Methods

Results

The pooled population comprised 106,307 individuals in 2019 and 113,343 in 2023. Of these, a fourth were children; most were between 19 and 65 years old, while 7%–8% were over 65 years old. Gender was equally distributed. Most patients lived in urban and peri-urban areas (90%). Most (92%) had no PCGs (Table 1). These patients were prescribed 24,601 prescriptions for antibiotics in 2019 and 26,479 in 2023.

Table 1.

Demographic characteristics of the pooled analytical population, 2019 and 2023—full data for 2019–2023 can be found in appendix 3. Values are weighted n (%). PCG, pharmaceutical cost group. Weighted counts were rounded to the nearest whole number

Table 1. Long description

A table with 11 rows and 4 columns comparing demographic characteristics of a population in 2019 and 2023. The columns are labeled ‘Characteristic’, ‘2019 (N=106,307)’, and ‘2023 (N=113,343)’. The table includes data on age groups, gender, municipality types, and the number of pharmaceutical cost groups (PCGs). Row 1: Age group 0-18, 24,641 (23.2 percent), 29,990 (26.5 percent). Row 2: Age group 19-65, 74,722 (70.3 percent), 74,156 (65.4 percent). Row 3: Age group over 65, 7,708 (7.3 percent), 10,016 (8.8 percent). Row 4: Gender (male), 55,272 (52.0 percent), 59,115 (52.2 percent). Row 5: Urban municipality, 72,987 (68.7 percent), 77,136 (68.1 percent). Row 6: Peri-urban municipality, 23,036 (21.7 percent), 24,599 (21.7 percent). Row 7: Rural municipality, 11,047 (10.4 percent), 12,427 (11.0 percent). Row 8: Number of PCGs = 0, 98,189 (92.4 percent), 102,390 (90.3 percent). Row 9: Number of PCGs = 1, 6,940 (6.5 percent), 9,237 (8.1 percent). Row 10: Number of PCGs = 2, 993 (0.9 percent), 1,460 (1.3 percent). Row 11: Number of PCGs >= 3, 184 (0.2 percent), 256 (0.2 percent).

We observed a change in antibiotic prescription fill rate and prescriber groups over time. In our data set, we saw a decline in antibiotic prescription fills in 2020. After that, prescription fills increased, surpassing 2019 levels by 2023 (Figure 1).

Figure 1.

Change of antibiotic prescription fills per 1,000 enrollees from 2019 to 2023; Q1: first quarter of the year, Q2: second quarter of the year, Q3: third quarter of the year, Q4: fourth quarter of the year.

The registered PCPs had been responsible for 40% of antibiotic fills in 2019 but, by 2023, their share decreased to 36% while the percentage of other providers issuing antibiotic dispensations increased from 60% to 64% (P value < .001, Figure 2).

Figure 2.

Rate of antibiotic prescription fills by prescriber in 2019 and 2023.

In 2019, other mixed-specialty group practices and general practitioners were responsible for approximately 23% of antibiotic dispensations, compared with 24% in 2023. Hospital outpatient services accounted for approximately 15% in 2019 and 16% in 2023. The proportions attributable to urologists (2%), gynecologists (5%), and other specialists (6%) remained largely unchanged (Figure 3).

Figure 3.

Antibiotic prescription fills by prescriber group in 2019 and 2023.

When we used the WHO’s guidelines and categorized the antibiotics by stewardship category, we observed shifts over time. In 2019, averages were 57% “Access,” 35% “Watch,” and 8% “Reserve.” In 2023, this shifted to 62% “Access,” 30% “Watch,” and 8% “Reserve.” For prescriptions issued by other physicians and hospitals, 2019 averages were 66% “Access,” 25% “Watch,” and 9% “Reserve.” In 2023, these proportions shifted to 71% “Access,” 21% “Watch,” and 8% “Reserve” (Figure 4).

Figure 4.

Antibiotic prescription fills by WHO classification in 2019 and 2023.

Discussion

Our main finding was a shift in outpatient antibiotic prescribing between 2019 and 2023. Although the total number of prescription fills increased, the proportion issued by registered PCPs decreased from 40% to 36%, while the contribution of other healthcare providers increased. Registered PCPs issued a higher proportion of Watch-classified antibiotics than other providers. Previous studies have highlighted that primary care is a significant contributor to outpatient antibiotic prescriptions, with 80% of such prescriptions being attributed to this sector. ^{Reference Daouk, Heim and Kronenberg8,10,Reference ECfDPa.25} Although efforts have been made to inform PCPs of the risks of inappropriate prescribing ^{Reference Butler, Simpson and Dunstan6,Reference Avent, Cosgrove and Price-Haywood26} , our findings show that a substantial portion of antibiotic use comes from other providers. The most recent analysis from data of another health insurer in Switzerland found a similar percentage of antibiotic prescriptions for general internal medicine doctors, which includes PCPs. ^{Reference Graber, Stollberg and Plüss-Suard21} Our study extends these findings by showing that substantial prescribing outside the registered PCP also occurs among patients enrolled in a managed-care plan who already have a designated gatekeeping PCP.

The decline in the proportion of prescriptions issued by registered PCPs coincided with an increase in total antibiotic prescriptions and a higher contribution from other healthcare providers. These concurrent trends do not establish that prescribing was displaced from registered PCPs to other providers or that they resulted from previous stewardship activities. Rather, they demonstrate that outpatient antibiotic prescribing is distributed across a broader care network, including other primary care practices, walk-in services, specialists and hospital outpatient departments. Interventions should also be designed to target in-hospital prescribers and walk-in practices that may see many patients with acute infections. Given the increasing shortage of PCPs in Switzerland, ^{Reference Stierli, Rozsnyai and Felber27} an important feature of our population was that all patients were enrolled in a managed-care plan and had a designated gatekeeping PCP. Therefore, the substantial proportion of antibiotic prescriptions issued by other healthcare providers is less likely to be explained solely by patients not having an established PCP. Nevertheless, registration with a gatekeeping PCP does not necessarily guarantee timely access to primary care and limited appointment availability may still have contributed to patients seeking care in other outpatient services.

Gatekeeping PCPs prescribed a higher proportion of Watch and a lower proportion of Access antibiotics than other healthcare providers, while the proportion of Reserve antibiotics was similar. The underlying reason remains unclear, and the available claims data cannot determine whether this difference reflects inappropriate antibiotic selection. Differences in infection type, patient characteristics or clinical case mix may contribute. To reduce the prescription of “Watch” antibiotics, antibiotic stewardship should focus not only on the appropriate timing of prescriptions but also on selecting the most suitable drug. ^{Reference Dryden, Johnson and Ashiru-Oredope28}

Our study benefits from comprehensive data provided by SWICA, a major health insurance provider in Switzerland, which enabled a robust analysis of outpatient antibiotic prescribing patterns. The overall prescription rates from 2019 to 2023 were comparable to data gathered by the Swiss Federal Office of Public Health, which supports the validity of our data. ^{Reference Daouk, Heim and Kronenberg8} The sharp decline in antibiotic prescriptions in 2020 is explained by the COVID-19 pandemic. ^{Reference Plüss-Suard, Mueller and Plate29}

Our study has several limitations. First, our study was conducted in the German-speaking region of Switzerland using claims data from a single health insurance company. Therefore, the findings may not be generalizable to other insurers or to Switzerland’s other language regions, where populations and prescribing practices may differ. ^{Reference Daouk, Heim and Kronenberg8,Reference Szymczak and Linder30,Reference Friedli and Labutin31} Second, we only included patients who had a managed-care contract, so our study group includes a lower proportion of older adults than found in the general population. Third, we could not precisely identify the specialists in the “group practice” and “hospital outpatient” categories, so we could not determine which specialists were responsible for prescribing antibiotics within these settings. Finally, we decided not to calculate Defined Daily Doses (DDDs) as we were interested in the frequency of antibiotic prescription, rather than the dosage per use. However, a comparable analysis using DDDs showed a nearly identical distribution of prescriber categories when examining overall antimicrobial use. ^{Reference Graber, Stollberg and Plüss-Suard21}

We found that the proportion of antibiotic prescriptions among healthcare providers shifted over time. Because the proportion of prescriptions issued by gatekeeping PCPs declined between 2019 and 2023, we suggest extending antimicrobial stewardship interventions to other groups of providers in both hospital and outpatient settings. Antibiotic stewardship in PCPs should additionally focus on reducing the number of broad-spectrum antibiotics prescribed.