Liang Liu

Western medicine depends on science to create and assess drugs at the molecular level. In Asia, there is a commonly held belief that there is an art to healing too, and that both art and science should cooperate to help eradicate illness and relieve suffering.

Part of this art is seen in the way that Asian doctors perform a diagnosis. Doctors identify multiple signs and symptoms across the whole body and then consider these as a constellation or pattern that is unique to each person. Prescription, therapy and dosages are tailored to each patient and are often modified as the course of treatment progresses. Such complexity and changeability has been difficult, if not impossible, to assess through the current standard method of randomized controlled clinical trials (RCTs). Therefore, there is a huge need to design a rigorous yet appropriate clinical trial that can assess the safety and efficacy of traditional Asian medicines.

RCTs are the gold standard for testing new medicines. Within an RCT, a person will be randomly allocated to a group that receives one of the treatment options or a placebo. Typically, each group should contain the same distri-bution of relevant demographics, such as age, sex and ethnicity, among others.

RCTs can be classified into two subtypes depending on what they seek to prove. Explanatory RCTs (ERCTs) evaluate the efficacy of a pharmaceutical agent under highly controlled conditions that do not necessarily reflect real life. Pragmatic RCTs (PRCTs) assess the effectiveness of a treatment in everyday situations, meaning that a less-than-strict adherence to the treatment regimen is not a reason to exclude a subject. Other than that, both subtypes of RCT share the same principles: randomization of subjects into treatment and control groups; blinding all participants as to which group subjects are in; the need for follow up; and the handling of drop-outs and withdrawals1. ERCTs can produce more precise assessments of a pharmaceutical agent, although the trial design often deviates significantly from daily routine. PRCTs, on the other hand, generate weaker evidence but better reflect the circumstances of patients as they go about their daily lives.

The rigidity of ERCTs has necessitated the concept of the 'average patient'. Yet such a patient does not really exist, and designing a trial in this manner is not ideal for tailoring treatment to an individual. Without prior consideration of the underlying genetics, patients in a particular trial group might receive a drug that is ineffective for their genotype — or worse still, dangerous1. With the growing appreciation that a person's genetic variation affects response to treatment, trial designs must adapt.

Knowledge that dysregulated genes and proteins underlie the pathogenesis of many diseases has led to the development of pharmacogenomics: targeted molecular or genotype-based therapies that improve treatment and reduce adverse effects. However, RCTs have struggled to keep pace with the emerging understanding of individual pharmacogenomics2. For example, gefitinib, a lung-cancer drug, inhibits epidermal growth factor receptor (EGFR) signalling in target cells, providing patients with a significantly longer period without disease progression and fewer side effects than with conventional chemotherapy3. Only patients with a specific mutation in the gene that encodes EGFR get any benefit from gefitinib. Yet initial clinical trial results were negative because none of the patients in the trials had been selected based on EGFR activity. Positive clinical results were seen only after these subgroups were identified4,5.

Personalized medicine shares similarities with the methods of traditional Asian medicines.

Integrating the principles of personalized medicine into RCTs leads to personalized explanatory RCTs (PERCTs) and personalized pragmatic RCTs (PPRCTs). These types of trials select patients according to their genotype, rather than the normal methods based on shared clinical diagnostic and biological characteristics. Moreover, this concept of personalized medicine has similarities with the individualized diagnostic and treatment methods of traditional Asian medicines (notwithstanding the fact that the molecular mechanisms have yet to be elucidated). Therefore, it is reasonable to assume that any clinical trial designed for personalized medicine should be adaptable for testing traditional Asian medicines.

The diagnostic principles of Asian and Western medicine can be bridged by systems biology, which is a trend in biomedical research to examine complex systems in their entirety, rather than take a more reductionist, molecular viewpoint. Work is ongoing to link traditional Asian diagnoses and treatments with proteomic and biochemical signatures (see 'All systems go', page S87), with the promise that these can be applied in PERCTs and PPRCTs, as part of the criteria for assessing treatments6. For example, a recent report showed that the traditional Asian pattern diagnosis of 'heat' or 'cold' in rheumatoid arthritis patients was associated with different underlying genomic and metabolomic profiles, and different treatment outcomes7.

For traditional Asian approaches to become integrated into modern medicine, we need to consider alternative inclusion and exclusion criteria for RCTs. We also need to review how outcomes are assessed. These concepts are neither new nor unproven. Both Western-style modern medicine and traditional Asian medicine aim to heal patients in a harmonized way and can be developed together into an integrated form of personalized medicine. Redesigning clinical trials will accelerate the blending of these two styles of healing, for the benefit of humankind.