• I still think it’s outpatient care

    Cost chart week continues. I’ve posted graphs with outpatient and inpatient care, but people want more.  So here are more data on how US spending has changed in a number of areas over the last few decades (all in US$ at 2000 PPP rates):

    I’ve added expenditures on home health care (yellow), pharmaceuticals and other non-durable goods (red), and durable goods (purple).

    One of these things is not like the other.  For the most part, growth was consistent for inpatient and outpatient care until the 1980s.  Growth in the other sectors was much slower.  Starting in the early 80’s, though, outpatient care takes off in a much different way than other areas.  Outpatient spending increased in the 90’s even when other areas flattened.  By 2000, drug spending took off too.

    One argument is that outpatient growth is due to a massive shift from inpatient care.  We do know there was a shift in surgical setting in the 1980s and 1990s. I’m not seeing a very large reduction in inpatient spending though.  Yes, inpatient spending slowed slightly when outpatient care took off, but not enough to compensate for the steep green line there. To argue for a huge shift in volume from inpatient to outpatient, one would have to also find a big increase in inpatient prices to keep total inpatient costs growing at nearly the same rate. Is there evidence for large increases in inpatient prices in the 1980s?

    Inpatient prices paid by private plans did go up in the late 1980s as Medicare’s prospective payment system phased in (this was the golden era of cost shifting). This wouldn’t explain the early 1980s though. All in all, it seems that there was probably an increase in outpatient volume that did not substitute for inpatient care. I have no good answer as to why this would have occurred at that specific time and not before. Did something happen in the organization, politics, and economics of medical care in the US in or just preceding 1982?  Was there some other change?  Someone ought to know.

    Share
    Comments closed
     
    • It seems pretty simple to me. That is that era of the proliferation of managed care plans with no deductible and low co-pays for outpatient services. These conprehensive first-dollar coverage third-party payment plans replaced major medical indemnity policies that applied highdeductbles to outpatient services and did not routinely process or review outpatient claims. The introduction of third-party payment dramatically increased both demand for outpatient services and the administrative cost of processing outpatient claims for both payers and providers. Physician practices, for example, increased their non-physician employees from an average in 1980 of 2 per physician to the current average of 5 non-physician employees per physician. The added employees are necessary to handle administrative tasks imposed by third-party payers. Note that Medicare also changed to third-party payment during this time period.

    • Aaron, I certainly agree that a portion of the outpatient growth did not substitute for what would have been incurred as inpatient services.

      But let’s say that portion was as high as 50%. That means 50% would have been inpatient before the market shift and we should see a corresponding drop. We do see a substantial drop between 1980 and today. though I don’t have data on the exact timing. Despite that total inpatient costs kept going up, as you note.

      Some combination of the following must have happened for inpatient: 1) new business partially replaced that lost to outpatient care and 2) prices increased faster than utilization declined–not on a service by service basis, but overall.

      So the question is, how large are these two effects and how do they compare to the size of pure outpatient growth that wasn’t just a transfer from inpatient?

      I exclude inpatient-to-outpatient migration spending because I think it can be misleading to treat this as cost “growth” when attributing blame for the 80s. After all, outpatient is supposed to be cheaper than inpatient, so whatever that migration portion is of the outpatient curve, it should actually be conceptualized as a decline vs what would have been total healthcare spending. The real difficulty here is estimating the size of a counterfactual,

      in any case, once all the numbers are run, my money is still on inpatient cost growth (compared to what would be expected given the loss of businss to outpatient) being nearly as large a factor as outpatient growth. Thus, for now at least I think this was and remains a facility cost problem rather than outpatient alone.

      As for Donna’s comment about managed care, she is off by a decade. HMO/PPO took over the market between 1990 and 1995.

    • My guess is there are two things going on here:

      [1] One, as you mentioned at some point, is a secular component related to technological improvements (in anesthesia, minimally invasive surgical techniques, and etc.).

      [2] The other is the enactment of case-based cost restraints in TEFRA (1982) and the PPS (1983).

      The cost limits in TEFRA applied only to inpatient services. Outpatient costs continued to be reimbursed based on some type of retrospective cost-based system until the Outpatient-PPS began in 2000. Consequently, hospitals had a strong financial incentive to shift procedures to the outpatient setting.

      I found a number of contemporary studies that discuss this issue. Sulvetta (Achieving Cost Control in the Hospital Outpatient Department, Health Care Financing Review, 1991 Annual Supplement, p. 95-106) looks at the issue from a PPS perspective. Chulis (Assessing Medicare’s Prospective Payment System for Hospitals, Med Care Res Rev, Summer 1991, p. 167-206) looks at it in the context of PPS performance.

      In particular, Table 5 in Chulis (p. 193) shows that inflation-adjusted Medicare payments for outpatient services increased 74% between 1984 and 1988 compared to a 3% increase for inpatient services.

      The 1984-88 change in outpatient costs in your graph looks to me to be about 40%, so it seems reasonable that the shift in Medicare services to the outpatient setting can explain a significant part of the outpatient “hockey stick” effect.

      The story would be something like the following. Technological improvements permitted shifting inpatient procedures to the outpatient setting. TEFRA and PPS supplied financial incentives for hospitals to initiate a shift to the outpatient setting. Once the process was set in motion, additional technological innovation and the implementation of managed care systems in the private sector sustained the trend.

      • @Rob – The paper I was writing for which you supplied some key early Medicare payment information goes over hospital PPS based on the work of Rick Mayes (good stuff, worth reading). I’ll share my paper with you and the world when it is out. From my research, it seems from that that PPS didn’t have a big impact for a few years due to a gradual phase in and fairly generous payments based on historical costs. My guess is that it may have exacerbated or supported some shift to outpatient that was already in the works due to some forces a few years prior. Maybe the writing was on the wall? Maybe technology is the key? Maybe it was a shift to “corporate medicine” (a theme from Starr’s book). I think there are still a few dots to connect to tell a clear story from 1980 forward.

    • Austin:

      Thanks for the link to the Rick Mayes papers — I will definitely look at them. Also, as I recall, the 1991 Sulvetta paper seemed to emphasize the role of TEFRA more than PPS in the shift to outpatient. The idea being that TEFRA placed cost-per-case limits on inpatient reimbursement but not on outpatient. I think she might agree with you that PPS played mainly a supporting role.

      I’m a big fan of Starr’s book. Had the AMA been less successful in the 1920s and 30s in constructing bulwarks against corporate medicine, it seems to me that we might have seen ACO-type integration begin to take hold in the late 1970s and 80s.

      I look forward to your paper.

    • @Rob – I can’t find the Sulvetta paper online, just references to it. Do you have a copy to share?

    • Having lived and practiced through that period, a few points come to mind. First, PPS capped the inpatient payments, so there was great pressure to get patients out of the hospital ASAP. Many costs that would have been in the inpatient box previously were shifted to outpatient, so the outpatient box exhibits most of the overall growth in spending. Inpatient v. outpatient is really a pretty artificial division–many expenses can be shifted from one to another.

      Second, the push for discharge from the inpatient setting let to a change in medical practice. The early 80’s were the inflection point where inpatient care changed from a relatively relaxed, exploratory process (diagnosis could take a week or more; patients were kept in the hospital to ‘rest’, etc.) to one that was hell-bent to get things done fast. As a result, instead of pursuing lines of diagnosis sequentially, everything in the differential was pursued at once. Even as the duration of an inpatient stay decreased, the overall amount of ‘work’ and cost per stay increased dramatically.

      Third, the 80’s were an inflection point in diagnostics, especially radiology (CTs). It was interesting to compare the practice in a VA (at that time still in late 60’s, 70’s mode) with a ‘modern’ teaching hospital. At the VA there was one blood draw a day; very limited X-rays, etc. At the teaching hospital a ‘regular’ bed had diagnostic services equal to the ICU at the VA: there were routine blood draws 4 times a day; bedside serial X-rays, etc. I doubt there are many hard statistical data points, but if you look in the medical literature there are many discussions about the escalating ‘acuity’ of inpatients.

      Fourth, the early 80’s were the beginnings of AIDS. I am not sure how much this affected the statistics, but it tremendously affected inpatient care, from a marked increase in expenses for protective gear and processes, to large increases in very expensive, long ICU stays.