Article Text

Download PDFPDF

Who makes prescribing decisions in hospital inpatients? An observational study
  1. Sarah Ross1,
  2. Laura Hamilton1,
  3. Cristin Ryan2,
  4. Christine Bond3
  1. 1Division of Medical and Dental Education, University of Aberdeen, Aberdeen, UK
  2. 2School of Pharmacy, Queen's University Belfast, UK
  3. 3Centre for Academic Primary Care, University of Aberdeen, Aberdeen, UK
  1. Correspondence to Dr Sarah Ross, Senior Clinical Teaching Fellow, Division of Medical and Dental Education, University of Aberdeen, Polwarth Building, Foresterhill, Aberdeen, AB25 2ZD, UK; s.ross{at}abdn.ac.uk

Abstract

Purpose of the study Errors involving drug prescriptions are a key target for patient safety initiatives. Recent studies have focused on error rates across different grades of doctors in order to target interventions. However, many prescriptions are not instigated by the doctor who writes them. It is important to clarify how often this occurs in order to interpret these studies and create interventions. This study aimed to provisionally quantify and describe prescriptions where the identity of the decision maker and prescription writer differed.

Design of the study Observational data was collected in six wards, 2 weeks per ward, at a single large UK teaching hospital over a 12-week period from January to April 2011.

Results In 112/183 (61%) cases where a new medicine was prescribed, the decision maker was not the prescription writer.

Conclusions Decision making and the writing of prescriptions are generally not undertaken by the same doctor. Moreover, communication about prescriptions is poor. Further research in a larger sample of hospitals is required to confirm generalisability of the results, and to inform educational interventions to reduce error rates.

  • Prescribing
  • junior doctors
  • decision making
  • medical education
  • clinical pharmacology
  • medical education training
  • internal medicine

Statistics from Altmetric.com

Introduction

Errors in prescribing drugs are an important cause of adverse events in hospitals. In the UK, the majority of prescribing in hospitals is undertaken by junior doctors1 (ie, doctors in any training grade), therefore, targeting interventions to this group could make a substantial impact on error rates. Studies in the UK have shown, for example, that 90% of errors were in medication orders written by junior staff,2 and that foundation doctors (those in the first 2 years of postgraduate training) had an error rate of 7.4–10.3% of prescribed items.3 ,4 However, it is recognised that many prescriptions in hospitals are not instigated by the doctor who writes them, thus raising issues about legal and clinical responsibilities in addition to training.5 There is no available data on how often this occurs, or under what circumstances. This limits the value of the data reported in studies noted above which have attributed errors made to different grades of doctors. Further understanding of this process is necessary prior to designing interventions to reduce errors in prescribing, as these would differ between instances where the doctor is instigating the prescription and where they are merely following instructions.

The objectives of this small observational study were to identify prescriptions where the decision maker and prescription writer were not the same doctor, and in these cases, to describe the level of detail about the recommended medication regime (eg, name, dose, frequency, formulation, route, duration) communicated between the decision maker and the prescription writer. In this paper, we adopt the UK convention of using the term prescription to mean the medication order written on the inpatient drug chart and signed by a doctor.

Research approvals were granted by the North of Scotland Research Ethics Service Committee and local National Health Service Research and Development. Support was also gained from the Medical Director of the hospital and hospital consultants working in each ward.

Methods

Setting

The study took place within six purposively sampled wards at a large UK teaching hospital and included adult medical and surgical units involved in emergency and elective admissions (General Medicine, Trauma Orthopaedics, General Surgery, Acute Medical Assessment, Stroke and Urology). All medical staff, nursing staff and pharmacists on each participating ward were provided with an information sheet about the study (via the charge nurse) and written consent was obtained from a consultant-grade doctor in each ward on behalf of all the ward consultants.

Data collection

Data was collected during normal working hours (09:00–17:00) for consecutive periods of 2 weeks per ward, (ie, 12 weeks in total) between January and April 2011. Medical training starts in August; thus, the study was undertaken halfway through the training year. Each medical ward team was consistent throughout every 2-week period of data collection; the 4-monthly change of placement for Foundation doctors at the start of April 2011 did not interrupt any 2-week period. Wherever logistically possible, the researcher (LH) attended daily ward rounds (in the three units where two concurrent rounds occurred, the researcher attended each on alternate days) and then shadowed a member of the medical team observing decision making and writing of any prescription. All doctors working on the ward were eligible for inclusion; a different grade of doctor was shadowed on each subsequent day to observe equal numbers of doctors from each grade. It was recognised that not all prescriptions written on the ward could be observed in each time period. Data was collected on a study-specific form, previously piloted and revised during a 2-week period of familiarisation with all the study wards. Data collected included: grade of doctor who made the prescribing decision; grade of doctor who wrote the prescription; any information communicated between decision maker and prescription writer; and details of the drug as actually prescribed (including dose, route, frequency and duration for all drugs).

Data analysis

Data were entered into an SPSS database. Frequencies were calculated for: grade of doctor who made the prescribing decision; grade of doctor who wrote the prescription; and the instances where these differed. In cases where the decision maker and prescription writer were different, frequencies of individual components of information (drug name, dose, etc) communicated were calculated.

Results

All wards approached agreed to take part in the study.

Prescribing decisions and prescription writing

Two hundred and thirty-six prescribing decisions were observed during the study period. Of these, 146/236 (62%) were then written by a different doctor. There were 183 decisions (77.5%) about initiation of medication, where 89 decisions were made by a different doctor (49%), and 53 decisions (22.5%) were about discontinuation of medications, where 23 of these prescribing decisions were written by a different doctor (43%).

The prescribing decisions made by each grade of doctor are shown in table 1. Foundation doctors made 59 (25%) prescribing decisions overall, but wrote 105 (44.5%) prescriptions. Two hundred and twenty-eight (97%) prescription decisions were documented by doctors in training grades, but the actual decisions were made by these doctors in only 156 (66.1%) cases.

Table 1

Grade of doctor making prescription decisions and writing prescriptions

There were marked differences observed between wards in the proportions of prescribing decisions made by foundation-year doctors. The lowest percentage was observed in Acute Medicine, where foundation doctors made 5.5% of prescribing decisions, compared with the highest in Trauma Orthopaedics, where they made 71.1% of decisions.

Communication of instructions for prescriptions

In 88 of the 89 cases (99%) in which the decision to initiate a medicine was made by someone other than the prescription writer, the decision was made by the more senior doctor. The information communicated between the decision maker and the prescription writer is shown in table 2.

Table 2

Components of prescriptions communicated from decision maker to prescription writer for medicine initiations (n=89)

The specific name of the drug was stated in 73/89 (82%) of prescriptions. In the remaining cases, the drug was identified by class (eg, β blocker) or type (eg, antibiotic). Other drug information was less frequently verbalised, with no decision makers explicitly specifying all the required details.

Discussion

Overall, doctors in training grades (ie, foundation or specialty doctors) wrote 97% of prescriptions, but were only responsible for 66% of prescribing decisions. Foundation doctors wrote 44.5% of prescriptions and were responsible for 25% of observed prescribing decisions. In cases where the decision maker and prescription writer differed, the information that was generally communicated only included the name of the medication, with the prescription writer expected to know/decide the other core data requirements of the prescription (such as dose, frequency and route).

The main implication of our findings is that a prescription different from that intended by the initiator may be written, resulting in a suboptimal prescription, or an error. This is especially true for doctors in their first foundation year, who wrote over a third of prescriptions and have least knowledge and experience, and who may be unwilling to admit it.6

Medical culture is hierarchical, and junior doctors can find it challenging to clarify or question instructions. This may mean that when unsure about the details of a specific prescription, junior doctors do not ask. Moreover, even when all the necessary information is provided many junior doctors are uncritical when prescribing as directed, feeling that it is ‘not their place’ to criticise those more senior. This could be seen to be an opportunity for a useful safety check if critical review of prescriptions was encouraged. The literature on prescribing suggests that it is more senior doctors that influence prescribers, and that other members of the multidisciplinary team can prompt prescriptions,6 whether in general or for a specific drug. Additionally, there may be situations when more senior doctors may not be aware of current practice (eg, when this is outside their area of expertise), and may deliberately rely on their juniors to prescribe appropriately from a general instruction. Overall, this suggests that an emphasis on any one grade of doctor alone may not be sufficient to reduce errors.

Considerable variation in the dynamics of decision making and writing of prescriptions were seen across the range of wards in the study. While this may reflect inherently different cultures and practices, it may also be a pragmatic response to the availability of staff. For example, while consultants are present most of the day in the Acute Medicine ward studied, they may be mostly occupied in the theatre in surgical wards. These different situations present different challenges for newly graduated doctors, who may be expected to prescribe with very different levels of supervision in different units. Moreover, the finding that core prescription information is explicitly communicated between decision maker and prescription writer reveals other possible sources of error. While this has been previously reported,2 it is disappointing that little has changed in the intervening decade. This finding is in keeping with the literature on the importance of communication as a human factor in error causation,7 but has not, to our knowledge, been specifically investigated in prescribing.

We believe this is the first study to specifically quantify the separation of decision making and prescription writing in hospitals. Our findings have important implications for the interpretation of studies, which focus on the prescription writer alone, and more importantly for those focusing on junior doctors as a target for interventions to reduce prescribing errors. There are also several important training implications for both undergraduate and postgraduate medical education focused around prescribing, which has been the subject of much recent debate. While improving knowledge is important, and the introduction of the prescribing skills assessment8 is welcome, care must be taken to address other issues relating to safety. This study has highlighted the variability in the junior prescriber's role and senior supervision between wards where new graduates are asked to start prescribing, which should be emphasised in undergraduate training. It also demonstrates the need to address training in prescribing to all grades of staff, perhaps to encourage more senior staff to discuss prescribing instructions and ensure that appropriate decisions are made regarding dose, frequency, route of administration, etc. These areas should also be emphasised in undergraduate and early postgraduate training, where junior doctors are already aware that they have gaps in knowledge.9 Attitudes to prescribing should also be addressed by encouraging critical appraisal of other doctors' prescriptions, so that this is the norm rather than being potentially socially embarrassing. All involved in prescribing should be reminded that the prescription writer is legally accountable for the prescription, and knowing this might strengthen junior doctors to resolve to seek clarification from their seniors, and make senior doctors more supportive of their juniors' need for support.

This was a small study with a number of limitations. It was conducted in a selection of wards in a single teaching hospital; results were markedly different across different specialities; a single observer could not assess all prescriptions being written on the ward, and we have no information on ‘out-of-hours’ practices (although other data suggest that 50% of prescribing errors occur ‘in-hours’).4 We also do not know what proportion of the total prescribing was observed in our study. All these limitations may have introduced unknown biases into the findings. Fewer prescriptions than might be expected were written by foundation year 2 (F2) doctors (ie, those in their second year of postgraduate medical training), due to the fact there were no F2 doctors working on three of the six wards sampled. There may also have been a Hawthorne effect resulting from the non-participant observation method used, but we believe this will be minimal. As such, the results from this single site may not be generalisable in the case of other hospitals; however, they do effectively confirm that prescription writers are not always the source of the errors observed in the prescriptions they write, and suggest the need to routinely collect this information in future studies relating to prescribing errors.

The complex workings of many wards, where multiple concurrent ward rounds can occur and different doctors may be prescribing at the same time for different patients, make study design challenging. However, this type of snapshot sample is commonly used in patient safety studies, and the frequency with which the same finding is shown across the range of wards studied suggests that it is an issue that needs to be addressed. This study is a first step in developing methodology to investigate this area, but requires further refinement.

Further study on a larger scale in a representative range of hospitals is required to confirm the results of this study and to examine association of different models of prescribing in decision making—writing dynamic with different factors such as setting (ward type, hospital type), and geographical location. Investigation of the specific role of communication of individual prescription components in causing errors is also needed. Investigation of the relative strength of associations of omission of different components of participant information with a subsequent error is also needed. This in-depth information is critical in determining how and why prescribing errors occur, and to inform designing appropriate interventions in undergraduate and postgraduate training, which might include the use of explicit communication frameworks/aide memoirs as a tool to reduce errors.

Main messages

  • Most prescriptions in hospitals are written by junior doctors, and in the majority of cases this is not the same doctor who made the prescribing decision. There is marked variability in the decision maker–prescription writer dynamic across ward types.

  • Communication of information from the decision maker to the prescription writer is inconsistent and sub-optimal.

  • Prescribing errors may be wrongly attributed to junior doctors. Studies of prescribing must differentiate between those making prescribing decisions and those writing the prescriptions if interventions to reduce error are to be introduced.

Current research questions

The findings from this small study need to be confirmed in a larger representative sample to answer the questions:

  • What proportions of prescriptions in hospitals are written by a doctor different from the one who made the decision?

  • Is there any association between different decision making–prescription writing dynamics and settings such as ward type, hospital type?

  • How can communication of essential components of individual prescription orders be improved to reduce error?

Key references

▶ Dean B, Schachter M, Vincent C, et al. Prescribing errors in hospital inpatients: their incidence and clinical significance. Qual Saf Health Care 2002;11:340–4.

▶ Dornan T, Ashcroft D, Heathfield H, et al. An In-Depth Investigation into Causes of Prescribing Errors by Foundation Trainees in Relation to Their Medical Education: EQUIP Study. Final Report to the General Medical Council. University of Manchester: School of Pharmacy and Pharmaceutical Sciences and School of Medicine. 2009. http://www.pharmacy.manchester.ac.uk/cip/CIPPublications/commissionedreports/prescribing_errors_prevalence_incidence.pdf (accessed Jul 2011).

▶ Dean B, Schachter M, Vincent C, et al. Causes of prescribing errors in hospital inpatients: a prospective study. Lancet 2002;359:1373–8.

Acknowledgments

The authors wish to thank the staff in Aberdeen Royal Infirmary for their support and cooperation with the study.

References

Footnotes

  • Funding This study was undertaken by LH as part of her intercalated BSc studies. There was no external funding.

  • Competing interests None.

  • Ethics approval Ethics approval was provided by North of Scotland Research Ethics Service.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.