Safeguarding the process of drug administration with an emphasis on electronic support tools

Abstract

Aims

The aim of this work is to understand the process of drug administration and identify points in the workflow that resulted in interventions by clinical information systems in order to improve patient safety.

Methods

To identify a generic way to structure the drug administration process we performed peer-group discussions and supplemented these discussions with a literature search for studies reporting errors in drug administration and strategies for their prevention.

Results

We concluded that the drug administration process might consist of up to 11 sub-steps, which can be grouped into the four sub-processes of preparation, personalization, application and follow-up. Errors in drug handling and administration are diverse and frequent and in many cases not caused by the patient him/herself, but by family members or nurses. Accordingly, different prevention strategies have been set in place with relatively few approaches involving e-health technology.

Conclusions

A generic structuring of the administration process and particular error-prone sub-steps may facilitate the allocation of prevention strategies and help to identify research gaps.

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT

• Similar to prescription errors also drug administration errors contribute to a large fraction of preventable medication errors and adverse drug events.

• Drug administration can be exceedingly complex with regard to dose form, route of administration, number of co-medications, and setting and accordingly, error rates may vary.

WHAT THIS STUDY ADDS

• The drug-specific drug administration process can be generically structured for distinct dosage forms to allow identification of particularly error prone steps and allocated prevention strategies.

• Potential research gaps in error epidemiology and prevention are highlighted.

Introduction

Drug administration as a relevant part of the drug treatment process

In contrast to drug prescription, which mostly lies in the hands of health care personnel, drug administration is everyday practice for almost any human [1]. Hence, in ambulatory care, only a small fraction of drugs are administered by trained personnel such as mobile nursing services, while most drugs are actually administered by patients, family members, or even teachers and school secretaries, all of whom are untrained and have no medical background [2, 3].

Drug administration is an inherent part of everyday life for both adults and children and when asked, every second child (50.8% of 0–17 year olds) [4] and more than two-thirds of adults (71.5% of 18–79 year olds) in Germany stated that they administered drugs during the previous week (data from 2003–2006 [5]). More than 50% of adults administer drugs daily [6]. This fact, however, does not automatically make drug administration a safe and straightforward process and, indeed, drug administration errors are frequent and in the inpatient setting, roughly 30% of errors resulting in adverse drug events (ADE) happen during drug administration [7]. Comparably susceptible to errors is the prescription process [7]. Whereas for the prescribing process electronic prescription platforms with enhanced clinical decision support tools were implemented as promising error prevention strategy [8], similar and theory-driven approaches are only scarcely available for the administration process [9], even though introduction of electronic support in the drug administration process has shown to reduce error rates [10].

Drug treatment is a rather complex and demanding task (Figure 1) and depending on the setting and co-medication a great number of different errors may occur. The aim of this study was to structure the drug administration process (process 5 in Figure 1) in a way that allowed the allocation of current knowledge on frequent sources of errors as well as successful methods to prevent such errors with a particular emphasis on electronic tools. Concurrently, known unexplored sources of errors in the administration process are described to highlight critical, error prone steps and promising measures of error prevention. Therefore this paper should encourage research in the many areas of uncertainty, help generate knowledge of the administration process and ultimately stimulate system changes to eliminate the very numerous errors in this field.

Figure 1.

Whenever a drug is prescribed, 10 processes must be carried out in a well-organized sequence to make treatment successful. Each of these processes makes special demands, may be diverse and complex and may be flawed by a range of errors thus requiring rigorous quality management. After selection of an appropriate drug and its prescription (1), the drug has to be dispensed (2) and the patient or health care provider has to be informed about its proper use (3). Then the patient has to be motivated to adhere to this treatment regimen (4) and to ultimately perform a more or less complex sequence of preparation and administration steps (5). In the subsequent pharmaceutical process (6) the drug has to be released from the formulation to get absorbed, distributed and ultimately eliminated (7). Only if the drug reaches the target compartment will it elicit its pharmacodynamic effect (8), which will ultimately produce the intended therapeutic response (or adverse events) (9). Finally, each drug therapy should be monitored appropriately (10) to differentiate between success, excessive exposure and nonresponse

Methods

Description of the drug administration process

To describe the drug administration process, we discussed in peer groups consisting of clinical pharmacists and a clinical pharmacologist drug administration processes for relevant drug forms and routes of administration and described evident sub-steps in detail. Based on this specification we could identify similarities and differences. In order to check whether we missed important steps, we screened the literature on drug administration errors for sub-steps that reportedly trigger administration errors (Medline search term: (‘Drug Administration Routes/instrumentation'[Mesh] OR ‘Drug Administration Routes/methods’[Mesh] OR ‘medication administration’) AND (‘Medication Errors’[Mesh] OR ‘drug administration error'[all fields]) (last search performed on 6 March 2013).

Within the process of drug administration, we determined that a correct drug administration process would not only require the appropriate preparation of the correct single drug dose, but also the consideration of time, route of administration, administration technique, context factors and appropriate monitoring (Table 1). Even if the distinct application step was performed correctly, drug administration could be erroneous in the context of the individual situation. A typical example is the simultaneous administration of infusions that are pharmaceutically not compatible and will therefore precipitate [11–13]. While we aimed to describe a drug-centred administration process, there are also setting-specific constraints (e.g. dispensing or picking errors in the pharmacy as well as selection of the wrong patient or faulty documentation of the administration process) that may be considered as administration errors and are therefore commonly reported in observational trials. Moreover, we put the main emphasis of this study on regular drug administration and we did not specifically consider particular drugs that require well-defined application techniques and have standardized or particular constraints, such as chemotherapeutics.

Table 1.

Process steps in drug administration, sources and frequencies of errors and options for error prevention

Epidemiology of drug administration errors and options for prevention

We employed the above mentioned literature search to identify the frequency and nature of drug administration errors. We aimed to describe the epidemiology of drug administration errors according to the above mentioned scheme of relevant sub-steps in drug administration (without focusing on specific indications but only on specific dosage forms) and allocated prevention strategies where appropriate. Hence, this approach would regroup different types of errors such as wrong dose errors, wrong time errors, omission and commission errors which served in many earlier studies as error categories [14].

To identify solutions based on electronic information and technology systems for prevention of administration errors we performed a specific search (‘Drug Administration Routes’[Mesh] OR ‘self administration'[Mesh] OR ‘medication administration’) AND ‘Medication Errors’[Mesh] AND ‘Decision Support Systems, Clinical'[Mesh]) (last search performed on 6 March 2013).

Results

Description of the drug administration process

We analyzed drug administration processes for frequent (according to [15]) or specific dosage forms, i.e. solid and liquid oral drugs, inhalers, eye drops, nasal sprays, ointments, transdermal systems, suppositories, infusions and injections. As a result, we concluded that the drug administration process might consist of up to 11 drug-related sub-steps that can be grouped into the four sub-processes of preparation, personalization, application and follow-up and that are framed by three setting-related sub-steps (i.e. delivery of drug, identification of patient (before the drug-related administration process) and documentation (after drug-related administration process) (Figure 2, Table 1)). Hence, this description of the drug administration process complements the error- and process-related description of the administration process with the nine Rs [16].

Figure 2.

Depiction of different variants of process sequence in the drug administration process. Described are different drug dose forms and aligned medical aids. Dotted lines are sub-steps that are performed automatically or are not required

Epidemiology of drug administration errors and options for prevention

Errors in drug handling and administration are diverse and frequent (Table 1) and in many cases not caused by the patient him/herself, but by family members [17] or the professional provider administering the drug [11–13, 17–19]. Accordingly, different prevention strategies have been set in place (Table 1). A substantial fraction of errors ultimately leading to flaws in the administration process is not linked to different steps of the administration process itself but rather to setting-related constraints such as the identification of the right patient [20, 21] or the documentation of drug administration (responsible for 13% of all errors classified as administration errors in one study [20]), or the administration at the right time. Most of the e-health technologies focus on these setting-related constraints, i.e. barcoding patients, automated dispensing systems and medication administration records.

Error frequency might vary along with the medication error detection method. In the inpatient setting, the most common technique is direct observation, where a third person, openly or disguised, observes the process of drug administration and documents important steps by using checklists [18, 22]. Direct observation can both reduce and increase error rates. On the one hand, observation can lead to increased attention of the person being observed (‘Hawthorne’ effect) and on the other, it can lead to increased nervousness and insecurity. However, while Hawthorne effects have been described [23], these effects seem to be reducible if the observation spans across a longer period of time [24]. Direct observation can also be employed with patients or family members to detect administration errors when using inhalers [25–28], insulin pens [29], eye drops [30, 31], parenteral drugs [17, 18, 32, 33] or per oral drugs [17], for instance. Such assessments form the basis to describe the epidemiology of administration errors as well as other influencing variables and are essential to describe the benefit of potential error prevention strategies.

Discussion

When looking at the entire drug treatment process, early studies suggested that in inpatient care, one in three medication errors was related to the drug administration process [7]. At this time, there are few corresponding studies for the ambulatory sector. However in one paediatric study, 70% of preventable ADEs were related to drug administration by parents [34]. Given the frequency of drug administration, the complexity and multitude of required sub-steps, and the number of involved people some of whom are not trained in drug administration, it seems reasonable to assume that drug administration errors might even be far more prominent in the ambulatory setting than in the hospital setting. Moreover, in contrast to the prescription process, the administration process is the last step before the drug actually reaches the patient. Whereas drug prescription errors can be detected and corrected during the course of the treatment process (near-misses), drug administration errors are costly to detect, difficult to intercept, and may, therefore, have a high risk of reaching the patient and translating into an adverse patient outcome. Indeed, flaws and deficits in each sub-step of drug administration might entail adverse outcomes, while different outcomes are typically linked to distinct sub-steps. For instance, problems in correct identification of drug packages are likely to be associated with decreased adherence [35], as are difficulties in handling drugs during the personalization process (e.g. tablet splitting [36, 37]), whereas errors in drug dosage and drug application are related to non-response [38] or toxicity [39, 40]. Hence, the prevention of drug administration errors appears important, particularly in the ambulatory care setting, where the incidence rate might be higher and subsequent strategies to counteract adverse outcomes following administration errors might be less accessible. To minimize sources of error in drug administration, their identification and specification is crucial. Hence, various efforts were made to render the general classification of ‘prescription and administration errors’ more precise [14, 41] and develop the basis for a targeted implementation of error prevention strategies. Both explorative and qualitative assessments [42] as well as structured (semi-) theoretical approaches such as failure mode and effects analysis (FMEA) [43] with a specific emphasis on reliability [44] have been employed to describe drug administration processes. The more precisely the administration process is divided into its sub-steps, the easier observation techniques can be employed to assess drug administration [18] and the effect of intervention strategies. This will also form the basis for targeted interventions aimed at improving such deficits. Generally, administration errors can either result from treatment- or process-related faults (e.g. picking the wrong drug or treating the wrong patient) or from drug-specific challenges (e.g. crushing a sustained-release tablet). Hence, approaches to reduce administration errors may focus on the drug itself, the entire process in a given setting as well as the individual patient or a specific patient population. In either case, the error source (e.g. slips or oversights, gaps in knowledge, violations of established rules, lack of skills or false beliefs) will determine the most appropriate set of prevention strategy.

To date, most interventions have aimed at reducing the overall rate of administration errors without focusing specifically on critical sub-steps. Such global measures were mainly intended to optimize or disentangle the entire process by improving the surrounding conditions (e.g. reducing disturbances) [45], introducing in-process controls [46], or selecting [29] or modifying [47] medical aids and appliances. These should, however, be chosen individually and integrated carefully to reduce errors successfully [26, 27, 48–51]. If the error is not linked to a particular step of the drug-specific administration but rather to the setting-specific treatment process (e.g. selection of the wrong patient), the implementation of barcoding has shown benefits both in reducing dispensing errors [52] as well as administration errors, particularly if the patient was also scanned before drug administration [20, 21]. This benefit is not seen in all studies [53] and is potentially more evident if barcoding is linked to electronic medication administration records [54]. Moreover, a simplification of the entire process by reducing or automating the necessary sub-steps can reduce error rates [55] or avoid errors altogether. For instance, pre-filled multidose inhalers [25] or pre-filled insulin pens [28], which can be more easily administered, show lower error rates than single dose inhalers or refill pens. Moreover, drug adherence is better with insulin pens than with syringes and ampoules [56].

Electronic information and technology systems have only been scarcely evaluated as approaches for error prevention within the distinct drug administration process and focus most often on the administration of intravenous drugs (smart pumps). However, there are a variety of electronic tools that have been assessed as reminder or alert systems before drug administration actually takes place with varying results. In one study, complete oblivion could be decreased by reminders via voice mail [57]. Text messages showed both higher persistence of adherence rates [58] as well as no impact in another patient population where text messages did not improve adherence [59]. Reminding systems might be particularly successful, if they consider context factors and issue dynamic alerts [60]. For ambulatory patients, complete medication management systems that are supposed to provide the patient with the right drug at the right time and remind him/her to take the medicine have been developed but not yet evaluated to measure their potential benefit [61].

Table 2 offers an overview of potentially risky sub-steps and currently available data on error frequencies and options for preventions. Hence, this table might guide both future research activities in order to close the gaps and serve as checklist to identify potentially error prone processes in an individual setting.

Table 2.

Alignment of potentially error prone sub-steps of drug administration with reported errors and published evidence on error prevention

Competing Interests

All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare no support from any organization for the submitted work, no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years and no other relationships or activities that could appear to have influenced the submitted work.

This work was funded by a grant of the Federal Ministry for Education and Research (BMBF) (#01GK0801 and #01ET1004B). JS has a scholarship of the GRK793 of the Deutsche Forschungsgemeinschaft (DFG). AL has a scholarship from the Dr August and Dr Anni Lesm????ller Stiftung.