Evidence-Based Plastic Surgery: Its Rise, Importance, and a Practical Guide

Abstract

There is a perfect storm developing in 21st century healthcare; rising complexity and patient expectations in the context of fiscal restraint. Evidence-based medicine (EBM) may be the best-kept secret in dealing with the “storm.” Such an approach prefers management pathways that deliver better outcomes at less relative cost. In this article, the rise of EBM, its significance, a guide to practicing it, and its future in the field of plastic, reconstructive, and aesthetic surgery are presented.

Healthcare in the 21st century is a perfect storm, with unprecedented demands being placed on healthcare systems globally. In the developed world, there are soaring rates of obesity, diabetes, and chronic disease in the context of an ageing society; thus, patients are becoming increasingly complex.^1,2 In the developing world, there are high rates of trauma and tobacco use, infant and maternal mortality, infectious diseases, and malnutrition.^3,4 Such trends are occurring within the context of rising global demand for surgery/interventions,⁵ costly technology and drugs, costs of healthcare,⁶ and geopolitical instability against the backdrop of tight fiscal restraint resulting from a global financial crisis.⁷ There has also been a concomitant surge in public and patient expectations, as well regulatory demands.⁸ Juxtaposed to this is the growing realization and demand for patient choice with patient-centered care that respects patient autonomy and that is high quality, safe, timely, efficient, and equitable and combined with less variation and greater reliability.⁹ Indeed, this “perfect storm” extends to aesthetic surgery, as demonstrated in part by poor evidence for the efficacy of new technologies and the significance of industry.^10,11

THE EVIDENCE-BASED MEDICINE REVOLUTION AND ITS IMPORTANCE

Over the last 20 years, a quieter revolution has taken place, which might just have presented us with our most powerful weapon against these trends—evidence-based medicine (EBM).^12-14 There are powerful examples of how the evidence has influenced us. In 1984, Haagensen and Bodian wrote in support of Halsted's radical mastectomy.¹⁵ Defending it against less radical techniques and the criticism that it “cripples the arm and shoulder,” they stated: “My patients swim and play golf and tennis very well” and that “Detailed and very long-term results with these newer methods are not yet available and may never be because of the great expense and effort required.” Indeed, results did become available, as evidenced by Veronesi's seminal randomized controlled trial (RCT) with 20-year follow-up, which found that wide local excision had comparable outcomes to mastectomy for breast tumors less than 2 cm in diameter.¹⁶ The significance of this is that the Halsted radical mastectomy was the standard of care in the 1960s, but now the majority of breast cancer patients would have a wide local excision, and those needing a mastectomy would have a skin or even nipple-sparing one.

Another powerful example is melanoma excision margins. In 1962, Petersen was advocating 15-cm margins, but today the largest is 3 cm from the British Association of Dermatology Guidelines (Figure 1).^17,18

Figure 1.

Relative 3-cm and 15-cm excision margins for melanoma.

Donald Lalonde has eloquently discussed how he changed his own practice in light of the evidence across a range of areas.¹⁹ EBM is a logical approach to patient management that integrates a step, which ensures we can learn from others and their collective experiences of treating hundreds and thousands of patients and integrate that with our own. Clinical judgment is therefore based on a larger sample, has greater validity, and has more statistical power.

Sackett et al have defined EBM as the integration of clinical expertise and patient preferences with the best available research evidence to make decisions about the care of patients.²⁰ This logical triple-lock is both a movement, a philosophy, and something that can be utilized in individual doctor-patient consultations when reviewing treatment options. EBM cannot therefore be done well simply by reviewing the evidence; it takes time for clinicians to build “muscles” around judgment and to develop their expertise. The “this works well in my hands” argument is still valid within an EBM framework. For instance, a Cochrane systematic review analyzing 3 RCTs found that drains had no benefit in reduction mammaplasty, but some surgeons still prefer to use drains for fear of hematoma.²¹ Indeed, anecdotally, we understand some surgeons stopped using drains on such evidence, and their hematoma rate increased; thus, they reverted to their previous practice. If a person's experience is discordant with good evidence, this may represent poor luck, but it should trigger a search for the reasons why. Are patients being correctly selected, is the technique being executed in the same way, is the postoperative setting and support comparable? The same goes for integrating patient preferences and judging the impact of a treatment on a patient. Judging the best available research evidence is a matter of critical appraisal.

So how should one interpret research? How should one decide the ramifications for clinical practice? These are questions that clinicians and policy makers have been grappling with since the dawn of EBM and before.²² “Don't believe everything you read” is what our parents tell us as children. So what should we believe and how do we go about it? Indeed, this year is the 350th anniversary of the scientific journal—what better time to consider such issues?²³

EBM IN PRACTICE: HOW TO REVIEW THE LITERATURE CRITICALLY

This is a vast topic. But here is a simple framework, which provides guiding principles:

1. Is the research question worth asking? Is it the correct one to ask?

2. Are the methods appropriate to answer the stated research question(s)?

3. Are the conclusions justified by the data and the methods?

These are the broad “brush-strokes” of an initial analysis. Other questions may include

1. Are the groups appropriately balanced to minimize confounders?

2. Is the patient population sufficiently representative?

3. Were the groups treated equally and steps taken to minimize bias at all stages of the study?

4. Are the outcomes suitably objective? Can they be consistently measured?

5. Is the follow-up period sufficiently long?

6. Are the limitations of the work discussed?

7. Does this manuscript truly add to the literature in a meaningful way?

8. Is there overall clarity of the presentation and flow of logic?

Deconstructing the research question is a useful initial step. A useful framework is PICOST (ie, patient/population, intervention, comparator, outcome, setting, and time horizon).^38,53 These details should be appropriately described in the objective and the inclusion/exclusion criteria. The analogy of a race is useful here:

• Fair start: the groups are equal in baseline prognostic variables

• Fair race: the groups are treated equally

• Fair finish: the outcomes are objective

The next steps (assuming point 1 is satisfied) are to look closely at the methodology. There are a variety of standard scoring systems one can use. An early one for RCTs was the Jadad scale.²⁴ Research by our own group on the methodological quality of RCTs in plastic surgery has shown that many fall short.²⁵ In this case, we used an extended version of the Linde Internal Validity Scale; editors and peer reviewers would be wise to have similar scoring systems, such as the Cochrane risk of bias tool, on hand when evaluating such articles. Readers may not have these on hand at the time they read an article, but, again, such knowledge should permeate the clinical community more and more with time.

An important issue in conducting systematic reviews is to assess for publication bias. This is especially important in aesthetic surgery in which device manufacturers play a big role and so does United States Food and Drug Administration (FDA) regulation. Negative studies are not always published, so bias exists and opportunities for learning are lost. Readers should be aware of the inherent bias within our literature. The technical aspects of funnel plots—which are used to evaluate publication bias, are outside the scope of this article.

The other side of this “coin” is reporting issues. The Consolidated Standards of Reporting Trials or CONSORT was one of the first reporting criteria developed in a systematic way and has aided in the assessment of RCTs since its launch in 2001 and through subsequent updates.²⁶ CONSORT has opened a new field of meta-research, examining the quality of research and its reporting. Work by our own group in these areas has found poor reporting of RCTs across numerous surgical specialties.^27-30 Such reporting problems have been shown in observational studies too.³¹ Such work has now been confirmed by others, and we have continually raised these issues.^32,33 The problem is not confined to surgery, but permeates across the whole of clinical medicine.^34-36 Most of the aforementioned work shows that the trend over time is not improving and that there is no link between reporting quality and the journal impact factor.

Journals have an important role here. Editors and peer reviewers are guardians of the scholarly literature, discerning the best research from what they receive and ensuring its reporting is optimal. In our experience, peer reviewers often became bogged down with asking questions about information that authors should have included in the first place—this is where reporting guidelines such as CONSORT and STROBE can be so useful. Some journals insist on authors submitting a completed CONSORT or STROBE checklist for RCTs and observational studies, respectively, prior to peer review of the article.³⁷ This ensures the reporting quality of the article is optimal prior to peer review, thus allowing peer reviewers to focus on what is there, not what is missing. For our most prevalent article type—the humble case series—help is at hand through basic guidance provided by Coroneos et al.³⁸ Methodologically robust guidelines named PROCESS (preferred reporting of case series in surgery) are currently in development by our research group.³⁹ Guidelines are also in development for certain clinical areas. For instance, our group is developing a core outcome set for fat grafting (the VOGUE Study; Valid Outcomes in the Grafting of aUtologous fat to the brEast). This work was necessary following a systematic review that showed that 35 studies were reporting a total of 51 different outcomes, defined in various ways over the course of 1986 to 2014.^40,41

Transparency is important too. Research Protocols should be prospectively registered and articles should publish conflicts of interest, ethical approval, and funding sources, as well as describe author contributions.⁴² Some journals have taken the step of publishing peer reviewer reports, aiding transparency, accountability, and openness.⁴³

The beauty of EBM is that, ultimately, we get closer to the truth than we otherwise would have done. “Light” is shone onto bias (anything that causes a systematic deviation of the results from the truth). The risk of breast, uterine, and ovarian cancer with hormone replacement therapy (HRT) is one such example.⁴⁴ Other findings have been surprising, such as evidence that oxygen administration in acute uncomplicated myocardial infarction may actually increase mortality, contrary to traditional teaching.⁴⁵

THE BEST AVAILABLE EVIDENCE

Part of determining the best available research evidence is to categorize research into the appropriate level of evidence (LOE) as defined by the American Society of Plastic Surgeons (Figure 2).^46-48 This aids searching and transparency, with higher levels of evidence associated with less bias and greater validity if performed well. The performance and reporting of higher LOE plastic-surgery clinical research was stressed in an edition of the Clinics in Plastic Surgery in 2008.⁴⁹ In January 2011, a specialty-wide EBM initiative in plastic surgery was launched.^50,51 This included the listing of the LOE on all articles amenable to it.⁵²

Figure 2.

The Levels of Evidence pyramid (adapted from Chung KC, Swanson JA, Schmitz D, Sullivan D, Rohrich RJ. Introducing evidence-based medicine to plastic and reconstructive surgery. Plast Reconstr Surg. 2009;123:1385-1389). *Assuming high methodological quality.

There are 3 important points we wish to emphasize when it comes to LOE:

1. As one ascends the pyramid, there is less inherent bias in the study design.

2. The specialty needs to drive toward more level 1 and 2 studies (RCTs and Prospective Cohorts and systematic reviews of such studies).

3. One should not discard lower LOE studies, for hypothesis testing, for new treatments, new diseases, rare diseases, and rare or emergency scenarios they may be the most appropriate study to perform, at least in the first instance. The old adage “every dog has its day” comes to mind.

4. As Jenicek points out, the discussion around case reports and case series often revolves around their relevance to a potential cause-effect relationship. In reality, we can learn a great deal from such evidence (or building blocks) if we accept them for what they are and respect their limitations.⁵³

LOE is only a guide to judging a study design's relative quality and bias. High LOE studies can be badly conducted and reported, lack a power analysis, and suffer from type II statistical error; conversely, low LOE studies may still reliably answer the research question they set out to. Furthermore, we would like to highlight that conducting high LOE studies in surgery is challenging. Lack of funding, the need for multicenter collaboration to accrue enough cases and to demonstrate external validity, difficulties with randomization, allocation concealment, blinding, and therapeutic equipoise are all significant barriers.

THE FUTURE

Choudhry et al's systematic review showed that clinical performance can deteriorate over time after initially peaking in the years after graduation.⁵⁴ Indeed, overall, they observed a 0.5% increase in mortality for every year since the treating physician had graduated from medical school, despite controlling for a patient's probability of death, hospital location and practice environment, physician specialty, board certification, and the volume of patients seen. They postulated that older physicians may be less likely to update their “toolkits,” less likely to adopt new therapies, and less receptive to new clinical guidelines and standards.⁴³ A commitment to lifelong learning and keeping up to date must be integral to the foundations of ethical professional practice. Governments also want enduring, sustainable, evidence-based policies toward disease prevention and management that integrate scientific and clinical evidence with health economics and consumer research.⁵⁵ More cost-effectiveness studies are needed to ensure the dimension of cost can be considered given limited resources and therefore the need to prioritize.

Medical knowledge is growing at a phenomenal rate, doubling every 18 months.⁵⁶ At any given moment, there are 55 new clinical trials taking place. Each year, 800 new primary-care guidelines are added and more than 2000 new research articles are added to MEDLINE each day, this is a considerable challenge.⁵⁷ Glasziou has argued²²:

The search engine is now as essential as the stethoscope … a 21st century clinician who cannot critically read a study is as unprepared as one who cannot take a blood pressure or examine the cardiovascular system. The medical curriculum should reflect this importance of changing information for today's practitioner – the necessary skills must be taught and assessed with the same rigour as the physical examination.

Physicians and surgeons now find themselves in a perfect storm: stretched clinically, dealing with a high turnover of patients of increasing complexity, and information overload. Strong leadership and educational reform is now required.⁵⁸

Critically appraising research is difficult. Findings should be interpreted cautiously, and one should look for multiple studies showing a similar effect. Changing clinical practice should be done on the basis of robust evidence, appropriately interpreted and with patient safety of upmost concern. New interventions should be developed in EBM-rich pathways like IDEAL (idea, development, exploration, and long-term follow-up).⁵⁹ Recent guidance has been provided on adopting new aesthetic surgical innovations as well.¹¹

MOVING EVIDENCE-BASED PLASTIC SURGERY FORWARD

The future, we hope, will include mandatory adherence by authors to reporting guidelines with enforcement by peer reviewers, editors, and journal submission systems and publishers. The Declaration of Helsinki now mandates that all research is registered prior to recruitment of the first patient.⁶⁰ Here, too, there has been innovation recently, with the launch of the Research Registry.⁴⁰ We encourage plastic surgeons to register their work prospectively and submit their protocol for peer review, hence opening it up to discussion, feedback, and potentially a course correction, preventing wasted time, funding, and cynicism. Authors and journals need to publish negative studies to prevent repetition, allow for learning, and prevent “back to the future” studies. A system to transparently report complications (like the yellow-card system for drugs) without fear of legal ramifications is needed. Above all, this all means greater effort, less haste, and a stronger focus on quality of research, collaboration, and a multicenter approach. A pathway for the evidence-based plastic surgeon is given in Figure 3.

Figure 3.

A pathway for developing higher quality research projects. *Assuming it's conducted well.

Over 2500 years ago, Hippocrates declared that the physician needed to “rely on actual evidence rather than on conclusions resulting solely from reasoning, because arguments in the form of idle words are erroneous and can be easily refuted.”⁶¹ Healthcare today is far more complex than at Hippocrates' time, and doctors have to be able to justify their treatment decisions robustly. Pooling expertise through multidisciplinary teams has shown significant benefits for patient outcomes and for dealing with increasingly complex patients and diverse therapeutic modalities that no one physician or surgeon can master.⁶² EBM is not just the right thing to do by the patient, as its use will likely result in the best outcomes, but it is a clear and present answer (at least in part) to the perfect storm. The first published cost-effectiveness study of EBM clinical pathways in 2010, found that costs were 35% lower for those patients on-pathway rather than off, with no difference in overall survival.⁶³

CONCLUSIONS

EBM education is clearly needed to deal with the problems and the pressures of 21^st-century healthcare. Such skills must complement the traditional focus on craft skills to enable the modern-day plastic surgeon to consistently deliver high-quality care.

Disclosures

Dr Agha is Managing and Executive Editor of the International Journal of Surgery and Founder and Director of the ResearchRegistry.com.

Funding

The authors received no financial support for the research, authorship, and publication of this article.