There were 1,854 adverse outcomes in the period before the letter, and 3,271 in the period after the letter.
The rate of patient and process-of-care outcomes per 100 EOC remained the same for amoxicillin, at 3.8 before and after the letter. The rate changed from 9.3 to 11.5 for macrolides, from 6.9 to 7.1 for cefaclor, from 9.0 to 9.6 for cephalexin, and from 7.9 to 11.3 for AC.
Of all EOCs that involved AC before the letter, 0.14 per 100 EOCs had a hospitalisation. Of all EOCs that involved AC after the letter, 1.84 per 100 EOCs had a hospitalisation, (p=0.0011).
For the EOCs in which the macrolides cefaclor, cephalexin and amoxicillin were given alone, there were no significant changes in the rate of each outcome.
When all EOCs for which at least two of these antibiotics were given as a total group were analysed, there were significant increases in the rate of adverse outcomes per 100 other antibiotic-related EOCs for the following:
hospitalisation, 0.44 before the letter and 0.86 after the letter, (p=0.0054);
radiologic investigations, 3.27 before the letter and 4.87 after the letter, (p=0.00001); and
pathologic investigations, 2.73 before the letter and 3.62 after the letter, (p=0.005489).
There was a significant association between the increase in the rate per month for all outcomes and the decrease in AC-prescription share, (p=0.011), with a 3-month lag.
There was also a significant association between AC-prescription share and the overall outcome rate weighted by relative cost, (p=0.0024).
The rate of patient outcomes, including return visits, was not significantly associated with AC-prescription share.
The rate of process-of-care outcomes was significantly associated with AC-prescription share also at the 3-month lag, (p=0.006).
After the first visit, there was no significant increase in the proportion of EOCs with a patient outcome before and after the letter, (p=0.1776). There was a significant increase in the proportion of EOCs with a process-of-care outcome after the letter, (p<0.0001). There was a significant increase in the proportion of EOCs in which tetracyclines were prescribed, (p=0.0304).
There was a significant increase in the proportion of EOCs that involved a patient outcome before and after the letter (2.3% before and 6.1% after; p<0.0001).
There was also a significant increase in the proportion of EOCs with a process-of-care outcome after the letter (4.3% before and 10.2% after; p<0.0001).