Most patients give consent to use their data for record linkage
MedWire News: With small exceptions, patients tend to give their general consent to use their data for record linkage, as noted in a systematic review published in the Journal of Medical Ethics.
With the growing use and dissemination of computerized medical records, it is often beneficial to use record-linkage techniques to ensure that as much medical information pertaining to a single individual is available at the same time. However, because this information is highly personal, there is often a question about whether the patient needs to give his or her written consent for this linkage.
For secondary data research, patients do not need to give consent, as the number of study participants usually makes this impractical. Furthermore, the data can be used in ways that may not require consent, but those ways are not often clear at the time of data collection.
Márcia Elizabeth Marinho da Silva (Agency for Health Plans and Insurance, Rio de Janeiro, Brazil) and colleagues conducted a systematic review of 11 studies that assessed consent proportions (using traditional "opt-in" consents) to record linkage. All except one were published after 2000, and the studies originated in Taiwan, Canada, the UK, and Australia. Over 108,000 patients were represented and came from either population surveys or health services organizations. Consent was requested by letter or face-to-face approaches.
Consent rates ranged from 39% up to 97%; seven studies had a consent rate of 88% or greater. Neither subgroup analyses of the studies' characteristics nor a meta-regression found an explanation for this heterogeneity. The three studies with the lowest consent rates were carried out in very different populations: patients with diabetes, participants in a women's health longitudinal study, and patients enrolled in a stroke registry.
The authors note that one limitation of their analysis is the small number of articles included, but concede that there is a complex interplay between individual and contextual factors, such as age, race, and income. They acknowledge that there needs to be a balance between no or minimal consent, which could expose patient information to privacy breaches, and consent or security matters restrictive enough to encumber potential research. Other researchers have recommended developing more complex procedures for safeguarding privacy, rather than just relying on simple patient consent.
The authors conclude that their results, as well as earlier findings, "lend support to policies that, while keeping relevant ethical controls in place, do not require individual informed consent for each and every study that relies on secondary data."
By Stephanie Leveene