Poverty, Primary Care and Mortality: The View from Across the Pond

It is an article of faith that, in Barbara Starfield’s words, adults whose regular source of care is a primary-care physician rather than a specialist have lower mortality, even after accounting for differences in income, and she draws upon studies at both the county and state levels to prove it. Now a new paper in JAMA about England’s Primary Care Trusts refocuses the discussion on poverty.

While Starfield’s county-level studies are often cited as evidence that more primary care physicians and fewer specialists lead to lower mortality, they actually showed virtually no differences at all. And when repeated by Ricketts, the small differences noted were not consistent throughout various regions of the US. On the other hand, “counties with high income-inequality experienced much higher mortality.” So, in reality, the county studies demonstrated the strong impact of poverty and the marginal impact (if any) of primary care.

But there is a problem associated with measuring the association between physicians and mortality among counties. Patients in one county often receive care from physicians in another. This same limitation does not hold for measuring the association between income and mortality. They are both measured in the same person, who is in the same place.

States offer the opportunity to collect physicians and patients within units that generally include both, although the large size of states poses other limitations. In examining the effect of primary care at the state level, Starfield found that more primary care physicians were also associated with lower mortality, but that was true only for family physicians. More internists and pediatricians offered no such advantage. However, as in her county studies, she also found a strong association between greater income-inequality and higher all-cause mortality.

Why was the association of lower mortality and more primary care physicians limited to family practice? The answer is that family physicians are mostly in the upper-Midwest, where there are few blacks (whose mortality is double that of whites) and even fewer poverty ghettos (where mortality is greatest). The east and south, which have plenty of both, have preferred internists. The association of more “primary care” (aka, family practice) with lower mortality is a geographic artifact. Indeed, there is an even stronger association between lower mortality and more snow!

In this week’s JAMA, Levene and his colleagues have reported their study of variation in mortality among England’s 152 Primary Care Trusts, which, of course, collect patients and physicians together. They found that the strongest predictors were socioeconomic deprivation, race, smoking and diabetes. Neither the number nor the clinical performance of primary care providers predicted mortality.

The consistent theme is that poverty matters and primary care does not, at least for mortality. But everyone knows that primary care matters anyway, and so does specialty care. And both function better when they work together, and they function best when there are sufficient numbers of both kinds of providers. Yet, we are on a path toward increasing shortages across the board.

Health care reform is off the track for many reasons, not the least of which is that the role of primary care has been so misrepresented and the drive to denigrate specialists has been so vigorous. The tasks at hand is to end the “verbal propaganda” that has so divided disciplines and concentrate on assuring that the supply of physicians and other advanced health care professionals, both specialists and PCPs, will be sufficient for the future. And the broader task is to build the social infrastructure of housing, education and more that allows the poor to function and stops the cycle of poverty that plagues America.


  1. Michael Lyon

    Joshua Horn’s book “Away with All Pests,” describing medical care in China shortly after the revolution, presents a picture that is similar to Cooper’s idea about primary care docs and specialists.

    After the revolution in China, debate raged over whether the chosen medicine should be Western medicine, which was considered science-based but was associated with the attitudes of the rich people who could afford it, or traditional medicine, which was based on folk-knowledge but was trusted by the people. Health people argued back and forth, until they came to the realization that they didn’t have enough of either kind of practitioner, and they needed to use everyone they had.

    Once that decision was made, they decided to mix traditional and western practitioners in each of the health stations, clinics, and hospitals, so they could learn from each other, and learn which situations were best for which forms of practice. In Horn’s medical work, trauma, they discovered compound fractures were best treated with western casts which totally immobilized the break, but minor fractures were best treated with a traditional wrapping which allowed a small amount of movement. Medical Science was also transformed, and became a search for the biological mechanism by which known treatments worked, with the aim of improving them.

  2. Louis Levene

    As the lead author of the JAMA paper you refer to, may I make a few comments?

    Whilst most of the variation in CHD mortality rates in our model was explained by the 4 population factors you mention, you must not forget the 5th factor: that areas with higher rates of detection of hypertension had lower CHD mortality rates, so we believe that primary care still does have a role to play.

    White ethnicity was a positive predictor in multivariable analysis, but in England, non-whites are more like to have diabetes and live in deprived areas. In univariable analysis South Asian ethnicity rates correlated positively with CHD mortality rates.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s