Today's Managing Health Care Costs Number is 1.5% |
Exchange plans have gravitated to narrow networks - and many have been encouraged that narrow networks allow higher value while being an important element of leverage with provider systems that have excellent reputations, high market leverage and command premium pricing. Most of these high-leverage systems are anchored by academic medical centers (AMCs). AMCs have higher costs than non-teaching hospitals because:
- They do teaching - which increases costs as trainees order extra tests or are less efficient
- They offer highly specialized services which have low volume
- They take care of the sickest of the sick - and even risk adjustment probably doesn't fully reflect the real resource costs of the care they provide.
Nonetheless, it's been hard to show any correlation between this higher cost and better quality or better outcomes -and narrow networks that restrict access to the highest priced systems have proliferated, and have seemed like a good idea.
Employer plans have followed - narrow networks are frankly much more attractive than additional cost shifting to members who already face substantially more out of pocket spending than they did a few years ago.
The Commonwealth Fund showed that in Massachusetts tiered networks were associated with 5% lower costs - with no apparent loss of quality. In Massachusetts, though, academic medical centers deliver an outsized portion of total care, and few narrow networks exclude all major teaching hospitals.
Lower metallic level plans across the country are increasingly offering no access to AMCs. This could be a challenge for people who have severe illnesses where AMCs are required. Presumably these plan do pay for AMC care for patients who need liver transplants or the most specialized of care -- but patients and their physicians need to jump through difficult hoops that are a special challenge to patients and their families confronting catastrophic illness.
JAMApublished research late last month that represents a challenge to narrow networks. Researchers reviewed over 21 million Medicare admissions from 2012-2014 to major teaching hospitals (6%), minor teaching hospitals (20%), and nonteaching hospitals (74%). They found:
Unadjusted 30-day mortality was 8.1% at major teaching hospitals, 9.2% at minor teaching hospitals, and 9.6% at nonteaching hospitals, with a 1.5% (95% CI, 1.3%-1.7%; P<.001) mortality difference between major teaching hospitals and nonteaching hospitals.
The difference persisted after risk adjustment - and was true regardless of the size of the hospital. (Austin Frakt, one of the researchers, summarized this in the NYTimes Upshot).
Ouch. This research looks solid. The sample size is huge (100% of the Medicare database). The mortality difference isn't from more aggressive diagnosis coding - and the differences are statistically significant in 13 of 15 top DRGs and 5 of 6 surgical categories.
This is Medicare data - the difference might not be applicable to the population under 65. Differences among hospitals within each category might still be more important than the difference between teaching hospitals and nonteaching hospitals. But I think this research poses a serious challenge to the effort, which I've vigorously supported, to move "community" medical care away from the major teaching hospitals.
I believe that adult inpatient care will eventually follow pediatric inpatient care. Very little inpatient pediatric care is done in non-teaching hospitals at this point. If kids are sick enough to be in the hospital they are in a teaching hospital - otherwise they are cared for at home. Many of the conditions that required multiple day hospitalizations when I was in training are now treated on an ambulatory basis. That trend will continue.
But in the meantime, these researchers have shown us a caution flag in the effort to move inpatient care away from the most expensive inpatient providers. This will make efforts to control unit prices even more important.
EmoticonEmoticon