1/ BC's privacy scheme relating to the release of small data numbers in communities makes NO sense. Yesterday, the docs said that neighborhood data could be released only when numbers grew high enough. Otherwise it compromised privacy.

2/ This is a govt idea. On both overdoses and COVID deaths, there is an idea that being specific about the number of people in a community to be afflicted with a problem compromises privacy IF that number is between 1 and 4.

3/ When that number gets high enough, it isn't as much of a privacy problem. It sounds, when you hear it, like it makes sense when applied to communities. It makes NO sense.

4/ The idea is that if you know if there are 2 or 3 people in your community, you might be able to figure out who those people are, or if your neighbour is one of those 2 or 3 people.

5/ Now, the identification of the first case does make a dent in privacy. (Though it's low.) If you know someone in your community has X disease, you know it's possible that person is your neighbour. But gov't doesn't have an issue with identifying one case.

6/ They are also fine to distinguish between there being 10 cases in a community or 100 cases. Their issue is with being specific about small numbers. And that is where logic goes out the window.

7/ The neighbourhoods are often 30k in size. But let's use an apartment building as an example. The BC CDC says it's harmful to specify the exact numbers of people in that community with X disease, b/c others might figure things out IF numbers are small.

8/ But if numbers are bigger, there is less of a privacy problem. But that's totally wrong. Imagine an apartment building. If you hear a neighbour cough and know there are 2 cases in the building, there is only a small chance that neighbour has it. It's hard to guess.

9/ Knowing whether there are 2, 3 or 4 doesn't change things hugely. But. When you get into bigger numbers, it makes it much more likely, and much easier to assume, your neighbours are affected. Bigger specific numerators in this way can be worse for privacy, not better.

10/ That is directly opposite to what the doctors were saying yesterday. They were saying that giving specific numbers about small numbers of cases is bad for privacy. But that when those numbers rise, it's good.

11/ A similar logic underlies overdose numbers. We are only told when deaths reach 5 in a community. Which makes little sense, given that high numbers of deaths from any one cause make it more likely that you can guess the cause of death of an individual.

12/ Why might health officials figure this? Well, because they are working in reverse from normal people. They don't know the name first. They know the disease. So as the number of people associated with a disease grow, they become more anonymous to the system.

13/13 Anyways. There may be better ways to put that, but, PLEASE no one tell anyone in the BC government about the census, a huge effort to record and release data about local populations at neighbourhood levels that would make officials quake with fear.

14/13: One more thing: The docs said the data was a draft and needed verification. Given they seemed to totally miss the math behind the variant-driven 3rd wave, I'm not at all comfortable with leaving COVID data analysis solely to the BC CDC.