What Is Best Practice in Healthcare?

Best practice in healthcare refers to a clinical approach, procedure, or protocol backed by the strongest available research evidence, filtered through a clinician’s professional judgment and aligned with what the patient actually wants. The concept grew out of the evidence-based medicine movement, which formally defined good clinical care as the integration of the best research evidence with real-world clinical expertise and patient values. That three-part definition sounds simple, but each piece introduces tensions that make best practice far more contested and harder to implement than most people realize.

Where the Idea Comes From

The phrase “best practice” gets thrown around loosely, but in healthcare it has a specific intellectual lineage. Evidence-based medicine, or EBM, emerged as a formal discipline in the mid-nineteenth century and has evolved continuously since then. Its core argument is that clinical decisions should rest on high-quality research rather than tradition, authority, or habit alone. The framework identifies three pillars: the best available research evidence, the clinician’s own expertise, and the individual patient’s values and preferences.

That third pillar is easy to overlook but matters enormously. A treatment that works well on paper might be wrong for a particular patient because of their life circumstances, their priorities, or conditions the research population didn’t share. Clinical expertise acts as the bridge, helping a clinician figure out which research findings apply to the person sitting in front of them and how to weigh the patient’s own goals against the statistical evidence.¹ Best practice, then, is not a single universal recipe. It is a process of reasoning that tries to get all three pillars working together for a specific clinical decision.²

How Clinical Guidelines Get Built

Best practices are most often codified as clinical practice guidelines, the documents that tell clinicians what to do for a given condition. These are not opinion pieces. The development process is structured and labor-intensive. The American College of Physicians, for example, follows a multistep method that starts with a systematic review of the evidence, moves through committee deliberation, and ends with graded recommendations indicating how strong the evidence is and how confident the panel feels about each suggestion.³

One of the most widely used systems for grading evidence is the GRADE approach, which evaluates recommendations based on the balance between benefits and harms, the quality of the underlying evidence, patient values and preferences, and resource considerations.⁴ A “strong recommendation” means the panel is confident the benefits clearly outweigh the harms for nearly all patients. A “weak” or “conditional” recommendation means the trade-offs are closer and the right choice depends more heavily on individual circumstances. Readers of guidelines sometimes miss this distinction and treat every recommendation as equally iron-clad, which it isn’t.

The Delay Between Discovery and Bedside

Even once the evidence is clear and the guideline is published, getting it into everyday practice is notoriously slow. A widely repeated figure in healthcare is that it takes about 17 years for research findings to reach routine clinical use. That number traces back to a study that analyzed how long it took evidence across nine clinical disciplines to reach a 50 percent adoption rate. The researchers found an average annual uptake increase of about 3 percent, which translated into roughly six years for evidence to appear in reviews and textbooks and another nine years for clinicians to actually implement the findings in practice.⁵

The real picture is more complicated than the headline number suggests. The gap varies significantly by clinical specialty, by what counts as “implementation,” and by local context. Some findings get adopted quickly because they are cheap and easy to integrate. Others stall because they require new equipment, workflow redesign, or a cultural shift in how clinicians think about a problem. The existence of this delay is one reason implementation science has become its own field, focused specifically on how to close the gap between what the research says and what actually happens in clinics and hospitals.⁶

Safety Checklists as a Concrete Example

If you want to see what best practice looks like in action, surgical safety checklists are one of the clearest cases. The WHO Surgical Safety Checklist, introduced in 2008, is a straightforward tool: before an operation begins, the surgical team pauses to confirm the patient’s identity, the procedure, the surgical site, and anticipated complications. It was tested in a landmark study across eight hospitals worldwide, and the results were striking. The death rate dropped from 1.5 percent to 0.8 percent, and the complication rate fell from 11 percent to 7 percent.⁷

A review of the broader literature on surgical checklists confirmed these patterns. Checklists were consistently linked to better detection of potential safety hazards, fewer surgical complications, and improved communication among operating room staff.⁸ A separate analysis of two outcome studies found relative reductions in surgical death of 47 percent and 62 percent, and relative reductions in complications of 36 percent and 37 percent.⁹

What makes the checklist story instructive is how simple the intervention is. No new drug, no expensive device, just a structured pause. That gap between simplicity and impact is a recurring theme in best practice: the barriers to better care are often organizational and behavioral, not technological.

Why the Patient’s Voice Matters

Best practice is not just about clinicians applying the right evidence. It increasingly means involving patients in the decision itself. Shared decision-making is a process where physicians bring the scientific evidence and patients bring their own goals, preferences, and values, and together they arrive at a plan.¹⁰ This matters most for preference-sensitive decisions, situations where more than one reasonable option exists and the right choice depends on what the patient cares about. Cancer screening, joint replacement, and long-term medication use are common examples.

Patient decision aids are tools designed to help people participate in these choices by laying out the options, the evidence behind each one, and prompting the patient to think through what they personally value.¹¹ In parallel, patient-reported outcome measures and patient-reported experience measures are becoming standard tools for tracking whether care is actually improving things that matter to people, not just moving lab values in the right direction.¹² The shift here is subtle but important: best practice used to mean “the treatment supported by the best evidence.” Now it increasingly means “the treatment supported by the best evidence that also fits this patient’s life.”

Communication Between Clinicians

A lot of what goes wrong in healthcare has nothing to do with whether the right treatment was chosen. It has to do with whether critical information made it from one person to another. Miscommunication during handoffs, when one nurse passes a patient’s care to the next shift or a surgeon updates the primary physician, is a persistent source of error. One response has been the adoption of structured communication tools, the most widely used being SBAR, which stands for Situation, Background, Assessment, and Recommendation.¹³

The evidence on SBAR is encouraging but uneven. A systematic review found that of 26 different patient outcomes measured across the studies, eight improved significantly and 11 were described as improving without a formal statistical test. None of the studies found that SBAR worsened patient safety.¹⁴ In one study from an anesthetic clinic, the proportion of incident reports caused by communication errors dropped from 31 percent to 11 percent after SBAR was introduced, and staff perception of safety climate improved significantly.¹⁵ Structured communication isn’t glamorous, but it is a genuine best practice in the sense that it addresses a known source of harm with a tool that works.

Unwarranted Variation in Care

If best practices existed and everyone followed them, you would expect patients with the same condition to receive similar treatment regardless of which hospital they visited. They do not. Research on clinical variation has consistently shown that much of the difference in how healthcare is used has less to do with differences in illness or patient preference and more to do with the willingness and ability of individual doctors to offer treatment.¹⁶

Some variation is perfectly appropriate. Patients differ, and adapting care to individual needs is part of best practice. The problem is unwarranted variation, differences in care that cannot be explained by patient factors and that do not improve outcomes. A framework for distinguishing warranted from unwarranted variation highlights the roles of motivation, evidence, judgment, and both personal and organizational capacity in clinical decisions.¹⁷ Understanding why care varies is the first step toward deciding which variation to reduce and which to preserve.

Stopping Low-Value Care

Best practice is not only about doing the right thing. It is also about stopping the wrong thing. Healthcare is full of practices that were once considered standard but have since been shown to offer little benefit or, worse, to cause harm. De-implementation, the deliberate process of withdrawing low-value care, turns out to be psychologically and organizationally much harder than adopting new treatments.

A qualitative study of hospitals participating in Canada’s Choosing Wisely campaign found that habit was the dominant provider factor sustaining low-value care. Clinicians knew intellectually that certain tests or treatments weren’t helpful, but the routine of ordering them was deeply ingrained. The availability of performance data, the influence of the campaign itself, and the absence of firm targets or automated safeguards all shaped whether hospitals actually managed to reduce unnecessary care.¹⁸ The lesson is that evidence alone rarely changes behavior; you need systems that make the right choice the easy choice.

Audit and Feedback

One of the most studied tools for nudging clinician behavior closer to best practice is audit and feedback: giving healthcare professionals summaries of how their performance compares to a standard or to their peers.¹⁹ A large Cochrane review synthesizing hundreds of studies found that audit and feedback produced a median absolute improvement in desired practice of about 3 percent, with a mean improvement of about 6 percent.²⁰

That sounds modest, and it is. But applied across thousands of clinical encounters, even small shifts in practice can translate into meaningful improvements in outcomes. The effect is also quite variable: feedback works better when baseline performance is low, when it comes from a trusted source, when it includes specific action plans, and when it is delivered repeatedly rather than as a one-off report. When feedback feels coercive, it tends to produce compliance without deep engagement, giving clinicians a sense of what they should change and why, but not much help with the how.²¹

Technology at the Point of Care

Electronic health records have made it possible to embed best-practice prompts directly into clinical workflows. Clinical decision support tools can flag drug interactions, suggest evidence-based screening tests, or alert a physician when a patient’s situation doesn’t match the planned treatment. A randomized trial tested whether decision support could help physicians contextualize care, meaning recognize and respond to the individual circumstances that might make a standard recommendation inappropriate. Physicians who received the alerts were roughly twice as likely to investigate patient-specific red flags and about two and a half times as likely to incorporate those factors into the care plan. Across both study groups, when care plans were contextualized to the patient, the odds of the flagged problem improving roughly doubled.²²

The promise here is real, but so is the risk of alert fatigue. When clinicians are bombarded with pop-ups, they start clicking past them without reading. Designing these systems well means choosing carefully which alerts actually deserve to interrupt a busy clinician and which can be delivered more quietly.

Building Organizations That Sustain Best Practice

Individual tools and guidelines matter, but whether best practice sticks often comes down to organizational culture. The concept of high reliability organizations, borrowed from industries like nuclear power and aviation where failure is catastrophic, has been increasingly applied to healthcare. The idea is to build systems that expect things to go wrong and have layers of defense in place to catch errors before they reach the patient.²³

A rapid evidence review identified five common strategies for moving toward high reliability: developing leadership commitment, supporting a safety culture, providing training and learning opportunities, building robust data systems, and implementing quality improvement interventions. Multicomponent efforts delivered over at least two years were linked to improved process and patient safety outcomes, though the reviewers noted the overall level of evidence remained low.²⁴ The US Department of Veterans Affairs began a system-wide push toward high reliability in 2019, centering on foundational practices designed to advance a culture of safety across all its facilities.²⁵

Conflicts of Interest in Guideline Development

A persistent vulnerability in the best-practice pipeline is the influence of financial and intellectual conflicts of interest on the people who write guidelines. The American College of Physicians has a relatively rigorous process: anyone involved in developing a guideline must disclose all healthcare-related financial and intellectual interests from the previous three years. A review panel grades each conflict as low, moderate, or high, with high-level conflicts resulting in recusal from authorship, voting, and discussion.²⁶

Not every organization is that careful. A study examining conflict-of-interest policies among organizations that produce large numbers of clinical practice guidelines found that fewer than half had a policy specifically tied to guideline development. Among those that did, the policies varied widely, and not a single organization met all of the standards recommended by the Institute of Medicine. Nine organizations didn’t meet even one.²⁷ This gap matters because guidelines shape prescribing, testing, and surgical decisions for millions of patients. When the people writing those recommendations have undisclosed ties to the companies selling the recommended products, the credibility of the entire best-practice framework is at stake.

When Guidelines Meet Malpractice Law

Clinical practice guidelines were created to improve patient care, but they have taken on a second life in courtrooms. Both plaintiffs and defendants in medical malpractice cases increasingly use guidelines as evidence of what the standard of care should be.²⁸ A plaintiff might argue that a doctor fell below the standard by not following a published guideline. A defendant might argue that the guideline supports the decision they made.

This creates a tension. Guidelines are written for populations, not individuals, and every guideline panel recognizes that some patients will need care that deviates from the recommendation. But in a legal setting, the nuance of weak versus strong recommendations and the idea that clinical judgment sometimes justifies departure can get flattened. For clinicians, this means that guidelines carry a kind of soft regulatory power even when no law mandates following them.

Adapting Evidence Across Populations

Evidence-based guidelines are typically developed from research conducted in well-resourced settings with relatively homogeneous patient populations. What happens when a community health team tries to apply those guidelines in a context the researchers never studied? A study examining how mental health agencies adapted an evidence-based fitness program for people with serious mental illness found that all of the adaptations were unplanned and reactive. The teams hadn’t set out to modify the program; they did so because participants faced social and economic barriers to accessing the community fitness facilities the program assumed they could use. Agencies ended up offering sessions at the mental health center instead.²⁹

This kind of adaptation raises a question that runs through all of best practice: how much can you change an intervention before it stops being the evidence-based program you tested? The field is still working that out. But the example illustrates something important. Best practice is not just about having good evidence. It is about having the organizational capacity, the resources, and the local knowledge to put that evidence to use in ways that actually reach the people who need it.