Background

A substantial proportion of hospitalised patients experience medication-related harm that is preventable [1]. Drug errors have been estimated to account for over a quarter of causes of adverse drug events [2]. Strategies to prevent such problems are being developed. One such strategy is the structured review of patient medication by pharmacists to identify patients with medication errors that may lead to harm. The advantage is that the complete clinical status of each patient is taken into account when identifying problems [3]. In a study, clinical pharmacists performed better than the decision support system in identifying drug–drug interactions clinical interventions [4]. The authors realised that clinical pharmacists looked at individual administration intervals and drug sequence to determine the clinical relevance of the interactions.

The Harvard Medical Practice Study analysed error recovery, i.e. the circumstances under which errors were detected and corrected. The study acknowledged the important role pharmacists play in identifying and correcting drug errors from reaching patients.

The Institute of Medicine’s report includes recommendations for health systems to implement error reporting reviews [5]. In the inpatient setting, clinicians have used various drug event reporting programs to better understand and prevent drug errors [6, 7]. The interception of medication errors is thus essential for improving patient safety. The aim of the study was therefore to evaluate the clinical interventions of hospital pharmacists.

Methods

The study was a retrospective review of reports, which had no names, nor personal information of patients. No consent was taken from patients since they could not be traced to any reports.

Study setting

The study took place at Korle Bu Teaching hospital which is a 2000 bed tertiary teaching hospital located in the capital city of Ghana. At the time of the study, the hospital had about 80 pharmacists. The main pharmacy services provided in the hospital were dispensing, clinical, drug information, research and small scale manufacturing. There were about 30 pharmacists who actively undertake clinical duties across the various wards of the hospital.

Data collection

To evaluate the clinical interventions of pharmacists working in the hospital, copies of reports for the period January 2011–December 2013 were made and relevant data extracted using a specially designed data collection sheet. Pharmacists had previously identified drug errors and manually reported the clinical interventions. The pharmacist interventions accepted and the actions taken were reported on the submitted manual reports. Pharmacists had discovered drug errors during their normal duties from review of patient medical records, laboratory reports, interactions with other health care professionals, patients, caregivers or family members.

Data analysis

The extracted clinical intervention data was entered into and analysed using SPSS (Statistical Package for Social Sciences) version 16 for Windows. Descriptive analysis was performed on all the data to obtain the frequency of clinical interventions, drug characteristics and pharmacist characteristics. Aggregate data were tabulated and summarized using frequency statistics such as count, range, mean and standard deviation. Descriptive analyses of all drug error types and related interventions were also tabulated. Pharmacist Clinical intervention data were compared between drug classifications, drug error types and whether pharmacist interventions were accepted or not using Chi square test dichotomous variables. Any p < 0.05 was considered as statistically significant.

Results

Study participants

The evaluation revealed that 24 pharmacists made 529 paper-based reports over the 3 years. Majority of them were female (70.8%) and more than half had less than 10 years experience (53.3%). The basic characteristics of pharmacists who made the reports are presented in Table 1.

Table 1 Characteristics of pharmacists involved in reporting (N = 24)
Full size table

Drug error reports

Of the 529 paper-based drug error reports, 448 contained complete information and hence were included in the study. Reasons for not including the 79 were no drug name (n = 67), no reason for error (n = 6) and no recommendation (n = 6). Drug errors were reported from all the units of the hospital that pharmacists worked; surgery (24%), medicine (22%), paediatric (21%), obstetrics and gynaecology (17%) and others (16%). Pharmacists discovered drug errors from review of patient medical records (74%), from other health care professionals (10%), laboratory reports (8%), patients (6%), caregivers (1%), and other unspecified sources (3%). The frequently occurring therapeutic drug categories with errors were cardiovascular (44.4%), infections (22.8%), nutrition (12.9%) and musculoskeletal (6.6%). Table 2 describes the therapeutic drug categories and the degree of acceptance. The five most frequently reported classes of drugs associated with drug errors were antibiotics (20.2%), anticoagulants (19.9), iron supplement (16.3%), diuretics (9.4%) and non-steroidal anti-inflammatory drugs (4.4%). The most frequently reported drugs with error were warfarin (9.5%), potassium chloride (6.0%) and potassium citrate (5.5%). The drug error types identified were categorised as prescribing, dispensing/implementing, administering/patient receiving and monitoring (see Table 3). Majority of reported drug errors were due to prescribing (70.9%) and least due to dispensing/implementing (2.0%). The most frequently reported drugs associated with prescribing errors included cardiovascular (42.6%), anti-infectives (22.9), and nutritional agents (10.5%). During dispensing or implementation, the frequently reported drug errors were anti-infectives (50.0%), endocrine (35.0%) and cardiovascular (15.0%). The most frequently reported drug category associated with administration or patient receiving included anti-infectives (45.5%), cardiovascular (22.7%) and central nervous system (15.2%). The common reasons pharmacists provided for drug errors included untreated indication (18.9%), wrong dose prescribed (12.5%), wrong drug prescribed (11.4%), medicine interactions (10.7%) and duplication of therapy (9.8%).

Table 2 Therapeutic category of drugs with intervention reports
Full size table
Table 3 Drug error types and reasons
Full size table

Clinical interventions

The twenty most frequently drugs occurring in intervention reports and their potential risk are presented in Table 4. The pharmacists made 1019 interventions and recommendations in 448 handwritten reports. The average intervention per report was 2.5, standard deviation (±0.67), range (2–4), and mode (2). The interventions and recommendations made have been categorised as drug regimen change (76.1%), monitoring required (13.0%), communication (5.4%), counselling required (5.0%) and adverse drug reporting (0.6%). The intervention types have been summarised in Table 5. Monitoring-required based interventions were significantly more likely to be accepted (130 vs 38; p <0.0001). Drugs involving drug regimen adjustment by pharmacists included potassium citrate (n = 56), enoxaparin (n = 54), warfarin (n = 42), diclofenac (n = 40), and morphine (n = 37). Monitoring required interventions were made for potassium chloride (n = 46), frusemide (n = 22), warfarin (n = 20), gentamicin (n = 19) and metolazone (n = 14). Drugs requiring counselling included warfarin (n = 44), iron supplement (n = 12), inhaled steroid (n = 6), insulin (n = 4), and lamivudine (n = 3). Drugs involving communication between pharmacist and other healthcare professionals included frusemide (n = 31), diclofenac (n = 18), iron supplement (n = 16), warfarin (n = 11) and antacid (n = 7). Majority (90.5%) of the recommendations and interventions made by pharmacists were accepted by prescribers and other healthcare professionals (see Table 5). These interventions were communicated via the following means: verbal (76.4%), write in patient medical notes (16.3%), acted upon by reporting pharmacist (6.1%), prepare formal note (0.7%) and prescribe/procure for patient (0.5%).

Table 4 Drugs most frequently occurring in intervention reports and their potential risk
Full size table
Table 5 Types of pharmacist clinical interventions and degree of acceptance
Full size table

Discussion

This part of the study evaluated the clinical intervention reports submitted by pharmacists working in a tertiary hospital. The pharmacists identified drug related problems in the management of patients and made interventions to prevent these errors from reaching patients. Twenty-four pharmacists made 1019 clinical interventions in 448 handwritten reports. Majority of the interventions related to drug therapy changes. Though this study evaluated handwritten reports, it is comparable to evaluations done on electronic incident reports [8, 9].

The categories of drugs most often associated with drug error reports were similar to those reported from previous studies and included cardiovascular agents [10, 11], anti-infectives [11, 12], and central nervous system agents [10, 13], suggesting that future strategies for reducing drug errors could target these agents. This study also found challenges with the use of nutritional supplements. Most of the challenges with nutritional supplements had to do with untreated anaemia, which physicians had overlooked. Iron deficiency anaemia is a serious nutritional problem in developing countries given its impact on increased mortality or serious morbidity in patients [14].

The frequently reported drug was warfarin as found in other studies [15]. The use of warfarin presents substantial safety concerns for patients. Adverse events associated with warfarin therapy are common [16]. This will require prevention strategies targeted at the prescribing and monitoring stages of warfarin management.

Though this study concentrated on pharmacists identifying inpatient drug errors as in other studies, results are comparable with studies conducted in outpatients [11, 17–20]. The drug errors assessed in this study were reports from only pharmacists although physicians [11, 21], nurses [22, 23] and others [24, 25] had reported drug errors in other studies.

The most frequently reported drug errors found in our study were drug regimen change and originated from drug prescribing. This finding is consistent with findings from other studies conducted in clinical centres [26], tertiary [27], hospital inpatient [28] and ambulatory care settings [11]. Inappropriate prescribing predicts the risk of adverse drug events [29]. The most commonly reported prescribing drug error was untreated indication. This was followed by prescribing wrong dose as seen in other studies [11, 30]. Children are particularly at risk of wrong dose errors [31].

The top two most frequently reported drug type associated with prescribing, dispensing and administration errors were cardiovascular and anti-infective agents. The most commonly reported dispensing error was dispensing wrong drug. Previous studies have reported dispensing wrong drugs in all types of inpatient settings [25, 32–35]. Omission due to drug unavailability was also a common error identified. Drug unavailability is common and poses a major challenge to healthcare systems in transitional and developing countries [36, 37]. In addition to drug omissions, administering wrong drug followed by duplication were common reported administration errors. The most commonly reported monitoring error was omitting relevant laboratory test. Monitoring errors had been previously reported [11].

More than 90% of interventions and recommendations by pharmacists were accepted and implemented. Over 70% of the interventions involved drug regimen change. Studies have reported prescribing errors as a major contributor to patient harm in hospitals [10, 26, 28, 29, 38]. Most (76%) of the interventions were communicated verbally. This would require an operational collaborative working relationship between pharmacists and other healthcare professionals to enhance patient care [39]. Previous studies reported that pharmacists in a collaborative team in hospitals helped reduce adverse drug events by 30–86% [19, 40–42]. Pharmacists’ close proximity with physicians provides opportunity for timely verbal communications on error interceptions. Some of the potential risk prevented by pharmacists included bleeding, anaemia, nephrotoxicity, electrolyte imbalance, severe diarrhoea etc.

Moreover, other interventions by pharmacists in this study related to patient counselling. It has been reported that patient counselling prevents adverse drug events during and after hospitalisation [43].

The study had some limitations. The study evaluated voluntary incident reports. Voluntary reports could be underreported and did not provide the actual frequency of the total denominator of errors that were detected and ameliorated. Secondly, reports did not contain information on whether errors reached patients and the effects of any harm. Moreover, the study could not assess the outcomes of the interventions performed by pharmacist.

Conclusions

Hospital pharmacists identified a lot of drug use challenges and subsequently prevented errors from reaching patients. Majority of pharmacist interventions and recommendations to prevent or ameliorate drug errors were accepted and implemented.