Nevertheless, testing capacity remains woefully inadequate to the scale of the problem. The outbreak remains broadly entrenched across all age groups (Figure 2).

Note: Data for this report was updated Friday, January 15 allowing 4 full days to adjudicate cases and keep week-over-week backfill <10%. This allows more interpretable comparisons and graphics. All comparisons are week-over-week changes. Future updates will be released on Saturdays.

Test positivity among those undergoing traditional nasopharyngeal PCR testing decreased from 35% the week ending January 3rd to 33% the week ending January 10th (Figure 3). The decline in positivity indicates that viral transmission is growing slower than testing capacity is increasing. Nevertheless, 33% positivity reflects a woefully inadequate testing capacity as positivity should be less than 5 – 10%.

As of Jan 15th, 4849 (56%) of Arizona’s 8664 general ward beds were occupied by Covid-19 patients, a 1% decrease from the previous week’s 4907 occupied beds (Figure 4 and Figure 5 Panel A). Another 656 (8%) beds remained available for use. The number of available beds is higher than the previous week’s 618 beds.

At 4849 beds and 56% occupancy, Arizona has seen the first, albeit small, week-over-week decline in Covid-19 patients and percent occupancy in several months. While still incredibly high, the change is welcomed.

As of January 15th, 1144 (64%) of Arizona’s 1800 ICU beds were occupied with Covid-19 patients, a 2% increase from the prior week’s count of 1122 patients (Figure 4 and Figure 5 Panel B). An additional 138 (8%) ICU beds remain available. This is higher than the prior week’s 131 beds. Covid-19 occupied ICU beds and occupancy remain at or near record levels.

Arizona hospitals’ safety margin continues to be dangerously low in both the ward and ICU (Figure 6). Most hospitals are still postponing scheduled procedures to ensure adequate capacity exists for patients with Covid-19. The hospitalization numbers and occupancy trends suggest some moderation of demand for Covid-19 care. While this is encouraging, it is difficult to accept these numbers at face value owing to capacity constraints and subsequent changes in care practices. For example, patients may be discharged from the ED with higher acuity illness because hospitals have reached capacity for beds or personnel. Even if this moderation persists, hospitals will remain dangerously overcrowded for the foreseeable future compromising access to care for Covid-19 and non-Covid patients alike.

The week ending December 20^th has recorded 759 deaths, making it Arizona’s deadliest week so far (Figure 7). The tally for the week of December 20^th will likely be surpassed by in the coming weeks as Covid-19 deaths are on track to exceed 700 per week (100 per day) for the foreseeable future. Tallies in excess of 800 deaths per week are likely in January.

Covid-19 deaths will also exceed 700 per week for the foreseeable future. The uniformly increasing mortality curve in Figure 7 should give us pause.

Using case counts to predict the course of the Covid-19 outbreak is intuitively obvious but statistically challenging. Case counts can provide misleading signals because of changes unrelated to viral transmission such as testing capacity, testing accuracy, testing uptake, and reporting delays. Figure 8, an accelerometer of week-over-week changes in the 7-day moving average of cases, illustrates this point. Until Thanksgiving, week-over-week changes tracked closely with long-term trends. Since Thanksgiving, week-over-week changes have oscillated widely making them much less useful barometer.

The second data item, I would like to revisit is serology testing. Earlier this summer, I discounted its usefulness due to sampling bias. However, given the breadth of this fall outbreak, I wanted to show them again (Figure 9). Serology positivity reached 36% this week establishing a likely upper bound on the percentage of Arizonans who have been infected. In September, an ASU-Maricopa County collaboration indicated 11% of Arizonans were seropositive leading to a 3.5X multiplier of reported cases to presumed cases. Given cumulative incidence is now 8900 cases per 100,000 residents, the multiplier would yield 31% seroprevalence. If these estimates are substantiated, they indicate we may be getting near the point at which natural immunity begins to act as a drag on the pace of viral transmission.

For the week ending January 10th, 8274 Pima County residents were diagnosed with Covid-19 (Figure 10). This represents an 11% increase from the 7470 cases initially reported last week and a new record. Test capacity and/or uptake has returned to mid-December levels. Last week’s initial tally of 7470 cases was upwardly revised by 2% (147 cases) this week.

Last week’s initial tally of 7470 cases was upwardly revised by 2% (147 cases) this week. Trends are similar across various age groups (Figure 11).