Introduction

Diseases affecting the ear, nose and throat (ENT) constitute 20–50% of disorders treated at health facilities [1]. In most resource-limited settings, they are associated with poor outcomes [2]. Despite their high prevalence, ENT diseases have been disregarded in global health [3]. In Zambia and other developing countries, ENT health care is poorly funded and has inadequate infrastructure, equipment, medication, human resources and training facilities [4, 5]. The few available ENT specialists are distributed to limited overwhelmed urban units. Thus, most ENT patients are treated by clinicians not trained in ENT disease management [4]. Although these clinicians must be competent in the basic management of ENT conditions, many countries have deemed them inadequately trained to treat ENT diseases [6]. As a result, the true incidence of diagnostic error (delayed, missed or wrong diagnosis) [7] and inappropriate or late ENT patient referrals for treatment among non-ENT clinicians is likely high.

Diagnostic error contributes up to 70% of all medical errors [8]. In well-resourced countries, the incidence of diagnostic error varies between 0.7 to 15%, of which half is harmful [9, 10]. However, in low-resource settings, data relating to harmful diagnostic error remains sparse due to limited access to diagnostic resources, shortage of qualified medical professionals and poverty of electronic record-keeping systems [11]. While inappropriate referral of patients can undermine the quality of patient care [12], misdiagnosis increases the risk of death, elevates the financial strain on the already stretched health systems and may result in significant litigation [13]. In Zambia, there has not been an investigation into the prevalence and impact of diagnostic errors in health facilities.

In this study, we determined the diagnostic accuracy and estimated the appropriateness of referrals among patients treated at Zambia’s highest ENT treatment facility. In addition, we determined the level of agreement of ENT diagnoses between the ENT specialist and referring clinicians.

Methods

Results and analysis

Diagnosis accuracy and mismatch

Overall, 67.4% (n = 916) of all referral diagnoses did not match the ENT specialists’. Most (36.5%, n = 333) of the misdiagnosed patients had sinonasal diseases and the least (26.2%, n = 239) had ear diseases. Further, the majority (25.1%, n = 230) of misdiagnosed patients were aged 0–5 years, while the minority (4.4%, n = 40) were aged 13–18 years. Most (74.1%, n = 80) misdiagnoses were from level 1 hospitals (S2 Table). Despite non-ENT clinicians correctly diagnosing 72.0% (n = 18) of nasal polyps and 51.5% (n = 35) of acute tonsillitis in patients referred for specialist care, only 3.3% (n = 4) of them correctly diagnosed allergic rhinitis. Table 1 outlines the common diagnoses made by the ENT specialist at UTH ENT clinic, and proportions of their corresponding referring (non-ENT) clinician diagnoses.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Disease diagnosis comparison between ENT specialist and non-ENT clinicians.

https://doi.org/10.1371/journal.pone.0281686.t001

In the multivariate logistic regression model, patients aged 51–87 years were 1.77 (aOR: 1.03–3.04) fold more likely to be misdiagnosed by the referring clinician than those aged 0–5 years. Patients with ear diseases (aOR: 1.63; 95% CI: 1.14–2.33), those with sinonasal diseases (aOR: 1.80; 95% CI: 1.14–2.45) and those with no ENT pathology (aOR: 3.08; 95% CI: 2.58–4.48) were more likely to have a wrong pre-ENT referral diagnosis compared to those with head and neck diseases (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Univariate and multivariate logistic regression of the pre-ENT referral misdiagnosis among the different patient groups.

https://doi.org/10.1371/journal.pone.0281686.t002

Appropriateness of referrals

Of the total 1372 non-ENT specialist clinician referrals, 56.6% (n = 777) were inappropriate. Of these, 67.3% (n = 523) could have been treated by non-ENT clinicians at lower-level hospitals, 23.9% (n = 186) required trial of medical treatment before referral to UTH and 7.2% (n = 56) had no indication for ENT treatment. Patients aged 0–5 years were the most (27.4%, n = 213), while those aged 13–18 years were the least (4.6%, n = 36) inappropriately referred. The majority (39.1%, n = 302) of inappropriate referrals had head and neck diseases, while the minority (0.4%, n = 3) had an ENT-unrelated medical problem (S3 Table). Univariate and multivariate logistic regression showed that referrals from within the UTH complex of hospitals were 1.35 (95% CI: 1.03–1.75) and 1.34 (95% CI: 1.03–1.75) more likely to be inappropriate than those from other facilities (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Univariate and multivariate logistic regression of inappropriate referrals among the different patient groups.

https://doi.org/10.1371/journal.pone.0281686.t003

Discussion

In this study, we determined that the diagnostic error and inappropriate patient referrals for specialist ENT care in our resource-limited setting were high. Our findings provide evidence of and highlight the need to improve the diagnostic accuracy among medical doctors to improve the quality of ENT patient care. Although ENT is a surgical speciality, the skills it encompasses are relevant to all medical and surgical disciplines [19]. Therefore, all doctors are required to be competent to diagnose and treat basic ENT diseases and refer patients appropriately [19].

Worldwide, most health-care visits by patients are due to ENT diseases, most of which relate to the ear [20]. In contrast, this study determined that the majority (34.8%) of patients attending the UTH ENT clinic had diseases affecting the head and neck and the minority (25.7%) had those affecting the ear. This finding was inconsistent with evidence from the developing world, including India [21] and Nigeria [22], where ear diseases form the majority of ENT diseases treated in health facilities. While bigger multi-centre studies may be required to verify this finding, a possible explanation could be that many people with ear diseases (e.g., hearing loss, chronic otitis media) in Zambia do not readily access ENT services due to poor socioeconomic conditions, unavailable ENT services in health care facilities and poor health-seeking behaviour [4, 5, 23]. Considering the high (77.9%) poverty levels and large (56.4%) population of rural Zambia [15], and that ENT services are only found in few urban hospital units [4], many ear disease-affected patients in rural areas may not have insight into the potential adverse outcomes of these diseases to warrant long-distance travel to access an ENT service [24, 25]. Therefore, more effective universal health education, improved socioeconomic conditions and disadvantaged communities’ improved access to ENT services may result in a different epidemiological outlook of ENT diseases from what has been observed in the current study.

Our results suggest a high prevalence of diagnostic error among non-ENT trained health workers in low-resource settings. While data comparing diagnostic error between medical and surgical subspecialties is scarce, available evidence suggests a lower prevalence of diagnostic error among physicians and paediatricians. For instance, a survey conducted on 583 physicians across the United States of America found diagnostic error of 4.5% [26]. Similarly, diagnostic error of 3.6% among primary care physicians was observed in Malaysia [27]. In another study, more than half of paediatricians reported diagnostic error at least once or twice a month, and they made harmful errors at least once or twice a year [28]. Further, a study in the Netherlands reported diagnostic accuracy of 0.42 among medical residents, scored as either 0 (incorrect), 0.5 (partially correct) or 1 (correct) [29]. With 96.7% of allergic rhinitis, 95.7% of laryngopharyngeal reflux and 87.0% of ear wax impaction missed by non-ENT clinicians in our study, most of the patients with ENT conditions are likely inappropriately treated with the wrong medication and surgery. Further, some patients may not be referred early for management of life-threatening conditions like cancer. This may worsen the disease prognosis and increase the ultimate cost of patient care due to prolonged hospital stay and more advanced, costly treatment. In addition, delayed diagnosis of benign ENT conditions may result in life-threatening complications (airway obstruction in laryngeal papilloma, intracranial sepsis in acute otitis media and rhinosinusitis).

The poor agreement between the referring clinicians and ENT specialist (κ = 0.0001) and the high rate (56.6%) of inappropriate referrals for ENT specialist care may indicate inadequate training and poor knowledge of basic ENT disease management among most clinicians. In Zambia and the rest of the world, ENT is given little weight in medical school curricula, with most ENT undergraduate curricula offering a 1–2 week predominantly observational rotation [30]. As such, most clinicians lack the necessary skills to confidently manage ENT diseases upon completion of their undergraduate medical training [31]. Even if some literature emphasizes the superiority of adequate undergraduate ENT medical training over the duration spent as a general practitioner in the acquisition of competence to treat ENT conditions [32], the majority of clinicians often do require additional training to improve their competence following their undergraduate medical education [31]. Therefore, improving ENT medical education and clinical exposure at undergraduate, graduate and in-service training in Zambia would give clinicians better knowledge and skill to handle ENT conditions [32, 33]. A study assessing the knowledge, attitudes, and current practices of health workers with regard to the basic management of ENT diseases in Zambia is ongoing. We hope that the study will provide more insight into why some uncomplicated and correctly diagnosed conditions (i.e., acute tonsillitis, allergic rhinitis) are sent for ENT specialist treatment.

Literature has conflicting evidence on the rate of disease misdiagnosis in different age groups, with some reporting higher rates in older age groups [34] and some in younger ones [35]. In our study, patients aged 51–87 years were 1.8 times more likely to be misdiagnosed by the non-ENT specialist than those aged between 0–5 years. While we cannot provide a reason for this finding, it is worth exploring further with a more robust study with more participants. However, the widespread shortage of ENT diagnostic instruments across Zambia may explain the higher likelihood of misdiagnosis of patients with ear and those with sinonasal diseases. As nasal specula, otoscopes and endoscopes are seldom available in health facilities [5], clinicians seldomly perform complete ear and nasal examinations for diagnostic purposes. Except that of the larynx, much of the head and neck pathology requires no instrumentation to complete. In addition to the lack of sufficient skills and equipment at lower-level hospitals, patient referral networks in Zambia and many other low-income nations are undermined by poor coordination between lower and upper-level hospitals [36]. This may, in part, explain the high rate of inappropriate referrals noted in this study.

Even though debatable, available evidence shows that the knowledge and experience gained from regular practice of the clinician matters in reducing diagnostic errors, with more experienced clinicians likely making fewer diagnostic errors [37]. In our study, however, this trend was not apparent as patients referred from level 3 hospitals were not the least misdiagnosed (66.8%) [even though those referred from level 1 hospitals were the most misdiagnosed (74.1%)]. This finding supports the evidence that doctors’ medical knowledge stagnates after completing undergraduate training, but we suggest gathering better evidence with a more robust review of the literature. Our study also suggests that the referral practices in high-level hospitals (level 3) ought to be audited to ascertain whether junior doctors do some referrals to ENT without more experienced senior input, a practice that may elevate the rate of diagnostic error and inappropriate referral. In Zambia, level 1 hospitals are manned by Senior Resident Medical Officers (doctors who have completed at least 1.5 years of consultant-supervised practice following completion of undergraduate medical training but not yet commenced specialist training), while level 3 hospitals are training sites for medical students, Junior Resident Medical Officers (newly qualified doctors undergoing consult supervision and not fully registered as medical practitioners with the Health Professions Council of Zambia), as well as specialist trainees.

To make quality ENT services more available to the Zambian public, non-ENT specialist cadres can be trained to diagnose and manage basic ENT diseases, a strategy that has been successful in other countries within the region, such as Malawi [38]. Some literature from the developing world suggests that health workers can be trained to reliably diagnose and treat ENT diseases [39], with one study in Malawi [40] reporting good (κ = 0.7) and moderate (κ = 0.5) agreements in ear pathology diagnosis between the ENT specialist and the ENT clinical officer (mid-level health care providers with less training and more restricted scope of practice than medical doctors) and nurse respectively. The investigators noted the potential for non-specialist health workers to be involved in the patient assessment of ENT disease. Accordingly, training these non-specialist health workers in the use of basic ENT equipment (otoscopes, tuning forks, nasal specula), performance of basic procedures (ear syringing, indirect laryngoscopy) and non-expert interpretation head and neck x-rays would improve the quality of patient care in many places lacking an ENT specialist. In addition to training non-ENT specialists in ENT disease management, establishing international partnerships to train specialists locally would be successful in steadily increasing the pool of ENT specialists in Zambia. Such initiatives have been effective in advancing ENT in Africa, including The CBM International assistance in the establishment of ENT units in Central Africa, The University of Cape Town Karl Storz Head, Neck and Paediatric ENT Fellowships training African head, neck and paediatric ENT Fellows, and Operation Ear Drop Kenya equipping the temporal bone laboratory in Nairobi. Further, the introduction of ENT telemedicine to Zambia may improve patients’ access to quality health care. While telemedicine has been successfully implemented in other medical subspecialties in Zambia [41], it is yet to be used in ENT health care. However, to ensure success of telemedicine, the government must develop its implementation strategies that are tandem with Zambia’s needs [42].

The World Health Organisation (WHO) recognizes the importance of errors in diagnosis in its prioritization of safety errors in primary health care [43]. With up to 50% of GP consultations due to ENT conditions, the accuracy of diagnosis of ENT diseases remains a vital component of quality health care [43]. As serious harm rates of misdiagnosis per incident disease can be as high as 36% [44], all efforts to increase the ENT workforce in Zambia must be accompanied by effective mechanisms to guarantee knowledgeable and competent clinicians for ENT health care. While Cannon et al. (2004) estimated that one otolaryngologist for every 40,000 population is sufficient to provide comprehensive patient care [45], Zambia’s 0.007/40,000 ENT specialist/population ratio is well below the proposed minimum requirement [4]. Considering that diagnostic errors are the largest proportion of medical errors (up to 70%) [8], lowering ENT clinicians’ quality would be additional harm to the Zambian health system. While doctors often work towards correcting health-care system factors that lead to medical error, they are reluctant to acknowledge their diagnostic errors because mistakes may be perceived as professional lapses or personal failings [46]. This practice may be one of the largest compromises of good health-care delivery.

Conclusions

In low-resource settings, more effective, accelerated ENT training of clinicians may improve the diagnostic accuracy and reduce the high non-specialist clinician diagnostic error and inappropriate patient referrals. Findings of this study may be used to guide policy aimed to improve ENT health care in Zambia.

Supporting information

Acknowledgments

We would like to thank the management of the University Teaching Hospitals (UTH) for their permission to conduct the research. Our gratitude also goes to Kasoka Lukama and Chabota Mungambata for transferring the patient information hard copies to the excel spreadsheet. Further, we would like to thank all members of staff at the UTH ENT clinic for helping to retrieve the patient records for the study. In addition, we would like to thank the EDCTP-TESA II/UTH Project in Zambia for sponsoring Lufunda Lukama to their 2019 Data Analysis and Manuscript Writing Bootcamp in Livingstone, Zambia.