Risk reduction in a cancer chemotherapy production unit: Contribution of an automated drug dispensing system

Abstract

Objective

The purpose of our study was to evaluate the contribution of an automated drug dispensing system in securing cancer chemotherapy production process at the pharmacy of the National Institute of Oncology in Rabat.

Methods

The failure modes and effects analysis method was applied to the chemotherapy production process in two phases: Phase 1, using an open shelf for storage then phase 2, using an automated drug dispensing system. The failure modes were defined and their criticality indexes was calculated on the basis of the likelihood of occurrence, the potential severity for the patients and/or the impact on the process and the detection probability. The criticality indexes of the two phases were prioritized and compared.

Results

We identified 35 failure modes for phase 1 and 37 for phase 2. The sum of criticality indexes was 5957 and 4586, respectively, for phase 1 and phase 2, corresponding to a criticality reduction of −23%. The greatest improvements concerned that the needed drug is missing during the picking, storage of potential expired drugs, and double compounding.

Conclusion

Our study highlighted the contribution of automated drug dispensing system in risk minimization. The use of automated drug dispensing system is a part of security improvement in chemotherapy production unit.

Keywords

Automated drug dispensing system chemotherapy production FMEA risk analysis security improvement

Introduction

Chemotherapy cancer process is a complex process.¹ In general, it is formed of three main steps: protocol prescribing, chemotherapy production and administration.² Chemotherapy production step appears to be crucial in the care of cancer patients, both for the effectiveness of care and patient safety. However, the chemotherapy production process is not immune to medication errors (ME). ME is defined as a deviation from what should have been done during the therapeutic drug management of the patient.^3,4 It is used to refer to an unintentional omission or completion of a medication act which may cause a risk or an adverse event for the patient. By definition, ME is avoidable.^3,4 Many errors are made in the production step of chemotherapy process.⁵ According to Ashokkumar et al.,⁶ chemotherapy MEs ranged from 0.40% to 0.50% in the production step. These errors can be for example incorrect volume or type of diluent, incorrect drugs, wrong patient-specific drug label.⁷ They can lead to ineffective treatment or serious damage to patient’s health.

Risk reduction in the cancer chemotherapy process has become a major objective for all hospitals.⁸ In fact, the number of cancer cases worldwide is constantly increasing.^9,10 In Morocco, the annual national cancer incidence rate is estimated at 30,500 new cases of cancer per year, including 16,775 new cases for women and 13,725 new cases for men.¹¹ The more the number of cancer patients increases the more the quantity of chemotherapy compounding increases and the risk of error increases.^12,13 To deal with this situation, the National Institute of Oncology in Rabat (NIO) after centralizing the chemotherapy production process at the centralized chemotherapy production unit (CCPU) of the pharmacy in 2012¹⁴ and computerizing the process introduced an automated drug dispensing system (ADS) in December 2019. Many studies were interested in the contribution of centralization and computerization in the risk minimization of chemotherapy production process.^8,15,16 Some studies showed that the centralization of chemotherapy production process at the pharmacy allowed risk reduction of the process around 25%.^4,15,16 The use of technology information (computerized prescription, code bar for each patient) allowed risk reduction around 20% compared to the centralized non-computerized system.¹⁶

The aim of our study is to evaluate the contribution of ADS in the risk minimization of chemotherapy production process through risk analysis using the failure modes and effect analysis (FMEA) method.

Materials and methods

Setting of the study

Our study is a comparative risk analysis in two phases. The study took place at the hospital pharmacy of NIO in Rabat. NIO is a reference center for the treatment of cancer in Morocco and Africa. On an average, 3000 cytotoxic preparations were made per month in the CCPU.¹⁴ The risk analysis concerned the chemotherapy production step in the chemotherapy process.

Phase 1

In phase 1, the pharmaceutical validation and the labelling were computerized using the medical software named ENOVA. The storage and picking of raw materials are done using an open storage shelf. Drugs were stored according to the frequency of use. Compounding operations are carried out in two laminar flow hoods extemporaneously for workdays and in advance for weekends. A final control by a visual double check was done before the delivery of the product to care service. Figure 1 describes the chemotherapy production process at NIO.

Figure 1.

Chemotherapy production process at the National Institute of Oncology in Rabat.

Phase 2

In phase 2, the process was the same (Figure 1), it was computerized with instructions on ADS, and compounding operations and control were done with the same organization. The major difference was the use of ADS in this phase instead of the shelf. The ADS of our CCPU was STOCKART type. The ADS was interfaced with the patient’s barcode and with the software (ENOVA) for prescription, pharmaceutical validation and stock management. Several access levels are configured according to the user’s authorization (pharmacist, pharmacy technician, pharmacy resident, maintainer). Users can access their session after recognition of their personal code or fingerprints.

The pharmacist predefined a list of drugs to be configured according to the chemotherapy protocols used at NIO.

The pharmacy technician responsible for loading the ADS, consulted the server daily and defined the list and the quantities of drugs below the stock threshold defined. The ADS was loaded on average three times a week. During restocking, the pharmacy technician entered the expiry dates of drugs into STOCKART. During validation, the pharmacist followed the stock status of different drugs specialities present in the ADS and their batches. During drugs picking, the pharmacy technician selected for a given patient (according to the validated protocol), the drugs one after the other, respecting the trade name and the quantity validated. In addition to the expiration date (close or far), he was informed of the specific terms of use of the selected product by messages configured by the pharmacist such as the dilution solvent, the storage conditions after dilution. Upon request, the drawers and shelves opened one after the other, with light-emitting diodes indicating the location of drugs to be withdrawn.

Choice of the method used

Our analysis was conducted by the FMEA method. FMEA is applied to the analysis and evaluation of the criticality of all the potential failures of a product caused by the processes it underwent.¹⁷ FMEA is an inductive quality tool. Its objective is to identify steps where additional preventive measures may be appropriate to minimize the risks.¹⁸ The FMEA method is described by the ISO 31000, 2009 standards and is one of the tools adapted by the Institute of Healthcare Improvement to help hospitals improve patient safety by optimizing their own organization.¹⁶ This method was chosen due to its practicality and its relative ease of implementation in addition to allowing risk quantification.

Failure modes, effects and criticality analysis

To carry out this analysis, a multidisciplinary team of six persons w was formed. This team was constituted of: a professor of pharmacotechny, a hospital pharmacist (head of chemotherapy production process and quality control), an oncology clinical pharmacist, pharmacy resident, a pharmacy technician, and a nurse.

Four work sessions of approximately 90 min were carried out with the aim to:

Describe the steps of the chemotherapy production process.

Identify through brainstorming possible failure modes during these steps that could harm the patient by asking the question “what could possibly go wrong with this step’’? It was not necessary for a participant to have witnessed the occurrence of a failure mode to include it.

Consensually rate the likelihood of occurrence (from 1 to 10), the severity of the potential effect on patient (from 1 to 10) and the likelihood to detect the failure (from 1 to 10). The evaluations were carried out on the basis of explicit criteria published by Boulé et al.¹⁹ The frequency is an estimate of the number of cases per unit of time. We determined potential causes and consequences of the failure modes, which allowed us to rate the severity and the detection scores. To rate the severity, we took into account the possible consequences for the patient and/or the impact on the process. For detectability, we estimate the probability of failure detection by control measures. To be considered detected, the failure mode must be perceived before it has an impact on patient. We took into account the chance of detection in the others steps of the process.

Calculate the criticality index (from 1 to 1000): The CI of each failure mode was calculated by multiplying the frequency, the severity and the detection scores.

Suggest improvement measures: Identify and make recommendations on steps where additional preventive measures may be appropriate to minimize the risks.

Results

Description of the chemotherapy production process

Chemotherapy production in our CCPU takes place in several steps: Storage of raw materials; pharmaceutical validation; labelling; raw materials picking; compounding and final control. Our comparative risk analysis focused on these steps.

Failure modes definition

In phase 1, the work team identified 35 failure modes in the six steps of the production process – including 6 failure modes at the raw materials storage step; 8 failure modes at the pharmaceutical validation step, 3 failure modes at the labelling step; 6 at the raw material picking step;11 and 1 failure modes respectively at the compounding and final control steps. In phase 2, we identified in addition to failure modes of phase 1, two failure modes specific to the use of ADS. These failure modes were the non-opening of the user’s session and the non-opening of the ADS drawers and the ADS refrigerator door (Table 1).

Table 1.

Failure modes, their criticality indexes in phases 1 and 2 and their criticality indexes variations from phase 1 to phase 2.

Steps	Failure modes	Criticality indexes phase 1	Criticality indexes phase 2	Criticality indexes variations	Percentage of criticality indexes variations
Storage of raw materials	Packaging degradation	98	48	−50	−51.02
	failing drugs retention conditions	120	30	−90	−75
	Failing thermolabile drugs retention conditions	120	90	−30	−25
	Inventory difference between the software ENOVA and physical stock	192	72	−120	−62.5
	Storage of potential expired drugs	180	40	−140	−77.77
	Stock-out	32	32	0	0
Pharmaceutical validation	Non-detection of prescription errors (dose, protocol, way of administration)	420	420	0	0
	Wrong specialty choice	36	18	−18	−50
	Bottle number error calculating	240	240	0	0
	Dispensation of past cures	400	200	−200	−50
	Forgotten validation	128	64	−64	−50
	Incomplete validation	120	80	−40	−33.33
	Non-opening ENOVA session	21	21	0	0
	Non-joining prescriber	32	32	0	0
Labelling	Illegible labelling	144	72	−72	−50
	No labelling	160	160	0	0
	Double labelling	78	48	−30	−38.46
Picking of raw materials	Need product is missing	96	12	−84	−87.5
	Non-opening of the automated drug dispensing system drawers and refrigerators door	–	24	24	100
	Non-opening of Stockart user’s session	–	24	24	100
	Picking of potential expired drugs	240	80	−160	−66.66
	Product error	120	80	−40	−33.33
	Bottle number error picking	72	40	−32	−44.44
	Solvent error	64	64	0	0
	Forgotten drugs or materials	120	45	−75	−62.5
Preparation	Forgotten preparation	64	64	0	0
	Labelling error (inversion)	120	120	0	0
	Using wrong product	270	270	0	0
	Using wrong solvent	168	168	0	0
	Solvent volume error	270	270	0	0
	Product quantity error	630	630	0	0
	Chemical contamination	210	210	0	0
	Incomplete preparation	150	120	−30	−20
	Traceability error	120	120	0	0
	Microbiological contamination	210	210	0	0
	Double preparation	192	48	−144	−75
Control	Double check missing before the delivery to care services	320	320	0	0

Criticality indexes analysis

In phase 1, the sum of CIs for the 35 failure modes was 5957. The average CI was 170. In phase 2, we found a sum of CIs = 4586 for the 37 failure modes, with an average CI = 124, corresponding to a sum of CIs reduction of −1371 (−23%) from phase 1 to 2.

In phase 1 as in phase 2, the major critical steps in chemotherapy production process were the compounding step (CI = 2404 in phase 1 and 2230 in phase 2) and the pharmaceutical validation step (CI = 1397 in phase 1 and 1075 in phase 2). The sum of CIs at the storage of raw materials step was 742 in phase 1 and 312 in phase 2. At the picking raw materials step, it was 712 and 369, respectively, in phases 1 and 2. At the labelling step, the sum of CIs was 382 in phase 1 and 280 in phase 2. At the final control step, CI still remains 320 in phase 2 as in phase 1. The CI variation from phase 1 to phase 2 were −430 (−58%) for the storage of raw materials step, −322 (−23%) for pharmaceutical validation step; −102 (−27%) for the labelling step; −343 (−48%) and −174 (−7%), respectively, for the raw material picking and compounding steps. At the final control, there was no CI variation (Figure 2).

Figure 2.

Percentage of criticality indexes variation at the chemotherapy production steps from phase 1 to phase 2.

From phase 1 to phase 2, 18 failure modes saw their CI decrease, whereas the risk remained unchanged for 17 and 2 failure modes specific to the use of ADS appeared. The most remarkable CIs reductions concerned: the needed product is missing −84 (−87%); storage of potential expired drugs −140 (−77%); and the double compounding −144 (−75%). The details of CIs variation of the failure modes are shown in Table 1.

Discussion

Our study showed a total CI reduction to −23% by using an ADS in the chemotherapy production process. Eighteen failure modes saw their CI decrease and two failure modes specific to the use of ADS appeared. The highest steps of CI reduction concerned the storage of raw materials step (−58%) and raw material picking (−48%) step. These results showed that the ADS has an important impact in the risk minimization of chemotherapy production process.

Our analysis demonstrated that the use of ADS improves the management of the stock of drugs reserved to chemotherapy production by the interface with ENOVA software. It reduced the stock-outs during the picking (−87%) and reduced the inventory difference between ENOVA and the physical stock (−62%). According to Kheniene et al.,²⁰ the use of ADS can reduce the stock-outs and reduced the stock value of drugs by 56% in an intensive care unit. Chapuis et al.³ showed that the storage errors were drastically decreased by ADS (−96%) in an intensive care. The ADS allowed the reduction of criticality of incomplete validations (−33%), dispensing of past treatments (−50%) and double labelling (−75%). Indeed, the ADS is equipped with a barcode reader that allows the pharmacy technician to scan the patient ID during drugs picking. The patient's treatments validated are preselected allowing additional control.

Our results suggest that ADS reduced considerably the confusion risk between the drugs which have similar packaging (−33%); the risk of bottle number errors (−44%); forgotten drugs or materials (−62%). These results lead to the same conclusion as a study in Wales which showed that the ADS improves efficiency (+43%) and the rate of avoidable dispensations errors.²¹ Cousein et al.²² suggested a 53% reduction in the risk of ME during the picking before drugs administration. In such a process, we cannot avoid all errors; nevertheless, it is clear that the risk of making wrong chemotherapy compounding was markedly reduced.

Some failure modes still had high CI in phase 2 and two news failure modes appeared. The CI of the double check missing before the delivery to care services remains unchanged. Moreover, the visual double checking has limits. According to Carrez et al.,²³ it did not improve the accuracy of final production. It is difficult for an operator to accurately distinguish low volume and doses variations. Improving the final control could certainly improve the detectability of certain failure modes in our CCPU. Different methods of chemotherapy production control are available. Gravimetric control allows the detection of large compounding errors such as forgetting or doubling doses.²⁴ Bonnabry et al.⁸ showed that the gravimetric control assisted by computer (CATO® software) allowed CI reduction to −13%. The method of video analysis (Drugcam^®) allows to control 100% of the production without additional need for staff or premises.²⁵ However, this latter method is expensive for developing countries. Improving the safety of the compounding step requires other important approaches such as a training program for chemotherapy manipulators (pharmacy technicians) to respect good manufacturing practices.²⁶ Concerning the failure modes specific to ADS using, their frequency could be further reduced by repeating user training and improving ADS maintenance.

We should note that in such a process, errors cannot completely be eliminated. Continuous risk assessment is required to improve the safety of the process. A study carried out in the Netherlands²⁷ showed that 0.4% of dispensing errors remain in pharmacies equipped with an ADS.

The major limitation of our study is human intervention in the FMEA method, generating a degree of subjectivity in the rating of each risk. To reduce this subjectivity, our team was large and multidisciplinary. In addition, we used an explicit scoring grid as recommended by Boulé et al.¹⁹ Another limitation of our study is that it did not include an economic assessment. The cost of the ADS and its interfacing has been evaluated at USD 61.700. Considering its impact on reducing drug expiration and optimizing stock management on the one hand, and the high prices of anti-cancer drugs on the other hand, we estimate the return of investment after two years of use.

Conclusion

This study allowed to highlight the impact of ADS on securing the chemotherapy production process. With the FMEA method used, the analysis of the six steps suggested a reduction in criticality of −23%. Like the centralization and computerization, of chemotherapy production process, the implementation of ADS should be a part of the quality assurance policy in CCPU. FMEA is one of the traditional methods of risk analysis applied in the health field. It has the advantage of being able to be implemented throughout the life cycle of a system. It would be interesting to do it again after the implementation of corrective measures.

Footnotes

Acknowledgments

We would like to thank the work team and all the pharmacy staff at the National Institute of Oncology for their collaboration

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Jean-Marie Ouedraogo

Casimir Adade Adade

References

Sarsi

Benabbes

El Harrass

, et al. Development of an alternative manual traceability system in the case of computer system malfunction and risk analysis using FMECA method: case of a central unit for cytotoxic preparation. J Pharm Clin 2019; 38: 91–100.

Schulmeister

Chemotherapy medication errors: descriptions, severity, and contributing factors. Oncol Nurs Forum 1999; 26: 1033–1042.

Chapuis

Roustit

Bal

, et al. Automated drug dispensing system reduces medication errors in an intensive care setting. Crit Care Med 2010; 38: 2275–2281.

Nicolas

Sécurisation du circuit hospitalier de préparation des chimiothérapies: application de l’analyse de risque a priori. Sciences pharmaceutiques. HAL Id:dumas-01496003, https://dumas.ccsd.cnrs.fr/dumas-01496003/document.(2017, accessed 8 March 2020) [French].

Bonnabry

Information technologies for the prevention of medication errors. Chim Int J Chem 2005; 59: 359–361.

Ashokkumar

Srinivasamurthy

Kelly

, et al. Frequency of chemotherapy medication errors: a systematic review. J Pharma Pharm 2018; 9: 86–91.

White

Cassano-Piché

Fields

, et al. Intravenous chemotherapy preparation errors: patient safety risks identified in a pan-Canadian exploratory study. J Oncol Pharm Pract 2014; 20: 40–46.

Bonnabry

Cingria

Ackermann

, et al. Use of a prospective risk analysis method to improve the safety of the cancer chemotherapy process. Int J Qual Health Care 2006; 18: 9–16.

Farmer

P FJ

Knaul

, et al. Expansion of cancer care and control in countries of low and middle income: a call to action. Lancet 2010; 376: 1186–1193.

10.

Bray

F FJ

Soerjomataram

, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. Cancer J Clin 2018; 68: 394–424.

11.

Amadou

MD.

etude economique des cancers au maroc: estimation à partir des référentiels internationaux, https://toubkal.imist.ma/handle/123456789/8090 (accessed 8 March 2020). [French]

12.

Carrez

Bouchoud

Fleury

, et al. Work overload is related to increased risk of error during chemotherapy preparation. J Oncol Pharm Pract 2019; 25: 1456–1466.

13.

Limat

Drouhin

Demesmay

, et al. Incidence and risk factors of preparation errors in a centralized cytotoxic preparation unit. Pharm World Sci 2001; 23: 102–106.

14.

Adade

Benabbes

Belahcen

, et al. Centralization impact and cost-saving study in a Moroccan hospital's centralized unit of chemotherapy preparation. J Oncol Pharm Pract 2020. Published as OnlineFirst DOI: 10.1177/1078155220901336

15.

Bonnabry

Despont-Gros

Grauser

, et al. A risk analysis method to evaluate the impact of a computerized provider order entry system on patient safety. Am Med Inform Assoc 2008; 15: 453–460.

16.

Djermoune

Oukkal

Bouzid

, et al. Utilisation d’une méthode d’analyse prospective des risques pour améliorer la sécurité du processus chimiothérapie anticancéreuse dans un service d’oncologie médicale en Algérie. Le Pharm Hosp Clin 2016; 51: 40–50. [French]

17.

Chrysler

Motors

Potential failure mode and effect analysis (FMEA): reference manual. United States of America:Edisi Ke-4 Ford Motor Company, 2008.

18.

Alou

Harti

Mapping risk management in predisinfection of medical devices in a hospital environment. The example of the central sterilization laboratory of the Ibn Sina Hospital, Rabat, Morocco. Le Pharm Hosp Clin 2018; 53: 279–287.

19.

Boulé

Lachapellea

Collin-Lévesquea

, et al. A commented step-by-step approach to realize a FMEA as part of the drug circuit. Le Pharm Hosp Clin 2018; 53: 315–324.

20.

Kheniene

Bedouch

Durand

, et al. Economic impact of an automated dispensing system in an intensive care unit. Ann Franc d'anesth Reanim 2008; 27: 208–215.

21.

James

Barlow

Bithell

, et al. The impact of automation on workload and dispensing errors in a hospital pharmacy. Int J Pharm Pract 2013; 21: 92–104.

22.

Cousein

Mareville

Lerooy

, et al. Effect of automated drug distribution systems on medicationerror rates in a short-stay geriatric unit: drug distribution and medication error. Eval Clin Pract 2014; 20: 678—684.

23.

Carrez

Bouchoud

Fleury-Souverain

, et al. Reliability of chemotherapy preparation processes: evaluating independent double-checking and computer-assisted gravimetric control. J Oncol Pharm Pract 2017; 23: 83–92.

24.

Le Garlantézec

Aupée

Bohand

, et al. Choix et mise en place d’un contrôle pondéral libératoire au sein d’une unité hospitalière centralisée de préparation des cytotoxiques. Le Pharm Hosp 2008; 43: 19–27. [French]

25.

Deldicque

Bourhis

Cadot

, et al. Analyse des différentes méthodes de contrôle des chimiothérapies: quel choix pour un centre hospitalier général? Le Pharm Hosp Clin 2016; 51: 85–86. [French]

26.

Cazin

Gosselin

Implementing a multiple‐isolator unit for centralized preparation of cytotoxic drugs in a cancer center pharmacy. Pharm World Sci 1999; 21: 177–183.

27.

Cheung

K-C

Van Den Bemt

Bouvy

, et al. Medication incidents related to automated dose dispensing in community pharmacies and hospitals-a reporting system study. PLoS One 2014; 9: e101686.