Abstract

Background.

Chronic diseases affect more than half of the population ≥75 years of age in developed countries. Prescription medication use increases with age. Depending on definition, 25–80% of elderly are exposed to polypharmacy. Polypharmacy increases the risk of hospitalization, interactions and adverse drug reactions.

Objective.

To examine the frequency of medication errors in patients with polypharmacy treated in general practice.

Methods.

The medications of 169 patients with polypharmacy treated in 22 GP surgeries in Austria were analysed. The analysis identified (i) medication errors, including non-evidence-based medications, dosing errors and potentially dangerous interactions in all patients and (ii) potentially inappropriate medications (PIMs) in the subgroup of elderly patients (≥65 years).

Results.

The patients took on average 9.1±3.0 medications per day. The maximum, in one patient, was 20 medications per day. Some 93.5% had at least one non-evidence-based medication. On average, 2.7±1.66 medications per patient were found to be not indicated. At least one dosing error was found in 56.2% of all patients. One potential interaction of the most severe degree (category X interaction) was detected in 1.8% (n = 3) and two such interactions in 0.6% (n = 1). These combinations should have been avoided. Of the 169 patients, 158 were elderly (≥65 years). Of these seniors, 37.3% (n = 59) had at least one PIM according to the PRISCUS list for the elderly.

Conclusion.

The frequency of medication errors is high in patients with polypharmacy in primary care. Development of strategies (e.g. external medication review) is required to counteract medication errors.

Introduction

Multiple chronic diseases affect more than half of the population ≥75 years of age in developed countries.1 Prescription medication use increases with age, because, amongst others, guideline adherence is endorsed by medical policymakers. Depending on definition, 25–80% of elderly are exposed to polypharmacy. On average, elderly take 7.5±3.8 medications daily. Polypharmacy increases the risk of hospitalization, interactions and adverse drug reactions (ADRs).2–4

There is no common definition of polypharmacy. In this study, we followed the suggestion of Junius-Walker et al.3 and defined it as the daily intake of five or more medications. Polypharmacy has been increasing in recent years. The increase was most distinct in the elderly.5 However, an increase was seen in all age groups, including children6 and adults with specific diagnoses.7

A study of 65 GPs and their patients showed both ‘believe’ in their medications and such ‘belief’ constitutes a psychological barrier for discontinuation.8 The GPs acknowledged that due to a lack of time, they do not regularly review their patients’ medications. The problem is present not only in family practices but also in hospitals. Schuler et al.9 examined 543 elderly inpatients and found that polypharmacy, inappropriate prescribing and ADRs are highly prevalent.

Strategies to reduce polypharmacy have been proposed, including the Beers Criteria,10 the PRISCUS list for the elderly,11 the Good Practice Geriatric Palliative Algorithm12 and the STOPP Criteria.13 However, few studies investigated whether polypharmacy can be reduced in daily clinical practice. In a non-randomized study in geriatric nursing departments, Garfinkel et al.14 discontinued on average 2.8 medications in 119 patients (intervention group) and matched them to 71 comparable patients. The 1-year mortality rate was 21% in the intervention group, compared with 45% in the matched controls. In a randomized controlled study in an ambulatory health care centre, Williams et al.15 were only able to discontinue an average of 1.5 medicines, which did not lead to a change in patient outcome.

We, therefore, investigated in this descriptive study the frequency of medication errors, including non-evidence-based medications, dosing errors and potential interactions in patients with polypharmacy treated in family practices, as well as potentially inappropriate medications (PIMs) in the subgroup of elderly patients. Our goal hereby is to quantify the potential of reducing polypharmacy in the primary care setting in a future prospective intervention study.

Methods

Study design, population and setting

After approval of the study by the ethics committee, the authors asked 22 personally known GPs to participate. The GPs were approached at two continuing medical education activities. All 22 GPs asked were willing to recruit patients for the study. We chose this convenience sample with a 100% participation rate to avoid the bias of selective participation in a random sample. The GPs were asked to invite 10 consecutive patients taking 5 or more medicines to take part in the study. Patients with cancer and elderly patients were not excluded, in order to better reflect problems GPs may encounter while treating patients with polypharmacy. All participating GPs and patients received an information pack and signed a consent form according to the Declaration of Helsinki.

Data collection and analysis

Each GP filled out a case report form (CRF) for each patient. Each CRF included two sections.

The information section consisted of GP ID number, patient ID number (anonymized), sex and age of the patient (in years, or as date of birth), weight (kg), height (cm) and serum creatinine (mg/dL).

Six GPs personally known to the institute additionally provided their names, addresses and telephone numbers to be able to receive advice about the medication of their patients and to give feedback via an unstructured interview regarding the advice given in the study.

In the medication section, a maximum of 20 medications could be listed per patient. GPs were instructed to submit an extra form for patients taking more than 20 medications. For each medication, the following had to be specified: (i) name of the medicine (generic or trade name). (ii) Whether the medicine was prescribed by the GP or a specialist, or was an over-the-counter self-medication (OTC). (iii) Total daily dose in milligrams (total daily dose had to be indicated in detail if not in milligrams). (iv) Indication for use. (v) Any additional information the GPs wished to specify, e.g. ‘in the morning’.

All data were analysed at Paracelsus Medical University. At first, we checked the evidence base and indication of each medication, then we analysed dosing and interactions. At last, we checked for appropriateness of the medication for elderly patients.

Each analysis was performed thrice, independently by a medical doctor, a pharmacist and a senior researcher of the institute. Any discrepancies in the analyses were noted and resolved by consensus, and if necessary by arbitration of the senior researcher.

Analysis of non-evidence-based medications and dosing errors

Evidence base and indications for use were analysed using UpToDate®.16 A medication was classified as non-evidence based if the indication for use indicated by the GP was not mentioned in any peer-reviewed chapter of UpToDate®.

Doses were analysed also with UpToDate®, taking age, sex, height, weight, creatinine, body mass index and glomerular filtration rate (GFR) into account. GFR was calculated with the CKD-EPI creatinine equation.17 The doses were analysed independently of the indications for use.

Dosing information was classified as (A) ‘Correct dose’. (B) ‘Missing data’. (C) ‘Overdose’. (D) ‘Underdose’. (E) ‘Otherwise wrong’, entered, e.g., when a diuretic was given at night. (F) ‘Two or more medications from the same drug class’. (G) ‘As needed’, entered when the GPs indicated this explicitly. C, D and E were classified as dosing error. In case of F and G, a case-by-case decision was made whether to classify the dose as a dosing error or not (in case of lercanidipine, a calcium-channel blocker not approved in North America and hence not discussed in UpToDate®, indication and dose were analysed according to the Summary of Product Characteristics in Austria).

Analysis of interactions

Potential interactions were identified using the Lexi-Interact® database.18 All medications were entered for every patient.

The database Lexi-Interact® divides potential interactions into five categories. Only the two ‘dangerous’ categories were taken into account, category D (consider therapy modification) and category X (avoid combination) (in case of lercanidipine, amlodipine, also a dihydropyridine class calcium-channel blocker, was entered into the database).

Analysis of PIMs in the elderly

PIMs in seniors (≥65 years) were identified using the PRISCUS list for the elderly.11 Although this list does not cover every drug class, it has been evaluated by the consortium that created it to be useful in the elderly.19

The PRISCUS List was created by a German consortium. At the time of this study, this was the only PIM list originating from a German-speaking country and deemed appropriate. An Austrian counterpart was subsequently published and does not differ significantly from the German list.20

Unstructured interviews

The six GPs who received the results of the analysis gave their feedback after the study in unstructured interviews about the usefulness of the advice. The interview comprised the following questions: ‘You received advice from the Institute of General Practice regarding the evidence base of the medication of your patients. Did you act upon this information? Was it helpful for the treatment of your patients?’

The objective of the interview was to estimate whether GPs would react upon the advice given in the planned intervention study.

Results

The mean age of the 169 patients included in this study was 76.4±8.5 (SD) years. All patients took five or more medications per day. The patients were recruited in 22 GP surgeries. Six of these 22 GPs received advice and gave feedback.

Descriptive data of all patients are shown in Table 1.

Table 1

Descriptive data of all patients

 Age Sex Weight (kg) Height (cm) Creatinine (mg/dL) 
Mean 76.4 68 male (34.7%), 100 female (59.2%) 78.9 166.4 1.06 
SD 8.5 – 15.8 8.9 0.38 
Maximum 96 – 125 190 3.00 
Minimum 47 – 38 146 0.50 
Missing value or illegible – 3  3  3  
 Age Sex Weight (kg) Height (cm) Creatinine (mg/dL) 
Mean 76.4 68 male (34.7%), 100 female (59.2%) 78.9 166.4 1.06 
SD 8.5 – 15.8 8.9 0.38 
Maximum 96 – 125 190 3.00 
Minimum 47 – 38 146 0.50 
Missing value or illegible – 3  3  3  

Polypharmacy

The patients took on average 9.1±3.0 (SD) medications daily. The maximum, in one patient, was 20 medications per day.

The 10 commonest drug classes were beta-blockers, ACE inhibitors, PPIs/H2 antagonists, statins, antiplatelet drugs, thiazide diuretics, vitamins and enzymes, calcium-channel blockers, loop diuretics and non-steroidal anti-inflammatory drugs.

Psychoactive drugs also were very common: 97 of the 169 patients (57.4%) took at least one psychoactive drug, and 39 (23.1%) took at least two such drugs. The maximum was five psychoactive drugs in one patient.

OTC drugs: Only 38 [22 female, 16 male, mean age 74.7±10.0 (SD) years] of the 169 patients (22.5%) provided information about self-medication. Eighteen of these (47.4%) reported to take an OTC medication [vitamins and minerals (six patients), gingko (two patients), aloe vera, arginine, diclofenac, a dietary fibre supplement, ginseng, a homeopathic remedy, nose drops, paracetamol, passiflora and valerian (1 patient each)]. In addition, 6 of these 38 patients (15.8%) also reported the use of a prescription medicine without a prescription from the GP: amlodipine, furosemide, gentamycin eye drops, nitroglycerin spray, tramadol and vardenafil. We did not inquire how the patients obtained these prescription drugs without a prescription. Fourteen patients (36.8%) explicitly stated that they did not use any self-medication.

Non-evidence-based medications

One hundred and fifty-eight of the 169 patients (93.5%) had at least one non-evidence-based medication using the criteria mentioned above. On average, 2.7±1.66 non-evidence-based medications were found per patient. The relative maximum of non-evidence-based medications was 83.3% in one patient (five of six medications were non-evidence based). The absolute maximum of non-evidence-based medications was 8 medications in 1 patient (8 of 13 medications were non-evidence based). There was no patient in whom all medications were classified as non-evidence based. The percentage of non-evidence-based medication for all drug classes used by more than 20% of the patients are provided in Table 2 (details for all drug classes are available from the author upon request).

Table 2

Percentage of the most common non-evidence-based medications in all patients

Drug classa Percentage of patients taking a medication from this drug class Percentage of non-evidence-based medications within this drug class 
Beta-blockers, e.g. metoprolol 55.0 54.8 
ACE inhibitors, e.g. lisinopril 52.1 2.3 
PPIs/H2 antagonists, e.g. pantoprazole and famotidine 50.3 50.6 
Statins, e.g. simvastatin 46.2 64.1 
Antiplatelet agents, e.g. clopidogrel 45.6 28.6 
Thiazide diuretics, e.g. HCT 42.6 9.7 
Vitamins and enzymes, e.g. folic acid 35.5 30.0 
Calcium-channel blockers, e.g. amlodipine 34.3 34.5 
Loop diuretics, e.g. furosemide 29.0 10.2 
Non-steroidal anti-inflammatory drugs, e.g. ibuprofen 28.4 12.5 
Anticoagulants, e.g. acenocoumarol 23.1 5.1 
Metformin 23.1 2.6 
Benzodiazepines, e.g. oxazepam 22.5 97.4 
Thyroid hormone and antithyroid agents, e.g. levothyroxine and carbimazole 20.7 2.9 
Drug classa Percentage of patients taking a medication from this drug class Percentage of non-evidence-based medications within this drug class 
Beta-blockers, e.g. metoprolol 55.0 54.8 
ACE inhibitors, e.g. lisinopril 52.1 2.3 
PPIs/H2 antagonists, e.g. pantoprazole and famotidine 50.3 50.6 
Statins, e.g. simvastatin 46.2 64.1 
Antiplatelet agents, e.g. clopidogrel 45.6 28.6 
Thiazide diuretics, e.g. HCT 42.6 9.7 
Vitamins and enzymes, e.g. folic acid 35.5 30.0 
Calcium-channel blockers, e.g. amlodipine 34.3 34.5 
Loop diuretics, e.g. furosemide 29.0 10.2 
Non-steroidal anti-inflammatory drugs, e.g. ibuprofen 28.4 12.5 
Anticoagulants, e.g. acenocoumarol 23.1 5.1 
Metformin 23.1 2.6 
Benzodiazepines, e.g. oxazepam 22.5 97.4 
Thyroid hormone and antithyroid agents, e.g. levothyroxine and carbimazole 20.7 2.9 

aDrug classes sorted according to frequency of use by the patients. Drug classes used by 20% or less of the patients are not shown (a more detailed table is available from the author upon request).

The most frequently taken drugs in our study were beta-blocking agents: 93 of the 169 patients (55.0%) took a beta-blocker. From these, 51 prescriptions (54.8%) were analysed to be non-evidence based according to the recommendations of UpToDate. To judge a beta-blocker as non-evidence based, three conditions had to apply: (i) hypertension was the only indication provided. (ii) The beta-blocker was the only medication, or one of only two medications, for hypertension. (iii) A more specific indication (like heart failure) for the use of a beta-blocker could not be inferred from any other information provided.

Dosing errors

Seventy-four of the 169 patients (56.2%) had at least one dosing error. Details of dosing errors are provided in Table 3 for the drugs used by more than 20% of the patients (all dosing errors detected are available from the author upon request).

Table 3

Dosing errors in all patients

Drug classa Percentage of over- or underdose within this drug class Percentage of other dosing errorsb within this drug class 
Beta-blockers, e.g. metoprolol 6.4 0.0 
ACE inhibitors, e.g. lisinopril 3.4 0.0 
PPIs/H2 antagonists, e.g. pantoprazole and famotidine 2.4 2.4 
Statins, e.g. simvastatin 0.0 0.0 
Antiplatelet agents, e.g. clopidogrel 3.9 3.9 
Thiazide diuretics, e.g. HCT 1.4 1.4 
Vitamins and enzymes, e.g. folic acid 0.0 0.0 
Calcium-channel blockers, e.g. amlodipine 8.6 5.2 
Loop diuretics, e.g. furosemide 0.0 16.3 
Non-steroidal anti-inflammatory drugs, e.g. ibuprofen 4.0 0.0 
Anticoagulants, e.g. acenocoumarol 0.0 0.0 
Metformin 5.1 0.0 
Benzodiazepines, e.g. oxazepam 5.9 10.5 
Thyroid hormone resp antithyroid agents, e.g. levothyroxine and carbimazole 5.8 0.0 
Drug classa Percentage of over- or underdose within this drug class Percentage of other dosing errorsb within this drug class 
Beta-blockers, e.g. metoprolol 6.4 0.0 
ACE inhibitors, e.g. lisinopril 3.4 0.0 
PPIs/H2 antagonists, e.g. pantoprazole and famotidine 2.4 2.4 
Statins, e.g. simvastatin 0.0 0.0 
Antiplatelet agents, e.g. clopidogrel 3.9 3.9 
Thiazide diuretics, e.g. HCT 1.4 1.4 
Vitamins and enzymes, e.g. folic acid 0.0 0.0 
Calcium-channel blockers, e.g. amlodipine 8.6 5.2 
Loop diuretics, e.g. furosemide 0.0 16.3 
Non-steroidal anti-inflammatory drugs, e.g. ibuprofen 4.0 0.0 
Anticoagulants, e.g. acenocoumarol 0.0 0.0 
Metformin 5.1 0.0 
Benzodiazepines, e.g. oxazepam 5.9 10.5 
Thyroid hormone resp antithyroid agents, e.g. levothyroxine and carbimazole 5.8 0.0 

aDrug classes sorted according to frequency of use by the patients. Drug classes used by 20% or less of the patients are not shown (a more detailed table is available from the author upon request).

bWrong interval or time, no regular dosing scheme, more than one drug out of the same drug class.

‘Missing data’ was common. The principal reason was that the GPs almost always entered the dose, however often in a format similar to: ‘Brand Name, 1× mornings, 1× evenings’. If, however, ‘Brand Name’ is available on the market in different strengths, the dose remained unknown.

It was noticed that non-evidence-based prescribing and dosing errors were not as common with ACE inhibitors as with for example benzodiazepines. In patients taking ACE inhibitors, non-evidence-based prescribing was found in 2 out of 88 patients (2.3%) and a dosing error in 2.3%. In patients taking benzodiazepines, non-evidence-based prescribing was found in 37 out of 38 patients (97.4%) and a dosing error in 19.5%.

Potential interactions

Of the 169 patients, 69 (40.8%) had no category D or X interaction.

Category D interactions: Ninety-nine patients (58%) had at least one category D interaction. Of these, 60 patients (35.5%) had one category D interaction, 20 patients (11.8%) two, 10 patients (5.9%) three, 3 patients (1.8%) four, 3 patients (1.8%) five and 2 patients (1.2%) six such interactions.

Category X interactions: Four patients (2.4%) had at least one category X interaction. Of these, three patients (1.8%) had one, and one patient (0.6%) two such interactions.

The five category X interactions were (comments verbatim from the database) as follows:

  • Clopidogrel + fluoxetine: CYP2C19 inhibitors (moderate) may decrease serum concentrations of the active metabolite(s) of clopidogrel.

  • Metoclopramide + quetiapine: Metoclopramide may enhance the adverse/toxic effect of antipsychotics.

  • Tamsulosin + terazosin: Alpha-1 blockers may enhance the antihypertensive effect of other alpha-1 blockers.

  • Lorazepam + olanzapine: Olanzapine may enhance the adverse/toxic effect of benzodiazepines.

  • Triazolam + olanzapine: Olanzapine may enhance the adverse/toxic effect of benzodiazepines.

Feedback about the medication error analysis

The six GPs who participated in the feedback procedure underlined that the analysis of evidence base and dosing was ‘useful’, but that especially the analysis of potential interactions was ‘very useful’. They emphasized they acted almost immediately upon this information.

PIMs in elderly

One hundred and fifty-eight of the 169 patients (93.5%) were ≥65 years. Fifty-nine of these seniors (37.3%) had at least one medication that was inappropriate according to the PRISCUS list for the elderly.

The commonest PIMs were tricyclic antidepressants, benzodiazepines, hypnotics, non-benzodiazepines, typical antipsychotics and alpha-1 blockers (Table 4).

Table 4

PIMs in the subgroup of elderly patients

Drug class Number of elderly taking an appropriate drug Number of elderly taking a PIM The exact PIM(s) in question Percentage of PIMs within this drug classa 
Tricyclic antidepressants Amitriptyline 100.0 
Benzodiazepines 32 Multiple benzodiazepines  86.5 
Hypnotics, non-benzodiazepines 10 Zolpidem >5mg/day  76.9 
Alpha-1 blockers 10 10 Doxazosin, terazosin  50.0 
Typical antipsychotics Haloperidol >2 mg  50.0 
Cardiac glycosides Acetyldigoxin, metildigoxin  40.0 
SSRIs 27 Fluoxetine  10.0 
Vasodilators 28 Pentoxifylline  9.7 
Calcium-channel blockers 49 Nifedipine, non-sustained release formulation  5.8 
Anticholinergic agents 29 Oxybutynin  3.3 
Beta-blockers 86 Sotalol  1.1 
Drug class Number of elderly taking an appropriate drug Number of elderly taking a PIM The exact PIM(s) in question Percentage of PIMs within this drug classa 
Tricyclic antidepressants Amitriptyline 100.0 
Benzodiazepines 32 Multiple benzodiazepines  86.5 
Hypnotics, non-benzodiazepines 10 Zolpidem >5mg/day  76.9 
Alpha-1 blockers 10 10 Doxazosin, terazosin  50.0 
Typical antipsychotics Haloperidol >2 mg  50.0 
Cardiac glycosides Acetyldigoxin, metildigoxin  40.0 
SSRIs 27 Fluoxetine  10.0 
Vasodilators 28 Pentoxifylline  9.7 
Calcium-channel blockers 49 Nifedipine, non-sustained release formulation  5.8 
Anticholinergic agents 29 Oxybutynin  3.3 
Beta-blockers 86 Sotalol  1.1 

aThe table is sorted according to ‘Percentage of PIMs’.

Discussion

The frequency of non-evidence-based medications, dosing errors and potential interactions in patients with polypharmacy is alarmingly high in this descriptive study. This was especially true for psychoactive drugs.

Thus, there appears to be a large potential for the discontinuation of non-evidence-based medicines, and to correct dosing errors as well as avoid potentially dangerous interactions, if it could be achieved to regularly review the medication of patients on polypharmacy. The questions that need to be addressed are, why polypharmacy is not avoided more consequently in every day practice, and why the medication errors detected in this study remain largely undetected in routine primary care.

There are several issues that need to be taken into account. Firstly, GPs are unable to perform evidence-based medication reviews with their patients because of a lack of time and a lack of evidence-based resources. Secondly, GPs often do not feel that they have sufficient expertise to discontinue medications originally prescribed by a specialist. This is especially true for psychoactive drugs.

For example, 39 patients (23.1%) took at least 2 psychoactive drugs, although it is widely accepted that switching to psychoactive monotherapy is a more rational choice.21

This may very well be due to the fact that GPs often cannot discontinue psychoactive drugs easily. All six GPs who participated in the feedback procedure mentioned that discontinuation of psychoactive drugs on their part may rather lead to the respective patients simply switching to another GP than discontinuing the drug. This seems to be especially true for patients taking benzodiazepines. GPs need assistance from specialists in patients with psychiatric diagnoses, and cooperative care can lead to better outcomes.22

Thirdly, GPs may not feel that they are able to deviate from guidelines. Rigid guideline adherence inevitably leads to polypharmacy in patients with multiple chronic diseases. Of the six GPs who participated in the feedback procedure, all but one unsolicitedly voiced such concerns about guidelines. This problem definitely needs to be addressed by medical societies that publish guidelines, which frequently have only weak evidence regarding the treatment of elderly patients with comorbidities.23

Our study thus reveals the urgent necessity to develop and provide tools to the GPs to rapidly analyse the medications of patients with polypharmacy. Such a tool could be electronic clinical decision support integrated in the patient health record, which could be used to discuss the discontinuation of medicines with the patient in a shared decision making process.

We performed this study with the objective to quantify the potential for drug discontinuation in patients with polypharmacy as a basis for the development of such tools and testing their implementation in a future prospective intervention trial.

Limitations and strengths of this study

This study is limited mainly because of its size. Furthermore, the participating GPs represent a convenience, not a randomly selected, sample, implying the risk of selection bias, as highly motivated physicians probably were more likely to participate. Yet, this might also implicate that the problem of medication errors might be even greater in a larger, non-selective sample of GP surgeries.

A further limitation of this study is that no in-depth analysis of individual, patient-oriented reasons for non-evidence-based medications was performed. The focus in this study lay on discrepancies between the stated indication for use and the prescribed medicine.

Also, there was no 100% concordance regarding medication errors between the three independent reviewers from the same institute, but these differences were relatively minor. Overall, inter-rater correlation, although not measured, appeared high, and any disagreement could be settled easily by discussion.

The strengths of this study include that at each GP’s office, the patients were included into the study consecutively to decrease the risk of selection bias, that the GPs who participated in this study came from inner cities as well as rural communities, and that data in this field of medical research are still scarce.

Conclusion

The frequency of medication errors and potential interactions in patients with polypharmacy in this small descriptive study is alarming and in line with other studies.2–7 Thus, our study strongly indicated that there is a large potential for reducing polypharmacy in primary care.

These results also strengthen the claim that polypharmacy constitutes a real problem and decreases patient safety. Physicians in family practices—often under pressure of time—can certainly overlook medication errors or interactions. Therefore, they probably would benefit from external analyses of the medications of their patients with polypharmacy.

Declaration

Funding: This work was supported by the European Union’s Seventh Framework programme FP7/2007–2013 under grant agreement no. 223424 (LINNEAUS: Learning from International Networks about Errors and Understanding Safety in Primary Care).

Ethical approval: The Ethics Committee for the Federal State of Salzburg (Ethikkommission für das Bundesland Salzburg), Austria, approved this study.

Conflict of interest: There was no conflict of interest. None of the authors have to disclose any ownership of shares, consultancy, speaker’s honoraria or research grants from commercial companies or professional or governmental organizations with an interest in the topic of the paper. Although UpToDate® and Lexi-Interact® are mentioned in this paper, none of the authors have to disclose any financial relationship of any kind with UpToDate®, Lexi-Comp® or Wolters Kluwer Health.

References

1
Marengoni
A
Winblad
B
Karp
A
Fratiglioni
L
.
Prevalence of multiple chronic diseases among the elderly population in Sweden
.
Am J Public Health
 
2008
;
98
:
1198
200
.
2
Slabaugh
SL
Maio
V
Templin
M
Abouzaid
S
.
Prevalence and risk of polypharmacy among the elderly in an outpatient setting: a retrospective cohort study in the Emilia-Romagna region, Italy
.
Drugs Aging
 
2010
;
27
:
1019
28
.
3
Junius-Walker
U
Theile
G
Hummers-Pradier
E
.
Prevalence and predictors of polypharmacy among older primary care patients in Germany
.
Fam Pract
 
2007
;
24
:
14
9
.
4
Thomsen
LA
Winterstein
AG
Sondergaard
B
Haugbolle
LS
Melander
A
.
Systematic review of the incidence and characteristics of preventable adverse drug events in ambulatory care
.
Ann Pharmacother
 
2007
;
41
:
1411
26
.
5
Hovstadius
B
Hovstadius
K
Astrand
B
Petersson
G
.
Increasing polypharmacy—an individual-based study of the Swedish population 2005–2008
.
BMC Clin Pharmacol
 
2010
;
10
:
16
.
6
Feudtner
C
Dai
D
Hexem
KR
Luan
X
Metjian
TA
.
Prevalence of polypharmacy exposure among hospitalized children in the United States
.
Arch Pediatr Adolesc Med
 
2012
;
166
:
9
16
.
7
Wong
CY
Chaudhry
SI
Desai
MM
Krumholz
HM
.
Trends in comorbidity, disability, and polypharmacy in heart failure
.
Am J Med
 
2011
;
124
:
136
43
.
8
Anthierens
S
Tansens
A
Petrovic
M
Christiaens
T
.
Qualitative insights into general practitioners views on polypharmacy
.
BMC Fam Pract
 
2010
;
11
:
65
.
9
Schuler
J
Dückelmann
C
Beindl
W
Prinz
E
Michalski
T
Pichler
M
.
Polypharmacy and inappropriate prescribing in elderly internal-medicine patients in Austria
.
Wien Klin Wochenschr
 
2008
;
120
:
733
41
.
10
The American Geriatrics Society 2012 Beers Criteria Update Expert Panel
.
American Geriatrics Society Updated Beers Criteria for potentially inappropriate medication use in older adults
.
J Am Geriatr Soc
 
2012
;
60
:
616
31
.
11
Holt
S
Schmiedl
S
Thürmann
PA
.
Potentially inappropriate medications in the elderly: the PRISCUS list for the elderly
.
Dtsch Arztebl Int
 
2010
;
107
:
543
51
.
12
Garfinkel
D
Mangin
D
.
Feasibility study of a systematic approach for discontinuation of multiple medications in older adults: addressing polypharmacy
.
Arch Intern Med
 
2010
;
170
:
1648
54
.
13
Gallagher
P
Ryan
C
Byrne
S
Kennedy
J
O’Mahony
D
.
STOPP (Screening Tool of Older Person’s Prescriptions) and START (Screening Tool to Alert doctors to Right Treatment). Consensus validation
.
Int J Clin Pharmacol Ther
 
2008
;
46
:
72
83
.
14
Garfinkel
D
Zur-Gil
S
Ben-Israel
J
.
The war against polypharmacy: a new cost-effective geriatric-palliative approach for improving drug therapy in disabled elderly people
.
Isr Med Assoc J
 
2007
;
9
:
430
4
.
15
Williams
ME
Pulliam
CC
Hunter
R
et al
The short-term effect of interdisciplinary medication review on function and cost in ambulatory elderly people
.
J Am Geriatr Soc
 
2004
;
52
:
93
8
.
16
http://www.UpToDate.com/ (last accessed on 29 October 2012).
17
Levey
AS
Stevens
LA
Schmid
CH
et al
A new equation to estimate glomerular filtration rate
.
Ann Intern Med
 
2009
;
150
:
604
12
.
18
http://webstore.lexi.com/lexi-interact(last accessed on 29 October 2012).
19
Thiem
U
Theile
G
Junius-Walker
U
et al
Prerequisites for a new health care model for elderly people with multimorbidity: the PRISCUS research consortium
.
Z Gerontol Geriatr
 
2011
;
44
:
115
20
.
20
Mann
E
Böhmdorfer
B
Frühwald
T
et al
Potentially inappropriate medication in geriatric patients: the Austrian consensus panel list
.
Wien Klin Wochenschr
 
2012
;
124
:
160
9
.
21
Essock
SM
Schooler
NR
Stroup
TS
et al
Effectiveness of switching from antipsychotic polypharmacy to monotherapy
.
Am J Psychiatry
 
2011
;
168
:
702
8
.
22
Katon
WJ
Lin
EH
Von Korff
M
et al
Collaborative care for patients with depression and chronic illnesses
.
N Engl J Med
 
2010
;
363
:
2611
20
.
23
Wehling
M
.
Guideline-driven polypharmacy in elderly, multimorbid patients is basically flawed: there are almost no guidelines for these patients
.
J Am Geriatr Soc
 
2011
;
59
:
376
7
.