Inadequate PCR testing capacity and testing remains our central problem

The increasing COVID-19 deaths in the past few days is a sign that the outbreak may be larger than our current constrained PCR testing is able to track. As I and others warned six months ago, we needed to ramp up PCR testing to keep the virus at bay and prevent future outbreaks. The President and PM gave instructions to that effect, but in reality this did not happen. Why remains a mystery, but the current outbreak is the inevitable consequence.

Increasing PCR testing reduces COVID-19 transmission (Reff). In combination with contact tracing and isolation, it is the most effective intervention we have to control the virus. Better by far than lockdowns, masks, school closures and asking people to wash their hands… Other countries that were doing well back in May did continue to ramp up PCR testing, despite in some cases having no local cases. All these countries have managed to avoid a second wave.

Whatever happens, the most important gap in our current strategy and by far the most important one we need to fix remains PCR testing. It’s not ventilators, it’s not quarantine or contact tracing, and it’s not our border controls.

This is going to be hard and it will be impossible to fix in days or weeks. As of September, my understanding is that the health authorities had not placed orders for any of the large throughput machines we need (I hope I am wrong and I am happy to be corrected). The waiting lists on most of these machines now stretch to many months, so we are not likely to be able to fix this anytime soon.

Other than placing orders for those machines now, we need to focus on optimizing our existing capacity and especially throughput in MOH labs, as well as looking at use of methods such as pooled testing and saliva testing. When I met the Health Secretary, Dr Munasinghe, a few weeks ago, he understood the need to trial saliva testing, and I hope that he and the new DG follow-up on this.

MOH is already being forced to bring in rapid antigen tests out of desperation, but we should have no illusion. The available research and data indicate these are only a second best to the optimal control strategy of intensive PCR testing. If we want to avoid lockdowns, which the President has pointed out causes great hardship to the public, we need to replace it with something else. That something else remains widespread PCR testing on a routine basis.

7 thoughts on “Inadequate PCR testing capacity and testing remains our central problem

  1. Welcome back. Many things have changed since the first wave though. Remember you wrote cautioning against not testing a contact or a returnee from abroad before releasing at the end of the 14 quarantine period – that it could potentially release an infected (and infective) person into the society But current state of knowledge holds that if exposure to virus happened prior to the start of quarantine (not during) and if the quarantine passed without symptoms, the person can be released without a test and any risk of infectivity to the society? A lot of nonessential testing could be avoided this way? (Provided that an initial PCR was done and was negative) But wait. If the PCR gave a false negative, the person even though non-infective at the end of 14 (or even 10) days, could have passed it on to his contacts before he was placed in quarantine. So not testing him again during or just before the end of his quarantine ensures that he gets counted as a negative contact and as such his pre-quarantine, post exposure contacts would never get traced thus leaving active infection cells out in the society. But isn’t there a school of thought which holds that asymptomatic contacts should not be tested nor traced as it is a waste of resources? rapid antigen test is not intended as a substitute for PCR but as a screening tool. For example, however inferior may be its sensitivity, what are the chances that a garment factory or a busload of people or a elder’s home etc screened with the rapid antigen test will fail to raise any alarm bells if the infection was present ? It will reveal the danger much much more effectively than routine temperature checks right? And these alarm bells would signal the need for PCR testing of that site and thus help target the PCR tests to sites that really need them.

    • You need to be specific about what you mean by “current state of knowledge” and what the overall goal is that we use to assess the evidence. If the strategy is elimination, which I strongly believe is the only strategy compatible with avoiding lockdowns as the President wants, and with reducing damage to the economy, then all the evidence I have seen indicates that tough quarantine measures with PCR testing remains critical.

      In some contexts, eg the UK, the health authorities have recommended less draconian quarantine protocols, including reducing quarantine below 14d/and/or not doing PCR. However, this has to be read in context. The UK Public Health England modeling that supported this assumed that a 4% failure rate, ie 4% of cases getting through, was acceptable. That only makes sense because the UK is not aiming at elimination, since they seem to be following the Imperial College influenced strategy of repeated lockdowns. There are other countries that achieved elimination and then relaxed protocols as you suggest, but have since suffered second waves. This includes Iceland and Faroes.

      In all the countries that have achieved and maintained elimination (=near zero local transmission), they continue to do 14d quarantine with PCR testing. Eg China, Taiwan, Australia, New Zealand, Vietnam, etc. However, all of them have reported outbreaks which may have been caused by asymptomatic cases that got through 14d quarantine and PCR testing. This was always anticipated since even 14d quarantine and PCR is not 100% perfect. If we know that 14d quarantine + PCR testing still has leakages which lead to outbreaks which then take immense effort to control, the question we have to answer is this. If we adopt less draconian border controls, are we willing to accept and are we able to control the increased frequency of outbreaks caused by quarantine leakages? Based on current evidence where we are battling such an outbreak probably caused by a leakage more than two months ago, I really wonder.

  2. I was referring to the growing body of evidence which holds that beyond the tenth day of symptom onset or positive pCR result, the replication competent virus is absent in infected people with no symptoms or mild-moderate symptoms. SL discharge criteria have already been updated to reflect this new evidence
    Altered discharge criteria present interesting implications for releasing of contacts from home quarantine too right? Supposing the person in [home] quarantine is really an infected individual presenting no symptoms…supposing no pcr is done on him during, before or after quarantine? Supposing he was released from home quarantine on the 14th day…what harm can he do if no longer has the infective, replicating virus?

    • Thanks CT. Without doing a full review, I doubt that the data and evidence indicate an absolute zero (0%) risk of infectiousness. It may be small, but it is not zero. I say this because we have known for some months that cases – both asymptomatic and symptomatic – can remain infectious for weeks, and since PCR tests are not 100% sensitive, some of these will end up as false negatives. So it’s a question of probabilities, and how much risk you can tolerate. In USA and Europe where the virus is rampant, one can tolerate a high level of risk, because the occasional resulting outbreak won’t effect the general situation. However, if your goal is local elimination and allowing society to function as normal behind secure borders, ie no lockdowns, minimal social distancing, no masks, and partying as much as you like, then your risk tolerance should be very low. In that scenario, you should continue to test. That is in fact what all the near elimination countries who don’t have lockdowns are doing.

      In our situation, it is not clear to me that MOH really believes that elimination is possible or feasible. Despite the evidence of Vietnam, China, etc, they might even believe that we are too poor to do this – that of course would be terribly wrong. Even if they did believe that it is possible and are doing everything possible in their view to achieve this, I don’t think they really understand what is needed. So it would not surprise me that they accept a higher level of risk than I would, or Vietnam would, or Australia would.

  3. From

    “Prior to 4 June 2020, the Coronavirus Disease 2019 (COVID-19) CDNA National
    Guidelines for Public Health Units (SoNG) recommended that a person who had
    COVID-19 needed to:
    • be free of COVID-19 symptoms; and
    • have two negative PCRa results collected 24 hours apart
    before they could go into a high-risk setting, for example, persons going into a
    residential aged care facility, or healthcare workers returning to work.
    This was a precautionary approach to minimise the risk of an outbreak in a high-risk
    setting, where the consequences could be catastrophic.
    After assessing international evidence, and Australian public health experience,
    CDNA and PHLN have agreed to revised criteria. People who are eligible for release
    from isolation, based on the clinical criteria, do not pose a risk of onward infection.
    People who have recovered from COVID-19 are no longer required to meet
    additional laboratory testing criteria prior to going into high-risk settings.
    CDNA and PHLN considered more conservative approaches. This included extending
    the period from onset of symptoms to release from isolation beyond the currently
    advised 10 days, or reverting to negative PCR tests for persons to be released from
    isolation. However, current evidence is that people are no longer infectious after
    approximately 10 days since becoming ill with COVID-19 and more conservative
    approaches are not warranted.”

    And in NZ the pre release PCR was never a requirement for people after ten (not 14 days as in SL) days since symptom onset, who had been symptomless for 72 hours
    “Recovered cases are people who had the virus, where at least 10 days have passed since their symptoms started and they have not had symptoms for 72 hours, and they have been cleared by the health professional responsible for their monitoring.”-

    • CT: Busy weekend. Not sure what the main point here is and I suspect there is more than on issue here. So some quick observations:

      1. CDC released guidance suggesting that mild to moderate cases are not infectious 10d after symptom onset, which a lot of people refer to. However, we need to read the fine print. First, the single Taiwanese study that CDC called a “large study” and gave as evidence as it found no evidence of transmission after 6d was under-powered. It actually did report that the confidence interval for infectiousness after 6d was more than zero. CDC also caveated its assessment by noting another study that reported infectiousness substantially longer after symptom onset. What worries me is that there are other reports from countries suggesting that either symptomatic or asymptomatic cases in border quarantine may have transmitted after 14d. So my feeling is that we don’t really know. There is a possibility and some basis that a low probability of infectiousness might exist more than 14 after symptoms in mild cases. So if we are aiming at elimination, then for now we should err on the side of caution. Unfortunately, the navy cluster would have been a really good situation to have studied transmission dynamics but MOH and Navy did not really do this.

      2. Regarding NZ. This requires more investigation. If you check their border control policy, international arrivals with no symptoms/evidence of COVID-19 infection are sent to mandatory isolation. They are only released after 14d and a negative PCR test.

      So judgements have to take into account the strength of the data, what the overall strategic goal is (elimination or just low incidence), our risk tolerance for new outbreaks and for some of the published empirical work being wrong, etc.

      We can also take into account practical constraints such as we don’t have enough quarantine capacity. Australia/NZ are all honest about this and govt has explained openly why they can’t repatriate all citizens. So yes if we have problems, let’s relax the protocols, but be honest that we are going to accept higher risks because we have no choice.

Comments are closed.