Can we Reconcile Evidence-based Medicine with Personalized Medicine: Poised on a Cusp!

Abstract

How to cite this article: Kulkarni AP, Mishra RC. Can we Reconcile Evidence-based Medicine with Personalized Medicine: Poised on a Cusp! Indian J Crit Care Med 2022;26(S2):S1–S2.

Aristotle, the ancient Greek philosopher, advocated the move away from superstition or ideas without proof and suggested that scientific thinking should be based on logic. This affected almost all walks of science and he, therefore, is probably the originator of the idea of evidence-based medicine (EBM).¹ Professor Cochrane, in the 1970s, suggested that the practice of medicine should be based on evidence rather than only experience. Guyatt first used the term EBM in 1992.² The National Cancer Institute at the National Institutes of Health defines EBM as “a systematic approach to medicine in which doctors and other healthcare professionals use the best available scientific evidence from clinical research to help make decisions about the care of individual patients”.³ The EBM working group suggested relying on evidence from clinical research to individual care rather than intuition, clinical experience, and pathophysiological rationale alone.⁴

According to the tenets of EBM, the basis of therapy, i.e., the evidence, comes from clinical research. Clinical research, in the form of randomized controlled trials (RCTs), however, is well-conducted and ultimately tests the efficacy or harm of an intervention in a large number of patients (the higher the number, the better) who are randomly allocated to two or more groups. When the outcomes suggest the (new) intervention to be effective in a higher number of participants than the other (older) intervention, we say that it is more effective and can substitute for the other (older) intervention. Thus, RCTs, when available, form the basis of guidelines of professional organizations, which suggest as to how a group of patients with similar problems should be treated. The Institute of Medicine defines clinical guidelines as “systematically developed statements to assist practitioner and patient decisions about appropriate health care for specific clinical circumstances”.⁵

There are many potential benefits of the clinical practice guidelines such as improving the quality of care by advocating proven therapies and eliminating use of unproven or ineffective therapies. Guidelines may also help by improving the consistency of interventions in patients with similar clinical conditions. The main concern with guidelines may be that the protocols may be wrong for individual patients, possibly because (a) the evidence for recommendations may be wrong (particularly if the study design is flawed), lacking (not studied), misleading, or misinterpreted, (b) the recommendations may be inappropriate due to personal bias or opinions of the guidelines experts, and (c) recommendations may be flawed since other priorities have influenced the guidelines group rather than the best interests of the patients.⁶ There is also the danger, in particular, with trainee physicians simply doing things because they are part of a protocol or a guideline, in short, stifling individual thinking.⁷

David Sackett wrote that “Without clinical expertise, practice becomes tyrannized by epidemiological evidence, for even excellent external evidence may be inapplicable to or inappropriate for an individual patient”.⁸

Physicians know quite well that different patients respond differently to the administration of the same drug in the same dose. Science has now realized that this results from many different reasons such as variations in environmental exposure, genetic and epigenetic factors, dietetics, and response to concurrently administered medications. The study of these changes gave birth to the discipline of pharmacogenetics.⁹ In an excellent review about the role of genetics and antibodies in sepsis, the authors discuss the variable response of the patients to sepsis. In the immunosuppressive phase of sepsis, the concentrations of immunoglobulins are ultimately determined by genes with single-nucleotide polymorphisms (SNP). This means treatment efficacy of externally administered IgG cannot be determined without adjusting for SNPs for all genes involved in production of immunoglobulins.¹⁰ Georgescu and colleagues assessed a single SNP of tumor necrosis factor-α (TNF-α) gene in 163 patients with sepsis and compared the risk, severity, and outcomes of sepsis in them with over 200 controls.¹¹ They found that one of the SNPs (TNF-α +489G/A SNP A-allele) conferred protection against development of sepsis and septic shock without affecting mortality. In a Colombian study, Montoya-Ruiz et al. studied if variations in SNPs of lymphtoxin α, (LTA), TNF, InterLeukin (IL)1B, and IL10 genes were associated with clinical course and outcomes of sepsis. They found that various SNPs were associated with development, and/or severity, need for intensive care unit (ICU) admission, or mortality of sepsis.¹² The success of Imatinib in the treatment of chronic myeloid leukemia, in patients with dysfunctional tyrosine kinase, is a pointer toward better outcomes with personalized medicine. While these findings suggest that we are moving toward a therapeutic approach dominated by personalized medicine, the route of population-based randomized control trials for evaluating new therapies is diametrically opposite. How do we reconcile these two approaches?

We obviously need to abandon the old ways and adopt new routes to evaluating and validating personalized therapies. One such new way is n-of-1 studies (single-subject studies).^13,14 In n-of-1 studies, in a single subject, the serial effects of multiple interventions administered one at a time, comingled with wash out periods and placebo therapy, in a randomized manner, are observed. There are several drawbacks of this approach. N-of-1 studies cannot be performed in patients with acute or life-threatening conditions since there is not enough time to try multiple interventions or allow washout periods.¹⁵ N-of-1 studies are useful in chronic conditions, but only those with an appropriate biomarker, to assess patient response to multiple sequential therapies. These therapies should have rapid onset and rapid washout for obvious reasons.

Another approach is aggregating n-of-1 studies. Blackston et al. used simulation studies to compare aggregated n-of-1 with parallel and crossover RCTs. They concluded that n-of-1 trial designs trials are useful in testing interventions equally rigorously, in a smaller number of patients, at a lower cost with adequate power.¹⁶

Personalized medicine is the way of the future, till that becomes a standard of care across the world, the clinicians still need guidance in their day-to-day management of patients. This special issue of the Indian Journal of Critical Care Medicine contains several such guidelines and position statements for the intensivists. We hope this issue helps all our readers to improve the quality of care and outcomes of their patients.

Orcid

Atul Prabhakar Kulkarni https://orcid.org/0000-0002-5172-7619

Rajesh Chandra Mishra https://orcid.org/0000-0001-6305-5998