Pharm World Sci DOI 10.1007/s11096-007-9136-9

RESEARCH ARTICLE

Evaluation of pharmacist clinical interventions in a Dutch hospital setting Liesbeth Bosma Æ Frank G. A. Jansman Æ Anton M. Franken Æ Johannes W. Harting Æ Patricia M. L. A. Van den Bemt

Received: 7 December 2006 / Accepted: 13 April 2007
Springer Science+Business Media B.V. 2007

Abstract Objective Assessing the relevance of a clinically active pharmacist method compared to the traditional working method. Method The study was carried out in a general internal/ gastro-enterology unit during two 8-weeks periods in 2004. It was an observational, non-randomized prospective study. Outcome measures were compared before and during the intervention. The intervention was the active presence of a junior hospital pharmacist on the unit. The pharmacist focused on the pharmacotherapy of the individual patient. Patients were included when they used 5 or more medicines on day 1 or 2 of their stay at the ward and/or used at least 1 high-risk drug. L. Bosma
F. G. A. Jansman
J. W. Harting Hospital Pharmacy, Isala Klinieken, Zwolle, The Netherlands F. G. A. Jansman Department of Social Pharmacy, Pharmacoepidemiology and Pharmacotherapy, Groningen University Institute for Drug Exploration, Groningen, The Netherlands A. M. Franken Department of Internal Medicine, Isala Klinieken, Zwolle, The Netherlands P. M. L. A. Van den Bemt Hospital Pharmacy, St. Lucas Andreas Hospital, Amsterdam, The Netherlands P. M. L. A. Van den Bemt Department of Pharmaco-epidemiology & Pharmacotherapy, Utrecht Institute for Pharmaceutical Sciences (UIPS), Utrecht University, Utrecht, The Netherlands Present Address: L. Bosma (&) Apotheek Haagse Ziekenhuizen, Escamplaan 900, PO box 43100, 2504 AC, The Hague, The Netherlands e-mail: [email protected]

Clinical pharmacist interventions were counted and classified. A hospital pharmacist and an internal medicine specialist assessed the clinical relevance of all clinical pharmacist interventions retrospectively. The degree of acceptance of the interventions by physicians was measured. Finally, time associated with the clinical activities was measured. Main outcome measures Number of interventions (related to number of medication orders), clinical relevance and degree of acceptance. Results In the pre-intervention period 79 patients were included versus 84 in the during-intervention period. About 82 interventions in the pre-intervention period were made compared to 173 during the during-intervention period. There was little agreement between the professional raters (weighted jA–E = 0.30 and weighted j1–5 = 0.20). Nevertheless both ratings showed a substantial increase of clinically relevant interventions. The number of interventions accepted by the physician increased from 16 in the preintervention period to 75 in the during-intervention period. Working with this method took over 4 h a day. Conclusion Clinical pharmacy services provided by a junior hospital pharmacist on an internal medicine ward contribute to rationalization of drug therapy and are therefore likely to increase medication safety. Keywords Clinical pharmacy
Clinical pharmacist intervention
Drug related problem
Hospital pharmacy
Medication Error
Pharmaceutical care
Prescription error
The Netherlands

Impact of findings on practice •

Clinical pharmacy services on an internal ward contribute to rationalization of drug therapy.

123

Pharm World Sci

• •

Ward pharmacy is time consuming compared to the traditional way of pharmacy practice. Efforts should be made to implement ward pharmacy in daily practice of European Hospital pharmacy.

Introduction Medication errors are known to occur frequently. Prescribing errors are among the most common types of medication errors [1]. Traditionally most drug therapy decisions are made by the physician, leaving the pharmacist’s role more reactive; i.e., responding to prescribing errors long after the decision has been made, and without having direct clinical knowledge of the patient. This leads to inappropriate or contraindicated prescriptions that remain undetected by pharmacists [2]. It is thought that the pharmacist’s impact can be of more importance if he or she provides input earlier and more directly (at the time of prescribing by being present on the ward). Studies on an ICU and on a general medicine unit have shown that pharmacist participation in medical rounds contributes to a significant reduction in (preventable) adverse drug events (ADEs) [3–5]. Although clinical pharmacy services have been practiced since many years in UK hospitals, this is in general not the case in other European countries [6]. Most studies about the value of clinical pharmacy services originate from the USA and UK, while European studies on the clinical relevance of pharmacy interventions are mostly focused on community pharmacy [7–10]. The situation in USA and the UK hospital pharmacies cannot be compared to the situation in for example a Dutch hospital pharmacy. In general, in the USA and UK more pharmacists are employed in hospital pharmacies, which makes it easier to allocate time to clinical services. Furthermore, Dutch hospital pharmacies have additional tasks such as the preparation of drugs (i.e., actual preparation, not just making them ready for use) and laboratory medicine (therapeutic drug monitoring, toxicological screening). Comparable differences with USA and UK hospital pharmacies exist in many other European countries. Yet, as in the Netherlands and the rest of Europe the awareness of the importance of medication error prevention is increasing, this creates opportunities for expansion of clinical activities of hospital pharmacists. Evidence supporting the value of clinical services in Dutch hospitals could further aid in this expansion and thus in the improvement of patient safety. Therefore, we conducted a prospective study into the clinical relevance of pharmacist interventions, comparing the clinically active pharmacist method with the traditional reactive pharmacist method.

123

Method Setting The study was carried out in a general internal/gastroenterology unit at the ‘‘Isala klinieken,’’ a large secondary care hospital in Zwolle, the Netherlands, during two 8-week periods: February 2, through March 28, 2004 (phase I, pre-intervention), and April 12, through June 6, 2004 (phase II, during-intervention). On the unit 3 gastroenterologists (20 beds) and 4 internal medicine specialists (20 beds) were active, supported by 3 junior doctors (two registrars and a junior house officer). Doctor’s prescriptions are collected from the ward twice a day. Pharmacy technicians enter the prescriptions the pharmacy computer system. Besides supporting the logistics and distribution of drugs, the system builds up a patient drug history and generates alerts in case of duplicate medication, abnormal dosages, drug–drug interactions and non-formulary drugs. The pharmacy technician sees all alerts generated by the system while entering the prescription data in the computer. Depending on the nature of the alert, he or she can ignore the alert, resolve the problem himself or consult the hospital pharmacist. Every day one hospital pharmacist is engaged with the so-called ‘clinical duty.’ This means he or she answers questions from doctors, nurses, and pharmacy technicians. After the entire logistic process has finished, the pharmacist performs a final check on all prescriptions. At the end of the day the pharmacist checks a list of all alerts generated by the system during that day. Also the way technicians handled a certain alert is visible on the list. Depending on the nature of the signal, the hospital pharmacist can either ignore the alert, call a doctor, or make a ‘doctors note,’ which is sent to the ward at the end of the day. Drug–drug interactions with coumarin anticoagulants are sent to the anticoagulation clinic. The Isala klinieken have a drug substitution policy; the pharmacy is authorized to change non-formulary drugs to formulary drugs. For this purpose authorized substitution lists are used. Study design It was an observational, non-randomized, prospective study, divided into two phases of 8 weeks. We compared outcome measures before (‘pre-intervention period’) and during the intervention (‘during-intervention period’). Study population During the first 7 weeks of each phase patients were included, when they met the following criteria:

Pharm World Sci

– using 5 or more medicines on day 1 or 2 of their stay on the ward; – using at least 1 high-risk drug (see below). In this way patients at risk for ADEs [11, 12] were selected. Patients were excluded if they were discharged from the ward within 3 days. Patients were followed during their entire stay on the ward. During the last week of the study period no more patients were included; this week was used to follow the patients already included. In the preintervention period there were 9 ‘included’ patients still admitted to the ward when the 8 weeks study period finished, in the during-intervention period there were 7 of such patients. All of these patients were followed until discharge.

The assessments were made in accordance with the method described by Overhage et al. [14]. This method for rating the value of pharmacists’ clinical services consists of an instrument to measure the severity of medication errors and the value of pharmacists’ clinical interventions. The instrument has specific examples to assist in the proper application. A summary of the instrument is shown in Table 1. Each assessment was carried out based on a description of the intervention event, a complete medication chart and the discharge letter for the general practitioner (which contains general data on the medical history of the patient and more specific details on the period of hospitalization). Time associated with the clinical activities was a secondary outcome measure.

The intervention: a different working method

List of high-risk medicines

The intervention was the active presence of one junior hospital pharmacist on the ward. The pharmacist focused on the pharmacotherapy of the individual patient. A special working method was developed for the pharmacist. This method was based on the ASHP ‘‘Clinical skills’’-module [13] and contains the following activities: collecting patient information, pharmaceutical care evaluation and participation on physician rounds. The pharmaceutical care evaluation consisted of: evaluation of the choice of medical therapy, dosing regimen, therapeutic duplication, side effects, and drug–drug interactions. This clinically active method was compared with the traditional reactive method, as described in the ‘setting’ section above.

We compiled a list of 14 high-risk medicines, based on a list made by the Dutch Scientific Institute for Pharmacy (WINAp) [15] and the updated Beers criteria [12] as well as a Canadian list [16]. Based on the fact that half of the study population was gastro-enterologic, typical high-risk gastroenterology medicines like cisapride, azathioprine, and cyclosporine were added. High-risk medication that was not part of the hospital formulary was excluded (Table 2).

Outcome measures As outcome measure, the effect of the intervention was assessed by measuring the number of interventions (related to the number of medication orders) and the clinical relevance of pharmacist interventions. Also the number of physician implemented interventions was measured. A clinical pharmacist intervention was defined as a non-routinely action by the pharmacy technician or the pharmacist based on a doctors prescription. The type of intervention was classified as ‘‘prescribing error,’’ ‘‘distribution error,’’ ‘‘formulary intervention,’’ or ‘‘pharmaceutical care intervention’’ [3, 14]. Clinical interventions that were recommendations with the intention to change drug treatment were counted separately and were marked as ‘‘accepted,’’ ‘‘not accepted’’ or as ‘‘ acceptance unknown.’’ Two independent professionals (one senior hospital pharmacist and one internal medicine specialist–clinical pharmacologist) assessed the significance of all clinical pharmacist interventions retrospectively.

Data analysis Using an alfa of 0.05, a power of 0.80 and a pre-intervention proportion of clinical pharmacist interventions of 10% of prescriptions [17] and a during-intervention proportion of 15% a sample size of 686 prescriptions can be calculated, i.e., 350 prescriptions per period. Based on a worst case scenario (each patient using exactly 5 drugs), this would mean that at least 70 patients per period needed to be included. Based on 40 beds, a bed-occupancy of 90%, a mean admission period of 14 days and inclusion of 2 out of 3 patients, an inclusion period of 7 weeks was chosen. Patient data and clinical pharmacist intervention data were entered into a database (MS Access version 1997). Patient variables for the pre-intervention and during-intervention period were compared using the two-sample t-test for continuous variables like age, length of stay and number of drugs prescribed. We used chi-square analysis for dichotomous variables and proportions (e.g., for gender, medical specialty, type of censoring [discharge to another ward, to a nursing home, to patient’s own home or death] and proportion of patients on high-risk medication). Significance ranking scores of the clinical pharmacist interventions for the pre-intervention and during-intervention period were compared using chi-square analysis. A P-value