More and more evidence is coming out that points to the importance of contact tracing for control of coronavirus. A roundup of some recent published work on this topic, and challenges raised by their findings.
A recently released pre-print from @AdamJKucharski et al uses a unique dataset of residents’ movements and social interactions in the UK to model and compare different approaches to controlling #COVID19. cmmid.github.io/topics/covid19…
Authors compare how the effective reproductive rate of the epidemic changes based on implementation of a variety of control approaches, including: no control, self-isolation based on symptoms, mass testing regardless of symptoms, social distancing, and testing/tracing of cases.
The worst performing approach was mass testing regardless of symptoms. Testing 5% of the population/week & self-isolating positives led to a 2% reduction in R0. Relatively few infections are detected this way, and most would have already spread infection to others when detected.
Combined testing and tracing strategies performed best. Self-isolation (SI) of those testing positive alone reduced R0 by 32%, but SI plus manual tracing of all contacts reduced it by 61%; reducing contacts (i.e. social distancing) on top of that reduced R0 even further.
The authors point out that with lots of transmission, the number of contacts to be traced and tests performed can become overwhelming and strain resources. The estimate 20-30 contacts on average being traced and tested per case, through manual contact tracing.
A key takeaway: manual contact tracing and self-isolation can help reduce transmission, but likely not enough to fully interrupt spread. Maintaining some social distancing, in addition to testing, tracing, and isolation, will be important.
Another study by Cheng et al from JAMA Internal Medicine described the epidemiological investigations of the first 100 cases of covid-19 and their contacts in Taiwan. jamanetwork.com/journals/jamai…
These 100 cases had >2700 identified close contacts (average # of contacts was around 27 per case). 22 secondary cases occurred among the contacts, for an infection risk of 0.8% overall.
Household contacts were more likely to become infected by cases: around 5% of HH contacts developed the disease from a case in their HH. Much of the transmission occurred fairly early on in cases’ disease, before the onset or right around the onset of symptoms.
Authors conclude the characteristics of Covid-19 means interrupting transmission through contact tracing is difficult because of pre-symptomatic transmission and a rapid serial interval (short time between generations of cases).
They suggest that finding and isolating symptomatic patients alone may not suffice to contain the epidemic, and suggest maintenance of social distancing measures will also be needed.
Another contact tracing study from Bi et al in Shenzen China also found most infections occurred in households (11% HH secondary attack rate). Symptomatic contacts were identified through tracing an average of 2.7 days after symptom onset. thelancet.com/journals/lanin…
An accompanying editorial says the study shows “contact-based interventions are more efficient than case-based interventions to reduce transmission, since infected contacts are typically isolated earlier in their infection history than index cases.” thelancet.com/journals/lanin…
In other words, contact tracing to ID and isolate cases is key: “we contend that enhanced case finding and contact tracing should be part of the long-term response to this pandemic—this can get us most of the way towards control.”
These studies show that what's worked in the countries that have seen success is not a mystery. Good public health practices are what work. And we're going to have to do that everywhere if we want to control this pandemic.
Another new study: identifying 50% of symptomatic infections & tracing 40% of their contacts reduces transmission in a 2nd wave enough to allow the reopening of economic activities while keeping impact on the health care system manageable.
