Optimizing influenza vaccine allocation: A predictive analytics approach for informed public health planning.

Disclaimer: In an effort to expedite the publication of articles, AJHP is posting manuscripts online as soon as possible after acceptance. Accepted manuscripts have been peer-reviewed and copyedited, but are posted online before technical formatting and author proofing. These manuscripts are not the final version of record and will be replaced with the final article (formatted per AJHP style and proofed by the authors) at a later time.

Purpose: Excessive purchasing of influenza vaccine can lead to costly overages and waste of resources. Insufficient quantities, however, can jeopardize population health. Our project aimed to use predictive analytics to determine the influenza vaccine quantities that would be needed for the next influenza season while minimizing vaccine waste and meeting patient care demands.

Methods: Several data sources were evaluated to develop a predictive analytics model to better estimate future influenza vaccine orders during upcoming influenza seasons. A retrospective analysis of influenza vaccine administrations over the last 4 influenza seasons allowed the team to develop an algorithm to predict influenza vaccine needs. Two regions within Mayo Clinic were selected to determine the validity of our ordering process. These 2 regions, identified as regions 3 and 5, ordered influenza vaccines based on the algorithm, while the other 3 regions acted as control groups, ordering though traditional methods based on purchasing data.

Results: Predictive analysis for the 2 intervention regions resulted in a savings of over $1 million when compared to traditional ordering methods. The model predicted that the quantity of vaccine ordered should be 17,574.16 and 9,164.29 quadrivalent influenza vaccines for regions 3 and 5, respectively. On the basis of actual administration data, 15,902 vaccines for region 3 and 9,016 vaccines for region 5 will be administered by the end of the season, both of which are less than the predicted amount needed, demonstrating the accuracy of the analytics.

Conclusion: Compared to the traditional ordering method, ordering using predictive analytics allowed the team to more accurately determine future order volumes and spend, yielding significant cost savings.