Patient demographics and psychotropic medication prescribing in Australian general practices: pre- and during COVID-19 pandemic

Data source and study design

The study utilised data from electronic health records of 888 general practices in the Australian states of Victoria and New South Wales. Data were obtained using the Population Level Analysis and Reporting (POLAR) tool by Outcome Health, the data custodians, with identifying information removed and unique patient de-identified keys assigned. The methods used by the POLAR tool are detailed elsewhere.¹⁰

The study timeframe encompassed all psychotropic medication prescriptions from January 2018 to December 2022. Ethical approval to use general practice data for research purposes was obtained by the data custodian, from the ethics committee of the Royal Australian College of General Practitioners under reference number 17-008. Consent was obtained from general practitioners for the inclusion of their electronic health records in the study, and patients were provided with the opportunity to opt-out of having their data used for research. Ethics approval to use the data for this research was obtained from the Macquarie University Human Research Ethics Committee (Reference number: 52020675617176).

Population and outcomes

The target population for this study comprised all general practice patients aged 20+ years who had received a prescription for psychotropic medications with at least one repeat. The primary outcome was frequency of prescriptions for four Anatomical Therapeutic Chemical (ATC) categories of psychotropic medications: antidepressants, antipsychotics, anxiolytics, and hypnotics and sedatives. Additionally, the trends of first time prescribing of psychotropic medicines pre- and during the COVID-19 pandemic (2018 22) were evaluated. Demographic characteristics such as patient age (categorised into 20–39, 40–59, 60+ year age groups), sex, socio-economic status (SES) (low, low-mid, mid, high-mid, high), place of residence (metropolitan vs regional/remote), state of residence (NSW vs Victoria), and years of first-time medication prescribing were assessed for differences in psychotropic medication prescribing. We used the Modified Monash Model (MMM) to determine remoteness of areas.¹¹ The MMM measures remoteness and population size on a scale of seven categories where MM 1 is a major city (categorised as metropolitan in this research) and MM 7 is very remote. SES was determined using the Australian Bureau of Statistics’ Socio-Economic Indexes for Areas,¹² based on the relative socio-economic disadvantage index (IRSD).

Statistical analysis

We summarised categorical variables as counts and proportions. A mixed effects negative binomial regression model was employed to analyse the association between demographic variables with the rates of psychotropic medication prescribing. The negative binomial model was selected over other model families due to its capability to account for data dispersion, and the mixed effects model is particularly adept at addressing potential clustering and interdependencies observed among patients from the same practice.¹³ We also used patient follow-up length between their first and last prescription as an offset to standardise the counts and allow meaningful comparisons across different observation time intervals. Incidence risk ratios (IRRs) with corresponding 95% confidence intervals (CIs) were presented as measures of association.

We also examined the trends of first-time monthly prescriptions of each psychotropic medication as a percentage of total GP prescriptions for any drug in NSW and Victoria from January 2018 to December 2022. A seasonally adjusted autoregressive moving average (SARIMA) time series model was fitted using data spanning from 1 January 2018 to 28 February 2020 to forecast trends after the start of the COVID-19 pandemic in Australia. A comparison was plotted between the actual data and the predicted trend, highlighting the disparities due to the pandemic versus the hypothetical scenario where it did not occur. Before forecasting, the model was evaluated using residual plots, the Ljung-box test, and inverse autoregressive (AR) plots to check if it best fits the data¹⁴ (Supplementary Fig. S1a, b and Supplementary Table S1). The proportion of missing values in our dataset was less than 3% and we applied listwise deletion, as it would have little effect on the outcomes of our study. All statistical analyses were conducted using R statistical software version 4.3.1.¹⁵

Demographic characteristics

The study included a total of 154 528 patients (Supplementary Fig. S2) aged 20 years and above who had received prescriptions of more than one psychotropic medication in any of the ATC categories. Among these patients, 56.2% were female, 41.4% were aged between 20 and 39 years, 22.2% were aged 60+ years, 83.7% resided in metropolitan regions, and 62.8% were from the high SES category (Table 1).

Demographic factors associated with rates of psychotropic medication prescribing

Prescribing of Antidepressants

A total of 68 203 patients were prescribed antidepressants within the study period. Male patients had fewer antidepressant prescriptions than female patients (IRR = 0.95; 95% CI: 0.94, 0.97). Antidepressant prescribing rates increased with age, with an IRR of 1.07 (95% CI: 1.05, 1.08) for age group 40–59 years and 1.54 (95% CI: 1.51, 1.56) for age group of 60+ years compared with age group 20–39 years. Patients in the low (IRR = 1.14; 95% CI: 1.10, 1.18), low-mid (IRR = 1.09; 95% CI: 1.06, 1.12), mid (IRR = 1.24; 95% CI: 1.18, 1.30), and high-mid (IRR = 1.08; 95% CI: 1.06, 1.10) SES categories had higher rates of antidepressant prescribing compared to patients in the high SES category (Fig. 1).

Fig. 1.

Demographic characteristics associated with prescribing psychotropic medications (antidepressants, antipsychotics, anxiolytics, and hypnotics and sedatives) at general practices in NSW and Victoria, 2018–2022. Note: IRR, incidence risk ratio (adjusted); Metro, metropolitan; Reg, regional; SES, socio-economic status; *Categories to the right of vs are reference.

Prescribing of Antipsychotics

A total of 24 577 patients had antipsychotics prescribed within the study period. Prescribing rates of antipsychotics were lower in the age group 40–59 years (IRR = 0.95; 96% CI: 0.94, 0.99) and higher in the age group 60+ years (IRR = 1.34; 95% CI: 1.30, 1.38) compared to the age group 20–39 years. Patients in the low (IRR = 1.27; 95% CI: 1.20, 1.36), low-mid (IRR = 1.15; 95% CI: 1.09, 1.21), mid (IRR = 1.37; 95% CI: 1.24, 1.52), and high-mid (IRR = 1.12; 95% CI: 1.08, 1.17) SES categories had higher rates of antipsychotic prescriptions compared to patients in the high SES category. The rate of prescribing antipsychotics steadily increased every year with an IRR of 1.40 (95% CI: 1.35, 1.45) in 2020, 2.0 (95% CI: 1.89, 2.08) in 2021, and 4.48 (95% CI: 4.40, 4.78) in 2022 compared to 2018 (Fig. 1).

Prescribing of Anxiolytics

A total of 22 547 patients had anxiolytics prescribed within the study period. Prescribing rates of anxiolytics were higher in the age group of 60+ years compared to age group 20–39 years (IRR = 1.41; 95% CI: 1.36, 1.45). Patients in the low (IRR = 1.12; 95% CI: 1.04, 1.21), low-mid (IRR = 1.14; 95% CI: 1.07, 1.20), mid (IRR = 1.36; 95% CI: 1.23, 1.51), and high-mid (IRR = 1.16; 95% CI: 1.12, 1.21) SES categories received more frequent anxiolytic prescriptions compared to patients in the high SES category. The rate of prescribing also increased every year compared to the 2018 reference with an IRR of 1.11 (95% CI: 1.07, 1.14) in 2019, 1.46 (95% CI: 1.40, 1.52) in 2020, 2.09 (95% CI: 1.99, 2.20) in 2021, and 4.79 (95% CI: 4.48, 5.12) in 2022 (Fig. 1).

Prescribing of Hypnotics and sedatives

A total of 39 201 patients had hypnotics and sedatives prescribed within the study period. Male patients had fewer hypnotics and sedatives prescriptions than female patients (IRR = 0.97; 95% CI: 0.96, 0.99). Patients in the age group of 60+ years had higher prescribing rates of hypnotics and sedatives compared to age group 20–39 years (IRR = 1.46; 95% CI: 1.43, 1.50). Patients in the low (IRR = 1.14; 95% CI: 1.07, 1.20), low-mid (IRR = 1.14; 95% CI: 1.09, 1.18), mid (IRR = 1.24; 95% CI: 1.15, 1.34), and high-mid (IRR = 1.11; 95% CI: 1.08, 1.14) SES categories received more frequent hypnotics and sedatives prescriptions compared to patients in the high SES category. Compared to 2018, the prescribing rates demonstrated a consistent rise each year. Specifically, the IRRs were 1.03 (95% CI: 1.01, 1.05) in 2019, 1.38 (95% CI: 1.34, 1.42) in 2020, 1.90 (95% CI: 1.84, 1.97) in 2021, and 4.60 (95% CI: 4.38, 4.82) in 2022 (Fig. 1).

Trends of first-time psychotropic medication prescribing

Between 2018 and 2020, there was a gradual increase in first-time antidepressant prescriptions, with a substantial increase coinciding with the commencement of the COVID-19 pandemic. For the other psychotropic medications (antipsychotics, anxiolytics, and hypnotics and sedatives) prescribing trends decreased between 2018 and 2020, but they spiked around the period of the COVID-19 pandemic. First time prescribing of psychotropic medication trends showed a decreasing pattern beginning in 2022 (Fig. 2).

Fig. 2.

Trends of first-time psychotropic medication prescribing at general practice in NSW and Victoria, 2018–2022.

Strength and Limitations of the study

Our results are based on a large sample of general practices across two states in Australia, with data extracted from electronic health records and therefore closely representing general practice activity. The data reflects the vast majority of mental health prescriptions, since GPs prescribed over 85% of all mental health medications in Australia in 2021–22. However, the study results are based on medication prescriptions and do not necessarily reflect patient behaviour. Therefore, we do not know if the patients are filling the prescriptions and using the medications as directed by their general practitioners. These medications could also be prescribed for medical indications other than mental.