In total, 5,528,911 BDZ prescriptions were dispensed to 397,436 individuals (260,591 females, 136,845 males) during the 10-year study period.

During the study period the mean monthly prevalence rate of BDZ use in Estonia was 28.0 users per 1000 inhabitants, being lowest in July 2012 and highest in March 2020, with 23.3 and 32.7 users per 1000 inhabitants, respectively. BDZ use is twice as common in females as in males, the mean monthly prevalence rates were 36.5 and 18.5 users per 1000, respectively. BDZ use increased with age, the mean monthly prevalence ranged from 0.3 users per 1000 among children under 15 years to 92.8 users per 1000 among those aged 75 years and older. The most frequent clinical indication for BDZ prescription was sleep disorders (with a mean monthly prevalence of 12.5 users per 1000), followed by anxiety, depressive, and other mental and behavioral disorders (with mean monthly prevalence of 3.5, 4.2, and 4.5 users per 1000, respectively), and epilepsy was rare (with a mean monthly prevalence of 0.3 users per 1000). BDZs are most frequently prescribed by GPs (with a mean monthly prevalence of 21.4 users per 1000), and more rarely prescribed by psychiatrists and other specialties (with mean monthly prevalence of 4.4 and 1.9 users per 1000, respectively).

We observed a significant temporary (pulse) increase in the overall prevalence rate of BDZ use in March 2020 (2.698 users per 1000, 95% CI 1.408–3.988), but no significant change in the time trend (slope) from March 2020 until the end of the study period, December 2021 (p = 0.64) (Table 1).

A significant temporary increase in the prevalence of BDZ use in March 2020 was observed across all subgroups examined, except for new users, users aged 15–29 years, and those prescribed BDZs by specialties other than GPs and psychiatrists (Table 1). For indications, a temporary increase in the use for anxiety and depressive disorders was the highest, at 0.620 users per 1000 (95% CI 0.424–0.816) and 0.565 users per 1000 (95% CI 0.347–0.782), respectively.

However, those aged 15–29 were the only age group that showed a significant increase in the slope (0.032 users per 1000 per month, 95% CI 0.014–0.050), which resulted in 15.8% more users by the end of the study period (observed in December 2021 6.818 users per 1000 vs. predicted in December 2021 5.888 users per 1000). Also, prescribing by other specialties increased in longer term (Table 1).

Adolescents and young adults aged 15–29 years were the only group that experienced significant gradual long-term effect in BDZ use. Stratifying by gender, the increase in slope was significant among young females (0.056 users per 1000 per month, 95% CI 0.033–0.079), but not among males (0.009 users per 1000 per month, 95% CI – 0.017–0.035) (Fig. 1). Additionally, a significant temporary increase in March 2020 was observed for females aged 15–29, but not for males of the same age.

Interrupted time-series analysis of prevalence rate of benzodiazepine and Z-drug use among females (a) and males (b) aged 15–29 years in Estonia, January 2012 to December 2021 (red line—observed rate; green line—counterfactual predicted by ARIMA model in absence of the COVID-19 pandemic)

Among females aged 15–29 with prior BDZ use, both a significant temporary increase (0.712 users per 1000, 95% CI 0.398–1.027) and a significant increase in the slope were observed (0.047 users per 1000 per month, 95% CI 0.027–0.067), which resulted in 28.0% more users by the end of the study period. However, among new users, only an increase in the slope was present (0.020 users per 1000 per month, 95% CI 0.012–0.028), which resulted in 20.7% more users by the end of the study period.

Among females aged 15–29, a significant increase in slope was observed for depressive disorders (0.021 users per 1000 per month, 95% CI 0.012–0.030), other mental and behavioral disorders (0.020 users per 1000 per month, 95% CI 0.010–0.030), and anxiety disorders (0.017 users per 1000 per month, 95% CI 0.010–0.023) (Fig. 2), which resulted in 45.1%, 31.5%, and 19.9% more users by the end of the study period, respectively. No significant change in use for sleep disorders was detected (– 0.008 users per 1000 per month, 95% CI– 0.018 to 0.002).

Interrupted time-series analysis of prevalence rate of benzodiazepine and Z-drug use among females aged 15–29 years in Estonia, January 2012 to December 2021, by main indications (red line—observed rate; green line—counterfactual predicted by ARIMA model in absence of the COVID-19 pandemic)

By prescribers’ specialties, a significant increase in the trend of prescribing BDZ for females aged 15–29 was observed only for psychiatrists (0.026 users per 1000 per month, 95% CI 0.016–0.037), but not for GPs or other specialties.

Among females aged 15–29, the increase per month in the prevalence of BDZ use after the COVID-pandemic onset was significant for the age groups 15–19 and 20–24, which resulted in 36.3% and 29.6% more users, respectively, by the end of the study period (Table 2).

A significant increase in the trend of new users was detected among females aged 15–19 and 20–24 years, which corresponded to a 75.0% and 57.7% increase in the prevalence of use at the end of the study period, respectively. However, among prior users, a significant increase per month was observed for females aged 15–19 years and 25–29 years, but not for those aged 20–24 years (Table 2).