China’s New Cooperative Medical Scheme, launched in 2003, was designed to protect rural households from the financial risk posed by health care costs and to increase the use of health care services. This article reports on findings from a longitudinal study of how the program affected the use of health care services, out-of-pocket spending on medical care, and the operations and financial viability of China’s township health centers, which constitute a middle tier of care in between village clinics and county hospitals. We found that between 2005 and 2008 the program provided some risk protection and increased the intensity of inpatient care at township health centers. Importantly, the program appears to have improved the centers’ financial status. At the same time, the program did not increase the overall number of patients served or the likelihood that a sick person would seek care at a township center. These findings serve as a benchmark of the program’s early impact. The results also suggest that the composition of health care use in China has changed, with people increasingly seeking outpatient care at village clinics and inpatient care at township health centers.

Clinical practice guidelines aim to help providers make decisions that optimize patient care (1). Both developers and users of guidelines understand that guidelines could be improved by tailoring the recommendations to the specific circumstances of an individual patient. Tailored guideline recommendations may improve health outcomes when a group of patients can be divided into subgroups in which the tailored recommendations would increase benefits, reduce harms, or save costs relative to more generic recommendations. For example, a guideline for hypertension may recommend antihypertensive treatment in patients whose blood pressure is higher than 140/90 mm Hg. However, patients with diabetes or high cardiovascular risk may benefit from receiving treatment at lower blood pressures, whereas other patients may benefit from less aggressive treatment.

However, for both guideline developers and clinicians, tailoring recommendations is easier said than done. Developers of guidelines face a difficult tradeoff: Issue simplified, easy-to-follow guidelines or, instead, guidelines targeted more precisely to each patient, at the cost of greater complexity and more challenging implementation. In addition, little evidence may exist to guide how to tailor recommendations, because trials may exclude patients who have the comorbid conditions and other factors that would warrant tailoring a recommendation.

Clinicians also may have far more information about an individual patient than even complex guidelines can accommodate. Should a patient with atrial fibrillation receive guideline-concordant anticoagulation therapy even if he or she is at increased risk for falls? Such …