The results from our study show significant student interest in olfactory training, with almost 90% of students believing such exposure would make them more competent physicians. Yet, despite being able to connect “textbook” descriptions of odors to medical diagnoses, our study suggests that most medical students do not feel comfortable utilizing smell for diagnostic purposes, despite seeing a large potential for its use in their future practices. 78% of students felt either slightly competent or not competent at all in making accurate diagnoses when presented with a medically relevant scent. 62.9% of students ranked the sense of smell as either of moderate or high importance in the process of reaching a diagnosis. Therefore, we argue that there is a need for specific, targeted training in recognizing medically relevant scents. We propose that healthcare educational institutions consider the implementation of smell training into their existing curricula. It was interesting to note that for certain diseases, namely diabetic ketoacidosis, skin necrosis, and bacterial vaginosis, fourth-year medical students’ competencies just about doubled upon comparison to their first- and second-year counterparts. Though the reason could be the strong odor associated with these conditions, one might expect similar results for pseudomembranous colitis caused by Clostridium difficile, whose competencies remained stable among all years. It is likely that diabetic ketoacidosis, skin necrosis, and bacterial vaginosis are extremely commonplace in both the clinic and hospital setting, more so than pseudomembranous colitis, and students were able to make strong associations during their clinical rotations through repeated exposure. As expected, competencies were lowest for the musty/mousy odor of phenylketonuria as the disorder is very rare and unlikely to be encountered.
Smell is often overlooked as an important clinical skill (Hayden, 1980). Yet, smell has been used in clinical education settings in the past. A “sniffing bar” was created at one time to train clinicians in emergency department settings to correctly recognize and treat ingestion of lethal toxic substances before lab identification is performed (Goldfrank et al., 1982). More recently, there has been interest in creating a diagnostic electronic nose. There is exciting and ongoing research surrounding artificial intelligence and nanotechnology coming together to create sensors that can detect volatile organic compounds on patients’ breaths (Nakhleh et al., 2017).
There are limitations to this observational study. The first and foremost is the small size of the cohort; only 139 out of 720 contacted medical students responded to the survey. Among the participants, the overwhelming majority of responses was from medical students. Our study also lacks representation from students of other healthcare professions who are trained to diagnose diseases, such as nurse practitioner or physician assistant programs. Furthermore, the survey was conducted at only a single medical school in West Texas. Therefore, the results of our study may not represent the experiences and opinions of students attending other medical schools with different curricula in different parts of the country. In addition, the respondents' unequal experience, competence, and opinions in scent recognition may be influenced by the randomness of patient encounters during clinical MS3 and MS4 years. Finally, in retrospect, we do believe that the question, “Do you feel with your current healthcare educational training that you can accurately make diagnoses based on medically relevant scents?” was poorly worded in that it appears to imply there was an expectation to be able to diagnose certain odorous diseases on the sole basis of scent when that was never the case. Diagnosis is an extensive process that involves multiple modalities of analysis; we are only suggesting that scent be considered as one of many components.
We hope that the results from this study will encourage future research on the incorporation of olfactory training in medical education. We believe that such future studies should investigate the impact of targeted training of medically relevant smells upon students’ ability to learn and recognize medical diagnoses. We are of the opinion that utilizing scent as a diagnostic aid can be useful in providing future clinicians with additional competencies that can benefit patients. The vast majority of medical students acknowledge the utility of smell in the clinical setting, but the same students do not feel comfortable using their knowledge of pathognomonic scents to aid in diagnosis. We venture to argue that this gap in medical education can be easily addressed by formally including scent exposure in the clinical years. Further research can be targeted toward determining effective methods of providing scent exposure and in which manner it can be incorporated into medical curricula.