Factors affecting interpretation of dental radiographs

Abstract

Objectives:

To identify the factors influencing errors in the interpretation of dental radiographs.

Methods:

A protocol was registered on Prospero. All studies published until May 2022 were included in this review. The search of the electronic databases spanned Ovid Medline, PubMed, EMBASE, Web of Science and Scopus. The quality of the studies was assessed using the MMAT tool. Due to the heterogeneity of the included studies, a meta-analysis was not conducted.

Results:

The search yielded 858 articles, of which eight papers met the inclusion and exclusion criteria and were included in the systematic review. These studies assessed the factors influencing the accuracy of the interpretation of dental radiographs. Six factors were identified as being significant that affected the occurrence of interpretation errors. These include clinical experience, clinical knowledge, and technical ability, case complexity, time pressure, location and duration of dental education and training and cognitive load.

Conclusions:

The occurrence of interpretation errors has not been widely investigated in dentistry. The factors identified in this review are interlinked. Further studies are needed to better understand the extent of the occurrence of interpretive errors and their impact on the practice of dentistry.

Introduction

Making a diagnosis is a critical clinical decision and has implications for clinicians and their patients. The diagnostic process involves gathering data from patient history and clinical examination, performing diagnostic tests, and arriving at a diagnosis by interpreting and integrating the findings. ¹

In clinical dentistry, radiographs provide useful diagnostic information about diseases and pathologies of the teeth and jaws. They are routinely used in the clinical diagnosis of problems and contribute to treatment planning. ^2,3 Evaluation of a radiographic image involves visual inspection and interpretation. ⁴ Making a diagnosis from an image requires four steps: (a) detection- identifying a finding that would require further analysis. (b) recognition-identification of pathology, (c) discrimination- characterisation of the lesion, and (d) diagnosis and differential diagnosis. ⁵ The detection and discrimination of pathology in a radiographic image involves both perceptual (recognising a difference or change in appearance) and cognitive (understanding the significance of these changes) processes. The clinician must first detect pathology on a radiograph and then characterise it, leading to a diagnosis. However, this is a complex process, and decisions are often made based on incomplete clinical information. In some situations, inaccurate thinking processes can lead to errors in diagnosis. Due to the analytical complexity of working with partial information, diagnostic errors and errors of interpretation of radiographs are frequently unavoidable in day-to-day practice. ⁶

Errors of interpretation have been described as a discrepancy in the interpretation that significantly differs from the consensus of one’s peers. ⁷ The errors in interpreting radiographs have been studied in medicine. ^8–12 While interpretive errors and the factors affecting them could be the same in dentistry, they have not been assessed with the same rigour as in medicine and are therefore not well understood.

Aims

This systematic review aimed to identify the factors contributing to errors in interpreting radiographs by dental practitioners.

Objectives

1. Identify factors that influence or cause the errors in interpretation

2. Identify the relative significance of those factors

Methods

Protocol and registration

This systematic review followed PRISMA guidelines ¹³ and was registered on the National Institute of Health Research Database (www.crd.york.ac.uk/prospero, protocol registration number CRD42020207998).

Eligibility criteria

Inclusion criteria

1. Studies that examined factors influencing interpretation errors or diagnostic errors among dental professionals

2. Published in the English language

3. Articles published in peer-reviewed journals.

4. All clinical settings.

5. Reference lists of selected papers from the search

Exclusion criteria:

1. Case reports, conference proceedings, letters to the editor and news articles

Data sources

The search focused on the factors influencing interpretation errors of radiographs in clinical environments with no restrictions to study design. Ovid Medline, PubMed, EMBASE, PsycINFO, the Cochrane Library (Cochrane Database of Systematic Reviews, Cochrane Central Register of Controlled Trials (CENTRAL), Cochrane methodology register), Scopus, Web of Science (including Science Citation Index Expanded) and EBSCO databases were systematically searched. The search was conducted between February 2021 and March 2021 and then updated in May 2022. The databases were searched for articles published between 1946 and May 2022. The search was limited to articles published in English.

Search strategy

The search strategy was developed to include relevant concepts specific to dental radiographic interpretation errors. The databases were searched using keywords, Medical Subject Headings (MeSH), Emtree terms and free-text terms relating to radiology, dentomaxillofacial radiology, dentistry, errors of interpretation, diagnostic errors and clinical decision making. The Ovid Medline search used a combination of Mesh terms, and the same strategy was used for searching other databases. The Ovid database search strategy is included in Table 1. The potential factors affecting errors of interpretation were identified from the medical radiology literature and adapted to the dental context. The search strategy was validated by checking if seminal papers were captured in the database search. The conceptual structure of the search strategy is described in Table 2.

Table 1.

search strategy

Dentistry/ or dentists/ or dentist* or medical radiologist* or dental radiolog* or dentomaxillofacial radiolog* or dental imaging or dental X-ray* or radiographer* or panoramic radiograph* or OPG or orthopantomograph* or intraoral radiograph* or intraoral radiograph*).mp.
Anxiety/ or Burnout, Professional/ or Occupational Stress/px or fatigue or anxiety or depression or stress or burnout or workload or occupational stress or occupational burnout or job-related stress or job stress* or human error* or professional error* or time pressure or “not taking adequate time” or professional experience.mp.
Clinical Decision-Making/ or Decision Making/ or Diagnostic Self Evaluation/ or Diagnostic Errors/ or (diagnostic error* or misdiagnos#s or mis-diagnos#s or missed diagnos#s or delayed diagnosis or diagnos* mistake* or radilogical error* or interpretive error*).mp.

Table 2.

Conceptual structure of search strategy

Concept 1	Concept 2	Concept 3
dentist and dental imaging	factors causing errors of interpretation	consequences
Keywords: dentist, dentomaxillofacial radiology, dental radiology, dental imaging, panoramic, intraoral imaging	Keywords: stress, occupational stress, fatigue, burnout, excessive workload, attentional bias, case complexity, inattention, time pressure, clinical experience	Keywords: errors in decision making, problems in diagnosis, diagnostic errors, missed diagnosis, errors of omission, interpretive errors, errors of interpretation clinical decision making, image perception, delayed diagnosis, observer variation, radiological error,
MeSH terms: “Dentists”[Mesh] OR “Radiography, Dental, Digital”[Mesh] OR “Radiography, Panoramic”[Mesh]	MeSH terms: “Occupational Stress”[Mesh] OR “Stress, Psychological”[Mesh] OR “Fatigue”[Mesh] OR “Burnout, Professional”[Mesh] OR “Workload”[Mesh] OR “Attentional Bias”[Mesh]	MeSH terms: ((((((“Diagnostic Errors”[Mesh]) OR “Missed Diagnosis”[Mesh]) OR “Medical Errors”[Majr]) OR “Medical Errors”[Mesh]) OR “Clinical Decision-Making”[Mesh]) OR “Delayed Diagnosis”[Mesh]) OR “Observer Variation”[Mesh]

Data collection process and data items

Full texts of those articles that met the inclusion criteria were obtained, and data were extracted. A data extraction template was developed on Covidence (Covidence systematic review software, Veritas Health Innovation, Melbourne, Australia. Available at www.covidence.org). The data extraction form was tested using two included papers, and necessary modifications were made before extraction. Two reviewers (SH and JG) independently extracted the data using Covidence, and any conflict was resolved through discussion.

Methodological quality

The methodological quality was assessed using the Mixed Methods Appraisal Tool (MMAT). ¹⁴ The MMAT is a validated critical appraisal tool designed to appraise systematic mixed studies reviews. It permits the assessment of the methodological quality of five categories of studies: qualitative research, randomised controlled trials, non-randomised studies, quantitative descriptive studies, and mixed methods studies. ¹⁴ One reviewer (SH) independently performed the quality assessment, and then the second reviewer (JG) validated the information, and any disagreement was resolved through discussion.

Depending on the MMAT criteria, papers were grouped into high, medium, or low quality. For each study, the overall quality score was calculated using the MMAT. Based on the MMAT tool, studies were scored as unclassified, 20%, 40%, 60%, 80%% or 100%, with scores of 80–100% considered high quality. ¹⁴

Publication and reporting bias were minimised by searching multiple databases and including studies from those sites. Studies with low methodological quality were not excluded but were described in the analysis and synthesis.

Results

The results of this review were tabulated and described narratively. The eight studies in the systematic review highlighted the following factors as being significant for interpretation errors of dental radiographs: clinical experience, clinical knowledge and technical ability, time pressure, cognitive load, case complexity on dental radiographs, and training (duration of the didactic program). One study ¹⁵ did not fit these categories and was described separately.

Meta-analyses were not attempted due to the heterogeneity of the raw data due to widely varying study settings and study designs.

Study selection

Eight hundred fifty-eight articles were obtained from searching the databases and reference lists of selected papers. These articles were imported into Covidence for screening and data extraction. After the duplicates were removed, 336 articles were available for the title and abstract screening. Two reviewers (SH and JG) independently screened the titles and abstracts to identify studies that met the inclusion criteria. All studies deemed eligible by both reviewers were subjected to a full article review to determine inclusion in the final analysis, and eight articles met the inclusion criteria. They were included in the final data extraction. Any differences about the eligibility of inclusion were resolved by discussion between the two reviewers. The screening process is shown in the PRISMA flow diagram, ¹³ Figure 1.

Figure 1.

Prisma flow diagram for study identification, screening, eligibility, and inclusion.

Study characteristics

An overview of the eight studies included in this review is shown in Table 3.

Table 3.

data extraction table

	Author, year and country	Aim	Study design	Participant sample size	Image type and image sample size	Participants	Exposure (factor investigated)	Comparator	Outcome (Effects of the factor studied)	Conclusion
1	17(17), The Netherlands	To compare the diagnostic accuracy of dental students and general dental practitioners	Nonrandomised quantitative	Dentists = 273 Dental students = 259	Bitewing radiographs of extracted teeth showing 105 interproximal surfaces	Dentists	Clinical experience Years of experience = not stated	fourth-year dental students from three consecutive cohorts	diagnostic accuracy of dentinal caries on bitewing radiographs	Clinical experience increased specificity but inversely affected the sensitivity of radiographic diagnosis of dental caries, with dental students having higher sensitivity scores than dentists.
2	18(18), Brazil and Switzerland	To assess the influence of clinician experience on the reproducibility and accuracy of radiographic examination for detection of occlusal caries.	Nonrandomised quantitative study	Dentists = 10 Dental students = 10	Bitewing radiographs of 166 extracted permanent teeth	Brazilian and Swiss dentists	clinical experience Years of experience of dentists = 5 to 7 years location of dental education/training programs	Brazilian and Swiss dental students	accuracy (reproducibility and validity) of caries diagnosis at two thresholds	The clinical experience affected the sensitivity and specificity of radiographic caries diagnosis Clinical experience increased specificity. This paper did not specify how the education program (location) affected diagnostic accuracy
3	16 (16), USA	To identify and distinguish treatment related injuries that are avoidable from those that are sequelae of a dental problem	Retrospective cohort study (qualitative)	92 dental malpractice claims Poor interpretation and incorrect prescription of radiographs = 8 cases	not applicable	92 dental malpractice claims filed between 1900 and 1974	Knowledge and technical ability in radiography	No comparator	Dental adverse events and degree of avoidability of adverse events	Poor quality films and interpretation resulting in diagnostic errors were considered avoidable outcomes.
4	19(19), USA	To compare the effect of differing levels of didactic education and clinical experience among first- and senior-year dental students and the dental faculty on the diagnosis of occlusal caries on radiographs	Nonrandomised quantitative	Dentists = 15 Dental students in the first year = 15 Dental students in senior year = 14	Bitewings of 100 extracted teeth for assessment of occlusal caries	Dental faculty	clinical experience Years of experience of the dental faculty-17 years levels of didactic education (first and senior year students)	first-year and senior-year dental students	Confidence of presence/absence of caries using a 5-point Likert scale)	Clinical experience affected the sensitivity and specificity of caries diagnosis, with the specificity of radiographic caries diagnosis increasing with increasing experience. There was no significant difference in levels of accuracy between the first year and senior students
5	20(20), Germany	To compare the pupillary responses (pupil diameter, gaze hit mapping) of experts and student dentists to panoramic radiographs of varying difficulties.	Nonrandomised quantitative study	Dentists = 26 Dental students = 50	20 patient panoramic radiographs for students and 15 for experts	Dentists (GP and specialists) working at the university clinics	Cognitive load clinical experience case (image) difficulty/complexity years of experience of dentists = 10 years on average	Third-year dental students	Cognitive load was measured using pupil diameter. An increase in median pupillary diameter from baseline and variation with the level of difficulty on the radiograph was measured	More experienced clinicians show a proportional increase in pupillary response to increasing case difficulty. Dental students had a consistent increase in pupillary diameter regardless of difficulty level.
6	23(21), Brazil	To assess the ability of three observers to accurately confirm the existence and absence of MB2 canals in human first upper molars with different root conditions and the agreement among three observers.	Nonrandomised quantitative study	Dentists (OMR registrars) =3	CBCT of 82 extracted maxillary first molars was used. The teeth were grouped into three depending on the condition of MB1 as non-filled, filled and deobturated	Dentists (OMR registrars) in the oral and maxillofacial radiology program.	clinician experience Years of experience = reviewer one five years’ experience, reviewer two three years’ and reviewer three one year	Comparison of interobserver agreement	Agreement on identifying the presence of MB2 (second mesiobuccal canal) under different conditions of MB1 (non-filled, filled, and deobturated)	Agreement of MB2 diagnosis increased with clinical experience
7	22(22), UK	To explore the role of time pressure on the radiographic diagnostic performance when viewing bitewing radiographs among dentists	Randomised cross-over study	Dentists = 40 randomly assigned to four groups of ten each	12 patient bitewing radiographs Each bitewing had a range of difficulty (three easy, three difficult)	Primary care dentists (GP) Years of experience = 17 years (average)	time pressure	time pressure versus no time pressure time	radiographic diagnostic performance- diagnostic error	Under time pressure, the sensitivity was lower, but the specificity was not affected. Time pressure increased the incidence of diagnostic errors and errors of omission.
8	Bussaneli DG et al, 2014 (23), Brazil	To evaluate the influence of the examiner’s clinical experience on the detection and treatment decision of caries lesions in primary molars.	Nonrandomised quantitative study	Dental faculty = 3 Dental students = 3	Bitewing radiographs of 77 recently extracted or exfoliated primary molars.	Dentists from the department of paediatric dentistry.	clinical knowledge (ICDAS classification on radiographs) Clinical experience Years of experience of the dental faculty = 10 years (group A). Dental students (group B) were familiar with the ICDAS criteria	Second-year dental students	Accuracy of caries diagnosis using clinical evaluation and bitewing radiographs.	The professional experience did not affect the accuracy of caries diagnosis on bitewings.

Participants and settings

Dentists (clinicians) and dental students were participants in five of the included studies. ^16–20 One study was a retrospective analysis of dental records. ¹⁵ The study location varied among the included articles, with two studies based in the USA, ^16,17 two in Brazil, ^18,20 one in the Netherlands, ¹⁶ one in Germany, ²¹ one in the UK, ¹⁹ and 1 in two locations-Brazil and Switzerland. ²² 5 of the studies used radiographs of extracted teeth. ^{16–18,20,22} Two studies used patient radiographs. ^19,22 The outcomes were similar for five studies, and they examined the diagnostic accuracy of dental caries on bitewing radiographs. One study ¹⁹ used panoramic radiographs for patients, and the other ²¹ used CBCT images to detect MB2 of maxillary molars.  The final study ¹⁵ used malpractice claims to assess the extent of preventable adverse events and included radiographic interpretation errors.

The majority (62%) of the studies were conducted entirely in a university clinic setting, ^17–21 with one using both private practice and a university setting. ¹⁶ One study ¹⁵ was a retrospective audit conducted by a panel of dentists and dentists and physicians with legal training in private practice.

Quality assessment

The quality assessment of all included studies is shown in Table 4. This review has three categories of studies: non-randomised quantitative studies, ^16–19,21 , a quantitative randomised cross-over study ²² and a quantitative descriptive study. ¹⁵ In the MMAT checklist, there were five questions for each study type, to which the response was ‘yes’, ‘no’ or ‘can't tell’.

Table 4.

Quality assessment

	Quantitative randomised Cross-over					Quantitative non-randomised					Quantitative descriptive					Quality
References	2.1	2.2	2.3	2.4	2.5	3.1	3.2	3.3	3.4	3.5	4.1	4.2	4.3	4.4	4.5
Mileman 2002						*	*	*	-	-						medium
Diniz 2010						-	*	*	-	*						medium
Lazarchik 1995						-	*	*	-	*						medium
Castner 2020						-	*	*	-	*						medium
Vizzotto 2015						-	-	*	-	*						low
Bussaneli 2015						-	*	*	Can't tell	*						medium
Plessas 2019	*	*	*	-	Can’t tell											medium
Milgrom 1975											*	*	Can't tell	*	-	medium

1*meets 20% of MMAT criteria

2**meets 40% of MMAT criteria

3***meets 60% of MMAT criteria

4****meets 80% of MMAT criteria

5*****meets 100%of MMAT criteria

20-40%=low quality; 60–80% = medium quality; 80–100% = high quality

Seven ^15–20,22 of the eight studies were deemed medium quality with a score of 60%, and one study ²¹ was rated low quality with a score of 40%.

The participants did not represent the target population in four studies due to limited sample sizes. ^17–19,21 While all seven studies ^16–22 had dentists and dental students as participants, the study design did not specify the inclusion and exclusion criteria used, what methods were used to achieve a sample size of participants representing the target population, sample size calculation and whether the sample size was adequate for the target population. In addition, no attempt was made to minimise confounding bias either by using large sample sizes, stratification, or statistical control.

Five of the six non-randomised quantitative studies ^16–19,21 conducted measurements for outcomes and intervention (or exposure). Clinical experience was considered as exposure in these studies. The outcome was the diagnostic accuracy of radiographic findings and was measured using sensitivity and specificity of diagnosis. However, none of the five studies in this category accounted for confounders in design and analysis, such as factors influencing clinical experience, namely type of clinical practice, the dental clinic’s location (urban, rural), number of patients seen each day, number of radiographs diagnosed each day. The influence of changing diagnostic standards and criteria and evolving scientific knowledge on clinical experience were also not considered.

The quantitative randomised cross-over study by Plessas et al. (2019) ²² did not include blinding in the study design, the outcome assessors (participants) were not blinded to the intervention, and it was unclear whether the participants adhered to the assigned intervention.

The quantitative descriptive study ¹⁵ did not describe a sampling strategy and did not contain statistical analysis, and it was unclear whether the outcome measurements were appropriate. The variables were defined, but there was no description of how the variables were measured. The quality assessment of all studies using MMAT ¹⁴ is shown in Table 4.

Results of Synthesis

This review identified six factors influencing interpretation errors: the complexity of dental radiographs, clinical knowledge and technical ability, clinical experience, cognitive load, time pressure and dental training and education programs. These factors appear interrelated and contribute to the occurrence of interpretation errors.

The complexity of dental radiographs

The complexity of a radiograph describes the level of difficulty in detecting pathology and interpreting the findings. The complexity of a radiograph depends on the type (bitewings, panoramic radiographs, CBCT) and quality of radiographs (image contrast, presence of film faults), the type and incidence of pathology identified on the radiographs, and the conspicuity of the findings. ²³ This review found that interpretation errors were more likely to occur with radiographs of higher complexity.

In the studies in this systematic review, dental radiographs with various complexities were considered to evaluate the influence of complexity on interpretation errors. Most studies (50 %) used in vitro bitewing radiographs of extracted teeth to evaluate a single type of pathology, dental caries. ^16–18,20 The study by Plessas et al ²² used patient bitewing radiographs, and pathological findings included dental caries and periodontal disease.

In contrast, Vizzotto et al ²¹ utilised cone beam computed tomography (CBCT) images of extracted maxillary molars to study the diagnostic accuracy of identifying the second mesiobuccal canal (MB2) of maxillary molars. The study by Castner et al ¹⁹ directly assessed the effect of case complexity on diagnostic accuracy. They used patient panoramic radiographs and categorised the images depending on the difficulty of interpretation, which relied on the prevalence of the lesions and ease of detection (lesion conspicuity). They found that the diagnostic accuracy depended on the level of complexity of findings on the radiographs.

The seven studies described here measured the diagnostic accuracy using sensitivity and specificity. The type of image- bitewings, panoramic radiographs or CBCT, in vitro or patient radiographs and a variety of pathologies studied varied in the studies included in this review. These factors can affect the case complexity of dental radiographs and influence the accuracy of diagnosis and the occurrence of interpretation errors. In a clinical situation, the interpretation of patient radiographs may have a higher complexity than the radiographs of extracted teeth (in vitro images) used in experimental settings, as patient radiographs may show both expected and incidental findings. In addition to the artificial reading environment, this could influence diagnostic accuracy.

Clinical knowledge and technical ability in dental radiography

Clinical knowledge is described as scientific knowledge about diseases, their pathophysiology, bodily processes, appropriate diagnostic tests and therapeutic measures. ²⁴ In contrast, technical ability in dental radiology refers to the skills of dental radiography and patient management in dental radiology. This review found that clinical knowledge and technical skills affected the accuracy of radiographic diagnosis.

Two studies ^15,20 considered the influence of clinical knowledge and technical ability on the occurrence of interpretation errors.  The study by Milgrom et al ¹⁵ found that clinical knowledge and technical ability in dental radiography influenced the occurrence of diagnostic errors. The attributes related to clinical knowledge included poor interpretation and inappropriate use of radiographs. Taking poor quality films and failing to protect the patient were identified as technical abilities affecting the incidence of errors. All these were considered avoidable events. ¹⁵ The study by Bussanelli et al ²⁰ discussed the effect of knowledge in a limited manner, focusing only on the extent of knowledge of ICDAS coding and experience using it for caries diagnosis. They found that the understanding of and experience in using the caries classification increased the accuracy of caries diagnosis.

Time pressure

Time pressure is described as the psychological stress resulting from a time shortage to complete tasks. ²⁵ Time pressure has been identified as a contributor to diagnostic errors in medical radiology. ²⁶ Plessas et al ²² investigated the effect of time pressure on the accuracy of the radiographic interpretation of bitewing radiographs. They found that time pressure resulted in lower sensitivity and diagnostic accuracy, leading to an increased occurrence of interpretive errors. None of the other studies considered the effect of time pressure by including a time limit in their study design; therefore, a comparison of the impact of time pressure could not be made in those studies.

Dental training and education programs (location and duration)

Dental training programs vary in content and length in different parts of the world. The location and extent of training were discussed in two papers. ^17,18 In the study by Diniz et al, ¹⁷ dentists and dental students in Brazil and Switzerland were compared. The two education programs varied in length of training. The authors concluded that the differences in sensitivity of radiographic caries diagnosis between the student groups could be attributed to the differences in training in the two countries. Lazarchik et al ¹⁸ studied the effect of the level of training by comparing the accuracy of radiographic caries diagnosis among dentists, first year and senior dental students. They found that increase in years of training (first-year versus senior-year dental students) was associated with improved accuracy of radiographic diagnosis of caries. The remaining studies in this systematic review did not consider the effect of the location and duration of dental training and education program on diagnostic accuracy and the occurrence of interpretation errors. It is, therefore, not possible to conclude if the type and location of dental education influenced the accuracy of radiographic diagnosis.

Clinical experience

Six studies examined the impact of clinical experience on diagnostic accuracy. ^16–21 Clinical experience was described as years since graduation ¹⁸ or the number of years in clinical practice. ¹⁹ The influence of clinical experience on diagnostic accuracy was measured using sensitivity and specificity of radiographic caries diagnosis. Of the six studies, five (over 80 %) found a strong correlation in that clinical experience increased diagnostic specificity but not sensitivity. These studies also found higher sensitivity of radiographic caries diagnosis among participants with less clinical experience (dental students). Overall, these studies found a higher rate of errors in interpretation among less experienced clinicians. However, Bussanelli et al ²⁰ found no correlation between clinical experience and the accuracy of radiographic caries diagnosis. They reported that experienced clinicians were more confident about their diagnosis than novice clinicians. In Mileman et al, ¹⁶ the dentist cohort data was historical, and the dentist’s clinical experience was not specified.

Cognitive load

Cognitive load has been described as the mental load experienced by a clinician when faced with a challenging situation or a clinical problem. Lack of conceptual knowledge, inefficient reasoning strategies, and task difficulty increase cognitive load. ^23,27 Using pupillary response as a measure of cognitive load, Castner et al ¹⁹ reported that the cognitive load varied among novice and experienced clinicians depending on panoramic radiographs of different difficulty levels. They found that the cognitive load among students was consistently high regardless of the difficulty of diagnosis (case complexity). However, the cognitive load increased proportionately to the case difficulty among clinicians. The remaining studies in this review did not directly discuss the influence of cognitive load on the occurrence of errors of interpretation.

This systematic review recognised that the factors affecting the interpretation of dental radiographs identified in the included studies are interlinked and have a follow-on effect on each other, as demonstrated in Figure 2. The figure shows that the factors identified in this review either directly or indirectly affect cognitive load, and an overload on the cognitive functions leads to errors of interpretation.

Figure 2.

Concept map of factors affecting interpretation of dental radiographs

Discussion

In medical radiology, several factors influence the interpretation of radiographs, and failure of these processes leads to errors. ^6,8,28 Interpretation errors can lead to consequences such as delayed diagnosis of life-threatening diseases, unnecessary or harmful treatment and patient mortality. ^{4,6,7,11,12,29,30}

The primary goal of this systematic review was to identify from the literature, factors contributing to errors in interpreting dental radiographs. The factors identified from this review included clinical experience, clinical knowledge and technical ability, the complexity of dental radiographs, cognitive load, time pressure and geographic location and length of dental training and education programs. These factors have also been recognised in medical radiology as impacting the accuracy of radiographic diagnosis and interpretive errors. ^6,31–34 Research on interpretive errors in medical radiology has shown that interpretive errors are common. ^35–37

This systematic review found that clinical knowledge and technical skills in dental radiography were closely related to clinical experience. Understanding anatomy and pathologic processes, their consequences on oral health and their management are acquired during training, and the knowledge improves over time as a clinician gains more experience. ³⁸ However, this systematic review was inconclusive about how the duration and location of the dental education and training program impacted clinical knowledge and technical skills. Research has shown that experienced clinicians develop clinical gestalt, a heuristic decision-making approach. ^38–40 The ability to recognise patterns on radiographs and gain a holistic understanding of a radiographic image develops over time. Therefore, experienced clinicians demonstrate diagnostic acuity and higher levels of diagnostic accuracy. Experienced clinicians may also manage time pressure and case complexity differently than novice clinicians. Experienced clinicians are more likely to handle stress (cognitive overload), time pressure and case complexity better.

Several techniques or strategies have been studied in medicine, including templates, checklists, clinical decision support systems and machine learning algorithms to reduce the incidence of errors of interpretation. ^41–43 However, in dentistry, there is insufficient data regarding the use of cognitive aids to assist with clinical decision-making.

This systematic review had methodological limitations, including a small number of studies and limiting to papers published only in English. The exclusion of non -English studies may be a source of publication bias. In addition, publication bias in the included studies could not be assessed due to the heterogeneity of the papers in relation to their study design and significant differences in the effect sizes reported. The quality of evidence was rated as medium because the papers addressed only three of the five criteria described by MMAT. The quality assessment revealed that the study designs did not take appropriate measures to control confounding bias. The studies also had methodological issues, including small sample sizes and inaccurate diagnostic accuracy measurement, which impacted the statistical significance and generalizability of the results. Studies with small sample sizes reduce the power of the study, and the results cannot be extrapolated to a larger population. Diagnostic accuracy studies measure the accuracy of a diagnosis against a gold standard using parameters such as sensitivity and specificity. Studies in this review used different methods to establish gold standard diagnoses, including histopathological sections of teeth and expert panel diagnoses. In one study, gold-standard diagnoses were not used.

Conclusions

This review identified factors affecting the accuracy of radiographic diagnoses in dental radiology. However, it also revealed a notable lack of literature studying the causes and factors influencing errors in interpreting radiographs in a dental setting. All the factors described in this review directly or indirectly impact cognitive load. In turn, cognitive overload can lead to interpretive errors. Errors of interpretation can have a substantial impact on diagnostic and treatment decisions. These errors can potentially affect both the patient and the dental clinician adversely. Well-designed studies in clinical settings are needed to gain insights into how and why interpretive errors occur. Further research will also increase awareness about interpretive errors and pave the way for developing strategies to minimise them in dental practice. Strategies to close these gaps include identifying factors associated with interpretive errors and efforts to build interpretive error epidemiology that will account for the different underlying processes that affect them.