A multi-faceted analysis of in-hospital medication discontinuation
practices: a retrospective four-year time trend analysis and
socio-technical evaluation of best practices.
Joost Blom
10507876
October 2016
Master thesis
A multi-faceted analysis of in-hospital medication discontinuation practices: four-year time trend analysis and social technical model.
Author
J.A. Blom (Joost)
Department Medical Informatics, University of Amsterdam Meibergdreef 9, 1105 AZ Amsterdam
Supervisors
Linda Dusseljee-Peute
Department Medical Informatics, University of Amsterdam Meibergdreef 9, 1105 AZ Amsterdam
SRP mentors Remco Piening Furore Informatica (+31) (0)20 346 71 71
Bos en Lommerplein 280, Amsterdam Willeke Mennen
Furore Informatica (+31) (0)20 346 71 71
Bos en Lommerplein 280, Amsterdam SRP Duration February 2015 - October 2016 SRP Locations Furore Informatica Bos en Lommerplein 280 1055 RW Amsterdam (+31) (0)20 346 71 71 Netherlands
Academic Medical Center/University of Amsterdam Department of Medical Informatics
Meibergdreef 15 1105 AZ Amsterdam Netherlands
Keywords:
Not including medication reconciliation in the discharge process may foster the misperception of medication reconciliation as extra work that purely exists for administrative reasons [1].
1. Fernandes, O., & Shojania, K. G. (2012). Medication reconciliation in the hospital: what, why, where, when, who and how. Healthc Q, 15, 42-49.
Preface
This Master thesis is the final assignment for the Master Medical Informatics at the University of Amsterdam which I worked on at Furore in Amsterdam and the Academic Medical Center (AMC) in Amsterdam.
First I would to express my sincerest gratitude to my supervisor, Linda Dusseljee-Peute who supported me throughout my thesis. She is without a doubt one of the best supervisors and has impressed me with her enthusiasm for this topic but also for research as a whole. I admire her for her problem-solving skills and her perseverance.
Furthermore, I would like to thank my tutors at Furore, Willeke Mennen and Remco Piening. They helped me to think critically but also to develop my professional skills, which I highly appreciate. I would also like to thank everyone working at Furore for their input and some well needed distraction from time to time.
I would also like to express my gratitude towards, Paula van der Hilst and Petra van der Raad for giving me the opportunity to be part of the team. Furthermore, they learnt me the tricks of the trade of working in a hospital environment and off course, the importance of correct definitions. I would also like to thank everybody from the team for their help during my internship.
Abstract
The continuation of in-hospital medication after discharge of the patient might impact patient safety by increasing the risks on an ADE at hospital readmission. This study therefore investigated in what manner standard and high risk medication is discontinued at patient discharge between 2012 and 2015 and how and why departments with a high frequency of medication prescription differ in medication discontinuation practices.
In the second chapter of this thesis, the frequency of patients with incorrectly ‘continued’ medication after discharge between three clinical departments was analyzed. To assess the risk associated with the continuation of medication at readmission, an analysis of the continued medication performed by three clinical experts was performed. Finally, the difference in continued medication pre and post implementation of a protocol related to patient discharge was measured.
Overall, a high variability was visible between the percentage of patient with continued medication on both yearly basis as on inter-departmental basis. Seven medications were identified as high risk medications that might lead to ADE’s, of which two showed prevalence higher than 2 percent in the included study period. A statistical reduction of patients with continued medication was found after implementation of the protocol.
In the third chapter, an analysis of the gaps and best practices in medication discontinuation after discharge of the patient on three clinical departments is provided. The Punctuated Socio-Technical Information System Change Model (PSIC) was used as a basic theoretical framework to interpret the result of semi-structured open-ended interviews with healthcare professionals at the three departments. Three main factors identified at all three departments that influenced medication discontinuation were: 1) The EPR did not adequately and efficiently support the healthcare professionals in performing the task of medication closure, which was related to the system properties. 2) Healthcare professionals were unaware of the potential efficiency and patient safety risks associated with incorrect medication registration at discharge. 3) problems in timing and structuring of the task led to low task stability which appeared to be a reason for the continuation of medication after discharge. Departments differed in practices, workarounds, and technological infrastructure. These factors all influence the medication discontinuation practices.
In chapter four, the main research question is answered and recommendations to improve the medication discharge process are presented. Differences in medication continuation between departments were apparent. These differences can possibly be explained by the following reasons; one department made use of second EPR and departments were divided between separate floors. However, the socio technical analysis showed that there were gaps between the actor and technology and the actor and task. These gaps could be explained by absence of intuitiveness of the EPR during discharge of the patient. Recommendations towards improvement of the discharge process related to accurate medication registration at discharge are: 1) Perform a usability analysis of the EPR medication module and redesign the module so that it supports the healthcare professional in the process of medication discontinuation, by reducing redundant tasks and improving its intuitiveness. 2) Increase awareness among healthcare professionals. 3) Provide performance feedback to the healthcare professional. 4) Analyse the medication list before discharge and discontinue medication which cannot be continued at home. 5) Implement and communicate a hospital wide protocol in which the necessity of medication discontinuation is stated to increase the awareness. 6) assess the medication reconciliation process and implement an assistant pharmacist who performs medication reconciliation at admission
Samenvatting
Openstaande medicatie na ontslag van de patiënt heeft mogelijk invloed op de patiënt veiligheid doordat het risico op ADE’s wordt vergroot tijdens heropname. Deze studie onderzoekt het aantal openstaande standaard en hoog risico medicamenten in het electronische patientendossier na ontslag van de patiënt in de periode van 2012 tot 2015. Daarnaast wordt er onderzoek gedaan naar hoe afdelingen verschillen in het continueren van medicatie na ontslag en welke sociologische en technologische factoren van invloed zijn op de gemeten verschillen.
Het tweede hoofdstuk verdiept zich hierin door de frequentie van patiënten met onterecht gecontinueerde medicatie na ontslag te meten op drie klinische afdelingen die veel medicatie voorschrijven. Er wordt een analyse gedaan om de mogelijk risico’s in te schatten van de gecontinueerde medicatie door klinisch experts. Als laatste worden verschillen in gecontinueerde medicatie voor en na een protocol implementatie gemeten, welke gericht is op het reduceren van openstaande medicatie. Een hoge variabiliteit in het percentage patiënten met gecontinueerde medicatie was zichtbaar op jaar niveau als wel tussen afdelingen. Zeven medicamenten zijn geïdentificeerd als hoog risico welke mogelijk kunnen leiden tot ADE’s, waarvan twee medicamenten een prevalentie hoger dan twee procenten hebben. Tevens is een statische vermindering aantoonbaar van patiënten met gecontinueerde medicatie na implementatie van het protocol.
In het derde hoofdstuk wordt er een analyse van de hiaten en de best practices van het proces van medicatie afsluiten bij ontslag van de patiënt op drie klinische afdelingen gegeven. Een bekend socio-technologisch raamwerk genaamd PSIC wordt gebruikt als basis om de resultaten van semigestructureerde interviews te interpreteren. Drie factoren zijn gevonden die van invloed zijn op het onterecht continueren van medicatie: 1) Het patiëntendossier ondersteunt de zorgverlener niet adequaat en effectief genoeg tijdens het proces van medicatie afsluiten, 2) Er is een gebrek aan inzicht in de gevolgen van gecontinueerde mediatie bij ontslag van de patiënt en gevolgen hiervan op de risico’s voor de patiënt, 3) De taak van het discontinueren is onstabiel; continue veranderingen in werkwijze en uitvoering van de taak ondermijnt de stabiliteit en is eventueel een reden waarom medicatie na ontslag van de patiënten onterecht wordt gecontinueerd.
In hoofdstuk vier, worden de resultaten van beide studies bediscussieerd. Het doel van het onderzoek was om de verschillen tussen de afdelingen met betrekking tot gecontinueerd medicatie in kaart te brengen. Er was een zichtbare variatie tussen de verschillende afdelingen maar ook tussen de jaren op een afdeling. Deze verschillen kunnen mogelijk worden verklaard door volgende factoren; zo was er bijvoorbeeld op één afdeling een tweede patiëntendossier en waren de afdelingen opgesplitst op verschillende vloeren. Het socio-technologische onderzoek laat echter zien dat er hiaten zijn tussen de gebruiker en de technologie en de gebruiker en de taak. Deze hiaten kunnen worden verklaard door de verminderde bruikbaarheid van het patiëntendossier tijdens ontslag van de patiënt. Daarnaast speelt het gebrek aan inzicht in de gevolgen ook een rol.
De volgende aanbevelingen zijn gedaan om het proces rondom het afsluiten van medicatie te verbeteren: 1) Het onderzoeken van de bruikbaarheid van de medicatiemodule van het patiëntendossier en deze her-ontwikkelen zodat het de zorgverlener beter ondersteunt wordt in het afsluiten van medicatie bij ontslag. 2) Het verhogen van het bewustzijn van medicatie afsluiten onder de zorg professionals. 3) Aanbieden van prestatie feedback. 4) Het voor ontslag screenen van de medicatielijst en het afsluiten van medicatie die niet thuis voortgezet kan worden. 5) Het implementeren en communiceren van een ziekenhuis breed protocol in welke de noodzaak van het afsluiten van medicatie wordt aangekaart. 6) De implementatie van een apothekersassistente welke ondersteunt tijdens medicatie controle bij opname.
Table of Contents
1. General Introduction ... 1
Main research question: ... 2References: ... 4
Retrospective analysis of medication discontinuation at discharge at three clinical
departments between 2012 and 2015. ... 5
Abstract ... 62.1 Introduction ... 6
2.3 Methods ... 7
2.3.1 Setting ... 7
2.3.2 Retrospective study protocol ... 7
2.3.3 Data abstraction ... 7
2.3.4 Discharged patients with continued medication ... 8
2.3.5 High risk continued medication ... 9
2.3.6 Selection of protocols including medication discontinuation practices ... 9
2.3.7 Statistical analysis before-after process protocol implementation ... 9
2.4 Results ... 10
2.4.1 Frequency of patients with continued medication ... 10
2.4.2 Results of medication risk analysis ... 12
2.4.3 Results of previous implemented protocols ... 12
2.5 Discussion ... 14
2.6 Conclusion ... 15
References ... 16
A social technical evaluation of medication discontinuation practices after discharge of the
patient: identifying gaps and best practices. ... 19
Abstract ... 203.1 Introduction ... 20
3.2 Theoretical background: The Punctuated Socio-technical IS Change (PSIC) model ... 21
3.3.1 Interviews ... 22
3.3.2 Participants ... 23
3.4 Results ... 23
Department A ... 23
Department B ... 24
Department C ... 25
3.5 Discussion ... 26
3.6 Conclusion ... 28
References: ... 28
Overall Discussion and conclusion. ... 31
4.1 Discussion ... 324.2 Recommendations ... 33
4.3 Future research ... 34
4.4 Conclusion ... 34
References: ... 35
Appendix 1 ... 36
Appendix 2 ... 38
Appendix 3 ... 39
1. General Introduction
Various studies have shown that medication errors during hospital treatment are common and can lead to adverse drug events (ADE) putting the patient at risk of clinical harm [2, 5, 7-9]. These medical errors are said to find their cause in variances in medication orders and registration practices during the time of hospital admission and discharge [6]. However, specifically patient discharge is one of the unrecognised high risk care processes influencing patient safety. For, not only do patient discharge practices affect the transfer of information from the hospital to the General Practioner (GP), they may also impact patient safety at hospital re-admission.
For one, the transfer from hospital to home leads to a transition in responsibility from the hospitals’ healthcare professional to the patient or GP [1]. During this transfer there is a high risk for medication discrepancies. According to Foss et al. in 48 percent of all medications a discrepancy could be found in the medication overview between different sources (Hospital, General Practioner and Patient) one month after discharge [4]. Also, several studies have focused on reconciliation of medication variances during hospital stay [6]. They showed that at hospital admission 59 percent of the patients had medication discrepancies in their medication overview and they would have been harmed if these discrepancies had not been discovered. Furthermore, readmitted patients are specifically at risk of harm. In the discharge overview in the Electronic Patient Record (EPR) system administered medication should be recorded as ‘home-medication’ or as ‘discontinued in-hospital’ for transfer. If the discharge medication list is not a closed loop, potentially harmful medication may still be set as ‘continued’, leading to administering the medication at re-admission. This implies that medication which is not continued at home should always be set as discontinued, for it could lead to medication errors. It is therefore of major importance to not only perform reconciling medication practices during hospital admission and transfer, but to ensure that the medication overview is correct and complete at discharge of the patient from the hospital. Moreover, GPs found that receiving a complete medication overview, including discontinued medication, is vitally important to safely continue the treatment at home [10, 11].
As medication administration errors are common in hospitals [12], hospitals implement process protocols on both a hospital and a department level. These protocols aim to improve the quality of medication processes by standardising and assisting healthcare professionals in their work processes. The goal of this standardisation is to reduce the medication discrepancies leading to incorrect medication use or medication errors in the hospital and at discharge. However, it is unclear whether these medications related process protocol implementations are indeed adopted by the healthcare professionals. Next to this, if adopted, whether these protocols are effectively improving medication discontinuation processes is unknown. At last, since 2012 the Health Care Inspectorate (IGZ), Dutch Institute for Accreditation in Healthcare (NIAZ) and the insurance companies give more attention to safe use of medication. Several hospitals acted upon this request and started projects to increase this medication safety for vulnerable patients. The overall goal of these projects is to decrease the number of medication errors significantly within 5 years.
In addition, the influence of Information Systems (IS), in this case the EPR, on the processes surrounding medication discontinuation cannot be underestimated. During the implementation of a Computerised Physician Order Entry System (CPOE) in another academic hospital, socio-technical problems during the design and implementation of the system led to problems in the user-system interaction, low user satisfaction and incorrect system use [13]. Furthermore, numerous researchers have shown that there is growing evidence that the implementation of Healthcare Information Technologies (HIT) lead to undesired and unanticipated consequences impacting patient safety [3, 16-19]. However not all undesirable consequences can solely be related to the implementation of HIT, many of these consequences can also be related to the social and technical interactions within the entire work system
explored. Possible gaps in these interrelations might offer new insights into how to redesign the EPR to fit discharge processes and vice versa.
The study described in this thesis is performed within one of the large academic hospital in The Netherlands that participated in the overall project to increase medication safety. One of the hospital’s medication safety aims was specifically focussed on decreasing the number of medications continued after discharge of the patient which pose a potential risk for medication errors during transfer and readmission. The overall aim of this thesis is to retrospectively describe medication discontinuation practices between 2012 and 2015, and to analyse socio-, technical and organizational factors that explain potential differences in medication discontinuation on a department level that use the same system for recording medication discharge. Table 1 gives an overview of the main research question and the sub questions answered in this thesis.
To answer the main question, two interrelated studies were performed. Figure 1 depicts an overview of the overall study design. Chapter 2 describes the first study which provides insight into: the percentage of patients with continued medication after discharge between three departments with a high number of prescriptions, the number of high risk medications which are incorrectly continued after discharge, and the differences in continued medication performance pre and post process protocol implementation. Chapter 3 describes the second study and reveals socio-technical factors impacting medication discontinuation based on the punctuated socio-technical change model (PSIC) [15] to identify gaps between the actors, structure, task, and technology in the medication discontinuation practices that explain inter department differences. Chapter 4 discusses both study findings and their interrelation, proposes several process changes to augment current medication discharge practices and gives new directions for follow-up research.
Main research question:
In what manner is standard and high risk medication discontinued at patient discharge between 2012 and 2015 and how and for why do departments differ in medication discontinuation practices that lead to medication continuation after discharge?
Research questions study 1:
§ Q1 - What is the percentage of patients with continued medications per month between 2012 and 2015
and how do departments differ?
§ Q2 - Of these continued medications, what type of medication is considered high risk medication at
re-admission by medical experts and what is the prevalence of these high risk medications in the retrospective study-period?
§ Q3 – How do continued medication performances differ pre and post process protocol implementation
that aims to improve discontinuation practises?
Research questions study 2:
§ Q5 - What are the socio-technical factors that influence medication discontinuation practices at
included departments?
§ Q6 - Which gaps exists between the actors, structure, task, and technology that negatively impact the
medication discontinuation practices of the included departments and lead to medication continuation after discharge?
References:
1. Manian, F. A. (1999). Whither continuity of care?. New England Journal of Medicine, 340(17), 1362-1363.
2. Wong, J. D., Bajear, J. M., Wong, G. G., Alibhai, S. M., Huh, J. H., Cesta, A., ... & Fernandes, O. A. (2008). Medication reconciliation at hospital discharge: evaluating discrepancies. Annals of Pharmacotherapy, 42(10), 1373-1379.
3. Berg, M., Langenberg, C., vd Berg, I., & Kwakkernaat, J. (1998). Considerations for sociotechnical design: experiences with an electronic patient record in a clinical context. International journal of medical informatics, 52(1), 243-251.
4. National coordinating council for medication error reporting and prevention. About Medication Errors. Available from: http://www.nccmerp.org/aboutMedErrors.html.
5. Rozich, J. D., & Resar, R. K. (2001). Medication safety: one organization's approach to the challenge. JCOM-WAYNE PA-, 8(10), 27-34.
6. Vira, T., Colquhoun, M., & Etchells, E. (2006). Reconcilable differences: correcting medication errors at hospital admission and discharge. Quality and Safety in Health Care, 15(2), 122-126. 7. Bates, D. W., Cullen, D. J., Laird, N., Petersen, L. A., Small, S. D., Servi, D., ... & Vander Vliet,
M. (1995). Incidence of adverse drug events and potential adverse drug events: implications for prevention. Jama, 274(1), 29-34.
8. Lazarou, J., Pomeranz, B. H., & Corey, P. N. (1998). Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. Jama, 279(15), 1200-1205. 9. de Vries, E. N., Ramrattan, M. A., Smorenburg, S. M., Gouma, D. J., & Boermeester, M. A.
(2008). The incidence and nature of in-hospital adverse events: a systematic review. Quality and safety in health care, 17(3), 216-223.
10. Munday, A., Kelly, B., Forrester, J. W., Timoney, A., & McGovern, E. (1997). Do general practitioners and community pharmacists want information on the reasons for drug therapy changes implemented by secondary care?. Br J Gen Pract, 47(422), 563-566.
11. Balla, J. I., & Jamieson, W. E. (1993). Improving the continuity of care between general practitioners and public hospitals. The Medical Journal of Australia, 161(11-12), 656-659. 12. Haw, C., Stubbs, J., & Dickens, G. (2007). An observational study of medication administration
errors in old-age psychiatric inpatients. International Journal for Quality in Health Care, 19(4), 210-216.
13. Peute, L. W., Aarts, J., Bakker, P. J., & Jaspers, M. W. (2010). Anatomy of a failure: a sociotechnical evaluation of a laboratory physician order entry system implementation. International journal of medical informatics, 79(4), e58-e70.
14. Berg, M., Aarts, J., & van der Lei, J. (2003). ICT in health care: sociotechnical approaches. Methods Archive, 42(4), 297-301.
15. Lyytinen, K., & Newman, M. (2008). Explaining information systems change: a punctuated socio-technical change model. European Journal of Information Systems, 17(6), 589-613. 16. Wachter, R. M. (2006). Expected and unanticipated consequences of the quality and information
technology revolutions. JAMA, 295(23), 2780-2783.
17. Ash, J. S., Berg, M., & Coiera, E. (2004). Some unintended consequences of information technology in health care: the nature of patient care information system-related errors. Journal of the American Medical Informatics Association, 11(2), 104-112.
18. Campbell, E. M., Sittig, D. F., Ash, J. S., Guappone, K. P., & Dykstra, R. H. (2006). Types of unintended consequences related to computerized provider order entry. Journal of the American Medical Informatics Association, 13(5), 547-556.
19. Han, Y. Y., Carcillo, J. A., Venkataraman, S. T., Clark, R. S., Watson, R. S., Nguyen, T. C., ... & Orr, R. A. (2005). Unexpected increased mortality after implementation of a commercially sold computerized physician order entry system. Pediatrics, 116(6), 1506-1512.
20. Harrison, M. I., Koppel, R., & Bar-Lev, S. (2007). Unintended consequences of information technologies in health care—an interactive sociotechnical analysis. Journal of the American medical informatics Association, 14(5), 542-549.
21. Berg, M. (1999). Patient care information systems and health care work: a sociotechnical approach. International journal of medical informatics, 55(2), 87-101.
2.
Retrospective analysis of medication discontinuation at discharge at three
clinical departments between 2012 and 2015.
Abstract
The continuation of in-hospital medication after discharge of the patient might impact patient safety by increasing the risks on an ADE at hospital readmission. To provide further insights into this risk a retrospective analysis on the frequency of patients with incorrectly ‘continued’ medications was performed at three departments. To assess the patients’ risk, a risk assessment was conducted by three medical experts on the type of high frequency continued medications. Finally, an analysis was performed on the difference in continued medication pre and post implementation of a protocol related to patient discharge. Overall, a high variability was visible between the percentage of patient with continued medication on both yearly basis as on inter-departmental basis. Seven medications, were identified as high risk medications that might lead to ADE’s, of which two had a prevalence higher than two percent. Further research needs to focus on identifying factors underlying the high frequencies and the variability in the data of continued medication in patient discharge practices.
2.1 Introduction
Discharging patients from the hospital is a multifaceted process with many challenges. In 2013, more than 2 million hospital discharges were registered in the Netherlands [2]. When a patient is being prepared for discharge, the patient’s care team decides on the follow-up patient care plan. The prescribed patients’ hospital medication regime is one of the important aspects of this care-plan to uphold qualitative and safe continuous care to patients. During admission and discharge, ideally the medication regime is reviewed, also called medication reconciliation. This reconciliation process is considered a critical process potentially impacting a patient´s safety [3]. It consists of several phases which include the verification of the patient medication lists. Medications which have been added, discontinued, or changed relative to preadmission medication lists, are being checked. A systematic review performed in 2012 showed that medication reconciliation is associated with actual and potential adverse drug events (ADE) [4]. However, research on the effect of medication reconciliation after discharge remains inconclusive. Therefore, though most hospitals claim that they in general conduct medication reconciliation at patient discharge, it is unclear if and how accurately these processes are actually performed.
Whether a medication list in the medication reconciliation process is accurate depends on several aspects: the accurateness of the pre-admission medication list, an accurate list of the medication taken by the patient at discharge, and an overview of changes in the in-hospital medication regime and reasons to do so. Increasingly hospitals have electronic patients’ records (EPR) in place, in which the patient discharge plan is fully registered including the medication lists as part of the patient record. A small but essential part of the medication reconciliation, which is done in the EPR, is the registration and identification of medication that remains pending or continued at the time of discharge [5, 6]. This medication includes oral medication, topical medication, inhalation and injections. At discharge of the patient the physician thus needs to change the status of the in-hospital medication in the EPR medication list. Medication in the EPR which is continued at home (i.e. Nursery home, patients’ home and different hospital) is an essential part of follow-up care plan. However, the in-hospital medication that a patient received during admission should not remain registered as ‘continued’ with the exception of some medications for patients with chronic diseases that frequently visit the hospital. The physician therefore needs to manually check and register al in-hospital medication as ‘discontinued’ by the physician in the patients’ EPR medication list. Because the main focus of medication reconciliation is to have an accurate medication list in the follow-up care plan, the ‘discontinuation’ status in the system is considered less important. Physicians are well aware that not all in-hospital medication that remains registered as ‘continued’ directly poses a threat for the patient. However, research conducted on readmission percentages showed that almost 20 percent of discharged hospital patients are readmitted within 30 days, often unplanned [7]. If in theory the status of a medication given in the previous patient hospital admission is still set on ‘continued’ in the EPR at re-admission, this could result in inefficiency in medication reconciliation work practices and increased patient risk for an ADE. The edifice of an accurate patient medication list at discharge should therefore take into account precise medication discontinuation practices.
As part of a larger project in the Netherlands regarding medication safety, one academic hospital in the Netherlands aimed to gain insight into the medication discontinuation practices. In one year,
approximately 22.000 in-hospital medications were registered in this hospital as ‘continued’ in the EPR after patient discharge [8]. To accurately assess whether the ‘continued’ medications can be justified, this study retrospectively compares continued medication frequencies at three departments with a high number of patient admissions, and assesses the type of and risks associated with high frequency ‘continued’ medication. This assessment will provide further insight in the possible risks of discontinuation practices in the EPR for patients when readmitted.
Next to this, during the period included in the retrospective analysis several process protocols were implemented at the departments. These process protocols focused on the reduction of patients with continued medication after discharge. However, it is unclear whether medication related process protocol implementations positively impacted discontinuation practices. To conclude, the first study of this SRP will focus on gaining sight on the frequency of patients with ‘continued’ medication between three departments, associated high risk medication ´continuation´ and if before and after a medication discontinuation related protocol is implemented a significant difference is visible in medication continuation.
2.3 Methods
2.3.1 Setting
This study was conducted in a large Dutch academic hospital with approximately 800 beds. The hospital makes use of an EPR, which was implemented in 2011. The EPR enabled the physicians to record medication before admission, in-hospital medication and discharge medication. All physicians received training on how to use the EPR and were obliged to finish the training before working in the hospital 2.3.2 Retrospective study protocol
A retrospective study including a four-year period, 2012-2015, was conducted to analyse frequencies in continued medication, potential risks associated with high frequency continued medication and if before and after a protocol implementation a significant difference was visible in medication discontinuation. Data was obtained from the hospital’s EPR from 2011 until 2015, which contains information filled in by a healthcare professional during treatment of the patient as well as admission data. A specific query was developed to query data on patient admission date, medication type, medication description, medication status (continued after discharge of the patient from the hospital) prescription data, and department information. All patient information was encrypted and only contained a hashed patient number, which was used to identify different cases within the data selection. To increase the accuracy of the data, the selection criteria (Included years, threshold for included departments, removal of chronic medication) were reviewed by a second experienced analyst.
To include high prescribing departments in the analysis, more than 1500 discharges per year was set as a threshold. Based on this inclusion criteria three departments were selected for further analysis. Afterwards, this data was prepared for analysis: duplicate data in the records were excluded, data was also excluded if information was missing on one of the included data variables. Chronic medications were excluded from the analysis with the help of an assistant pharmacist. More than 1000 data points per included departments per year were set as requirement to perform a yearly analysis. In the beginning of 2011 the EPR was implemented at the departments. Therefore, 2011 did not meet the inclusion criteria of more than 100 data points and was excluded from the study. Data was further analysed per month. 2.3.3 Data abstraction
The data was extracted in separate time-frames from the hospitals’ EPR by a pharmacy employee before the start of the study. This employee had access to the EPR and was authorized to extract medication data from the EPR. Data was saved in a Microsoft Excel file and send via hospital mail to the first author. Table 1 shows the variables which were extracted from the EPR. The start date of prescription was used to merge data of each year into a separate file by the first author. At last, Duplicates were removed on
Variables:
• Last department identification * • Location identification
• Hashed Patient Identification (ID) * • Treatment name
• Medication type *
• Start date of prescription * • Stop date of prescription * • Status
• Mutation date • Mutation time
• Name of person who performed mutation
Table 1: Variables extracted from EPR database. Variables with * were included in the analysis.
A hashed patient ID was included in the analysis to be able to identify different patient cases within the data. To secure the privacy of the patients, the patient ID was hashed using Secure Hash Algorithm (SHA). This hash function was created in excel using an open-source Visual Basic for Applications (VBA). Last department ID was used as a variable in order to identify the selected departments. The start date of the prescription was used to construct a timeline on continued medication. At last, the medication type was also used to analyse the risks of medication continuation. Although data was only selected when it contained no stop date of prescription, some records did include a stop date. These records were excluded by the first author with the help of a pharmacist. The number of discharged patients per included departments per month were extracted from the hospital information system (HIS) by the first author.
2.3.4 Discharged patients with continued medication
To analyse the frequency of discharged patients with continued medication, data had to be deduced to patient level percentages. First duplicate data was removed using the variables, department identification, patient identification and start date of prescription. However, within one treatment medication could be prescribed on different dates. Therefore, duplicates were removed using only the following two variables: department identification and patient identification. This however, also removed patients with more than one treatment on the same department within one month. To validate the resulting data, the first author randomly assessed two datasets of different years and departments. Respectively, three and four percent of the patients had more than one treatment on the same department. At first abstraction, data from the period June 2014 until December 2014, was not included because the variable ‘start date of prescription’ was missing. Because all medication of this period were already discontinued by a pharmacist it was not possible to abstract this data again. A new data file was abstracted which included data on discontinued medication, instead of continued medication data. Important however was the filter which was used in the abstraction. Because the medication which was missed was discontinued by a pharmacist, which is not standard practice in the hospital, only data with the pharmacist as medication discontinuer was included. This file was then merged with the other data of 2014 and duplicates were removed in the same way.
Significance of differences between ranked means between three departments per year was measured by using the Kruskal-Wallis one-way analysis of variance using SPSS version 23. To assess whether departments significantly differed in patients with continued medication on yearly basis. The null-hypothesis is that the three departments are equal. To further specify the difference between two departments, a Mann-Whitney U test was conducted. This test compares the ranked means between two departments and tests the null-hypothesis signifies that the two groups are equal.
2.3.5 High risk continued medication
All data was merged in to one dataset in order to identify high risk medications. Medication names were reduced to brand name. Initially the name also consisted of frequency of admission, and way of administering the drug. To identify which medication were most frequently continued, a pivot table was used. The top-15 most frequently continued medications were included for further analysis. These medications were categorised by three expert reviewers: a doctor, pharmacist and nurse. Experts were individually asked to provide their expert opinion on risks associated with the 15 most frequently continued medications. The following risk assessment was made: no risk - there is no risk for the patient when medication is readmitted, low risk- the risk is dependent on the specific patient case and high risk - there is always a risk for a patient when re-admitted. For high risk medications where consensus was achieved by all three participants, and for medication for which two expert rated the medication ad high risk and one as low risk further analyses on prevalence and potential patient safety risks was performed. The quadratic weighted kappa is used to test interrater agreement between healthcare professionals. A weighted kappa was used to correct for the ordinal data. The test was conducted on data of all medication but also on high risk medication. To assess the results of the kappa test, a benchmark of Landis et al. is used [20].
2.3.6 Selection of protocols including medication discontinuation practices
Between 2011 and 2015 the hospital implemented protocols to reduce the percentage of patient with continued medication(s) after discharge. These protocols were stored in the hospital’s protocol database. The search terms used to query the database are shown in table 2.
• “Medicatieveiligheid” • “Medicatie veiligheid” • “Medicatie stoppen” • “Medicatie ontslag” • “Ontslag” • “Ontslag patiënt” • “Medicatie ontslag”
Table 2: Search terms used to query for protocols, which aimed to reduce the percentage of patients with continued medication, in the in-hospital protocol database.
The protocol search was performed by the first author. The abstracted protocols were analysed by the first author and were included in the study when they stated that after discharge the healthcare professional should discontinue the medication of a patient. The protocols are not included in the appendix but can be shown on request. All the abstracted protocols were written in Dutch. Besides the protocols, also the date of implementation and which specialties were addressed were abstracted from the hospitals’ intranet. The date of implementation and specialties is used to analyse the effects of the implementation on the number of continued medication.
2.3.7 Statistical analysis before-after process protocol implementation
A period of 12 months before and after implementation were included in the analysis. A period of two months surrounding the implementation date was excluded from the analysis to correct for a potential cool down period. To assess differences between ranked means in two periods (before and after the process protocol implementation), a Mann-Whitney U test was conducted. This is a non-parametric test to compare if two groups, derived from the same population, are equal or not. First, the groups of patients with continued medication were compared and secondly, the groups of patients without continued medication were compared. This analysis assesses whether continuation of medication differed
2.4 Results
In the results section below the results of the subsequent research questions will be addressed. First, the percentage of patients with continued medication after discharge between three clinical departments with a high number of prescription will be addressed. Secondly, the number of high risk medications which were incorrectly ‘continued’ after discharge will be assessed and analysed. Finally, the impact of previous protocol implementations is analysed.
2.4.1 Frequency of patients with continued medication
In table 3 an overview of the data per included department per year is shown. There are visible differences in the number of patients with continued medications between the three departments. First, department C has a higher number of discharged patients in comparison to department A and B, except for the year 2013. In 2013 department C discharged 2366 patients, were department A and B discharged respectively 2822 and 2878 patients. The number of patients with continued medications per year is also higher for department C in comparison to the other departments, except for the year 2013. The same holds for the average number of patients with continued medication per month is over all years higher at department C. Overall years combined, department A had a mean of 23 percent, department B had a mean 16 percent and department C had a mean of 35 percent of patients with continued medication after discharge.
A significant inter-departmental difference in patients with continued medication was measured using the Kruskal-Wallis test, the p-value of this test is shown in table 3. All p-values were lower than p = 0,05 and therefore the null hypothesis is rejected. A notable difference is that on department B the ranked mean of patients with continued medication is lower for all years except for 2012, were department C had the lowest ranked mean (7,669).
Table 3: Overview of included data per department per year: Total number of discharged patients, total number of patients with continued medication after discharged, ranked mean of patients with continued medication per year, Chi squared of Kruskal-Wallis test and P-value.
The results of the frequency in percentage of patients with continued medication after discharge per department between January 2012 and December 2015 are shown in figure 1. Despite the substantial fluctuations over time in percentage of patients with continued medication after discharge, differences in time trends between the three departments were apparent. First, percentages of patients with continued medication declined on all departments between May 2013 and June 2013. However, this was followed by an incline in percentages between July 2013 and August 2013. Secondly, in November 2012 all departments showed a decline in percentage of continued medications; departments A and B also had a decline in December 2013. Furthermore, between January and July 2013, department B had a low percentage of patients with continued medication. After July 2013 this percentage increased to 19 percent in October. Another distinctive result is the substantial fluctuation of department A between November 2013 and February 2014, with an incline of 31 percent between December and January. Overall, department C had a higher percentage of patients with continued medication after discharge. Seasonal influences were not apparent in data. In appendix 3, a complete overview with the percentage of patients with continued medication per department per month is given.
Figure 1: Graph of percentage of patients with continued medication after discharge of the three included departments between January 2012 and December 2015. Vertical dashed lines show the two implementations of protocols on department A and B.
2.4.2 Results of medication risk analysis
In table 4 the results of the most frequently continued medication after discharge and the prevalence of these medications on all three departments over all four years. The top 15 medication resembled 46 percent of all continued medications on the included departments. Each of these medications were categorised into three different categories, no risk, low risk- dependent on patient and high risk - always a risk. The experts agreed on six medications as high risk and one medications as a potential high risk. Medications were rated as high risk medication when, all three participants categorised the medication as always a risk, these medications are highlighted in bold. A potential high risk was when at least two experts rated the medication as always a risk and the third rated the medication as risk dependent on patient, this medication is highlighted in italic. The results of this classification is shown in table 4.
Table 4: Results of top-15 most continued medication after discharge on all three departments and the prevalence of these medication in comparison to all continued medication after discharge on all three departments over all four years (n = 5308). Medications in bold are classified as high risk by all experts. Medication in italic is classified as a high risk by two experts.
Seven out of the 15 medications were identified as risk medication. Nadroparin poses a high risk; all three participants categorised this medication as always a risk and it was continued 429 times; 8 percent of all continued medication. Metoprolol was continued 128 times, which is 2,4 percent of all continued medication. This medication was categorised as always a risk by both the nurse and the pharmacist, however the doctor categorised the medication as risk dependent on patient. Oxycodone was categorised as always a risk by all three participants and had a prevalence of 2,0 percent. The other four medication had a prevalence below two percent.
The interrater agreement between the healthcare professionals was tested using the quadratic weighted kappa. A kappa test on all medication data showed a kappa of 0,8774 between doctor and nurse, 0,68 between nurse and pharmacist and 0,725 between doctor and pharmacist. The test on the high risk medication showed slightly different results, a kappa of 0,69 between doctor and nurse, 0,7778 between nurse and pharmacist and 0,8571 between doctor and pharmacist was measured.
2.4.3 Results of previous implemented protocols
Table 5 gives an overview of implemented protocols that were implemented during the period covered in this retrospective study and which aimed at improving discharge practices. The first protocol was implemented on department A in May 2014. This protocol is a manual in which the medication module of the EPR is explained with the help of screenshots. This manual explicitly mentions that medication which is not continued at home should be discontinued. The second protocol was implemented on department B in December 2015. This is a starting manual for new ward doctors, in which tips and instructions are provided. This document mentions that medication should either be converted to home
medication or discontinued at discharge. In figure 1 the implementation dates are represented as a vertical line.
Table 5: Overview of the implemented protocols on the three included departments during the study period, date of implementation of the protocol and a short description of the content addressed in the protocol.
The Mann-Whitney U test was conducted to assess differences between ranked means in two periods (before and after the process protocol implementation). Results are shown in table 6. The test showed that there was a statistically significant difference between the two periods of patients with continued medication before and after implementation (p=0,046), with a mean rank of 15,38 before implementation and 9,63 after implementation. There was no statistical difference between the two periods of patients without continued medication (p=0,204), with a mean rank of 10,67 before implementations and 14,33 after implementation.
Table 6: Mann-Whitney U test on two periods, 12 months before (May 2013-April 2014) and after (July 2014-June 2015) implementation (with exclusion of 2 months cool-down period), of patients with and without continued medication.
Department Date of implementation Short description of content
Department A 13-05-2014 Manual for medication in
hospitals EPR.
2.5 Discussion
This study aimed to provide insight into 1) the percentage of patients with continued medication after discharge between three clinical departments with a high number of prescriptions, 2) the number of high risk medications which are incorrectly continued after discharge, and 3) difference in continued medication pre and post process protocol implementation. The study results show high percentages of patients with continued medication at all three included departments in the four-year period included in this study (department A – 23 percent, department B – 16 percent and department C – 35 percent of all discharged patients were registered with continued medication). To ensure safe medication practices for patients, these medications should have been discontinued in the EPR systems as in-hospital medications. Departments also significantly differed in medication discontinuation; department B showed a lower ranked mean in medication ‘continuation’ compared to the other departments in the years between 2013 and 2015 using the Kruskal-Wallis test. Department A had lower ranked mean of 7,669 in 2012. Furthermore, there is a clear inter-departmental variance in percentages of patients with continued medication on a yearly basis, which is shown by the Kruskal-Wallis test, p > 0,005 on all four included years. This fosters the perception that there is no clear directive on medication discontinuation at discharge. Gleason et al. showed that one-third of the admitted patients had a medication error and 85 percent of these errors were due to wrong medication histories [18]. According to Mueller et al. many hospitals implement simple measures to perform medication reconciliation, however most of these implementations do not increase patients’ safety because they fall short of rigorous improvements [19]. A possible explanation for the high variability in percentage on medication continuation at the three departments is the variance in occupation of the healthcare professionals. Research showed that physicians’ turnover is related to the continuity of care [14]. Due to the retrospective nature of this study, data was not available on the change of occupation. A second possibility was the influence of the emailed list of continued medication by the hospital pharmacy. This list was frequently mailed to the departments and could have functioned as a reminder for good working practices in patient medication discontinuation in the system and led to a reduction in percentages of continued medication on a periodical basis. However, dates on when these emails were send was not available and an additional analysis on potential influence of these feedback email practices was not possible. Though, previous research has shown that performance feedback can have a positive influence on the work practice [9]. Research by Van der Veer et al. also confirmed that care processes are positively influenced by performance feedback [16].
The risk on a medication that is still registered as ‘continued’ in the EPR when a patient is readmitted to the hospital is high. Out of the top-15 continued medications, seven medications where categorised as high risk medication by the healthcare professionals. To analyse the interrater agreement between the healthcare professionals, a quadratic weighted Kappa was conducted. The results showed that according to the benchmark of Landis et al., the strength of the agreement between the healthcare professionals was ‘moderate’ to ‘almost perfect’, with Kappa statistic values ranging between 0,68 and 0,8774 for both the groups (all medication and high risk medication) [20]. Nadroparin was identified as having the highest risk, because 8,2 percent of the continued medications after discharge were Nadroparin. Nadroparin is an anticoagulant which is most frequently used to prevent thromboembolic disorders [13]. It reduces the ability of blood clotting and therefore reduces the risk of harmful blood clots in blood vessels. Metoprolol was also identified as high risk, however lower than Nadroparin. Metoprolol is used for the treatment of high blood pressure, angina, hearth rhythm disorders, migraine, rapid thyroid function, heart failure and after a heart attack. It reduces the heart rate, blood pressure and reduces the oxygen demand of the heart [13]. The last medication which posed a significant risk was Oxycodone, is a heavy painkiller, which belongs to the opioid group [13]. It is used for severe pain for instance after operations, cancer and injuries. Though medication reconciliation practices are presumably performed when the patients are at risk for ADE’s when receiving these medications at re-admission. Furthermore, a significant proportion of the elective patients visit a pre-operative screening before they are administered to the hospital. During this pre-operative screening, an assistant pharmacist performs the medication reconciliation of the patient, medication of previous admissions which are no longer used are discontinued. If medication is still registered as ‘continued’ in hospital, it would take additional time to assure the correct medication list leading to unnecessary work.
Regarding protocol implementations, on 13-05-2014 a protocol was implemented at department A in which medication registration was explained in the hospitals EPR. In this protocol, medication discontinuation after discharge is specifically explained using print screens of the hospital’s EPR. However, besides a bold text explaining the essence of discontinuing in-hospital medication after discharge no further explanation is given in the protocol. A Mann-Whitney U test showed that in a 12-month period before and after introduction of the process protocol the mean number of patients with continued medication was statistically higher compared to the mean number of patients with continued medications after implementation. The mean number of patients discharged on the department before and after was comparable (statistically equal). However, this does not imply that these results are solely due to the implementation of the protocol. Though, research has shown that implementation of a discharge protocol can improve patients’ safety [17]. But still, more research is needed to gain sight on potential other factors that could have influenced this result.
This study has several limitations. The retrospective nature of the study restricted the analysis to the amount of variables which were available in the extracted data. A data set with more variables, such as age, treating specialism and admission number could help identify more factors which influence the discontinuation practices of in-hospital medication prescriptions.
Works changes on the departments could have influenced the results of this study. However, no data was available to assess the influence of these work changes on the percentage of patients with continued medication. A socio-technical analysis is needed to address the influence of these issues on discontinuation practices.
Potentially seasonal trends might explain a difference in the variance between years included in the study [15], for example during summer and around the end of year. However, none were discernible in the data. This however does implicate that there were none.
The protocols were queried from the hospital’s protocol database. Unfortunately, the database does not allow an advanced search, such as in PubMed. This means that only simple search terms could be used resulting in an unstructured result. The first author had to analyse all resulting protocols by hand to evaluate whether or not they should be included in the study. This could have led to missing protocols which have been implemented. Another way of missing protocols could have been that the protocols were not uploaded to the protocol database but only distributed within a department, for example through the usage of a shared network disc. This resulted in only one applicable protocol which was used in the research.
Data could be missed when the department filled in a stop date which was not realistic and was only used as a workaround. For example, a discontinuation date which was 20-1-2077. In this study no correction for outliers was made, although these could have had influence on the analysis. A sensitivity analysis could have conducted to analyse the effects of these outliers. However, data of this study was already analysed per month, therefore the effects of these potential outliers was already minimised. The authors specified the period of delayed-implementation at 2 months. However, one can argue whether this should have been longer or shorter.
Future research should focus on which socio-technical factors influence the variability between departments and years to discern best practices and gaps in working practices. These factors could help decrease the percentage of patients with continued medication after discharge.
2.6 Conclusion
On a yearly basis there is a high variability between percentage of patients with continued medication on a department. Also, there is a clear inter-departmental variance in percentages of patients with continued medication on yearly basis. High risk medications are frequently continued after discharge of the patient, which poses risks for the patients’ safety at hospital re-admission. The difference in patients with continued medication before and after a protocol implementation on department A was measured. A statistically significant difference was found: after implementation the ranked mean of patients with continued medication was reduced, but still present. However, other factors could have influenced this. Therefore, it is important to study the factors potentially influencing this variability from a
socio-References
1. Hospital Utilization (in non-Federal short-stay hospitals). (2016). Retrieved October 06, 2016, from http://www.cdc.gov/nchs/fastats/hospital.htm
2. "Hospital Discharges and Length of Stay Statistics." - Statistics Explained. N.p., n.d. Web. 06 Oct. 2016.
3. Duguid, M. (2012). The importance of medication reconciliation for patients and practitioners. Australian Prescriber, 35(1).
4. Mueller, S. K., Sponsler, K. C., Kripalani, S., & Schnipper, J. L. (2012). Hospital-based medication reconciliation practices: a systematic review. Archives of internal medicine,
172(14), 1057-1069.
5. Roy, C. L., Poon, E. G., Karson, A. S., Ladak-Merchant, Z., Johnson, R. E., Maviglia, S. M., & Gandhi, T. K. (2005). Patient safety concerns arising from test results that return after hospital discharge. Annals of Internal Medicine, 143(2), 121-128.
6. Walz, S. E., Smith, M., Cox, E., Sattin, J., & Kind, A. J. (2011). Pending laboratory tests and the hospital discharge summary in patients discharged to sub-acute care. Journal of general
internal medicine, 26(4), 393-398.
7. Jencks, S. F., Williams, M. V., & Coleman, E. A. (2009). Rehospitalizations among patients in the Medicare fee-for-service program. New England Journal of Medicine, 360(14), 1418-1428.
8. Vira, T., Colquhoun, M., & Etchells, E. (2006). Reconcilable differences: correcting medication errors at hospital admission and discharge. Quality and Safety in Health Care, 15(2), 122-126.
9. Gude, W. T., van Engen-Verheul, M. M., van der Veer, S. N., de Keizer, N. F., & Peek, N. (2016). How does audit and feedback influence intentions of health professionals to improve practice? A laboratory experiment and field study in cardiac rehabilitation. BMJ quality & safety, bmjqs-2015.
10. Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Houghton, Mifflin and Company.
11. Penfold, R. B., & Zhang, F. (2013). Use of interrupted time series analysis in evaluating health care quality improvements. Academic pediatrics, 13(6), S38-S44.
12. Taljaard, M., McKenzie, J. E., Ramsay, C. R., & Grimshaw, J. M. (2014). The use of
segmented regression in analysing interrupted time series studies: an example in pre-hospital ambulance care. Implementation Science, 9(1), 1.
13. Apotheek.nl. (n.d.). Retrieved October 06, 2016, from http://www.apotheek.nl/medicijnen 14. Misra-Hebert, A. D., Kay, R., & Stoller, J. K. (2004). A review of physician turnover: rates,
causes, and consequences. American Journal of Medical Quality, 19(2), 56-66.
15. Björ, O., & Bråbäck, L. (2003). A retrospective population based trend analysis on hospital admissions for lower respiratory illness among Swedish children from 1987 to 2000. BMC
Public Health, 3(1), 1.
16. van der Veer, S. N., de Keizer, N. F., Ravelli, A. C., Tenkink, S., & Jager, K. J. (2010). Improving quality of care. A systematic review on how medical registries provide information feedback to health care providers. International journal of medical informatics, 79(5), 305-323.
17. Clancy, C. M. (2009). Reengineering hospital discharge: a protocol to improve patient safety, reduce costs, and boost patient satisfaction. American Journal of Medical Quality, 24(4), 344-346.
18. Gleason, K. M., McDaniel, M. R., Feinglass, J., Baker, D. W., Lindquist, L., Liss, D., & Noskin, G. A. (2010). Results of the Medications at Transitions and Clinical Handoffs (MATCH) study: an analysis of medication reconciliation errors and risk factors at hospital admission. Journal of general internal medicine, 25(5), 441-447.
19. Mueller, S. K., Sponsler, K. C., Kripalani, S., & Schnipper, J. L. (2012). Hospital-based medication reconciliation practices: a systematic review. Archives of internal medicine, 172(14), 1057-1069.
20. Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. biometrics, 159-174.
3.
A social technical evaluation of medication discontinuation practices after
discharge of the patient: identifying gaps and best practices.
Abstract
The aim of this paper is to identify factors and reveal socio-technical system gaps influencing medication discontinuation practices that impact correct medication registration at discharge of the patient at three clinical departments. A Punctuated Socio-Technical Information System Change Model (PSIC) was used as a basic theoretical framework to interpret the result of semi-structured open-ended interviews with healthcare professionals at the three departments. On all three departments gaps in the socio-technical system of medication discontinuation were identified between the actor and technology and between the actor and task components. The analysis of the discontinuation process indicated that the EPR functionality in which medication was registered did not adequately and efficiently support the healthcare professional in performing the task of medication closure. Also the lack of awareness among healthcare professionals on the potential efficiency and patient safety risks associated with incorrect medication registration at discharge and problems in task stability appeared to be reasons for the continuation of medication after discharge. To prevent medication errors this chapter elaborates on possible changes in the socio-technical system to enhance medication registration of discontinued in-hospital medication.
3.1 Introduction
The development of sustainable healthcare is interrelated with information technologies (IT) such as the Electronic Patient Record (EPR), but every IT implementation may lead to new unanticipated consequences [6, 14]. As berg et al. mentioned; the EPR should support the physicians’ work, not generate work [16]. However, though EPRs are being implemented to increase patient safety and improve quality of care [10], their implementation is fraught with problems that limit their adoption and which may result in incorrect workings. For an EPR, its main goal is to supports healthcare professionals keep track of the patient during their stay in the hospital, for example by providing an overview of medications which are administered to the patient. Accurate registration of medications administered to a patient helps to identify and prevent adverse drug events (ADE’s) by for example automatically check the registered medication with medication allergies of a patient. Hence, it is crucial that the medication overview is correct and complete at discharge of the patient. By reviewing medication pre- and during hospitalization, healthcare professionals are improving the safety of the patients. Currently this process is not well standardised and incorporated in the workflow of the healthcare professionals.
Studies have shown that most healthcare professionals have a high workload [27]. With this in mind, to reduce the workload, the usage of an EPR should effectively support the healthcare professional in completing its task [1]. However, it is unclear if the EPR is solely the influencing factor or if other factors, such as organisational, cultural and workflow related issues influence the process of medication discontinuation.
Healthcare systems (i.e. EPR) are very dependent on the complex human organisational environment, which is why they seem particularly appropriate for a socio-technical analysis [13]. Socio-technical evaluation focusses on the interrelation of technology and organisational environment [6]. It seeks to identify the dynamics between the technology, in this case the EPR, and the social, professional, and cultural environment (the hospital and its healthcare professionals) in which it takes place [7].
For decades, frameworks and models are being used to analyse factors which influence the adoption of Information and Communication Technology (ICT). Several theory-based models have been developed to understand the factors influencing system implementation. For example, the models of Grover (framework for measuring IS effectiveness) [8], Delone and McLean (IS success model) [9], Seddon (variance model of Information System success) [10], Mirani and Lederer’s Framework (to measure benefits derived from IS projects) [11] and Smithson and Hirscheim (conceptual framework for IS evaluation) [12] have been applied to analyse the effects, success and impact of system implementations [15]. However, these models of ICT success cannot be utilised to identify the variables and factor which influence the process of system implementations [15]. Making use of socio-technical models could offer a more in-depth insight in the dynamics which take place between technology and the organisational environment. Furthermore, the success models and frameworks do not take organisational environment and culture into account [15], for which a social technical perspective is needed. To interpret the results of this study the Punctuated Socio-technical IS Change (PSIC) model [1] is usedin this study to analyse factors influencing in-hospital medication registration in the existing EPR system on three clinical
