As part of my literature review, I have identified a number of papers examining the attitudes and preferences to sharing patient data for secondary use purposes. One of the themes of these studies is a high number of respondents reporting their support for the use of data in medical research, while a much lower number reporting a willingness to share their own data. In this blog post, I suggest bystander intervention, a social-psychology theory to explain these somewhat paradoxical views.

1984, in the New York City borough of Queens, Kitty Genovesa, a young women of Italian decent was fatally stabbed in an overlooked parking lot, close to where she lived. While there reported to be many witnesses to the crime, nobody reported it, leading to numerous studies into the cause of this social inaction.

Darley and Latane (1968) developed a theory known as bystander intervention, or the bystander effect. Their theory suggests that the greater the number of bystanders witnessing an incident, the less likely it is that anyone will intervene. Further studies into the effect have shown that the knowledge of other bystanders does not have to be explicit, but implicit knowledge of their existence is enough for this effect to be observed.

The bystander effect typically requires three elements: (1) an intervenable act, (2) a perceived cost associated with intervening and (3) other bystander actors.

So, how does this all relate to the sharing of medical records? Good question.

Being asked for consent to share medical data to help in the research of a problem is arguable an (1) intervenable act, an act that comes with (2) a cost to privacy that (3) involves many other bystanders who are able to provide their data to help.

When forming a consent decision, the implicit knowledge that others are able to help may provide the internal justification for inaction to protect against the potential costs of intervening.

If this theory is relevant to the health-care data disclosure process, how can it be minimised to increase opt-in rates? One counter theory is that when people are exposed to, or have knowledge of others performing prosocial acts, it may encourage them towards similar prosocial acts.

Could the use of online social messaging platforms be used to create a prosocial data sharing behaviour model that others follow?