Multiple Myeloma Vaccination Patterns from a Large Health System: An Example of Cancer Care Delivery Research (CCDR)

Background

We have a poor understanding of the vaccination immune response and outcomes in multiple myeloma (MM). As MM patients (Pts) are living longer and therapies are immunomodulatory there is an unmet medical need to further characterize the role of the immune system. A common reason for hospitalization or death in MM Pts is infection. As an initial step in MM Cancer Care Delivery Research (CCDR), we evaluated the current vaccination practice patterns in MM Pts at Aurora Health Care using the EMR and data analytics.

Methods

An IRB approved study reviewed MM Pts from 5/15/2012 to 5/15/2014. Data collected included demographics, influenza (FV) and pneumonia vaccination (PV) history, hospitalization episodes, cost associated with hospitalization, and admission and discharge diagnoses. Pts were considered PV positive if vaccinated within 5 years prior to study with any PV type. FV was none (no FV in 2012-2014), optimal [FV in 2012 and (2013 or 2014)], or suboptimal [FV in 2012 or (2013 or 2014)]. Data was analyzed using SAS and STATA 12.

Results

A total of 1131 MM Pts were identified. Race included 70% white, 13% black, and 17% mixed, other or information not available. MM median age at diagnosis was 71 and only 4% (47) had prior autologous stem cell transplantation. PV rate was 30%. FV was 55% none, 24% suboptimal and 20% optimal. There was no statistically significant difference in the rate of PV and FV when stratified vs age, gender, and race. Over two years there were a total of 662 hospitalization events involving 317 MM Pts. The total cost of hospitalization was approx $35M. The average charge per hospitalized patient was $110K (range: $2K -1.3M) with an average $52K per hospitalization encounter (range: 2K – 648K). The rate of PV and FV vaccination among Pts with index hospitalization is significantly higher than non-hospitalized patients. There was no difference in hospitalization cost based on vaccination status. (See Table 1)

Discussion

Vaccination rates were low and did not correlate with hospital outcomes. This may be explained as a limitation for a retrospective EMR analysis without accounting for temporal relationship of vaccines – i.e. possible vaccination after admission. Alternatively, this may indicate that our current methods of vaccination in MM are not effective. Other limitations include need for a more granular review of treatment regimens and infectious complications. Additional surrogate markers are needed to understand the effect of vaccines and the immune system on health care outcomes such as hospitalizations, cost, and survival. This will be addressed in prospective registries and immunologic studies at our center and may be queried at other health systems.