Avoiding premature closure in sequential diagnosis

doi:10.1016/S0933-3657(97)00396-5

Artificial Intelligence in Medicine

Volume 10, Issue 3, July 1997, Pages 269-283

https://doi.org/10.1016/S0933-3657(97)00396-5 Get rights and content

Abstract

An important aspect of diagnostic reasoning is the ability to recognise when there is sufficient evidence to enable a working diagnosis to be made and thus avoid the unnecessary risks and costs of further testing. On the other hand, the reasoner must be careful to avoid the error, known as premature closure, of accepting a diagnosis before it is fully verified. In the absence of a more rigorous approach to verification, a pragmatic approach adopted in many programs for sequential diagnosis is to discontinue testing when the probability of the leading hypothesis reaches an arbitrary threshold. Experimental results are presented to illustrate the potential unreliability of this approach. A more reliable way to avoid premature closure is to discontinue testing only when the lower bound for the probability of the leading hypothesis reaches an acceptably high level. For example, a lower bound of 70% means that its probability can never be less than 70% regardless of any evidence that further testing may reveal. Another reason to discontinue testing may be that further evidence can at best increase the probability of the leading hypothesis by a small amount. For example, if the probability of the leading hypothesis is 72%, with a lower bound of 65% and an upper bound of 75%, further testing may be difficult to justify. As these examples illustrate, a termination strategy informed by upper and lower bounds for the probability of the leading hypothesis may help to avoid both premature closure and undue prolongation of the testing process. Finding upper and lower bounds for the probability of a diagnostic hypothesis as each new piece of evidence is obtained is feasible by existing techniques only when the number of remaining tests is small. However, new techniques are presented which can often produce a dramatic reduction in the computational effort required to find the upper and lower bounds. Based on the independence Bayesian framework, the theory presented extends a probabilistic model of hypothetico-deductive reasoning designed to enable programs for sequential diagnosis in medicine to emulate the reasoning processes of human diagnosticians.

References (22)

J. de Kleer et al.
Diagnosing multiple faults
Artif. Intel.
(1987)
G.A. Gorry et al.
Decision analysis as the basis for computer-aided management of acute renal failure
American Journal of Medicine
(1973)
J.P. Kassirer et al.
Towards a theory of clinical expertise
American Journal of Medicine
(1982)
P. Szolovits et al.
Categorical and probabilistic reasoning in medical diagnosis
Artif. Intel.
(1978)
I.D. Adams et al.
Computer aided diagnosis of acute abdominal pain: a multicentre study
Br. Med. J.
(1986)
F.T. de Dombal
The O.M.G.E. acute abdominal pain survey — progress report, 1982
Scandinavian Journal of Gastroenterology
(1984)
F.T. de Dombal
F.T. de Dombal et al.
Computer-aided diagnosis of acute abdominal pain
Br. Med. J.
(1972)
A.S. Elstein et al.
J. Fox et al.
Alternatives to Bayes?
Meth. Info. Med.
(1980)

D.E. Heckerman

Cited by (29)

Current smartphone-assisted point-of-care cancer detection: Towards supporting personalized cancer monitoring
2024, TrAC - Trends in Analytical Chemistry
With the rising incidence of cancer-related mortality, new enabling technologies are necessary to offer comprehensive molecular profiles of patients in order to assist clinicians to establish an early presumptive diagnosis. Biosensors that are technically comparable to conventional laboratory diagnostics have been developed. However, because the manufacture and operation processes of these biosensors were not well adjusted for end-users as patients, the approach is not optimized in these newly built platforms. Hence, smartphone-assisted biosensors have been developed for point-of-care utilization. They are faster, simpler, and more affordable than standard examinations and first-generation biosensors, however, they raise numerous concerns regarding their applicability for early cancer detection. Therefore, this review focuses primarily on cutting edge developments in smartphone-assisted biosensing platforms that are most relevant to early cancer diagnosis, including optical and electrochemical biosensors, and cancer imaging. What is needed to bring this important technology to realization as early cancer diagnosis tool is discussed.
Prevertebral abscess associated with meningitis: Double cause of neck stiffness
2022, IDCases
Citation Excerpt :
For these reasons, we had difficulty in promptly diagnosing the prevertebral abscess early in the course of management. Once a patient is diagnosed with meningitis, cognitive biases may make it difficult to find other sources of infection (the most common cognitive bias responsible for misdiagnosis is “premature closure,” a failure to continue considering reasonable alternatives after an initial diagnosis is reached) [11]. Consequently, the delay of timely drainage could be fatal [5].
While neck stiffness belongs to the classic triad of meningitis symptoms together with fever and altered mental status, it can also be attributed to inflammation from prevertebral space infection. We describe a difficult-to-diagnose case of prevertebral abscess associated with meningitis. Prevertebral space infection, vertebral osteomyelitis, and meningitis are reported to be associated with each other. When a patient presents with an altered mental status due to meningitis, signs and symptoms may be obscured and physicians may be unable to conduct detailed physical examinations or identify symptoms beyond neck stiffness. The threshold for imaging evaluation may need to be lowered for patients at high risk for prevertebral abscess or vertebral osteomyelitis. Physicians need to recognize this clinical entity, as prompt referral to specialists in head and neck surgery is essential for timely drainage.
Overconfident health workers provide lower quality healthcare
2020, Journal of Economic Psychology
Citation Excerpt :
We hypothesize that overconfidence makes providers more likely to mismanage patients in clinical practice, because overestimating one’s medical acumen leads to two undesirable behaviours. First, overconfident providers may gather information less thoroughly, as they overestimate the likelihood that their intuition is correct, and therefore underestimate and possibly neglect other plausible causes of illness (McSherry, 1997; Voytovich, Rippey, & Suffredini, 1985). Second, overconfident providers are unlikely to feel the need for assistance or input, in the form of clinical guidelines or advice from peers (Baumann, Deber, & Thompson, 1991; Berner & Graber, 2008).
While a growing body of evidence suggests that healthcare workers in low and middle-income countries often provide poor quality of care, the reasons behind such low performance remain unclear. The literature on medical decision-making suggests that cognitive biases, or failures related to the way healthcare providers think, explain many diagnostic errors. This study investigates whether one cognitive bias, overconfidence, defined as the tendency to overestimate one’s performance relative to others, is associated with the low quality of care provided in Senegal. We link survey data on the overconfidence of health workers to objective measures of the quality of care they provide to standardised patients – enumerators who pose as real patients and record details of the consultation. We find that about a third of providers are overconfident – meaning that they overestimate their own abilities relative to their peers. We then show that overconfident providers are 26% less likely to manage patients correctly and exert less effort in clinical practice. These results suggest that the low levels of quality of care observed in some settings could be partly explained by the cognitive biases of providers, such as overconfidence. Policies that encourage adequate supervision and feedback to healthcare workers might reduce such failures in clinical decision-making.
History and Physical Examination
2015, Murray and Nadel's Textbook of Respiratory Medicine: Volume 1,2, Sixth Edition
Overconfidence as a Cause of Diagnostic Error in Medicine
2008, American Journal of Medicine
Citation Excerpt :
Of these, premature closure and the context errors are the most common causes of cognitive error in internal medicine.86 Premature closure is narrowing the choice of diagnostic hypotheses too early in the process, such that the correct diagnosis is never seriously considered.117–119 This is the medical equivalent of Herbert Simon's concept of “satisficing.”120
The great majority of medical diagnoses are made using automatic, efficient cognitive processes, and these diagnoses are correct most of the time. This analytic review concerns the exceptions: the times when these cognitive processes fail and the final diagnosis is missed or wrong. We argue that physicians in general underappreciate the likelihood that their diagnoses are wrong and that this tendency to overconfidence is related to both intrinsic and systemically reinforced factors. We present a comprehensive review of the available literature and current thinking related to these issues. The review covers the incidence and impact of diagnostic error, data on physician overconfidence as a contributing cause of errors, strategies to improve the accuracy of diagnostic decision making, and recommendations for future research.
Set theory in diagnostic reasoning
2003, Clinica Chimica Acta
Background: Regarding diagnostic reasoning, the currently taught Bayesian theory is a form of hypothetical-deduction reasoning. Using set theory, we offer syllogism reasoning instead of hypothetical-deductive reasoning and establish an online diagnostic expert system model based on this diagnostic methodology. Methods: Concepts of set theory were employed to demonstrate diagnostic reasoning. ASP, Vbscript and Microsoft Access were used to establish the expert system and data of 50 cases from Shanghai pulmonary hospital were put into the program to test its efficiency. Conclusion: Diagnostic procedure is type of syllogism rather than hypothetical-deductive reasoning.

View all citing articles on Scopus

View full text

PaperAvoiding premature closure in sequential diagnosis

Abstract

Artif. Intel.

American Journal of Medicine

American Journal of Medicine

Artif. Intel.

Computer aided diagnosis of acute abdominal pain: a multicentre study

Br. Med. J.

The O.M.G.E. acute abdominal pain survey — progress report, 1982

Scandinavian Journal of Gastroenterology

Computer-aided diagnosis of acute abdominal pain

Br. Med. J.

Alternatives to Bayes?

Meth. Info. Med.

Paper
Avoiding premature closure in sequential diagnosis