The influence of gender concordance between general practitioner and patient on antibiotic prescribing for sore throat symptoms: a retrospective study

Bacterial resistance is an important topic in today’s health policy [1]. Although growing microbacterial resistance is partly a natural process, it can be accelerated with the inappropriate prescription of antibiotics by health care professionals [2]. With this in mind, factors which influence the prescription of antibiotics, other than patient’s clinical presentation, became an object of study. Indeed, research suggests the importance of such non-medical factors. General practitioners’ (GPs) attitudes such as fear and complacency affect prescribing behavior [3, 4]. Akkerman showed that physicians with more years of practice were more likely to prescribe antibiotics, especially if they felt they had little time per patient [5]. Moreover, GP’s perception of patients’ expectations concerning medication prescription influences the actual prescribing behavior [6‐8]. Overall, these studies imply that doctor-patient interaction has an important influence on antibiotic prescribing behavior.

When focusing on the doctor-patient relationship, the gender of both the doctor and patient are known to influence their interaction [9, 10]. Among others, the GP’s gender may determine the communication style, the contents of the consultation, which drugs are indicated and whether they are prescribed [11‐13]. In turn, the gender of the patient also affects the interaction between doctor and patient. For example, female patients are more likely to express their emotions during the consultation [14] and had fewer discussions about addictive behavior or heart disease risk [9, 15]. This implies that GPs might make medical decisions which are affected by gender-related considerations and gender stereotypes.

Since both patient’s and doctor’s gender play a part in the medical process, it is not unlikely that the combination of a patient and doctor of the same gender (compared to dyads of opposite gender) may have additional effects. In interactions between patient and doctor, four gender dyads can be distinguished (male-male, female-female, male-female, female-male) of which two are concordant and two are discordant. It has been found that female-female consultations contain more affective talk and less analytical talk. The opposite occurs in male-male consultations [16]. Moreover, it appears that female concordance leads to communication that is most patient centered [17, 18], which may enhance health outcomes by elevated patients’ trust, improved communication and patient satisfaction. Indeed, female gender concordance is associated with more effective treatment of cardiovascular risks [19] and male gender concordance is positively associated with measures on diet, nutrition and exercise counseling [20]. It is however not known whether gender concordance influences prescribing behavior of antibiotics. Possibly, concordance is an additional non-medical factor that affects (inappropriate) antibiotics prescription. Creating awareness of such non-medial factors could result in GPs being less biased, more objective and consistent in their treatments.

Since the 1980’s there is an increase in the extent to which primary care practice is influenced by clinical guidelines [21]. For example, some diagnoses are followed by guidelines that leave the physician room to follow a treatment of choice (‘non-protocolled’ guidelines). Other diagnoses have guidelines in which a ‘protocolled’ treatment (e.g. antibiotic prescription) is strongly recommended. The power of a clinical guideline is to make sure the patient gets the most effective, evidence-based treatment. A guideline also reduces the variation in treatment for the same diagnoses between GPs.

In this study, we will explore whether the prescription of antibiotics depends on the patients or physicians gender and/or on their gender interaction (concordance). To do this, we will study prescribing behavior concerning patients presenting sore throat symptoms in primary care. This patient group was chosen since the inappropriate prescription of antibiotics is very common in sore throat symptoms [22] and because the majority of antibiotic prescriptions in The Netherlands is issued in primary care [23]. In addition, we will investigate whether gender concordance is less influential in diagnoses corresponding to a strong antibacterial protocol than in diagnoses that leave the GP more room for interpretation and choice.

The specific objectives of this study were to explore:

Of all GP practice consultations, 52.1% were handled by male GPs. In 39.2% of the consultations the patient was male. Of all cases 27.6% of the patients were prescribed antibiotics. The concordant dyads are represented with 52.8% of the consultations, while 47.2% of the consultations contained discordant couples. For all consultations, 8.5% got assigned an ICPC code labeled as protocolled and 91.5% as non-protocolled.

Table 2 shows that patient age, presence of comorbidity and the amount of consultations without a real diagnosis were similar in consultations with male and female GPs. Male GPs prescribed antibiotics in 29.5% of their consultations. Female GPs did so in 25.7% of their consultations. For female GPs, 64.1% of their consultations were concordant (i.e. with a female patients), while male GPs had same-sex consultations in 42.3% of their consultations. The concordant couples had an antibiotic prescription rate of 26.8% of the consultations, against 28.6% in discordant couples.

Figure 2 shows the percentage of antibiotics prescribed per gender dyad. In accordance with Table 2, it illustrates that female GPs prescribe less antibiotics than male GPs. Female GPs especially prescribe less antibiotics for female patients (24.8% in concordant couples versus 27.3% in discordant couples). In consultations with male GPs, on the other hand, there is no such difference (concordant couples 29.6% versus 29.4% in discordant couples).

Table 3 shows the outcome of the multilevel regression analysis. In the total population, concordance (OR 0.92, OR 95% CI: 0.83–1.01), female GP (OR 0.88, 95% Cl: 0.67–1.09) and female patient (OR 0.93, 95% Cl: 0.84–1.02) show lower rates of antibiotic prescription, however not with statistical significance. Similar results were found in both types of prescriptions (protocolled and non-protocolled). Patient age was significantly associated with antibiotics prescription (OR 1.00, 95% CI: 1.00–1.00). This was equally true for protocolled and non-protocolled consultations. Comorbidity was a significant predictor in non-protocolled consultations (OR 1.21, 95% CI: 1.01–1.32), but not in the protocolled consultations (OR 0.92, 95% CI: 0.57–1.26).

Table 4 shows that for female GPs concordance is associated with antibiotic prescription after controlling for comorbidity and patient age. In female concordant consultations less antibiotics were prescribed (OR 0.85, 95% CL: 0.72–0.99). The lower section of Table 4 shows that for male GPs concordance is not associated with antibiotic prescription after controlling for comorbidity and patient age (OR 0.99, 95% CL: 0.86–1.11).

In this study, we found that gender concordance among females is associated with less prescription of antibiotics for sore throat complaints. In other words, female GPs appear to issue fewer antibiotics prescriptions to female patients than to male patients. Among male GPs, concordance did not appear to have a similar effect. Possibly, our findings can be explained by a different communication style that is handled by female GPs, especially when attending with female patients. Future studies should aim to describe and understand female concordance in more detail, since the findings may suggest that training doctors to adjust their communication style could prevent the prescription of unnecessary antibiotics.