Accuracy of medication documentation in hospital discharge summaries: A retrospective analysis of medication transcription errors in manual and electronic discharge summaries

https://doi.org/10.1016/j.ijmedinf.2009.09.002Get rights and content

Abstract

Background

Medication errors in hospital discharge summaries have the potential to cause serious harm to patients. These errors are generally associated with manual transcription of medications between medication charts and discharge summaries. Studies also show junior doctors are more likely to contribute to discharge medication error rates. Electronic discharge summaries have the potential to reduce discharge medication errors to ensure the safe handover of care to the primary care provider.

Objectives

(1) Quantify and compare the medication transcription error rate from handwritten medications on manual discharge summaries to typed medications on electronic discharge summaries, and (2) examine the quality of medication documentation according to the level of medical training of the doctors who created the discharge summaries.

Methods

A retrospective examination of 966 handwritten and 842 electronically generated discharge summaries was conducted in an Australian metropolitan hospital. The electronic discharge summaries at the study site were not integrated with an electronic medication management system and hence discharge medications were typed into the electronic discharge summary by the doctor. The discharge medication documentation in both types of summaries was transcribed, either handwritten or typed, from inpatient medication charts in paper-based medical records. Documentation differences between medications in discharge summaries and inpatient medication charts constituted medication errors.

Results

12.1% of handwritten and 13.3% of electronic summaries contained medication errors. The highest number of errors occurred with cardiovascular drugs. Medication omission was the commonest error. The confidence intervals of all odds ratios indicate handwritten and electronic summaries were similar for all areas of medication error. Error rates regarding all 13,566 individual medications for the 1808 summaries were similar by doctor medical training level (intern, resident, and registrar).

Conclusion

Similar medication error rates in handwritten and electronic summaries may be due to the common factor of transcription, either handwritten or typed, known to be associated with medication errors. Clinical information systems evolve and often in the early stages of implementation electronic discharge summaries are integrated with existing paper-based patient record systems. Automatic transfer of medications from an electronic medication management system to the electronic discharge summary holds the potential to reduce medication errors through the elimination of the transcription process.

Introduction

Discharge medication errors are well-recognised problem in hospitals and many have the potential to cause serious adverse events for patients [1], [2], [3]. Methods commonly cited to reduce discharge medication errors include information technology for prevention [4], [5] and hospital pharmacist intervention for correction [6], [7]. Electronic discharge summaries are used in some Australian hospitals, but mostly in conjunction with paper-based inpatient records which contain manual documentation of medications on medication charts filed in the medical record. Therefore, discharge medication data in electronic discharge summaries is usually manually typed by the physician into the electronic discharge summary by transcription from the patients’ paper-based medical record. Implementation of clinical information systems is an iterative and dynamic process which can go through multiple stages. Growing information systems [8], [9] rather than viewing them as a static piece of technology acknowledges that clinical information systems will continue to evolve and be refined. This is particularly the case for electronic discharge summaries. In the early stages of implementation electronic discharge summaries include some information which is automatically drawn from other electronic systems, for example test results from diagnostic test management systems, whilst some information is typed into the electronic discharge summary from existing paper-based medical records.

Commonly junior doctors [1], [7], [10], [11], [12], [13] produce discharge summaries and some studies have shown that errors are associated with this junior status [1], [7], [12]. Medication error rates have also been shown to vary according to the hospital doctors’ level of medical training [14]. Studies of varying sample sizes and settings have reported a range of discharge medication error rates, from 6% [15] to 66% [16], [17]. There are no studies which show the extent of the problem of medication transcription, the type of transcription errors and the medications involved using a large sample comparing discharge medications which have been handwritten in manual discharge summaries and typed into electronic discharge summaries. This study attempts to fill that gap by undertaking a retrospective audit of manual and electronic discharge summaries to:

  • 1.

    assess the quality of transcribing discharge medications, both manual and typed, from paper-based hospital medical records both to handwritten and electronically created discharge summaries, and

  • 2.

    examine the quality of medication documentation according to the level of medical training of the doctors who created the discharge summaries.

Handwritten discharge summaries in the sample were created during a 12-month period 7 months prior to the implementation of the electronic discharge summary system whilst the sample of electronic discharge summaries was created over a 12-month period 1 year following the implementation of the electronic system. Doctors responsible for creating the discharge summaries were unaware that their work was to be examined in this retrospective study. Discharge medication documentation in both types of summaries was transcribed, either handwritten or typed, from inpatient medication charts in paper-based medical records.

Section snippets

Methods

The study was conducted in a 78-bed hospital located in metropolitan Sydney, Australia specialising in acute general geriatric and rehabilitation services. The study hospital was chosen given the large percentage of elderly patients with multiple medications prescribed on discharge. The study sample spanned two 12-month time periods: March 2003 to February 2004 and March 2006 to February 2007. All discharge summaries were handwritten during the 2003/4 period, and all were created electronically

Results

Of the 1006 eligible patients admitted during the 2003/4 study period, 40 records were unavailable, and four had either no discharge summary or no discharge medications, leaving 966 records with handwritten discharge summaries available for study. Of the 946 eligible patients admitted during the 2006/7 period, the records of 49 were unavailable, 31 had a handwritten discharge summary, 15 had no discharge summary, and nine had no discharge medications, leaving 842 records with electronic

Discussion

The finding of an appreciable number of discrepancies when medications are transcribed, either handwritten or typed, from inpatient hospital records into discharge summaries re-affirms the problems caused by transcription, and corroborates results from our previous study [18] which showed that electronically produced discharge summaries can be as erratic with regard to medication documentation as handwritten summaries when transcription by typing is involved. Past studies have shown no

Conclusion

The manual process of medication transcription negates the hypothesised improvements in the quality of electronic discharge summaries. Manual transcription of discharge medications, whether typed or handwritten, is prone to error. Integrated clinical information systems within a complete electronic health record which allows the automatic population of discharge medications from an electronic medication management system to an electronic discharge summary hold the potential to reduce discharge

Authors’ contributions

Joanne Callen made substantial contribution to the conception and design of the study, analysis and interpretation of data and to drafting the article. Jean McIntosh and Julie Li made substantial contributions to: the acquisition of data and analysis and interpretation of data, and critically revising the article. All authors have approved the final version of the submitted paper.

Summary points

What was known before the study

  • Studies have reported wide variations in rates of discharge medication

References (35)

  • B.C. Grasso et al.

    Reducing errors in discharge medication lists using personal digital assistants

    Psychiatr. Serv.

    (2002)
  • K.K. Scarsi et al.

    Pharmacist participation in medical rounds reduces medication errors

    Am. J. Health-Syst. Pharm.

    (2002)
  • T.J. O'Shea

    Checking prescriptions before dispensing

    Aust. J. Hosp. Pharm.

    (1997)
  • C. Atkinson et al.

    Growing, not building, the electronic patient record system

    Methods Inf. Med.

    (1998)
  • J. Sexton et al.

    Ensuring seamless care at hospital discharge: a national survey

    J. Clin. Pharm. Ther.

    (2000)
  • S. Wilson et al.

    General practitioner-hospital communications: a review of discharge summaries

    J. Qual. Clin. Pract.

    (2001)
  • E.R. Warin

    Impact of a pharmacist on the medical admission and discharge process

  • Cited by (107)

    • Polypharmacy in the Homebound Population

      2022, Clinics in Geriatric Medicine
      Citation Excerpt :

      Older adults with dementia are often newly prescribed antipsychotics secondary to hospital-acquired delirium, and these may not be indicated to continue after discharge.20 Finally, documentation errors in the discharge summary can also contribute to polypharmacy.21 Collectively, this can lead to medication misuse, patient or caregiver confusion, adverse drug events, and rehospitalization.

    • Inpatient Hyperglycemia and Transitions of Care: A Systematic Review

      2021, Endocrine Practice
      Citation Excerpt :

      The ADA recommends a structured plan for continued diabetes management tailored to the individual patient and documented in the ER at discharge. Patients and family members should receive clear printed and oral instructions regarding home diabetes management that includes a reconciled list of medications and related diabetes devices, as well as instructions to arrange appointments with the outpatient diabetes providers.3,4 In 2017, the annual per capita cost of care in patients with diabetes in nursing homes/residential facilities was 4.2 times greater in patients with diabetes compared with those without diabetes, so appropriate diabetes care should also be extended in these locations.1

    • Information Architecture, and Innovation

      2019, Nurse Leader
      Citation Excerpt :

      It required 30 or more nurses moving through the hospital and examining patients to chart pressure-injury specifics on paper forms. It required manual entry and translation for analysis and reporting, a process that has legibility and translation errors,10,11 regardless of the type of data. We saved money and removed data entry errors by building HAPI (Hospital Acquired Pressure Injury), a custom application to streamline our quarterly surveys using iPads to chart each patient’s pressure injuries with an interface customized for the task.

    View all citing articles on Scopus
    View full text