Labour’s tweet about Serco is missing some important context
21 October 2020
What was claimed
Statistics show that Serco is much worse at contact tracing than local councils.
This comparison needs context. Public and private contact tracers do different jobs, and there’s reason to believe it may be easier for local teams to reach a higher proportion of their contacts.
The Labour party claimed on Twitter on 14 October that contact tracers working for local authorities are more effective than those employed by Serco, an outsourcing company that runs part of the NHS Test and Trace service in England.
However, you can’t make a direct comparison by looking at the proportion of contacts reached, because public and private contact tracers are doing different kinds of work, which is measured in different ways.
What jobs do Serco and the councils have?
All coronavirus cases in the English test and trace system are classed as either “complex” or “non-complex”.
Complex cases managed by PHE and local health teams relate to a particular place where there might be an outbreak, such as a care home or a prison, and therefore require special local attention. In general, they account for less than 10% of the cases in the system.
When a case is designated “complex” it is passed to local teams to deal with. These cases may involve an infected person having large numbers of close contacts, and are often managed “at a situation level”, rather than individually.
For example, the manager of an office might be asked to close it, and inform all staff. When this happens, “all associated contacts will be counted as having been reached and asked to self-isolate”, meaning that many contacts might be “reached” without each one being contacted individually by a contact tracer.
It may also be easier to trace contacts in complex cases, because they are often linked to a place where they live, or which they visit regularly, such as a prison, a workplace or care home. By contrast, the contacts in a non-complex case may more often be passersby that the infected person has no way to get in touch with.
Since the launch of the English test and trace system, local teams have reached about 98% of the close contacts identified in complex cases.
As the name suggests, their work is complex, but it involves fewer cases, and they can sometimes be recorded as “reaching” many contacts in one go.
Non-complex cases, traced by Serco and Sitel
Non-complex cases are the other confirmed coronavirus cases in the community, generally accounting for more than 90% of those in the test and trace system.
When someone tests positive, and the case is “non-complex”, they are required to give details of their recent close contacts to medically trained staff who work for the NHS. These details are then passed to national contact tracers, some working for Serco, some for Sitel.
Contacts of non-complex cases have to be reached one by one, either by text message or email, or on the phone. However, in some cases this is not possible. About 20% of the contacts passed to national contact tracers so far did not come with any communication details, and so could not be traced.
Infected people also effectively do some of the contact tracing work themselves—for instance by telling people they live with to self-isolate before a contact tracer has a chance to call them. In these cases, “This results in these contacts not being recorded as reached and asked to self-isolate.” The Department for Health and Social Care says this “may be a contributing factor to a lower proportion of household contacts being reached in comparison to non-household contacts.”
Non-complex contact tracers have consistently reached a smaller percentage of their contacts than local teams. However, in the week’s data used by Labour, about 20% of the contacts passed to contact tracers did not come with any communication details.
If you set aside the contacts provided without details, 78% were reached successfully in that week. This is lower than the 97% reached by local teams, but these two numbers need to be seen in context, given the different nature of the tasks involved.
We also don’t know how many of these contacts were handled by Serco’s own contact tracers, as opposed to Sitel’s. The published data isn’t broken down in a way that would allow us to check.