Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Adverse Drug Reactions in Hospital In-Patients: A Prospective Analysis of 3695 Patient-Episodes

  • Emma C. Davies,

    Affiliations The Royal Liverpool and Broadgreen University Hospitals Trust, Liverpool, United Kingdom, School of Pharmacy and Chemistry, Liverpool John Moores University, Liverpool, United Kingdom

  • Christopher F. Green,

    Affiliation Countess of Chester NHS Foundation Trust, Pharmacy: Martindale House, Countess of Chester Health Park, Chester, United Kingdom

  • Stephen Taylor,

    Affiliation Centre for Medical Statistics and Health Evaluation, University of Liverpool, Liverpool, United Kingdom

  • Paula R. Williamson,

    Affiliation Centre for Medical Statistics and Health Evaluation, University of Liverpool, Liverpool, United Kingdom

  • David R. Mottram,

    Affiliation School of Pharmacy and Chemistry, Liverpool John Moores University, Liverpool, United Kingdom

  • Munir Pirmohamed

    munirp@liv.ac.uk

    Affiliation Department of Pharmacology and Therapeutics, The University of Liverpool, Liverpool, United Kingdom

Adverse Drug Reactions in Hospital In-Patients: A Prospective Analysis of 3695 Patient-Episodes

  • Emma C. Davies, 
  • Christopher F. Green, 
  • Stephen Taylor, 
  • Paula R. Williamson, 
  • David R. Mottram, 
  • Munir Pirmohamed
PLOS
x

Abstract

Adverse drug reactions (ADRs) are a major cause of hospital admissions, but recent data on the incidence and clinical characteristics of ADRs which occur following hospital admission, are lacking. Patients admitted to twelve wards over a six-month period in 2005 were assessed for ADRs throughout their admission. Suspected ADRs were recorded and analysed for causality, severity and avoidability and whether they increased the length of stay. Multivariable analysis was undertaken to identify the risk factors for ADRs. The 5% significance level was used when assessing factors for inclusion in multivariable models. Out of the 3695 patient episodes assessed for ADRs, 545 (14.7%, 95% CI 13.6–15.9%) experienced one or more ADRs. Half of ADRs were definitely or possibly avoidable. The patients experiencing ADRs were more likely to be older, female, taking a larger number of medicines, and had a longer length of stay than those without ADRs. However, the only significant predictor of ADRs, from the multivariable analysis of a representative sample of patients, was the number of medicines taken by the patient with each additional medication multiplying the hazard of an ADR episode by 1.14 (95% CI 1.09, 1.20). ADRs directly increased length of stay in 147 (26.8%) patients. The drugs most frequently associated with ADRs were diuretics, opioid analgesics, and anticoagulants. In conclusion, approximately one in seven hospital in-patients experience an ADR, which is a significant cause of morbidity, increasing the length of stay of patients by an average of 0.25 days/patient admission episode. The overall burden of ADRs on hospitals is high, and effective intervention strategies are urgently needed to reduce this burden.

Introduction

Adverse drug reactions (ADRs) in hospitalised patients can be divided into two broad categories: those that cause admission to hospital, and those that occur in in-patients after hospital admission. In a meta-analysis, using a random-effects model to reduce heterogeneity, Lazarou et al [1] showed that the total incidence of both categories of serious ADRs was 6.7%, of which 4.7% were responsible for admission and 2.1% occurred after admission, with an overall fatality rate of 0.32%. A recent Swedish study has also implicated ADRs as 7th most common cause of death [2]. In a study of almost 19000 admissions, we were able to show that 6.5% of patient admissions to two National Health Service (NHS) hospitals in the UK were related to an ADR [3]. This incidence figure is broadly compatible with pooled data from older studies [1], [4], and with more recent studies [5], [6].

By contrast, data on ADRs occurring after hospital admissions are poor. Older studies have suggested that between 10–20% of patients suffer ADRs in hospital [7][10], while Lazarou et al suggested that 10.9% of patients suffer ADRs of all severities as in-patients [1]. A systematic review by Wiffen et al estimated that in the NHS in England, 1.6 million bed days, equivalent to 13.6 (400-bed) hospital equivalents annually are due to in-patient ADRs [4]. It is important to note that most of these data relate to studies that are decades old. With the changing demographics, the well-known predisposition of the elderly to ADRs, and the changes in medical practice that have occurred over the last few decades, there is a need for more data on the ADR burden in hospital in-patients.

As part of our overall strategy to determine the burden of ADRs in hospitals, after the completion of our ADR hospital admission study [3], we undertook a pilot study to establish the methodology for determining the burden of ADRs in in-patients. This pilot study of 125 in-patients showed that 19% of patients suffered ADRs, with patients experiencing an ADR spending 6.5 days longer in hospital than those without ADRs [11]. In this paper, we report the results of our large-scale prospective study which further explores the impact of ADRs on NHS hospital in-patients in terms of incidence, length of stay, costs involved, and factors that predispose patients to ADRs.

Methods

Patients and settings

The study was conducted on 12 wards (9 medical and 3 surgical) at the Royal Liverpool University Hospital (RLUH) over a six-month period between June and December 2005. The RLUH is a teaching hospital which serves a population of about 0.5 million with a total annual activity of 90,000 admissions. The study protocol was assessed and approved by the Liverpool Local Research Ethics Committee and the audit department at the RLUH, and the Research Ethics Committee at Liverpool John Moores University.

Methods

For the purposes of this study, an ADR was defined according to the definition of Edwards and Aronson [12]. ADRs were identified on the basis that they were well recognised as evidenced by their inclusion in either the Summary of Product Characteristics [13] and/or the British National Formulary [14]. Only ADRs that occurred during admission as a result of drugs initiated or continued in hospital were included, while community acquired longstanding ADRs that were treated during the hospital stay were excluded (n = 17, 2.3% of all ADRs detected). ADRs that manifested no clinical signs, for example, suspected drug-induced abnormalities in blood test results were differentiated from those which caused clinical symptoms.

The study wards were a sample representative of the medical to surgical ward ratio in the hospital. Intensive care units and more specialist units such as the renal dialysis unit were excluded as our focus was on ADRs occurring in wards that are found in most UK general hospitals. The RLUH does not have psychiatric, paediatric or obstetrics and gynaecology wards, and thus our estimate of the incidence of ADRs excludes such patient groups. Patients admitted to the study wards during the data collection period were identified daily (Monday to Friday) by the research pharmacist using the hospital Patient Administration System (PAS). Patients whose admission did not include a weekday were therefore excluded, as were patients recorded on the PAS system following the daily check of ward lists, and discharged within one day prior to the next morning. Study wards were visited daily by the research pharmacist, and patients' drug charts, medical and nursing notes were reviewed for evidence of an ADR. Objective markers of ADRs, e.g. laboratory results, were identifiable from the patient notes and the hospital computer system, while subjective markers of ADRs, for example headache, nausea and rash were identified through patient notes, discussion with the ward team and, where appropriate, discussion with the affected patient. Clinical staff were informed that the study was taking place and could also refer directly either in person or through notification cards that were made available on the wards. The clinical ward pharmacists were consulted regularly regarding the possibility of ADRs on their designated wards. Following completion of the ward based data-collection period, retrospective case note analysis was performed to assess patient outcomes and to ensure that all available details regarding the ADR had been collected.

Suspected ADRs were classified in terms of causality [15] and avoidability [16] according to validated algorithms and assessed for suitability for Yellow Card Reporting to the Commission on Human Medicines and to the Medicines and Healthcare products Regulatory Agency (CHM/MHRA) [17]. ADRs were also classified as either type A (dose-dependent and predictable from the known pharmacology) or type B (idiosyncratic, no clear dose response relationship, and not predictable from the known pharmacology) according to the system introduced by Rawlins and Thompson in 1977 [18]. We chose this classification instead of the more recent DoTS classification [19] so that our data could be compared with previous studies. We also recorded severity of the ADRs according to the Hartwig severity scale [20], which was adapted from our pilot study [11], and subsequently modified to include two level 7 ADRs in order to differentiate between ADRs which directly, and those which indirectly, cause death.

Analysis for causality, avoidability, severity and suitability for yellow card reporting was done independently by investigators ED and CG. Overall there was approximately 60% agreement in the causality and avoidability assessments. Any discrepancies in scoring were then discussed before consensus was achieved between two investigators (ED and CG) in conjunction with the chief investigator (MP). Nine (1.2%) suspected ADRs were subsequently excluded by the investigating team, on the basis that an ADR was unlikely to be drug related according to the Naranjo algorithm [15]. The overall incidence of in-patient ADRs was defined as the total number of in-patient episodes which resulted in ADRs in relation to the total number of in-patient episodes in the study wards during the study period.

The length of stay for each patient was recorded using PAS data, enabling comparisons between patients with and without ADRs. Analysis of whether the ADR directly increased the length of stay, and the duration of this increase, was made following an assessment of the clinical features of the underlying disease and ADR, and after discussion with the ward team including the ward pharmacist and medical staff, and assessment of relevant case-notes. Clinical judgment was used to assess the additional length of stay attributable to the ADR. Thus, for example, if a patient had an ADR whilst waiting for nursing home placement, e.g. antibiotic-related C. difficile diarrhoea and the wait for placement independently exceeded the duration of the ADR, no additional length of stay was attributed to the reaction. Conversely, if a patient was ready for discharge, but an ADR occurred which required the patient to stay in hospital, the additional length of stay until recovery from the ADR was attributed to the reaction. All drugs including the causative drug(s) were recorded for all patients with ADRs. Given that there is no electronic system to capture information on medicines prescribed within our hospital, with the resources available, capturing information from all the patients would have led to incomplete and inaccurate information. We therefore decided to capture high quality data from a random sample of 1 in 10 in-patients on the same study wards.

ADRs which occurred despite specific prophylaxis against the ADR were recorded while the potential effect of polypharmacy on ADRs was measured by comparing the number of regular medicines taken by ADR patients on the first day of ADR with the number of regular medicines taken for the control sample (1 in 10 patients), assessed on the day of the in-patient stay where the patient received the maximum number of medicines. The most frequent ADRs relative to usage were calculated by using data of all drugs administered to one tenth of patients admitted. The frequency of the drug group causing a suspected ADR was divided by the number of times a drug in that class was administered in the control sample of patients (if≥1). The resulting ratio allowed drug groups to be further ranked by frequency of ADRs relative to drug use. The costs to the NHS were estimated using number of bed-days for additional length of stay based on the standard daily costs of NHS hospital episodes (£228) [21], consistent with the estimates used in our previous study [2].

Statistical methods

The data were hierarchically structured, in that multiple ADR episodes can occur both within patients with multiple admissions to hospital and within a particular patient admission, hence the study collected data at the patient, admission, and ADR episode level. To compare ADR incidence between hospital wards, a generalised estimating equation (GEE) [22] model with compound symmetry was used to account for the within-patient correlation. This was considered more appropriate than a random effects model when there are small numbers of observations within patients [22].

For all other analyses, where a patient had multiple admissions or multiple ADRs, we used a patient's first ADR episode and analysed at the patient level only. The first ADR episode was used to simply assess the affected patient population and their risk factors (age, gender, number of medicines and placement on a medical or surgical ward) were identical or assumed to be broadly similar for patients who had multiple admissions. Binary outcomes were compared between groups using the chi-square statistic for assessing significance. Comparisons between groups with respect to continuous measures were made using the t-test or the Mann-Whitney U-test, depending on skewness, for assessment of statistical significance. The 5% level was used for assessing significance.

Risk factors for ADRs were identified by investigating the effects of age, gender, number of drugs prescribed and placement on a medical or surgical ward, on the time to ADR. Regression analysis was undertaken via the Cox proportional hazards model. Results are given in terms of the hazard ratio (HR) for a particular factor with accompanying 95% confidence interval (95%CI). The 5% significance level was used when assessing factors for model inclusion. The risk factor ‘number of regular drugs prescribed’ (calculated on day of admission where patient was taking most regular medications for control sample patients) had data available for 10% of the total sample (n = 374), and therefore was only investigated in this sample.

Results

Over six months, there were a total of 3695 patient episodes assessed for ADRs involving 3322 patients. Out of these patient episodes, 545 (14.7%, 95% CI 13.6–15.9%) resulted in one or more ADRs. At the patient level, 524 of 3322 (15.8%) patients experienced at least one ADR. Women experienced significantly more ADRs (n = 308, 17.8%) than men (n = 216, 13.5%; χ2 = 11.6, df = 1, p<0.001). The median age was significantly higher in the ADR group at 72 years (Q1–Q3 56–81 years) compared with 61 years in the non-ADR group (Q1–Q3 41–77 years; U = 109, p<0.0001). More medical patient episodes (n = 406, 16.0%) than surgical episodes (n = 139, 12.0%) resulted in ADRs (χ2 = 10.12, df = 1, p<0.01). The incidence of ADR episodes varied further according to the specialty of the wards studied as shown in Table 1.The median length of stay for patient episodes that resulted in an ADR was 20 days (Q1–Q3 12–35 days), compared to 8 days (Q1–Q3 5–13 days; U = 138, p<0.0001) for those episodes without ADRs. Within the group of patients experiencing an ADR, the mortality was higher (n = 58, 10.7%), compared with 3.9% (n = 126) of patients who did not experience an ADR (χ2 = 42.4, df = 1, p<0.0001). ADRs contributed to 14 out of the 184 deaths (0.4% of patients admitted, 8.2% of all deaths), with one death (0.03% of patients admitted, 0.5% of all deaths) being directly attributable to the ADR, specifically GI bleed with diclofenac and dalteparin (see Table 2).

A total of 733 ADRs were identified. Type A ADRs accounted for 690 (94.1%) of the ADRs while 232 (31.7%) ADRs fulfilled the requirements for reporting to the UK regulatory agency. Drug-drug interactions were linked to 433 (59.1%) of the ADRs; the majority of these were pharmacodynamic in origin (91.7%), for example, the use of multiple diuretics in patients with renal failure; 5.3% were pharmacokinetic, and 3% had a mixed pharmacokinetic-pharmacodynamic mechanism. The interactions will be described in more detail elsewhere.

The majority (n = 602, 82.1%) of the ADRs occurred as a result of initiation of the causative drug in hospital, of which 390 (65%) showed clinical signs. Similarly, of the 131 (62%) ADRs where the drug had been initiated prior to hospital admission, with the ADR occurring during admission, patients showed clinical signs in 81 (62%) cases. Tables 3 and 4 show detailed results of causality, severity and avoidability assessments.

thumbnail
Table 3. The adapted Hartwig severity scale and corresponding adverse drug reaction (ADR) frequency.

https://doi.org/10.1371/journal.pone.0004439.t003

thumbnail
Table 4. Causality and avoidability assessments of the adverse drug reactions in hospital in-patients.

https://doi.org/10.1371/journal.pone.0004439.t004

In patient episodes associated with an ADR, the number of medicines taken was significantly higher (median 9, Q1–Q3 6–13) than in those episodes not associated with an ADR (median 6, Q1–Q3 4–10; U = 92644; p<0.0001). The drug groups most frequently implicated in the ADRs, and the frequency of their use in the study population are shown in Table 5. The most frequent causative drugs relative to usage were anticoagulants (warfarin), fibrinolytics (streptokinase) (4 ADRs), unfractionated heparin (3 ADRs), loop diuretics and allopurinol (5 ADRs).

thumbnail
Table 5. Drugs most frequently implicated in causing the adverse drug reactions (ADRs).

https://doi.org/10.1371/journal.pone.0004439.t005

ADRs occurred despite prophylaxis in 67 (9.1%) cases involving 10 types of ADR (constipation (35), electrolyte disturbances (10), renal failure (8), bleeding (5), raised INR (3), nausea (2), opioid withdrawal, opioid dependence, oral Candidal infection, and diarrhoea (all 1)). ADRs directly increased length of stay in 147 (27.0%) episodes, equating to 4.0% of all inpatient episodes and accounting for 934 out of 50145 (1.9%) bed days or 0.25 days/patient admission episode. For those episodes where length of stay was directly increased by an ADR, the median increase was 4 days (Q1–Q3, 2–7 days).

Table 6 gives results from multivariate risk factor analyses. The only significant predictor was the number of medicines (p<0.0001; HR 1.14; 95% CI 1.09, 1.20). Multivariable Cox regression confirmed these results, with the number of medicines as the only significant predictor. Therefore, on average, each additional medication multiplies the hazard of an ADR episode by 1.14. There may be a power issue in using the 10% sample, since the full dataset showed that both gender and age were significant risk factors for the ADR episode (p = 0.001 for both factors; HR (95% CI) 1.33 (1.12, 1.59) and 1.01 (1.0, 1.01) respectively). Comparing these with the 10% sample showed that the mean values from the full dataset were consistent with the 95% confidence intervals of the 10% sample. Furthermore, the 10% sample was representative of the whole sample set and unlikely to have introduced bias (Table 7).

thumbnail
Table 6. Risk factors for adverse drug reaction assessed by multivariate analysis.

https://doi.org/10.1371/journal.pone.0004439.t006

thumbnail
Table 7. Comparison of demographics of patients in 10% sample with the remainder of the study population.

https://doi.org/10.1371/journal.pone.0004439.t007

Discussion

This is the largest prospective study of adverse drug reactions in UK hospital in-patients. Our data suggest that at least 1 in 7 in-patient episodes is complicated by an adverse drug reaction. The incidence figure of 14.7% is consistent with our pilot study [11]. However, our figure is higher than the 3.5–7.3% incidence suggested in a systematic review [4]; this may be explained by the fact that pooling data from ADR studies with different designs can be problematical [23], [24] as illustrated by the widely differing estimates of ADR incidence determined in different studies in different populations (from 0.86% [25], to 37% [26]). In order to improve the accuracy of our assessments, individual causality assessments were undertaken using the Naranjo causality assessment tool [15]. There was 60% agreement in causality assessment and any disagreements were resolved by consensus discussions with a third assessor. Causality assessments are difficult, and inter-rater agreements vary enormously [27], but we feel that our methods were as robust as is possible in relation to assessment of individual cases. Thus, we feel that the prospective nature of this study, and the intensive nature of data collection and follow-up, similar in nature to two major recent studies of ADRs causing admission [3], [6] has resulted in, we believe, an accurate assessment of ADRs in adult hospital inpatients.

A clear limitation of our study is that it was conducted in one hospital and there is likely to be variation between different hospitals because of differences in the local population characteristics and the specialties within the hospitals. Nevertheless, we feel that the study population was selected from 12 wards which were representative of the clinical specialties (Table 1) commonly found in most acute hospitals and the age distribution of our study population was comparable to figures for all in-patient admissions in England [28]. Hospital episode statistics from the Department of Health in England state that in 2006–7 there were over 9 million adult admission episodes to NHS hospitals (excluding maternity admissions) [28]. There are 126,976 NHS beds in England, [29]. In our study, 1.9% of bed days were due to an ADR. Therefore we estimate that approximately 2000 bed days are due to an ADR at any one time, equivalent to approximately three 800-bed hospitals at 85% capacity. This figure, added to the estimate that the equivalent of seven 800-bed hospitals are filled with patients admitted with ADRs [3], suggests that ADR-related admissions and ADRs during hospitalization lead to the occupancy of ten 800-bed NHS hospitals. An accurate assessment of the financial cost of these ADRs is difficult, but a crude estimate based on an average cost of a bed day in the NHS in England suggests that the total costs are likely to exceed £171 million annually for ADRs occurring during admission. This is however likely to be an underestimate since the direct and indirect costs to patients such as loss of earnings due to extended stay or increased morbidity have not been measured, and neither were the costs which could be attributed to treating ADRs such as the prescribing of more medication and investigations, and involvement of clinical team external to the specialty to which the patient was admitted, all of which add to the overall ADR burden. As with our previous study [3], however, the figures provided here need to be interpreted cautiously as they represent an extrapolation from a single hospital to the NHS as a whole. However, taken together with our figure of £466 million for ADR-related admissions [3], we would estimate that ADRs cost the NHS in England in excess of £637 million annually, or approximately £5000 per hospital bed per year. This figure is comparable with adverse drug event (ADE) research from the United States ($8000 per hospital bed per year) [30], and with ADR research from mainland Europe (£4700 per hospital bed per year) [31].

Implicated drugs and severity of reactions

The most frequently implicated drugs were opioid analgesics, diuretics, systemic corticosteroids, anticoagulants and antibiotics. This is in accordance with several other studies of hospital inpatients [32][34]. When adjusted for the frequency of prescription, warfarin, fibrinolytics and unfractionated heparin were the top three causes of ADRs. It is worrying to note that the same drugs, warfarin, loop diuretics and opioids, are being consistently implicated in different studies of ADRs; this may partly reflect their high usage, but nevertheless suggest that lessons have not been learnt from previous studies, and effective preventive strategies have not been put in place.

Approximately three quarters of adverse drug reactions were scored at level 3 or below on the Hartwig scale (Table 3) and by definition required intervention. These interventions ranged from stopping the causative medicine(s) to administration of specific antidotes, for example, naloxone for opioid-induced respiratory depression. The remaining ADRs were sufficiently serious to result in an increase in length of stay or admission to intensive care, and in some cases, death. The assessment of whether an ADR has increased the length of stay or caused death, and in particular whether it is due to the underlying disease or due to an ADR, can be extremely difficult; in our study, we undertook careful assessment involving the clinical team whenever possible, and taking into account individual patient factors such as the nature and severity of the underlying disease, and social factors which may have contributed to the length of stay more significantly than the ADR itself. Our estimate therefore is likely to have been relatively conservative.

Prevention of adverse drug reactions

In this study, just over half of the ADRs were deemed possibly or definitely avoidable, which is consistent with the broad range of figures (30–70%) suggested in the literature [35], [36]. Given the considerable burden of ADRs, there is a need to put into place preventive strategies. Given the wide variety of drugs implicated, and the huge array of ADRs that we identified affecting almost every organ system in the body, prevention is likely to require complex multi-faceted intervention strategies. In our study, increasing age, admission to a medical ward, female gender, and number of regular medicines were identified as risk factors. Multivariable analysis of a representative sample of the data-set (Tables 6 and 7) showed that the only significant risk factor was the number of medicines the patient was taking, which may in itself be a reflection of age, gender and status as a medical patient. This is consistent with a number of previous studies [37][40]. Nevertheless, in order to confirm the results of our risk factor analysis it would be desirable to repeat the study in a second cohort. Given the increasing age of the population, and the trend towards polypharmacy, even in younger patients, the problem of ADRs is likely to remain a significant, if not increasing burden on our hospitals. Computerised prescribing and monitoring systems [41][43], the presence of pharmacists on ward rounds [44], [45], the need for better monitoring [46], and enhanced education of prescribing, leading to error reduction [47], are amongst the possible intervention strategies that have been suggested to be important in reducing the burden of ADRs. There is however a need for further research in this area, not only for the development of a robust evidence base to allow for prevention of ADRs, but also in the implementation of these strategies into hospital healthcare systems. Although it would be prudent to initially focus on the more serious ADRs, it is important to remember that even so-called non-serious ADRs, for example constipation from using opioids, can have a significant impact on the patient's quality of life, and also require the development of preventive strategies.

In conclusion, our study shows that ADRs are a significant problem in hospital inpatients, contributing to morbidity and mortality and resulting in considerable financial burden. Over half are definitely or potentially avoidable, and steps should be taken to introduce strategies to reduce their impact.

Acknowledgments

The authors wish to thank all healthcare professionals at the Royal Liverpool and Broadgreen University Hospitals Trust who helped with the study, and all the patients.

Author Contributions

Conceived and designed the experiments: ECD CFG DRM MP. Performed the experiments: ECD. Analyzed the data: ECD CFG ST PW DRM MP. Contributed reagents/materials/analysis tools: ST PW. Wrote the paper: ECD CFG ST PW DRM MP.

References

  1. 1. Lazarou J, Pomeranz BH, Corey PN (1998) Incidence of adverse drug reactions in hospitalized patients. A meta-analysis of prospective studies. JAMA 279: 1200–1205.J. LazarouBH PomeranzPN Corey1998Incidence of adverse drug reactions in hospitalized patients. A meta-analysis of prospective studies.JAMA27912001205
  2. 2. Wester K, Jonnson AK, Sigset O, Druid H, Hagg S (2008) Incidence of fatal adverse drug reactions: a population based study. Br J Clin Pharmacol 65: 573–579.K. WesterAK JonnsonO. SigsetH. DruidS. Hagg2008Incidence of fatal adverse drug reactions: a population based study.Br J Clin Pharmacol65573579
  3. 3. Pirmohamed M, James S, Meakin S, Green C, Scott AK, et al. (2004) Adverse drug reactions as a cause of admission to hospital: prospective analysis of 18 820 patients. BMJ 329: 15–19.M. PirmohamedS. JamesS. MeakinC. GreenAK Scott2004Adverse drug reactions as a cause of admission to hospital: prospective analysis of 18 820 patients.BMJ3291519
  4. 4. Wiffen P, Gill M, Edwards J, Moore A (2002) Adverse drug reactions in hospital patients. A systematic review of the prospective and retrospective studies. Bandolier Extra 1–16.P. WiffenM. GillJ. EdwardsA. Moore2002Adverse drug reactions in hospital patients. A systematic review of the prospective and retrospective studies.Bandolier Extra116
  5. 5. Winterstein AG, Sauer BC, Hepler CD, Poole C (2002) Preventable drug-related hospital admissions. Ann Pharmacother 36: 1238–48.AG WintersteinBC SauerCD HeplerC. Poole2002Preventable drug-related hospital admissions.Ann Pharmacother36123848
  6. 6. Howard RL, Avery AJ, Howard PD, Partridge M (2003) Investigation in to the reasons for preventable drug admissions to a medical admissions unit: observational study. Qual Saf Health Care 12: 280–285.RL HowardAJ AveryPD HowardM. Partridge2003Investigation in to the reasons for preventable drug admissions to a medical admissions unit: observational study.Qual Saf Health Care12280285
  7. 7. Seidl LG, Thornton GF, Cluff LE (1965) Epidemiological studies of adverse drug reactions. Am J Public Health 65: 1170–1175.LG SeidlGF ThorntonLE Cluff1965Epidemiological studies of adverse drug reactions.Am J Public Health6511701175
  8. 8. Smith JW, Seidl LG, Cluff LE (1966) Studies on the epidemiology of adverse drug reactions. V- Clinical factors influencing susceptibility. Ann Int Medicine 65: 629–639.JW SmithLG SeidlLE Cluff1966Studies on the epidemiology of adverse drug reactions. V- Clinical factors influencing susceptibility.Ann Int Medicine65629639
  9. 9. Hurwitz N, Wade OL (1969) Intensive hospital monitoring of adverse reactions to drugs. BMJ 1: 531–536.N. HurwitzOL Wade1969Intensive hospital monitoring of adverse reactions to drugs.BMJ1531536
  10. 10. Ogilvie RI, Ruedy J (1967) Adverse drug reactions during hospitalisation. Canadian Medical Association Journal 97: 1450–1457.RI OgilvieJ. Ruedy1967Adverse drug reactions during hospitalisation.Canadian Medical Association Journal9714501457
  11. 11. Davies EC, Green CF, Mottram DR, Pirmohamed M (2006) Adverse drug reactions in hospital in-patients: a pilot study. J Clin Phar Ther 31: 335–341.EC DaviesCF GreenDR MottramM. Pirmohamed2006Adverse drug reactions in hospital in-patients: a pilot study.J Clin Phar Ther31335341
  12. 12. Edwards IR, Aronson JK (2000) Adverse drug reactions: definitions, diagnosis and management. Lancet 356: 1255–9.IR EdwardsJK Aronson2000Adverse drug reactions: definitions, diagnosis and management.Lancet35612559
  13. 13. Association of the British Pharmaceutical Industry (2005) ABPI data sheet compendium. London: Datapharm Communications Ltd. Association of the British Pharmaceutical Industry2005ABPI data sheet compendiumLondonDatapharm Communications Ltd
  14. 14. British Medical Association, Royal Pharmaceutical Society of Great Britain (2005) British National Formulary. London: BMA, RPS, (No 49 and 50). British Medical Association, Royal Pharmaceutical Society of Great Britain2005British National FormularyLondonBMA, RPS, (No 49 and 50)
  15. 15. Naranjo CA, Busto U, Sellers EM, Sandor P, Ruiz I, et al. (1981) A method for estimating the probability of adverse drug reactions. Clin Pharmacol Ther 30: 239–45.CA NaranjoU. BustoEM SellersP. SandorI. Ruiz1981A method for estimating the probability of adverse drug reactions.Clin Pharmacol Ther3023945
  16. 16. Hallas J, Harvald B, Gram LF, Grodum E, Prosen K, et al. (1990) Drug related hospital admissions: the role of definitions and intensity of data collection and the possibility of prevention. J Int Med 228: 83–90.J. HallasB. HarvaldLF GramE. GrodumK. Prosen1990Drug related hospital admissions: the role of definitions and intensity of data collection and the possibility of prevention.J Int Med2288390
  17. 17. Medicines and Healthcare Products Regulatory Authority. Medicines and Healthcare Products Regulatory Authority.What to report. http://www.mhra.gov.uk/home/idcplg?IdcService=SS_GET_PAGE&nodeId=750. Accessed 2008 Jan 8. What to report. http://www.mhra.gov.uk/home/idcplg?IdcService=SS_GET_PAGE&nodeId=750. Accessed 2008 Jan 8.
  18. 18. Rawlins MD, Thompson JW (1977) Pathogenesis of adverse drug reactions. In: Davies DM, editor. Textbook of adverse drug reactions. Oxford: Oxford University Press. MD RawlinsJW Thompson1977Pathogenesis of adverse drug reactions.DM DaviesTextbook of adverse drug reactionsOxfordOxford University Press10
  19. 19. Aronson JK, Ferner RE (2003) Joining the DoTS: a new approach to classifying adverse drug reactions. BMJ 327: 1222–5.JK AronsonRE Ferner2003Joining the DoTS: a new approach to classifying adverse drug reactions.BMJ32712225
  20. 20. Hartwig SC, Siegel J, Schneider PJ (1992) Preventability and severity assessment in reporting adverse drug reactions. Am J Hosp Pharm 49: 2229–32.SC HartwigJ. SiegelPJ Schneider1992Preventability and severity assessment in reporting adverse drug reactions.Am J Hosp Pharm49222932
  21. 21. Chartered Institute of Public Finance and Accountancy (CIPFA) (2002) The health service financial database and comparative tool. Croydon: Institute of Public Finance. Chartered Institute of Public Finance and Accountancy (CIPFA)2002The health service financial database and comparative toolCroydonInstitute of Public Finance
  22. 22. Brown H, Prescott R (1999) Applied Mixed Models in Medicine. London: Wiley. H. BrownR. Prescott1999Applied Mixed Models in MedicineLondonWiley
  23. 23. Kvasz M, Allen E, Gordon MJ, Ro EY, Estok R, et al. (2000) Adverse drug reactions in hospitalised patients: A critique of a meta-analysis. Medscape Gen Med 2: E3.M. KvaszE. AllenMJ GordonEY RoR. Estok2000Adverse drug reactions in hospitalised patients: A critique of a meta-analysis.Medscape Gen Med2E3
  24. 24. Davies EC, Green CF, Mottram DR, Pirmohamed M (2007) Adverse drug reactions in hospitals: A narrative review. Current Drug Safety 2: 79–87.EC DaviesCF GreenDR MottramM. Pirmohamed2007Adverse drug reactions in hospitals: A narrative review.Current Drug Safety27987
  25. 25. Simmons C, Georgeson EM, Hill RC (1998) Adverse drug reactions: can we reduce the risk? Hosp Pharm 33: 1568–1576.C. SimmonsEM GeorgesonRC Hill1998Adverse drug reactions: can we reduce the risk?Hosp Pharm3315681576
  26. 26. Van Kraaij Dj, Haagsma CJ, Go IH, Gribnau FW (1994) Drug use and adverse drug reactions in 105 elderly patients admitted to a general medical ward. Neth J Med 44(5): 166–73.Dj Van KraaijCJ HaagsmaIH GoFW Gribnau1994Drug use and adverse drug reactions in 105 elderly patients admitted to a general medical ward.Neth J Med44(5)16673
  27. 27. Agbabiaki TB, Savovic J, Ernst E (2008) Methods for causality assessment of adverse drug reactions. Drug Safety 31(1): 21–37.TB AgbabiakiJ. SavovicE. Ernst2008Methods for causality assessment of adverse drug reactions.Drug Safety31(1)2137
  28. 28. Department of Health, England (2008) Department of Health Headline statistics 2006–07. Department of Health, England2008Department of Health Headline statistics 2006–07.http://www.hesonline.nhs.uk/Ease/servlet/ContentServer?siteID=1937&categoryID=193. Accessed 2008 Jan 8. http://www.hesonline.nhs.uk/Ease/servlet/ContentServer?siteID=1937&categoryID=193. Accessed 2008 Jan 8.
  29. 29. Department of Health, England (2008) Department of Health, England2008Average daily number of available and occupied beds by sector, England, 2006–07 http://www.performance.doh.gov.uk/hospitalactivity/data_requests/download/beds_open_overnight/bed_07_Summary.xls. Accessed 2008 Jan 8. Average daily number of available and occupied beds by sector, England, 2006–07 http://www.performance.doh.gov.uk/hospitalactivity/data_requests/download/beds_open_overnight/bed_07_Summary.xls. Accessed 2008 Jan 8.
  30. 30. Bates DW, Spell N, Cullen DJ, Burdick E, Laird N, et al. (1997) The costs of adverse drug events in hospitalized patients. JAMA 277: 307–311.DW BatesN. SpellDJ CullenE. BurdickN. Laird1997The costs of adverse drug events in hospitalized patients.JAMA277307311
  31. 31. Moore N, Lecointre D, Noblet C, Mabille M (1998) Frequency and cost of serious adverse drug reactions in a department of general medicine. Br J Clin Pharmacol 45: 301–308.N. MooreD. LecointreC. NobletM. Mabille1998Frequency and cost of serious adverse drug reactions in a department of general medicine.Br J Clin Pharmacol45301308
  32. 32. Leach S, Roy SS (1986) Adverse drug reactions: an investigation on an acute geriatric ward. Age and Ageing 15: 241–246.S. LeachSS Roy1986Adverse drug reactions: an investigation on an acute geriatric ward.Age and Ageing15241246
  33. 33. Classen DC, Pestotnik SL, Scott Evans R, Lloyd JF, Burke JP (1997) Adverse drug reactions in hospitalized patients. JAMA 277: 301–306.DC ClassenSL PestotnikR. Scott EvansJF LloydJP Burke1997Adverse drug reactions in hospitalized patients.JAMA277301306
  34. 34. Suh D-C, Woodall BS, Shin S-K, Hermes-De Santis (2000) Clinical and economic impact of adverse drug reactions in hospitalized patients. Ann Pharmacother 34: 1373–9.D-C SuhBS WoodallS-K ShinSantis Hermes-De2000Clinical and economic impact of adverse drug reactions in hospitalized patients.Ann Pharmacother3413739
  35. 35. Bates DW, Cullen DJ, Laird N, Peterson LA, Small AD, et al. (1995) Incidence of adverse drug events and potential adverse drug events. Implications for prevention. JAMA 274: 29–34.DW BatesDJ CullenN. LairdLA PetersonAD Small1995Incidence of adverse drug events and potential adverse drug events. Implications for prevention.JAMA2742934
  36. 36. Ducharme MM, Boothby LA (2007) Analysis of adverse drug reactions for preventability. Int J Clin Pract 61(1): 157–61.MM DucharmeLA Boothby2007Analysis of adverse drug reactions for preventability.Int J Clin Pract61(1)15761
  37. 37. Hoigne R, Lawson DH, Weber E (1990) Risk factors for adverse drug reactions – epidemiological approaches. Eur J Clin Pharmacol 39: 321–325.R. HoigneDH LawsonE. Weber1990Risk factors for adverse drug reactions – epidemiological approaches.Eur J Clin Pharmacol39321325
  38. 38. Carbonin P, Pahor M, Bernabei R, Sgadari A (1991) Is age an independent risk factor of adverse drug reactions in hospitalised patients? J Am Geriatr Soc 39: 1093–1099.P. CarboninM. PahorR. BernabeiA. Sgadari1991Is age an independent risk factor of adverse drug reactions in hospitalised patients?J Am Geriatr Soc3910931099
  39. 39. van den Bemt PMLA, Egberts ACG, Lenderink AW, Verzijl JM, Simons KA, et al. (2000) Risk factors for the development of adverse drug events in hospitalised patients. Pharm World Sci 22: 62–66.PMLA van den BemtACG EgbertsAW LenderinkJM VerzijlKA Simons2000Risk factors for the development of adverse drug events in hospitalised patients.Pharm World Sci226266
  40. 40. Camargo AL, Cardoso Ferreira MB, Heineck I (2006) Adverse drug reactions: a cohort study in internal medicine units at a university hospital. Eur J Clin Pharmacol 62: 143–149.AL CamargoMB Cardoso FerreiraI. Heineck2006Adverse drug reactions: a cohort study in internal medicine units at a university hospital.Eur J Clin Pharmacol62143149
  41. 41. Evans RS, Pestotnik SL, Classen DC, Horn SD, Bass SB, et al. (1994) Preventing adverse drug events in hospitalized patients. Ann Pharmacother 28: 523–7.RS EvansSL PestotnikDC ClassenSD HornSB Bass1994Preventing adverse drug events in hospitalized patients.Ann Pharmacother285237
  42. 42. Raschke RA, Gollihare B, Wunderlich TA, Guidry JR, Leibowitz AI (1998) A computer alert system to prevent injury from adverse events. JAMA 280(15): 1317–20.RA RaschkeB. GollihareTA WunderlichJR GuidryAI Leibowitz1998A computer alert system to prevent injury from adverse events.JAMA280(15)131720
  43. 43. Dormann H, Criegee-Rieck M, Neubert A, Egger T, Levy M, et al. (2004) Implementation of a computer-assisted monitoring system for the detection of adverse drug reactions in gastrenteroloy. Ailment Pharmacol Ther 19: 3039.H. DormannM. Criegee-RieckA. NeubertT. EggerM. Levy2004Implementation of a computer-assisted monitoring system for the detection of adverse drug reactions in gastrenteroloy.Ailment Pharmacol Ther193039
  44. 44. Kucukarslan SN, Peters M, Mlynarek M, Nafziger DA (2003) Pharmacists on rounding teams reduce preventable adverse drug events in hospital general medicine units. Arch Intern Med 163: 2014–2018.SN KucukarslanM. PetersM. MlynarekDA Nafziger2003Pharmacists on rounding teams reduce preventable adverse drug events in hospital general medicine units.Arch Intern Med16320142018
  45. 45. Leape LL, Cullen DJ, Dempsey Clapp M, Burdick E, Demonaco HJ, et al. (1999) Pharmacist participation on physician rounds and adverse drug events in the intensive care unit. JAMA 282: 267–70.LL LeapeDJ CullenM. Dempsey ClappE. BurdickHJ Demonaco1999Pharmacist participation on physician rounds and adverse drug events in the intensive care unit.JAMA28226770
  46. 46. Pirmohamed M, Ferner RE (2003) Monitoring drug treatment. BMJ 327: 1179–1181.M. PirmohamedRE Ferner2003Monitoring drug treatment.BMJ32711791181
  47. 47. Dean B, Schachter M, Vincent C, Barber N (2002) Causes of prescribing errors in hospital inpatients: a prospective study. Lancet 359: 1373–8.B. DeanM. SchachterC. VincentN. Barber2002Causes of prescribing errors in hospital inpatients: a prospective study.Lancet35913738