This from Sam Metz who has reviewed this and whose opinion I trust.
It investigates a phenomenon that I did not entirely appreciate before – the proportion of income devoted to healthcare across economic classes. One could use these data to demonstrate the value of tax-based financing of healthcare rather than premium-based financing. Premiums do not vary by income and in fact probably are regressive: more expensive for low-income earners. Premium subsidies can correct that burden to neutral, as demonstrated here. But only taxes can be made progressive: more expensive for high-income earners. It’s a good case for single-payer rather than private insurance.