In This Issue

December 2008 | Bradford H. Gray

The use of the methods and strategies of research can be beneficial in the care of patients who are not participating in research. That is the topic of the first article in this issue of The Milbank Quarterly, “What Ever Happened to N-of-1 Trials? Insiders’ Perspectives and a Look to the Future,” by Richard Kravitz, Naihua Duan, Edmund Niedzinski, M. Cameron Hay, Saskia Subramanian, and Thomas Weisner. Kravitz and his colleagues describe the benefits of bringing rigorous scientific methods to the care of individual patients with chronic diseases and how that can be done. They also explain why it does not happen routinely.

Kravitz and his colleagues note that the methods of clinical research are invaluable to determining which therapeutic alternatives work best, on average. However, not all patients respond the same way to the same treatment. This phenomenon is called the heterogeneity of treatment effects (Greenfield et al. 2007), and it is particularly important for chronic diseases for which multiple treatment options, often involving combinations of drugs, may be available. Although clinicians commonly use trial and error to determine which alternative works best for individual patients, this process generally is unsystematic, and the effects of different treatments often are ambiguous. This is the type of problem for which n-of-1 trials were designed.

Invented some twenty years ago by a group of researchers at McMaster University (Guyatt et al. 1986), n-of-1 trials use such tools as randomization, placebos, crossover methods, blinding, and the systematic measurement of outcomes to determine which therapeutic alternative works best for individual patients facing an extended period of treatment, as is the case with many chronic diseases. The patient, in effect, serves as his or her own control.

Kravitz and his colleagues show that n-of-1 trials have a unique capacity for establishing the best treatment for an individual patient. Based in part on interviews with n-of-1 pioneers and practitioners in Canada, Australia, and the United States, they provide both an intellectual history and an analysis of why, then, the approach has had only limited use. Resistance to change is one explanation, but some policy barriers are also involved. Some of these pertain to payment policies in medical care. Others arise from the fact that n-of-1 trials use methods normally associated with research, even though the purpose is not the development of generalizable knowledge, as research is defined in the Common Rule (Title 45, Part 46 of the Code of Federal Regulations), which governs research involving human subjects in the United States. Even though informed consent is essential in n-of-1 trials, the fact that they are not designed as research projects may lead to problems if institutional review boards assume jurisdiction.

Kravitz and his colleagues nevertheless are optimistic about the possibility of increased use of such an approach to care evaluation. They suggest that the growing interest in therapeutic precision, personalized medicine, and the acceptance of evidence-based decision making may stimulate interest in the n-of-1 approach, and they recommend ways to increase its use.

The next two articles in this issue of The Milbank Quarterly— “The Relative Merits of Population-Based and Targeted Prevention Strategies” by Donna Zulman, Sandeep Vijan, Gilbert Omenn, and Rodney Hayward and “Revisiting Rose: Comparing the Benefits and Costs of Population-Wide and Targeted Interventions” by Jennifer Ahern, Matthew Jones, Erin Bakshis, and Sandro Galea—both offer analytic approaches for exploring the important policy issue defined by the late Geoffrey Rose (1985, 1992): the relative merits in disease prevention of population-wide approaches and interventions targeted at high-risk individuals. Both groups of researchers used simulations to assess the two approaches to preventing cardiovascular disease, and both focused on the variables determining the relative effectiveness or cost-effectiveness of targeted and population-wide approaches.

Zulman and her colleagues concentrated on high- and low-intensity population-based approaches and two targeted approaches aimed at reducing low-density lipids. Using outcome measures that included cardiovascular events and mortality, as well as quality-adjusted life years, they found that although relative merits varied according to which strategies and outcomes were compared, the risk-targeted approach was superior in reducing the number needed to treat to prevent a cardiovascular event or death. Their results were particularly sensitive to the presence of treatment-related adverse effects. Zulman and her colleagues conclude with an overall analysis of the two prevention strategies, including a discussion of the implications of their approach for the development of practice guidelines to prevent cardiovascular disease.

Ahern and her colleagues analyzed population-wide and targeted interventions to reduce blood pressure to prevent cardiovascular disease. Their research focused primarily on cost-effectiveness, using estimates of the cost of treatment and the average benefit per case of disease prevented. They discovered that the comparative cost-effectiveness of different approaches was sensitive to the cost of the intervention and the extent to which the intervention was targeted. After presenting the results of their simulations, Ahern and her colleagues discuss several issues in assessing the relative merits of various approaches to reducing risk and preventing disease.

Preventable disease is also the underlying theme of the next article, “Helping Smokers Quit: Understanding the Barriers to Utilization of Smoking Cessation Services,” by Sarah Gollust, Steven Schroeder, and Kenneth Warner. Their concern is that despite the terrible health consequences of smoking and the growing availability of ways to help smokers quit, cessation services are used less often than might be expected. They cite barriers to the use of such services in many parts of the health care system—health plans, employers, physicians, patients, and government. The article ends with a discussion of policy alternatives at both the federal and state levels.

This issue of The Milbank Quarterly concludes with an article about health care costs. It is well known that the United States spends much more per capita than any other country, but with poorer results by measures such as mortality and citizen satisfaction. This suggests that there is much waste in the system. But where? In “Waste in the U.S. Health Care System: A Conceptual Framework,” Tanya Bentley, Rachel Effros, Kartika Palar, and Emmett Keeler distinguish among administrative, operational, and clinical waste; summarize the evidence about these types of waste; and analyze the advantages and disadvantages of different approaches to addressing them. The right incentives will be helpful, they say, but more knowledge and better tools are needed as well.

Bradford H. Gray
Editor, The Milbank Quarterly


Greenfield, S., R. Kravitz, N. Duan, and S.H. Kaplan. 2007. Heterogeneity of Treatment Effects: Implications for Guidelines, Payment, and Quality Assessment. American Journal of Medicine 120(4, suppl.1):S3–S9. Abstract available at (accessed August 28, 2008).

Guyatt, G., D. Sackett, D.W. Taylor, J. Chong, R. Roberts, and S. Pugsley. 1986. Determining Optimal Therapy—Randomized Trials in Individual Patients. New England Journal of Medicine 314(14):889–92. Abstract available at (accessed August 28, 2008).

Rose, G. 1985. Sick Individuals and Sick Populations. International Journal of Epidemiology 14(1):32–38. Available at (accessed August 28, 2008).

Rose, G. 1992. The Strategy of Preventive Medicine. Oxford: Oxford University Press. Available at (accessed August 28, 2008).

Author(s): Bradford H. Gray

Read on Wiley Online Library

Read on JSTOR

Volume 86, Issue 4 (pages 529–532)
DOI: 10.1111/j.1468-0009.2008.00538.x
Published in 2008