In some cases, it is worth considering the reverse situation (a higher % of income becoming taxable, or a higher than expected tax rate applied to the same income), and if applicable this could argue for the Roth conversions in early retirement even if it pushes the person into a higher tax bracket. The most common example might be a MFJ couple and the ramifications of one partner passing away. The drastically reduced standard deduction and exemptions (= higher taxable income on virtually the same RMD) and the reduced tax brackets (=higher tax rates on that taxable income) can be a real eye opener. Converting a big chunk of money to a Roth while both partners are still alive might prove to be a really smart move. But, a single surviving spouse in most cases would be expected to have lower expenses than 2 people, so it would be an unusual case in which these tax issues would significantly hurt. Still, worth considering for some people.
In addition to the case you cite (higher medical expenses), retirees could also find themselves in low tax brackets if their investments do poorly over the coming decades. It could clearly happen to most of us, and it would reduce the absolute value of RMDs and also of any cap gains from taxable accounts. I'd sure feel dumb if I went crazy and converted a lot of funds, paid tax at the 25% rate, and then our investments hit the skids and I found I'd be in the 15% bracket for a long, long time. That extra money paid in taxes earlier might come in handy.
I plan to do Roth conversions to the top of my "regular" expected bracket in early retirement (before I start taking SS). After that, I'll leave things alone and risk paying higher taxes well down the line --hopefully when it is apparent I've "won the game" and my nest egg is really going to outlast me.
Bogle along with several other well regarded investment "Gurus" have repeatedly made the point that investment returns for the typical diversified portfolio are likely to be considerably lower going forward for both stocks (given current valuations) and bonds (given current yields) and that this situation may well extend for a decade +. I am skeptical of an approach that predicates payment of large taxes now in anticipation of significant future growth of a tax free portfolio. Also, for those of us in high income tax rate states such as Oregon or California the Fed 15% rate is really more like a 25%+ when the state income tax is taken into account.