Health Behavior, Health Promotion, Influenza, Risk Reduction, Vaccination
Conditions
Keywords
Flu Vaccine, Choice Architecture, Machine Learning, Perceived Credibility
Brief summary
This study will test the relative efficacy of high-risk messages in increasing flu shot rates in patients at moderately high risk for flu and complications (those in the top 11-20% of risk). It will also examine whether informing patients that their high-risk status was determined by analyzing their medical records or by an artificial intelligence (AI) / machine-learning (ML) algorithm analyzing their medical records will affect the likelihood of receiving a flu vaccine.
Detailed description
Almost everyone age 6 months or older can benefit from the vaccine, which can reduce illnesses, missed work, hospitalizations, and death by reducing the likelihood of contracting influenza. Flu shots are particularly important for patients at high risk of experiencing severe outcomes. In the 2020-21 and 2021-22 flu seasons, the study team sent messages to Geisinger patients in the top 10% of risk for flu and complications according to an artificial intelligence algorithm. Messages that disclosed patients' risk status significantly increased flu vaccination rates. Additionally, messages that included risk information were most effective in patients at relatively lower risk (those in the top 4-10%) compared with those at the highest risk (top 3%). The present work will test the effectiveness of high-risk messages in patients who are in the top 11-20% of risk, at high risk but lower than previous studies. These communications will inform patients they are at high risk with either (a) no additional explanation, (b) an explanation that this determination comes from an analysis of their medical records, or (c) the additional explanation that an AI or ML algorithm made this determination.
Interventions
BEHAVIORALRisk Reduction
Letter, patient portal, SMS and/or another modality
Sponsors
National Bureau of Economic Research, Inc.
Massachusetts Institute of Technology
Geisinger Clinic
Study design
Allocation
RANDOMIZED
Intervention model
PARALLEL
Primary purpose
PREVENTION
Masking
SINGLE (Caregiver)
Masking description
Providers who prescribe vaccination and diagnose conditions will not be randomized to study arms or informed of patient assignment. Although patients will not be explicitly informed which arm they have been randomized to, they will be aware of the messages they receive.
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* Included on a list of active Geisinger patients (all patients on this list attended at least one primary care appointment at Geisinger between 10/1/2008 and 4/13/2022, and either had a Geisinger primary care provider assigned as of April 2022, or were in the Electronic Health Record \[EHR\] since at least September 2021 and had at least one encounter in 2020-2022) * Aged 18 or older * In the top 11-20% of risk for flu and flu complications, according to Medial's flu complications machine learning algorithm (which operates on coded EHR data) * Has a Geisinger PCP assigned as of August 2022 * Has had an encounter in the last 2 years as of August 2022
Exclusion criteria
\- Cannot be contacted via any of the communication modalities (e.g., letter, patient portal, SMS) being used in the study, either due to insufficient/missing contact information in the EHR or because they opted out of all modalities
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Flu vaccination | Within 6 weeks of the patient's study start date | Received a flu vaccination within within 6 weeks of the patient's study start date |
Other
| Measure | Time frame | Description |
|---|---|---|
| Likely flu diagnosis | Up to 8 months | Received a high confidence flu diagnosis (with positive PCR/antigen/molecular test) and/or likely flu diagnosis (as assessed via International Classification of Disease \[ICD\] codes or Tamiflu administration or positive PCR/antigen/molecular test) (yes/no) during the 2022-23 flu season (from the patient's study start date through April 30, 2023). Note that likely flu is a superset of the high confidence flu diagnoses. |
| Flu complications | Up to 11 months | Diagnosed with flu-related complications (yes/no) from the patient's study start date through July 31, 2023. |
| High confidence flu diagnosis | Up to 8 months | Patient received a flu diagnosis via a positive polymerase chain reaction (PCR)/antigen/molecular test (yes/no) during the 2022-23 flu season (from the patient's study start date through April 30, 2023). |
| Hospitalizations | Up to 11 months | Number of hospitalizations from the patient's study start date through July 31, 2023. |
| COVID-19 vaccination rates | Up to 8 months | Received at least one COVID-19 vaccination (yes/no) during the 2022-23 flu season (from the patient's study start date through April 30, 2023). |
| ER visits | Up to 11 months | Number of ER visits from the patient's study start date through July 31, 2023. |
Countries
United States
Outcome results
None listed