Abstract

Background: inappropriate prescribing encompasses acts of commission i.e. giving drugs that are contraindicated or unsuitable, and acts of omission i.e. failure to prescribe drugs when indicated due to ignorance of evidence base or other irrational basis e.g. ageism. There are considerable published data on the prevalence of inappropriate prescribing; however, there are no recent published data on the prevalence of acts of omission. The aim of this study was to calculate the prevalence of acts of prescribing omission in a population of consecutively hospitalised elderly people.

Methods: a screening tool (screening tool to alert doctors to the right treatment acronym, START), devised from evidence-based prescribing indicators and arranged according to physiological systems was prepared and validated for identifying prescribing omissions in older adults. Data on active medical problems and prescribed medicines were collected in 600 consecutive elderly patients admitted from the community with acute illness to a teaching hospital. On identification of an omitted medication, the patient's medical records were studied to look for a valid reason for the prescribing omission.

Results: using the START list, we found one or more prescribing omissions in 57.9% of patients. In order of prevalence, the most common prescribing omissions were: statins in atherosclerotic disease (26%), warfarin in chronic atrial fibrillation (9.5%), anti-platelet therapy in arterial disease (7.3%) and calcium/vitamin D supplementation in symptomatic osteoporosis (6%).

Conclusion: failure to prescribe appropriate medicines is a highly prevalent problem among older people presenting to hospital with acute illness. A validated screening tool (START) is one method of systematically identifying appropriate omitted medicines in clinical practice.

Introduction

Inappropriate prescribing of medications in older people is an important cause of morbidity and mortality and has been studied well in Europe and the United States [ 1–9 ]. It encompasses the overuse of drugs, prescribing drugs that are predictably tolerated poorly by most older people, prescribing drugs that are likely to exacerbate a clinical problem in an older person (e.g. benzodiazepines in the presence of recurrent falls) and the underuse of appropriate medication [ 2 ]. Most of the published literature on inappropriate prescribing in late life deals with acts of commission, i.e. the prescribing of drugs that should be avoided. There is also literature on underprescribing in the elderly although most studies identify single instances only. In contrast, there are very few published data on screening tools that measure multiple acts of prescribing omission i.e. the failure on the part of doctors to prescribe drugs that are clearly indicated and likely to benefit the patient [ 10–12 ]. This may, in part, result from the lack of suitable screening tools designed to alert the clinician to consider indicated drugs when identified in individual patients. Screening tools designed to detect acts of commission of inappropriate prescribing in elderly patients, such as Beers' criteria [ 13–15 ] and the inappropriate prescribing the elderly tool (IPET) [ 16 , 17 ] have been researched widely, although not used routinely in clinical practice.

The aims of this study were therefore threefold: (i) to devise and validate an evidence-based screening tool for indicated medicines of particular relevance to older people and (ii) to determine the prevalence of omission of indicated medicines in a population of older people hospitalised with acute illness. As a corollary to these aims, we also sought to (iii) calculate the cost of prescribing the indicated but omitted medicines in this population of patients.

Methods

A senior academic geriatrician (DOM) composed the original list of prescribing criteria on the basis of extensive literature review, recent texts on geriatric pharmacotherapy and clinical experience. Twenty two evidence-based common prescribing indicators for elderly patients were identified and arranged according to the relevant physiological systems into a systematic list called screening tool to alert doctors to the right treatment (i.e. indicated, but not prescribed) (START) for older people. Eighteen experts, with recognised credentials in their specialist areas, were invited by letter to participate in the Delphi process [ 18 ]. Study design and aims were explained. The panel comprised teaching hospital consultants in geriatric medicine ( n  = 9), clinical pharmacology ( n  = 3) and old age psychiatry ( n  = 1), two senior academic primary care physicians and three senior hospital pharmacists with an interest in geriatric pharmacology. In addition to seeking consensus from the panel on the list of 22 specific evidence-based prescribing indicators, we asked each panel member to suggest any further important prescribing indications.

The first round questionnaire was posted to each panellist and START criteria were presented as statements describing each instance of potentially inappropriate prescribing in people aged 65 years. The Delphi process was completed in two rounds and full consensus was reached without the need to proceed to a third round. Subsequently, inter-rater reliability was addressed by the review of 100 charts by two observers using the START tool. A κ-coefficient of 0.68 was calculated. This suggested that the tool performed well with substantial agreement obtained.

The local research ethics committee approved the patient study protocol. The validated version of START (Figure 1 ) was applied to concurrent medical diagnostic and prescription information in a prospective, unselected consecutive cohort of 600 community-dwelling patients (aged 65 years and over) on admission to hospital with acute illness. The mean age (SD) of the patients was 77.9 (6.8) years. Two hundred and one patients were aged 65–74 years (33% of total), 299 patients were aged 75–84 years (50%) and the remaining 100 patients were aged 85 years or over (17% of total). Fifty-six per cent of the patients were female. Patients who were resident in local community hospitals were excluded because of the possible influence of hospital-based consultant geriatricians on their medications (regular input into community hospital care of elderly people being part of the remit of geriatricians in the catchment area of the teaching hospital where the research was carried out). Baseline demographic information was obtained as well as the results of relevant baseline investigations from each patient's hospital case records. Medical co-morbidities and the full list of current medications list were documented following detailed clinical assessment and prescription review at the time of admission to hospital, and before any changes to medications were made by the attending physician in the hospital. These lists were documented from a number of sources including General Practitioners referral letters, the patients' own medication list, pharmacy records where necessary and the hospital admission records and notes. Data capture occurred once for each patient and was completed by a specialist registrar in geriatric medicine, supported by an experienced research nurse. Medication details were corroborated from as many sources as possible. The START criteria were then applied to the defined handwritten list of co-morbidities on the day of hospital admission and the patients' medication lists on admission. The number of omitted appropriate prescriptions was identified and recorded accordingly. The precise definition of co-morbidities on admission facilitated deployment of the START tool in less than 3 minutes in the majority of cases.

Figure 1

Screening tool to alert doctors to the right (i.e. indicated, but not prescribed) treatment for older people (START)

Figure 1

Screening tool to alert doctors to the right (i.e. indicated, but not prescribed) treatment for older people (START)

The contra-indications to the medicines in the START tool refer to the clearly defined contra-indications specified in the British National Formulary [British National Formulary: 48th edition (Sep 2004)]. Formalised assessment tools scoring was not used in the tool. This was partly to allow for variations in assessments used in different centres but also to prevent a screening tool from becoming an over-elaborate document requiring inclusion of multiple other assessment tools.

In addressing the secondary aim of calculating the financial cost of those indicated but omitted medicines, current drug-manufacturing costs (wholesale costs) were derived from a national formulary of prescription medicines. [Medical Publications (Ireland) Limited, Monthly index of medical specialties. Dublin; September 2004.] Where possible, the costs of the cheapest generic formulation of the indicated but omitted medicines were calculated. Cost calculation was based on 30 days' prescription of each indicated but omitted medicine, excluding pharmacist's dispensing charges which are variable.

Results

A total of 3,234 medications were prescribed to the 600 patients up to the point of acute admission to hospital. The median number of medications per patient was five. Using the START criteria, one or more appropriate medicines were omitted in 347 (57.8%), where no contra-indication existed. The probability of omission of an appropriate medicine increased with advancing age (detailed below). The medications omitted are detailed here. These were in order of frequency: statins for atherosclerotic cardiovascular disease, warfarin thrombo-embolic prophylaxis in chronic atrial fibrillation, angiotensin coverting enzyme (ACE) inhibitor therapy for congestive cardiac failure, aspirin for symptomatic stenotic arterial disease and calcium supplementation for established symptomatic osteoporosis (i.e. prior history of fragility fractures). See Table 1 .

The top five 30-day prescription costs (rounded off to the nearest Euro) associated with these indicated but omitted medicines for all 600 patients were as follows in order of decreasing cost:

  • Statins in symptomatic cardiovascular disease    €3926

  • Bisphosphonates with long term corticosteroid treatment     €1056

  • ACE inhibitors in congestive cardiac failure     €738

  • Statins in diabetes mellitus with hypercholestero laemia     €604

  • ACE inhibitors with a prior history of myocardial infection     €446

Table 1

Itemised 30-day cost of indicated, but omitted generic drug therapy in 600 acutely- ill hospitalised elderly people, based on START criteria

Indication Medication No. of subjects Daily dose (mg) 30 day cost (€) Omitted med's cost (€) 
• Statin therapy in patients with documented history of vascular disease Atorvastatin 156 10 25.17 3926.52 
• Warfarin in the presence of chronic atrial fibrillation Warfarin 57 10 5.70 303.90 
• ACE inhibitor in chronic heart failure Ramipril 48 15.38 738.24 
• Aspirin with a documented history of coronary, cerebral or peripheral vascular disease Aspirin 44 75 2.24 98.56 
• Calcium and vitamin D supplement in patients with osteoporosis Calcium/vitamin D 35 1 g/800 iu 10 350 
• Metformin with type 2 diabetes +/− metabolic syndrome Metformin 34 1500 2.88 97.92 
• Bisphosphonate in patients taking glucocorticoids >1 month Alendronate/weekly 30 70 35.26 1057.80 
• ACE inhibitor following acute myocardial infection. Ramipril 29 15.38 446.02 
• β-blocker in chronic angina Bisoprolol 28 8.51 238.28 
• Regular inhaled β 2-agonist or anti-cholinergic agent for mild to moderate asthma or COPD Salbutamol 24 0.2 3.31 79.44 
• Inhaled steroid in moderate–severe asthma or COPD Budesonide 24 0.4 24.02 576.48 
• Statin therapy in diabetes mellitus if fasting serum cholesterol >5.0 mmol/l Atorvastatin 24 10 25.17 604.08 
• ACE inhibitor in diabetes with overt proteinuria or microalbuminuria Ramipril 23 15.38 353.74 
• Aspirin therapy in diabetes mellitus with well controlled BP Aspirin 16 75 2.24 35.84 
• Aspirin in the presence of chronic atrial fibrillation, where warfarin contra-indicated Aspirin 14 75 2.24 31.36 
• Disease-modifying drug with known, moderate-severe rheumatoid disease Methotrexate 13 7.5 2.00 26.00 
• Antidepressant in the presence of clear-cut depression Citalopram 10 10 17.39 173.90 
• Antihypertensive therapy Bendrofluazide 2.5 1.40 11.20 
• Fibre supplement for chronic, symptomatic diverticular disease with constipation. Fybogel 2 sachets 2.92 8.76 
• PPI in the presence of chronic severe gastro-oesophageal acid reflux or peptic stricture requiring dilatation. Omeprazole 20 41.46 165.84 
L -DOPA in Parkinson's disease with definite functional impairment  Levodopa + Carbidopa 62.5 × 3 10.12 40.48 
• Home continuous oxygen where chronic type 1 or type 2 respiratory failure has been well documented. Oxygen concentrator N/A 76.82 
Total cost of medications/month     €9364.34 
Indication Medication No. of subjects Daily dose (mg) 30 day cost (€) Omitted med's cost (€) 
• Statin therapy in patients with documented history of vascular disease Atorvastatin 156 10 25.17 3926.52 
• Warfarin in the presence of chronic atrial fibrillation Warfarin 57 10 5.70 303.90 
• ACE inhibitor in chronic heart failure Ramipril 48 15.38 738.24 
• Aspirin with a documented history of coronary, cerebral or peripheral vascular disease Aspirin 44 75 2.24 98.56 
• Calcium and vitamin D supplement in patients with osteoporosis Calcium/vitamin D 35 1 g/800 iu 10 350 
• Metformin with type 2 diabetes +/− metabolic syndrome Metformin 34 1500 2.88 97.92 
• Bisphosphonate in patients taking glucocorticoids >1 month Alendronate/weekly 30 70 35.26 1057.80 
• ACE inhibitor following acute myocardial infection. Ramipril 29 15.38 446.02 
• β-blocker in chronic angina Bisoprolol 28 8.51 238.28 
• Regular inhaled β 2-agonist or anti-cholinergic agent for mild to moderate asthma or COPD Salbutamol 24 0.2 3.31 79.44 
• Inhaled steroid in moderate–severe asthma or COPD Budesonide 24 0.4 24.02 576.48 
• Statin therapy in diabetes mellitus if fasting serum cholesterol >5.0 mmol/l Atorvastatin 24 10 25.17 604.08 
• ACE inhibitor in diabetes with overt proteinuria or microalbuminuria Ramipril 23 15.38 353.74 
• Aspirin therapy in diabetes mellitus with well controlled BP Aspirin 16 75 2.24 35.84 
• Aspirin in the presence of chronic atrial fibrillation, where warfarin contra-indicated Aspirin 14 75 2.24 31.36 
• Disease-modifying drug with known, moderate-severe rheumatoid disease Methotrexate 13 7.5 2.00 26.00 
• Antidepressant in the presence of clear-cut depression Citalopram 10 10 17.39 173.90 
• Antihypertensive therapy Bendrofluazide 2.5 1.40 11.20 
• Fibre supplement for chronic, symptomatic diverticular disease with constipation. Fybogel 2 sachets 2.92 8.76 
• PPI in the presence of chronic severe gastro-oesophageal acid reflux or peptic stricture requiring dilatation. Omeprazole 20 41.46 165.84 
L -DOPA in Parkinson's disease with definite functional impairment  Levodopa + Carbidopa 62.5 × 3 10.12 40.48 
• Home continuous oxygen where chronic type 1 or type 2 respiratory failure has been well documented. Oxygen concentrator N/A 76.82 
Total cost of medications/month     €9364.34 

Details of the total 30-day costs of omitted but indicated medicines are shown in Table 1 . The total cost of all omitted medications for the patient group was €9364.34. The probability of omission of a potentially appropriate medication was significantly related to patients' age. The likelihood of having an appropriate medication omitted did not change in the 65–74 age group (55.2% with one medication omitted) and the 75–84-year-old group (54.8% with one medication omitted)—Odds Ratio 1.0512, CI 0.71–1.45, P  = 0.93. However there was a significant likelihood of omission of appropriate medication in the group older than 85 years (72.2% with one medication omitted)—Odds Ratio 2.08, CI 1.24–3.50, P < 0.01. The likelihood of having an appropriate medication omitted in females compared to males was also increased (Odds Ratio 2.29, CI 1.65–3.19, P  < 0.01).

Discussion

This study describes the development and validation of a new systems-based tool to help medical practitioners identify possible appropriate medications in older adults. It was validated using a technique described previously in other similar tools [ 15–17 ]. It may prove useful if applied in a prospective manner in primary care and in an acute general medical setting among hospital in-patients. One of the major limitations in the use of tools to identify prescription of inappropriate medications or underutilisation of appropriate drugs is that studies to date have in essence calculated only prevalence rates and have not identified if these tools can actually influence prescribing in the longer term. There is scant evidence to support the clinical benefit of these tools and well-designed pragmatic randomised controlled trials are required to evaluate this.

This is one of the few studies to report the rate of omission of appropriate, evidence-based medicines in elderly people [ 10–12 , 19 , 20 ] In the current study, over half of the elderly acutely-ill newly hospitalised patients—57.9%—had at least one appropriate medication omitted from their list of regular prescription medicines. The probability of not receiving an appropriate medication increased with age over 85 years and female gender. Failure to prescribe appropriate medicines, which have a proven important role in primary and secondary disease prevention, could have a substantial clinical and economic impact over time, although there are, as yet, no prospective randomised controlled trial data to support this suggestion. Our data show that the over 85 age group were less likely to be prescribed appropriately; this may reflect a desire to avoid polypharmacy or lack of clear-cut evidence of efficacy of particular therapeutic interventions in this age group. Limitations in data collection could have underestimated use of appropriate medications also and complete records of patients' medication lists depended on access to referring general practitioners' letters or the patients' own list of medicines. However, where there were doubts about prescribed agents, confirmation was sought by telephone from the patients' general practitioners and in some instances, from the dispensing pharmacy to ensure the list of medications included in the study were complete and accurate.

The financial cost of the omitted medicines was not large; in this study it was calculated at €112,745 per year for the 600 subjects, this being the wholesale cost of the omitted drugs in generic form (and not including extra costs such as pharmacists' dispensing fees and use of non-generic drugs). This may seem substantial until viewed with the perspective of secondary prevention. For example, 71 subjects with chronic atrial fibrillation did not receive warfarin or aspirin, despite the absence of clear-cut contra-indications to these medicines. In this patient age group not receiving thrombo-embolic prophylaxis, the annual risk of stroke is approximately 10–15% [ 21 ]. Therefore, in the 71 patients with chronic atrial fibrillation and intrinsic heart disease, approximately 7–11 stroke events would be expected in this group each year. Warfarin therapy would be expected to reduce the annual stroke risk by approximately 60% i.e. to prevent 4–7 cases of avoidable stroke [ 21 ]. The total cost of treating these 4–7 stroke cases in a teaching hospital in 2006 is calculated at €38,000 (4 cases) to €66,500 (7 cases). This estimate is based on recent cost estimations and taking annual healthcare inflation taken into account [ 22 ]. However, it is recognised that warfarin therapy in this age group does carry a risk of bleeding and this cost may also needed to be taken.

However, omission of evidence-based appropriate medicines by physicians in almost 58% of elderly patients being hospitalised with acute illness remains unacceptably high. The present study was not designed to identify the precise reasons for this high omission rate. There are several possible reasons for this finding. These include a lack of knowledge of evidence-based secondary preventive therapies, low levels of therapeutic expectation in frail people aged over 80 years, a desire to avoid polypharmacy, greater focus on palliation of symptoms than on secondary disease prevention and, in some cases, negative ageist and sexist attitudes leading to therapeutic nihilism. There is limited evidence for many drug therapies in the over 80 year age group and some decisions not to prescribe certain drugs in the over 80s may have been rational, based on the lack of high quality evidence, in certain age groups. Further work needs to be done to clarify the main causes of inappropriate omission of medicines, since these causes will determine the necessary interventions needed for corrective action. This is particularly relevant given that, ironically, the most frequent omissions of evidence-based preventive drug therapy in the present study related to cardiovascular disease, the leading cause of death in older people globally.

The real clinical value of screening tools for inappropriate medication remains unclear. Several studies describe the prevalence of prescription of inappropriate drugs in older people, using screening tools such as Beers' criteria and IPET [ 13–17 ]. However, whether the regular use of such screening tools in day-to-day clinical practice results in significantly reduced morbidity, hospitalisation and mortality remains speculative. Similarly, it is unknown whether the routine use of inappropriate prescription-screening tools leads to significant drug expenditure savings. Future large scale randomised controlled studies will need to focus on these fundamental research questions, if screening tools are to have any clinical value in routine practice.

In conclusion, the present study shows that the prevalence of omission of evidence-based appropriate drug therapy in acutely-ill hospitalised elderly people was high, particularly in relation to cardiovascular disease prevention. START is a validated, effective and easy-to-use systems-based screening tool that is capable of identifying inappropriate omission of essential drug therapy in a matter of minutes. Further work needs to be done using START in other populations of elderly people, such as non-hospitalised elderly people living at home and in nursing homes. Ideally, START would be combined with a complimentary screening tool for identifying drug therapy that may need to be avoided in the same patient. To this end, we have completed preliminary validation work on such a complimentary screening tool to START, called screening tool of older persons' potentially inappropriate prescriptions (STOPP). Similar to START, STOPP is systems-based and evidence-based, and is quick and easy to use in day-to-day practice. Our research group is currently completing the Delphi validation of STOPP.

Conflicts of interest

None

Key points

  • START (screening tool to alert doctors to the right treatment) is an evidence-based, systems-defined screening tool to detect prescribing omissions in elderly patients and has been validated using best practice.

  • The use of the START tool identified that approximately 57% of older adults admitted to a teaching hospital had at least one appropriate medication omitted.

  • Older female adults were significantly more likely not to have appropriate medication prescribed. This may reflect the age bias of most large scale randomised controlled trials.

Acknowledgements

The authors wish to thank the Consultant medical staff of Cork University Hospital for granting access to their patients and their medical records for the purpose of this research. The research was funded by the Health Research Board of Ireland.

References

1.
Barry
M
Drug expenditure in Ireland 1991–2001
Ir Med J
 , 
2002
, vol. 
95
 (pg. 
294
-
5
)
2.
Kinirons
MT
Wood
AJJ
George
CF
Woodhouse
K
Denham
MJ
Pharmacokinetics
Drug Therapy in Old Age
 , 
1998
Wiley
3.
Juurlink
DN
Mamdami
M
Kopp
A
, et al.  . 
Drug-drug interactions among elderly patients hospitalised for drug toxicity
JAMA
 , 
2003
, vol. 
289
 pg. 
1652
 
4.
Onder
G
Pedone
C
Landi
F
, et al.  . 
Adverse drug reactions as a cause of hospital admissions: results from the Italian Group of Pharmacoepidemiology in the Elderly (GIFA)
J Am Geriatr Soc
 , 
2002
, vol. 
50
 pg. 
1962
 
5.
Goldberg
RM
Mabee
J
Chan
L
, et al.  . 
Drug-drug and drug-disease interactions in the emergency department: analysis of a high-risk population
Am J Emerg Med
 , 
1996
, vol. 
14
 (pg. 
447
-
50
)
6.
Hanlon
JT
Schmader
KE
Kornkowski
MJ
, et al.  . 
Adverse drug events in high risk older outpatients
J Am Geriatr Soc
 , 
1997
, vol. 
45
 (pg. 
945
-
8
)
7.
Bootman
JL
Harrison
DL
Cox
E
The health care cost of drug related morbidity and mortality in nursing facilities
Arch Intern Med
 , 
1997
, vol. 
157
 (pg. 
2089
-
96
)
8.
Bates
DW
Cullen
DJ
Laird
N
, et al.  . 
Incidence of adverse drug events and potential adverse drug events: Implications for prevention. ADE prevention Study Group
JAMA
 , 
1995
, vol. 
274
 (pg. 
29
-
34
)
9.
Bates
DW
Leape
LL
Petrycki
S
Incidence and preventability of adverse drug events in hospitalised adults
J Gen Intern Med
 , 
1993
, vol. 
8
 (pg. 
289
-
94
)
10.
Oborne
CA
Batty
GM
Maskrey
V
Swift
CG
Jackson
SH
Development of prescribing indicators for elderly medical inpatients
Br J Clin Pharmacol
 , 
1997
, vol. 
43
 (pg. 
91
-
7
)
11.
Oborne
CA
Hooper
R
Swift
CG
, et al.  . 
Explicit evidence based criteria to assess the quality of prescribing to elderly nursing home residents
Age Ageing
 , 
2003
, vol. 
32
 (pg. 
102
-
8
)
12.
Batty
GM
Grant
RL
Aggarwal
R
, et al.  . 
Using prescribing indicators to measure the quality of prescribing to elderly medical in-patients
Age Ageing
 , 
2003
, vol. 
32
 (pg. 
292
-
8
)
13.
Beers
MH
Ouslander
JG
Rollingher
I
, et al.  . 
Explicit criteria for determining inappropriate medication use in nursing home residents
Arch Intern Med
 , 
1991
, vol. 
151
 (pg. 
1825
-
32
)
14.
Beers
MH
Explicit criteria for determining potentially inappropriate medication use by the elderly. An update
Arch Intern Med
 , 
1997
, vol. 
157
 (pg. 
1531
-
6
)
15.
Fick
DM
Cooper
JW
Wade
W
, et al.  . 
Updating the Beers criteria for potentially inappropriate medication use in older Adults–Results of a US consensus panel of experts
Arch Intern Med
 , 
2003
, vol. 
163
 (pg. 
2716
-
24
)
16.
McLeod
PJ
Huang
AR
Tamblyn
RN
, et al.  . 
Defining inappropriate practices in prescribing for elderly people: a national concensus panel
CMAJ
 , 
1997
, vol. 
156
 (pg. 
385
-
91
)
17.
Naugler
CT
Brymer
C
Stolee
P
, et al.  . 
Development and validation of an improved prescribing for the elderly tool
Can J Clin Pharmacol
 , 
2000
, vol. 
7
 (pg. 
103
-
7
)
18.
Dalkey
N
Brown
B
Cochran
S
The Delphi Method, III: Use of Self Ratings to Improve Group Estimates
1969
Santa Monica
Rand Corp
 
NovemberPublication RM-6115-PR
19.
Batty
GM
Grant
RL
Aggarwal
R
, et al.  . 
National clinical sentinel audit of evidence-based prescribing for older people
J Eval Clin Pract
 , 
2004
, vol. 
10
 (pg. 
273
-
9
)
20.
Steinman
MA
Seth Landefeld
C
Rosenthal
GE
, et al.  . 
Polypharmacy and prescribing quality in older people
J Am Geriatr Soc
 , 
2006
, vol. 
54
 (pg. 
1516
-
23
)
21.
Wolf
PA
Abbott
RD
Kannel
WB
Atrial fibrillation as an independent risk factor for stroke: the Framingham Study
Stroke
 , 
1991
, vol. 
22
 (pg. 
983
-
8
)
22.
McGowan
B
Heerey
A
Tilson
L
, et al.  . 
Cost of treating stroke in an Irish teaching hospital
IMJ
 , 
2003
, vol. 
96
 (pg. 
234
-
6
)

Comments

0 Comments