5%) (0.20 per 100,000 patient-years) asthma-confirmed deaths. Individuals with asthma-confirmed deaths were younger, less likely to have ever smoked, and had fewer emergency visits in the 6 months before death compared with those with asthma excluded. Individuals with asthma-confirmed deaths used preventative inhalers at very low rates. We unexpectedly found that ever inclusion in the 2016 National Committee for Quality Assurance health effectiveness data and information set (HEDIS) for persistent asthma was associated with a higher risk of all-cause early death. Individuals with asthma-confirmed deaths were also unlikely to be in the HEDIS asthma dataset in the year they died, thus not targeted for outreach.

Audit-confirmed fatal asthma is more likely to occur in younger, nonsmoking, individuals, using very low rates of preventive inhalers. This will be a very difficult group to prospectively identify and manage effectively. Further research into the reasons for early death after HEDIS asthma dataset inclusion is warranted.