Discussion

Main findings

In this study, the medication review intervention with focus on deprescribing was feasible as part of the developed Chronic Care Model in real-life primary care. The intervention led to 255 medication changes, of which more than 80% were deprescribing. The medication changes were maintained during the 3-4 months follow-up period for nearly all patients. At follow-up, we found that patients’ self-reported health status had increased, while general condition and functional level remained stable. Generally, similar trends were observed in the sub-group analyses for both medication-related and health-related measures.

Comparison with existing literature

In recent years, several systematic reviews have synthesised the evidence on the effectiveness of deprescribing interventions. These reviews have focused on older people in general18 or in different settings such as hospitals19, nursing homes20,21, or primary care22,23. Overall, the existing evidence suggests that deprescribing is feasible, safe, and ,generally, effective in reducing the number of inappropriate prescriptions22,24. A systematic review of deprescribing trials in primary care showed that the proportion of patients who successfully stopped their medication varied from 20% to 100%22. In 19 of the 27 included studies, more than half of the participants had successfully stopped medications. In our study, the majority of the medication changes were deprescribing. Comparable to our findings, studies of deprescribing trials have shown average discontinuations per patient between 2.8 and 4.425–28. Additionally, we found that medication changes were maintained for nine of ten patients at 3-4 months follow-up indicating successful deprescribing.
It is well-known that deprescribing can also lead to patient-harm in terms of adverse drug withdrawal events or return of symptoms (e.g., increased pain levels or mood changes), for which the medication was originally prescribed. Importantly, the majority of these harms can be minimized or even prevented by using a patient-centred deprescribing process with planning, tapering, and close monitoring during and after medication withdrawal29. This was possible in our study where a patient-centred deprescribing process was undertaken as part of routine chronic care management in general practice in close collaboration with the Centre for Health and Care in the Municipality of Frederikshavn. We found that the intervention led to an increase in self-reported health status from baseline to follow-up. Additionally, general condition and functional level remained stable. In the subgroup analyses, similar trends were seen in health-related outcomes among care home residents and community-dwelling patient.
However, an interesting finding was the considerable, non-significant increase in general condition among care home residents. Although non-significant results should be interpreted with caution, this signals that it may be possible to improve general condition through medication reviews with focus on deprescribing in this vulnerable patient group. Oppositely, no signal of change in general condition was observed among community-dwelling patients, which may partly be explained by the high proportion (80%) rated as “average or above” at baseline, which left limited room for improvement.
Few studies have been able to demonstrate an effect of medication review interventions on health-related outcomes of importance to patients. A recent example is the DREAMeR study, in which community-dwelling older persons with polypharmacy were offered patient-centred medication reviews versus usual care30. This study showed improved quality of life measured by the EQ-Visual Analogue Scale and reduced health problems with a moderate to severe impact on daily life. However, no effect was seen on quality of life measured by the EQ-5D-5L or on total number of health problems. This highlights the complexity of measuring improvement in the wellbeing of older and multimorbid patients.
In a recent review by Ibrahim et al., the current evidence for deprescribing among older people living with frailty was reported24. Of six included studies, three reported a positive impact on clinical outcomes such as depression, mental health status, function, and frailty. However, results were mixed on falls and cognition, and no significant impact was demonstrated on quality of life24. The latter echoes previous findings across a range of studies conducted in primary care using various quality-of-life measures10,31–33. These mixed results call for consideration regarding whether we are using the right measures to capture potential benefits of interventions at a patient-level. Moreover, they call for consideration regarding whether a lack of statistically significant improvements in health-related outcomes should be viewed more positively, as deprescribing without deterioration of patient health may also be a desirable outcome.

Primary care as a setting for deprescribing

In many countries, GPs are responsible for chronic care management in primary care and the relational and managerial continuity in this setting provide an optimal basis for deprescribing34. In this real-life quality improvement project, the GPs in the GPF decided to construct a new Chronic Care Model, including the person-centred medication review intervention, to systematize the care of patients with chronic diseases. It has been advocated to integrate clinical practice guidelines more systematically into existing care models to minimise the burden on health systems and primary care providers35. Thus, the developed Chronic Care Model employed in this study may have been an important enabler for intervention implementation.
It was originally planned that a pharmacist employed by the municipality would perform an initial medication review and present the findings for the GP, who would then implement clinically relevant medication changes. However, it soon became clear that the GPs performed the medication reviews themselves and took ownership of the process in close collaboration with the pharmacist, the nurses, and frontline staff at the care homes. Ownership, flexibility, and autonomy of the primary care providers have been identified as important enablers for implementation of clinical practice guidelines35. Additional enablers reported include a well-organised practice and clarification about the role of primary care providers in disease management. Importantly, multidisciplinary collaborations between different care levels should also be considered to support the primary care providers’ recognition of their role and responsibility for clinical practice guidelines implementation35. In our study, this was attempted through the cross-sectoral communication model that was established alongside the Chronic Care Model.
The approach taken in our study might be inspirational to other Danish municipalities as well as other countries with a similar organization of primary care. However, our results might not be directly transferable, as primary care medication management constitutes a complex health care system. It encompasses different types of healthcare organization (e.g., home care, care homes, general practices) and health care providers (e.g., nurses, pharmacists, GPs)36. Furthermore, both private and public stakeholders exist in most countries and may be highly dissimilar in their organization and available resources. Thus, the specific context, in which the intervention is to be implemented, should be fully considered, as adaptions may be needed to achieve success and sustainability35.

Strengths and limitations

A major strength of this study was the real-world primary care setting, in which the study was conducted. The recruitment and retention of elderly patients in clinical trials provide many challenges37. Thus, in contrast to the highly selected patient groups often included in randomized controlled trials, our study population more likely represent an unselected, real-world patient population, which strengthen the generalizability of our results. Further, it represents real-world implementation of a complex intervention, which suggests that our intervention is feasible and realistic in similar contexts. Even though the study was conducted during the COVID-19 pandemic and the associated restrictions, it was possible to implement the Chronic Care Model and include both care home residents and home-dwelling patients in the intervention.
A major limitation of the study is that no control group was included to compare results against usual care in similar GP clinics. Thus, no causal links can be made between the intervention and our results. Another limitation was the follow-up period of 3-4 months. As medical conditions in older patients are unstable37, more medication changes, incl. potential restarts as well as additional deprescribing, would have been captured if we had used a longer follow-up period. However, we expect that most potential harms of the implemented medication changes would have been manifested during the 3-4 months period. Four care home residents died before follow-up. However, by clinical evaluation it was concluded that none of these deaths were directly related to the intervention. Furthermore, the incidence of deaths was not higher than expected in care home residents in general 38.

Conclusion

In this real-world quality improvement study settled in primary care, we found that a systematic GP-led medication review intervention led to deprescribing and increased self-reported health status without deterioration of general condition or functional level among care home residents and community-dwelling patients with chronic disease. These results add new aspect to the existing literature and show that it may be possible to improve patients´ self-perceived health status through medication review interventions with a focus on deprescribing.