Int J Clin Pharm DOI 10.1007/s11096-014-9939-4
RESEARCH ARTICLE
Characterization of drug-related problems identified by clinical pharmacy staff at Danish hospitals Lene Juel Kjeldsen • Trine Birkholm • Hanne Fischer • Trine Graabæk • Karina Porsborg Kibsdal • Lene Vestergaard Ravn-Nielsen • Tania Holtum Truelshøj
Received: 12 December 2013 / Accepted: 24 March 2014 Koninklijke Nederlandse Maatschappij ter bevordering der Pharmacie 2014
Abstract Background In 2010, a database of drug related problems (DRPs) was implemented to assist clinical pharmacy staff in documenting clinical pharmacy activities locally. A study of quality, reliability and generalisability showed that national analyses of the data could be conducted. Analyses at the national level may help identify and prevent DRPs by performing national interventions. Objective The aim of the study was to explore the DRP characteristics as documented by clinical pharmacy staff at hospital pharmacies in the Danish DRP-database during a 3-year period. Setting Danish hospital pharmacies. Method Data documented in the DRP-database during the initial 3 years after implementation were analyzed retrospectively. The DRP-database contains DRPs reported at hospitals by clinical pharmacy staff. The analyses focused on L. J. Kjeldsen (&) H. Fischer The Danish Research Unit for Hospital Pharmacy, Amgros I/S, Dampfærgevej 22, 2100 Copenhagen Ø, Denmark e-mail: [email protected] T. Birkholm The Hospital Pharmacy, Roskilde, Region Zealand, Denmark T. Graabæk Centre for Clinical Research, Hospital South West Jutland, Esbjerg, Denmark K. P. Kibsdal The Hospital Pharmacy, Aalborg, North Denmark Region, Denmark L. V. Ravn-Nielsen Clinical Pharmacy Unit of Odense University Hospital, Odense, Denmark T. H. Truelshøj The Hospital Pharmacy, Aarhus, Central Denmark Region, Denmark
DRP categories, implementation rates and drugs associated with the DRPs. Main outcome measure Characteristics of DRPs. Results In total, 72,044 DRPs were documented in the DRP-database during the first 3 years of implementation, and the number of documented DRPs increased every year. An overall stable implementation rate of approximately 58 % was identified. The DRPs identified were multi-facetted, however evenly distributed for each of the 3 years. The most frequently identified DRP categories were: ‘‘Dose’’, followed by ‘‘Nonadherence to guidelines’’ and ‘‘Supplement to treatment’’. The highest implementation rates were found for the following DRP categories: ‘‘Non-adherence to guidelines’’ (79 %) followed by ‘‘Therapeutic duplication’’ (73 %) and ‘‘Dosing time and interval’’ (70 %). Even though the top 25 drugs were involved in 58 % of all DRPs, multiple drugs were associated with DRPs. The drugs most frequently involved in DRPs were paracetamol (4.6 % of all DRPs), simvastatin (3.0 %), lansoprazole (2.7 %), morphine (2.6 %) and alendronic acid (2.4 %). Conclusions The study found that a national database on DRPs contained multi-facetted DRPs, however evenly distributed for each of the 3 years. Even though the top 25 drugs were involved in 58 % of all DRPs, multiple drugs were associated with DRPs. The study emphasizes the importance of detecting and intervening for DRPs. Keywords Clinical pharmacy Database Denmark DRP Drug-related problem Hospital Medication review
Impact of findings on practice •
The analysis of the Danish DRP-database may assist in further developing clinical pharmacy in
123
Int J Clin Pharm
•
Denmark by identifying focus areas for intensified intervention. Data from the Danish DRP-database may be used for benchmarking between hospitals and over time.
all available data despite local differences, a national study on the quality and reproducibility of the database was conducted [6]. The results showed high inter-rater reliability indicating that local data may be pooled and used for national analyses [6].
Introduction
Methods
Suboptimal medication treatment may result in increased morbidity, mortality and costs [1, 2]. A method for improving the medication treatment in secondary care is by clinical pharmacists conducting medication related interventions such as medication reviews [3]. Medication reviews include identification of drug-related problems (DRPs), which may be defined as ‘‘a circumstance related to the patient’s use of a drug that actually or potentially prevents the patient from gaining the intended benefit of the drug’’ [4]. When conducting a medication review, recommendations of changes to the medication treatment based on the identified DRPs are made in order to optimize the medication treatment of the patient. A recent systematic review on medication reviews performed by clinical pharmacists in secondary care showed that a large variability existed in the proportion of DRPs identified in the 31 included studies; 0.13–10.6 per patients [3]. This variation may be explained by the heterogeneity of the study cohorts, the methodology used for the intervention including specific foci and the clinical experience of the clinical pharmacist. Usually, implementation of recommendations proposed by clinical pharmacists when conducting medication review is subject to acceptance by the treating physician. The systematic review also found a large variation in acceptance rates among the studies; 39–100 %, most likely due to the heterogeneity of the methods of the studies [3]. Studies in primary care have shown that categorization of DRPs may help developing counseling skills, assist in quality assurance, and is of value regarding pharmaceutical care practice and research [4, 5]. In Denmark, a national electronic tool to standardize DRPs was established in 2009/2010 (the DRP-database) and it currently contains 125,000 records. The DRP-database was established on request by clinical pharmacists throughout the country, and is available for all clinical pharmacists at hospital pharmacies in Denmark. The DRP-database is hosted and maintained by Amgros I/S, which is owned by the Danish regions. The DRP-database is used locally for documentation of activities at the wards, for quality assurance purposes and to identify focus areas to prevent DRPs. The DRP-database provides the opportunity to analyse data at a national level and develop national strategies to prevent DRPs. To explore whether analyses could be performed on
A project group consisting of five clinical pharmacists from hospital pharmacies representing various areas of the country and two clinical pharmacists from the Research Unit for Hospital Pharmacy designed and overviewed the current study. The five clinical pharmacists from hospital pharmacies ensured representation of issues related to regular use of the DRP-database.
123
The DRP-database DRPs are entered electronically by clinical pharmacy staff, and at present the database is predominantly used by clinical pharmacists. The DRP-database comprises information on each DRP by categorization, Anatomical Therapeutic Chemical (ATC) code [7], follow-up on the DRPs (documentation of whether the DRP was e.g. accepted or rejected) and a text field for documentation of additional information such as the recommendation made. In addition, the database contains information on demographics, date of registration of the DRP and number of drugs of the patient’s medication regimen. The DRPs can be categorized at two levels. The first level contains 14 categories; dose, dosing time and interval, side effect, interaction, drug form and strength, compatibility, non-adherence to guidelines, therapeutic duplication, drug allergies, length of treatment, supplement to treatment, electronic patient chart related (EPC-related), inappropriate drug, as well as no DRP identified. Each of the categories may be further characterized on a sublevel, e.g. ‘‘dose’’ may be subcategorized into ‘‘Dose too high’’ or ‘‘Dose too low’’ etc. ‘‘Non-adherence to guidelines’’ include several guidelines such as drug formularies, practice guidelines and institutional guidelines. ‘‘EPC-related’’ contain DRPs related to the set-up and use of the EPC, e.g. an unconscious patient with medications documented for self administration. Data for the analyses Data were retrieved retrospectively from the DRP-database for 3 years from the introduction of the database; July 2010 to June 2013, and comprised 124,718 records. The data were categorized into three time periods; (1) July 2010 to
Int J Clin Pharm Fig. 1 Actions taken on DRPs distributed on the three time periods. (Color figure online)
June 2011, (2) July 2011 to June 2012 and (3) July 2012 to June 2013. Results on the sublevel are not presented in the current study since the validity of the DRP-database has only been tested at the first level [6]. Additionally, the category ‘‘Compatibility’’ was excluded from the analyses, since it is rarely used (nine registrations during the study period). Finally, the category ‘‘No DRP identified’’ was excluded from the analyses (52,665 records). This category could have helped identify the proportion of patients with DRPs of the reviewed patients. However, a survey showed that all users do not consistently document patients with no DRPs identified in the database, hence the results would not be valid [6]. The remaining data, 72,044 records, were included for analyses. The DRP-database contains two unauthorized ATC codes; Z01 (=other) and Z02 (=not relevant). Z01 comprises a plethora of different DRPs—often related to supplement to treatment e.g. calcium, vitamin D etc. or recommendations regarding several drugs concomitantly, for which the clinical pharmacy staff may have found difficulties in identifying a specific ATC code. Z02 is primarily used in relation to lack of medication anamnesis, lack of registration of drug allergies or untreated indication, where no specific drug is involved. Data associated with these unauthorized ATC codes were excluded in the analyses when applicable. Approval of the study was received by the Danish Data Protection Agency.
Results During the study period, 72,044 DRPs were documented in the DRP-database. An increase in documentation of DRPs was reported during the 3 time periods, particularly from period 1–2; 15,901 in period 1, 27,203 in period 2 and 28,940 in period 3. About half of the DRPs were identified at emergency wards, approximately a third at medical wards, about 10 % at surgery wards, 1 % at psychiatric wards and only a few at intensive care units and paediatric wards. As well as for the sub-categories, the overall implementation rates of the DRPs were stable over the 3 years; approximately 58 % (Fig. 1). The prevalence of the DRP categories varied with the most frequently being ‘‘Dose’’, followed by ‘‘Non-adherence to guidelines’’ and ‘‘Supplement to treatment’’ (Table 1; Fig. 2). In addition, implementation rates of interventions also varied between the DRP categories with the highest being for ‘‘Non-adherence to guidelines’’ (79 %) followed by ‘‘Therapeutic duplication’’ (73 %) and ‘‘Dosing time and interval’’ (70 %) (Fig. 2). The lowest implementation rates were found for ‘‘Side effect’’ and ‘‘Drug allergies’’, both 37 % (Fig. 2). The drugs most frequently reported in the DRP-database were paracetamol (4.6 % of all DRPs), simvastatin (3.0 %), lansoprazole (2.7 %), morphine (2.6 %) and alendronic acid (2.4 %). Accordingly, these five drugs were present in top three for one or more of the individual DRP
123
Int J Clin Pharm Table 1 Prevalence of DRP categories of the three time periods pooled and including prevalence of the three most prevalent medications for each DRP category DRP category
Prevalence of DRP category
Dose
12,039 (17 %)
Non-adherence to guidelines
Supplement to treatment
11,179 (16 %)
10,997 (15 %)
Medication Paracetamol (N02BE01)
975 (8.1 %)
Zopiclone (N05CF01)
759 (6.3 %)
Pantoprazole (A02BC02) Lansoprazole (A02BC03)
Inappropriate drug
Electronic patient chart related
Interaction
Therapeutic duplication
Drug form and strength
Drug allergies
Side effect
Length of treatment
8,616 (12 %)
8,379 (12 %)
7,250 (10 %)
3,283 (4.6 %)
2,485 (3.4 %)
2,361 (3.3 %)
1,954 (2.7 %)
1,863 (2.6 %)
1,638 (2.3 %)
506 (4.2 %) 1,461 (13 %)
Paracetamol (N02BE01)
800 (7.2 %)
Omeprazole (A02BC01)
771 (6.9 %)
Calcium, combinations with vitamin D and/or other drugs (A12AX)
394 (3.6 %)
Simvastatin (C10AA01)
378 (3.4 %)
Paracetamol (N02BE01) Dosing time and interval
Most frequent within DRP category
Alendronic acid (M05BA04)
280 (2.5 %) 1,158 (13 %)
Cefuroxime (J01DC02)
613 (7.1 %)
Simvastatin (C10AA01)
596 (6.9 %)
Quinine (P01BC01)
278 (3.3 %)
Ibuprofen (M01AE01) Tramadol (N02AX02)
264 (3.2 %) 230 (2.7 %)
Warfarin (B01AA03)
635 (8.8 %)
Paracetamol (N02BE01)
265 (3.7 %)
Simvastatin (C10AA01)
169 (2.3 %)
Ciprofloxacin (J01MA02)
360 (11 %)
Warfarin (B01AA03)
218 (6.6 %)
Ferrous sulfate (B03AA07)
203 (6.2 %)
Paracetamol (N02BE01)
247 (9.9 %)
Morphine (N02AA01)
120 (4.8 %)
Tramadol (N02AX02)
109 (4.4 %)
Paracetamol (N02BE01)
317 (13 %)
Morphine (N02AA01)
282 (12 %)
Tramadol (N02AX02)
136 (5.8 %)
Morphine (N02AA01)
165 (8.4 %)
Benzylpenicillin (J01CE01) Beta-lactam antibacterials, penicillins (J01C)
138 (7.1 %) 109 (5.6 %)
Ibuprofen (M01AE01)
85 (4.6 %)
Citalopram (N06AB04)
72 (3.9 %)
Tiotropium bromide (R03BB04)
64 (3.4 %)
Prednisolone (H02AB06)
260 (17 %)
Pivmecillinam (J01CA08)
134 (8.2 %)
Ciprofloxacin (J01MA02)
49 (3.0 %)
ATC codes Z01 and Z02 not presented
categories. Indeed, paracetamol, the most frequent drug, could be found in top three for six DRP categories (Table 1). The high frequency of these individual drugs affected the prevalence of drugs at ATC level four (Table 2). The most prevalent drug class was ‘‘proton pump inhibitors’’, and lansoprazole comprised 41 % of this class followed by
123
pantoprazole (28 %) and omeprazole (21 %). Of the class ‘‘anilides’’, 97 % could be attributed to paracetamol, followed by paracetamol, combinations excl. psycholeptics (2 %) and paracetamol, combinations with psycholeptics (1 %). Finally, the third most prevalent class of ‘‘natural opium alkaloids’’ consisted of morphine (59 %), oxycodone (37 %), codeine, combinations excl. psycholeptics
Int J Clin Pharm Fig. 2 Number of implemented and not implemented DRPs distributed on DRP categories pooled for the three time periods. Not implemented ‘‘no action identified’’, ‘‘DRP accepted but no action identified’’, ‘‘DRP rejected’’ and ‘‘patient discharged’’. Implemented ‘‘DRP accepted and implemented by physician’’ and ‘‘DRP accepted and implemented by clinical pharmacy staff’’. (Color figure online)
(2 %) and a few other less prevalent drugs. The 25 most prevalent drugs associated with DRPs comprised 58 % of all DRPs. An example of a recommendation documented in the DRP-database of a frequent DRP related to paracetamol (ATC code N02BE01) categorized as ‘‘Dose’’ was: •
Reduction of paracetamol dose—risk of exceeding maximum recommended daily dose due to concurrent treatment with regularly prescribed as well as prn dose. The consequence might be effect on hepatic function.
Another example of a frequent case was one related to ibuprofen (ATC code M01AE01) and categorized as ‘‘Side effect’’: •
Recommend reviewing NSAID since this is not recommended for treatment of elderly due to increased risk of gastrointestinal side effects.
Clinical relevance of the DRP was not recorded in the DRP-database. However, a couple of examples of complex DRPs with significant clinical relevance are mentioned below (both categorized as ‘‘Inappropriate drug’’ with ATC codes A10BA02 and N06AA09, respectively): •
•
Recommend cessation of metformin due to poor kidney function (eGFR 24 ml/min). Metformin is contraindicated when eGFR is \30 ml/min due to risk of lactic acidosis. Recommend tapering off amitriptyline due to present acute myocardial infarction (AMI). Tricyclic antidepressants are contraindicated during the first 6 months following an AMI.
Discussion The study found 72,044 DRPs documented in the DRPdatabase during the first 3 years of implementation, and the number of documented DRPs increased every year. An overall stable implementation rate of approximately 58 % was identified. The DRPs identified were multi-faceted, however similarly distributed for each of the 3 years. Even though the top 25 drugs were involved in 58 % of all DRPs, multiple drugs were associated with DRPs. Frequency of DRP categories Comparison of DRPs between studies is a challenge, since a variety of DRP categorization methods exist in the literature by e.g. Hepler and Strand, Hanlon et al., American Society for Health-system Pharmacists, and the Pharmaceutical Care Network Europe [4, 7–11]. Of six selected Danish studies, two [12, 13] used a similar categorization system to characterize DRPs and allowed the results to be compared to the current evaluation [12–17]. The current study found the categories ‘‘Dose’’, ‘‘Nonadherence to guidelines’’ and ‘‘Supplement to treatment’’ as the most frequently identified DRPs. Somewhat similar findings were reported from a study at an emergency ward; ‘‘Non-adherence to guidelines’’ (67 %), ‘‘Inappropriate drug’’ (10 %) and ‘‘Supplement to treatment’’/‘‘Electronic patient chart related’’ (5 %) [13]. Contrarily, a study of seven wards in Horsens most frequently identified DRPs within the categories not specified as a category in the current categorization system (‘‘Medication reconciliation’’
123
Int J Clin Pharm Table 2 Prevalence of the 25 most prevalent drug classes at ATC level 4 during the 3 years Drug
ATC code at level 4a
Period 1
Period 2
Period 3
Total
Total (%) of all DRPs
Proton pump inhibitors
A02BC
1,031
1,849
1,762
4,642
6.4
b
Other
Z01
1,127
1,451
1,514
4,092
5.7
Anilides
N02BE
620
1,405
1,380
3,405
4.7
Natural opium alkaloids
N02AA
592
1,288
1,275
3,155
4.4
HMG CoA reductase inhibitors
C10AA
538
918
1,028
2,484
3.4
Platelet aggregation inhibitors excl. heparin
B01AC
572
835
857
2,264
3.1
Bisphosphonates Calcium, combinations with vitamin D and/or other drugs
M05BA A12AX
444 496
709 583
653 516
1,806 1,595
2.5 2.2
ACE inhibitors, plain
C09AA
384
490
545
1,419
2.0
Benzodiazepine related drugs
N05CF
353
535
516
1,404
1.9
Other opioids
N02AX
274
622
484
1,380
1.9
Selective serotonin reuptake inhibitors
N06AB
249
506
619
1,374
1.9
Second-generation cephalosporins
J01DC
196
476
563
1,235
1.7
Vitamin K antagonists
B01AA
348
413
456
1,217
1.7
Beta blocking agents, selective
C07AB
237
367
536
1,140
1.6
Benzodiazepine derivatives
N05BA
238
360
485
1,083
1.5
Beta-lactamase sensitive penicillins
J01CE
262
336
464
1,062
1.5
Sulfonamides, plain
C03CA
190
353
425
968
1.3
Osmotically acting laxatives
A06AD
175
345
414
934
1.3
Propionic acid derivatives
M01AE
175
373
349
897
1.2
Adrenergics and other drugs for obstructive airway diseases
R03AK
193
304
394
891
1.2
Glucocorticoids Anticholinergics
H02AB R03BB
183 187
372 319
288 290
843 796
1.2 1.1
Fluoroquinolones
J01MA
207
264
312
783
1.1
Other antidepressants
N06AX
155
274
336
765
1.1
a
Z02, i.e. DRPs non related to specific drugs such as e.g. lack of medication anamnesis, lack of registration of drug allergies or untreated indications were excluded from this table
b
Z01 is annotated to DRPs with no identifiable ATC code. However in practice, it is used for DRPs related to multiple medications and in particular to vitamins and minerals
and ‘‘Other’’) (35 %), followed by ‘‘Side effect’’ (18 %) and ‘‘Dose’’ (10 %) [13]. Despite using similar categorization systems, it is likely that the focus of the intervention, the ward type as well as the culture of the location lead to varieties in DRP identification within Denmark. Similar findings were identified in a selection of international studies [18–21]. Implementation of interventions To illustrate the proportion of DRPs, on which action was taken, implementation rates were presented i.e. DRP accepted and implemented by physicians as well as by clinical pharmacy staff. Acceptance rates are often used in the literature instead of implementation rates, however, when considering acceptance rates, the relevant DRPs are limited to recommendations requiring an action by physicians [3]. Additionally, only the acceptance of the
123
recommendation by the physician is recorded—not necessarily whether it was implemented. In the current study, an overall implementation rate of 58 % was discovered, which was similar to acceptance rates found in three Danish studies; 52, 53 and 59 %, respectively [13, 14, 16]. Another two Danish studies found higher acceptance rates; 76 and 80 %, respectively [12, 17]. Individual acceptance rates were available for two of the studies, and a significant cause for achieving an overall high acceptance rate in one of the studies [12] was attributed to the number of DRPs regarding ‘‘Non-adherence to guidelines’’, which to a certain extent may be implemented directly by clinical pharmacy staff (acceptance rate = 88 %) [12, 16]. Implementation of specific interventions may be specified in local contracts between the hospital pharmacy and the respective ward. Direct comparison of the individual acceptance rates was not possible due to the use of an alternative categorization
Int J Clin Pharm
method in one of the Danish studies, and low numbers for some of the categories in the other study [12, 16]. A review reported acceptance rates in international studies between 39 and 100 %, which may be explained by the different procedures in the studies [3]. In a Swedish national DRP-database in primary care, DRPs are categorized into actual and potential DPRs [22]. It is possible that acceptance rates may depend on whether the DRP is considered as actual or potential, however, the study of the Swedish DRP-database did not present any data on this. The Danish DRP-database contains a mixture of actual and potential DRPs, but no registration on this level of detail is recorded. Clinical significance Another factor, which might influence acceptance rates/ implementation rates, is clinical significance. The two Danish studies from Kolding assessed clinical significance of the identified DRPs, and they found in both studies that almost half of the DRPs (47 and 48 %, respectively) could be considered significant/critical [14, 15]. They also found a higher acceptance rate (67 %) among significant/critical DRPs than for the ones classified as minimal/moderate (52 %) [14]. The DRP-database does not contain information on clinical significance on the reported DRPs, which is a rather resource demanding process, hence the proportion of significant/critical DRPs is unknown. Even though these studies of clinical significance are small and may not be generalized to the DRP-database, they may indicate that a considerable amount of the DRPs reported in the DRPdatabase are of significant/critical clinical significance. This assumption is based on the fact that all DRPs identified in Kolding are reported to the DRP-database, and that the studies were performed at the emergency ward, which is the most common ward represented in the DRP-database. Consequently, whether the implementation rate of 58 % is a reasonable number is unknown. It is possible that a higher implementation rate could be achieved by focusing on significant/critical DRPs when identifying DRPs, which seems to increase the acceptance rate. Another strategy could be to increase the potential for clinical pharmacists themselves to implement actions on less significant DRPs. Development in DRPs and drug registrations over time As for DRP categories, no development over time was detected for drugs most frequently involved in DRPs. It is possible that this finding may be due to the stable distribution of ward types during the 3 years with emergency departments being the most dominant. Another possibility may be that identification of DRPs as a part of conducting medication reviews is a general service that despite local
and ward diversities do result in similar DRPs. In a Belgian study, the most frequently drugs associated with DRPs were ‘‘antithrombotic agents (B01; 9.1 % of all interventions), psycholeptics (N05—including antipsychotics, anxiolytics, hypnotics, sedatives; 8.8 %), psychoanaleptics, (N06— including antidepressants, antidementia drugs; 8.2 %), analgesics (N02; 6.9 %), and drugs for obstructive airway diseases (R03; 6.6 %)’’ [21]. These drug classes are presented at ATC level two, however drugs from the four most frequent drug classes are also present at the list of the 25 most frequently drugs at ATC level four associated with DRPs in the current study. This may further emphasize the similarities in drugs associated with DRPs, which could be explained by these drugs being of high risk to the patients. However, the similar distribution of DRPs during the 3 years may be important regarding the approach to DRPs in clinical practice and hence further development of the medication review service. It is likely that certain medications may continuingly be associated with DRPs e.g. high risk medications, but it is important to ensure continued development of the medication review service to include e.g. focus on DRPs associated with new drugs and new drug combinations to deliver the best possible service. Increase in documentation in the DRP-database over time During the 3 years studied, the number of DRPs documented in the DRP-database increased with more than 80 %, however, the increase between year 2 and 3 was modest compared to the increase between year 1 and 2. It is likely that the initial increase was caused by an increase in the number of clinical pharmacists using the DRP-database, while the increase between the second and third year may primarily be due to an increase in the number of medication reviews delivered and hence documented. Potential perspectives of the DRP-database An evaluation of the DRP-database revealed a good quality and generalisability of the data allowing for valid national analyses, even though the DRP-database was developed for local quality assurance activities and not for research purposes. The current study presented some data at the national level, which may inspire clinical pharmacists to make further use of the information in the DRP-database. Some examples include benchmarking over time and between locations and regions, identification of focus areas for interventions locally or nationally, sharing information about routine services between locations and regions, identification of potential patient safety issues when constructing or updating medication treatment guidelines or when making procurements of hospital medications.
123
Int J Clin Pharm
Limitations Since the DRP-database was developed on request as a tool to assist clinical pharmacists in their daily work, data entry in the DRP-database is based on voluntary reporting. Estimation indicates that about half of all clinical pharmacy services provided in Denmark are documented in the database, and some hospital pharmacies use the DRPdatabase to document 100 % of their recommendations [6]. Consequently, the DRP-database does not cover all services provided throughout the entire country, which may affect the generalisability of the data [6]. In addition, no central control is provided to ensure uniformity when entering data. Several issues may affect the recording of the data, which may be assessed differently between locations [6]. A proportion of data entry mistakes may be expected, since no double control and audits of the data are performed. In the current study, some records had to be excluded due to wrong dates being recorded, but the extent of data entry mistakes is unknown. Further limitations have been addressed elsewhere [6].
Conclusion The study found that a national database on DRPs contained multi-faceted DRPs, however similarly distributed for each of the 3 years. Even though the top 25 drugs were involved in 58 % of all DRPs, multiple drugs were associated with most DRPs. The study emphasizes the importance of detecting and intervening for DRPs. Acknowledgments Funding
None.
None.
Conflicts of interest of interest.
The authors declare that they have no conflict
References 1. van den Bemt PM, Egberts TC, de Jong-van den Berg LT, Brouwers JR. Drug-related problems in hospitalised patients. Drug Saf. 2000;22(4):321–33. 2. Kra¨henbu¨hl-Melcher A, Schlienger R, Lampert M, Haschke M, Drewe J, Kra¨henbu¨hl S. Drug-related problems in hospitals: a review of the recent literature. Drug Saf. 2007;30(5):379–407. 3. Graabaek T, Kjeldsen LJ. Medication reviews by clinical pharmacists at hospitals lead to improved patient outcomes: a systematic review. Basic Clin Pharmacol Toxicol. 2013;112(6):359–73.
123
4. Van Mil F, Westerlund T, Hersberger KE, Schaefer MA. Drugrelated problem classification systems. Ann Pharmacother. 2004;38:859–67. 5. Westerlund LOT, Handl WHA, Marklund BRG, Allebeck P. Pharmacy practitioners’ views on computerized documentation of drug-related problems. Ann Pharmacother. 2003;37:354–60. 6. Kjeldsen et al. Manuscript submitted to IJCP for review. 7. WHO Collaborating Centre for Drug Statistics Methodology. The anatomical therapeutic chemical (ATC) classification system. http://www.whocc.no/atc/structure_and_principles/. Accessed 16 Dec 2013. 8. Hepler CD, Strand LM. Opportunities and responsibilities in pharmaceutical care. Am J Hosp Pharm. 1990;47:533–43. 9. Hanlon JT, Schmader KE, Samsa GP, Weinberger M, Uttech KM, Lewis IK, et al. A method for assessing drug therapy appropriateness. J Clin Epidemiol. 1992;45:1045–51. 10. ASHP guidelines on a standardized method for pharmaceutical care. In: Deffenbaugh J, editor. Best practices for health-system pharmacy. Bethesda: American society of health-system pharmacists. 1996, pp. 109–11. 11. Pharmaceutical Care Network Europe. DRP-classification V6.2. http://pcne.org/sig/drp/documents/PCNE%20classification% 20V6-2.pdf. Accessed 16 Dec 2013. 12. Grønkjær LS, Jensen ML, Madsen H, Hallas J. Successful implementation of pharmaceutical intervention at an acute medical admission unit. Ugeskr Laege. 2011;173(19):1353–5. 13. Kjeldsen LJ, Olesen C, Truelshøj T, Nielsen LB. Quality assurance of medical treatment—an approach by Danish clinical pharmacists. EJHP Pract. 2011;17:31–4. 14. Mogensen CB, Olsen I, Thisted AR. Pharmacist advice is accepted more for medical than for surgical patients in an emergency department. Dan Med J. 2013;60(8):A4682. 15. Mogensen CB, Thisted AR, Olsen I. Medication problems are frequent and often serious in a Danish emergency department and may be discovered by clinical pharmacists. Dan Med J. 2012;59(11):A4532. 16. Munk CL, Bendixen HK, Kjeldsen LJ. Medication review with a focus on fracture prophylaxis among patients suffering collum femoris fractures. EJHP Pract. 2011;17:26–30. 17. Nielsen TR, Andersen SE, Rasmussen M, Honore´ PH. Clinical pharmacist service in the acute ward. Int J Clin Pharm. 2013;35(6):1137–51. 18. Bladh L, Ottosson E, Karlsson J, Klintberg L, Wallerstedt SM. Effects of a clinical pharmacist service on health-related quality of life and prescribing of drugs: a randomised controlled trial. BMJ Qual Saf. 2011;20:738–46. 19. Blix HS, Viktil KK, Moger TA, Reikvam A. Characteristics of drug-related problems discussed by hospital pharmacists in multidisciplinary teams. Pharm World Sci. 2006;28:152–8. 20. Khalili H, Farsaei S, Rezaee H, Dashti-Khavidaki S. Role of clinical pharmacists’ interventions in detection and prevention of medication errors in a medical ward. Int J Clin Pharm. 2011;33:281–4. 21. Spinewine A, Dhillon S, Mallet L, Tulkens PM, Wilmotte L, Swine C. Implementation of ward-based clinical pharmacy services in Belgium—description of the impact on a geriatric unit. Ann Pharmacother. 2006;40:720–8. 22. Westerlund T, Gelin U, Pettersson E, Ska¨rlund F, Wa˚gstro¨m K, Ringbom C. A retrospective analysis of drug-related problems documented in a national database. Int J Clin Pharm. 2013;35:202–9.
RESEARCH ARTICLE
Characterization of drug-related problems identified by clinical pharmacy staff at Danish hospitals Lene Juel Kjeldsen • Trine Birkholm • Hanne Fischer • Trine Graabæk • Karina Porsborg Kibsdal • Lene Vestergaard Ravn-Nielsen • Tania Holtum Truelshøj
Received: 12 December 2013 / Accepted: 24 March 2014 Koninklijke Nederlandse Maatschappij ter bevordering der Pharmacie 2014
Abstract Background In 2010, a database of drug related problems (DRPs) was implemented to assist clinical pharmacy staff in documenting clinical pharmacy activities locally. A study of quality, reliability and generalisability showed that national analyses of the data could be conducted. Analyses at the national level may help identify and prevent DRPs by performing national interventions. Objective The aim of the study was to explore the DRP characteristics as documented by clinical pharmacy staff at hospital pharmacies in the Danish DRP-database during a 3-year period. Setting Danish hospital pharmacies. Method Data documented in the DRP-database during the initial 3 years after implementation were analyzed retrospectively. The DRP-database contains DRPs reported at hospitals by clinical pharmacy staff. The analyses focused on L. J. Kjeldsen (&) H. Fischer The Danish Research Unit for Hospital Pharmacy, Amgros I/S, Dampfærgevej 22, 2100 Copenhagen Ø, Denmark e-mail: [email protected] T. Birkholm The Hospital Pharmacy, Roskilde, Region Zealand, Denmark T. Graabæk Centre for Clinical Research, Hospital South West Jutland, Esbjerg, Denmark K. P. Kibsdal The Hospital Pharmacy, Aalborg, North Denmark Region, Denmark L. V. Ravn-Nielsen Clinical Pharmacy Unit of Odense University Hospital, Odense, Denmark T. H. Truelshøj The Hospital Pharmacy, Aarhus, Central Denmark Region, Denmark
DRP categories, implementation rates and drugs associated with the DRPs. Main outcome measure Characteristics of DRPs. Results In total, 72,044 DRPs were documented in the DRP-database during the first 3 years of implementation, and the number of documented DRPs increased every year. An overall stable implementation rate of approximately 58 % was identified. The DRPs identified were multi-facetted, however evenly distributed for each of the 3 years. The most frequently identified DRP categories were: ‘‘Dose’’, followed by ‘‘Nonadherence to guidelines’’ and ‘‘Supplement to treatment’’. The highest implementation rates were found for the following DRP categories: ‘‘Non-adherence to guidelines’’ (79 %) followed by ‘‘Therapeutic duplication’’ (73 %) and ‘‘Dosing time and interval’’ (70 %). Even though the top 25 drugs were involved in 58 % of all DRPs, multiple drugs were associated with DRPs. The drugs most frequently involved in DRPs were paracetamol (4.6 % of all DRPs), simvastatin (3.0 %), lansoprazole (2.7 %), morphine (2.6 %) and alendronic acid (2.4 %). Conclusions The study found that a national database on DRPs contained multi-facetted DRPs, however evenly distributed for each of the 3 years. Even though the top 25 drugs were involved in 58 % of all DRPs, multiple drugs were associated with DRPs. The study emphasizes the importance of detecting and intervening for DRPs. Keywords Clinical pharmacy Database Denmark DRP Drug-related problem Hospital Medication review
Impact of findings on practice •
The analysis of the Danish DRP-database may assist in further developing clinical pharmacy in
123
Int J Clin Pharm
•
Denmark by identifying focus areas for intensified intervention. Data from the Danish DRP-database may be used for benchmarking between hospitals and over time.
all available data despite local differences, a national study on the quality and reproducibility of the database was conducted [6]. The results showed high inter-rater reliability indicating that local data may be pooled and used for national analyses [6].
Introduction
Methods
Suboptimal medication treatment may result in increased morbidity, mortality and costs [1, 2]. A method for improving the medication treatment in secondary care is by clinical pharmacists conducting medication related interventions such as medication reviews [3]. Medication reviews include identification of drug-related problems (DRPs), which may be defined as ‘‘a circumstance related to the patient’s use of a drug that actually or potentially prevents the patient from gaining the intended benefit of the drug’’ [4]. When conducting a medication review, recommendations of changes to the medication treatment based on the identified DRPs are made in order to optimize the medication treatment of the patient. A recent systematic review on medication reviews performed by clinical pharmacists in secondary care showed that a large variability existed in the proportion of DRPs identified in the 31 included studies; 0.13–10.6 per patients [3]. This variation may be explained by the heterogeneity of the study cohorts, the methodology used for the intervention including specific foci and the clinical experience of the clinical pharmacist. Usually, implementation of recommendations proposed by clinical pharmacists when conducting medication review is subject to acceptance by the treating physician. The systematic review also found a large variation in acceptance rates among the studies; 39–100 %, most likely due to the heterogeneity of the methods of the studies [3]. Studies in primary care have shown that categorization of DRPs may help developing counseling skills, assist in quality assurance, and is of value regarding pharmaceutical care practice and research [4, 5]. In Denmark, a national electronic tool to standardize DRPs was established in 2009/2010 (the DRP-database) and it currently contains 125,000 records. The DRP-database was established on request by clinical pharmacists throughout the country, and is available for all clinical pharmacists at hospital pharmacies in Denmark. The DRP-database is hosted and maintained by Amgros I/S, which is owned by the Danish regions. The DRP-database is used locally for documentation of activities at the wards, for quality assurance purposes and to identify focus areas to prevent DRPs. The DRP-database provides the opportunity to analyse data at a national level and develop national strategies to prevent DRPs. To explore whether analyses could be performed on
A project group consisting of five clinical pharmacists from hospital pharmacies representing various areas of the country and two clinical pharmacists from the Research Unit for Hospital Pharmacy designed and overviewed the current study. The five clinical pharmacists from hospital pharmacies ensured representation of issues related to regular use of the DRP-database.
123
The DRP-database DRPs are entered electronically by clinical pharmacy staff, and at present the database is predominantly used by clinical pharmacists. The DRP-database comprises information on each DRP by categorization, Anatomical Therapeutic Chemical (ATC) code [7], follow-up on the DRPs (documentation of whether the DRP was e.g. accepted or rejected) and a text field for documentation of additional information such as the recommendation made. In addition, the database contains information on demographics, date of registration of the DRP and number of drugs of the patient’s medication regimen. The DRPs can be categorized at two levels. The first level contains 14 categories; dose, dosing time and interval, side effect, interaction, drug form and strength, compatibility, non-adherence to guidelines, therapeutic duplication, drug allergies, length of treatment, supplement to treatment, electronic patient chart related (EPC-related), inappropriate drug, as well as no DRP identified. Each of the categories may be further characterized on a sublevel, e.g. ‘‘dose’’ may be subcategorized into ‘‘Dose too high’’ or ‘‘Dose too low’’ etc. ‘‘Non-adherence to guidelines’’ include several guidelines such as drug formularies, practice guidelines and institutional guidelines. ‘‘EPC-related’’ contain DRPs related to the set-up and use of the EPC, e.g. an unconscious patient with medications documented for self administration. Data for the analyses Data were retrieved retrospectively from the DRP-database for 3 years from the introduction of the database; July 2010 to June 2013, and comprised 124,718 records. The data were categorized into three time periods; (1) July 2010 to
Int J Clin Pharm Fig. 1 Actions taken on DRPs distributed on the three time periods. (Color figure online)
June 2011, (2) July 2011 to June 2012 and (3) July 2012 to June 2013. Results on the sublevel are not presented in the current study since the validity of the DRP-database has only been tested at the first level [6]. Additionally, the category ‘‘Compatibility’’ was excluded from the analyses, since it is rarely used (nine registrations during the study period). Finally, the category ‘‘No DRP identified’’ was excluded from the analyses (52,665 records). This category could have helped identify the proportion of patients with DRPs of the reviewed patients. However, a survey showed that all users do not consistently document patients with no DRPs identified in the database, hence the results would not be valid [6]. The remaining data, 72,044 records, were included for analyses. The DRP-database contains two unauthorized ATC codes; Z01 (=other) and Z02 (=not relevant). Z01 comprises a plethora of different DRPs—often related to supplement to treatment e.g. calcium, vitamin D etc. or recommendations regarding several drugs concomitantly, for which the clinical pharmacy staff may have found difficulties in identifying a specific ATC code. Z02 is primarily used in relation to lack of medication anamnesis, lack of registration of drug allergies or untreated indication, where no specific drug is involved. Data associated with these unauthorized ATC codes were excluded in the analyses when applicable. Approval of the study was received by the Danish Data Protection Agency.
Results During the study period, 72,044 DRPs were documented in the DRP-database. An increase in documentation of DRPs was reported during the 3 time periods, particularly from period 1–2; 15,901 in period 1, 27,203 in period 2 and 28,940 in period 3. About half of the DRPs were identified at emergency wards, approximately a third at medical wards, about 10 % at surgery wards, 1 % at psychiatric wards and only a few at intensive care units and paediatric wards. As well as for the sub-categories, the overall implementation rates of the DRPs were stable over the 3 years; approximately 58 % (Fig. 1). The prevalence of the DRP categories varied with the most frequently being ‘‘Dose’’, followed by ‘‘Non-adherence to guidelines’’ and ‘‘Supplement to treatment’’ (Table 1; Fig. 2). In addition, implementation rates of interventions also varied between the DRP categories with the highest being for ‘‘Non-adherence to guidelines’’ (79 %) followed by ‘‘Therapeutic duplication’’ (73 %) and ‘‘Dosing time and interval’’ (70 %) (Fig. 2). The lowest implementation rates were found for ‘‘Side effect’’ and ‘‘Drug allergies’’, both 37 % (Fig. 2). The drugs most frequently reported in the DRP-database were paracetamol (4.6 % of all DRPs), simvastatin (3.0 %), lansoprazole (2.7 %), morphine (2.6 %) and alendronic acid (2.4 %). Accordingly, these five drugs were present in top three for one or more of the individual DRP
123
Int J Clin Pharm Table 1 Prevalence of DRP categories of the three time periods pooled and including prevalence of the three most prevalent medications for each DRP category DRP category
Prevalence of DRP category
Dose
12,039 (17 %)
Non-adherence to guidelines
Supplement to treatment
11,179 (16 %)
10,997 (15 %)
Medication Paracetamol (N02BE01)
975 (8.1 %)
Zopiclone (N05CF01)
759 (6.3 %)
Pantoprazole (A02BC02) Lansoprazole (A02BC03)
Inappropriate drug
Electronic patient chart related
Interaction
Therapeutic duplication
Drug form and strength
Drug allergies
Side effect
Length of treatment
8,616 (12 %)
8,379 (12 %)
7,250 (10 %)
3,283 (4.6 %)
2,485 (3.4 %)
2,361 (3.3 %)
1,954 (2.7 %)
1,863 (2.6 %)
1,638 (2.3 %)
506 (4.2 %) 1,461 (13 %)
Paracetamol (N02BE01)
800 (7.2 %)
Omeprazole (A02BC01)
771 (6.9 %)
Calcium, combinations with vitamin D and/or other drugs (A12AX)
394 (3.6 %)
Simvastatin (C10AA01)
378 (3.4 %)
Paracetamol (N02BE01) Dosing time and interval
Most frequent within DRP category
Alendronic acid (M05BA04)
280 (2.5 %) 1,158 (13 %)
Cefuroxime (J01DC02)
613 (7.1 %)
Simvastatin (C10AA01)
596 (6.9 %)
Quinine (P01BC01)
278 (3.3 %)
Ibuprofen (M01AE01) Tramadol (N02AX02)
264 (3.2 %) 230 (2.7 %)
Warfarin (B01AA03)
635 (8.8 %)
Paracetamol (N02BE01)
265 (3.7 %)
Simvastatin (C10AA01)
169 (2.3 %)
Ciprofloxacin (J01MA02)
360 (11 %)
Warfarin (B01AA03)
218 (6.6 %)
Ferrous sulfate (B03AA07)
203 (6.2 %)
Paracetamol (N02BE01)
247 (9.9 %)
Morphine (N02AA01)
120 (4.8 %)
Tramadol (N02AX02)
109 (4.4 %)
Paracetamol (N02BE01)
317 (13 %)
Morphine (N02AA01)
282 (12 %)
Tramadol (N02AX02)
136 (5.8 %)
Morphine (N02AA01)
165 (8.4 %)
Benzylpenicillin (J01CE01) Beta-lactam antibacterials, penicillins (J01C)
138 (7.1 %) 109 (5.6 %)
Ibuprofen (M01AE01)
85 (4.6 %)
Citalopram (N06AB04)
72 (3.9 %)
Tiotropium bromide (R03BB04)
64 (3.4 %)
Prednisolone (H02AB06)
260 (17 %)
Pivmecillinam (J01CA08)
134 (8.2 %)
Ciprofloxacin (J01MA02)
49 (3.0 %)
ATC codes Z01 and Z02 not presented
categories. Indeed, paracetamol, the most frequent drug, could be found in top three for six DRP categories (Table 1). The high frequency of these individual drugs affected the prevalence of drugs at ATC level four (Table 2). The most prevalent drug class was ‘‘proton pump inhibitors’’, and lansoprazole comprised 41 % of this class followed by
123
pantoprazole (28 %) and omeprazole (21 %). Of the class ‘‘anilides’’, 97 % could be attributed to paracetamol, followed by paracetamol, combinations excl. psycholeptics (2 %) and paracetamol, combinations with psycholeptics (1 %). Finally, the third most prevalent class of ‘‘natural opium alkaloids’’ consisted of morphine (59 %), oxycodone (37 %), codeine, combinations excl. psycholeptics
Int J Clin Pharm Fig. 2 Number of implemented and not implemented DRPs distributed on DRP categories pooled for the three time periods. Not implemented ‘‘no action identified’’, ‘‘DRP accepted but no action identified’’, ‘‘DRP rejected’’ and ‘‘patient discharged’’. Implemented ‘‘DRP accepted and implemented by physician’’ and ‘‘DRP accepted and implemented by clinical pharmacy staff’’. (Color figure online)
(2 %) and a few other less prevalent drugs. The 25 most prevalent drugs associated with DRPs comprised 58 % of all DRPs. An example of a recommendation documented in the DRP-database of a frequent DRP related to paracetamol (ATC code N02BE01) categorized as ‘‘Dose’’ was: •
Reduction of paracetamol dose—risk of exceeding maximum recommended daily dose due to concurrent treatment with regularly prescribed as well as prn dose. The consequence might be effect on hepatic function.
Another example of a frequent case was one related to ibuprofen (ATC code M01AE01) and categorized as ‘‘Side effect’’: •
Recommend reviewing NSAID since this is not recommended for treatment of elderly due to increased risk of gastrointestinal side effects.
Clinical relevance of the DRP was not recorded in the DRP-database. However, a couple of examples of complex DRPs with significant clinical relevance are mentioned below (both categorized as ‘‘Inappropriate drug’’ with ATC codes A10BA02 and N06AA09, respectively): •
•
Recommend cessation of metformin due to poor kidney function (eGFR 24 ml/min). Metformin is contraindicated when eGFR is \30 ml/min due to risk of lactic acidosis. Recommend tapering off amitriptyline due to present acute myocardial infarction (AMI). Tricyclic antidepressants are contraindicated during the first 6 months following an AMI.
Discussion The study found 72,044 DRPs documented in the DRPdatabase during the first 3 years of implementation, and the number of documented DRPs increased every year. An overall stable implementation rate of approximately 58 % was identified. The DRPs identified were multi-faceted, however similarly distributed for each of the 3 years. Even though the top 25 drugs were involved in 58 % of all DRPs, multiple drugs were associated with DRPs. Frequency of DRP categories Comparison of DRPs between studies is a challenge, since a variety of DRP categorization methods exist in the literature by e.g. Hepler and Strand, Hanlon et al., American Society for Health-system Pharmacists, and the Pharmaceutical Care Network Europe [4, 7–11]. Of six selected Danish studies, two [12, 13] used a similar categorization system to characterize DRPs and allowed the results to be compared to the current evaluation [12–17]. The current study found the categories ‘‘Dose’’, ‘‘Nonadherence to guidelines’’ and ‘‘Supplement to treatment’’ as the most frequently identified DRPs. Somewhat similar findings were reported from a study at an emergency ward; ‘‘Non-adherence to guidelines’’ (67 %), ‘‘Inappropriate drug’’ (10 %) and ‘‘Supplement to treatment’’/‘‘Electronic patient chart related’’ (5 %) [13]. Contrarily, a study of seven wards in Horsens most frequently identified DRPs within the categories not specified as a category in the current categorization system (‘‘Medication reconciliation’’
123
Int J Clin Pharm Table 2 Prevalence of the 25 most prevalent drug classes at ATC level 4 during the 3 years Drug
ATC code at level 4a
Period 1
Period 2
Period 3
Total
Total (%) of all DRPs
Proton pump inhibitors
A02BC
1,031
1,849
1,762
4,642
6.4
b
Other
Z01
1,127
1,451
1,514
4,092
5.7
Anilides
N02BE
620
1,405
1,380
3,405
4.7
Natural opium alkaloids
N02AA
592
1,288
1,275
3,155
4.4
HMG CoA reductase inhibitors
C10AA
538
918
1,028
2,484
3.4
Platelet aggregation inhibitors excl. heparin
B01AC
572
835
857
2,264
3.1
Bisphosphonates Calcium, combinations with vitamin D and/or other drugs
M05BA A12AX
444 496
709 583
653 516
1,806 1,595
2.5 2.2
ACE inhibitors, plain
C09AA
384
490
545
1,419
2.0
Benzodiazepine related drugs
N05CF
353
535
516
1,404
1.9
Other opioids
N02AX
274
622
484
1,380
1.9
Selective serotonin reuptake inhibitors
N06AB
249
506
619
1,374
1.9
Second-generation cephalosporins
J01DC
196
476
563
1,235
1.7
Vitamin K antagonists
B01AA
348
413
456
1,217
1.7
Beta blocking agents, selective
C07AB
237
367
536
1,140
1.6
Benzodiazepine derivatives
N05BA
238
360
485
1,083
1.5
Beta-lactamase sensitive penicillins
J01CE
262
336
464
1,062
1.5
Sulfonamides, plain
C03CA
190
353
425
968
1.3
Osmotically acting laxatives
A06AD
175
345
414
934
1.3
Propionic acid derivatives
M01AE
175
373
349
897
1.2
Adrenergics and other drugs for obstructive airway diseases
R03AK
193
304
394
891
1.2
Glucocorticoids Anticholinergics
H02AB R03BB
183 187
372 319
288 290
843 796
1.2 1.1
Fluoroquinolones
J01MA
207
264
312
783
1.1
Other antidepressants
N06AX
155
274
336
765
1.1
a
Z02, i.e. DRPs non related to specific drugs such as e.g. lack of medication anamnesis, lack of registration of drug allergies or untreated indications were excluded from this table
b
Z01 is annotated to DRPs with no identifiable ATC code. However in practice, it is used for DRPs related to multiple medications and in particular to vitamins and minerals
and ‘‘Other’’) (35 %), followed by ‘‘Side effect’’ (18 %) and ‘‘Dose’’ (10 %) [13]. Despite using similar categorization systems, it is likely that the focus of the intervention, the ward type as well as the culture of the location lead to varieties in DRP identification within Denmark. Similar findings were identified in a selection of international studies [18–21]. Implementation of interventions To illustrate the proportion of DRPs, on which action was taken, implementation rates were presented i.e. DRP accepted and implemented by physicians as well as by clinical pharmacy staff. Acceptance rates are often used in the literature instead of implementation rates, however, when considering acceptance rates, the relevant DRPs are limited to recommendations requiring an action by physicians [3]. Additionally, only the acceptance of the
123
recommendation by the physician is recorded—not necessarily whether it was implemented. In the current study, an overall implementation rate of 58 % was discovered, which was similar to acceptance rates found in three Danish studies; 52, 53 and 59 %, respectively [13, 14, 16]. Another two Danish studies found higher acceptance rates; 76 and 80 %, respectively [12, 17]. Individual acceptance rates were available for two of the studies, and a significant cause for achieving an overall high acceptance rate in one of the studies [12] was attributed to the number of DRPs regarding ‘‘Non-adherence to guidelines’’, which to a certain extent may be implemented directly by clinical pharmacy staff (acceptance rate = 88 %) [12, 16]. Implementation of specific interventions may be specified in local contracts between the hospital pharmacy and the respective ward. Direct comparison of the individual acceptance rates was not possible due to the use of an alternative categorization
Int J Clin Pharm
method in one of the Danish studies, and low numbers for some of the categories in the other study [12, 16]. A review reported acceptance rates in international studies between 39 and 100 %, which may be explained by the different procedures in the studies [3]. In a Swedish national DRP-database in primary care, DRPs are categorized into actual and potential DPRs [22]. It is possible that acceptance rates may depend on whether the DRP is considered as actual or potential, however, the study of the Swedish DRP-database did not present any data on this. The Danish DRP-database contains a mixture of actual and potential DRPs, but no registration on this level of detail is recorded. Clinical significance Another factor, which might influence acceptance rates/ implementation rates, is clinical significance. The two Danish studies from Kolding assessed clinical significance of the identified DRPs, and they found in both studies that almost half of the DRPs (47 and 48 %, respectively) could be considered significant/critical [14, 15]. They also found a higher acceptance rate (67 %) among significant/critical DRPs than for the ones classified as minimal/moderate (52 %) [14]. The DRP-database does not contain information on clinical significance on the reported DRPs, which is a rather resource demanding process, hence the proportion of significant/critical DRPs is unknown. Even though these studies of clinical significance are small and may not be generalized to the DRP-database, they may indicate that a considerable amount of the DRPs reported in the DRPdatabase are of significant/critical clinical significance. This assumption is based on the fact that all DRPs identified in Kolding are reported to the DRP-database, and that the studies were performed at the emergency ward, which is the most common ward represented in the DRP-database. Consequently, whether the implementation rate of 58 % is a reasonable number is unknown. It is possible that a higher implementation rate could be achieved by focusing on significant/critical DRPs when identifying DRPs, which seems to increase the acceptance rate. Another strategy could be to increase the potential for clinical pharmacists themselves to implement actions on less significant DRPs. Development in DRPs and drug registrations over time As for DRP categories, no development over time was detected for drugs most frequently involved in DRPs. It is possible that this finding may be due to the stable distribution of ward types during the 3 years with emergency departments being the most dominant. Another possibility may be that identification of DRPs as a part of conducting medication reviews is a general service that despite local
and ward diversities do result in similar DRPs. In a Belgian study, the most frequently drugs associated with DRPs were ‘‘antithrombotic agents (B01; 9.1 % of all interventions), psycholeptics (N05—including antipsychotics, anxiolytics, hypnotics, sedatives; 8.8 %), psychoanaleptics, (N06— including antidepressants, antidementia drugs; 8.2 %), analgesics (N02; 6.9 %), and drugs for obstructive airway diseases (R03; 6.6 %)’’ [21]. These drug classes are presented at ATC level two, however drugs from the four most frequent drug classes are also present at the list of the 25 most frequently drugs at ATC level four associated with DRPs in the current study. This may further emphasize the similarities in drugs associated with DRPs, which could be explained by these drugs being of high risk to the patients. However, the similar distribution of DRPs during the 3 years may be important regarding the approach to DRPs in clinical practice and hence further development of the medication review service. It is likely that certain medications may continuingly be associated with DRPs e.g. high risk medications, but it is important to ensure continued development of the medication review service to include e.g. focus on DRPs associated with new drugs and new drug combinations to deliver the best possible service. Increase in documentation in the DRP-database over time During the 3 years studied, the number of DRPs documented in the DRP-database increased with more than 80 %, however, the increase between year 2 and 3 was modest compared to the increase between year 1 and 2. It is likely that the initial increase was caused by an increase in the number of clinical pharmacists using the DRP-database, while the increase between the second and third year may primarily be due to an increase in the number of medication reviews delivered and hence documented. Potential perspectives of the DRP-database An evaluation of the DRP-database revealed a good quality and generalisability of the data allowing for valid national analyses, even though the DRP-database was developed for local quality assurance activities and not for research purposes. The current study presented some data at the national level, which may inspire clinical pharmacists to make further use of the information in the DRP-database. Some examples include benchmarking over time and between locations and regions, identification of focus areas for interventions locally or nationally, sharing information about routine services between locations and regions, identification of potential patient safety issues when constructing or updating medication treatment guidelines or when making procurements of hospital medications.
123
Int J Clin Pharm
Limitations Since the DRP-database was developed on request as a tool to assist clinical pharmacists in their daily work, data entry in the DRP-database is based on voluntary reporting. Estimation indicates that about half of all clinical pharmacy services provided in Denmark are documented in the database, and some hospital pharmacies use the DRPdatabase to document 100 % of their recommendations [6]. Consequently, the DRP-database does not cover all services provided throughout the entire country, which may affect the generalisability of the data [6]. In addition, no central control is provided to ensure uniformity when entering data. Several issues may affect the recording of the data, which may be assessed differently between locations [6]. A proportion of data entry mistakes may be expected, since no double control and audits of the data are performed. In the current study, some records had to be excluded due to wrong dates being recorded, but the extent of data entry mistakes is unknown. Further limitations have been addressed elsewhere [6].
Conclusion The study found that a national database on DRPs contained multi-faceted DRPs, however similarly distributed for each of the 3 years. Even though the top 25 drugs were involved in 58 % of all DRPs, multiple drugs were associated with most DRPs. The study emphasizes the importance of detecting and intervening for DRPs. Acknowledgments Funding
None.
None.
Conflicts of interest of interest.
The authors declare that they have no conflict
References 1. van den Bemt PM, Egberts TC, de Jong-van den Berg LT, Brouwers JR. Drug-related problems in hospitalised patients. Drug Saf. 2000;22(4):321–33. 2. Kra¨henbu¨hl-Melcher A, Schlienger R, Lampert M, Haschke M, Drewe J, Kra¨henbu¨hl S. Drug-related problems in hospitals: a review of the recent literature. Drug Saf. 2007;30(5):379–407. 3. Graabaek T, Kjeldsen LJ. Medication reviews by clinical pharmacists at hospitals lead to improved patient outcomes: a systematic review. Basic Clin Pharmacol Toxicol. 2013;112(6):359–73.
123
4. Van Mil F, Westerlund T, Hersberger KE, Schaefer MA. Drugrelated problem classification systems. Ann Pharmacother. 2004;38:859–67. 5. Westerlund LOT, Handl WHA, Marklund BRG, Allebeck P. Pharmacy practitioners’ views on computerized documentation of drug-related problems. Ann Pharmacother. 2003;37:354–60. 6. Kjeldsen et al. Manuscript submitted to IJCP for review. 7. WHO Collaborating Centre for Drug Statistics Methodology. The anatomical therapeutic chemical (ATC) classification system. http://www.whocc.no/atc/structure_and_principles/. Accessed 16 Dec 2013. 8. Hepler CD, Strand LM. Opportunities and responsibilities in pharmaceutical care. Am J Hosp Pharm. 1990;47:533–43. 9. Hanlon JT, Schmader KE, Samsa GP, Weinberger M, Uttech KM, Lewis IK, et al. A method for assessing drug therapy appropriateness. J Clin Epidemiol. 1992;45:1045–51. 10. ASHP guidelines on a standardized method for pharmaceutical care. In: Deffenbaugh J, editor. Best practices for health-system pharmacy. Bethesda: American society of health-system pharmacists. 1996, pp. 109–11. 11. Pharmaceutical Care Network Europe. DRP-classification V6.2. http://pcne.org/sig/drp/documents/PCNE%20classification% 20V6-2.pdf. Accessed 16 Dec 2013. 12. Grønkjær LS, Jensen ML, Madsen H, Hallas J. Successful implementation of pharmaceutical intervention at an acute medical admission unit. Ugeskr Laege. 2011;173(19):1353–5. 13. Kjeldsen LJ, Olesen C, Truelshøj T, Nielsen LB. Quality assurance of medical treatment—an approach by Danish clinical pharmacists. EJHP Pract. 2011;17:31–4. 14. Mogensen CB, Olsen I, Thisted AR. Pharmacist advice is accepted more for medical than for surgical patients in an emergency department. Dan Med J. 2013;60(8):A4682. 15. Mogensen CB, Thisted AR, Olsen I. Medication problems are frequent and often serious in a Danish emergency department and may be discovered by clinical pharmacists. Dan Med J. 2012;59(11):A4532. 16. Munk CL, Bendixen HK, Kjeldsen LJ. Medication review with a focus on fracture prophylaxis among patients suffering collum femoris fractures. EJHP Pract. 2011;17:26–30. 17. Nielsen TR, Andersen SE, Rasmussen M, Honore´ PH. Clinical pharmacist service in the acute ward. Int J Clin Pharm. 2013;35(6):1137–51. 18. Bladh L, Ottosson E, Karlsson J, Klintberg L, Wallerstedt SM. Effects of a clinical pharmacist service on health-related quality of life and prescribing of drugs: a randomised controlled trial. BMJ Qual Saf. 2011;20:738–46. 19. Blix HS, Viktil KK, Moger TA, Reikvam A. Characteristics of drug-related problems discussed by hospital pharmacists in multidisciplinary teams. Pharm World Sci. 2006;28:152–8. 20. Khalili H, Farsaei S, Rezaee H, Dashti-Khavidaki S. Role of clinical pharmacists’ interventions in detection and prevention of medication errors in a medical ward. Int J Clin Pharm. 2011;33:281–4. 21. Spinewine A, Dhillon S, Mallet L, Tulkens PM, Wilmotte L, Swine C. Implementation of ward-based clinical pharmacy services in Belgium—description of the impact on a geriatric unit. Ann Pharmacother. 2006;40:720–8. 22. Westerlund T, Gelin U, Pettersson E, Ska¨rlund F, Wa˚gstro¨m K, Ringbom C. A retrospective analysis of drug-related problems documented in a national database. Int J Clin Pharm. 2013;35:202–9.
