As I write this post, Arizona is once again leading the nation in per capita rate of deaths. That troubles me. Deeply.
Back in July (States setting new fatality records so far in July) Arizona had a fatality rate of 11 people per million per day – see link. Today the fatality rate is over 19 people per million per day.

In July, Arizona recorded 2,050 deaths and 94,000 cases for the month. In December it recorded 2,200 deaths and 197,000 cases. So far, in the first 8 days of January it has recorded almost 1,100 deaths. Clearly January will almost certainly set new records for Arizona.


But lets also look to the future. The impact of record new cases now, is more deaths in the future. Arizona has one of the highest caseloads per capita (the number of people testing positive in the last month – and therefore used as a proxy for those actively coping with covid).
Root Cause
I’m starting to develop a theory that the performance of any given state is driven by 2 things
- the Quality of the leadership in the state (the willingness to take tough decisions ahead of time to protect hospital capacity)
- the collective ability and willingness of the people to understand their role in society to stop the spread
If I were to put that in a 2×2 grid, Arizona would definitely be in the worst quadrant. Poor quality leadership and seemingly low ability and willingness to stop the spread.
Education is key to #2. This is a virsus that is contagious before you exhibit symptoms. Therefore its not a case of just wearing a mask when you think you are sick.
When infection rates are so high, it becomes a civic responsibility to assume everyone has it – and to wear a mask, social distance etc.
Getting out of this mess
While vaccination may be the perceived solution – that won’t happen quick enough to curtail deaths in January and likely February. The only solution is to follow those countries and states that have the courage to take the action to lock down / curfew etc.
Quality of leadership in Arizona needs to improve. It also needs to improve elsewhere.