Minimising potentially inappropriate polypharmacy in community living older people: a multi-phase, mixed methods study to develop and pilot a general practitioner-led deprescribing intervention in primary care

BackgroundPolypharmacy-related harm affects many older Australians.  Clinician supervised withdrawal or dose reduction of potentially inappropriate medicines (PIMs), defined as deprescribing, aims to reduce this harm. General practitioners (GPs), with the support of consultant pharmacists (CPs), are well positioned to initiate deprescribing in community living older people with potentially inappropriate polypharmacy.  Effecting prescribing change is difficult however, and there is little research into GP-led deprescribing interventions in this patient group which evaluates clinicians’ barriers and enablers to change.  Research aims To: 1) investigate factors which shape prescribers’ behaviour towards continuing or discontinuing PIMs in adults; 2) explore the views of GPs and CPs about potentially inappropriate polypharmacy and the reasoning they apply to deprescribing in community living older people; and 3) investigate the feasibility, effectiveness and safety of a GP-led deprescribing intervention involving such people in primary care. MethodologyA sequential, exploratory mixed methods design with three phases which aligned to three research aims was used.  Phase One involved a systematic review and thematic synthesis of studies exploring prescribers’ perceived barriers and enablers to minimising chronically-prescribed PIMs in adults. Phase Two, a qualitative investigation, comprised seven focus group discussions involving 32 GPs and 15 CPs recruited from within metropolitan Southeast Queensland using a mix of convenience and snowball sampling. Data were analysed thematically using the Framework method. Phase One and Two findings informed the design of the multifaceted GP-led deprescribing intervention piloted in Phase Three.  A pragmatic, controlled pre-post mixed methods design was used to evaluate intervention feasibility, effectiveness and safety.  Convenience sampling was used to recruit five general practices and 22 clinicians.  Seventy-eight intervention and 67 usual care patients were consecutively sampled. Quantitative and qualitative data were collected at both clinician and patient levels. The primary outcome was the mean difference in the number of regular medications deprescribed (i.e. ceased or dose-reduced) per patient over the 18-week study period.  Secondary outcomes comprised intervention impact on patients’ medication regimens, self-reported health status, attitudes towards medicines and deprescribing, and GP/patient reports of actual or suspected adverse outcomes or experiences.  Qualitative data from semi-structured interviews were used to help explain quantitative results and the intervention’s feasibility, acceptability and sustainability.  Results                                                                  The systematic review addressing the first research aim of exploring prescribers’ perceived barriers and enablers to minimising PIMs comprised studies mostly exploring primary care physicians’ perspectives on managing older, community living adults.  Four major themes emerged: problem awareness; perceived value of ceasing versus continuing PIMs; self-efficacy regarding clinicians’ ability to alter prescribing; and feasibility of altering prescribing in routine care environments given external constraints. Two major themes were derived in response to the second thesis aim regarding GPs’ and CPs’ views and reasoning about potentially inappropriate polypharmacy: 1) Working through uncertainty encapsulated the immense complexity clinicians face when assessing an older person with potentially inappropriate polypharmacy, such that weighing harm against benefit in absolute terms at the level of the individual was perceived as unfeasible.  However, strategies and circumstances were identified that could mitigate this uncertainty; 2) Perceived risk as a frame of reference referred to the dichotomised view that deprescribing was a risk to be avoided or risk to be reconciled, with tipping points in risk perception identified which might trigger action towards deprescribing.The exploratory pilot study addressed the third aim of investigating the feasibility, effectiveness and safety of the multifaceted intervention in community living older people in primary care.  The mean difference between intervention and usual care groups in the number of regular medications deprescribed per patient was -0.55, 95%CI -0.897 to -0.212, p = 0.002.  The respective proportions of patients having at least one medication deprescribed were 52.6% versus 28.4%, p = 0.005, such that intervention patients were 2.3 times more likely to have at least one regular medication deprescribed (incidence rate ratio [IRR] 2.3; 95% CI 1.297-3.964, p = 0.004).  The intervention was not associated with any reported harm or deterioration in health-related quality of life which may have resulted from injudicious attempts at deprescribing appropriate medication.  A subset of intervention patients reported greater certainty in the necessity and appropriateness of their medications at study completion.  Qualitative data showed the majority of GPs and patients derived satisfaction from the process of medication review and consultation during dedicated deprescribing appointment/s, irrespective of whether the outcome of successful medication withdrawal was achieved.  Whilst seen as feasible in the short-term, GPs gave mixed responses regarding the sustainability of the intervention in routine care. Conclusions and recommendationsThis multi-phase, mixed methods study demonstrated immense diversity in clinicians’ perceived barriers and enablers to deprescribing in community living older people.  The multifaceted intervention addressing local barriers was feasible and safe in the short-term and conferred a clinically modest deprescribing effect.   Further research into the long-term effectiveness and safety of deprescribing interventions targeting community living older people at high-risk of medication misadventure is warranted.