Coronavirus: How to sort out the HSE’s contact tracing in six weeks
Covid-19 briefings should specify the time it is taking to trace contacts of positive cases
‘The Covid tracker app was launched after a long wait with much fanfare and promise but its usefulness remains unknown.’ Photograph: Niall Carson/PA Wire
Yet again the Health Service Executive Covid-19 contact-tracing system is in difficulty, something that is hardly surprising given the rapid rise in the number of cases testing positive over the past months. On August 17th, responding to a query from RTÉ, the HSE said it was “developing a new testing and tracing model to meet the demands of Covid-19. As part of this process there will be a new model for contact tracing.” Two months on, the same model is still in place.
Research suggests that for contact tracing to be in any way effective, the period between an individual developing symptoms and at least 80 per cent of their contacts being put into quarantine must be less than four days. This time span includes a test phase (time from symptom development to GP referral to test to result) followed by a trace phase (time taken to phone someone to inform them of a positive result and the actions they need to take, identify their close contacts and inform them that they must quarantine).
So how has the HSE contact-tracing system performed? Data on the HSE website is provided in a manner that makes it difficult to determine overall turnaround times. Average times are provided for individual stages but data for the critical initial stage from symptom development to referral for test is not provided.
The data does show that over the past two months the median time from test to result has remained steady at about 1.1 days. However, the mean time taken to complete subsequent contact-tracing calls has increased rapidly from 1.1 days in early September to 2.6 days in mid-October. More recent data (October 23rd) gives the number as 3.8 days. This is a rapid deterioration, with the numbers suggesting that instead of the recommended four days, the process is currently taking at least nine days, rendering test and trace ineffective. Over the course of the six-week lockdown, this must be improved.
Throughout this pandemic considerable attention has been given to the turnaround time from test to result and the HSE has been successful in increasing the volume and speed of the testing process. However, rather less attention has been paid to the tracing process (albeit there have been moves to recruit paid staff to the tracing centres) and how that might be adapted and improved to meet the changing requirements in the ebb and flow of this pandemic.
It may be fortuitous that the system breakdown last week led to people being asked to conduct their own contact tracing – at least it opens the door to other possibilities of how the process might be conducted. Manually phoning people is a laborious and time-consuming business. To reduce timescales more automation is needed for routine cases, allowing public-health experts time to deal with the more complex cases.
There are a number of key steps required to improve the speed and efficiency of tracing. We must first ensure the four-day metric becomes part of the public consciousness. “Eighty per cent end-to-end turnaround from symptom development to testing being carried out and contact tracing being completed” is quite a mouthful, and not particularly memorable.
Simplifying this target to “symptom to quarantine – 80 per cent”, or simply “SQ80” and making it a key metric on the HSE dashboard would generate meaningful data to enable tracking of progress on this crucial metric. Failure to achieve this metric means testing and tracing becomes ineffective in controlling the spread of the virus.
Along with figures for cases and deaths, the current SQ80 value should also be included in daily briefings. Public knowledge of the situation regarding contact tracing can help advocate for improvements if matters deteriorate or problems are not resolved. It might also encourage people who test positive to contact and inform their close contacts. Monitoring this will be particularly important during the lockdown.
There is also substantial scope for automation to support the somewhat slow and cumbersome process that is currently in place. The covid tracker app was launched after a long wait with much fanfare and promise but its usefulness remains unknown. While this was in development it is possible that other less spectacular technical improvements were mooted but not implemented. Indeed, the May HSE roadmap to enhance capacity and turnaround for contact tracing recognised that delivering positive results by SMS could speed up the process. However, this went live only last week.
Extending this SMS procedure could greatly enhance speed. For example, the text is currently sent with an embedded link to a HSE website providing information on what actions to take regarding isolation. This website could also contain a form to allow the recipient to enter the phone numbers of contacts. On submitting the form, a text would be sent by the HSE system to the contacts informing them that they must quarantine, thereby reducing this process from several days to a few minutes. Self-service technology could also be used. This will allow people, after an initial short contact-tracing call, to enter details such as symptoms, underlying conditions, etc, into a database without the involvement of a contact tracer.
The similarities between a contact-tracing centre and a telemarketing call centre may not be immediately apparent, but they are there. Both are involved with continuously making outbound calls. Extensive system support and technologies have been developed to ensure that call centres operate at peak efficiency by utilising features such as repeated auto-dialling to no response numbers, the ability to transfer calls seamlessly from one group of agents to a more expert group (in this case public-health specialists) and the automatic output of statistics such as number of calls made, type of calls and duration of calls, all essential for planning and operating the centre. What is needed here is to upgrade phone systems currently in use at contact-tracing centres to meet call-centre specification.
All efforts must be made to get SQ80 to four days or less. This is a minimum requirement. If lockdown has the desired effect of decreasing the number of positive cases, this is the opportune time to ensure our tracing system is ready for the next wave.
Dr David Joyce and Prof Eilish McAuliffe are health system researchers at the UCD Iris Centre