Potentially Inappropriate Prescribing Among People with Dementia in Primary Care: A Retrospective Cross-Sectional Study Using the Enhanced Prescribing Database

Abstract

Background: Little is known about prescribing appropriateness for community-dwelling people with dementia (PWD).

Objective: To estimate potentially inappropriate prescribing (PIP) prevalence among PWD in primary care in Northern Ireland, and to investigate associations between PIP, polypharmacy, age, and gender.

Methods: A retrospective cross-sectional study was conducted, using data from the Enhanced Prescribing Database. Patients were eligible if a medicine indicated for dementia management was dispensed to them during 1 January 2013–31 December 2013. Polypharmacy was indicated by use of ≥4 repeat medications from different drug groups. A subset of the Screening Tool of Older Persons Potentially Inappropriate Prescriptions (STOPP) criteria, comprising 36 indicators, was applied to the dataset. Overall prevalence of PIP and the prevalence per each STOPP criterion was calculated as a proportion of all eligible persons in the dataset. Logistic regression was used to investigate associations between PIP, polypharmacy, age, and gender.

Results: The study population comprised 6826 patients. Polypharmacy was observed in 81.5% (n = 5564) of patients. PIP prevalence during the study period was 64.4% (95% CI 63.2– 65.5; n = 4393). The most common instance of PIP was the use of anticholinergic/antimuscarinic medications (25.2%; 95% CI 24.2–26.2; n = 1718). In multivariable analyses, both polypharmacy and gender (being female) were associated with PIP, with odds ratios of 7.6 (95% CI 6.6–8.7) and 1.3 (95% CI 1.2–1.4), respectively. No association was observed between PIP and age, after adjustments for gender and polypharmacy.

Conclusion: This study identified a high prevalence of PIP in community-dwelling PWD. Future interventions may need to focus on certain therapeutic categories and polypharmacy.

Keywords

Dementia inappropriate prescribing pharmacoepidemiology polypharmacy primary health care

INTRODUCTION

The aging population is growing rapidly worldwide, and it is a trend reflective of the major advancements in healthcare over the last century. Consequently, prescribing for older people, conventionally defined as those aged 65 years and over, is becoming an increasingly important aspect of clinical care, and one that requires prudent consideration from prescribers [1]. The presence, and subsequent management of, multiple morbidities in older patients will often result in polypharmacy [2], which has frequently been described as the concurrent use of four or more medications [3, 4]. Use of ten or more medications has been termed “excessive polypharmacy” [5]. While polypharmacy may be appropriate and therapeutically beneficial where a number of medications are clinically indicated (such as patients with complex or multiple conditions) [2], it is known to be a risk factor for adverse drug events (ADEs), drug-drug and drug-disease interactions, and potentially inappropriate prescribing (PIP) [3 , 6]. PIP refers to the use of medicines that introduce a greater risk of adverse drug-related events where a safer, and just as effective alternative is available to treat the same condition [6]. PIP is associated with increases in negative outcomes, such as morbidity, ADEs, hospitalizations, and mortality [7, 8], and is reported to be common amongst older people [9 –11]. A myriad of tools have been developed to identify inappropriate prescribing [2, 7]. The recently updated Screening Tool of Older Person’s Prescriptions (STOPP) criteria is a screening tool comprising 80 clinically significant criteria for PIP in older people, primarily organized by physiological system, validated by a Europe-wide Delphi consensus panel [12]. These evidence-based criteria take drug-drug and drug-disease interactions, drug doses, duration of treatment, and clinical effectiveness into consideration when determining the appropriateness of the prescribed treatments. Each criterion is accompanied by a concise, evidence-based explanation as to why the prescribing practice is potentially inappropriate. The STOPP criteria have been extensively validated for use in the UnitedKingdom (UK) setting [2].

Consideration of the appropriateness of prescribing for people with dementia (PWD) is particularly important due to the unique medication needs that this vulnerable population have in comparison to the rest of the older population. The presence of other comorbidities and complex medication regimens with possible psychoactive drug use, together with deficits in cognition and communication and diminishing decision-making capacity, generate challenges with medication management, particularly adherence [13]. Such issues may also influence doctors’ prescribing behavior and the quality of chronic illness management [14, 15]. For example, Wood-Mitchell et al. reported that psychiatrists in England felt under pressure to prescribe for PWD experiencing behavioral and psychological symptoms and did not always adopt an evidence-based approach to prescribing activity [13]. Whilst a number of studies have reported on appropriateness of prescribing for PWD, these tend to focus on dementia patients living in long-term care facilities [16, 17], those at the end of life [18, 19], or those prescribed antipsychotic medications [20, 21]. Less attention has been paid to PWD living in their own homes within the primary healthcare setting. Studies that have specifically investigated inappropriate medication use within this dementia patient population have been small in size and relied on patient or caregiver reports of drug use [22 –26].

An assessment of the appropriateness of prescribing for PWD, especially those managed within the primary healthcare setting, may help to identify a population likely to benefit from interventions to optimize prescribing practices. Therefore, the aim of this pharmacoepidemiological study was to estimate the prevalence of PIP among PWD in primary care in Northern Ireland (NI), by applying a subset of the STOPP criteria to a prescribing database. We also sought to explore the association between PIP and factors such as polypharmacy, age and gender, to more precisely characterize those with dementia who might be at risk of PIP.

MATERIALS AND METHODS

Setting

Northern Ireland is part of the UK, has a population of 1.7 million, and primary healthcare is delivered through 330 general practices. Healthcare in NI is provided under the UK’s National Health Service (NHS), where health and social care is publicly funded through central taxation and is free-of-charge at the point of need to all citizens. Unlike some other countries in the UK (namely England and Scotland), prescriptions (and therefore all medications) have been free in NI since prescription charges were phased out in 2010.

Data source

Data were extracted from the Enhanced Prescribing Database (EPD), which securely holds information on drugs prescribed and subsequently dispensed to patients in primary care in NI. The EPD does not contain data relating to prescribing in the hospital setting or over-the-counter (OTC) medication use. Once prescriptions have been dispensed by community pharmacies, they are forwarded to the Health and Social Care (HSC) Business Services Organisation (BSO) at the end of each month for reimbursement. Computer-generated prescriptions contain a unique two-dimensional barcode which is scanned by the BSO during the reimbursement process. This barcode links a patient’s Health and Social Care Number with details of their prescribed medication and prescriber. Once this information is scanned by the BSO, it is held in a secure database, the EPD. At present, approximately 85–90% of all prescriptions forwarded to the BSO result in data of research standard, which has helped to generate a central database of approximately 1.9 million patients in NI [9]. Diagnoses and other clinical information are not recorded in the EPD.

Study design and population

This was a retrospective, cross-sectional study using data from the EPD. Ethical approval was received from the NHS Research Ethics Committee London – City Road and Hampstead (14/LO/1891). Study participants were identified by a computerized search of the EPD, which was conducted by BSO data custodians. The study population comprised all individuals in the EPD who were dispensed a drug for the management of dementia (donepezil, galantamine, rivastigmine, memantine) during the study period 1 January 2013–31 December 2013. These drugs were used as proxy measures for diagnosis of dementia in the absence of clinical information about individuals. Patients in the EPD who entered a care home on or before 31 December 2013 were excluded, as were patients who left NI or died during the study period. In order to apply certain STOPP criteria, all patients were required to have at least three months of lead-in data prior to 1 January 2013, to ascertain long-term use of certain medications. All data were anonymized and the research team had no access to any patient identifiable data.

The final version of the dataset that was available to the research team included a unique patient identifier and information on patients’ age (in years), gender, the month and year in which a prescription was scanned by the BSO, and data on all items prescribed (such as the drug name, strength, quantity, and date of issue) during the study period.

Exposures

Thirty-eight of the 80 STOPP indicators were deemed suitable by the research team for application to the EPD dataset in the absence of clinical or diagnostic information. Some indicators could not be applied due to the absence of clinical or diagnostic data and were therefore excluded. For example, “aldosterone antagonists with concurrent potassium-conserving drugs without monitoring of serum potassium” could not be operationalized due to the absence of data on biochemical monitoring, and, therefore, was not included. For some criteria, prescription drugs for the treatment of certaindisease conditions were identified in the EPD dataset and used as proxies for diagnosis, where possible, such as for glaucoma and gout (Supplementary Table 1). This method has been used in other studies [8, 9]. During analysis, the following two STOPP indicators were unable to be operationalized due to lack of long-term prescribing data: “long-term use of NSAID for symptom relief of osteoarthritis where paracetamol has already been tried” and “long-term NSAID or colchicine for chronic treatment of gout where there is no contraindication to a xanthine-oxidase inhibitor”. Therefore, a total of 36 STOPP indicators were applied to the final dataset.

Data on drug use were extracted using British National Formulary (a standard drug reference text used in the UK) codes [1]. Patients were categorized into those who received a STOPP criteria drug or drug combination. STOPP criteria which specified a particular duration, such as “benzodiazepines for ≥4 weeks”, were assessed by identifying individuals who used the drugs for durations exceeding these appropriate thresholds within the study period (using the month a prescription was scanned by the BSO). STOPP criteria which specified a particular dosage not to be exceeded, such as “oral elemental iron doses greater than 200 mg daily”, were evaluated by calculating the number of daily defined doses (DDDs) for each recipient using the strength and quantity of the dispensed medication for each prescription.

The total number of prescriptions dispensed for each different drug group (according to BNF code) was calculated for each individual, during the one-year study period. A “repeat medication” was defined as one for which the patient received three or more prescriptions for that agent in the study period. Polypharmacy was examined by the use of four or more repeat medications from different drug groups.

Outcomes

The primary outcome was the overall prevalence of PIP in PWD in primary care in NI in 2013, according to a subset of the STOPP criteria. Secondary outcomes measures were: (i) the prevalence of PIP per individual STOPP criterion, and (ii) the association between PIP and polypharmacy, gender, and age group.

Statistical analysis

The overall prevalence of PIP in the study population and the prevalence per individual STOPP criterion in 2013 (the study period) were calculated as a proportion of all eligible persons in the dataset and reported as percentage estimate and 95% confidence intervals (CI). Adjusted logistic regression analyses were used to calculate odds ratios (OR) and 95% CI to investigate the association between any (versus no) PIP and polypharmacy (categorized as 0–3 versus ≥4 repeat drug classes), age group (≤44, 45–64, 65–84,≥85 years) and gender (male, female). There were no missing data for the variables of interest. Analyses were performed using STATA SE v13 (StataCorp, College Station, TX, USA).

RESULTS

Characteristics of the study population

For the study period, 6826 persons identified in the EPD were eligible for inclusion in the study (Table 1). Of these, approximately two-thirds were female (n = 4393, 64.4%), with a mean age of 79.6 (standard deviation (SD) ± 8.0) years. Patients were taking a mean number of 6.8 (SD ± 3.5) repeat medications. Over three-quarters of patients (n = 5564, 81.5%) were receiving four or more repeat medications (the definition of polypharmacy adopted for this study), whilst the use of ten or more repeat medications was observed in one-fifth of patients (n = 1427,20.9%).

Overall prevalence of PIP in 2013

The overall prevalence of PIP in the study period, according to the 36 STOPP indicators that were applied to the dataset, was 64.4% (95% CI 63.2–65.5) (n = 4393). Over one-fifth of the population (n = 1571, 23.0% (95% CI 22.0–24.0)) was prescribed one potentially inappropriate medication, 1141 patients (16.7% (95% CI 15.8–17.6)) were prescribed two potentially inappropriate medications, and 1681 patients (24.6% (95% CI 23.6–25.7)) were prescribed three potentially inappropriatemedications.

Prevalence of PIP in 2013 according to individual STOPP criteria

Table 2 describes the prevalence for each STOPP criterion. The most common instance of PIP was the use of anticholinergic/antimuscarinic medications (n = 1718, 25.2%). The second most frequently prescribed potentially inappropriate medicines were proton pump inhibitors (PPIs) at full therapeutic dosage for >8 weeks (n = 1561, 22.9%), followed by acetylcholinesterase inhibitors with concurrent treatment with drugs that reduce heart rate (n = 1276, 18.7%), benzodiazepines for ≥4 weeks (n = 777, 11.4%), and use of regular opioids without concomitant laxative (n = 715, 10.5%). Duplication of therapy within drug classes was most frequently observed with opioid analgesics (n = 346, 5.1%) and benzodiazepines (n = 239, 3.5%). Many other STOPP criteria had a prevalence less than 1.0%, such as “thiazide diuretic with a history of gout” and “phenothiazines as first-line treatment, since safer and more efficacious alternatives exist”.

Factors associated with PIP

Univariate logistic regression confirmed that polypharmacy, age, and gender were significantly associated with PIP (Table 3). A strong association between PIP and polypharmacy was observed. Those receiving four or more repeat medications were seven and a half times more likely to be exposed to PIP compared to those on zero to three repeat medications(adjusted OR 7.6, 95% CI 6.6–8.7). PIP was more likely to occur in females than in males after adjusting for age and polypharmacy (adjusted OR 1.3, 95% CI 1.2–1.4). No association was observed between PIP and age after adjustments for gender and polypharmacy.

DISCUSSION

Findings

Based on the data from a comprehensive dispensing database of 6826 dementia patients in NI, we found that both polypharmacy and PIP were prevalent among this community-dwelling patient population during 2013. PIP occurred in nearly two-thirds of the population (64.4%), according to the subset of STOPP criteria applied. The most commonly prescribed potentially inappropriate medicines were anticholinergic/antimuscarinic medications, followed by PPIs at maximum therapeutic dosage for >8 weeks, acetylcholinesterase inhibitors with concurrent treatment with drugs that reduce heart rate, and benzodiazepines for ≥4 weeks. Polypharmacy and gender were significantly associated with PIP. Age was not associated withPIP.

To our knowledge, this is one of the first studies to apply the STOPP criteria to a large prescribing database in order to ascertain the prevalence of PIP amongst community-dwelling dementia patients. Previous studies have reported a lower prevalence of potentially inappropriate medication use (between 15% and 47%) among community-dwelling dementia patients, as reported using either the Beers criteria or PRISCUS list (a tool developed for use in Germany) [22 –26]. The prevalence of PIP in our study was nearly double that reported by Bradley et al. who investigated PIP in older people (aged ≥70 years) in NI using the STOPP criteria, but whose methodology did not focus specifically on PWD [10].

In addition, we found that the prevalence of polypharmacy, as defined by the use of four or more repeat medications, was high amongst this patient population (81.5%). Again, this is difficult to directly compare with previous studies which have used different numeric thresholds to define polypharmacy in their study populations. However, this finding is much greater than that reported by Montastruc et al. [26] and Lau et al. [23] who reported polypharmacy (≥5 medications) in 43% and 52%, respectively, of community-dwelling PWD. A high prevalence of polypharmacy is unsurprising in PWD, as often this patient population will suffer from a number of comorbidities due to their increasing age and frailty [27]. Whilst patients in the current study population ranged in age from 34 to 100 years, they had a mean age of 79.6 years, and would therefore be expected to be receiving a number of different medications for comorbid conditions. There has been discussion within the literature about reducing reliance on numeric thresholds for polypharmacy and considering instead the appropriateness of polypharmacy, taking into account the fact that the use of “many drugs” may be necessary for those with multimorbidities [2, 28].

This study revealed a number of instances of PIP; some of these, such as the use of PPIs at full therapeutic dosage for >8 weeks and benzodiazepines for ≥4 weeks, are unsurprising and are consistent with findings reported in other studies exploring PIP amongst older people [10, 11] and PWD in care homes [17]. The prescribing of anticholinergic/antimuscarinic medications in our study population, received by one-quarter of patients (25.2%), was a concerning finding. The use of these drugs in PWD is not recommended due to their association with decline in both physical and cognitive function [29], and yet other studies have found similarly prevalent use of anticholinergics in dementia patients [24 , 30]. A number of tools have been developed to measure the anticholinergic drug burden, such as the validated Anticholinergic Cognitive Burden Scale [31]. The availability of such tools to clinicians could prove invaluable during an in-depth medication review with dementia patients, and may help them to change patients to alternative drugs with a lower anticholinergic burden. In some situations, non-pharmacological measures could be used as alternatives to prescribing anticholinergic medications, for example scheduling regular toilet breaks and making dietary modifications instead of using bladder antispasmodics [32].

Practice implications

In our study, the high prevalence of both polypharmacy and PIP could serve as an indicator that review of these patients is required to fully assess the appropriateness of the medication regimens used, particularly considering the strong relationship we observed between polypharmacy and PIP, which has been reported previously [9–11 , 26]. This study also revealed that PIP among community-dwelling dementia patients was associated with female gender, but not age. Again, these relationships have been reported elsewhere [9–11 , 25–26] and would be of assistance to clinicians identifying patients at risk of PIP. These associations may be useful in generating hypotheses which could be explored in other datasets. Consideration of PIP, polypharmacy, and gender could be incorporated into clinicians’ prescribing systems in order to alert them to such high-risk patients and potentially inappropriate medication combinations [33]. Medication review is just one component of medicines optimization, employing a patient-focused and person-centered approach which ensures that patients obtain the best possible outcomes from their medicines [34]. Often general practitioners (GPs) find it difficult to incorporate robust medication review into consultations due to time constraints; opportunity therefore exists for other healthcare professionals such as community pharmacists and nurses to assist with this and examples of such interventions in a primary care setting have been reported in the literature [35 –37]. With respect to pharmacists, the role of the GP practice-based pharmacist is expanding and a pilot scheme will be launched in the UK during 2016 [38]. These pharmacists will be ideally placed to assist with medication review of patients and will also be able to identify patients at high risk from PIP and potentially inappropriatemedications.

Deprescribing is another way in which inappropriate medication use and polypharmacy may be managed [39] and could prove to be a useful intervention in this particular patient population. For example, “drug holidays” (where medication is stopped for a trial period to assess effectiveness of treatment and/or remission of symptoms [40]) could be advocated for anticholinergic medications, such as those for urinary incontinence. Deprescribing is an emerging areawithin the scientific literature and it has been acknowledged that a wider evidence-base is needed to support such an approach [41 –44]. It has been reported that deprescribing may be particularly complicated in PWD due to their diminishing capacity and involvement in decision-making about their medicines, and difficulties with communication and understanding [45]. Reeve et al. have called for further research into the beliefs and preferences of dementia patients and their caretakers in order to better understand how deprescribing can be of optimal benefit to this patient population [45].

Strengths and limitations

This is one of the largest epidemiological studies to use a prescription-based database to estimate PIP amongst community-dwelling dementia patients. The EPD holds information on all prescriptions dispensed in community pharmacies in NI, and the high scan rate of prescriptions has generated a reliable database of great use to researchers. Although we have confidence in the generalizability of the results to the wider dementia patient population within NI, there are a number of methodological limitations which may limit generalizability of the findings to other settings. The lack of clinical information within the EPD, notably diagnostic data, means there could be an underestimation of the prevalence of patients with dementia. We had to identify patients who had received one of four drugs used in the management of dementia, using these medications as a proxy for a dementia diagnosis. These drugs are licensed for the treatment of mild to moderate dementia in Alzheimer’s disease (donepezil, galantamine), moderate to severe dementia in Alzheimer’s disease (memantine) or mild to moderate dementia in Parkinson’s disease (rivastigmine) [1]. Whilst this may have excluded those with dementia of different etiologies or those with severe cases in whom the medication had been stopped, we had no alternative means of identifying the patient population for inclusion in the study in the absence of diagnostic information. In addition, the lack of clinical data within the EPD only allowed us to apply a subset of the STOPP criteria and some diagnoses had to be determined using drug proxies, an analytical approach which has been used previously [9 , 46]. Therefore some instances of PIP identified within this study may not be clinically relevant, and clinicians must ensure that prescribing decisions are also based upon their clinical and personal knowledge of each patient. A set of explicit prescribing criteria for dementia is under development in Australia [47] and may be useful to researchers carrying out similar epidemiological studies in the future. The EPD was chosen for its relevance to the NI setting over other databases such as the Clinical Practice Research Datalink (CPRD), which is not representative of NI prescribing data [48]. Other limitations of using drug dispensing data is that patient adherence to medication is assumed. Use of over-the-counter (OTC) medications purchased without a prescription is not accounted for, which may under-estimate or over-estimate PIP prevalence and use of anticholinergic/antimuscarinic medications in particular, due to the anticholinergic effect of many OTC sleeping aids andantihistamines.

Despite these limitations, polypharmacy and PIP are prevalent among community-dwelling dementia populations; female patients and those receiving four or more medications may be at particular risk from inappropriate prescribing practices. This study has added to the limited body of epidemiological work undertaken with the community-dwelling dementia population as its focus and may assist clinicians to identify “at-risk” dementia patients in need of medication review within the primary care setting. Further pharmacological studies should be undertaken to validate the findings from the present study in other settings, such as the rest of the UK or Europe. Future work should also focus on exploring GPs’ prescribing behaviors for these patients to further understand the factors influencing prescribingdecisions.

Footnotes

ACKNOWLEDGMENTS

The authors wish to thank the staff at the HSC Business Services Organisation, Information and Registration Unit for supplying data from the Enhanced Prescribing Database and providing technical support throughout the study. This work was funded by the HSC Research & Development Division of the Public Health Agency in Northern Ireland and The Atlantic Philanthropies (Ref: COM/5020/14). The funders had no role in the design or conduct of the study; in the analysis and interpretation of the data; or in the preparation or approval of the manuscript.

Authors’ disclosures available online ().

Appendix

References

Joint Formulary Committee (2015) British National Formulary, 69th ed, BMJ Group and Pharmaceutical Press, London, UK.

Duerden

, Avery

, Payne

(2013) Polypharmacy and medicines optimisation: Making it safe and sound. The King’s Fund, London, UK.

Rollason

, Vogt

(2003) Reduction of polypharmacy in the elderly: A systematic review of the role of the pharmacist. Drugs Aging 20, 817–832.

Patterson

, Cadogan

, Kerse

, Cardwell

, Bradley

, Ryan

, Hughes

(2014) Interventions to improve the appropriate use of polypharmacy for older people. Cochrane Database Syst Rev 10, CD008165.

Jyrkka

, Enlund

, Korhonen

, Sulkava

, Hartikainen

(2009) Patterns of drug use and factors associated with polypharmacy and excessive polypharmacy in elderly persons. Results of the Kuopio 75+ study: A cross-sectional analysis. Drugs Aging 26, 493–503.

O’Mahony

, Gallagher

(2008) Inappropriate prescribing in the older population: Need for new criteria. Age Ageing 37, 138–141.

Spinewine

, Schmader

, Barber

, Hughes

, Lapane

, Swine

, Hanlon

(2007) Appropriate prescribing in elderly people: How well can it be measured and optimised?. Lancet 370, 173–184.

Hamilton

, Gallagher

, Ryan

, Byrne

, O’Mahony

(2011) Potentially inappropriate medications defined by STOPP criteria and the risk of adverse drug events in older hospitalised patients. Arch Int Med 171, 1013–1019.

Cahir

, Fahey

, Teeling

, Teljeur

, Feely

, Bennett

(2010) Potentially inappropriate prescribing and cost outcomes for older people: A national population study. Br J Clin Pharmacol 69, 543–552.

10.

Bradley

, Fahey

, Cahir

, Bennett

, O’Reilly

, Parsons

, Hughes

(2012) Potentially inappropriate prescribing and cost outcomes for older people: A cross-sectional study using the Northern Ireland Enhanced Prescribing Database. Eur J Clin Pharmacol 68, 1425–1433.

11.

Bradley

, Motterlini

, Padmanabhan

, Cahir

, Williams

, Fahey

, Hughes

(2014) Potentially inappropriate prescribing among older people in the United Kingdom. BMC Geriatr 14, 72.

12.

O’Mahony

, O’Sullivan

, Byrne

, O’Connor

, Ryan

, Gallagher

(2015) STOPP/START criteria for potentially inappropriate prescribing in older people: Version 2. Age Ageing 44, 213–218.

13.

Maidment

, Fox

, Boustani

, Katona

(2012) Medication management – the missing link in dementia interventions. Int J Geriatr Psychiatr 27, 439–442.

14.

Wood-Mitchell

, James

, Waterworth

, Swann

, Ballard

(2008) Factors influencing the prescribing of medications by old age psychiatrists for behavioural and psychological symptoms of dementia: A qualitative study. Age Ageing 37, 547–552 .

15.

Imfeld

, Brauchli Pernus

, Jick

, Meier

(2013) Epidemiology, co-morbidities and medication use of patients with Alzheimer’s disease or vascular dementia in the UK. J Alzheimers Dis 35, 565–573.

16.

Zuckerman

, Hernandez

, Gruber-Baldini

, Hebel

, Stuart

, Zimmerman

, Magaziner

(2005) Potentially inappropriate prescribing before and after nursing home admission among patients with and without dementia. Am J Geriatr Pharmacother 3, 246–254.

17.

Parsons

, Johnston

, Mathie

, Baron

, Machen

, Amador

, Goodman

(2012) Potentially inappropriate prescribing in older people with dementia in care homes. Drugs Aging 29, 143–155.

18.

Parsons

, Hughes

, Passmore

, Lapane

(2010) Withholding, discontinuing, and withdrawing medications in dementia patients at the end of life: A neglected problem in the disadvantaged dying?. Drugs Aging 27, 435–449.

19.

Parsons

, McCorry

, Murphy

, Byrne

, O’Sullivan

, O’Mahony

, Passmore

, Patterson

, Hughes

(2014) Assessment of factors that influence physician decision making regarding medication use in patients with dementia at the end of life. Int J Geriatr Psychiatr 29, 281–290.

20.

Guthrie

, Clark

, McCowan

(2010) The burden of antipsychotic drug prescribing in people with dementia: A population database study. Age Ageing 39, 637–642.

21.

Monette

, Monette

, Sourial

, Vandal

, Wolfson

, Champoux

, Fletcher

, Savoie

(2013) Effect of an interdisciplinary educational program on antipsychotic prescribing among residents with dementia in two long-term care centers. J Appl Gerontol 32, 833–854.

22.

Fialová

, Topinková

, Gambassi

, Finne-Soveri

, Jónsson

, Carpenter

, Schroll

, Onder

, Sørbye

, Wagner

, Reissigová

, Bernabei

, AdHOC Project Research Group (2005) Potentially inappropriate medication use among elderly home care patients in Europe. JAMA 293, 1348–1358.

23.

Lau

, Mercaldo

, Harris

, Trittschuh

, Shega

, Weintraub

(2010) Polypharmacy and potentially inappropriate medication use among community-dwelling elders with dementia. Alzheimer Dis Assoc Disord 24, 56–63.

24.

Koyama

, Steinman

, Ensrud

, Hillier

, Yaffe

(2013) Ten-year trajectory of potentially inappropriate medications in very old women: Importance of cognitive status. J Am Geriatr Soc 61, 258–263.

25.

Fiss

, Thyrian

, Fendrick

, van den Berg

, Hoffman

(2013) Cognitive impairment in primary ambulatory health care: Pharmacotherapy and the use of potentially inappropriate medicine. Int J Geriatr Psychiatry 28, 173–181.

26.

Montastruc

, Gardette

, Cantet

, Piau

, Lapeyre-Mestre

, Vellas

, Montastruc

, Andrieu

, REAL.FR Group (2013) Potentially inappropriate medication use among patients with Alzheimer disease in the REAL.FR cohort: Be aware of atropinic and benzodiazepine drugs. Eur J Clin Pharmacol 69, 1589–1597.

27.

Formiga

, Fort

, Robles

, Riu

, Sabartes

, Barranco

, Catena

(2009) Comorbidity and clinical features in elderly patients with dementia: Differences according to dementia severity. J Nutr Health Ageing 13, 423–427.

28.

Hughes

, Cooper

, Ryan

(2013) Going beyond the numbers – a call to redefine polypharmacy. Br J Clin Pharmacol 77, 915–916.

29.

Fox

, Smith

, Maidment

, Chan

, Bua

, Myint

, Boustani

, Kwok

, Blover

, Koopmans

, Campbell

(2014) Effect of medications with anti-cholinergic properties on cognitive function, delirium, physical function and mortality: A systematic review. Age Ageing 43, 604–615.

30.

Sura

, Carnahan

, Chen

, Aparasu

(2013) Prevalence and determinants of anticholinergic medication use in elderly dementia patients. Drugs Aging 30, 837–844.

31.

Salahudeen

, Buffull

, Nishtala

(2015) Anticholinergic burden quantified by anticholinergic risk scales and adverse outcomes in older people: A systematic review. BMC Geriatr 15, 31.

32.

Specht

(2011) Promoting continence in individuals with dementia. J Gerontol Nurs 37, 17–21.

33.

Clyne

, Smith

, Hughes

, Boland

, Bradley

, Cooper

, Fahey

, OPTI-SCRIPT study team (2015) Effectiveness of a multifaceted intervention for potentially inappropriate prescribing in older patients in primary care: A cluster-randomized trial (OPTI-SCRIPT study). Ann Fam Med 13, 545–553.

34.

National Institute for Health and Care Excellence (2015) Medicines optimisation: The safe and effective use of medicines to enable the best possible outcomes, https://www.nice.org.uk/guidance/ng5/resources/medicines-optimisation-the-safe-and-effective-use-of-medicines-to-enable-the-best-possible-outcomes-51041805253, Accessed 03 December 2015.

35.

Zermansky

, Petty

, Raynor

, Freemantle

, Vail

, Lowe

(2001) Randomised controlled trial of clinical medication review by a pharmacist of elderly patients receiving repeat prescriptions in general practice. BMJ 323, 1340.

36.

Lenaghan

, Holland

, Brooks

(2007) Home-based medication review in a high risk elderly population in primary care – the POLYMED randomised controlled trial. AgeAgeing 36, 292–297.

37.

Milos

, Rekman

, Bondesson

, Eriksson

, Jakobsson

, Westerlund

, Midlöv

(2013) Improving the quality of pharmacotherapy in elderly primary care patients through medication reviews: A randomised controlled study. Drugs Aging 30, 235–246.

38.

NHS England, Health Education England, Royal College of General Practitioners, British Medical Association (2015) Clinical pharmacists in general practice pilot, https://www.england.nhs.uk/commissioning/wp-content/uploads/sites/12/2015/07/clinical-pharmacists-gp-pilot.pdf, Accessed 03 December 2015.

39.

Scott

, Hilmer

, Reeve

, Potter

, Le Couteur

, Rigby

, Gnjidic

, Del Mar

, Roughead

, Page

, Jansen

, Martin

(2015) Reducing inappropriate polypharmacy: The process of deprescribing. JAMA Intern Med 175, 827–834.

40.

Howland

(2009) Medication holidays. J Psychosoc Nurs Ment Health Serv 47, 15–18.

41.

Garfinkel

, Zur-Gil

, Ben-Israel

(2007) The war against polypharmacy: A new cost-effective geriatric-palliative approach for improving drug therapy in disabled elderly people. Isr Med Assoc 9, 430–434.

42.

Garfinkel

, Mangin

(2010) Feasibility study of a systematic approach for discontinuation of multiple medications in older adults: Addressing polypharmacy. Arch Int Med 170, 1648–1654.

43.

Scott

, Gray

, Martin

, Pillans

, Mitchell

(2013) Deciding when to stop: Towards evidence-based prescribing of drugs in older populations. Evid Based Med 18, 121–124.

44.

Reeve

, Shakib

, Hendrix

, Roberts

, Wiese

(2014) Review of deprescribing processes and developmentof an evidence-based, patient-centred deprescribing process. Br J Clin Pharmacol 78, 738–747.

45.

Reeve

, Bell

, Hilmer

(2015) Barriers to optimising prescribing and deprescribing in older adults with dementia: A narrative review. Curr Clin Pharmacol 10, 168–177.

46.

Cooper

, Moriarity

, Ryan

, Smith

, Bennett

, Fahey

, Wallace

, Cahir

, Williams

, Teeling

, Hughes

(2016) Potentially inappropriate prescribing in two populations with differing socio-economic profiles: A cross-sectional database using the PROMPT criteria. Eur J Clin Pharmacol. doi: 10.1007/s00228-015-2003-z

47.

Page

, Potter

, Clifford

, McLachlan

, Etherton-Beer

(2015) Prescribing for Australians living with dementia: Study protocol using the Delphi technique. BMJ Open 5, e008048.

48.

Herrett

, Gallagher

, Bhaskaran

, Forbes

, Mathur

, van Staa

, Smeeth

(2015) Data resource profile: Clinical Practice Research Datalink (CPRD). Int J Epidemiol 44, 827–836.