The Top Seven Analytics-Driven Approaches for Reducing Diagnostic Error and Improving Patient Safety
Diagnostic errors, from missed diagnoses to misdiagnoses, are all too common in healthcare—and they’re getting more attention as a prominent healthcare concern. In 2015, The National Academy of Medicine (NAM)—formerly known as the Institute of Medicine—published a report in the New England Journal of Medicine that described the industry’s growing concern as diagnostic errors become more clinically and financially damaging.
The commentary starts: “Diagnostic errors are clinically and financially costlier today than ever before. Efforts to identify, monitor, and reduce the rates of such errors therefore require greater attention and more dedicated resources than they’ve received in the past.” According to Bryan Oshiro, MD, “To get [the diagnosis] right, and to get it right the first time is one of the greatest challenges in healthcare. It is more important now than ever before due to the complexity and cost and available cures and treatments today.”
Accurate, timely diagnosis helps ensure patients receive the most appropriate care. Health systems that prioritize diagnostic process improvements make the most efficient use of resources and tests, and ensure the correct diagnostic procedures and protocols are used.
This article takes a closer look at diagnostic error, from how it’s defined to what the research reveals about its clinical and financial impacts. It emphasizes the importance of healthcare analytics, outlining seven analytics-driven approaches for reducing diagnostic error and improving diagnostic processes.
Back to Basics: Defining Diagnostic Error
NAM defines diagnostic error as: “the failure to (a) establish an accurate and timely explanation of the patient’s health problem(s) or (b) communicate that explanation to the patient.” A diagnosis is considered wrong when it’s inaccurate, incomplete (it doesn’t represent the patient’s true condition or doesn’t reveal enough detail for optimal treatment), or is determined too late to guide effective treatment decisions.
To explain the different types of diagnostic error, the Society to Improve Diagnosis in Medicine (SIDM) uses three main categories:
- Missed diagnosis: when the diagnostic tests don’t provide an explanation for a patient’s complaints (common in patients with chronic fatigue or chronic pain).
- Wrong diagnosis: when the original diagnosis is found to be incorrect because the true cause is discovered later (i.e., a patient having a heart attack is told their pain is from acid indigestion).
- Delayed diagnosis: the most common type of diagnostic error, in which the diagnosis should have been made earlier.
The SIDM also addresses the misconception that diagnostic errors primarily result from sub-optimal care. On the contrary, “Most diagnostic errors are made by conscientious clinicians practicing in first-rate medical organizations,” the SIDM states. SIDM emphasizes that diagnostic errors are more likely the consequence of the complexity of diagnosis and healthcare delivery, and what they call “cognitive error” (e.g., basic human mistakes). According to SIDM, most malpractice claims for diagnostic error involve heart attack, cancer, or stroke. The delayed diagnosis of breast cancer is also very common.
The Consequences of Diagnostic Error
The case for reducing diagnostic error intensifies when healthcare organizations realize how common it is. But, despite the prevalence of diagnostic error in the industry, many health systems don’t realize the severe consequences of these errors. Examples from recent research demonstrate how frequently diagnostic errors occur and how they impact patient safety and healthcare costs.
Patient Safety Consequences
The Patient Safety Network reported on two studies that assessed death tolls due to diagnostic error:
- In one study from the Harvard Medical Practice, diagnostic error accounted for 17 percent of preventable deaths in hospitalized patients.
- The second study, a systematic review of autopsy studies over four decades, found that approximately 9 percent of patients experienced a major diagnostic error that went undetected in their lifetime. Examples include abnormalities on mammograms that aren’t seen or followed up on, and abnormal pap smears that aren’t read correctly.
These studies indicate that diagnostic error is linked to thousands of deaths each year. One multicenter review, which examined 669 cases from 22 institutions, revealed that 28 percent of the study’s reported diagnostic errors were considered major (life-threatening or resulted in patient death or permanent disability). Forty-one percent of the reported errors were considered moderate. The same analysis of malpractice claims determined that the most frequent outcomes of diagnostic errors were death, significant permanent injury, major permanent injury, and minor permanent injury. Diagnostic errors more often resulted in death than other allegation groups and were the leading cause of claims-associated death and disability.
Another diagnosis-related factor that impacts patient outcomes is communication. NAM emphasizes the importance of communication, delineating that for a diagnosis to be useful to patients, providers must explain it in a way (according to the average patient’s level of medical literacy) that helps patients make the appropriate treatment decisions or lifestyle changes. Providers must explain why a particular test is so critical—and why it must be done in a timely fashion. The communication doesn’t end there; next, the provider makes sure the right tests are performed and that the results are evaluated and communicated to the patient.
Diagnostic error is responsible for costs associated with permanent disabilities, unnecessary testing, lost productivity, and increased insurance payouts. A 1986–2010 analysis of malpractice claims assessed the financial consequences of diagnostic error to find an inflation-adjusted, 25-year sum of diagnosis-related payments of $38.8 billion (or a mean per-claim payout of $386,849). Diagnostic errors were the leading type of claim (28.6 percent) and accounted for the highest proportion of total payments (35.2 percent). The costs associated with diagnostic error stem from more than just claims; they stem from the myriad of additional tests ordered and performed.
There’s no denying the significant impact diagnostic error has on patient safety and cost. The consequences of diagnostic error for patients are countless: the worry, angst, and stress patients feel when they miss work to come in for additional visits is just one of many examples. More severely, a delayed diagnosis could lead to an increase in morbidity and mortality. Given the significant patient outcome and financial consequences of diagnostic error, it is an industry wide problem health systems are compelled to solve.
Where to Start: The Top Seven Analytics-Driven Approaches for Reducing Diagnostic Error
NAM’s 2015 report described several strategies for improving diagnosis, from establishing formal medical school curricula about misdiagnosis to developing standard processes to determine how frequently diagnostic errors occur (and their downstream impact on health and economics). But improvements won’t happen without healthcare analytics.
Healthcare analytics is at the heart of reducing diagnostic errors, as it provides a source of truth about which diagnostic procedures are effective in which circumstances, as well as which ones are less likely to improve outcomes. Dr. Oshiro says, “With analytics, we don’t have to wait for the publication of huge studies before we can consider improvement work; we now have the power to review healthcare data on our local population and our current practices to glean insights quickly and to turn our facilities into continuous learning labs.”
Health systems need analytics before they can even embark on diagnostic process improvements. Without analytics, knowing where to begin is impossible (e.g., Is the error an outlier or is it part of a pattern of errors?). Health systems can reduce diagnostic error by implementing seven analytics-driven approaches:
Approach #1: Use Key Process Analysis (KPA) to Target Improvement Areas
Target improvement areas using an 80/20 Pareto analysis to identify cost-driving clinical areas and variation in care processes. The analysis should combine clinical and financial data to highlight the best opportunities for improvement and cost reduction, guiding the development of applications to support your improvement initiatives. The analysis should reveal areas with the greatest cost and volume, and show the degree of variation within subpopulations within select clinical programs. Health Catalyst has developed a tool called the Key Process Analysis (KPA) Application, which substantially automates this process for easy analysis.
Approach #2: Always Consider Delayed Diagnosis
Keeping in mind that the cause of delayed treatment is often delayed diagnosis, health systems should consider time-sensitive conditions in which they often fail to meet time targets for intervention (such as tissue plasminogen activator for myocardial infarction or stroke and the 3-Hour Bundle for sepsis). In incidents of delayed treatment, health systems should always ask, “Is delayed diagnosis the cause?”
Approach #3: Diagnose Earlier Using Data
Health systems should use analytics to examine to determine within their population or system those conditions that are being identified on the later side. For example, do you only identify diabetes patients when they have complications—not when they have prediabetes? In these cases, health systems should use data to determine what might be done to diagnose earlier, particularly for conditions in which early diagnosis is important to patient outcomes.
Approach #4: Use the Choosing Wisely Initiative as a Guide
The Choosing Wisely Initiative promotes patient-physician conversations about unnecessary medical tests and procedures, and delayed diagnosis. Diagnostic accuracy is critically dependent on the partnership between clinicians and patients. The initiative also promotes the use of evidence-based, safe, and necessary testing. Health systems should use Choosing Wisely recommendations as their guide, using analytics to answer several key questions:
- Are we doing too many unnecessary tests (i.e., routine X-rays/CTs/MRIs for low back pain)?
- Do patients who receive certain imaging or lab tests always meet indications?
- Are we taking too much time to get to the right tests?
- Are we mistaking symptoms for the wrong condition with a dramatically different treatment and risk (e.g., heartburn and heart attack can cause very similar symptoms)?
- How do the above factors influence a delayed diagnosis or wrong diagnosis?
Approach #5: Understand Patient Populations Using Data
Use analytics to understand characteristics of your patient populations (e.g., the social and economic factors that influence people’s health) and inform screening decisions. In the absence of specific guidelines or regulations, providers need to implement local standards that balance costs, potential benefits, and the potential for and impact of false-positives. For example, universal chlamydia screening may have high value and utility in high-prevalence areas (or with high-risk patients), but it may lead to a high false-positive rate in low-prevalence areas.
Approach #6: Collaborate with Improvement Teams
Connect with improvement teams already underway in a particular domain, such as cardiovascular or respiratory. Inquire about their greatest diagnostic challenges. Use analytics to understand the size and impact of delayed, mistaken, or missed diagnoses. Then investigate possible causes and quantify the impact of improvement.
Providers frequently re-order diagnostic tests because they don’t know if their patient has already had them at another facility or during another encounter, or because they can’t access test results and must therefore repeat tests. Consider ways to improve communication and reduce wasteful tests that may delay diagnosis. This could be something as simple as adopting the practice of always asking patients, “Have you been seen for this problem before? When and where?”
Approach #7: Include Patients and Their Families
Health systems must engage not only patients, but also patients’ family members. Family members (and even friends) have a global perspective—a big picture view—of patient’s lives. This understanding about a patient’s medical condition and other social determinants of their health will lead to better patient engagement, improved diagnostic accuracy, stronger compliance to treatment, and, ultimately, better outcomes. Enabling patients and their families to contribute valuable input will facilitate an accurate and timely diagnosis; it will improve care.
Reducing Diagnostic Error: A Success Story
In an effort to tackle the significant clinical and financial consequences of sepsis, one large medical center reached out to Health Catalyst to leverage advanced healthcare analytics solutions to improve sepsis care. Sepsis, a life-threatening disease caused by an infection, is responsible for approximately 200,000 deaths, more than 750,000 hospitalized patients, and 570,000 emergency department visits each year in the United States. Sepsis is costly, totaling $20.3 billion in 2011 or 5.2 percent of total inpatient costs in 2011.
The system had already partnered with Health Catalyst to implement a clinical, analytic, and process-based framework for improving the quality and cost of care, which gave it a foundation from which to deploy the Health Catalyst’s Population Analytics Advanced Application-Sepsis Module. This scalable platform integrated clinical, financial, operational, and other data sources to enable the health system’s teams to track interventions and their impact on sepsis rates and financial measures. The health system also formed a cross-functional team, who used the sepsis analytics platform to define cohorts and recommend best practices based on data-driven sepsis diagnosis and outcomes information.
Targeting provider compliance with the 3-hour sepsis bundle (including diagnosis), the health system was able to use its advanced healthcare analytics to track the rate at which nurses complied with the guidelines. The goal is full compliance with the assessment and 3-hour sepsis bundle program in order to reduce time diagnosis of sepsis and prevent progression to severe sepsis or septic shock.
Working closely with Health Catalyst advisors to facilitate implementation and adoption, the health system developed a prototype electronic algorithm for automating sepsis assessment. The results are impressive: they’ve tracked an 80 percent relative increase in compliance and a 41 percent absolute increase in nurses’ compliance with completing assessments. In less than 12 months, the health system significantly improved outcomes and reduced costs, with an average decrease in mortality of 22 percent and more than $1.3 million in cost savings.
Reducing Diagnostic Error Is a Healthcare Imperative
Simply put, reducing diagnostic error is a healthcare imperative because it will improve patient safety and outcomes, reduce costs, and protect resources. With accurate and timely diagnosis, healthcare systems can help patients get the care they need (when they need it), while avoiding spending money and wasting resources on treatments that are unlikely to help the patient. As Dr. Oshiro explains, appropriate, data-driven diagnosis is the first clinical step in outcomes improvement: “We talk about ‘having a single source of truth’ in our work; diagnostic accuracy is the [source of truth] for all treatment.”
With seven approaches grounded in analytics as their guide, health systems can start tackling diagnostic error. From using analytics tools such as the KPA application, to involving patients’ support networks (families and friends), health systems can begin to improve diagnostic accuracy.
Would you like to use or share these concepts? Download this heart failure readmission presentation highlighting the key main points