Family medicine residents’ skill levels in emergency chest X-ray interpretation

Family medicine is a unique specialty because the scope of practice is not limited to a particular age, gender, or organ system. As the first point of contact within the healthcare system, family medicine physicians may encounter a broad spectrum of medical conditions, with different complexities, ranging from minor complaints to life-threatening emergencies. The provision of emergency care for acute and urgent conditions is a pivotal role of family medicine practice. It is estimated that emergency cases constitute approximately 5% of all cases in primary healthcare centers [1]. The primary care centers may not be adequately equipped for proper management of emergency cases [2]. Physicians in primary healthcare centers depend largely on their skills in history taking and physical examination to establish a diagnosis considering limited access to other assessment tools [3].

The chest X-ray is a crucial diagnostic tool because it is widely available, relatively inexpensive, and has a low dose of radiation exposure. The chest X-ray remains the starting point among imaging modalities for the diagnosis and management of cardiac and pulmonary conditions. Several studies have demonstrated the fundamental role of chest X-ray in clinical decision-making [4‐6]. For instance, a prospective cohort study involving 78 general practitioners revealed that the proposed management was altered in 60% of patients who underwent a chest X-ray with a reduction in referrals, increased satisfaction of patients, and improved prescribing [4].

Most primary healthcare centers may not have access to radiologists for the interpretation of radiographs. Family medicine physicians, therefore, must rely on their own radiologic interpretive skills. Several studies have investigated the agreement rate between family medicine physicians and radiologists and found the discordance rates may be as high as 58.1% [7, 8]. While the majority of this discordance may not cause substantial changes in patient care [8, 9], misinterpretation of chest X-rays of some emergencies could result in serious outcomes [10].

According to the curriculum of the Saudi Board for Family Medicine Program, family medicine residents are expected to master the necessary knowledge and skills for requesting and interpreting chest X-ray images; however, little is known about the competency of family medicine residents in the interpretation of chest X-rays for emergency conditions. It should be pointed out that diagnostic radiology is not given adequate consideration in the undergraduate medical education and graduates may not feel sufficiently prepared for accurate radiological interpretation in medical practice [11].

Therefore, it would be of particular interest to assess the competency of family medicine residents in the interpretation of chest X-rays and to compare diagnostic accuracy and confidence with that of diagnostic radiology residents and physicians who have not undergone a postgraduate training program, including medical interns and general practitioners.

This study assessed the competency of family medicine residents in interpreting chest X-ray images of emergency conditions. There was a significant difference in the diagnostic accuracy between family medicine and diagnostic radiology residents. Furthermore, the competency of family medicine residents was comparable to physicians who had not undergone any postgraduate training, including medical interns and general practitioners.

It has been recognized that family medicine physicians often provide a lower quality of care at disease-specific levels with better outcomes for patients and lower costs than the other specialties. This phenomenon is called the “primary care paradox” and it has been well-described and investigated in the literature [13]. However, the clinical cases provided in the study were of common emergency conditions that all physicians should be able to recognize. It should be pointed out that the interpretation of chest X-ray by radiographers was found to be non-inferior to diagnostic radiologists [14].

Interestingly, the study showed that the lowest diagnostic accuracy of family medicine residents was in interpreting the case of left lower lobe pneumonia. It should be noted that pneumonia is the most common indication for obtaining a chest X-ray in general practice [4]. A chest X-ray is a valuable tool in the assessment of patients with suspected pneumonia because the clinical presentation alone is not predictive of pneumonia and a chest X-ray substantially reduce the number of misdiagnoses [15]. The low diagnostic accuracy for the interpretation of left lower lobe pneumonia in the present study could be attributable to the subtle abnormality on the frontal chest X-ray; however, the consolidation was clearly evident on the lateral view (Fig. 1). A similar study by Satia et al. [16] showed that lower lobe atelectasis is the least correctly interpreted chest X-ray. Indeed, lower lobe pathology is better evaluated by a lateral view chest X-ray [17].

Despite its importance, the lateral view is often ignored during routine chest X-ray evaluation. The lateral view is useful in interpreting findings that are not clearly shown on the frontal view because 15% of the lung is obscured by cardiovascular and mediastinal structures [17, 18]. Hence, the lateral view is of paramount importance to visualize the hidden areas of the lung in the frontal view. The low diagnostic accuracy of family medicine residents in recognizing pneumonia in the present study cannot be generalized to all cases of pneumonia considering the wide spectrum of radiological manifestations of pneumonia [19]. Similarly, lobar atelectasis has different radiological findings based on the anatomic location of the involved lobe; however, left upper lobe collapse has a unique manifestation as the lobe collapses anteriorly giving a characteristic veiling opacity (Fig. 2).

The pneumomediastinum case was interpreted correctly by less than one-half of the participants, and was the case interpreted correctly least often. This finding is consistent with a previous study that estimated 50% of pneumomediastinum cases are missed on frontal chest X-rays [20]. Several imaging findings of thoracic conditions have been linked to radiological signs and symbols. The knowledge of these associations is helpful, for both radiologists and non-radiologists, and could aid in the interpretation of chest X-ray images [21]. For example, a pneumomediastinum may manifest on frontal chest X-ray as a radiolucency between the heart and the superior surface of the diaphragm. Such a finding is referred to as a “continuous diaphragmatic sign” (Fig. 3) [22].

The COVID-19 pneumonia case (Fig. 4) was correctly answered by the highest number of participants. Considering that the study was conducted in the midst of the COVID-19 pandemic, this finding might not be surprising. Prior exposure to similar radiological images has been shown to be associated with an increased likelihood of accurate and correct interpretations [23]. This finding stressed the importance of having adequate exposure for family medicine residents to medical images to enhance their diagnostic abilities.

Notably, one-fifth of the participants did not interpret the normal chest X-ray correctly. Misdiagnoses of normal chest X-rays may subject patients to inappropriate management with potential adverse effects and complications. While the chest X-ray used in this study was entirely normal, it should be noted there can be significant differences in interpretation of parenchymal findings on chest X-rays regardless of the experience of the physician [24]. Furthermore, a previous study showed that 30% of patients who were assumed to have community-acquired pneumonia based on clinical and chest X-ray findings were found to have no evidence of pneumonia on computed tomography scans [25].

The current study showed a significant positive correlation between the diagnostic confidence and accuracy in chest X-ray interpretation. We assume that when a physician is confident in the interpretation of a chest X-ray, the interpretation is more likely to be accurate. Diagnostic radiology residents were shown to have the highest diagnostic confidence and accuracy; however, the diagnostic confidence was evaluated using a subjective 5-point scale, thus making the comparison among physicians of limited value. The current study did not demonstrate an association between diagnostic confidence or diagnostic accuracy and years in practice. Surprisingly, Satia et al. [16] reported that the diagnostic accuracy of specialist registrars was higher than non-respiratory medicine consultants. It was shown that the self-assessment of physicians is of limited accuracy and external assessment, as such in this study, is needed to correctly assess the competency and skills of physicians [26].

The current study findings demonstrated the need for interventions targeting family medicine residents to improve chest X-ray interpretation skills. Family medicine residents who reported having adequate training in chest X-ray interpretation had higher diagnostic accuracy and confidence. There is a pressing need to implement a radiology training course in the family medicine residency program. Today, radiology courses could be delivered in a wide variety of formats, including an online platform [27]. Several participants welcomed the idea of this survey and requested further similar online quizzes. Furthermore, introducing a rotation in the diagnostic radiology department during the family medicine residency program should be considered.

A particular feedback comment caught our attention: “the survey has a good selection of cases; however, the provision of clinical vignettes made the interpretation easier.” This comment is valid because previous studies showed that diagnostic accuracy and confidence increase when clinical information is provided [28, 29]. While the clinical vignettes we provided were not indicative of a specific diagnosis, the interpretation of a chest X-ray should not be made in isolation from the clinical information. For example, a patient with a pleural effusion will be diagnosed with a hemothorax when a clinical history of trauma is provided. Similarly, a chest X-ray with bilateral pulmonary infiltrates might be diagnosed as pulmonary edema. If a clinical history of fever and productive cough is provided, however, it is more likely to be diagnosed as multifocal pneumonia. Hence, providing clinical information is of paramount importance in chest X-ray interpretation. The clinical information, however, may be a distraction and lead to false-positive interpretations. Therefore, it is suggested to review the radiological images before reading the clinical data [30].

This is the first study which investigated the competency of family medicine residents in the Saudi Board of Family Medicine Program in interpreting chest X-rays for emergency conditions in comparison with diagnostic radiology residents. Several participants acknowledged the invitation for the survey and appreciated the option to receive the correct answers after the study was completed. We believe that this motivator along with the reminder e-mails helped to achieve a reasonable response rate because the use of incentive strategies has been shown to improve the response rates in online surveys [12].

The present study had some limitations. The survey-based methodology could have introduced some bias. Physicians who had a greater interest in chest X-ray interpretation might have been more likely to respond. Furthermore, the study used a limited number of chest X-rays to assess the interpretation skills. For example, skeletal radiographs and pediatric chest X-rays were not included. The radiological appearances of the pathologies on chest X-rays in the present study were classic, which might not be the case in real-life medicine. Lastly, the majority of participants completed the survey using their mobile phones, which could be less optimal than reading the images on computer monitors, as reported by one of the participants.

The competency of family medicine residents in the interpretation of chest X-rays for emergency conditions is far from optimal. A radiology training course focusing on the acute and life-threatening conditions seems a good starting point if the family medicine residents are expected to make clinical decisions based on chest X-ray interpretations. Alternatively, the use of tele-radiology in primary care could be considered. Further research is needed to re-assess the interpretation skills of family medicine residents after implementation of improvement measures.