Double-blind "N of 1" crossover trials—in which a patient alternates courses of active drug and placebo—are rarely used in clinical practice. But as a recent methylphenidate study illustrates, this approach can provide objective evidence that an individual patient is truly benefiting from a particular medication, rather than from nonspecific effects of treatment. Such trials may be particularly helpful when physicians, patients, or family members have concerns about a drug's efficacy or safety profile.

Peter Camfield, MD, FRCRP, who coauthored the methylphenidate study, believes that "N of 1" trials are greatly underutilized in patients with attention-deficit/hyperactivity disorder (ADHD). "I don't think you should start methylphenidate except in this kind of setting," he said in an interview. "I don't see a reason to do it any other way." Other medications—particularly those that have quick onset and offset—may also be suitable for the approach, he said.

Dr. Camfield and his wife, Carol S. Camfield, MD, FRCPC, both of whom are professors of pediatrics at Dalhousie University Medical School in Halifax, Nova Scotia, have made "N of 1" trials a routine part of their clinical practice. Families considering methylphenidate treatment receive three types of capsules, all identical in appearance: two dosages of methylphenidate (0.3 mg/kg and 0.6 mg/kg) and placebo. Each treatment, given double-blind, is tried for one week in random order; parents and teachers complete rating scales to assess the child's behavioral symptoms during each trial. At the end of the trial period, the Camfields meet with the families and reveal the identity of each week's treatment; families can then decide which methylphenidate dosage offers the best combination of maximum efficacy and minimal adverse effects—or they may choose to try another treatment instead.

"PARENTS FOUND IT HELPFUL"

In the December issue of the Archives of Pediatric and Adolescent Medicine,the Camfields, together with Mary Ann Kent, MD, described their experiences with 50 such families. Of the 43 families available for follow-up interviews, 37 completed the crossover trial, and 31 children responded to methylphenidate. Twenty (65%) of the responders were still using methylphenidate one year later, while eight responders and three nonresponders remained on the drug after the trial but discontinued it within a year. Only six families—those of three responders and three nonresponders—opted not to use methylphenidate at all.

"The families really like this approach," Peter Camfield said, noting that all of the families took the trial results into account when deciding whether to continue the medication for their child, and more than two thirds based their decision exclusively on the results. "Even in cases where the child was not a responder to methylphenidate," added Carol Camfield, "parents found it helpful to know this in a scientific fashion. Because parents can really understand this."

Of course, double-blind "N of 1" trials are useful for drugs other than methylphenidate. The Camfields and their colleagues have also used the approach to test the effects of melatonin in children with fragmented sleep, and Peter Camfield tried a trial of imipramine in a child with self-abusive behavior. The literature also includes reports on the use of the technique with oral ketamine (for chronic pain), amitriptyline (fibromyalgia), and pemoline and dextroamphetamine (ADHD).

Another opportunity for "N of 1" trials is with treatments for which efficacy is not well established, such as alternative therapies. "People were asking me to [try treatments] that made me feel like a quack," said Ellen Wiebe, MD, explaining why she used an "N of 1" approach for a patient who wanted to see if vitamin B12 injections would improve his chronic fatigue symptoms. (They didn't.)

However, another chronic fatigue patient had a "magical" response to nimodipine, said Dr. Wiebe, who is a clinical professor of family medicine at the University of British Columbia. In this case, she used the "N of 1" approach because she wanted to be sure that the medication's benefits justified its high price.

POTENTIAL PITFALLS

As in large-scale randomized trials, clinical "N of 1" studies must be properly designed to yield valid information. There are several keys to success with these trials:

• Be sure the drug and disorder are suitable for this approach. Medications that are quick-acting and have short half-lives allow time for multiple crossovers, increasing the validity of the results. In some cases, four trials each of medication and placebo may be desirable. "It depends on how dramatic and measurable the response is," said Peter Camfield. In addition, the technique shouldn't be used if placebo trials put the patient in danger or cause undue suffering.

• The patient should be unable to tell the placebo from the active drug. To conceal the identity of the treatment, the Camfields have pharmacists at their institution put methylphenidate or placebo inside a capsule; this technique is readily done, they say, and maintains the blind as long as patients follow instructions not to open the capsule. Other placebo preparations pose greater hurdles; for B12 injections, Dr. Wiebe had a pharmacist put the vitamin and placebo in opaque syringes that masked the bright red color of the vitamin solution.

• Rating scales must be completed accurately and on time. "That's really critical," said Peter Camfield. "Sometimes you get the rating scales back, particularly from schools, and you can tell that they were all done on the last day."

• If formal rating scales are unavailable for a particular condition, patients and clinicians may wish to improvise their own, tailored to the patient's symptoms. One of Dr. Wiebe's chronic fatigue patients was particularly plagued by what she called "brain fog." In her trial, she rated this symptom on a 10-point scale each day.

• Trials should be timed to avoid confounding factors. For example, behavioral ratings from methylphenidate trials conducted in December may be tainted by the child's excitement about the holiday season. In addition, the Camfields start each week's trial on a Saturday, so that parents can monitor for adverse effects.

While preparing one's first "N of 1" trial requires a good deal of planning and preparation, subsequent trials using the same drug quickly become routine. The Camfields put together a packet of instructions and information, including ratings scales, that they give to each family. "Once you've done that," said Peter Camfield, "it just becomes an automatic part of your discussion with parents."

And because the results of the trial often provide more reliable and objective assessment of a treatment's benefits than does standard clinical practice, "N of 1" trials can actually speed up the process of establishing a stable treatment regimen for a patient. "It can be a time saver, rather than a time consumer," Peter Camfield emphasized.

—Peter Doskoch

Suggested Reading
1. Kent MA, Camfield CS, Camfield PR. Double-blind methylphenidate trials: practical, useful, and highly endorsed by families. Arch Pediatr Adolesc Med.1999;153:1292-1296.

Return to table of contents