Abstract
Objective:
To describe completeness and accuracy of recording medication changes in progress notes during psychiatric inpatient admissions.
Method:
A retrospective audit of records of 54 randomly selected psychiatric admissions at a metropolitan tertiary hospital. Medication changes recorded on National Inpatient Medication Chart (NIMC) were compared to documentation in the clinical progress records and assessed for completeness against seven quality criteria.
Results:
With between one and 32 medication changes per admission, a total of 519 changes were recorded in NIMCs. Just over half were documented in progress notes. Psychotropic and regular medications were more frequently charted than ‘other’ and ‘if required’ medications. Documentation was seldom comprehensive. Medication name was most frequently documented; desired therapeutic effect or potential adverse effects were rarely documented. Evidence of patient involvement in, and an explicit rationale for, a change were infrequently recorded.
Conclusions:
Revealing substantial gaps in communication about medication changes during psychiatric admission, this audit sheds light on a previously undescribed source of medication error, warranting attention. Further research is needed to examine barriers to best practice, to support design and implementation of quality improvement activities but in the interim, attention should be addressed to development and articulation of content and procedures for documentation.
Any preventable event leading to inappropriate medication use or patient harm is a medication error. 1 Affecting 1–2% of people admitted to medical and psychiatric settings, the impact of individual errors ranges from inconsequential to fatal; collectively, medication errors are associated with substantial morbidity and costs.2–5 Having historically received limited attention, medication safety in psychiatry6,7 is currently accorded priority in Australia. 8 Services are obliged to act to reduce the risk of patient harm and increase the effectiveness of medicines. 9 Assessment of current medication-related practices is a critical first step to practice optimisation.
Errors can arise at various stages in the medication process, from system or human factors. Sources of provider-related errors include faults in prescribing, dispensing or administration. 6 With safe prescribing dependent on timely access to comprehensive information about the patient, condition, medication history and preferences, National Standards require that a comprehensive medication history and current medication record be available at the point of care. 9 Given the complexities of mental disorders, variable medication response, wide dosing intervals, frequent poly-pharmacy, involvement of multiple clinicians simultaneously and over time, and inconsistent adherence to medication, this is crucial in psychiatry.3,6,7,10 Recognising gaps in clinician–clinician and clinician–patient communication about medicines in inpatient settings (where patients are typically acutely unwell) as a potential source of errors, we assessed conformity to standards and the quality of communication in patient records.
Method
The method was an audit of clinical records of admissions to a public psychiatric inpatient unit, granted exemption from ethical review by Australian Human Research Ethics Committee EEOC172.
The study was conducted in mid-2013 at a metropolitan hospital providing inpatient psychiatric care for adults in three 25-bed wards. Care is overseen by 12 psychiatrists who, along with 12 registrars prescribe medication dispensed, for inpatients, by the hospital pharmacy. Service policy 11 embodies National Service and professional Standards9,12,13 specifying that medication be ‘completely and comprehensively documented in the clinical record with demonstration of the patient’s full involvement in treatment decisions and the health professional’s rationale for treatment decisions’. 11
The sample was drawn from patients discharged during January 2013. A random number generator was used to select 50% of discharges from each ward, excluding those following electroconvulsive therapy (medication changes not expected), including only the first of multiple discharges, if applicable. Progress notes (paper-based) and the National Inpatient Medication Chart (NIMC), accepted for study purposes as a complete and accurate record of medication, were obtained for each case.
Data were extracted and entered into an Excel spreadsheet by a final year pharmacy (JT) and a fifth-year medical student familiar with record format and clinical notation, in three steps. Firstly, extracted for each case were patient demographics, primary diagnosis, admission duration and treating psychiatrist. Next, changes (commencing/ceasing medication, increasing/decreasing dose and changing formulation) in medication were extracted from the NIMC. Finally, progress notes were searched, with any documentation related to medication changes during admission transcribed for assessment of accuracy (against NIMC) and quality using criteria specified below.
Where the documentation reviewer was uncertain about a criterion being met, consultation with either a senior pharmacist (EP) or clinician (VM) guided the decision. Accuracy of data extraction and analysis were checked by two researchers (SP, EP), who each reviewed six randomly selected cases over the course of data collection.
The first step in analysis involved reconciliation of medication changes identified in the NIMC with progress note documentation. The presence/absence of any documentation regarding given changes was established. Where documentation was found this was assessed against seven criteria. Marked as present or absent were (1) medication name, (2) dose, (3) frequency, (4) concordance (present where documentation included reference to involvement of patient in decision about given change), (5) desired therapeutic effect, (6) potential adverse effects (present where documentation referenced any potential therapeutic of adverse effect respectively) and (7) an explicit rationale (present where any reason for change was recorded). Descriptive analysis (frequencies and central tendencies) was undertaken.
Results
Records were obtained for 54 people aged 18–65 (mean 37) years, including 34 men. The majority (n = 30; 56%) were diagnosed with a psychotic disorder, predominately schizophrenia (n = 24; 44%), and nine (17%) with depression. Other diagnoses included personality, eating and anxiety disorders. Admission duration was 1–70 (mean 15) days.
The frequency and types of medication changes and comprehensiveness of documentation are summarised in Table 1, with detail related to psychotropic medications presented in Table 2. In summary, 519 medication changes were recorded in NIMCs, with 1–32 (mean 9.6) changes/ patient. The majority of changes (n = 362; 70%) were of psychotropic medicines and regular (not pro re nata or as needed) (n = 345; 66%) dosing regimens. Non-psychotropic medicines/formulations prescribed included antihypertensives, antihyperlipideamics, dietary supplements, laxatives and analgesia. More than two thirds (356; 69%) of changes were commencing medications; a minority related to frequency (n = 13) or formulation change (n = 2).
Medication changes recorded in National Inpatient Medication Chart (NIMC) and documentation in progress notes
Psychotropic medication changes in National Inpatient Medication Chart (NIMC) and documentation in progress
Recording of medication changes
Just over half (303; 58.4%) of the 519 changes were recorded at all in progress notes; documentation was seldom comprehensive. Changes were differentially recorded. Changes to regular and psychotropic medications were recorded more frequently than those to PRN and ‘other’ medications, respectively. Most commonly documented was medication name (n = 266/303; 87.8%). Rarely was desired therapeutic effect or potential adverse effects documented, for any psychotropic medication. Evidence of patient involvement in medication decisions (concordance), found in around a quarter (27%) of documented changes, was associated with a minority of doctors. A rationale for a medication change was included in less than half of documented changes of any type of medication. Typically, such notation was brief, for example, ‘alcohol withdrawal symptoms’ or ‘for anxiety’, but occasionally more information was provided, for example ‘restarting after mood deteriorated following ceasing lithium’. For 13 changes, illegibility made it impossible to assess whether a rationale was provided (marked absent).
Discussion
In demonstrating substantial gaps in documentation of medication changes in progress notes used by clinical teams during inpatient psychiatric admission, this audit evidences a previously undescribed potential source of medication errors. With medications in psychiatric units often given in challenging circumstances, and (re)admissions commonly for treatment of acute exacerbations of illness related to sub-optimal adherence or response to medications, 10 failure to communicate effectively places staff, patients and organisations at risk. While study strengths (random selection of charts, rigorous data extraction and validation, and the substantial number of medication changes examined) support internal validity, implications need to be considered in light of limitations that constrain generalisability. The study was conducted at a single site on records made in one month. While there is no reason to think local prescribers or patients differ systematically from their peers, site idiosyncrasies necessarily influence communication and charting practice. Moreover, in the absence of consensus definition of comprehensive medication documentation, we specified the criteria against which documentation was assessed. Inclusion of others, such as medication monitoring and route of administration, could have provided a more detailed account of documentation. Nonetheless we have, using a method appropriate to the task, 14 identified a problem requiring attention.
Applicable service and professional standards8,11,12,13 outline requirements – accurate and comprehensive documented of prescribed medications must be accessible at the point of care. However, they do not specify the constituents of comprehensive documentation.
We contend consensus about this (locally and more generally) is a critical first step to the mandated quality improvement activities. Beyond the basics – legible documentation of medication name, dose and route/timing of administration – we propose that to be comprehensive, documentation should include rationale for changes being made and concordance. As pointed out by one psychiatrist during presentation of audit findings, knowing why medications are changed is critical to avoiding ‘reinventing the wheel’, trying a medication that has previously proven unsatisfactory. In relation to concordance, we note that absence of evidence of patient involvement in medication decisions in this study does not equal evidence of absence; indeed, we were assured by psychiatrists with whom findings were discussed that they routinely engage patients in decisions. It may well be that patients are involved and garner needed information from various sources, including other clinicians and medication information leaflets. However, failure to record the patient’s views and experiences can disrupt continuity of care and lead to patients being asked similar questions by multiple clinicians, potentially hindering engagement. If a therapeutic relationship is not established, as argued by Mitchell and Selmes ‘then it should be of no surprise if (medication) concordance turns out to be less than ideal’. 10
While responsibility for communication of medication information and reducing risk of error must be shared amongst clinicians providing care, and pharmacists have a pivotal role to play, it can be argued that a particular responsibility adheres to prescribers. Thus, as observed elsewhere our findings may indicate that medical staff need to become increasingly mindful of medication safety. 3 Moreover, because many solutions to prevention of medication errors are found in standardisation and systemisation, 9 development and active promotion of processes and procedures for documentation should be considered. Local policies need to be clear about responsibility and accountability for documentation and mechanisms for ensuring that patients are aware of and are provided with written information about medications.
Given the acknowledged complexity of practice change and the need for multi-faceted targeted strategies to encourage the uptake of practice guidelines, however, quality improvement activities should be grounded in a detailed, contextualised understanding of influences on practice. 15 Locally we will use qualitative methods to examine knowledge and attitudes of prescribers regarding documentation to ensure that interventions (e.g. education) meet needs.
Medication errors, high amongst medical errors and causes of patient harm, are by definition preventable. 16 If the strategies needed to prevent them are to be effective, they must address (potential) problems in the medication management process. Notwithstanding the need for further study to examine the reasons for and outcomes of the identified communication gaps, findings of this study provide new evidence to prompt practice improvement. In concert with efforts to minimise prescribing and administration errors and promote concordance, this can support quality use of medicines in psychiatric inpatient care and patient safety.
Footnotes
Disclosure
The authors report no conflict of interest. The authors alone are responsible for the content and writing of the paper.