Data Misinterpretation

In Reply Drs Goodman and Valenti raise the important issue that social spending likely affects health outcomes, and we agree.While the somewhat lower-than-average investment in social spending in the United States may not be driving higher health care spending, it is likely a driver of worse health outcomes. 1 As their letter notes, although many poor individuals in the United States may have access to excellent health care, they cannot easily access food or safe housing, factors that have a profound effect on health.Given that the United States has a much higher poverty rate than many of the other countries we studied, its lower-than-average social spending is particularly problematic for population health.Instead of justifying social spending as a way to reduce health care spending, social spending should be invested in for the most important reason of all: it likely improves health and well-being.Dr Bazemore and colleagues raise important questions about our estimates of the proportion of physicians in the United States who are practicing primary care.They rightly note that our estimates rely on numbers taken from the Kaiser Family Foundation. 2 These estimates differ from other estimates such as those of the Bureau of Labor Statistics or the AAMC.Although each approach has its strengths and weaknesses, our primary motivation was to use an approach that mirrored closely how other countries count primary care physicians.Because the Bureau of Labor Statistics excludes all self-employed physicians or physicians who are owners or partners in unincorporated practices, it likely leaves out a substantial proportion of the US physician workforce. 3Data from the other countries do not have such exclusions.The AAMC estimates that 34.8% of total active physicians were in primary care in 2015. 4However, it appears that the AAMC categorizes physicians whose specialty is not listed (n=83 637) as "other specialties." 5If we remove these unclassified physicians, the AAMC's estimate of primary care physicians is 38.5%.The AAMC recently began excluding hospitalists from its primary care categories, a reasonable decision. 5However, we chose not to do the same for purposes of comparability between countries.If we had also excluded all hospitalists, our primary care proportion would have dropped to 39%, which is nearly identical to the AAMC estimates.Of course, we then would have had to identify numbers of physicians functioning as hospitalists in other countries and excluded them as well.Those numbers were not generally available.There is no gold standard for measuring primary care, and each country does it a bit differently.Our approach was to create the most comparable set of numbers to understand how the United States compares with other countries.While the specific numbers for any one nation can be debated, if we were to take a different approach for the United States, we would have to apply the same filters and exclusions for all the other countries.Whatever approach one takes, using the lens of comparable numbers across nations, there is little evidence that the United States is an outlier in its mix of primary care and specialist physicians.