Reducing Expectations for Antibiotics in Patients With Upper Respiratory Tract Infections

A Primary Care Randomized Controlled Trial

Anna Ishani Perera, MBChB; Mark Greenslade Thomas, MD; Keith James Petrie, PhD; Janet Frater, MBChB; Daniel Dang, PhD; Kiralee Ruth Schache, MHP; Amelia Frances Akroyd, MHP; Stephen Robert Ritchie, PhD

Disclosures

Ann Fam Med. 2021;19(3):232-239. 

In This Article

Results

In total, 325 participants completed both prepresentation and postpresentation questionnaires (Figure 1). Most (94%) also completed the postconsultation questionnaire. Roughly a third of participants (28%) were consulting the family practitioner because their child (aged 0 to 7 years) was unwell. There were no significant differences in clinical and demographic characteristics comparing the participants who viewed the futility, adverse effects, and control presentations (Table 1).

Figure 1.

CONSORT diagram showing participant flow through the study.
CONSORT = Consolidated Standards of Reporting Trials.
aA total of 19 participants were lost to follow-up at postconsultation (N = 306 were followed up).
bAn additional 2 participants did not have data available for analysis (N = 304 had complete data for analysis).

Expectations and Beliefs About Antibiotics for URTI

Likert scores for participants' agreement with the statement "I wish to receive antibiotics for my/my child's cold/flu" before presentation viewing did not differ significantly across the 3 groups viewing different presentations. The mean (SD) score was 3.0 (0.3) in the futility group, 2.6 (0.3) in the adverse effects group, and 2.7 (0.3) in the control group (P = .20).

After viewing the presentation, the mean reduction in score for those who viewed the futility presentation (1.1; 95% CI, 0.8 to 1.3) or the adverse effect presentation (0.7; 95% CI, 0.4 to 0.9) was significantly greater than that for those who viewed the control presentation (0.1; 95% CI, 0.0 to 0.3) (Kruskal-Wallis H = 36.7, Cohen d = 0.7, P <.001) (Figure 2). This effect was also observed when analysis was restricted to the parents of the 91 child participants; the mean reduction in score for those who viewed the futility presentation or adverse effect presentation (0.8; 95% CI, 0.5 to 1.4) was significantly greater than that for those who viewed the control presentation (0.1; 95% CI, −0.2 to 0.3) (Kruskal-Wallis P <.01).

Figure 2.

Participants' expectation of receiving an antibiotic.
Notes: Participants used a 7-point Likert scale to express level of agreement with the statement, "I wish to receive antibiotics for my/my child's cold/flu." For graphical purposes, the Likert scores were collapsed to a 5-point scale by combining responses of 1 and 2 to represent "strongly disagree" and 6 and 7 to represent "strongly agree." Values in circles are means.

Similarly, after viewing the presentations, the mean reduction in Likert scores for participants' agreement with the statement "I think antibiotics are a helpful treatment for cold/flu" were significantly greater for those who viewed the futility presentation (0.9; 95% CI, 0.6 to 1.2) and adverse effect presentation (0.7; 95% CI, 0.4 to 1.1) than for those who viewed the control presentation (−0.1; 95% CI, −0.2 to 0.1) (Kruskal-Wallis H = 31.4, Cohen d = 0.6, P <.001).

Factors Associated With a Continuing Expectation to Receive Antibiotics

We classified participants who agreed or strongly agreed with the statement "I wish to receive antibiotics for my/my child's cold/flu" on the postpresentation questionnaire as having a continued expectation for antibiotics. The univariate odds ratios for factors associated with a continued expectation to receive antibiotics are shown in Supplemental Table 1, available at https://www.AnnFamMed.org/content/19/3/232/suppl/DC1/. Relative to participants in the control group, participants in the futility and adverse effect groups were less likely to continue to expect antibiotics (unadjusted odds ratio [OR] = 0.45; 95% CI, 0.27 to 0.75; P <.01, and unadjusted OR = 0.53; 95% CI, 0.32 to 0.86; P = .01, respectively).

We refined the multivariate ordinal regression model until it included variables that did not display multicollinearity and until the assumption of proportional odds was met. The final model was a good fit to the data (P = .38) and the model predicted participants' expectations to receive antibiotics better than the intercept-only model (P <.001). Participants in the futility group (adjusted OR [aOR] = 0.58; 95% CI, 0.35 to 0.95; P = .03) and in the adverse effects group (aOR = 0.44; 95% CI, 0.26 to 0.74; P <.01) were less likely to continue to expect antibiotics compared with peers in the control group (Table 2). Males (aOR = 1.75; 95% CI, 1.12 to 2.74; P = .02) and participants who reported having a sore throat (aOR = 1.62; 95% CI, 1.06 to 2.48; P = .03) were more likely to continue to expect antibiotics.

Prescribing, Dispensing, and Participant Satisfaction

Of the 306 participants who completed the postconsultation questionnaire, 30% (92) reported receiving an antibiotic prescription. There was no significant difference in this proportion between the futility group (31%), the adverse effects group (28%), and the control group (32%) (χ 2 = 0.361; P = .84).

Pharmacy dispensing information was available for 304 of the 306 participants who gave consent for their records to be reviewed. Overall, 28% (84) of these participants were dispensed an antibiotic in the 7 days after their consultation. The proportion did not differ between the futility group (26%), the adverse effects group (24%), and the control group (33%) (χ 2 = 1.68, P = .43). Interestingly, 40% of participants (37 out of 92) who reported receiving an antibiotic prescription did not have an antibiotic dispensed in the following 7 days, and 32% of participants (27 out of 84) who were dispensed an antibiotic did not report receiving an antibiotic prescription after their consultation.

Participants' level of satisfaction with their consultation did not differ across the 3 groups viewing different presentations. The mean (SD) score for satisfaction out of a possible 5 points was 4.7 (0.6) in the futility group, 4.8 (0.5) in the adverse effects group, and 4.7 (0.5) in the control group (F (2, 303) = 0.59, P = .56).

Factors Associated With Prescribing or Dispensing

We used binomial logistic regression analyses to determine the effect of various factors on receipt of an antibiotic prescription and on the dispensing of an antibiotic in the week after the consultation. Linearity of the continuous variables (age, perceived severity) was confirmed for each model.

Participants who strongly agreed with the statement "I wish to receive antibiotics for my/my child's cold/flu" after viewing the presentation were significantly more likely to receive an antibiotic prescription (aOR = 2.69; 95% CI, 1.27 to 5.69; P = .01) and to have an antibiotic dispensed in the 7 days after the consultation (aOR = 6.76; 95% CI, 2.92 to 15.67; P <.01) (Table 3). (Univariate ORs are given in Supplemental Table 2, available at https://www.AnnFamMed.org/content/19/3/232/suppl/DC1/.) In addition, participants who reported that earache was a predominant symptom for them or their child were more likely to receive a prescription (aOR = 2.36; 95% CI, 1.30 to 4.30; P <.01), and participants who perceived their/their child's illness as being severe were more likely to have an antibiotic dispensed (aOR = 1.49; 95% CI, 1.12 to 2.00; P = .01 for each point increase on the 7-point scale).

processing....