By: Craig W. Robbins, MD, MPH
In August, the Guidelines International Network (G-I-N) held its 2014 Annual Conference in Melbourne, Australia. G-I-N's mission is to lead, strengthen, and support collaboration and work within the guideline development, adaptation, and implementation community (1). In this Expert Commentary, Dr. Craig Robbins, who attended the 2014 G-I-N conference agreed to share insights on some of the topics he found relevant to his role as medical director of the Kaiser Permanente National Guideline Program. This year's conference theme was "Creation and Innovation: Guidelines in the Digital Age."
During the G-I-N plenary session, "Guidelines in Practice: Making Recommendations for Patients, Not Conditions," Dr. Allen Frances, Professor Emeritus of Psychiatry and former Chair at Duke University, encouraged the audience to use its evidence-based approach in leading local, regional, and national health policy conversations. We were reminded that evidence-based medicine (EBM) is a phrase that gets bandied about with both aspiration and consternation in many settings. Those of us in the healthcare community often do not mean the same thing when we invoke the phrase.
Within the Kaiser Permanente National Guideline Program, we have at times struggled with whether to label individual recommendations as evidence-based or consensus-based. Through conversations over the last few years with fellow G-I-N members as well as the adoption of GRADE (2) as our core guideline development methodology, we have come to understand EBM as a process and not a specific outcome or recommendation. Essentially, all clinical practice recommendations, which are implementable at the point of care and have the opportunity to improve health outcomes, will involve some amount of consensus building by the guideline development group. Consensus is established based on the quality of the body of evidence. Although a body of higher quality evidence can provide us with greater confidence in the estimated treatment effects for an intervention, no treatment is 100% effective or 100% safe. For this reason, every treatment we consider recommending to individual patients or across populations involves weighing the balance of potential benefits versus potential harms.
At its core, an evidence-based guideline development process means that a comprehensive literature search has been conducted, through which studies are selected objectively rather than on the basis of supporting opinion or preconceived notions. The quality of the evidence depends on how relevant the completed research is to the clinical question. The quality of the evidence should be assessed by people skilled in critical appraisal, typically those with backgrounds in biostatistics and clinical epidemiology. Regardless of the evidence quality, a clinical question is chosen for investigation in response to a need for clinical guidance. We should, therefore, consider the balance between benefits and harms, patient values and preferences, as well as resource implications in developing a clinical practice recommendation. Such a focus will lead us to write recommendations of varying strength, in language that appropriately reflects the level of importance for implied compliance with their guidance. If developed under systematically transparent and explicit documentation of the process and method for decision-making, the resulting guideline will indeed be evidence-based, even when it contains recommendations based on a lower quality body of evidence.
Shared Decision Making
If we are honest with ourselves, our patients, and the communities we serve, there is little we know in health care with absolute certainty. Most states of health, risk, and illness need to be understood in terms of likelihoods and probabilities. It behooves those of us delivering care to help our patients understand these issues, factoring in their own values and preferences when making lifestyle and treatment decisions aligned with their own realistic desires and goals for care and health. Such an approach is referred to as "shared decision making." In "How We Do Harm," Dr. Otis Brawley (3), of the American Cancer Society, encourages us to share routinely and truthfully with our patients what we know, what we do not know, and what we believe about the best course of action in a particular clinical situation. If in our evidence-based approach to care delivery we follow this admonition, we will move positively in the direction of making recommendations for patients, not conditions.
Overdiagnosis and Overtreatment
The participants of the G-I-N conference spent significant time considering the concepts of overdiagnosis and overtreatment. Overdiagnosis has been defined as the diagnosis of conditions or detection of abnormalities that will never cause symptoms or death (4). Overdiagnosis leads to overtreatment, exposing patients to the risk of adverse effects while offering no health benefit. Overdiagnosis and overtreatment are impossible to predict a priori in an individual patient. Rather they are revealed when we examine populations of patients across screening, prevention, and disease state management efforts.
How might guidelines developed through an evidence-based process reduce overdiagnosis and overtreatment? For example, in the area of prostate cancer screening, we know that the majority of prostate cancers are slow growing and if left undetected or untreated will not lead to significant morbidity or premature death in men. On the other hand, there are some aggressive prostate cancers that do lead to significant morbidity and premature death. We don't yet have a reliable screening test or protocol to differentiate between malignant and benign prostate cancer in an individual man. Based on the most favorable prostate cancer screening trial result (5) over a 10 year period, we can expect that prostate screening will prevent prostate cancer death in 1 man out of 1000 at most. We also know that many men will have positive prostate-specific antigen (PSA) tests that will lead to further evaluation, and in some cases, treatment of prostate cancer that was never destined to cause symptoms or premature death. Balancing these issues, our Kaiser Permanente Prostate Cancer Screening guideline development team recommended offering screening in the context of a shared decision making process to men in appropriate age ranges and higher risk categories. Looking across the balance of potential benefits and harms, evidence quality, and patient values and preferences, we are making more recommendations for shared decision making with individual patients, especially in the realm of screening and prevention.
Guidelines for Patients with Multiple Chronic Conditions
The topic of guidelines for patients with multiple chronic conditions was addressed frequently at the 2014 Annual G-I-N Conference. The standard approach to guideline development is to consider topics or conditions in isolation. Systematic evidence review is simplified by narrowing down our clinical questions to focused patient populations, treatment and comparison interventions, and outcomes. But, as Dr. Cynthia Boyd, geriatrician at Johns Hopkins University, pointed out in the final plenary session, as people age they often exhibit multiple chronic conditions. Addressing one chronic condition in isolation can lead to more harms than benefits from treatment. These conditions can be in related areas such as those that affect cardiovascular health: hypertension, dyslipidemia, and diabetes. Or they can involve somewhat disparate conditions like hypertension, chronic obstructive pulmonary disease (COPD), and rheumatoid arthritis. In order to best care for patients with multiple chronic conditions, clinicians and care delivery teams must incorporate patient values and preferences regarding potential benefits and harms of various treatment options into the decision-making process. To better inform these conversations, we should consider the multiple chronic conditions perspective in defining our clinical questions and completing systematic evidence reviews.
Forums like the G-I-N Annual Conference offer great opportunities to influence health care delivery and outcomes for patients, their families, and the communities in which they live. As reflected in this year's conference theme, the digital age provides us with powerful tools to bring evidence-based guidance to the point of care. In the United States, these tools may enable us to better share evidence-based clinical guidelines with the goal that they will enjoy wide dissemination and adoption, and ultimately lead to improved health outcomes for all Americans. Key to reaching this goal is maintaining the focus on patients, rather than conditions, and helping them make health care decisions informed by the best available evidence that reconciles with their personal values and preferences. In this way, we will be truly making recommendations for patients, not conditions.
Craig W Robbins, MD, MPH
Medical Director, Kaiser Permanente National Guideline Program
Medical Director, Kaiser Permanente National Guideline Program
The views and opinions expressed are those of the author and do not necessarily state or reflect those of the National Guideline Clearinghouse™ (NGC), the Agency for Healthcare Research and Quality (AHRQ), or its contractor ECRI Institute.
Potential Conflicts of Interest
Dr. Robbins is a member of the NGC/NQMC Editorial Board and a Trustee on the G-I-N Board of Trustees. He declares no financial or personal conflicts of interest with respect to this commentary.
- Guidelines International Network. About G-I-N/Introduction. Available from: www.g-i-n.net/about-g-i-n/introduction
- GRADE working group. Welcome. Available from: www.gradeworkinggroup.org
- Brawley O, Goldberg P. How we do harm: A doctor breaks ranks about being sick in America. New York: St. Martin's Press; 2012 Jan.
- Welch H G, Schwartz LM, Woloshin S. Over-diagnosed: making people sick in the pursuit of health. Boston: Beacon Press; 2011.
- Schroder FH, Hugosson J, Roobol MJ, et al. Screening and prostate-cancer mortality in a randomized European Study. N Engl J Med 2009;360:1320-8.