Problem prescriptions in sweden necessitating contact with the prescriber before dispensing

Available online at www.sciencedirect.com Problem prescriptions in Sweden necessitating contact Anders Ekedahl, M.Sc.(Pharm.), Ph.D.(Med. aR&D department, National Corporation of Swedish Pharmacies (Apoteket AB), Apoteket Lejonet, bSchool of Pure and Applied Natural Sciences, University of Kalmar, Kalmar, Sweden Background: Pharmacists have an important role in detecting, preventing, and solving prescriptionproblems, which if left unresolved, may pose a risk of harming the patient.
Objectives: The aim was to examine prescription problems detected at pharmacies in Sweden, wherepharmacists consider it necessary to contact the prescribers for clarification, completion or correction ofthe prescriptions before dispensing, and to compare the intervention rates at public pharmacies at hospitals(PPHs) with those at city center pharmacies (CCPs).
Methods: All attempts to contact the prescriber about a prescription problem were recorded by trainedobservers (pharmacy students). Analyses were made of overall distribution of problem prescriptions,including data from all 14 participating pharmacies, and a comparison between CCPs and PPHs with datafrom the 5 areas, each consisting of 1 CCP and 1 PPH (10 pharmacies). Chi-square-analyses were used tocompare proportions, Spearman’s rank-correlation coefficient was used to test correlation between re-corded rates and dispensed volume, and Wilcoxon two-sample test was used to test differences betweenthe CCPs and PPHs. P ! .05 is regarded as statistically significant.
Results: The pharmacists contacted the prescribers for 1% of all new prescriptions before dispensing.
Errors that may compromise patient safety and medication outcome constituted almost 60% of theproblems. However, there was an inverse correlation between the intervention rates and the pharmacy’sdispensing volume. Significantly lower rates of problem prescriptions were recorded for women than formen. The highest rates were seen for prescriptions to patients younger than 15 years, and the ratesdecreased with increasing patient age. Pharmacists at PPHs contacted the prescribers about prescriptionproblems twice as often as those at large CCPs. Pharmacists spent an average of 5 minutes on the telephoneto solve the problem (median time), but 25% of the prescriptions took 10 minutes or more.
Conclusions: Computerized physician order entry (CPOE) and electronically transmitted prescriptions(ETP) can not only reduce the total rate of prescription problems, but also introduce new clinicallyimportant errors that may compromise patient safety and medication outcome. The prescription problemrates in the present study differed across prescriber groups and patient age and gender, and the inversecorrelation to pharmacy size indicates that all problems are not revealed and corrected and may thus reachthe patient. CPOE and ETP have been used extensively in Sweden for the past decade, but the present * Corresponding author. Tel.: þ46 70 545 1057.
E-mail address: .
1551-7411/09/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.sapharm.2009.09.001 Ekedahl/Research in Social and Administrative Pharmacy j (2009) 1–11 study indicates that there is still a potential and need for improvement for the vision of ‘‘no prescribingerrors/problems will reach the patient’’ to come true.
Ó 2009 Elsevier Inc. All rights reserved.
Keywords: Prescription errors; Prescriptions interventions; Pharmacy; Sweden form and how the prescription is generated(HWP; printed prescriptions; e-prescribing/com- According to the Institute of Medicine report puterized physician order entry (CPOE), or elec- ‘‘Preventing Medication Errors,’’ at least 1.5 tronically transmitted prescriptions [ETP]).
million preventable medication errors occur each Error rates are higher with HWPs, where ambigu- year in the United States, many of them in ous instructions/the risk for interpretation errors, and formal errors, such as missing prescriber or that w10% of all patients experience adverse patient data, may constitute a large part of the re- drug reactions, and more than one-third of all acute care admissions are due to drug-related In Sweden, all prescribers licensed to prescribe reimbursed drugs have a unique prescriber bar preventable. Prescribing errors are a common code (with identity, specialty, workplace, and cause for preventable medication errors and ad- employer/health care provider information), and verse drug events in primary care.Examples of it is mandatory that the prescriber bar code is present on the prescription. In Sweden, the first safety and medication outcome and requiring con- electronic prescription from the physician’s com- tact with the prescriber for clarification are puter directly to pharmacy computer for out- (1) wrong dosage of prescribed drug (wrong patients was transmitted in 1983. In 1995, virtually all primary health care centers in Sweden used computerized patient journals and electronic (3) wrong administration formula and strength; prescribing (e-prescribing)/computerized physi- (4) wrong amount of doses and/or duration of cian order entry (CPOE), and e-prescribing/ CPOE have constituted most prescriptions in (5) unclear or ambiguous instructions, where there is a risk of interpretation errors (hand- may choose a printout at the surgery (‘‘original written prescriptions, HWP; prescribers use paper prescription’’) or an ETP to the national of ‘‘self-defined abbreviations’’); and pharmacy server (the so-called national prescrip- (6) prescriptions issued to the wrong patient.
tion mailbox). CPOE may reduce the total num-ber of prescribing errors, formal errors, and Pharmacists have an important role in detect- errors of clinical importance. However, CPOE ing, preventing, and solving prescription prob- may introduce new clinically important errors, lems, which if left unresolved, may pose a risk of such as prescribing to the wrong patient or of harming the patient. In studies that judged the the wrong product.The national product regis- clinical importance of pharmacists’ actions, it was try (NPR) is a database with all licensed packs found that their actions were clinically relevant (and many pharmacy extemporaneous composi- and approved by the prescribers in most cases.
tions) available in Sweden. NPR is provided free However, the recorded rates of prescription prob- of charge to all pharmacies and prescribers/health lems at pharmacies reflect and are composed of care providers for e-prescribing/CPOE. Only prescription error rates, detection rates, and ac- packs present in NPR can be prescribed (no free tions taken as well as recording rates.
text for drug is possible or allowed), and all pre- Prescription error rates may vary with pre- scriptions have to comply with the national stan- scribers, licensing examination scores, medical dards. In Sweden, the licensed products in the specialty, experience, practice group structure computerized patient records are commonly pre- and culture, and the number of doctors involved sented in alphabetical lists, that is, pull-down menus (Ekedahl A, unpublished). Selection of may vary with the design of the prescription wrong line in the list (above or below the intended Ekedahl/Research in Social and Administrative Pharmacy j (2009) 1–11 product, a so called juxtaposition error) may con- contexts and varying numbers of participating phar- sequently result in a prescription for a different macies. Prescription forms vary between countries.
pack size, strength, administration formula or There are differences in used protocols, methods, and definitions. Inclusion criteria vary, some studies All ETPs in the national prescription mailbox include problems with all prescriptions (new pre- are accessible to dispense from any pharmacy in scriptions, renewals and refillsand others Sweden. However, ETP may introduce new er- only problems with new prescriptions25; some rors. Every pack of licensed drugs in Sweden (eg, studies include all problems (formal, technical, re- the Nordic countries) has a unique product and imbursement, delivery, and clinical problems), pack identifying number. When an ETP is trans- others some of the problems, and still other studies mitted from a physicians’ computer to pharmacy only ‘‘clinical problemsThere are also differ- computers in Sweden, only the product digit ences in how data are collected, prospective or number from the NPR for the prescribed product retrospectivand in how they are recorded, self- is transmitted. There are continuous changes in completed reports (when the dispensing pharmacist the list of licensed and marketed products, and does the classification and recording of the prob- packs provided to the Swedish drug market and the NPR is updated several times per week. This is Independent of the method used, the total incidence done automatically in pharmacy computers. How- of prescribing errors may be considerably higher ever, health care providers have to update their than that recorded in studies at pharmacies, as registries themselves because of legal reasons. If many errors are not detected and reach the pa- the NPR used at the physicians’ office is out of tientIt was hypothesized that prescription date, a prescription may be transmitted for problems rates differ across different prescriber a product (ie, product number), where there is groups and patient age and gender at pharmacist’s no matching product number in the actual NPR prescription review and that there are differences at the pharmacies, resulting in an empty field on in problem rates between different types of the ETPd‘‘no drug.’’ A printout at the office, on the other hand, will result in the product printed The aim was to study the prescription prob- lems detected at pharmacies in Sweden, where Pharmacists’ detection, recording, and action pharmacists’ consider it necessary to contact the rates of prescription errors are influenced by prescribers for clarification, completion, or cor- pharmacy size, location, organization, workload, rection of the prescriptions before dispensing, and available information (such as presence on the to compare the intervention rates at public prescription of indication or intended use), access pharmacies at hospitals (PPHs) with city center to patient record or patient medication profile, attendance of the patient or a representative atthe pharmacy to have the prescription dispensed,and the individual pharmacists’ education and traininHigher detection rates have beenrecorded when pharmacists have access to more information, such as the patients themselves, mainly in the mid and northern regions of medication profile, and data of the intended use/ Sweden, 7 PPHs (2 at university hospitals and 5 indication on the prescription.The detection of at county hospitals) and 8 large CCPs, including 1 problems as well as the actions the pharmacist takes CCP from the same or neighboring city, as each to solve the identified problems may depend on be- one of 6 PPHs, were invited to participate in the ing able to consult with the patient personally when study. One CCP declined to participate, and the the patient visits the pharmacy to have prescriptions material constitutes 14 pharmacies, 7 PPHs, and 7 dispensed, as patients increasingly initiate and make large CCPs, 5 areas with 1 CCP and 1 PPH.
pharmacists aware of problemActions are also influenced by the collaborate climate between phar- weeks per pharmacy, 15 weekdays, February 2007 to February 2008. Data for the PPH and There is a large variation in recorded prescrip- the CCP in the same area were collected concom- tion problems, not only between studies but also itantly. The areas were allocated to different collection periods to decrease variation because performed in different settings with differing of seasonal variations in prescribing.
Ekedahl/Research in Social and Administrative Pharmacy j (2009) 1–11 Recording was performed by trained observers to compare proportions; Spearman’s rank-corre- (pharmacy students). The observers followed the lation coefficient was used to test correlation be- pharmacists closely to record all contact attempts tween recorded rates and dispensed volume for with the prescribers and attached a copy of the the 14 participating pharmacies and Wilcoxon prescription to the protocol. Included were prob- two-sample test was used to test differences be- lem prescriptions for medicinal products licensed tween the CCPs and PPHs. P ! .05 is regarded for human use, which contained errors, ambigui- ties, or other problems, such that the pharmacistjudged it necessary to contact the prescriber forclarification before dispensing, correction, comple- tion, or change. All telephone call attempts tocontact the prescribers during the study period Pharmacists judged it necessary to contact the were included in the study, whether or not the prescriber for correction and/or clarification be- pharmacist succeeded to reach the prescriber. Pre- fore dispensing for 0.6 Æ 0.3% (mean Æ SD) of all scriptions for nonpharmaceuticals and prescrip- dispensed prescriptions and 1.0 Æ 0.5% of all new tions for animals were excluded from the study.
prescriptions (see There was a 6-fold var- Prescribing errors and the corresponding in- iation in rates and a significant, inverse correla- terventions were reported on a form originally tion between the pharmacy’s dispensing volume developed in the United States and translated and and the rates for all dispensed prescriptions adapted to the Nordic context.The author ex- (rs ¼ À0.6769; P ¼ .008), for all new prescriptions amined all cases. Irregularities in classification were discussed with the observers for consensus (rs ¼ À0.6072; P ¼ .021), lower the rates the larger Data on dispensed prescriptions at the partici- There were also differences across patient age pating pharmacies during the study periods were obtained from National Corporation of Swedish seen for prescriptions to patients younger than Pharmacies (Apoteket AB). Data were coded and 15 years, and the rates decreased with increasing entered into a database (Microsoft AccessÒ) for calculations and cross-tabulations. Two analyses were made: (1) overall distribution of problem (c2 ¼ 4.01; P ! .05) and women, had the rates prescriptions, where data from all 14 participating been proportional to all dispensed prescriptions pharmacies are included and (2) a comparison be- (59.4% for women; c2 ¼ 9.565; P ! .01).
tween CCPs and PPHs, wherein data from the 5 areas with both 1 CCP and 1 PPH (10 pharma- than 90% of all intervention contacts with the cies) are included. Chi-square-analyses were used prescribers. The most frequent problems (40%) Table 1Problem prescriptions where pharmacists judged it necessary to contact the prescriber before dispensing at 14 pharmaciesin Sweden All dispensed prescriptions (n ¼ 103,654) Electronically transmitted prescriptions, ETPs Previously dispensed prescriptions (n ¼ 46,569) a 131 (22.9%) general practitioners; 370 (64.6%) hospital physicians; 72 (12.6%) other prescribers.
b Prevalence problem prescription of dispensed prescriptions.
c The ETP share of all new prescriptions is 63.02 Æ 14.26 (mean Æ SD).
d 108 handwritten new prescriptions.
Ekedahl/Research in Social and Administrative Pharmacy j (2009) 1–11 Table 2Problem prescriptions across age and gender at 14 pharmacies in Sweden 63.4% of all dispensed prescriptions in Sweden 2007 were issued to females, compared with 60.8% at CCPs and 57.2% at PPHs in the study. 43.7% of all dispensed prescriptions in Sweden 2007 were issued to patients O65 years,compared with 38.2% at CCPs and 39.1% at PPHs in the study.
a Prevalence problem prescription of dispensed prescriptions.
were related to wrong product (wrong drug, ). All types of prescription problems were erroneous strength, wrong administration form, or wrong amount of doses) followed by insuffi- When the pharmacist contacted the prescriber, cient information concerning drug schedule in- the suggestion by the pharmacist was accepted in structions for use of medicine (24%) (see most cases, and only few (6%) were rejected One-third of the problems with ‘‘wrong drug’’ (); however, the pharmacists failed to get were the prescribing of drugs that were withdrawn in touch with the prescriber in one-fifth of the (not licensed anymore) from the market. Problems cases. These cases were discussed with either an- with availability of the prescribed drugs, namely other physician, a nurse at the ward/surgery, or out of stock at the wholesaler or at the pharmacy, a secretary. The pharmacists spent on average were common (14%). Problems with potential 5 minutes on the telephone to solve the problem clinical hazards constituted about two-thirds of (median time), but for 25% of the prescriptions, all recorded problems. Noteworthy are the pre- scriptions issued to the wrong patient. Few errorswere identified concerning potential risk for inter-actions, contraindications, and side effects of med- icines. Handwritten prescriptions constituted 108of 186 (58%) of the prescriber contacts for new The present study indicates that pharmacists in paper prescriptions (see ). The interven- Sweden consider it necessary to contact the pre- tions were more frequent for ETPs compared scriber for clarification, correction, completion, or with new CPOE prescriptions (printout of the pre- change before dispensing for about 1% of all new scription at the surgery), except for formal prescriptions, corresponding to approximately 330,000 prescription problems per year. This is Pharmacists at PPHs contacted the prescriber a somewhat higher rate than reported by Hulls significantly more often than pharmacists at CCPs for all dispensed prescriptions (z ¼ À2.611; Rupp et al.In one-fifth of the cases, the pharma- P ! .01), for all new prescriptions (z ¼ À2.402; cist could not reach the prescriber but talked to P ! .05), as well as for ETPs (z ¼ À2.402; P ! .05) someone else at the surgery, similar to the figures ). The contact rates were significantly higher reported by Rupp et al,but higher than that re- at PPHs than CCPs for both genders and the age groups 15-24; 25-44; 45-64, and 64-74 years used self-completed reports, whereas the present Ekedahl/Research in Social and Administrative Pharmacy j (2009) 1–11 Table 3Problems encountered on ETPs and new paper prescriptions at 14 pharmacies in Sweden HWP, handwritten prescriptions; nda, no data available.
a Prevalence problem prescription of dispensed prescriptions.
b More than one problem per prescription may have been recorded.
study and the study by Rupp et alused trained the investigation period. The overall intervention observers (students). On one hand, studies using rate was about 10 times higher than the contact self-completed forms may underestimate the prob- rate with prescribers. About 90% of all problems lem rate because of noncompliance with the study with ETPs were that the dosage text had to be protocol, due to high workload and forgetfulness.
clarified and edited, mainly due to prescribers’ On the other hand, independent observers can de- tect obvious actions, such as telephone contacts Many studies have reported a large variation, with the prescriber, but may have difficulties to 10-20 fold or higher, between participating phar- detect other corrections made. However, total rates of interventions in many studies are consid- est and the highest rates in the present material is erably higher than contact rates.At one of somewhat lesser. However, there was an inverse the participating pharmacies, prescription inter- correlation between pharmacy size (dispensing ventions made without contacting the prescriber volume) and contacts with the prescriber, in were also recorded (self-completed reports) during accordance with the findings by Rupp et Ekedahl/Research in Social and Administrative Pharmacy j (2009) 1–11 Table 4Dispensed prescriptions and prescription problems at 5 city center pharmacies and 5 public pharmacies at hospitals inSweden RS, rank-sum.
a Prevalence problem prescription of dispensed prescriptions.
contacted the prescriber about twice as often as pharmacists at CCPs. However, prescription Rupp et concluded that their findings sug- problems are common with HWPs, and 85% of gested that pharmacists’ willingness or ability to all HWP-problems occurred at PPHs, constituting intervene in problematic new prescriptions de- 27% of all problem prescriptions compared with creases as workload increases. Knapp et alhy- 7% at CCPs. Furthermore, 85% of the problem prescriptions at PPHs were issued by hospital a benchmark level of intervention rate. Although physicians compared with 31% at CCPs. The pa- it is reasonable to believe that the variation do re- tient selection and their drug treatment differ flect differences in training, work routines, and fo- between PPHs and CCPs, as many patients at cus by the pharmacists, the present material PPHs are home-going after hospital care.
indicates that there are differences in problem Computerized physician order entry may not rates between different pharmacies due to varia- only reduce the total number of prescribing errors tions among prescribers (GPs, specialists, hospital but also introduce new clinically important errors, physicians), prescriptions (HWP, CPOE, ETP), such as prescription of the wrong drug or to the and patients and the ability to consult with the pa- wrong patient. Problems with ETPs were slightly tient personally. In general, pharmacists at PPHs Ekedahl/Research in Social and Administrative Pharmacy j (2009) 1–11 Table 5Prescription problems across age and gender at 5 city center pharmacies and 5 public pharmacies at hospitals in Sweden However, problems with HWPs constituted most interventions for new paper prescriptions. No for patients older than 65 years and women were recording is made, but HWPs constitute a small recorded, had the problem prescriptions been pro- share of the dispensed prescriptions in Sweden, as portional to all dispensed prescriptions. The de- most new paper prescriptions are CPOE prescrip- tection of problems may depend on being able tions. In a sample at one of the PPHs, HWPs to consult with the patient personally, as patients constituted 12% of all dispensed prescriptions and increasingly initiate and make pharmacists aware 47% of the problem prescriptions compared with of problems.Many patients older than 65 years, 19% of the problem prescriptions at the other 13 many of them women, do not visit the pharmacy participating pharmacies, indicating that HWPs themselves to have their prescriptions dispensed; may constitute about 5% of all dispensed pre- rather the medicine is collected by a representative.
scriptions. This corresponds to an intervention The results could indicate that the lower detection rate for CPOE prescriptions of about half that of rates were due to lack of information.
ETPs in the present study, similar to the findings Problems that may compromise patient safety, in a study at 3 mail-order pharmacies in such as wrong product and ‘‘dosage/dosage One reason for the higher problem rate with ETPs schedule,’’ constituted almost 60% of the prob- is the presence of ‘‘out of date’’ NPRs at the phy- lems when the pharmacist contacted the pre- sicians’ offices. A prescription transmitted for scriber. The most frequent problems, 39% in the a product with no matching product number in present material, were related to wrong product the actual NPR at the pharmacies results in an (wrong drug, erroneous strength, administration empty field on the ETP (‘‘no drug’’). On a pre- form, package size), which is similar to the scription printout at the surgery on the other findings in previous studies.Examples in hand, the product is printed on the prescription, the present material of wrong drug (juxtaposition which usually also contains the full text corre- errors) are methotrexate instead for metformin; sponding to ‘‘self-defined abbreviations.’’ The pre- DurogesicÒ (fentanyl patches) instead of Duro- scriber may also discover and correct errors feronÒ (iron tablets). However, few errors were before signing and handing the prescription to identified concerning potential risk for contraindi- cations, adverse drug reactions, or drug-drug in- The highest intervention rates were seen for teractions, the latter contrasting to the reported prescriptions to patients younger than 15 years high incidence of drug-drug interactions in recent and rates decreased with increasing patient age, which is in accordance with the findings by pharmacies in Sweden did not have access to Ekedahl/Research in Social and Administrative Pharmacy j (2009) 1–11 Problems encountered on new prescriptions at 5 city cen- Prescriber action to suggestion by pharmacist for new ter pharmacies and 5 public pharmacies at hospitals in were only rejected in about 6% of the cases.
However, the pharmacists in this study spent on average about 5 minutes on the telephone to solve problems, increasing patients’ waiting time. Time is a cost to both pharmacies and the prescribers/health care providers, and the waiting time is in- creased, not only for the actual customer, but may affect waiting time for other customers as well. It is thus in the interest of all participants not only to increase quality in prescribing and de- crease error rates but also to limit contacts to problems where there is a need to involve the There are certain limitations of the study. One of them is the study sample. One aim was to compare 2 different types of pharmacies. The invited CCPs were selected from the same or neighboring city as the participating PPHs in the study, but pharmacies from the 3 largest cities in Sweden and small and mid-size pharmacies were not included. Another limitation is the point of a Prevalence problem prescription of dispensed measure in the present material and the decision by the pharmacist to contact the prescriber to b More than one problem per prescription have been solve one or more problems with the prescription.
It is emphasized that the material only representa fraction of all prescription problems, as not allproblems are revealed or result in a decision tocontact the prescriber before dispensing. How- patient medication profiles, and information in ever, telephone contacts with prescribers are easily the dosage text on the indication/intended use accessible by independent observers (students), was only present for about 50% of prescriptions.
and one observer can record all contacts with This may also indicate low detection or/and ac- the prescribers occurring at one pharmacy. As the tion rates at pharmacies and that many prescrip- incidence of problems requiring contact with the prescriber is low (occurred at an average of 1.5-2 Pharmacies have to balance expeditiousness times per day per pharmacy), each pharmacist had with safety when serving patients. In accordance only few problem prescriptions, where the pre- with previous studies, the present material in- scriber was contacted before dispensing. Self- dicates that pharmacists’ contacts with prescribers completed reports may consequently introduce are relevant, as the suggestions by the pharmacists problems of forgetting to record the contacts as Ekedahl/Research in Social and Administrative Pharmacy j (2009) 1–11 well as low adherence to the study protocol, due modifications of prescription errors. Br J Clin Phar- to high workload and inconvenience with the protocol. However, self-completed reports may 7. Claesson C, Burman K, Nilsson JLG, Vinge E. Pre- be advantageous if all detected problems and scription errors detected by Swedish pharmacists.
Int J Pharm Pract 1995;3:151–156.
8. Hulls V, Emmerton L. Prescription interventions in New Zealand community practice. J Soc AdminPharm 1996;13:198–204.
9. Tamblyn R, Abrahamowicz M, Brailovsky C, et al.
CPOE and ETP can not only reduce the total Association between licensing examination scores rate of prescription problems but also introduce and resource use and quality of care in primarycare practice. JAMA 1998;280:989–996.
new clinically important errors that may compro- 10. Tamblyn R, Abrahamowicz M, Dauphinee WD, mise patient safety and medication outcome. The et al. Association between licensure examination prescription problem rates in the present study scores and practice in primary care. JAMA 2002; differed across prescriber groups and patient age and gender, and the inverse correlation to phar- 11. Buurma H, de Smet PA, van den Hoff OP, macy size indicates that all problems are not Egberts AC. Nature, frequency and determinants of revealed and corrected and may thus reach the prescription modifications in Dutch community patient. CPOE and ETP have been used exten- pharmacies. Br J Clin Pharmacol 2001;52:85–91.
sively for the last decade in Sweden, but the 12. Shah NH, Aslam M, Avery AJ. A survey of prescrip- present study indicates that there is still a potential tion errors in general practice. Pharm J 2001;267:860–862.
and need for improvement for the vision of ‘‘no 13. Gron P, Kennbo I. [Errors in prescriptions and con- prescribing errors/problems will reach the pa- trol of prescriptions.] Ugeskr Laeger 1989;151:3385– 14. Kralewski JE, Dowd BE, Heaton A, Kaissi A. The influence of the structure and culture of medicalgroup practices on prescription drug errors. Med I am indebted to the student observers who collected the data for this project and to the staff at the pharmacies participating in the study. I also Crombie IK. A classification of prescription errors.
want to thank Dr Svein Haavik for his valuable J R Coll Gen Pract 1989;39:110–112.
16. Westein MP, Herings RM, Leufkens HG. Determi- nants of pharmacists’ interventions linked to pre-scription processing. Pharm World Sci 2001;23:98– 17. Bizovi KE, Beckley BE, McDade MC, et al. The 1. Aspden P, Wolcott J, Bootman JL, Cronenwett LR, effect of computer-assisted prescription writing on eds. Preventing Medication Errors. Washington, DC: emergency department prescription errors. Acad 2. Isacson D, Johansson L, Bingefors K. Nationwide 18. Cohen MR, Davis NM. Complete prescription survey of subjectively reported adverse drug reac- orders reduce medication errors. Am Pharm 1992; tions in Sweden. Ann Pharmacother 2008;42:347–353.
3. Fryckstedt J, Asker-Hagelberg C. [Drug-related 19. Koppel R, Metlay J, Cohen A, et al. Role of comput- problems common in the emergency department of erized physician order entry systems in facilitating internal medicine. The cause of admission in almost medication errors. JAMA 2005;293:1197–1203.
every third patient according to quality follow-up.] 20. Varkey P, Aponte P, Swanton C, Fischer D, Johnson SF, Brennan MD. The effect of computer- 4. Kuo GM, Phillips RL, Graham D, Hickner JM.
ized physician-order entry on outpatient prescription Medication errors reported by US family physicians errors. Manag Care Interface 2007;20:53–57.
and their office staff. Qual Saf Health Care 2008;17: 21. Oliven A, Michalake I, Zalman D, Dorman E, Yeshurun D, Odeh M. Prevention of prescription errors by computerized, on-line surveillance of Chrystyn H. Clinical pharmacy interventions by drug order entry. Int J Med Inform 2005;74:377– community pharmacists during the dispensing pro- cess. Br J Clin Pharmacol 1999;47:695–700.
6. Buurma H, De Smet PA, Leufkens HG, Egberts AC.
de Escalza P, Odriozola I, Gastelurrutia MA. Assess- Evaluation of the clinical value of pharmacists’ ment of poorly written prescriptions that reach Ekedahl/Research in Social and Administrative Pharmacy j (2009) 1–11 a community pharmacy. Seguimiento Farmacotera- 32. Haavik S, Horn AM, Mellbye KS, Kjonniksen I, Granas AG. [Prescription errorsddimension and 23. Astrand B. ePrescribing-Studies in Pharmacoinfor- measures.] Tidsskr Nor Laegeforen 2006;126:296–298.
matics [Ph D thesis]. University of Kalmar, Sweden, 33. Quinlan P, Ashcroft DM, Blenkinsopp A. Medica- tion errors: a baseline survey of interventions 24. Ballentine AJ, Kinnaird D, Wilson JP. Prescription recorded during the dispensing process in community errors occur despite computerized prescriber order pharmacies. Int J Pharm Pract 2002;10(Suppl):R67.
entry. Am J Health Syst Pharm 2003;60:708–709.
34. Leemans L, Veroeveren L, Bulens J, et al. Frequency 25. Rupp MT, DeYoung M, Schondelmeyer SW. Pre- and trends of interventions of prescriptions in Flem- scribing problems and pharmacist interventions in ish community pharmacies. Pharm World Sci 2003; community practice. Med Care 1992;30:926–940.
Foster S. Effect of patient information on the quality Kulkarni S, Formica RN Jr. Medication errors in of pharmacists’ drug use review decisions. J Am the outpatient setting: classification and root cause Pharm Assoc (Wash) 2000;40:500–508.
analysis. Arch Surg 2007;142:278–283. discussion 284.
27. Kuyper AR. Patient counseling detects prescription 36. Gandhi TK, Weingart SN, Seger AC, et al. Outpa- errors. Hosp Pharm 1993;28:1180–1181. 1184–1189.
tient prescribing errors and the impact of computer- 28. Benrimoj SI, Langford JH, Berry G, et al. Economic ized prescribing. J Gen Intern Med 2005;20:837–841.
impact of increased clinical intervention rates in com- 37. Kennedy AG, Littenberg B. Medication error report- munity pharmacy. A randomised trial of the effect of ing by community pharmacists in Vermont. J Am education and a professional allowance. Pharmacoe- 38. Astrand B, Montelius E, Petersson G, Ekedahl A.
Assessment of ePrescription quality: an observa- Albrant DH. Community pharmacist interventions tional study at three mail-order pharmacies. BMC in a capitated pharmacy benefit contract. Am J Health Syst Pharm 1998;55:1141–1145.
39. Astrand B, Astrand E, Antonov K, Petersson G.
30. Mandt I, Horn A, Ekedahl A, Granas A. Community Detection of potential drug interactionsda model pharmacists’ prescription intervention practicesd for a national pharmacy register. Eur J Clin Pharma- exploring variations in practice in Norwegian phar- macies. Res Social Adm Pharm; Online first, epub 40. Johnell K, Klarin I. The relationship between num- ber of drugs and potential drug-drug interactions in 31. Kennedy AG, Littenberg B. A dictation system for the elderly: a study of over 600,000 elderly patients reporting prescribing errors in community pharma- from the Swedish prescribed drug register. Drug Saf cies. Int J Pharm Pract 2004;12:13–19.

Source: http://itivarden.idg.se/polopoly_fs/1.272522.1259067284!ae920020d01.pdf