Table 3.

Methodological Details of Studies Measuring Diagnostic Uncertainty

Author, Publication Year	Article Type and Study Design	Study Objectives	Measurement Methods and Tools Used
Anstee, 197761	Longitudinal hospital-based study	Objective: To follow up patients with diagnostic uncertainty and to understand clinical and demographic factors associated with it	Unsolved diagnosis at discharge from general inpatient unit
Allman, 198543	Physician survey	Objective: To gain insight into the individual physician’s tolerance for diagnostic uncertainty	Pre- and post-liver-spleen scan probability estimates provided by physicians were used for identification of diagnostic uncertainty.
Johnson, 198862	Vignette study with patients	Objective: To evaluate the impact of communicating uncertainty on patient visit satisfaction	Video clips of patient–physician encounters denoting physician expressions of uncertainty were shown to patients, and a questionnaire was subsequently used to evaluate patient responses.
Moskowitz, 198863	Transcript of the physician’s encounter	Objective: To generate hypotheses regarding how physicians make difficult clinical decisions	Case presentations with details about difficult diagnoses were presented to 3 physicians and were used as measures for the presence of uncertainty.
Lossos, 198964	Retrospective analysis of patient records and administrative data	Objective: To operationalize a definition for deferred diagnosis (when diagnosis is unclear) and describe its clinical spectrum	The primary diagnosis for the visit denoted with ICD-9 codes for signs, symptoms and ill-defined conditions was used to identify uncertainty.
Buntinx, 199165	Physician survey	Objective: To compare the initial diagnosis made immediately after physical examination by a general practitioner with final diagnosis made between 2 weeks and 2 months later	Initial diagnosis was captured with a score on the certainty of the diagnosis: “unknown,” “suspected,” “probable” or “certain.”
Gerrity, 1990, 1992 & 199540,41,66	Physician survey	Objective: Development and refinement of an instrument to measure uncertainty in physicians	23-item Physicians’ Reaction to Uncertainty (PRU) scale was developed, containing subscales for stress from uncertainty, anxiety and reluctance to disclose uncertainty.
Zaat, 199210	Physician survey	Objective: To explore the relationship between physician uncertainty, risk-taking attitude and laboratory test use	Three categories reflecting different levels of diagnostic uncertainty (uncertain, moderately certain and certain) were evaluated using a questionnaire with two 5-point scales. The first scale (14 items) evaluated the extent of the physician’s self-reproach. The second scale (5 items) assessed opinions about risk avoidance.
Hewson, 199667	Study of patient–physician encounter using standardized patients	Objective: To identify strategies involved in diagnosis and treatment plans for primary care problems that are uncertain and complex	Study of primary care physician interactions with standardized patients who portrayed typical primary care problems involving uncertainty and complexity
Lave, 199768	Physician survey	Objective: To determine whether hospital staff’s diagnostic uncertainty is a predictor of hospital resource use	A visual analog scale (0, 25, 50, 75 and 100) was used to measure diagnostic uncertainty.
Allison, 19989	Physician survey	Objective: To explore the association between PCPs’ attitudes toward risk-taking and uncertainty	A 23-item Physicians’ Reaction to Uncertainty (PRU) scale was used.
McKinlay, 199869	Physician survey	Objective: To determine the presence of diagnostic uncertainty and its association with variation in diagnosis and patient factors	A scale of uncertainty from 0 to 100 was used.
Patton, 199870	Physician survey	Objective: To develop a 5-point scale to identify uncertainty in pediatrics	Levels of certainty of diagnosis were captured: very little, some, moderate, substantial, sufficient.
Shaner, 199871	Retrospective analyses of patient charts	Objective: to determine the sources and frequency of diagnostic uncertainty for patients with chronic psychosis and active cocaine abuse or dependence	Cases where a provisional diagnosis or a list of alternate diagnoses was used
Gordon, 200072	Physician survey Transcript analyses using audiotapes of patient–physician encounters	Objective: To identify physician expressions of uncertainty during actual clinic visits and to examine their associations with physicians’ and patients’ characteristics and communication behaviors	A 23-item Physicians’ Reaction to Uncertainty (PRU) scale was used. Physician expression of uncertainty was determined by trained coders. Direct and unambiguous statements included “I do not know” or “it’s not clear.” Indirect expressions (e.g., it was “pretty much” normal or there is a “good chance” it’s normal) were not used due to lack of reliability.
Ogden, 200299	Physician perspectives about behavioral and verbal expressions of uncertainty Vignette study with patients	Objective: To explore the impact of doctors’ expressions of uncertainty for a range of diagnostic and management aspects of the consultation	Behavioral expressions of uncertainty were used—e.g., used a book or computer to find out about a condition or a drug, asked another GP or nurse for advice or referred the patient to a hospital. Verbal expressions of uncertainty were also used—e.g., “I am not sure about this,” I need to find out more,” “let’s see what happens,” “I do not know,” “I have not come across this before,” “I think this might be….” Patients rated the expression of uncertainty.
Weijden, 200273	Qualitative interview with physicians regarding the patient–provider encounter	Objective: To understand the general practitioner’s (GP’s) motives for ordering laboratory tests for patients presenting with unexplained symptoms	Qualitative assessment of physicians’ perceptions of specific diagnosis was conducted immediately after a consultation with a patient.
Friedman, 200574	Physician survey	Objective: To explore the alignment between physicians’ confidence in their diagnoses and the “correctness” of these diagnoses, as a function of clinical experience, and whether subjects were prone to over- or under-confidence	To assess confidence, physicians rated the likelihood that they would, at the time they generated the differential, seek assistance in reaching a diagnosis.
Resar, 200675	Retrospective analyses of patient charts	Objective: Use of a trigger tool to identify adverse events in ICUs	Use of more than 3 consultations as an indicator of diagnostic uncertainty
Bovier, 200776	Physician survey	Objective: To describe sociodemographic and professional characteristics of reactions to uncertainty among physicians from all specialties, including physicians in training	Reaction to medical care uncertainty was measured with the Anxiety Due to Uncertainty and Concern About Bad Outcomes scales.
Green, 200816	Physician survey	Objective: To explore the association between the presence of diagnostic uncertainty and adverse events	A certainty estimate of either ≤20% or ≥80% was classified as clinical certainty, while estimates between 21% and 79% were defined as clinical uncertainty.
Farnan, 200877	Qualitative interview of physicians	Objective: To describe types of uncertainty faced by residents and strategies employed to manage uncertainty and effects on patient care	Using critical incident technique, residents were asked to recall important clinical decisions during a recent call night, with probes to identify decisions made during uncertainty.
Swoboda, 200878	Physician survey	Objective: To examine the clinical decision-making involved in diagnosing contested illnesses (chronic fatigue syndrome, multiple chemical sensitivities and Gulf War syndrome)	Physician respondents were asked whether sufficient knowledge existed for determining legitimacy of contested illnesses.
Blanch, 200979	Patient vignette study	Objective: To examine the consequences of expressions of uncertainty (EOUs) in medical student interactions, with a particular focus on the gender of the expressor	Videotaped interactions were shown to patients. Expressions of uncertainty were used to denote uncertainty. The patients were asked to rate whether the doctor sounded sure of himself/herself on a 10-point scale from 1 = definitely no to 10 = definitely yes.
Evans, 200957	Physician survey	Objective: To investigate the relationship between primary care physician’s stress reactions to uncertainty and conceptual resource of epistemology	Stress reactions to uncertainty were measured using subscales of the Physician Reaction to Uncertainty Scale (PRUS): “Anxiety due to uncertainty” (5 items) and “Concern about bad outcomes” (3 items).
Nevalainen, 201080	Qualitative study with medical students	Objective: To investigate how medical students experience uncertainty during their first clinical years and how their feelings develop with time as they progress from the 3rd to the 4th year	Qualitative assessment of uncertainty in reflective learning diaries and writings collected during 3rd and 4th years of medical studies
Politi, 201081	Physician survey	Objective: To identify variables associated with physicians’ anxiety from uncertainty and reluctance to disclose uncertainty to patients	Scales for “anxiety from uncertainty” and “reluctance to disclose uncertainty” were used.
Schneider, 2010, 201442,82	Physician survey	Objective: Development and refinement of the Dealing with Uncertainty Questionnaire (DUQ) and the Communicating and Dealing with Uncertainty (CoDU) questionnaire	The Dealing with Uncertainty Questionnaire (DUQ) was refined, and 4 CoDU scales were identified: “communicating uncertainty,” “diagnostic action,” “intuition” and “extended social anamnesis.”
Portnoy, 201183	Physician survey	Objective: To explore the association between physicians’ attitudes about communicating and managing scientific uncertainty, and their perceptions of negative patient reaction to uncertainty	Four items focusing on physicians’ attitudes toward communicating and managing scientific uncertainty on a 5-point Likert scale from “strongly agree” to “strongly disagree.” Ambiguity Aversion in Medicine Scale was also used.
Hanauer, 201284	Retrospective analysis of patient EHR notes	Objective: Quantified the use of uncertainty expressions from institutions’ EHR	Diagnostic uncertainty identified using 313 distinct uncertainty expressions, e.g., “could be,” “most likely,” “probably,” in the EHR note
Sarkar, 201285	Clinician survey	Objective: To survey primary care practitioners about potential barriers to timely diagnosis in outpatient settings and diagnostic difficulty faced in their practices	Extent of perceived diagnostic difficulty (uncertainty) was determined using the question, “In the past year, about what percentage of your patients did you consider difficult to diagnose?”, with 5 ordered responses of 0%, 1–5%, 6–10%, 11–15% and >15%.
Cousin, 201338	Experimental study with patients Field study with videos of physician during patient–physician encounter	Objective: To determine how physician-expressed uncertainty affects patient satisfaction in relation to both the physician and patient gender	Vignettes using statements of physician expression of uncertainty (e.g., “I cannot tell you” or probability words such as “maybe,” “probably”) Independent coder rated the presence of diagnostic uncertainty using 11-point Likert scale (0 = no uncertainty, 10 = total uncertainty).
Meyer, 201386	Physician vignette study	Objective: To evaluate how physicians’ diagnostic calibrations, defined as the relationship between diagnostic accuracy and confidence in that accuracy, change with evolution of the diagnostic process and with increasing diagnostic difficulty of clinical case vignettes	Diagnostic confidence was measured on a scale of 0 to 10.
Whaley, 201387	Retrospective analysis of patient EHR notes	Objective: To measure antibiotic prescribing rate, prevalence of diagnostic complexity and uncertainty that clinicians face when treating patients with acute cough	Documentation in the EHR note was used to evaluate the presence of diagnostic uncertainty, e.g., use of words such as “maybe” or “unclear” or question marks in association with the diagnosis. Differential diagnosis was also used (e.g., pneumonia vs. acute bronchitis).
Bosner, 201449	Qualitative interview of physicians	Objective: To study the early diagnostic phase of the decision-making process when no specific diagnosis is reached in patients presenting with headache	Qualitative assessment of physicians’ perceptions of uncertainty
Cunningham, 201488	Physician survey	Objective: To investigate whether physician anxiety due to uncertainty is associated with a higher propensity to use race in medical decision-making	Anxiety Due to Uncertainty (ADU), a 5-item measure of emotional reactions to clinical uncertainty, was used.
Lanham, 201489	Qualitative interview of physicians	Objective: To study differences in individual physicians’ EHR use patterns and identify perceptions of uncertainty as an important variable in understanding EHR use	Qualitative assessment of physicians’ perceptions of uncertainty
Nevalainen, 201490	Physician survey	Objective: To investigate medical students’ feelings about facing uncertainty in medical decision-making and the associations of tolerance of uncertainty using demographic factors, students’ fears of making mistakes and views of a GP’s work	Questionnaire about student’s views on how they felt about and tolerated uncertainty
Serbic, 201491	Experimental study with patients (mixed factorial design)	Objective: To examine the relationship between diagnostic uncertainty and recall bias in 2 groups of chronic low back pain patients	Patients reported diagnostic uncertainty using the question, “I think there is something else happening with my back which the doctors have not found out about yet.” (yes/no)