This interesting meta-analysis was just published (notably, in a very reputable open-access, advertisement-free journal). 20% seems a very conservative estimate…
The PDF is here:

Here is the abstract:

RESEARCH ARTICLE
Inappropriateness of Medication Prescriptions to Elderly Patients in the Primary Care Setting: A Systematic Review

Dedan Opondo1*, Saied Eslami1, Stefan Visscher2, Sophia E. de Rooij3, Robert Verheij2, Joke C. Korevaar2, Ameen Abu-Hanna1
1 Department of Medical Informatics, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands, 2 Netherlands Institute for Health Services Research (NIVEL), Utrecht, The Netherlands, 3 Department of Geriatrics, Academic Medical Center, University of Amsterdam, The Netherlands
Abstract Top
Background
Inappropriate medication prescription is a common cause of preventable adverse drug events among elderly persons in the primary care setting.

Objective
The aim of this systematic review is to quantify the extent of inappropriate prescription to elderly persons in the primary care setting.

Methods
We systematically searched Ovid-Medline and Ovid-EMBASE from 1950 and 1980 respectively to March 2012. Two independent reviewers screened and selected primary studies published in English that measured (in)appropriate medication prescription among elderly persons (>65 years) in the primary care setting. We extracted data sources, instruments for assessing medication prescription appropriateness, and the rate of inappropriate medication prescriptions. We grouped the reported individual medications according to the Anatomical Therapeutic and Chemical (ATC) classification and compared the median rate of inappropriate medication prescription and its range within each therapeutic class.

Results
We included 19 studies, 14 of which used the Beers criteria as the instrument for assessing appropriateness of prescriptions. The median rate of inappropriate medication prescriptions (IMP) was 20.5% [IQR 18.1 to 25.6%.]. Medications with largest median rate of inappropriate medication prescriptions were propoxyphene 4.52(0.10?23.30)%, doxazosin 3.96 (0.32 15.70)%, diphenhydramine 3.30(0.02?4.40)% and amitriptiline 3.20 (0.05?20.5)% in a decreasing order of IMP rate. Available studies described unequal sets of medications and different measurement tools to estimate the overall prevalence of inappropriate prescription.

Conclusions
Approximately one in five prescriptions to elderly persons in primary care is inappropropriate despite the attention that has been directed to quality of prescription. Diphenhydramine and amitriptiline are the most common inappropriately prescribed medications with high risk adverse events while propoxyphene and doxazoxin are the most commonly prescribed medications with low risk adverse events. These medications are good candidates for being targeted for improvement e.g. by computerized clinical decision support.

Citation: Opondo D, Eslami S, Visscher S, de Rooij SE, Verheij R, et al. (2012) Inappropriateness of Medication Prescriptions to Elderly Patients in the Primary Care Setting: A Systematic Review. PLoS ONE 7(8): e43617. doi:10.1371/journal.pone.0043617

Editor: Colin Simpson, The University of Edinburgh, United Kingdom

Received: January 3, 2012; Accepted: July 23, 2012; Published: August 22, 2012

Copyright: ? 2012 Opondo et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: These authors have no support or funding to report.

Competing interests: The authors have declared that no competing interests exist.

* E-mail: d.o.opondo@amc.uva.nl