In Sweden, use of multi-dose dispensing (MDD) amongst adults aged 70 years or older increased from 10.6% in 2014 to 12.1% in 2023. We did not observe an additional increase during the COVID-19 pandemic (2020–2022). Uptake increased in almost all subgroups, but particularly in people diagnosed with dementia (+45.3 percentage points), those receiving home help services (+21.4 percentage points) and those with five or more chronic conditions (+15.3 percentage points). Over time, the profile of MDD users shifted towards community-dwelling adults receiving home help services, younger older adults (70–79 years old) and individuals with higher education. Overall, the findings indicated that MDD has become the predominant mode of medication management in Swedish nursing homes, as  96.1% of residents used MDD by 2023. Results show increased prevalence in community settings over time, indicating that either MDD or its target population of older adults with complex health needs are increasingly present in the community. The COVID-19 pandemic had little discernible impact on MDD prescribing trends.

In addition to providing the first national estimates of MDD prevalence across clinically relevant subgroups and during the COVID-19 pandemic, our prevalence estimates serve as an update to those of Sjoberg et al., who reported that 11% of adults aged >65 years used MDD in Sweden in 2008 [10]. While there is no published cross-country comparison on prevalence, some articles report the number of MDD users at a national level which are substantially lower than our estimates for older adults in Sweden, for example, 2.2% in the Netherlands in 2011 [7], 0.4% in Finland in 2007 [7] and 0.8% in Denmark in 2010 [30]. However, these figures refer to the general population rather than older adults specifically, and differences in population subgroups, definitions, and time windows (both ages and calendar years) make direct comparisons with the aforementioned studies difficult.

The 1.5-percentage-point increase in crude MDD prevalence over 10 years represents a 14.2% relative increase, corresponding to almost 55,000 additional users. Notably, this increase occurred despite a decline in the proportion of older adults residing in nursing homes (from 5.9 to 4.5%) and receiving home help services (from 10.9 to 8.5%, see Supplementary Table 7), the two settings where MDD use is most concentrated. This suggests that the overall increase was driven by growing MDD adoption amongst community-dwelling older adults, who made up an increasing share of the population (from 83.2 to 87.0%). The relatively modest increase in crude prevalence contrasts with much steeper increases in certain high-risk subgroups, particularly people with dementia and those with ≥ 5 chronic conditions. Over time, these populations made up a progressively smaller share of the cohort, while MDD prevalence remained low or even declined in the expanding groups, such as those with ≤ 1 chronic condition. As a result, substantial increases in MDD use within some of these strata translate into only a modest rise in overall prevalence. Taken together with the general shift from nursing homes to home help and community living, these patterns are consistent with MDD use becoming increasingly concentrated amongst older adults with complex health needs living outside institutions. This could indicate that MDD is being used as part of broader efforts to support ageing in place, including home help services that frequently assist with medication management [5]. It should be noted that the shift in the profile of MDD users may partly reflect broader changes in where older adults receive care rather than changes in who is prescribed MDD, as the same patients who would previously have been in nursing homes are now increasingly managed in the community.

The larger increases in MDD prevalence amongst women and widowed individuals likely reflect that women live longer than men and are therefore more likely to survive into older age with multiple chronic conditions requiring medication support [31]. Widowed older adults, who are predominantly women, may also have a greater need for MDD as they lack a cohabiting partner who could assist with medication management. The apparent decline in the share of MDD users with dementia in nursing homes should be interpreted considering two concurrent trends. First, MDD prevalence amongst nursing home residents with dementia continued to increase over the study period, approaching near-universal use. Second, the overall profile of MDD users shifted from nursing homes towards community settings, meaning that while MDD use grew amongst people with dementia in nursing homes, it grew even faster in the community, diluting the relative share of nursing home residents amongst all MDD users.

The growth in MDD prevalence should also be considered in the context of increasing drug utilisation amongst older adults. In our data, the medication burden amongst MDD users shifted upwards over time, with the share using ten or more concurrent drugs increasing from 33.7 to 42.4%, while the median number of drugs rose from eight to nine (Supplementary Table 7). Amongst non-MDD users, the median number of drugs remained stable at four throughout the study period, suggesting that the increasing drug burden was concentrated amongst MDD users. This pattern aligns with broader national trends: a recent study using Swedish register data from 2006 to 2020 found that polypharmacy prevalence amongst older adults increased from 28.6 to 32.8%, driven primarily by faster medication accumulation amongst those aged 75–89 years [32]. Excessive polypharmacy (≥ 10 drugs) also increased nationally over the same period [33]. As therapeutic options expand and treatment guidelines evolve, the growing complexity of medication regimens is likely to increase both the need for and the actual use of MDD. However, whether the observed increase in drug counts amongst MDD users reflects more intensive treatment of underlying conditions, reduced flexibility in deprescribing within the MDD system, or a combination of both, cannot be determined from our data.

Several factors may have affected  the overall growth in MDD use over the study period. MDD is designed for patients with stable medication regimens, yet many older adults with multimorbidity require frequent dose adjustments, medication switches, or short-term treatments that are poorly suited to the fixed 2-week dispensing schedule [34]. In such cases, clinicians may prefer standard prescribing to retain flexibility, as MDD has been associated with fewer drug treatment changes compared with ordinary prescriptions [10]. Additionally, initiating MDD requires coordination between prescribers, pharmacies and care services, and its practical implementation is constrained by organisational and interprofessional challenges, which may act as a barrier to uptake, particularly for community-dwelling older adults without regular home help contact. Furthermore, growing evidence of poorer quality of drug treatment amongst MDD users, including higher rates of potentially inappropriate medications and drug–drug interactions [9, 35], may have contributed to more selective use of MDD by prescribers.

The lack of a spike during the COVID-19 pandemic suggests that MDD was not used as an emergency response to limit pharmacy visits during the pandemic and reflects established patterns of use. More broadly, demand for MDD is likely to increase in the future as a growing number of older adults with complex health problems continue to live at home, while the availability of nursing home care remains relatively stable despite population ageing [4, 10, 11, 36]. However, this trajectory is not certain. The same complexity that drives the need for medication support also makes MDD more challenging to implement, as patients with multimorbidity often require frequent medication changes that conflict with the fixed dispensing schedule. Whether MDD use will continue to grow therefore depends on how well the system adapts to serve patients with dynamic treatment needs, for example, better integration of medication reviews into the MDD workflow.

Taken together, these patterns suggest that medication review efforts within the MDD system should be particularly directed towards community-dwelling older adults with high medication burden, where MDD use is growing fastest and the risk of suboptimal prescribing may be greatest. Additionally, future research should quantify both the benefits and risks of MDD use. Potential benefits include reduction in medication errors and drug waste, and increase in medication adherence, while potential risks include reduced flexibility in dose adjustments, polypharmacy, delayed deprescribing and clinical inertia in drug reviews. Future studies should also describe commonly prescribed drugs amongst MDD users and compare rates of MDD use across countries.

This large nationwide study leverages the longitudinal nature of well-established national registers to provide 10 years of prevalence data on MDD use, overall and across subgroups, amongst all people aged > 70 years in Sweden.

Our study also has limitations. Our study may include misclassification bias due to the method used to define MDD use, dementia and multimorbidity. We may have missed MDD dispensations which are not repeated every 2 weeks and also misclassified dispensations as MDD in 2017 due to known recording issues in the Swedish Prescribed Drug Register [21], which likely explains the peak in prevalence that year. In addition, our operational definitions of MDD use, residence in nursing homes and receipt of home help services are imperfect proxies based on the available registers. We did not determine the exact temporal ordering of these events (e.g. whether MDD started before or after entry into nursing home care or home help services), which may introduce further misclassification. This is unlikely to substantially affect our conclusions, as our aim was to describe patterns rather than to draw causal inferences. We did not have information for diagnoses set in primary care (due to lack of nationwide primary care data in Sweden). Moreover, dementia is likely underdiagnosed in the older population, particularly amongst nursing home residents whose cognitive decline may never be formally evaluated [37]. However, we used the most recognised MDD prescribing schedule of 2 weeks [4], used a dementia diagnosis of ‘ever’ to ensure we captured all historical cases and pre-defined methods to define multimorbidity to make results as comparable as possible with existing literature. Moreover, because we analysed annual prevalence, we may have missed short-term fluctuations in MDD use around the onset of COVID-19. Additionally, the exclusion of individuals who died on or before 31 December of a given year from prevalence calculations may lead to underestimation of the true prevalence of MDD use. Finally, our results may only be generalisable in settings similar to Sweden.