Polypharmacy and Inappropriate Prescribing
Polypharmacy, or chronic use of multiple medicines, poses significant threats to patients’ health. A consensual definition of polypharmacy is lacking, but it is often described as the concurrent use of five or more different chronically used drugs . Polypharmacy has been associated with negative health consequences. Drugs may cause clinical interactions or adverse effects that may aggravate patients’ symptoms instead of relieving them. Medicine issues including underprescribing, overtreatment and decreased drug adherence have been associated with polypharmacy [2–9]. A 2008 study showed that in the Netherlands, 5.6 % of all acute hospital admissions had medication-related causes . For elderly patients, who constitute half of all chronically ill polypharmacy patients, this figure was twice as high .
The concurrent use of multiple medications is not entirely undesirable, as in many patient cases, polypharmacy is indicated or even unavoidable. However, inappropriate prescribing of medications is prevalent among elderly patients . An incidence-focused study found that inappropriate medication use increased elderly persons’ risks of hospitalization and mortality . Geriatric assessment and medication review have been shown to be effective methods in aiding prescribers with optimizing polypharmacy [14, 15].
A multitude of initiatives has been developed to assess the appropriateness of drugs prescribed for individual patients. These approaches can be divided into implicit and explicit methods. The former implicit methods use patient-specific information, combined with medical knowledge, to determine medication appropriateness, while the latter explicit methods provide screening tools, containing lists of clinical interactions or contraindications . Among the explicit methods are the Beers Criteria and the Screening Tool to Alert to Right Treatment (START) and Screening Tool of Older People’s Prescriptions (STOPP) criteria, while the implicit methods include the Medication Appropriateness Index and the pharmacotherapy review focused on drugs’ use, indication, safety and effectiveness (Gebruik Indicatie Veiligheid Effectiviteit; GIVE) [16–19]. The effectiveness of these interventions varies; generally they appear beneficial in terms of reducing inappropriate prescribing and medication-related problems, but they have not been proven to lead to clinically significant improvement .
In order to improve medication prescribing in primary care, several implicit and explicit methods have been combined into an all-encompassing systematic medication review approach—the Polypharmacy Optimization Method (POM). It has been shown to significantly improve general practitioners’ (GPs’) prescriptions for polypharmacy patients in an experimental setting .
A variety of barriers are impeding the widespread adoption of structured medication reviews in daily practice. Recently, Anderson et al.  conducted a systematic literature review on enablers and barriers to minimizing potentially inappropriate medications by GPs. Most factors revolved around physicians, and they included inertia (his or her attitudes towards discontinuation, such as fearing negative consequences), self-efficacy (his or her knowledge and available information on the topic) and awareness (his or her having poor insight or discrepant beliefs). Barriers that were not physician related included a lack of resources, patients resisting changes to their medication, and practical and cultural factors. A separate study focusing on barriers regarding pharmacist-led medication reviews reported lack of time and lack of self-confidence as the most commonly perceived barriers .
Recently, the POM, GIVE, and START and STOPP criteria have been combined into the Systematic Tool to Reduce Inappropriate Prescribing (STRIP), which has consequently been included as part of a Dutch multidisciplinary guideline on polypharmacy in elderly patients . The STRIP has been designed to be an all-encompassing drug optimization process in primary care, focusing not just on pharmacotherapeutic analysis but also on patients’ medication histories and preferences; Fig. 1 shows the STRIP method’s different steps.
The STRIP analysis is more extensive than its predecessors [14, 17, 19]. It combines both the implicit approaches of the POM and the GIVE, and the explicit lists of the first version of the START and STOPP criteria. The pharmacotherapeutic analysis in the STRIP includes checks on underprescribing, overtreatment, recommended dosage adjustments, drug effectiveness, potential adverse effects, dose frequency, clinical interactions and medication adherence, including practical problems with medication use. The START and STOPP criteria are implemented in the pharmacotherapeutic analysis. This extensive medication review results in a patient-specific treatment plan in which new drugs are gradually added and superfluous ones are discontinued. This approach to conducting structured medication reviews is based on consensus rather than evidence, synthesizing the results of the earlier optimization methods mentioned above. Currently, solid evidence for choosing specific strategies for the optimization of pharmacotherapy in the elderly over others is lacking .
Involvement of patients in the medication review is emphasized to ensure their therapy adherence; patients’ preferences are taken into account as much as possible. The pursuit of the treatment plan is monitored through regular communication between the practitioner, pharmacist and patient. The involvement of pharmacists in medication reviews, as part of multidisciplinary teams, has been shown to lead to improved pharmacotherapy for older patients . Educating patients on their medication use and treatment goals, simplifying their drugs regimens and preventing adverse drug reactions have all been identified as factors influencing patients’ adherence to their treatments .
Clinical Decision Support Systems
In recent years, computerized physician order entry (CPOE) systems have gradually changed in terms of functionality. From systems that were traditionally organizational in nature, they have been enhanced to facilitate management of electronic medical records and clinical decision support . There is consensus in the literature that clinical decision support has the potential to improve GPs’ and pharmacists’ decision-making : “Both commercially and locally developed CDSSs [clinical decision support systems] are effective at improving health care process measures across diverse settings”. The evidence for concurrent improvement in efficiency, cost effectiveness or clinical effectiveness is inadequate or ambiguous. A study investigating the attitudes of Dutch GPs to the introduction of a decision support system specifically aiding them with conducting medication reviews revealed that the majority were positively inclined towards using such a system .
In order to enable GPs and pharmacists to effectively and efficiently incorporate the STRIP method into their daily practice, the STRIP Assistant has been developed. The STRIP Assistant has been designed as a stand-alone web application, which aims to assist GPs and pharmacists with pharmacotherapeutic analysis of patients’ medical records. On the basis of patients’ records and the decisions that GPs and pharmacists make during the medication review, the application generates context-specific advice. The STRIP Assistant’s design decisions adhere to best practice in information science research; the user interface conception and decision rule implementation have been designed to balance efficiency and information completeness, aiming to minimize previously mentioned barriers such as users’ lack of confidence and lack of time .
The knowledge used to generate the STRIP Assistant’s advice consists of well-established guidelines on clinical interactions, double-medication, contraindications, dosage strength and frequency, and specific implementations of version 1 of the START and STOPP criteria [30, 31]. The rules incorporate not only patients’ diseases and drugs but also their contraindications, complaints and relevant physical properties (such as renal function and weight). This results in items of advice that recommend users to add new drugs or to remove superfluous ones, or to change dosages of existing medicines.
It has been planned that in the future, the STRIP Assistant will integrate with existing CPOE systems, thereby increasing the efficiency with which the method can be performed. Additionally, use of data-mining techniques on historical data should reveal patterns in users’ behaviour towards the generated advice, which could be used to improve recommendations .
A video demonstrating the use of the STRIP Assistant can be viewed at http://videodemo.stripa.eu/english/ .
Usability has long been regarded as an essential factor for the success of software applications. In the widely used definition issued by the International Organization for Standardization (ISO), usability is defined as “the extent to which a product can be used by specified users to achieve specified goals with effectiveness, efficiency, and satisfaction in a specified context of use” . In this context, effectiveness is understood as the “accuracy and completeness with which users achieve specified goals”, while efficiency consists of “resources expended in relation to the accuracy and completeness with which users achieve goals”. Finally, user satisfaction is the subjective “degree to which user needs are satisfied when a product or system is used in a specified context of use”.
A recent systematic literature review on clinical decision support systems showed that there is ample evidence that these systems can improve effectiveness; not enough research on efficiency and user satisfaction is available to make generalizations regarding these aspects . In the technology adoption literature, it has been shown that systems’ perceived usefulness and ease-of-use—aspects closely related to usability—are the major determinants of people’s attitudes towards using technology [35–37].
Hornbaek  described the current practices in evaluating usability. A multitude of metrics and instruments have been used to measure the three main factors of usability identified in the ISO definition. Measurements of effectiveness usually involve the degree to which a task has been successfully completed, leading to metrics such as accuracy, recall and completeness. Efficiency metrics mostly revolve around the time spent completing a task but can also involve mental efforts. The subjective user satisfaction criterion is often measured through standardized questionnaires or interface ranking.
The aforementioned considerations lead us to believe that clinical decision support has the potential to successfully aid GPs and pharmacists in incorporating structured medication reviews into daily practice. Therefore, we aimed to validate the STRIP Assistant instrument’s usability as a tool for physicians to optimize medical records for polypharmacy patients in an experimental setting. This main research question was divided into the following sub-questions:
Do GPs and pharmacists make significantly more appropriate decisions when optimizing the medical records of polypharmacy patients with the STRIP Assistant than without it?
Do GPs and pharmacists make significantly fewer inappropriate decisions when optimizing the medical records of polypharmacy patients with the STRIP Assistant than without it?
Do GPs and pharmacists take significantly less time to optimize prescribing for polypharmacy patients with the STRIP Assistant than without it?
Do GPs and pharmacists perceive use of the STRIP Assistant for optimizing the medical records of polypharmacy patients as satisfactory?
In this context, the term ‘appropriate decisions’ means decisions that correspond to those agreed upon by an expert panel.