On April 5th, I wrote that that the COVID-19 fatality and other data reported by the California Department of Public Health since the inception of the outbreak had demonstrated that California had flattened its COVID-19 curve in March. Just two days later, the data I was waiting to see have now revealed that the curve is not just “bending” as Governor Newsom stated yesterday, but very likely bent – just as I predicted it would be weeks ago. Accordingly, I am updating my prior Post with the: (1) good news arising from the latest fatality data; and (2) the great news reflected in the hospital admission data (the best data for assessing whether the State has achieved its goal of flattening the curve to the point where our healthcare system can consistently handle COVID-19 cases while providing the highest quality care to all patients).
The Good News: Daily Fatalities are Flat
Epidemiologists have stated that the most accurate way to chart the course of the pandemic in California is to examine the fatality rate. Accordingly, the chart below provides the 5-day moving average for fatalities since the pandemic began.
The COVID-19 fataltiy curve in California is now flat.
As I noted previously, the most important fact to understand in reviewing this chart is that a fatality reported today reflects an infection that occurred 15-25 days earlier. The data now clearly show that sometime roughly between March 8th and March 18th, the pandemic began to slow dramatically, producing an obviously flattened curve beginning on April 2nd. This flattening likely corresponds to the impact of the late stages of the decentralized public health measures that began across California on March 8th. In short, the data strongly suggest that these early public health measures substantially curtailed the spread of the virus. Once the full impact of the stay at home order is reflected in the fatality data, we should expect to see the curve start bending downward over the next several days.
The Great News: Daily Hospitalizations are Down
Speaking of bending downward: The State began reporting hospitalization data on March 31st. Before I get to the great news, as I need to explain the State’s opaque reporting system:
There are two classes of hospitalized patients: confirmed COVID-19 patients and “suspected” COVID-19 patients. The State reports both because “suspected” patients are utilizing precious COVID-19 healthcare resources just like confirmed patients. Once a “suspected” patient is confirmed to be positive, he or she is then (presumably) included in the confirmed hospitalizations count. If a “suspected” patient is confirmed negative, he or she is (presumably) no longer counted as a COVID-19 hospitalization (but still requires some level of care until discharged). On any given day, if the State receives more testing results for existing “suspected” patients than the number of new “suspected” patients it admits, the daily total of patients in the “suspected” category will decrease. We have seen this dynamic over the last 6 reporting days – “suspected” patients have decreased from 3481 to 2796.
Accordingly, to accurately depict the daily increase or decrease of COVID-19 hospital patients in California, we need to add the sum of the daily change in confirmed cases (newly admitted patients who have tested positive plus previously admitted “suspected” patients who have since tested positive minus discharges) to the sum of the daily change in “suspected cases” (newly admitted suspected patients minus patients for whom test results have been received) to generate the “net hospital admissions” per day for COVID-19. This chart shows the downward trajectory of net daily COVID-19 hospital admissions in California since the State began reporting this data:
The daily COVID-19 hospital admissions curve is now bent.
As you can see, daily hospital admissions are clearly dropping. This is really good news. At the same time, COVID-19 patients are being discharged every day as well. Discharges are going up as the rising hospital admissions in late March turn into discharges in early April. When daily admissions equal daily discharges, we have completely flattened the hospital bed curve. From the publicly reported data, I estimate that this flattening occurred between March 30th and April 3rd. Again, more really good news. The really great news: Since then, the curve has almost certainly been bending downward – discharges should now be outpacing daily admissions, and we will be able to confirm this in a few days.
The stated goal of the stay at home order was to flatten the daily hospital admissions curve to the point where our healthcare system could provide the highest quality care to both COVID-19 and all other patients by ensuring that new admissions never exceeded the number of discharges on a given day. The State’s early predictions were that we would need more than 50,000 beds to achieve this goal, with thousands of new admissions each day during the peak of the pandemic. We appear to have done it with roughly 6000 beds (including beds for “suspected” patients who later test negative) and less than 500 admissions per day, just as I predicted we would. We have more than “met the moment.” We should be enormously proud of our collective achievement and deeply grateful to our healthcare professionals.
Why am I confident that this is not just a blip on the radar, and that we have, in fact, bent the most important curve? Because both new confirmed admissions and “suspected” admissions are trending substantially downward, exactly when I predicted the full impact of the stay at home order would be manifest in the data. The curve depicted above reflects the progression of infections that likely began during the first days of the stay at home order as compliance was increasing. Over the next several days, we will see the full impact of the stay at home order in the data, and I expect to see continuous, but uneven, deceleration in the virus. In Italy, the most recent data show the nationwide lockdown has led to a continuous deceleration in the trajectory of the virus over the past several days. I am confident we will see the same here, which is why I am publishing this Post today.
Why does the State keeping saying the virus will peak in May? I believe the most likely explanation is that State is using models, like the one it used initially to predict the uncontrolled spread of the virus over 8 weeks, that: (1) fail to fully account for the effectiveness of the various responses to the pandemic that preceded the stay at home order; (2) disregard, inadequately weigh, and/or belatedly utilize the actual data regarding the trajectory of the pandemic in California in favor of estimates and theoretical assumptions (that may not apply neatly to California); and (3) ignore valuable data from other jurisdictions (e.g., the deceleration rate of the virus in Italy). I do not see any evidence in the publicly reported data to date to support an assertion the pandemic will peak in May, but I will update my analysis as warranted by any new data reported by the State.
What does this mean going forward? This is a question we need to put squarely to our elected officials. We have done what they asked of us. Now they must do the same. The data show that this moment is drawing to a successful close. We need to start preparing for the next one while we continue to clobber this pandemic together.
Be safe. Stay healthy.
April 8, 2020
Nice to see some fair, logical reporting. Of course, if everyone went back to work tomorrow (including large events like sports, etc.), we would probably see a dramatic increase. I think a slow, steady return to normalcy is the way to go, all the while encouraging safe behavior (hand washing, avoidance of large gatherings for a little while, etc.)
GREAT WORK!!
April 9, 2020
I agree with your conclusion. The return to something resembling normalcy is going to need to be slow and steady and carefully controlled. Social distancing measures exist on a spectrum from not shaking hands to stay at home orders. I’m not sure we’ll ever go back to shaking hands. But I am sure it is time to start transitioning away from the stay at home order in California, just as Austria is doing now (and Italy soon will be). Thank you for the compliment on my work.
April 9, 2020
Brian, what do you mean by your last paragraph by saying “Now they must do the same.” re: our public officials. You follow that with ” The data show that this moment is drawing to a successful close. ” Are you advocating an early relaxation of the social distancing protocols?
I have been sharing your projections as evidence that the protocols are working but urging people to stay vigilant in following the protocols as set by our state and cities.
April 9, 2020
Thank you for tracking my analyses and your question. In short, the stated purpose of the extraordinary stay at home order was to ensure that California had adequate healthcare capacity to withstand the peak of the COVID-19 pandemic. As you may recall, the Governor estimated we would need 50,000 or more hospital beds to achieve this goal. As of today, the data strongly suggest we will need about 6,000 beds at most so long as the stay at home order is in place. The data also strongly suggest that we had quite possibly flattened the curve with the public health and social distancing measures that were in place on March 18th before the stay at home order was issued. Given that we have achieved our goal with enormous room to spare, I believe California, like Austria and Italy (for example) is at the point where it needs to begin planning for the next phase in managing the COVID-19 pandemic. This will likely include a gradual loosening of certain public health measures and the maintenance of many others. I believe the State should be preparing such a plan and presenting it to the public in the very near term.
April 9, 2020
I saw a jump in covid deaths in Ca. today to 61. Is this significant?
April 11, 2020
No. Please don’t read too much into one day of data. You need to look at data over 3, 5, or 7 days. I believe fatalities have peaked. We’ll know for certain fairly soon.
April 11, 2020
Yes, now I understand the 5 day rolling average. But being in the “at risk” generation I can’t help but feel it is too early to think of relaxing the protocols that have worked so well this far. I feel that the people who are on the fence about obeying the protocols would only need a small nudge to resume a normal life.
April 9, 2020
Great analysis of what has happened in California. HOWEVER, the excellent efforts in social distancing which have led to a death rate per capita of 30% of the national average implies that the acquired immunity in the State must be significantly below the national average. What will happen when people start encountering others when the stay-at-home orders relax is anybody’s guess but will likely have significantly bigger impacts in states like California and Texas than ir the rest of the US. Deaths in California dropping below 5 by May 7 in California as predicted by the model (8 worst case) are totally unrealistic.
April 16, 2020
Thank you for your comment. It will be very interesting to see what the Stanford and USC studies tell us about the presence of antibodies in the general population. In the interim, I believe our transition out of the stay at home order will need to be carefully-crafted with careful monitoring, as I wrote today.