Table 1.
Summary of educational intervention studies for prescribing
| Authors | Setting | Study design | Number of participants | Intervention | Prescribing outcome measures | Results | Potential for bias |
|---|---|---|---|---|---|---|---|
| Akici et al25 | Turkey | Randomised controlled trial | 12 GPs in intervention group; 13 GPs in control group | Short rational pharmacotherapy course based on the ‘problem-based Groningen/WHO model’ | Written examination with open and structured questions based on hypertensive cases as well as a question on osteoarthritis (unexposed indication) | Significant improvement in the mean test scores post-training of the intervention group (p<0.05) for both questions, showing a transfer effect. The improvement was maintained for at least 4 months after training | None declared |
| Butler et al43 | UK | Randomised controlled trial | 34 medical practices with 139 GPs were in the intervention group; 34 medical practices with 124 GPs in the control group | The intervention contained 7 parts. Six of these were online and included a reflection on their own practice, evidence and guidelines, novel communication skills and sharing experiences. Last, a face-to-face presentation of resistance trends throughout Wales and actual practices | Total numbers of oral antibiotic items dispensed for all causes per 1000 practice patients adjusted for the previous year's dispensing | A significant reduction of total oral antibiotic dispensing for the intervention group was observed compared with the control group (664 vs 681.1, p=0.02) | None declared |
| Celebi et al26 | Germany | Randomised controlled trial | 36 medical students in early intervention group; 38 medical students in late intervention group | A 1-week prescribing training module which comprised a seminar on common prescription errors, a prescribing exercise with a standardised paper case patient, drafting of inoperative prescription charts for real patients and discussions with a lecturer | Students were asked to make prescriptions for two virtual cases on a standard patient chart. These prescription charts were subsequently analysed by two independent raters using a checklist for common prescription errors | Prior to training, students committed a mean of 69±12% of the potential prescription errors. This decreased to 29±15% after prescribing training (p<0.001) | None declared |
| De Vries et al27 | Eight countries in Asia and Europe | Randomised controlled trial | 194 medical students in personal formulary (PF) group; 198 in existing formulary (EF) group; 191 in control group | The PF and EF groups were given teaching sessions based on the WHO Guide to Good Prescribing model (PF group=whole manual; EF group=manual minus p-drugs), with and without use of PF | Written examination using 16 patient cases based on four topics: hypertension, osteoarthritis, acute bronchitis, gastroenteritis | A significant increase in mean scores of the intervention group compared with the control group (p<0.05). The increase in the PF group was significantly higher than in the EF group. However, this effect was only visible in the universities in Yemen, the Russian Federation, and Indonesia. No significant differences between PF and EF scores were found in the universities in the Netherlands, Slovakia, Spain, India or South Africa | Funded by the VU University Medical Center and by the Department of Essential Drugs and Medicines Policy of the WHO |
| Degnan et al28 | UK | Randomised controlled trial | 9 medical students in intervention group; 35 in the control group | An online teaching module consisting of an interactive tutorial of 12 multiple-choice questions and three case studies covering pharmacokinetics, adverse drug reactions and drug doses calculations | OSCE station requiring administration of lidocaine and adrenaline for a patient with laceration and anaphylaxis | The teaching module significantly improved the students’ ability to calculate the correct volume of lidocaine (p=0.005) and adrenaline (0=0.0002) | Funded by the Association of Anaesthetists of Great Britain and Ireland |
| Esmaily et al29 | Iran | Randomised controlled trial | 58 GPs in intervention group; 54 GPs in control group | Education with an outcome-based approach utilising active-learning principles | Multiple choice and short answer questions, with two case scenarios and three ‘irrational’ prescriptions | There was an overall improvement of 26 percentage units in the prescribing knowledge and skills of GPs in the intervention group. No such improvements were seen in the control group | Additional funding from the National Public Health Management Centre in Tabriz and the Ministry of Health and Medical Education of Iran |
| Fender et al30 | UK | Randomised controlled trial | 54 GPs in intervention group; 46 GPs in control group | An educational package based on principles of ‘academic detailing’ | The appropriate prescribing of tranexamic acid, nonsteroidal antiinflammatory drugs and norethisterone | A proportionately higher level of appropriate prescribing was found in the intervention group. An increase of 63% in the prescription of tranexamic acid, the most effective first line treatment for menorrhagia, was observed in the intervention group | None declared |
| Gordon et al42 | UK | Randomised controlled trial | 76 junior doctors in intervention group; 86 in control group | A 1–2 h e-learning course on paediatric prescribing | Total correct responses on each prescribing assessment. Drug selection, prescribing calculations for children, discussing therapies and sources of errors were assessed | A significant increase in correct responses by the intervention group compared with the control group at both 4 and 12 weeks after the intervention. At 4 weeks: 79% vs 63% (p<0.0001) At 12 weeks: 79% vs 69% (p<0.0001) |
None declared |
| Hassan et al31 | Yemen | Randomised controlled trial | 56 medical students in intervention group; 44 students in control group | A prescribing course based on WHO's Guide to Good Prescribing, the Yemen Essential Drug List and Yemen Standard Treatment Guidelines | Written examination based on eight patient problems where a complete treatment plan form must be completed | Students from the study group performed significantly better than those from control in all problems presented and also when compared with the results of the pretest (p<0.05) | None declared |
| Hux et al32 | Canada | Randomised controlled trial | 135 GPs in intervention group; 116 GPs in control group | Mailed packages of prescribing feedback and guidelines-based educational materials | Median antibiotic cost and proportion of episodes of care in which a prespecified first-line antibiotic was used first | The median prescription cost remained constant in the feedback group but rose in the control group (p<0.002). First-line drug use increased in the feedback group but decreased in the control group (p<0.01) | Author receives salary support from the Institute for Clinical Evaluative Sciences in Ontario |
| Kahan et al33 | Israel | Randomised controlled trial | 32 physicians exposed to both interventions; 130 physicians who only received personalised letter; 29 physicians who only attended the lecture; 107 in the control group | Interventions were in the form of a lecture at a conference and a letter with personalised feedback to improve physicians’ rates of prescribing in the treatment of acute uncomplicated cystitis in adult women | Outcome was the rate of adherence to the guidelines for appropriate treatment using nitrofurantoin or second-line therapy of ofloxacin for 3 days | The letter intervention significantly influenced physicians’ prescribing patterns. The lecture intervention was only effective in the short run, indicating that the effect of this technique does not last unless reinforced | Partially funded through a research grant from The Israel National Institute for Health Policy and Health Services Research and through an educational grant from Schering Plough Israel |
| Midlöv et al34 | Sweden | Randomised controlled trial | 23 GPs in the intervention group; 31 GPs in the control group | Educational outreach visits | Number of prescriptions of benzodiazepines (BDP) and antipsychotics to the elderly | One year after the educational outreach visits there was a significant decrease in prescribing of medium-acting and long-acting BDP and total BDP in the active group compared with the control group (p<0.05). For antipsychotics there were no significant differences between active and control group | Funded by the Department of Primary Care Research and Development in the county of Skåne, Apoteket AB and the Faculty of Medicine, Lund University |
| Nsimba35 | Tanzania | Randomised controlled trial | 20 pharmacists in intervention facilities; 20 in control facilities | Posters, individual information and one-to-one training sessions | Simulated clients assessed the drug seller/pharmacist's knowledge and prescribing choices. A short examination was also conducted to assess participants’ knowledge of appropriate treatments for common childhood conditions | 85% of simulated clients who went to the intervention facilities were sold the first line drug sulfadoxine/pyrimethamine (SP) compared with 55% at control facilities (p<0.01). The intervention group also performed significantly better on the knowledge examination (p<0.01) | Funded by COSTECH-Tanzania |
| Ochoa et al36 | Cuba | Randomised controlled trial | 4 groups of 10 physicians (A, B, C, D) with A receiving community education programme and refresher training, B receiving refresher training, C receiving community education, D was the control group | Refresher training based on teaching sessions and periodic advisory visits. Community education involved group discussions and distribution of educational materials | Rate of overprescription of antibiotics for mild-acute respiratory infection (ARI) cases | Following the interventions, antibiotic overprescription rates declined by 26% and 63% in groups A and B, while increasing by 2% and 48% in groups C and D | None declared |
| Odusanya and Oyediran37 | Lagos state, Nigeria | Randomised controlled trial | Number of participants not specified. Primary healthcare workers (no doctors) in Mushin were in the intervention group; health workers in Ikeja were in the control group | 4-week training programme on rational drug use | Prescriptions were evaluated according to compliance to ‘standing orders’, which are a set of treatment modules. Drug use indicators were also compared | At the 2-week evaluation, the intervention group achieved a significant reduction in the average number of medicines prescribed compared with the control group. There was also a significant increase in the percentage of patients rationally managed from 18% to 30% (p=0.0005) in the intervention group. Improvements were not sustained at the 3-month evaluation | None declared |
| Rothmann et al38 | South Africa | Randomised controlled trial | 35 primary healthcare nurses in the intervention group; 31 in the control group | A competency-based primary care drug therapy training programme in the treatment of acute minor ailments | Written examination with 8 case studies including scenarios on acute gout, congestive heart failure, acute tonsillitis and infectious arthritis | Post-test results of the intervention group indicated significant improvement towards correct diagnosis and management of the conditions (p<0.05) | Funded by Boehringer Ingelheim (Pty) Ltd (Self-Medication Division) |
| Sandilands et al18 | UK | Randomised controlled trial | 50 medical students in the intervention group; 28 students in control group | Focused doctor-led and pharmacist-led practical prescribing teaching | Written prescribing examination consisting of six scenario-based questions | Teaching improved the assessment score of the intervention group: mean assessment 2 vs 1, 70% vs 62%, p=0.007; allergy documentation: 98% vs 74%, p=0.0001 and confidence. However, 30% of prescriptions continued to include prescribing errors | None declared |
| Scobie et al39 | UK | Randomised controlled trial | 16 medical students in intervention group; 16 students in control group | Practical structured teaching sessions led by a pharmacist | Nine station OSCE examination covering topics such as anticoagulation, intravenous administration, discharge prescription and medication history | The intervention group achieved higher scores in 8 OSCE stations. Four of these were statistically significant (p<=0.005) | None declared |
| Smeele et al40 | The Netherlands | Randomised controlled trial | 17 GPs in the intervention group; 17 GPs in control group | Four sessions (lasting 2 h each) of interactive group education and peer-review programme aimed at implementing national guidelines | Data on prescription of inhaled and anti-inflammatory medications were collected through self-recording by GPs and recording of repeat prescriptions for patients | No significant difference was found in the pharmacological treatment between intervention and control groups (p>0.05) | None declared |
| Webbe et al41 | UK | Randomised controlled trial | 13 preregistration house officers | A clinical teaching pharmacist programme to improve prescribing skills | Number of prescribing errors | A 37.5% reduction (p=0.14) in prescribing errors after pharmacist intervention | None declared |
| Akici et al47 | Turkey | Non-randomised comparative control | 50 medical students (interns) in intervention group; 54 interns in control group; 53 GPs | Problem-based rational pharmacotherapy education via the WHO/Groningen model | A written examination with open and structured questions based on case scenarios of tonsillitis and mild-to-moderate essential hypertension patients | Mean scores of the interns in the intervention group were higher than GPs, which were in turn higher than those of interns in the control group for all cases | Funded by a grant from Marmara University Scientific Research Projects Commission |
| Akram et al56 | Malaysia | Non-randomised comparative control | 18 final year dental students in the intervention group; 19 in the control group | Didactic lecture on how to write a complete prescription | Three case studies including irreversible pulpitis associated with a child, a pregnant woman and periapical pulpitis for an adult man. Assessed according to WHO's Guide to good prescribing | Significant improvement in the intervention group occurred compared with the control group in the following areas; date of issue, Rx symbol present, medicine legible, direction to use medicines, refill instructions, prescriber's signature, prescriber's date and prescriber's registration | Funded by the faculty of medicine, Universiti Kebangsaan Malaysia |
| Al Khaja et al57 | Bahrain | Non-randomised comparative control | 460 medical students over different stages of the degree were in the intervention group; 450 in the control group | A 2 h interactive session on prescription writing skills is presented followed by 5–6 case scenarios given as homework. Formative feedback on these cases was given to the students | A written examination. Physician-related components of the prescription assessed legality of prescription writing while drug-related components relate to the rational and appropriate use of medicines | Significantly higher scores were achieved by those that attended the interactive sessions compared with those that did not. 73.5% vs 59.5% (p<0.0001) | No funding received |
| Al Khaja et al48 | Bahrain | Non-randomised comparative cohort | 539 medical students | Problem-based learning curriculum incorporating a prescribing programme | A written examination. Physician-related components of the prescription assessed legality of prescription writing while drug-related components relate to the rational and appropriate use of medicines | Rate of physician-related components by students (years 2–4) was 96.1 (CI 94.1 to 97.5). However, the rate of various drug-related components was 50.2 (CI 46.0 to 54.4). No significant difference in overall performance of years 4 and 2 students (p=0.237). However appropriateness of drug-related components was significantly higher in year 4 than 2 (p<0.05) | None declared |
| Celebi et al49 | Germany | Non-randomised comparative control | 18 medical students who had never completed an internal-medicine clerkship; 38 students who had completed 1–4 weeks of clerkship; 18 students who had completed>5 weeks of clerkship | Internal medicine clerkship based on one general learning objective of ‘students are to be familiarised with caring for patients in an outpatient and inpatient setting’ | A written test comprising of the completion of prescription charts for two standardised patient paper cases. These were marked using a checklist for common prescription errors | Students committed 69%±12% of all possible prescription mistakes. There was no significant difference between the group without clerkships in internal medicine (G1) (71±9%), the group with 1–4 weeks (G2) (67±15%), and the group with more than 5 weeks of clerkships (G3) (71±10%), p=.76 | None declared |
| Coombes et al50 | Australia | Non-randomised comparative control | 99 medical students in intervention group; 134 in control group | Eight interactive problem-based tutorials covering topics such as antibiotics, anticoagulants, intravenous fluids, analgesics, oral hypoglycaemics and insulin | A written examination consisting of short answer questions on ADR identification, anticoagulants and analgesics | A significantly higher score was found in intervention students compared with controls; mean score in intervention group 29.46; control group 26.35 (p<0.05) | None declared |
| Franson et al51 | The Netherlands | Non-randomised comparative cohort | 181 medical students in baseline 2003 cohort, 285 students in 2004, 275 students in 2005, 264 students in 2006 | Students were taught to use a structured format called the Individualised Therapy Evaluation and Plan (ITEP) to communicate a therapeutic plan including the writing of a prescription | Written examination involving two different therapeutic cases; a simple paediatric case and a complex geriatric case | Students’ scores improved significantly in the 3 years after the introduction of the ITEP in the curriculum. The average score of the 2006 cohort was 6.76 compared with 3.83 for the 2003 group (p<0.0001) | None declared |
| Kozer et al52 | Canada | Non-randomised comparative control | 13 trainees in intervention; 9 trainees in control | 30 min tutorial focusing on appropriate methods for prescribing medications followed by a written test | Main outcome measure was the number of prescribing errors on medication charts completed after the tutorial | No significant difference in errors was found between the intervention group (12.4%) and the control group (12.7%) | Funded by the Trainee's Start-up Fund, The Research Institution, The Hospital for Sick Children, Toronto Canada |
| McCall et al44 | Australia | Non-randomised comparative control | 14 GPs in intervention group; 14 in control group | Completion of a Graduate Certificate in General Practice Psychiatry conducted primarily via distance education programme | A clinical audit assessed GPs’ recognition, drug management, non-drug management of patients with depression and anxiety | No effect on the intervention GPs prescribing habits (p>0.05) | None declared |
| Pandejpong et al53 | Thailand | Non-randomised comparative control | 38 continuity of care (CCC) participants; 52 non-CCC participants | CCC curriculum | Medical chart audits were performed and scored with a 12-task checklist of cardiovascular risk management including appropriate prescribing | There was a significant increase in ability to properly adjust antihypertensive medication and in the prescribing of aspirin as primary prevention for cardiovascular disease in the CCC group (p<0.05) | Funded by a Faculty of Medicines Siriraj Hospital Medical Education Research Grant, Mahidol University |
| Richir et al54 | The Netherlands | Non-randomised comparative control | 197 medical students in the intervention group; 33 students in control | A context-learning pharmacotherapy programme with role-play sessions and OSCE | A written examination involving the formulation of a treatment plan for two patients using the WHO six-step guide of rational prescribing | The mean score on the six steps of the WHO six-step plan for prescribing increased significantly for students who have received the pharmacotherapy study (p<0.001) | None declared |
| Shaw et al45 | Australia | Non-randomised comparative control | The number of junior doctors in intervention and control hospitals was not specified | Academic detailing including the provision of a bookmark containing the requirements for addictive medicines | Prescription error rates of addictive medicines were assessed. Errors were defined according to legal requirements for prescription of addictive medicines | At the intervention hospital, there was a significant decrease in error rate (from 41% to 24%, p<0.0001). The control hospital did not show a significant change in error rate over the same study period (p=0.66) | Partially funded by the Postgraduate Medical Council of NSW |
| Tamblyn et al55 | Canada | Non-randomised comparative cohort | 751 doctors from four graduation cohorts; 600 from before the intervention and 151 after the intervention | A community-oriented problem-based learning curriculum | Annual performance in diagnosis (difference in prescribing rates for specific diseases and relief of symptoms), and management (prescribing rate for contraindicated medicines) assessed using provincial health databases for the first 4–7 years of practice | After the intervention, graduates showed a significant fourfold increase in disease specific prescribing rates compared with prescribing for symptom relief. No difference in rate of prescribing for contraindicated medicines was observed | Funded by the Canadian Institutes of Health Research and Fonds de Recherche en Sante du Quebec |
| Volovitz et al46 | Israel | Non-randomised comparative control | 83 physicians attended the education programme. Four groups of patients were included. The study group had patients whose physicians attended the education programme and completed two follow-up physician visits. Three control groups of patients were also included | Asthma education programme involving lectures on pathophysiology, asthma management and prevention. Physicians were also asked to invite patients for three visits to reinforce the principles highlighted in the education programme | Changes in asthma medicine use were analysed before and after the intervention. Data were derived from the central database of Maccabi, Israel | In all four patient groups, a smaller proportion of reliever medicines (SABA) and a greater proportion of controller medicines (ICS and LABA) were used in the follow-up period compared with before the intervention. Patients in the study group were twice more likely to decrease their use of SABA than patients from the control group (p=0.042) | None declared |
| Wallace et al58 | UK | Non-randomised comparative control | 20 final year medical students in the intervention group; 11 in the control group | 8 tutorials on prescribing in acute clinical scenarios using peer assisted learning | Accurate completion of a prescription chart | The intervention group significantly improved after the intervention; median score was 47 before; 66 after (p<0.01). No significant change occurred in the control group (p=0.17) |
None declared |
| Aghamirsalim et al69 | Iran | Before and after study | 72 orthopaedic surgeons | Formal 2 h lectures once a week for 4 weeks and a 30 min refresher course was offered at the 4th month. Also, simplified osteoporosis guidelines were distributed | Proportion of patients with fragility factures who received appropriate treatment for osteoporosis | Significantly more patients were appropriately prescribed calcium and vitamin D supplements on discharge. 10% vs 91% (p<0.05). Significantly more patients were appropriately prescribed a bisphosphonate on discharge. 0.1% vs 73% (p<0.05) | None declared |
| Bojalil et al59 | Tlaxcala, Mexico | Before and after | 72 private GPs; 44 public GPs | A training course based on in-service practice. Other materials included the official training manuals for the control of diarrhoea and ARI, training videos and wall charts | Aspects of diarrhoea and ARI treatment which were evaluated and scored using a checklist | Private practitioners showed significant improvements in prescribing practices for children with diarrhoea. For ARI management, decisions on antimicrobial therapy and symptomatic drug use improved for both groups but only reached statistical significance for public physicians | Funded by the Mexican Social Security Institute |
| Chopra et al60 | Cape Town, South Africa | Before and after | 21 nurse prescribers | WHO and UNICEF's Integrated Management of Childhood Illness implementation. Training used the WHO/UNICEF teaching and assessment modules | A structured observation checklist of the case management of sick children including rational prescribing | There were significant improvements in the appropriate prescribing of antibiotics, with a significant reduction of inappropriate antibiotic use (62% vs 84%). However, there was no change in the treatment of anaemia or the prescribing of vitamin A to sick children |
None declared |
| Davey et al 61 | UK | Before and after | The number of junior doctors included in the study was not specified | A paediatric junior doctor prescribing tutorial conducted by a pharmacist and a bedside prescribing guideline to encompass the most frequently prescribed medications utilised on the children's unit | Prescribing errors and preventable adverse drug events | The introduction of the prescribing tutorial decreased prescribing errors by 46% (p=0.023). The introduction of a bedside prescribing guideline did not decrease prescribing errors | Author's research position was funded by Airedale NHS Trust |
| Elkharrat et al62 | France | Before and after | 27 doctors | Doctors were informed of the Drug Regulatory Agency prescribing guidelines of NSAIDs. Group sessions were held, posters were displayed and pocket sized, 10-page manuals were distributed | The rate of NSAID prescribing errors was analysed | Prescribing errors declined from 20% to 14% and when prescriptions were stratified by cause, the quality of prescribing increased significantly | None declared |
| Guney et al64 | Turkey | Before and after study | 101 medical students | Rational pharmacotherapy training based on the Groningen/WHO model | Prescription audit and OSCE examination based on a simulated patient case with uncomplicated essential hypertension | A significant improvement in prescription audit scores was observed after the training (p: 0.022) | None declared |
| Gall et al63 | UK | Before and after study | 212 GPs; 139 community nurses | Training on the use of guidelines on prescribing supplements | Changes in prescribing practice of supplements | Education significantly reduced total prescribing by 15% and reduced the levels of inappropriate prescribing from 77% to 59% due to an improvement in monitoring of patients prescribed supplements | Funded by South Thames Health Authorities Clinical Audit Programme |
| Leonard et al65 | USA | Before and after study | The number of clinical staff (physicians, nurses, pharmacists) included in the study was not specified | Educational patient safety initiatives using multiple interrelated educational and behavioural modification strategies | Assessment of medication orders which were then used to calculate the absolute risk reduction from prescribing errors | The absolute risk reduction achieved after the interventions was 38/100 orders written (t=25.735; p=0.001). This yielded an overall relative risk reduction from prescribing errors of 49% (p<0.001) |
Funded by the New York State Department of Health 2003 Patient Safety Award and by a donation from Lexi-Comp of Pediatric Lexi-Drugs limited licenses |
| Minas et al70 | Australia | Before and after study | GPs and healthcare prescribers in Emergency Departments and Sexual Health Clinics. Number included not specified | Treatment guidelines were distributed and informed through professional development sessions, letters and newsletters | Proportion of patients receiving non-occupational postexposure prophylaxis (nPEP) according to the relevant treatment guidelines | Significantly more patients that received nPEP met the eligibility criteria as stated in the relevant treatment guidelines after the educational intervention. 61.2% vs 90% (p<0.001) | None declared |
| Otero66 | Argentina | Before and after study | Number of participants not specified. Prescriptions for 95 patients were analysed in 2002 and for 92 patients in 2004 | Educational programme developed by the Patient Safety Committee of the Department of Pediatrics including the implementation of the ‘10 steps to reduce medication errors’ checklist | Prevalence of medication errors detected in written prescription orders during June 2002 (before intervention) and May 2004 (after intervention) | Prevalence of prescription errors was significantly lower in 2004 compared with 2002; 11.4% vs 7.3% (p<0.05) | None declared |
| Taylor et al67 | UK | Before and after study | 242 junior doctors | 1 h interactive, case-based educational programme regarding inpatient diabetes care | Number of insulin prescribing errors on medication charts observed after the tutorial | Insulin prescription errors were significantly lower after the intervention; 15.4% vs 7.8% (p<0.05) | None declared |
| Zgheib et al68 | Lebanon | Before and after study | 127 final year medical students divided into 18 groups of 6–7 students | 5 team-based learning (TBL) based on WHO's Guide to Good Prescribing. A lecture on the role of the pharmacist, Pharmacy and Therapeutics committees, formularies and hospital policies were also presented | Group formulary and prescription writing exercises occurred after each TBL session. Rationale for the selection of a drug and the format of the prescription was included in these assessments | Significant improvement in mean group scores for both formulary and prescription-writing exercises occurred over the 5 TBL sessions (p=0.002) | None declared |
GP, general practitioner.