Growth in Inflation-Adjusted Health Output per Capita over 80 Years

Footnotes

1. The same is true of measuring economic output in general, i.e., gross domestic product.
2. This is identical to the approach used to estimate gross domestic product.
3. For example, assume physicians are paid 3 times as much as LPNs in Year 1 and 4 times as much in Year 10. Assume that health output consisted of 1 million hours of physician time and the equivalent number of hours of LPN time in both years. Even though output is unchanged in this example, the measured size of the health sector would be 25% larger in Year 10 [1m. x (1 LPN wage unit) + 1m. x (3 LPN wage units) =4m. LPN wage units, whereas the same calculation substituting 4 LPN wage units to value each hour of physician time yields 5m. LPN wage units.
4. Another important difference is that the PCE implicit price deflator is essentially equivalent to a chain-type price index. The BEA observes "chain-type estimates provide the best available method for comparing the level of a given series at two points in time" (BEA 2013a). The chief limitation of a fixed-weight price index such as the medical CPI is that it used fixed price weights in a base year and measures pure changes in prices. In the preceding example, where the weights (relative output of physicians vs. LPNs) did not change, such a price index is perfectly adequate since so long as we use either Year 1 or Year 10 prices, we will deduce there has been no change in output between the two years. But in reality, if physician prices go up faster than LPN wages, there may be some substitution of nurses for doctors (to the degree that's possible), so the weights would likely change. Total weighted output conceivably could decline in this instance. Thus, a fixed-price index would erroneously conclude that the "medical cost of living" had risen by 25% when in reality--once both substitution of lower-cost LPN services for higher-cost MD services was taken into account--the actual increase in the cost of living would be more modest. In contrast, a chain-type price index allows for weights to continuously change (e.g., annually) (technically, an implicit price deflator is calculated as the ratio of current-dollar value to the corresponding chained-dollar value multiplied by 100; according to the BEA, "the values of the IPD and of the corresponding "chain-type" price index are very close." (BEA 2013a). This is why Congress is debating whether to use the chained CPI rather than regular (fixed-weight) CPI in calculating cost-of-living increases for government programs. It would be more accurate and would save the federal government a growing amount with each passing year.
5. The BEA calculates implicit price deflators for a large number of components of GDP, not just health care (BEA 2013b).
6. The Bureau of Labor Statistics has been measuring prices of a broad range of items since 1913 (BLS 2013).
7. AHRQ (2013) provides an extensive explanation of which medical price index is appropriate to use in different contexts.
8. According to AARP, the economic value of family caregivers was $450 billion in 2009 (Feinberg et al. 2011). Inclusion of this cost would have added 18% to NHE that year.