|The aim of this research study was to conduct a randomised controlled trial (RCT) to evaluate the effect of pharmacy TECHnician supported MEDication administration rounds (TECHMED) on reducing the frequency of preventable omitted doses at a teaching NHS hospital in the North West of England. Objectives:
– preventable omitted doses (due to unavailable medicines/administrations not signed for);
– preventable omitted doses of critical medicines (as defined by NPSA rapid response report 2010),
– all omitted doses (due to any reason),
– all omitted doses of critical medicines (as defined by NPSA rapid response report 2010), and
Benefits of the proposed research
|Pharmacy Practice benefits:
A recent investigation into the role of pharmacy assistant supported medicines administration in the UK reported encouraging effects on omitted doses but had important methodological limitations (see section D). Our project intends to use a robust RCT design to provide a more definitive answer for hospital leaders and policy makers as to whether pharmacy technician supported medicines administration is both feasible, acceptable and effective in reducing preventable omitted doses. In a time when there is increasing pressure upon NHS institutions to use scarce health care resources efficiently and effectively, this project may provide welcome and important evidence on how this could be carried out.
In light of the increasing use of electronic patient record data and computerised prescribing and medicines administration in UK hospitals, we hope that this project will encourage others working within the NHS and academia to consider these tools as potential assets for conducting research/audits involving large volumes of data. Such studies may then help to answer important clinical questions with confidence given the potential for large sample sizes and up to date real world data.
As interest in the potential role of pharmacy teams in improving medicines optimisation and patient care increases (e.g. the recent recommendations for pharmacist involvement in GP surgeries), it is important to understand how and why any novel interventions such as TECHMED are implemented and work in everyday practice. Results from our process evaluation may inform the future development and optimal implementation of TECHMED, whilst also providing important learning for others wishing to trial novel pharmacy services in clinical settings.
If found to be effective in reducing omitted doses in practice, TECHMED:
Reducing the prevalence of omitted doses has the potential to increase patient exposure to drug treatment, which may then contribute to reduced length of hospital stay and costs (e.g. time away from employment, carer costs). The risk of adverse events associated with omitted doses may also be reduced (e.g. relapse of illness, emergence of antimicrobial resistance, inappropriate dosage adjustment). These benefits may be particularly visible for omissions involving critical list medicines, which evidence suggests are linked to patient harm (see section D).
An indirect benefit of this project may be to increase awareness and management of omitted doses amongst health care staff and patients. Part of the TECHMED intervention is that technicians work with ward staff when patients refuse doses to identify issues and determine their causes (e.g. due to dysphagia) which will involve additional conversations with patients. Such issues may not have otherwise been raised as quickly (if at all). Such activity may therefore promoted patient involvement and shared decision making concerning optimal medicines use.
|Whilst the majority of medicines are used safely in the National Health Service (NHS) patient harm can and does occur from using drug treatments (known as adverse drug events (ADEs)). Research evidence suggests that ADEs are a chief cause of patient harm in hospitals worldwide, with a median of more than one third being preventable in nature. Preventable ADEs are caused by medication errors, and are estimated to affect a median of 1 in 50 hospitalised patients and contribute to approximately 5% of UK hospital admissions.
Medication errors are a common occurrence in hospitals, and frequently affect the medication administration stage.[4,5] Worldwide, medication administration errors (MAEs) effect a median estimate of 19.1% doses administered or omitted in hospitals, though rates vary considerably.[4,6,7] These errors may result in ADEs, with the intravenous route of administration in particular associated with potentially higher MAE rates and risk of patient harm.[4-7] Nurses are placed at additional risk during medication administration due to the relatively fewer opportunities for checking by colleagues when compared to prescribing and medication dispensing stages.
Of the various types of MAE that occur in hospitals, doses given at the wrong time and dose omissions are consistently reported as the most common.[4,7] The patient safety risk posed by delayed and omitted doses was recognised in a 2010 UK National Patient Safety Agency (NPSA) rapid response report, which described 95 resultant cases of severe patient harm or death between September 2006 – June 2009 and made remedial recommendations for health care orgnaisations. Many of the published studies which focus on the prevalence and nature of omitted doses in hospitals originate from the UK and Australia. Through extremely heterogeneous in design and setting, these studies report omission rates of 1.9-12.4% of administerable doses,[10-14] 20-30% of drugs [14,15] and 17-79% of inpatients.[11,14-16] Prescriptions not signed to record administration, patient refusal to take the dose and/or drug not available on the ward are frequently among the three most common types of omitted medication dose observed, and with exception of patient refusals these could be considered to be preventable in nature.[11,13-15,17]
The issue of medication dose omissions has emerged as an important concern at Salford Royal NHS foundation trust (SRFT), the study site. Baseline data recently extracted from electronic prescribing and administration records on omitted doses for a random sample of 60 medical and surgical inpatients from SRFT reveals an omission rate of 14.2% of scheduled doses (1048/7388) over a total of 350 inpatient days [unpublished data, 2015]. A median of 13 omitted doses per patient were observed over a 7 day period and 2.2% of dose omissions involved critical medicines (according to NPSA criteria). Central nervous system and gastro-intestinal class medicines were most commonly omitted and frequent reasons were patient refusal (36%), drug not available on the ward (15%) and no reason documented (4%). This data also indicated that the first 48 hours of admission onto the ward was associated with 70% of the identified omitted doses over a 7 day period per patient.
A number of different interventions have been suggested to reduce the number of omitted doses in hospitals, including ward-based pharmacy team support [18-20], nurse education/training , patient self-administration programmes  and multimodal approaches . However, few of these studies used study designs such as RCTs or before and after studies with a control group, which increases the risk of bias and is also a common issue affecting the wider medication administration error intervention literature . When trials of interventions are conducted, concurrent process evaluations can help explain differences between study findings and expectations as well as the influence of context and how implementation of such interventions can be optimised. Recently the Medical Research Council (MRC) has issued updated guidance on how to conduct such investigations . These omitted dose intervention studies did not include process evaluation, and a lack of systematic evaluation of how interventions were implemented and why they produced the results observed has been previously reported for RCTs [26,27]. Without such evaluations researchers and clinicians may face uncertainty in knowing which modifyable factors influence successful application of TECHMED elsewhere, and whether failure of the intervention to reduce omitted doses is due to the design of the intervention or it’s implementation in practice.
In the UK, pharmacy technician support during medicines administration rounds may be of particular interest given how frequently ‘preventable’ omissions such as unavailable medicines on the ward and unsigned dose administration records cause omitted doses in practice. This is in part because an important role for many pharmacy technical staff is ensuring inpatient wards have regular medication supplies. A recent UK cross-sectional study  examined the impact of pharmacy assistant supported medicines administration on medical and surgical wards in a hospital, finding reduced rates of ‘unacceptable’ (i.e. preventable) dose omissions and unacceptable critical dose omissions compared to intra-and inter-ward controls. However, this study had a number of important limitations, including the use of manual chart review (which can introduce reviewer error/bias) and a lack of detail on the medication classes involved in omissions. Perhaps the most noteable limitations are that wards/participants were not assigned randomly and that there were no measures of omitted doses before and after intervention periods, which significantly limits the extent to which the observed findings can be attributable to the intervention, amongst other concerns.
With this in mind, a more robust randomised trial with concurrent process evaluation is warranted, taking advantage of using readily accessible computer generated data to determine whether and how best pharmacy technical staff can help reduce omitted doses in UK hospitals by supporting ward based nursing medication administration rounds. The findings of this research may be used to adopt the intervention more widely across the base hospital and other NHS hospitals in the future (using process evaluation data for support).
Plan of investigation
|This project consisted of two main workstreams: randomised controlled trial of TECHMED and process evaluation of TECHMED. Conducting and reporting of both workstreams was carried out according to recognised standards/guidance (CONSORT Statement  and MRC process evaluation guidance ).
Randomised controlled trial (RCT).
Setting and sampling: This single site study took place at Salford Royal NHS Foundation Trust (SRFT) hospital. A pool of medical and surgical wards were selected for inclusion in the sampling frame. This selection process was overseen by the Study Advisory Committee (see Section G) to ensure that wards shared similar medication use characteristics (acquired from electronic prescribing and medicines administration data) and had the staffing capacity to act as an intervention ward if required. From this group, 4 intervention and 4 control wards were selected such that each group contained 1 pair each of medical and surgical wards (e.g. 1x pair of medical wards for the intervention and 1xpair of medical wards for the control group). This selection process was led by EK and based on a minimisation algorithm to ensure the baseline characteristics of the two ward groups, especially in terms of the outcome, were comparable.
Trial Phases: The RCT was divided into three phases – pre-intervention/baseline, intervention (where intervention wards receive the TECHMED intervention) and post-intervention. The pre- and post-intervention phases involved no TECHMED support on the intervention wards, and will last 1 month with the intervention phase also lasting 1 month.
Data collected: Total number of regular prescribed doses on intervention and control wards. Number of scheduled doses that were recorded as omitted (all omitted doses or AOD) using the electronic administration records, including more specifically those considered to be ‘preventable’ (defined as dose omissions due to unavailable medication on the ward or omitted due to no specified reason, or PrOD) and those involving critical list medicines as defined by the NPSA RRR 009 [see background for reference]. Anonymised inpatient descriptive data pertaining to scheduled and omitted doses on intervention and control wards (e.g. gender, age in years, number of prescribed medicines, time since hospital admission).
Data source and collection method: Data was extracted and anonymised by SRFT information technology (IT) support staff from the Salford electronic patient record system (also known as Salford Integrated Record). Data extraction will occur immediately following each study phase (pre-intervention, intervention and post-intervention).
TECHMED intervention: Up to 6 trained pharmacy technicians already employed by SRFT performed the TECHMED intervention. These trained technicians will accompany nursing staff on two 2-hour medication administration rounds on all 5 weekdays during the 1 month intervention period (i.e. a total of 20 days). During the medicines administration rounds, pharmacy technicians accompanied nurses and directly supported them during this process. Specific duties which TECHMED technicians trained to carry out included:
These duties were chosen given the high frequency of preventable dose omissions and refused doses in baseline data (see Background data). One pharmacy technician was assigned to each intervention ward during the study and where appropriate alternative technicians will replace each other depending on sickness, annual leave or emergence of other critical duties (ward cover rotas will be kept). Using the input of the project advisory committee (see below) patient orientated materials will be developed (posters, leaflets) which explain the purpose behind the presence of technicians during ward medication rounds.
Pharmacy technician recruitment and training for TECHMED: Up to 6 pharmacy technicians will be recruited and trained in the TECHMED intervention process; this is to ensure that there are at least 2 additional staff to cover in case of sickness, annual leave and/or emergency capacity issues affecting the technicians assigned to regular intervention wards. All recruited technicans will be of the same AFC grade (4 or 5). Suitable technicians were identified by the Study Advisory Panel and invited to participate, before being offered information about the study and asked to sign a written consent form. Training in the TECHMED intervention was delivered by the research team. This training took place up to 2 weeks before the intervention period (to maximise retention of knowledge), and included attendance at one of two scheduled 3 hour face-to-face workshops (2 workshops will be run to accommodate for technician availability, where appropriate) which covered:
TECHMED pharmacy technicians also had access to a standardised data collection guidebook written by the research team at all times during the intervention period – this will reiterate the training points above and contain contact details for support.
Nursing staff engagement: Once suitable wards had been selected randomly for inclusion as intervention units, the research team approached ward nursing staff (including managers) before TECHMED was introduced on the ward to provide written information about the study and to provide a short presentation to staff.
Data analysis: Rates of primary and secondary outcomes were calculated by dividing the number of omitted doses of different kinds by the total of number of eligible scheduled doses and multiplying the answer by 100. Rates were presented using 95% confidence intervals. Additional descriptive data will also be extracted to support outcome rate measurement. Primary and secondary outcome rates were compared between intervention and control wards during the intervention period to determine the effectiveness of the TECHMED intervention. Additional inter- and intra-ward comparisons were made across all stages of the study to determine:
Adequate sample size calculations based on a 20% reduction in outcome rates with 90% power were informed and tested by extrapolating baseline data of dose omissions (see section D).
Based on extrapolation of our baseline data from SRFT, these figures are well within our conservative predictions of total scheduled doses expected during the intervention period (assuming only two drug administration rounds assessed on weekdays):
We used multilevel logistic regression to quantify the effect of the intervention on i) dose omissions and ii) preventable/unacceptable dose omissions. Accounting for the nested structure of the data (doses nested within patients; patients nested within wards), we performed the analyses at the lower level i.e. doses. We accounted for the data-structure using random-effects at the patient level and fixed effects for the ward level. The intervention was included as a binary variable in the model (0 for all wards not receiving the intervention; 1 for all wards receiving the intervention) and we also included additional covariates (such as age, sex,+++). We also used post-estimation commands to calculate and display the probabilities associated with each outcome, with and without the intervention. Any significant differences between outcome rates were presented as p<0.05 values.
Study sample and eligibility criteria: A purposive sampling approach was used for this workstream. TECHMED technicians, intervention ward qualified nursing staff (including managers) and pharmacy staff involved indirectly with the intervention (e.g. dispensary staff who process TECHMED medication orders and intervention ward based medical staff/pharmacists responding to TECHMED technican queries (e.g. to review patients refusing medicines) were all eligible to participate in the process evaluation.
Recruitment: TECHMED technicians and ward nursing staff (including managers) were approached by the research team in the clinical setting and provided information about the process evaluation; those who were interested were then invited to submit written consent to take part in an interview. These consent processes were separate from those of the main TECHMED trial. Pharmacy/medical staff involved in processing additional medication orders and/or responding to queries raised by TECHMED technicians during ward medication rounds were approached by the research team to participate in interview and sign written consent forms. Recruitment then took place before and during the pre-intervention baseline phase.
Data collection: Semi-structured qualitative interviews conducted by the researcher (LS) with participants took place, each lasting approximately 30 minutes (direct non-participant observation of care processes by the researcher is beyond resource capabilities given other trial activities). An interview schedule was developed which addressed the three broad areas of process evaluation (see MRC guidance, reference in background section) in the context of our study:
Interviews took place throughout and immediately after the study intervention period, so as to identify early ‘teething’ problems, ongoing day-to-day management problems and issues arising when the TECHMED intervention is removed. Where appropriate, quantitative data on scheduled and omitted doses was used alongside qualitative data to explain one another and identify inter-dependencies. All interviews were audio recorded by the researcher unless participants indicated that they preferred the researcher take written notes. Interviews took place at a time and private place convenient for the participant.
Data analysis: Interview audio files were transcribed verbatim using a university approved transcriber. Transcripts were coded and analysed by LS in an iterative fashion so that emerging themes can be explored in later interviews. Whilst interview analysis took place after the trial results were known, LS was not involved in a lead role with quantitative analysis and may therefore be less sensitive to the trial findings. Descriptive quantitative data was used alongside qualitative data where appropriate, as described above. Analysis of interview data proceeded in thematic fashion, utilising principles of normalisation process theory (to help understand implementation and post-removal of TECHMED)  and realist evaluation (to help understand mechanisms and context) . RK will independently analyse 25% of interviews and compare findings with LS – DA will resolve any disagreement between coders.
1. deVries EN, Ramrattan MA, Smorenburg SM, et al. The incidence and nature of in-hospital adverse events: a systematic review. Qual Saf Health Care 2008;17:216-23.
2. Kanjanarat P, Winterstein AG, Johns TE, et al. Nature of preventable adverse drug events in hospitals: a literature review. Am J Health Syst Pharm 2003;60(17):1750-9.
3. Kongkaew C, Hann M, Mandal J, et al. Risk factors for hospital admissions associated with adverse drug events. Pharmacotherapy 2013;33(8):827-37.
4. Keers RN, Williams SD, Cooke J, et al. Prevalence and nature of medication administration errors in health care settings: a systematic review of direct observational evidence. Ann Pharmacother 2013;47(2):237-56.
5. Cousins DH, Gerrett D, Warner B. A review of medication incidents reported to the National Reporting and Learning System in England and Wales over 6 years (2005-2010). Br J Clin Pharmacol 2012;74(4):597-604.
6. McLeod MC, Barber N, Franklin BD. Methodological variations and their effects on reported medication administration error rates. BMJ Qual Saf 2013;22(4):278-89.
7. Berdot S, Gillaizeau F, Caruba T, et al. Drug administration errors in hospital inpatients a systematic review. PLoS One 2013;8(6):e68856.
8. Anderson DJ, Webster CS. A systems approach to the reduction of medication error on the hospital ward. J Adv Nurse 2001;35(1):34-41.
9. National Patient Safety Agency. Rapid Response Report 009: Reducing Harm from Omitted and Delayed Medicines in hospital. 2010. Available from: http://www.nrls.npsa.nhs.uk/alerts/?entryid45=66720 [Last accessed 10/05/2015].
10. Lawler C, Brien JA, Welch SA. Omitted medication doses: frequency and severity. J Pharm Pract Res 2004;34(3):174-7.
11. Latimer SL, Chaboyer W, Hall T. Non-therapeutic medication omissions: incidence and predictors at an Australian hospital. J Pharm Pract Res 2011;41(3):188-191.
12. Nettleman MD, Bock MJ. The epidemiology of missed medication doses in hospitalized patients. Clin Perform Qual Health Care 1996;4(3):148-53.
13. Coleman JJ, McDowell SE, Ferner RE. Dose omissions in hospitalised patients in a UK hospital: An analysis of the relative contribution of adverse drug reactions. Drug Saf 2012;35(8):677-83.
14. Warne S, Endacott R, Ryan H, et al. Non-therapeutic omission of medications in acutely ill patients. Nursing in Critical Care 2010;15(3):112-7.
15. Green CJ, Du-Pre P, Elahi N, et al. Omission after admission: failure in prescribed medications being given to inpatients. Clinical Medicine 2009;9(6):515-8.
16. Shandilya S, Nizamuddin K, Waqar Faisal M, et al. Omitted medications: a continuing problem. Clinical Medicine 2015;15(1):12-4.
17. Keers RN, Williams SD, Cooke J, et al. Causes of medication administration errors in hospitals: a systematic review of quantitative and qualitative evidence. Drug Saf 2013;36(11):1045-67.
18. Baqir W, Jones K, Horsley W, et al. Reducing unacceptable missed doses: pharmacy assistant-supported medicine administration. IJPP 2015: doi 10.1111/ijpp.12172.
19. Marconi GP, Claudius I. Impact of an emergency department pharmacy on medication omission and delay. Pediatr Emer Care 2012;28:30-33.
20. Seaton SM, Adams RC. Impact of a Hospital Pharmacy Technician Facilitated Medication Delivery System. J Pharm Pract Res 2010;40(3):199.
21. O’Shea TJ, Spalding AR, Carter FA. Impact of nurse education on the incidence of omitted medication doses in hospital inpatients. J Pharm Pract Res 2009;39:114-6.
22. Richardson SJ, Brooks HL, Bramley G, et al. Self-administration of Medication (SAM) Schemes in the Hosptial Setting: A Systematic Review of the Literature. Plos One 2014;9(12):e113912.
23. Coleman JJ, Hodson J, Brooks HL, et al. Missed medication doses in hospitalised patients: a descriptive account of quality improvement measures and time series analysis. Int J Qual Health Care 2013;25(5):564-72.
24. Keers RN, Williams SD, Cooke J, et al. Impact of interventions designed to reduce medication administration errors in hospitals: a systematic review. Drug Saf 2014;37(5):317-32.
25. Moore G, Audrey S, Barker M, et al. Process evaluation of complex interventions: UK Medical Research Council (MRC) guidance. Available from: http://www.populationhealthsciences.org/MRC-PHSRN-Process-evaluation-guidance-final-2-.pdf [last accessed 10/05/2015].
26. Oakley A, Strange V, Bonell C, et al. Process evaluation in randomised controlled trials of complex interventions. BMJ 2006;332:413-6.
27. Grant A, Treweek S, Dreischulte T, et al. Process evaluations for cluster-randomised trials of complex interventions: a proposed framework for design and reporting. Trials 2013;14:15.
28. Reeves BC. Principles of research: limitations of non-randomized studies. Surgery 2008;26(3):120-4.
29. Schulz KF, Altman DG, Moher D. CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. BMJ 2010;340:c332.
30. May C, Finch T. Implementing, embedding, and integrating practices: an outline of normalization process theory. Sociology 2009;43(3):535-54.
31. Salter KL, Kothari A. Using realist evaluation to open the black box of knowledge translation: a state-of-the-art review. Implement Sci 2014;9:115.