[Goose]

The CDC changed the guidelines without citing any data. Newsom and Cody both rejected the change without citing any data. That is disturbing IMHO. SCC, California, and the CDC should be able to supply two numbers:

1. The number of people who were tested due to an initial test given because they were named as a contact and were also asymptomatic. Call this number X.
   
2. The percentage of X that tested positive. Call this number P.
   

If X is a very small number compared to the number of total tests given in a jurisdiction, the argument is moot. You might as well test because it isn't a resource driver. One would expect this is not the case, but it must be considered. Next, if P is greater than the nominal percent positive test results in a jurisdiction, you should continue to use the old guidelines. The "expected value" of the test is greater than the average test you are doing, so you want to continue doing them. If P is less than the percentage of COVID-19 in the general population, you for sure should not test these people because they are LESS likely to test positive than a random person. One would expect this won't be the case, but it must be considered. If the results are "in between", the decision becomes more difficult.

One might expect that since the contact is known to have been exposed to COVID-19 that it would be much more probable that a test would produce a positive result than the nominal percent positive in a jurisdiction, but there are good reasons that may not happen. The people going for COVID-19 tests on their own accord are often symptomatic and therefore more probable to test positive. The asymptomatic person who is a contact may not actually have COVID-19. He also may be in the incubation period and will not have enough viral load yet to test positive. The exact profile of how quickly one test positive after initial exposure is not well known yet, and depends on many things that may be specific to a given case and "local" to a given jurisdiction. However, all jurisdictions should know P. If they don't, we are all just guessing.

If P is "in between" (which is likely IMHO), second order effects come into play. The benefit of an "early" test is that if it comes up positive, the contact knows he must isolate instead of quarantine. This might prevent/reduce spread within the family group. However, if the contact is already testing positive, it may be too late. While it would reduce the total number of cases, this "branch" of the infection is terminated in any case.

The second benefit is that contact tracking can start immediately based on the positive test. If any are positive it yields even more contacts. It may be that the contact will remain asymptomatic during his entire infection. In that case, the early test would be the only way that person would ever have been identified as positive and asymptomatic spread has been reduced. However, one must realize the person would be in quarantine anyway, so the reduction generally would be minimal.

The trade-off in deciding to not test all contacts would be if so doing would decrease the time it takes to get test results back to those being tested. If it doesn't do that, there is little reason not to test. I am sure there are estimates of this effect that can be cited to justify (or not) a policy of not automatically testing all contacts. If that is the justification, I would like to see it made public. If P is "almost" big enough, and the projected speed improvement is "small", it isn't worth not testing. If we all knew these numbers we could at least have an informed opinion.

It also must be factored in that none of this really matters if the contact isn't going to quarantine/isolate anyway. In SCC, 32% of the positive tests are "contacts to a case", which I believe is defined as contacts that are in quarantine and test positive, either initially or later. MT can correct me if that isn't right. 32% isn't insignificant, but it also isn't great. In South Korea, the corresponding number is 95%. Second, of that 32%, what is P? If it is "small" (say 5%), not much was accomplished by the initial testing. Based on what San Mateo County is reporting, a large percentage of positive tests, let alone contacts, are not quarantining/isolating. If so, all this is much ado about little.

[Goose]

The CDC changed the guidelines without citing any data. Newsom and Cody both rejected the change without citing any data. That is disturbing IMHO. SCC, California, and the CDC should be able to supply two numbers:

1. The number of people who were tested due to an initial test given because they were named as a contact and were also asymptomatic. Call this number X.
   
2. The percentage of X that tested positive. Call this number P.
   

If X is a very small number compared to the number of total tests given in a jurisdiction, the argument is moot. You might as well test because it isn't a resource driver. One would expect this is not the case, but it must be considered. Next, if P is greater than the nominal percent positive test results in a jurisdiction, you should continue to use the old guidelines. The "expected value" of the test is greater than the average test you are doing, so you want to continue doing them. If P is less than the percentage of COVID-19 in the general population, you for sure should not test these people because they are LESS likely to test positive than a random person. One would expect this won't be the case, but it must be considered. If the results are "in between", the decision becomes more difficult.

One might expect that since the contact is known to have been exposed to COVID-19 that it would be much more probable that a test would produce a positive result than the nominal percent positive in a jurisdiction, but there are good reasons that may not happen. The people going for COVID-19 tests on their own accord are often symptomatic and therefore more probable to test positive. The asymptomatic person who is a contact may not actually have COVID-19. He also may be in the incubation period and will not have enough viral load yet to test positive. The exact profile of how quickly one test positive after initial exposure is not well known yet, and depends on many things that may be specific to a given case and "local" to a given jurisdiction. However, all jurisdictions should know P. If they don't, we are all just guessing.

If P is "in between" (which is likely IMHO), second order effects come into play. The benefit of an "early" test is that if it comes up positive, the contact knows he must isolate instead of quarantine. This might prevent/reduce spread within the family group. However, if the contact is already testing positive, it may be too late. While it would reduce the total number of cases, this "branch" of the infection is terminated in any case.

The second benefit is that contact tracking can start immediately based on the positive test. If any are positive it yields even more contacts. It may be that the contact will remain asymptomatic during his entire infection. In that case, the early test would be the only way that person would ever have been identified as positive and asymptomatic spread has been reduced. However, one must realize the person would be in quarantine anyway, so the reduction generally would be minimal.

The trade-off in deciding to not test all contacts would be if so doing would decrease the time it takes to get test results back to those being tested. If it doesn't do that, there is little reason not to test. I am sure there are estimates of this effect that can be cited to justify (or not) a policy of not automatically testing all contacts. If that is the justification, I would like to see it made public. If P is "almost" big enough, and the projected speed improvement is "small", it isn't worth not testing. If we all knew these numbers we could at least have an informed opinion.

It also must be factored in that none of this really matters if the contact isn't going to quarantine/isolate anyway. In SCC, 32% of the positive tests are "contacts to a case", which I believe is defined as contacts that are in quarantine and test positive, either initially or later. MT can correct me if that isn't right. 32% isn't insignificant, but it also isn't great. In South Korea, the corresponding number is 95%. Second, of that 32%, what is P? If it is "small" (say 5%), not much was accomplished by the initial testing. Based on what San Mateo County is reporting, a large percentage of positive tests, let alone contacts, are not quarantining/isolating. If so, all this is much ado about little.