- Diagnosis is crucial for effective communication and documentation of a patient’s condition, guiding the selection of optimal treatment strategies. It facilitates inter-professional dialogue among clinicians and promotes consistent care delivery.
- Higher-order thinking, involving advanced cognitive processing beyond rote memorization of facts and concepts, is essential in the diagnostic process, particularly after an initial diagnosis is formulated.
- Diagnostic metrics serve as indicators, either internally focused on test performance or externally oriented towards informing post-test clinical decisions. The most valuable diagnostic tests are those that effectively guide subsequent decision-making.
- Overdiagnosis poses a risk, potentially leading to unnecessary and excessive treatment interventions. The pursuit of diagnoses must be carefully balanced to avoid potentially adverse patient outcomes.
- A single diagnosis can encompass diverse phenotypes, meaning patients sharing the same diagnosis may exhibit considerable variations in their clinical presentation and response to treatments.
Keywords: Diagnosis, Sensitivity, Specificity, Higher order thinking
Abstract
Background
Differential diagnosis, in medical terms, is a structured approach employed to distinguish between conditions with similar signs and symptoms to arrive at the correct diagnosis. This process is fundamental in clinical practice, ensuring accurate identification of a patient’s health issue from a range of possibilities.
Methods
This masterclass aims to explore the critical role of higher-order thinking within the framework of differential diagnosis. It delves into the cognitive skills and reasoning processes necessary for effective diagnostic decision-making.
Conclusions
For healthcare practitioners, diagnosis represents a vital component of the broader clinical decision-making process. It is characterized by the careful differentiation of potential conditions to achieve a definitive understanding of the patient’s underlying health issue. The diagnostic journey encompasses the identification of the etiology of a disease or condition through a thorough evaluation of patient history, physical examination findings, and the interpretation of laboratory and imaging results, culminating in a descriptive diagnostic label. While the proficiency in differential diagnosis varies among healthcare providers, the foundational concept of diagnosis remains universally pertinent across all medical disciplines. Ideally, a precise diagnosis enhances the utilization of classification systems, promotes clear communication, charts a course for treatment, improves prognostic understanding, and, in certain instances, facilitates preventative strategies. Realizing these benefits necessitates a comprehensive grasp of the clinical utility of diagnostic tests and measures in relation to diagnosis and the effective application of these insights in clinical settings. This demands a sophisticated level of cognitive engagement—higher-order thinking—regarding the function of diagnosis in patient management.
The Foundation of Diagnosis: Context and Evolution
Background
The diagnostic process is fundamentally about pinpointing the nature and cause of a patient’s health issue. It’s achieved through a systematic review of their medical history, a thorough physical examination, and the analysis of laboratory results or diagnostic imaging findings, ultimately leading to a descriptive diagnosis.1 Diagnoses are essential for clear communication—among healthcare providers, with patients, and with health systems and payers. According to Walker,2 the evolution of diagnosis over the past three millennia has been marked by significant advancements. These include the establishment of a rational basis for medicine as a profession, the invention of diagnostic tools, the use of autopsies to validate diagnoses, anatomical dissection for medical education, the increasing reliance on physical and laboratory examinations, and the systematization of diagnostic classifications.
In 1893, the International Statistical Institute adopted the first edition of the International Classification of Diseases (ICD),3 initially known as the International List of Causes of Death. The World Health Organization (WHO) released the 11th Revision of the ICD in May 2018. This system is designed to standardize the definitions of diseases, disorders, injuries, and health conditions globally. The ICD classification organizes health information into standardized categories, which facilitates: efficient storage, retrieval, and analysis of health data for evidence-based decision-making; consistent sharing and comparison of health information across different healthcare settings and countries; and longitudinal data analysis within the same region over time. Furthermore, the ICD coding system offers a greater level of detail and clinical specificity, enhancing the ability to document patient encounters comprehensively and compare health outcomes at a systemic level. At its core, the ICD diagnostic system is aimed at improving communication amongst healthcare providers and is considered a fundamental competency for all diagnosticians in healthcare.
While improved communication and a shared language for categorizing diseases are invaluable in differential diagnosis, effectively utilizing these disease categories and understanding the inherent limitations of diagnostic labels requires higher order thinking. Higher order thinking is rooted in the idea that certain learning processes demand more complex cognitive processing, extending beyond simple memorization of facts and concepts. These advanced cognitive skills include conceptualization, analysis, and evaluation, engaging higher levels of reasoning that involve productive, analytical thought rather than just reproductive recall.4 Essential skills for higher order thinking encompass analogical and logical reasoning.5 Analogical reasoning involves drawing parallels and analyzing similarities, while logical reasoning uses prior knowledge to infer and solve problems. Critical thinking is a key component of higher order thinking.
A fundamental understanding of diagnosis is a complex, iterative, and indispensable process. This masterclass emphasizes that higher order thinking transcends the rote application of tests and measures, sensitivity and specificity calculations, and the mere listing of ICD-based diagnoses. Specifically, we posit that effective higher order differential diagnostic reasoning necessitates a clinician’s focused attention on: (1) the potential for test metrics to be misleading; (2) the risk of diagnostic labels to complicate patient care; and (3) the benefits of alternative diagnostic classification methods for improved patient management.
The Deceptive Nature of Test Metrics
Interpreting Test Metrics
To arrive at a diagnosis—determining whether a patient has a particular condition—clinicians depend on diagnostic tests, such as clinical examinations and imaging. A crucial aspect of any diagnostic study is the comparison between a test being evaluated (the “index test”) and an established, definitive method for diagnosis (the “reference standard”). This comparison yields key test metrics.6, 7
Table 1 outlines the most frequently used metrics in diagnostic assessment. Sensitivity (SN) and Specificity (SP) are calculated within specific populations in case-control studies. For instance, Sensitivity is calculated only among patients confirmed to have the disease, while Specificity is calculated only in individuals known not to have the disease. Similarly, Positive Predictive Value (PPV) and Negative Predictive Value (NPV) are derived from case-control study populations. PPV is calculated exclusively from those who test “positive,” and NPV from those who test “negative.” SN, SP, PPV, and NPV are considered internal test metrics and are not designed for direct application in post-test decision-making.
Table 1.
Common test metrics for differential diagnosis.
| Metric | Abbreviation | Definition |
|---|---|---|
| Sensitivity | SN | Percentage of people who test positive for a specific disease among a group of people who have the disorder. |
| Specificity | SP | Percentage of people who test negative for a specific disease among a group of people who do not have the disorder. |
| Positive Predictive Value | PPV | Probability that subjects with a positive test truly have the disorder. |
| Negative Predictive Value | NPV | Probability that subjects with a negative test truly don’t have the disorder. |
| Positive Likelihood Ratio | LR+ | The odds of a patient to have a disorder if the test is positive compared to the probability for someone who does not have the disorder. |
| Negative Likelihood Ratio | LR− | The odds of a patient not having the disorder if the test is negative compared to the probability for a patient who has the disorder. |
Likelihood ratios (LRs), in contrast, are calculated using data from the entire case-control study population, making them valuable for assessing clinical utility—specifically, their usefulness in making informed diagnostic decisions. An LR+ greater than 1.0 increases the post-test probability of disease given a positive test result, whereas a low LR− (close to 0) increases the post-test probability of the absence of disease given a negative test result.8 Both LR+ and LR− are dependent on the pretest probability and are used to determine the post-test probability of a diagnosis, whether to confirm or exclude it. While benchmark values exist (e.g., LR+ >5 and LR− 9 But in reality, each likelihood ratio is dependent on the pretest probability and should be considered individually to guide decision-making.
The current diagnostic paradigm often relies on interpreting these metrics to assign the most probable diagnostic label. However, this approach has limitations. As mentioned, Sensitivity, Specificity, PPV, and NPV are internal metrics, not intended for standalone decision-making. Relying solely on these values can be misleading because they may not accurately reflect the diverse patient populations seen in clinical practice. The concepts of “SPin” (ruling in with high specificity) and “SNout” (ruling out with high sensitivity) are outdated and can lead to misinterpretations.9 For example, a study on diagnosing subacromial pain found that combining three clinical signs achieved 100% Specificity but only 9% Sensitivity.10 Even though this combination is highly specific for subacromial pain, it would only identify a small fraction (9%) of all patients with the condition, making it clinically impractical as a standalone diagnostic tool. The “SPin” and “SNout” approaches are only reliable when other metrics are also at acceptable levels to minimize diagnostic errors.
Likelihood ratios are beneficial as they quantify the change in post-test probability, but they too can be misinterpreted. For instance, the Lachman test’s accuracy for diagnosing anterior cruciate ligament (ACL) tears has been established in both primary care and orthopedic settings.11, 12 However, the prevalence of ACL tears differs significantly between these settings. In primary care, it’s around 4%, while in orthopedic clinics, it’s closer to 20–25% due to the different patient populations seen in each setting.11 Even with a high LR+ for the Lachman test,13 the post-test probability varies significantly depending on the prevalence in the cohort, affecting decisions about further imaging or specialist referrals. Prevalence also influences the interpretation of red flags. Recent studies question the utility of historical factors in ruling out serious causes of low back pain due to their poor negative likelihood ratios,14, 15 which is linked to the very low prevalence of serious conditions causing back pain in primary care.
Study Design and Condition Severity: Influential Factors
The interpretation of test metrics is also heavily influenced by the quality of the research supporting the test. For example, the Thessaly test for meniscal tears was initially validated in a study with methodological weaknesses, and subsequent studies have failed to replicate its findings.6, 16, 17 Clinicians should critically evaluate study designs, reference standards, and test descriptions to identify potential biases in diagnostic accuracy studies.18 Furthermore, the severity of the condition within the studied population can alter outcomes. Advanced conditions with significant disability and pain tend to show higher sensitivity and lower specificity in diagnostic tests. Conversely, milder conditions may exhibit lower sensitivity and higher specificity.
Implications for Clinical Decision-Making
Decision-making in clinical practice involves a balance between evidence-based analytical approaches (like test metrics) and intuitive judgment derived from clinical experience.19 Clinicians must navigate the challenges of test metric interpretation daily. All diagnostic tools, whether clinical examinations or imaging, have inherent strengths and limitations. Misinterpreting test accuracy, misunderstanding probabilities, and relying on low-quality evidence can impede the analytical process.20 Intuitive judgment can be skewed by cognitive biases, such as confirmation bias, anchoring bias, or premature closure, where a clinician latches onto an initial diagnosis and prematurely ends the diagnostic process once a seemingly fitting explanation is found.20
Ultimately, diagnostic test results guide clinicians in making decisions about further investigations and treatments. Clinical reasoning is, therefore, paramount in linking test results to appropriate management within a comprehensive care pathway. Higher order thinking requires clinicians to look beyond mere test metrics and consider the consequences of misdiagnosis and how diagnostic decisions impact healthcare utilization.
Key Takeaway: Most test metrics are internal and not directly applicable for determining post-test probability. Metrics are susceptible to biases from study design and patient condition severity. Even likelihood ratios, designed for post-test probability, must be applied with a thorough understanding of pretest probability to avoid misinterpretations.
The Pitfalls of Over-Reliance on Diagnostic Labels
Diagnostic codes, often rooted in patho-anatomical concepts, have led to an emphasis on tissue-based diagnoses for musculoskeletal disorders. However, diagnosing and classifying patients solely based on this model can result in overly complex or clinically insignificant diagnostic labels that may not improve patient outcomes.
Overuse of Diagnostic Testing and Overdiagnosis in Musculoskeletal Care
Diagnostic tests and metrics are widely used across medical fields to guide clinical decisions.21 However, there is growing recognition that an over-dependence on diagnostic labeling contributes to the overuse of diagnostic tests and the problem of overdiagnosis. Overdiagnosis occurs when a patient is given a diagnostic label for a condition that would never have caused them harm,22 often when tests identify abnormalities or risk factors that are unlikely to cause symptoms or functional impairments.23 Overdiagnosis is fundamentally linked to how diagnostic labels are defined and how test metrics are interpreted.
When patients present with pain in the spine, knee, hip, or shoulder, clinicians frequently initiate a series of inquiries, physical tests, clinical measures, and imaging studies to pinpoint the source of their symptoms.24 Musculoskeletal care is particularly prone to overuse of diagnostic tests, with up to 50% of imaging referrals deemed inappropriate.25 This is compounded by the high prevalence of asymptomatic structural abnormalities detected on imaging in musculoskeletal conditions. Common examples of such diagnostic labels include “lumbar degeneration,” “disc bulges,”26 “disc herniation,”27 “degenerative meniscal tears,”28 “degenerative labral tears,”29 “subacromial bursal thickening,”30 or “rotator cuff tendinosis.”30
From a healthcare pathway perspective, overuse of diagnostic tests and subsequent overdiagnosis can trigger a cascade of potentially inappropriate treatments, such as orthopedic surgery, over-prescription of opioids, or premature and intensive rehabilitation protocols as initial management strategies.24, 31 Differentiating between specific patho-anatomical diagnoses may be less critical than selecting appropriate first-line treatment options. It is crucial to evaluate whether current diagnostic practices actually improve patient outcomes.
Evidence Linking Diagnostic Tests to Patient Outcomes
Evidence supporting the idea that diagnostic tests improve patient outcomes in musculoskeletal disorders is limited. A meta-analysis examining the impact of routine diagnostic imaging on patient-reported outcomes for musculoskeletal conditions found 11 trials focusing on low back pain and knee issues. The analysis provided moderate evidence suggesting that routine diagnostic imaging does not significantly improve pain outcomes.32 One study showed that replacing spine radiographs with early magnetic resonance imaging (MRI) in primary care did not lead to improved back-related disability, but it did increase costs and potentially the rate of spine surgery based on MRI findings.27 Another trial found that patients receiving early MRI for low back pain were more likely to be out of work due to disability one year later.24 Similarly, a study found that adding MRI in primary care for younger patients with traumatic knee injuries did not improve knee function after one year.33
These studies indicate that incorporating imaging tests, which are known to frequently reveal asymptomatic structural findings, into the care pathway for musculoskeletal disorders does not translate into better patient outcomes. Instead, it may contribute to overdiagnosis and the overuse of subsequent interventions, such as surgery. Future research should investigate whether implementing current and emerging diagnostic methods (e.g., ultrasound), classification and biomechanical systems (e.g., McKenzie,34 movement system35, or prediction algorithms (e.g., clinical prediction rules) can enhance the overall clinical pathway, leading to improved patient outcomes without the harms of overdiagnosis. In essence, identifying the precise structural or movement impairment may not alter the selection of high-quality first-line treatments necessary for improving outcomes.
Prognosis: An Underutilized Yet Crucial Element
Prognosis, a method of classification focused on predicting future health outcomes,36 is often underemphasized in clinical practice. Prognostic research seeks to determine if a clinical decision will positively influence a patient’s future health trajectory. It is argued that prognostic considerations should be as central as diagnostic investigations because in many cases, “no intervention” is as beneficial as active treatment. Neglecting prognosis in clinical care can lead to adverse effects and overtreatment, as previously discussed.
Much of medical education emphasizes disease diagnosis and treatment. Historically, there has been a strong focus on educating both healthcare professionals and the public about the causes and mechanisms of diseases, as well as refining diagnostic approaches and prescribing effective, diagnosis-linked treatments. We contend that placing equal emphasis on prognosis could mitigate overdiagnosis and overtreatment. For instance, adopting a “watchful waiting” approach for benign conditions that often resolve spontaneously can reduce the risk of harm, unnecessary interventions, and increased patient anxiety. By accurately predicting disease trajectories, we can develop personalized rehabilitation strategies more likely to improve patient outcomes. This approach allows us to differentiate between patients who benefit from watchful waiting and those who require intensive rehabilitation, potentially reallocating healthcare resources to improve access to rehabilitation services.
Interestingly, studies have shown that both physicians and patients often prefer advanced imaging techniques and express greater satisfaction with their care, even when patient outcomes are not improved.27, 33 This presents a significant challenge for clinicians. Conceptual models suggest that receiving a diagnostic label can have physical, psychosocial, and financial repercussions, along with increased treatment burden, exposure to unnecessary tests and treatments, and adverse events that may lead to dissatisfaction with care.31 Patients are often unaware of the potential harms associated with diagnostic labeling. Given that many common musculoskeletal disorders are self-limiting, it is crucial to explore how a watchful waiting strategy can be effectively integrated into clinical practice.
Key Takeaway: Overzealous diagnostic efforts can lead to overdiagnosis and subsequent overtreatment. Emphasizing prognosis, especially for self-limiting conditions, can lead to better overall outcomes and more appropriate care.
Phenotyping: Enhancing Diagnostic Precision for Improved Management
We have demonstrated that current diagnostic labels in musculoskeletal disorders can sometimes negatively impact patient outcomes. To bridge the gap between diagnosis and improved outcomes, it’s essential to address the inherent complexity and variability within broad diagnostic categories. Phenotyping offers a promising approach to better understand musculoskeletal disorders.
Traditionally, “phenotype” referred to the observable characteristics of an organism resulting from the interaction of its genotype and environment.37 In modern science, phenotyping has expanded to include physical, biochemical, and genetic traits, as well as environmental interactions, that contribute to unique observable characteristics.37 In shoulder injury research, phenotyping that considers the interplay between genetic and psychological factors has been used to predict persistent shoulder pain. George and colleagues found that specific single nucleotide polymorphisms[38](#bib0420], when combined with psychological factors, could predict six distinct shoulder impairment phenotypes. They also identified that pain-related genes interacting with psychological factors could predict four shoulder impairment phenotypes.
Other researchers have focused on clinical findings for phenotyping. In knee osteoarthritis (OA), longitudinal studies have significantly advanced our understanding of phenotyping in the last decade. Two research groups, utilizing data from the Osteoarthritis Initiative study and the Amsterdam OA cohort, which included 3494 and 551 participants with knee OA respectively, identified up to five knee OA phenotypes.39, 40 These phenotypes were defined by factors such as radiographic severity of knee OA, knee muscle strength, body mass index, comorbidities, psychological distress, and alterations in pain neurophysiology. The phenotypes were categorized as “minimal joint disease,” “strong muscle strength,” “severe radiographic,” “obese,” and “depressive mood,” all within the same overarching diagnosis of knee OA.
Other researchers have also identified four pain susceptibility phenotypes in 852 participants from the Multicenter Osteoarthritis Study, a cohort of individuals with or at risk of knee OA.41 These phenotypes were based on clinical measures such as pressure pain threshold and temporal summation. The “high sensitization” phenotype was predictive of developing persistent knee pain over two years.
Another study group examined pain and function trajectories following total knee arthroplasty over five years in 689 patients.42 They identified subgroups of patients who experienced persistent pain or functional deficits post-surgery. These trajectories could be predicted by comorbidities and patient-reported psychological or physical measures.
In low back pain research, one group identified five pain trajectories over 12 weeks in 1585 patients seeking care for low back pain (recovery at week 2 or 12, pain reduction without full recovery, fluctuating pain, and persistent high-level pain for 12 weeks).43 Longer pain duration and beliefs about the likelihood of persistence were predictive of delayed or non-recovery from low back pain. High initial pain intensity and longer pain duration were associated with persistent high pain at 12 weeks.
Another study identified up to nine subgroups using 112 characteristics based on patient history and physical examination in primary care settings for low back pain.44 While the predictive capacity of these subgroups for pain intensity, frequency, and disability over 12 months was somewhat improved compared to earlier subgrouping methods, they were also more complex to apply in clinical practice. The authors suggested that future research should focus on determining if these subgroups respond better to targeted treatment approaches.
In a cohort of 682 participants with non-traumatic arm, neck, and shoulder complaints in primary care, one group identified three disability trajectories over two years using the Disabilities of the Arm, Shoulder and Hand questionnaire (DASH).45 They found that certain prognostic variables from the clinical examination, such as high levels of somatization, could predict a trajectory of continuous high disability. In a study of 127 patients with patellofemoral pain, researchers identified three subgroups: “strong,” “weak and tighter,” and “weak and pronated foot,” based on six common clinical measures including flexibility, strength, patellar mobility, and foot posture.46 The authors proposed that these subgroups could guide the development of targeted rehabilitation strategies to enhance patient outcomes.
These studies collectively suggest that multiple phenotypes can exist within a seemingly “single” diagnosis. This implies that patients within the same diagnostic category may experience different outcomes from identical treatments. We also propose that test results may vary within a single diagnosis depending on the patient’s phenotype.
Key Takeaway: Phenotyping, based on patient characteristics, reported outcome measures, and clinical examination findings, can refine our understanding of patient profiles and disease trajectories within a given diagnostic label. Ongoing large-scale longitudinal studies in musculoskeletal disorders promise to yield more insights into identifying clinically relevant patient subgroups.
Conclusion
Higher order thinking is indispensable for diagnostic clinicians, as it enables decision-making that extends beyond mere memorization of facts and application of routine concepts. This masterclass has highlighted how higher order thinking can mitigate interpretation errors associated with standard diagnostic metrics, reduce the risk of overdiagnosis, and clarify that a single diagnosis may encompass multiple distinct phenotypes. Looking ahead, it is crucial to move beyond conventional diagnostic metrics and explore the integration of phenotyping and prognostic evidence to refine targeted care strategies, ultimately enhancing patient outcomes. We are at the nascent stage of understanding the diverse profiles of patients with musculoskeletal disorders. Large cohorts, comprehensive databases, and advanced data analysis tools like artificial intelligence will accelerate our progress in elucidating the intricate links between diagnosis and patient outcomes.
Conflicts of interest
The authors declare no conflicts of interest.
References
1 World Health Organization. International classification of diseases 11th revision. Geneva: World Health Organization; 2018.
2 Walker HK. The origins of disease classification in western civilization. Bull World Health Organ. 2002;80(2):185–6.
3 National Center for Health Statistics. International classification of diseases, 9th revision, clinical modification. Washington, DC: Public Health Service; 1978.
4 Lai ER. Critical thinking: A literature review. Pearson’s Research Reports. 2011;
5 Sternberg RJ. Critical thinking: Its nature, measurement, and improvement. National Inst. of Education; 1986.
6 Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig LM, et al. STARD 2015: reporting diagnostic accuracy studies: explanation and elaboration. BMJ open. 2015;5(10):e007208.
7 Scholten RJ, Lijmer JG, тер Braak EW, де Vet HC, Bossuyt PM, Bouter LM. Diagnostic accuracy of physical examination findings in patients with shoulder disorders: a systematic review. J Clin Epidemiol. 1999;52(11):1081–92.
8 McGee S. Simplifying likelihood ratios. J Gen Intern Med. 2002;17(8):646–9.
9 Riegelman RK, Hirsch RP. Studying a study and testing a test: how to read the medical literature. Little, Brown; 1989.
10 Hermans J, Luime JJ, тер Hofstede Bartelds A, Verhaar JA, Bierma-Zeinstra SM. Does adding physical examination tests improve the diagnostic accuracy of history taking alone in rotator cuff disorders? J Shoulder Elbow Surg. 2006;15(3):333–41.
11 Osti L, Papalia R, Del Buono A, Osti R, Denaro V. Accuracy of clinical tests for anterior cruciate ligament rupture in primary care: a systematic review. Br J Sports Med. 2017;51(4):295–300.
12 Benjaminse A, Gokeler A, van der Meer M, Kerkhoffs GM, Benjaminse A, Gokeler A, et al. Diagnostic accuracy of clinical tests for anterior cruciate ligament rupture: a systematic review. J Orthop Sports Phys Ther. 2006;36(5):267–88.
13 Scholten RJ, тер Riet G, тер Braak EW, тер Weijden PF, тер Vet HC. Accuracy of physical diagnostic tests for assessing rotator cuff pathology: systematic review. J Fam Pract. 1998;47(6):443–54.
14 Williams CM, Henschke N, Maher CG, van Tulder MW, Koes BW, Macaskill P, et al. Red flags to screen for malignancy in patients with low-back pain. Cochrane Database Syst Rev. 2013(1):Cd008636.
15 Verhagen AP, Downie A, Popay SD, Maher CG, Koes BW. Red flags presented in current clinical practice guidelines for low back pain: a systematic review. Eur Spine J. 2016;25(7):2163–75.
16 Scholten RJ, Engelbert RH, тер Riet G, тер Weijden PF, тер Vet HC. Interobserver reproducibility of orthopaedic physical examination in patients with knee complaints. J Manipulative Physiol Ther. 2001;24(2):77–84.
17 Hegedus EJ, Cook C,ựaun M, Beattie P, Nicholas S. Clinical tests for assessing meniscal tears: a systematic review with meta-analysis. J Orthop Sports Phys Ther. 2007;37(10):576–88.
18 Whiting PF, Rutjes AW, Westwood ME, тер Riet G, тер Braak EW, Reitsma JB, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011;155(8):529–36.
19 Elstein AS. Clinical judgment in medicine and health care. Cambridge university press; 2009.
20 Croskerry P. Cognitive forcing strategies in clinical decisionmaking. Ann Emerg Med. 2003;41(1):110–20.
21 Woolf SH, Kuzel AJ, Dovey SM, Phillips Jr RL. Health promotion and disease prevention in primary care: putting it all together. Lea & Febiger; 1996.
22 Moynihan R, Brodersen J. Overdiagnosis and screening: how to recognize the potential for harm. BMJ. 2006;332(7543):605–9.
23 Welch HG, Schwartz L, Woloshin S. Overdiagnosed: making people sick in the pursuit of health. Beacon Press; 2011.
24 Bedigrew AL, Yousefi S, Goertz M, Ladeira CE, Fritz JM, Childs JD. The influence of early MRI on outcomes for patients with acute low back pain in primary care: a prospective observational study. BMC Musculoskelet Disord. 2018;19(1):1–8.
25 Mafi JN, McCarthy EP, Davis RB, Landon BE. Widespread overuse of health care services in U.S.
