Shopping for HI on the marketplace can be quite different from the limited choices that your company may have offered. Here I'll outline how I proceeded.
I started out with the spreadsheet in this link:
http://www.early-retirement.org/foru...ml#post1487700
There is also another one in this link,
http://www.early-retirement.org/foru...ml#post1510282, which I haven't tried.
But if you look at the inputs, you'll see that you'll need to specify how much and what kind of services you expect to use, so I had to step-back and look at history.
With my family of four, I saw that over the past 22 years, we really went to the doctor very few times. The records were spotty because of how things got paid for, but I did keep track over the years of every visit, even if it didn't have an associated payment (for instance preventive stuff, dental cleanings with the company plan, etc). But whatever YOUR case, you need to look back in order to look ahead for your expected usage.
Then I created a second scenario where somebody needed to go to the hospital on top of our regular usage. I presumed in this scenario that the person incurred expense beyond the out of pocket max.
I got two approximate values for a few categories of policies. The value was the sum of the premiums, plus the amount I'd be paying for services. Then I'd use a rough likelihood to get a single value.
So, for instance, selecting a policy that pays nothing until the deductible is reached and the premiums were $1000/mo, and I'd have 10 doctor visits at $135, that expected cost would be $13,350. And then for the SAME policy, I'd run a case for all of the above, plus, say, $10,000 for a hospital bill (but we'd hit the out of pocket maximum, so I'd only spend $5000). Then I'd guess that it's 5% likely we'll need hospitalization during the year, so that policy would have an expected cost to me of .05*17000 + .95*13350 = $13,532. This is the comparison number.
Then I would take another policy, say one that had a similar out of pocket limit, but covered 5 doctor visits a year in full, then a 20% copay for doctor visits. Since the math gets complicated, I just assumed that all visits would be free (giving this analysis an upward boost in the rankings). But still, if the expected scenario happened (10 doctor visits for free), the additional premiums on the policy exceeded the value of the free doctor visits. For instance, if the premiums were $1200 per month, then the expected value would be .05*19400 + .95*14400 = $14,650.
Since the second policy (free doctor visits) has an expected cost that exceeds the first policy (pay 100% up until deductible), I'd choose the first policy.
In this example, I had two scenarios. You could have more, so you might say scenario1*.80 + scenario2*.10 + scenario3*.10, but that was too much work, even for me!
The hardest part for me was pricing the scenarios against the various policies. This requires that you can translate between the words on the summary of benefits and how that affects what you pay for services. For me, with light utilization, the policy that paid nothing until the deductible was reached was hands-down the best bet. In fact, I'm going to research the over 30 catestrophic policies this year because what I'm buying is the out of pocket max, and they're all capped! So for me, at this juncture, no expensive meds, no frequent doctor visits, the insurance is only for "if something unexpected and bad happens".
On thing to remember when you're doing your historical medical utilization is that you need to subtract out anything that is considered "preventive" under PPACA; no matter what policy you get, that's not something you lay-out cash for.
So, there you have it, it's not easy. This will never happen, but it would be cool if you could build likely scenarios and drop each scenario on various policies and get an expected value for that scenario. Sort of like the example that's on the summary of benefits, but dynamic (no, I don't care how much it costs to have a baby!) At this point, you have got to know all the mechanics and slog through it.