Determinants of consent for electronic health information exchange: an observational retrospective study.

backgroundSharing of patient electronic health record (EHR) data between healthcare providers can enhance quality and efficiency of healthcare provision, and patient safety. Health information exchange is allowed only with explicit patient consent. In the Netherlands, patient consent and the exchange of information is organized nationally. Consent status is recorded routinely in general practice EHRs. This study examines how various characteristics at the individual and general practice level influence this consent for electronic health information exchange (HIE).

methodsRoutine EHR data from general practices and out-of-hours primary care services participating in the Nivel Primary Care Database were analysed for the period 2017-2019, just before the coronavirus disease 2019 (COVID-19) pandemic outbreak. Bivariate chi-squared test analysis and multilevel logistic regression analysis, adjusted for practice-level clustering, were conducted for each of the 3 observed years to assess whether consent for electronic HIE ("yes" or "no") was associated with the individual's health and healthcare utilization, demographics, neighbourhood characteristics (socioeconomic position, degree of urbanization, area deprivation), and general practice characteristics (practice type, EHR system, practice size).

resultsBetween 2017 and 2019, 38-45% of the individuals provided consent for electronic HIE. Individuals with a higher number of different prescriptions or those with long-lasting health problems or chronic diseases had lower odds of providing consent, in each of the 3 years. This result also applied to female and older individuals (aged ≥ 65 years). In contrast, individuals from below-average socioeconomic position neighbourhoods living in deprived urban or hardly urbanized regions generally had higher odds of providing consent.

conclusionsIn contrast to what was expected, patient groups with higher healthcare utilization were less, or as likely to provide consent for HIE compared with individuals with no or lower healthcare utilization. This implies that the population most likely to benefit from HIE is, in fact, less likely to profit from it. Further research is needed to determine whether these differences arise from individual trust, privacy concerns, transparency issues or other factors such as physicians' HIE beliefs. Optimizing the national HIE system in the Netherlands requires considering multiple influencing factors, on both the individual level and practice level.