Improving pediatric emergency department physicians’ adherence to clinical practice guidelines on the diagnosis and management of group A beta-hemolytic streptococcal pharyngitis—a cross-sectional study

Pharyngitis, colloquially known as “sore throat,” is a frequent complaint in the pediatric healthcare setting instigating hospital visits, with 1% to 2% constituting emergency department and outpatient hospital visits in the USA [1, 2]. While much of acute pharyngitis episodes are viral in origin, group A beta-hemolytic streptococcus (GABHS) accounts for about 37% of bacterial causes in children older than 5 years of age [3]. The isolation of GABHS is crucial in clinical practice to prevent suppurative and some non-suppurative post-infectious complications, and these include peri-tonsillar abscess, otitis media, and rheumatic fever [4‐6].

Treatment with antibiotics for other common forms of acute bacterial pharyngitis other than GABHS is not warranted [7]. This offers an emphasis on the importance of the accurate diagnosis of GABHS, to avoid the gratuitous use of antibiotics for other causes of acute pharyngitis. The inappropriate prescription of antibiotics for respiratory infections in the emergency department raises the likelihood of bacterial drug resistance [8]. In Riyadh, Saudi Arabia, antibiotics represent the third most commonly prescribed drug group [9]. In one study describing antibiotics overuse in children with upper respiratory tract infections in Saudi Arabia, Alumran and his group had found that antibiotics were prescribed for 44 to 88% of patients presenting with upper respiratory tract infections (URTIs) [10]. Furthermore, the excess use of antibiotics was found to be more common in children with URTIs caused by a viral pathology [11]. Evidence concerning the usage of antibiotics among children in Saudi Arabia with URTIs is limited, placing an importance on the need to study its prevalence [10].

It is important to note that several strategies were used to decrease the use of antibiotics in children with respiratory tract infections. These include patient education and delayed prescription of antibiotics, which shows a reduction of up to half in the latter [12]. It is also worthy to mention that antibiotics are often prescribed in the presence of factors related to clinical presentation, and these include an ill-looking child, more particularly of younger age, fever, sore throat, and earache, along with other factors unrelated to signs and symptoms, and these include the lack of national or local guidelines, parental pressure, inaccurate diagnosis, and the fear of unfavorable clinical outcomes [13, 14]. Furthermore, an emphasis of guidelines has been on the limitation of the use of antibiotics for the treatment of acute pharyngitis in both adults and children [1]. One of the commonly used systems is the Centor score (Modified/McIsaac), which is designed to predict the likelihood of streptococcal pharyngitis in children, from ages of 3 to 17 years, presenting with a sore throat, mainly to guide the testing and treatment. The score is composed of several factors: absence of cough, swollen tender anterior cervical lymph nodes, fever of 38 °C or more, and tonsillar exudate, each account for 1 point. For a child with a score of 1 or less, neither a throat swab nor antibiotic therapy was recommended, while testing is recommended for score 2 or above. It has been shown with a modified Centor score of 1, the probability of GABHS pharyngitis is 14% and 55% of which scored 4 and above [15]^.

In this study, we aim to assess adherence and enumerate the reasons behind the suboptimal adherence to guidelines (Centor/McIsaac score) of pediatric emergency department physicians in the diagnosis and management of GABHS pharyngitis to lay the groundwork for future actions and to employ educational programs and implement local guidelines for the prevention of the development of multi-drug resistant microorganisms.

Table 1 shows the sociodemographic data of our sample. Sixty-six percent of the physicians are between the ages of 25–35 years, with half of our sample being males. Saudis constitute more than half of the responders. Eighty-two percent of the studied population are non-consultants. Fifty-two percent of the responders attend 16 shifts a month, 8 h per shift. Most of them seeing an average of seven cases of acute pharyngitis a week.

Table 2 presents the knowledge scores of our physicians. A total of 9.1% (n = 22) of the ED physicians scored 0, while the majority, of both consultant and non-consultant groups earned a score of 1. The remainder of responders, making up 44.4% (n = 108), have a score of TWO.

Adherence to guidelines was defined as when diagnostic tests (throat culture or rapid antigen detection test) were always performed/requested prior to prescribing antibiotics when acute bacterial pharyngitis was suspected. In accordance with this, Fig. 1 demonstrates that only 27.3% (n = 67) are adherent to guidelines, whereas the majority, 72.7% (n = 175), are non-adherent.

Adherent participants were asked to rationalize always performing diagnostic tests. Guideline recommendations presented 70.6% (n = 47) of the answers, while hospital policy only presented 17.7% (n = 12). Insufficiency of signs and symptoms to reach the diagnosis represented the remaining 11.8% (n = 8). However, this is only for our adherent participants.

The majority of both consultants and non-consultants were non-adherent, 33 and 143 physicians, respectively. Sharing management decisions with senior physicians showed to be a statistically significant factor (p = 0.007), along with knowledge scores (p = 0.001) as elucidated by Table 3.

As shown in Table 4, the claim that “signs and symptoms are sufficient to make the diagnosis of acute bacterial pharyngitis” is most elected as the rationale behind not following guidelines, in both consultant and non-consultant groups, followed by difficulties in ensuring proper follow-up of patients and parental pressure to prescribe antibiotics, respectively. Unavailability of diagnostic tests is the least of all factors behind suboptimal adherence. Awareness about the clinical scoring systems that can predict GABHS is not statistically significant, in contrast to the presence of local hospital guidelines for the management of acute upper respiratory tract infections in the emergency department (p = 0.0004).

It is unquestionably essential to be more mindful when it comes to the prescription of antibiotics for URTIs, including pharyngitis, and this is to slow the rising rates of resistance [6]. With that in mind, physicians should be more adherent to guidelines, and not over/underprescribe antibiotics to children [16]. Utilizing the Centor score (Modified/McIsaac) when diagnosing pharyngitis improves the sensitivity and specificity when approaching these cases, and this is because viral etiologies are difficult to differentiate from bacterial etiologies clinically [17, 18].

In our study, we have found disappointing adherence rates of 27.3% of physicians. Even consultants, who should have superior knowledge and experience, are not adherent (76%). Alas, these low adherence rates have also been reported by other articles. Crocker et al. found an adherence rate of only 24% [19], and Linder et al. found that in more than half of the cases, physicians did not adhere to any guidelines in managing pharyngitis [20]. Nevertheless, our assessment of adherence did not detail any other clinical predictors that may direct physicians towards prescribing antibiotics.

Knowledge of guidelines is an important factor in adherence. Almazrou states: “Knowledge such as lack of familiarity and awareness, volume overload, time needed to stay informed and guideline accessibility are important in modifying physician practice patterns” [21]. In this study, we evaluated consultants and non-consultants’ knowledge. Surprisingly, even consultants who should have extensive knowledge of guidelines did not have a complete understanding of it, as 55.8% of consultants did not achieve a full score in the knowledge assessment, which could lead to an inappropriate practice of requesting throat cultures or prescribing antibiotics, which was identified by another study [22] to be a major contributing factor in non-adherence. However, a different study provides an opposing statement that argues that even when physicians were encouraged after educational interventions to use Centor score to differentiate between bacterial and non-bacterial pharyngitis, there was no decrease in the prescription of antibiotics [23]. It is correct to assume that knowledge influences guideline adherence, and it is also correct to assume that behavior also influences adherence, as reported by Fischer et al. [24], which we have failed to assess. Furthermore, this study was conducted in teaching hospitals; yet, most of our non-adherent group do not consult senior physicians regarding cases of pharyngitis, which could be a major problem, especially because junior physicians would treat even viral causes with antibiotics as discovered by Fakih and his group [25]. To further emphasize our point, physicians do believe resistance to antibiotics to be a definite problem but still fail to estimate the prevalence of resistance in their institutions [25], and this demonstrates how imperative it is to educate and further train physicians on the proper prescription of antibiotics.

To our knowledge, there are no articles that discussed the reasons behind suboptimal adherence in our setting. We enumerated the suspected reasons and found that “sufficiency of signs and symptoms” in the diagnosis of bacterial pharyngitis was on top of the list, which contradicts the findings of Barbosa et al. [26] and the recommendations of the Infectious Diseases Society of America (IDSA), which states that signs and symptoms are not enough [17, 18, 27]. Additionally, difficulty in ensuring proper follow-up is a major issue that physicians face when making decisions in managing cases because of the higher load on public hospitals and patients that live far away or in rural areas with little accessibility to hospitals due to the lack of public transportation systems.

Factors pertaining to patients such as parental pressure on physicians to prescribe antibiotics are important elements that should not be ignored when assessing adherence, along with other factors (e.g., beliefs that antibiotics treat viral infections, not wanting another visit) that influence physicians’ decision-making as mentioned by Alweis et al [22]. Nevertheless, according to Hedin et al., physicians who adhered to guidelines had fewer problems convincing patients not to take antibiotics, while their non-adherent counterpart had issues regarding this matter [23].

Thankfully, organizational factors of implementing guidelines were not an issue in our settings. Most of the hospitals have the diagnostic tests listed by the Centor score (Modified/McIsaac), but we have not studied this intensely. Other factors including the adoption of the guidelines by departments and evaluation of the quality were not assessed.

A review article declares that the dissemination of educational materials is a vital element of effective guideline implementation. Thus, providing educational materials (e.g., written materials, presentations, interactive conferences) is indispensable to increase awareness, understanding, and agreement with a guideline and its recommendations. Moreover, constant efforts in the training and education of health professionals are required, which may be achieved by educational assemblies, audit and feedback, workshops, and small-group collaborative training sessions [24]. Luckily, the awareness rates in both consultant and non-consultant groups of this study are high; yet, a further increase is still favorable.

To sum up, “knowledge-attitude-behavior framework” [24] dictates that physicians require awareness of guidelines and knowledge of their content. Subsequently, knowledge impacts attitude, and attitude will most definitely impact practice.

A significant number of our non-adherent group say that their institutions do not have local guidelines for the management of pharyngitis, which is not advantageous, as recommended by Fischer et al. that when guidelines are tailored to specific settings, a change in behavior could be achievable [24].

Our study implies that more efforts must be taken to change physicians’ level of knowledge by different strategies, including classes, workshops, and mandating guidelines implementation. Junior clinicians and medical students should be more involved in programs of evidence-based medicine and conducting researches that would improve the medical care system and reduce the burden of costs. Family pressure is an issue that faces those who adhere to the guidelines but could be dealt with by changing the public’s behavior towards antibiotics and influencing their awareness.

Our study had many limitations. Our data was only gathered from physician’s in the form of electronic questionnaire as it is difficult to obtain IRB from ten different hospitals; hence, access to medical files is not obtained, not to mention that some of our hospitals are still using paper files which is another obstacle. We believe that a multi-center chart review would be more helpful and authentic. Selection bias was not eradicated, given the fact that we used convenience sampling. We have not carried any interventions to investigate whether classes and discussions concerning resistance to antibiotics and utilization of guidelines increase their level of knowledge, and with this, changes to the physicians’ behaviors when it comes to guidelines. We also did not take in consideration the years of experience and undergraduate training as factors in our study.

Despite these limitations, we think our results provided a better understanding regarding the issue of not following guidelines, and the regrettable gaps in knowledge that our residents have and the importance of locally developed guidelines and how they aid in physicians’ adherence and attitude.

Unfortunately, we found low rates of adherence (27.3%) for guidelines in pediatric emergency departments of teaching hospitals in Riyadh, Saudi Arabia. Both consultants and non-consultants were mostly not knowledgeable enough to achieve a full score when assessed about Centor score (Modified/McIsaac). Awareness, knowledge, and behavior of physicians are the main factors behind low adherence, emphasizing the importance of local guideline implementation and the need of employing strategies to increase physicians’ awareness and ultimately changing their behavior. Future studies should aim to implement local guidelines and assess local physicians’ adherence to evaluate whether it will improve adherence or not.