The new HHS website replacing the previous CDC data is displaying per-state hospital capacity data based on a statistical method outlined near the bottom of the page. Can anyone here speak to the effectiveness of the statistical method they chose?<p>Looking at AZ, their numbers are a good 8 - 10% lower than the AZ Department of Health Services numbers. And a quick glance at other hotspots like Florida, it seems their capacity numbers may be under-represented in the HHS data.<p>I don't know enough about statistical analysis to know if their method improves what may be imperfect data, or if it may be politically motivated to purposefully show lower numbers.<p>EDIT: Here is the method as outlined on their website:<p>Statistical Method: For hospitals that report both the total number of beds and the total number of occupied beds, calculate the percentage % of beds that are occupied. Next, weights are calculated for reporting facilities by taking the % of total beds associated to hospitals that report their numbers. Then, a total estimated number of inpatient beds is derived by dividing the total number of beds by the weight. Next, the number of unreported beds is derived by subtracting the number of reported beds from the total estimated beds to get the estimated number of unreported beds. Lastly, the total estimated number of occupied beds is calculated by adding the number of reported occupied beds to the expected number of unreported occupied beds (based on the binomial distribution). The above calculation is done by state and nationwide for all three of the metrics of interest.