FormalPara Key Points

Almost half of community-dwelling older people used one or more psychotropic drug.

Despite a limited sample, we found notable sex differences in psychotropic drug use.

Psychotropic drug use was also associated with multi-dose dispensing.

1 Introduction

Older people are on average more susceptible to the adverse effects of psychotropic drugs; for example, many psychotropics are associated with drug-induced cognitive impairment [1, 2], and antipsychotic drugs increase the mortality risk [3, 4]. However, addressing older people as a homogenous group based on age alone can be misleading since there are many patient factors that could predict psychotropic drug use. There are well-documented differences related to sex and gender, as females use more psychotropic medicines than do males in various countries and settings [5,6,7]. Among other factors, multi-dose dispensing (MDD) is commonly used by older people and has previously been associated with poor quality of drug treatment in terms of psychotropic polypharmacy [8]. Excessive use of psychotropics has previously been documented in various nursing home settings [9,10,11], and the association with that type of residency is notably strong in comparison with other factors [12]. Still, older people in ordinary private homes are a loosely defined group where psychotropic drug use may also be extensive. One example of such individuals is people who are the subjects of hospital-based care, since somatic disease has a detrimental impact on mental health [13].

This study aimed to describe psychotropic drug use and associated factors among community-dwelling older people who have been admitted to hospital. Our primary objective was to show the prevalence of psychotropic drug users in this study population. A second objective was to analyse patient factors associated with psychotropic use.

2 Method

2.1 Study Population and Data Sources

The study comprised a sample of 300 community-dwelling people 75 years or older who had been admitted to the acute medical ward at Umeå University Hospital from September 2018 to November 2021. The full eligibility criteria are available in a published version of the study protocol for the clinical trial in which the study participants were enrolled [14]. During that period, other people had either been given the opportunity to participate but declined (n = 71) or withdrawn their consent to participate after enrolment (n = 19). Data for the study were extracted from electronic medical charts, using the epicrisis, the overview of diagnoses and the list of medicines based on the day before hospital admission. For each participant, personal information was pseudo-anonymized.

2.2 Statistics

Psychotropic drugs were categorized according to the Anatomical Therapeutic Chemical (ATC) classification system [15] as (1) N05A, antipsychotics; (2) N05B, anxiolytics; (3) N05C, sedatives and hypnotics; (4) N06A, antidepressants; and (5) N06D, antidementia drugs. Associations between psychotropic drug use and independent variables comprising sex, age, cohabitation, cardiovascular disease, diabetes type 2, major neurocognitive disorder (NCD) and MDD were analysed through logistic regression. In this analysis, psychotropic drug use functioned as a dependent variable and was treated as one single category that excluded antidementia drugs. We performed separate sensitivity analyses of antidepressants and sedatives or hypnotic drugs. We used IBM SPSS Statistics, Version 28.0 for all descriptive and analytical statistics. For all inferential statistical tests, p values < 0.05 were considered significant. We used variance inflation factor (VIF) to assess multicollinearity in our model; variables with values < 5 were not further checked.

2.3 Ethical Considerations

All study participants gave their consent to provide data for the study, and in cases of confirmed diagnosis of major NCD, the next of kin also approved. Personal data handling was performed within the framework of the concurrent clinical trial. The study was approved by the Swedish Ethical Review Authority (registration numbers 2020-05426, 2020-03699, 2017-69-31, 2018-254-32, 2018-83-32 and 2024-03795-02).

3 Results

The sample of 300 individuals was predominantly female and had a mean age of 84 years (Table 1). More than 60% of the patients had at least one recorded cardiovascular disease, here defined as coronary heart disease, heart failure or stroke, and almost one-third had records of diabetes type 2. There were 29 confirmed cases of major NCD, of which 20 were females. Alzheimer’s disease (n = 11) was the most common NCD subtype, followed by vascular dementia (n = 10) and Lewy body dementia (n = 3). In the subgroup diagnosed with major NCD, 71% of the people were female.

Table 1 Sample characteristics
Full size table

Almost half of the 300 patients had at least one psychotropic drug on their list of medicines, and 17.7% of the whole sample used two or more psychotropics (Table 2). In that respect, we observed that the proportions of individuals using two or more psychotropics were 24.7 and 14.9% among people with and without MDD service, respectively (p = 0.044). Sedatives and hypnotics where the most common drugs at the subgroup level, of which zopiclone was the individual drug displaying the highest user proportion of all psychotropics. Antidepressants comprised the second most prevalent drug class, whereas the user proportions of anxiolytics and antipsychotics were relatively low in comparison.

Table 2 Proportions of patients using psychotropic drugs
Full size table

Regarding notable patterns of drug use between the sexes, 54.8% of the females used at least one psychotropic drug compared with 37.5% of the males (p = 0.004). A similar observation (34.6 vs 23.2%, p = 0.048) was observed at the ATC subgroup level for sedatives and hypnotic drugs. When exploring the user proportions of sedatives and hypnotics further, zolpidem was notably more prevalent among the females than among the males, with percentages of 13.3 and 3.6%, respectively (p = 0.006). As a final notion about drug treatments, eight of the 14 individuals who had been diagnosed with either Alzheimer’s disease or Lewy body dementia used antidementia drugs.

Among the factors analysed (Table 3), female sex and MDD were positively associated with psychotropic drug use, both in simple and multiple logistic regression models. In the sensitivity analysis, we could not confirm an association between female sex and antidepressants (see eTable 1 in the electronic supplementary material), but female sex was positively associated with use of sedatives and hypnotics (eTable 2). Regarding MDD, on the other hand, there was an association with antidepressants (eTable 1) but not sedatives or hypnotics (eTable 2).

Table 3 Logistic regression analyses regarding factors associated with psychotropic drug use
Full size table

4 Discussion

4.1 Findings

This cross-sectional study was based on a sample of 300 community-dwelling people 75 years or older admitted to the acute medical ward. We found that approximately half of the sample used at least one psychotropic drug, and the use of such medicine was positively associated with female sex and MDD system.

As seen in Table 1, the study sample represented a multimorbid population with an extensive use of medicines, including psychotropics. For example, the user proportion of sedatives and hypnotics was higher than the percentages reported at the national level for older people in general [12]. This is concerning since hypnotic drug use (e.g. zolpidem and zopiclone) in older people is associated with apparent risks [16,17,18], even though there is no evidence pointing at hypnotic drugs being a common cause of hospital admission in this specific study population [19]. Previous research has reported on associations between poor mental health and somatic diseases, for example, cardiovascular disease [20] and diabetes type 2 [21]. In that respect, no associations between psychotropics and somatic comorbidities could be statistically confirmed in our study. Neither could we see any significant differences in psychotropic drug use in relation to major NCD, perhaps due to the limited study sample.

This study showed a pronounced difference between the sexes, with female and male user proportions of 54.8 and 37.5%, respectively (adjusted odds ratio [OR] 2.05 [95% confidence interval {CI} 1.22–3.42]). According to the sensitivity analysis, this result was likely due to a pronounced sex difference regarding sedatives and hypnotic drugs (eTable 2; see the electronic supplementary material). This dissimilarity between the sexes has, as mentioned, been observed before across various countries [5,6,7] and also in different subgroups of older people at the national level [12], and the explanation behind this pattern is likely multifactorial. According to previous research, women have a more positive help-seeking attitude regarding mental health problems compared with men, which in turn contributes to the higher rates of psychotropic drug use [22]. Furthermore, women live on average longer than do men and have more disabling chronic conditions [23]. Major NCD is one such example, and management of neuropsychiatric symptoms could have been a plausible indication for treatment that contributed to our observed discrepancy between the sexes.

Among the different ATC subgroups, sedatives and hypnotics were the most prevalent psychotropic medications among both males and females; however, females used them to a larger extent. This result was expected given that women seem to have a higher occurrence of troublesome sleep disorders (e.g. insomnia) compared with men [24, 25]. The greatest sex difference at the substance level was found among zolpidem users, since the use of zolpidem was 3.7 times more common among females compared with males (p = 0.006). These findings regarding hypnotics, and zolpidem specifically, are in line with a registry-based study from Sweden on gender and hypnotic drug use [26] and, although of uncertain relevance, important from a safety perspective since the plasma concentration resulting from a given dose of zolpidem is generally higher in females than in males [27, 28]. Taken together, our observed sex differences in combination with the high user proportions of hypnotic drugs are an indication to further study sex and gender differences in psychotropic drug use to increase knowledge about possible over- or undertreated symptoms within this vulnerable patient group.

Finally, the odds of using psychotropic drugs were more than twice as high among patients who had an MDD system compared with those who did not receive that service (adjusted OR 2.20 [95% CI 1.23–3.93]), and were even higher when analysing antidepressants separately (eTable 1). We also found that using two or more psychotropics was more common among those with an MDD system (p = 0.044). The latter observation was similar to observations from an earlier registry-based Swedish study [8], in which the burden of disease could not fully explain why many patients with an MDD system had multiple psychotropic drugs. Similar results have also been reported for MDD users and potentially inappropriate medicines in a Norwegian context [29]. Moreover, MDD has been associated with fewer changes compared with ordinary prescriptions, although that Swedish study did not specifically target psychotropic drugs [30]. Hence, our observation is in line with the possible explanation that renewing all prescriptions together as a single package might, over time, impede re-evaluation and withdrawal of psychotropics. Still, cross-sectional comparisons have limitations, and it could also be that polypharmacy inevitably increases the need for MDD. Longitudinal studies that investigate MDD systems in relation to number of medicines over time might provide further clues to the exact direction of the association.

4.2 Strengths

The broad inclusion criteria applied in this study enabled us to study a clinical sample of patients with high representativeness of vulnerable community-dwelling older people who are encountered within acute medical care. Moreover, the information on medicines, living situation and diseases was based on up-to-date medical records, matching the date of the index hospital admission.

4.3 Limitations and Risk of Bias

A limitation of the dataset was that there was no exact information on indication for treatment; hence, it is difficult to draw any conclusions concerning medication appropriateness. For example, the results do associate female sex with psychotropic drug use, but it does not say whether psychotropics are overprescribed or if there are underlying differences in disease occurrence and severity that ultimately warrants the current pharmacological approach. Second, the patients’ medication lists might not reflect drug use in practice; psychotropic drug use could consequently have been under- and overestimated, most likely the latter due to possible adherence issues.

5 Conclusions

The most common psychotropic drugs used by community-dwelling older people admitted to the acute medical ward were antidepressants and hypnotic drugs. Female sex and MDD system were positively associated with psychotropic drug use. Further studies concerning those two factors in relation to potential overprescribing could provide a better picture on how to optimize psychotropic drug use among vulnerable older people.