Work in Progress
with David C. Chan, Nancy L. Keating, Bruce E. Landon, and Michael L. Barnett
Working Paper · NBER SI Presentation
Abstract
Specialization can raise productivity, but its benefits depend on access to the relevant expertise. In oncology, subspecialization—the narrowing of clinical focus within cancer care—has become increasingly common. This paper estimates the impact of physician subspecialization in medical oncology on mortality, Medicare spending, and access to innovation. Using US Medicare claims covering 2.2 million first chemotherapy episodes from 2008 to 2020, we exploit quasi-random variation in access to subspecialized oncologists using a differential-distance instrument. Access to a subspecialist reduces three-year mortality by 4.5 percentage points (10 percent relative to the mean) without increasing total Medicare spending per episode. Mortality gains emerge at the end of the first year following therapy initiation, and the estimated effects are comparable in magnitude to survival improvements reported in some major recent oncology drug trials, despite arising from differences in physician specialization and care delivery rather than a specific therapeutic intervention. Subspecialist care also increases the use of newer chemotherapy agents and more than doubles enrollment in clinical cancer trials, particularly trials aligned with the patient’s cancer type. Falsification and selection analyses support the identifying assumptions and suggest that the estimated gains reflect differences in treatment pathways rather than patient selection. These findings provide new evidence on the productivity effects of specialization in a non-routine, knowledge-intensive profession and highlight organizational trade-offs in the delivery of complex care.
Working Paper
Abstract
The establishment of hydroelectric power plants resulted in a rapid structural transformation of Norwegian municipalities around the beginning of the 20th century. Using a novel dataset linking individuals born between 1890 and 1910 to historic death data, I find that experiencing childhood in rapidly transforming local areas leads to an increase of ten months in age at death for men. This effect is entirely driven by individuals born into higher socioeconomic status households. I find that incomes, manufacturing, immigration and economic inequality in local areas in the short/medium-term increase after the introduction of hydropower, while public health deteriorates at the same time. This suggests that, in the long term, economic development through structural transformation outweighs the negative consequences of a deteriorating public health environment and thereby increases the lifespan of individuals.