Does the use of personal electronic health records increase vaccine uptake? A systematic review

Although Personal Electronic Health Records (PEHR) have been identified as innovative tools enabling the provision of patient-centered care and prevention, evidence on the impact of their use is scant. With PEHRs being more and more marketed as easily implementable and cost-effective instruments to provide people with direct control on their health, the question on whether their use might be associated with the priority to improve vaccine coverage arises.

We conducted a systematic review following the PRISMA guidelines to retrieve, quantitatively pool and critically appraise the effectiveness of PEHR access on vaccine uptake. Analysis on PEHR effectiveness were carried out for the following comparison strata: i) PEHR access vs no intervention (standard care, no access to PEHR), ii) PEHR access only vs access to PEHR with additional features (e.g. health education materials, active reminders).

Of 3114 identified citations, 8 studies were included, the majority published in the US and before 2015; 62% were randomized trials, the rest having an observational study design. Evidence suggests a moderate positive impact of PEHR access in increasing vaccine uptake, with data available for influenza and pneumococcal vaccines, diabetic patients and childhood immunization. Pooled data report the addition of digital communication features, i.e. the delivery of educational messages, reminders and availability of scheduling features might increase vaccine uptake, as compared to PEHR access alone. However, evidence is not conclusive.

While immunization programs are struggling to achieve optimal coverage targets, it seems the potential of PEHRSs supporting informed adherence to vaccines recommendations is neither fully exploited nor explored. Which factors mediate the association between PEHRs access and vaccine uptake? Which PEHRs' design and functional components can maximize their impact? On which target populations? Which PEHR models works better for high-risk populations? Our findings can only partially answer those questions and further experimental research is needed.