As the Prime Minister, premiers and chief ministers plan to make decisions about a staged relaxation of the COVID-19 rules, they are entitled to feel very satisfied with their efforts and their political courage. The proof is to be seen in the statistics of Australia compared with other countries, particularly in North America and Europe. We have had only a fraction of the mortality per million residents , and only a fraction of the incidence of mild and severe cases of the disease.

The credit goes to early political action to quarantine Australia, containment measures, and in securing the general co-operation of the population in measures to reduce the possibility of transmission of the coronavirus. The nations which appear, so far, to have done best have been in south-east Asia and Oceania, particularly Taiwan, South Korea, Hong Kong, New Zealand and Australia. Had Europe and North America been as successful in mobilising their resources against the virus, perhaps 200,000 more people would still be alive.

But it is far too early to celebrate, here or anywhere else in the world where the curve is going down - in our case to very low numbers of new cases, active cases, and deaths. There is, first, the job of cranking up the economy again, with a very close eye on the continuing risk from a virus which is still present in the population. Experiences with hotspots, whether with cruise ships, meatworks or nursing homes, illustrate the potential of small numbers of cases to suddenly duplicate exponentially again, undoing all of the good work. Already some of the nations with a fairly good record - Singapore, for example - are back fighting the smouldering remnants of a fire they thought they had put out. There, as here, the general population is as susceptible to the virus as ever - only a tiny fraction of the population has acquired any immunity, the strength of which we do not yet know - by having had the virus, perhaps asymptomatically, and then "recovering". The hopes of the rest of us that we do not catch it depend on quarantine, isolation and containment programs that make exposure to the virus rare, and catching it less likely.

Doctors working on severe cases know by now a good deal about the way that the virus affects the susceptible - particularly those with compromised immune systems, or those who are old. Screening programs are typically finding a physical or serological case of the virus to every 99 cleanskins. From milder cases, which involve hospitalisation but not the need for transfer into intensive care units, doctors have found a wider array of symptoms than those usually described, including diarrhoea, digestive tract problems, and, in some cases, loss of a sense of smell and taste. A good many mild cases - about 40 per cent - do not have respiratory or flu-like symptoms, and, to date, have not been picked up by most screening programs, unless they have been pre-identified as suspects because of overseas travel, known exposure to a case or by case-tracking.

In Canberra, for instance, words such as mass screening or random screening have been thrown around, but no member of the general public has been tested unless they have respiratory symptoms when they drive to locations such as EPIC for a test. The result is that none of the thousands of tests performed have much epidemiological significance: we do not know how many cases had the virus without symptoms, or had it with symptoms so mild that they did not seek medical attention (or, if they did, did not arouse the suspicions of their doctor). Many of those who had flu-like symptoms of a non-respiratory nature - upset stomach, loss of appetite, vomiting, nausea and fatigue - may have figured, on the basis of the information distributed by public health authorities, that they simply had a flu and would not, in the middle of a well-publicised public health crisis, trouble the doctors.

This might well have suited those managing medical services to those who were seriously ill. We had only a fraction of the cases for which the system - after seeing what happened in China and early in the European pandemic - had organised. Nonetheless the seriously ill were difficult cases, and those concerned with them did not need the distraction of less serious cases, or cases that would resolve either with no treatment or only routine treatment. Indeed many of those managing the system rationalised their relative lack of interest in vigorous case-finding in the general population by saying that anyone with the virus, but no symptoms at all, or only mild symptoms, would be sent home for isolation. That is, they would be subjected to much the same regime being asked of the general population - staying at home as much as possible, keeping their distance from others, and using good hygiene practices.

That may be all right so far as any immediate drag on hospital resources was concerned. But such people still have the power to transmit the virus to others, whether from person to person, or by shedding the virus on to some surface touched by another. The social containment measures may have had a pronounced effect in reducing such transmission, and the length of time the measures have been in force may mean that anyone who had the virus at the time such measures began is no longer spreading it. But that is not necessarily true for a person to whom they passed it while they were contagious, or to a third- or fourth-generation recipient.

We know that the greater proportion of cases are asymptomatic - even if the carrier is spreading the virus to others. The evidence suggests the asymptomatic are as virulent as any other. With the absence of good epidemiological studies anywhere, no one has much idea of the proportion of asymptomatic cases, but most assume that it is at least 10 and up to 100 times the number of serious cases. It may even be higher. It is also, probably, at least 10 times the number of mild cases requiring hospitalisation, or at least treatment that involves screening.

One reason that the public health authorities are reasonably bullish about a very severe dent on incidence and prevalence of COVID-19 is that there are now relatively few new cases, considerably fewer than at the peak weeks ago. If we assume, say, that 5 per cent of the infected end up being hospital cases then the number of fresh cases suggests not many new people are being infected.

Against that is the fact that transmission in the community is creating an ever-increasing pool of cases, only a tiny proportion of whom have obvious symptoms or any at all. By now these may have been in the community long enough for many generations of transmissions. One study in the American Journal of Gastroenterology published on April 15 suggests that mild cases are likely to be out in the community, shedding and spreading the virus, for up to a week longer than those whose case becomes severe. Likewise, those with digestive tract symptoms only may be carrying the highest viral load, and appear to be infectious for significantly longer than cases with digestive and respiratory symptoms, and those with respiratory symptoms only.

For all of this set of patients, all of whom were hospitalised and later followed up, the mean interval between the first onset of symptoms and viral clearance (demonstrated by two negative tests at least 24 hours apart) was 38 days. But those with digestive symptoms only had an average interval of 41 days, compared with 33 days for respiratory cases.

To me this suggests that the exiting screening in the community might be the wrong way around. We can expect that almost all serious cases will get thorough screens regardless. If the focus of screening is to isolate virus carriers, it would be better to spend as much time looking for digestive tract symptoms as respiratory ones. The failure to do this over recent weeks, as well as the strict insistence on respiratory symptoms, may well mean that an ill-judged policy has prevented the finding of carriers.

The authors of the study do note that digestive tract symptoms are not uncommon, and that most cases of new-onset diarrhoea, nausea or vomiting are not due to COVID-19. Nevertheless, doctors should consider the possibility in the course of the pandemic.

"Failure to recognise these patients early may often lead to unwitting spread of the disease among outpatients with mild illness who remain undiagnosed and unaware of their potential to infect others," the study concludes.

When the public was being persuaded of the virtues of isolation and distancing policies, it was said that an asymptomatic person could spread COVID-19 to more than 9500 people in 40 days (that's with an R0 of 2.5 and 10 generations). If a high proportion of them have no symptoms, the impetus on them to conform to strict rules, or to stay on course, is much lower. As Andy Slavitt pointed out on Medium, the higher the proportion of asymptomatic or mild cases, the lower the case-fatality rate. But that can make it more deadly, given one's guard is lower in the presence of those who appear uninfected.

The lockdown rules would have dramatically cut these numbers, but they are unlikely to have reduced it altogether. Given that government seems to intend to bar air travel (except possibly to New Zealand) and maintain barriers to entry for non-Australians, it seems likely that it will be from this group in the population - whose numbers are unknown given the failure to conduct surveys - that second and third waves of the disease will come. With Spanish flu a century ago - which involved a fair measure of social distancing - the second wave killed many times more than the first. Those who had not had the virus were not in the least immune, and the only way in which they benefited from the earlier phases was that treatment was more organised. On the other hand, one might expect that a certain fatigue with social distancing, and unwillingness to revert to it once the controls have been loosened, may mean that compliance is weaker.

It's all very well to say that the longer the lockdown, the more damage is being done to the economy, and the longer it will take for it to revive. But premature reopening, followed by an emergency lurch back into quarantine if the epidemic flares up again, will make those costs, and those delays, seem like a picnic.

The argument of the government health advisers against random mass screening -- or selective random screening among vulnerable groups, such as Aboriginal people, those in aged care, the disabled, people with diabetes, or primary and secondary schoolchildren, is, they say, both the problem of small numbers and low prevalence rates and the damnable problems of false positives and false negatives. The overwhelming proportion of the population does not have the virus, yet at least. Even if the true rate of infection is somewhat higher than expected, it is still only a small percentage of the population. If the whole population is screened, then the fact that some of the testing machines are unreliable may mean that many people are identified as having (or having had) the virus when, as further testing will show, they have not been exposed. Others will be said not to have the virus, when they are in fact positive.

That seems to me a lazy excuse, particularly given the increasing sophistication of the testing equipment and given that the call on resource for a medical (as opposed to social) fight against the virus is only a fraction of what was planned or expected. Nor should there be any great problem of doing double, or triple, tests, say, directly for presence of the virus, and a serology test for antibody response. Perhaps a faecal sample would be most reliable. Increasingly there is equipment of high specificity and sensitivity, with far fewer wrong calls.

MORE JACK WATERFORD:

Australian public health authorities routinely screen large populations for conditions of lower prevalence than coronavirus, and in situations in which accurate diagnosis is far more difficult than with coronavirus. An example might be with PSA tests for prostate disease, of the most doubtful efficacy. Even Pap smears suffer from major problems of false positives and negatives, often leading to unnecessary further medical interventions.

What's important to remember is that we are not simply looking for cases who should be put immediately in an ambulance for treatment, or should be told to return to lockdown. While random mass testing should involve giving results to patients as promptly as possible, a result for epidemiological purposes is not as urgent as one for clinical screening, allowing for examinations to be conducted centrally. We are trying to get a picture of disease in the community.

In the US there are plans to have two or three major random mass screenings, each involving about 50,000 people, and each involving the collection of blood samples. With most of the US not yet at its peak incidence, analysis of the data, but not the data collection itself, may be delayed.

Modern epidemiology began in the 1950s with doctors such as Archie Cochrane, firmly committed not only to mass surveys and providing service but also to research. Cochrane is now commemorated by the Cochrane Collection - a database of international case studies, closely analysed for weaknesses of design and conclusions drawn.

I am beginning to suspect that newer technology, and the computer, has tended to make the typical epidemiologist of today more office-focused, and more statistically oriented. Something like econometrics, perhaps for shy doctors and bureaucrats with a fear of coming into contact with actual patients. Cochrane used to stress that the very word epidemiology meant the study of disease or epidemics in one's community, among one's people. Like many non-bureaucratic epidemiologists, here and abroad, he would be scathing of the low priority given by bureaucratic advisers to government to the need for surveys.

As we cautiously begin to lift the lockdown, if we don't know who the silent carriers are, how can they play an active role in keeping the community safe?