Share this article on:

Safety Management of a Clinical Process Using Failure Mode and Effect Analysis: Continuous Renal Replacement Therapies in Intensive Care Unit Patients

Sanchez-Izquierdo-Riera, Jose Angel*; Molano-Alvarez, Esteban*; Saez-de la Fuente, Ignacio*; Maynar-Moliner, Javier; Marín-Mateos, Helena*; Chacón-Alves, Silvia*

doi: 10.1097/MAT.0000000000000286
Clinical Critical Care

The failure mode and effect analysis (FMEA) may improve the safety of the continuous renal replacement therapies (CRRT) in the intensive care unit. We use this tool in three phases: 1) Retrospective observational study. 2) A process FMEA, with implementation of the improvement measures identified. 3) Cohort study after FMEA. We included 54 patients in the pre-FMEA group and 72 patients in the post-FMEA group. Comparing the risks frequencies per patient in both groups, we got less cases of under 24 hours of filter survival time in the post-FMEA group (31 patients 57.4% vs. 21 patients 29.6%; p < 0.05); less patients suffered circuit coagulation with inability to return the blood to the patient (25 patients [46.3%] vs. 16 patients [22.2%]; p < 0.05); 54 patients (100%) versus 5 (6.94%) did not get phosphorus levels monitoring (p < 0.05); in 14 patients (25.9%) versus 0 (0%), the CRRT prescription did not appear on medical orders. As a measure of improvement, we adopt a dynamic dosage management. After the process FMEA, there were several improvements in the management of intensive care unit patients receiving CRRT, and we consider it a useful tool for improving the safety of critically ill patients.

From the *Intensive Care Department, University Hospital 12 de Octubre, Madrid, Spain; and Intensive Care Department, University Hospital de Álava, Vitoria, Spain.

Submitted for consideration April 2015; accepted for publication in revised form September 2015.

Disclosure: The authors have no conflicts of interest to report.

Correspondence: Jose Angel Sanchez-Izquierdo-Riera, Chief of the Polivalent Section, Intensive Care Department, University Hospital 12 de Octubre. Avda de Córdoba s/n, 28041 Madrid, Spain. Email:

Acute kidney injury (AKI) is a common and severe problem in intensive care unit (ICU) patients,1,2 with a reported incidence of 5.6%. The incidence increases to 8.6% when coronary patients are not considered.1 The prevalence of AKI is even higher, reaching rates of greater than 40% in one recent Spanish study,3 resulting in higher mortality in these patients.

Because AKI usually occurs in the context of multiple organ dysfunction syndrome,1 continuous renal replacement therapies (CRRT) plays a key role in the management of the renal dysfunction of the ICU. Because of its characteristics, CRRT requires close monitoring to avoid possible adverse effects, such as circuit clotting that blocks the ability of blood to return to the patient, bleeding, metabolic disturbances, loss of heat, vascular complications, and human error.4

The aim of the present triphasic study was to develop a project to improve the safety of receiving CRRT in the surgical-medical ICU of the Doce de Octubre Hospital, Madrid. In parallel, a safety framework was established based on tools for detecting and analyzing patient safety problems; the incorporation of improvement actions; and a further evaluation of the outcomes, following Donabedian’s classic model for evaluation of care.5 This model thoroughly distinguishes efficacy, effectiveness, and efficiency and the other four components of quality. This model is also known for its quality triad, consisting of three areas under which quality may be measured: “structure” includes facilities and personnel; “process” refers to diagnosis, prognostication, and treatment; and “outcome” concerns the health benefits experienced.

The target in phase I was the detection of possible risks and mistakes related to CRRT and the determination of their prevalence6 in our sample. According to the potential safety issues identified in phase I, failure mode and effect analysis (FMEA) was performed. Failure mode and effect analysis was the cornerstone of phase II. Failure mode and effect analysis is a tool to identify and evaluate potential process mistakes, their causes, and their possible effects.7 It is very useful for healthcare processes.7,8 This five-step process uses an interdisciplinary team to proactively evaluate a health care process. In this case, we applied a process FMEA to identify and remove CRRT-related safety mistakes. The team uses process flow diagramming, a hazard scoring matrix, and the FMEA decision tree to identify and assess potential vulnerabilities. The team’s assessment, proposed actions, and outcome measures are recorded on the FMEA worksheet. Failure mode and effect analysis includes testing to ensure that the system functions effectively and that new vulnerabilities have not been introduced elsewhere in the system. After implementation of the improvement measures identified, we performed a prospective analysis of the patients receiving CRRT.

The objective of the final phase of the project was to evaluate the CRRT safety measures implemented in the surgical-medical ICU of the Doce de Octubre Hospital after the application of the improvement measures identified by the FMEA.

Back to Top | Article Outline


Phase I

This retrospective observational study was conducted in the surgical-medical ICU of the Doce de Octubre Hospital, Madrid, Spain. All patients older than 18 years receiving CRRT admitted to the ICU from January 2010 to December 2010 were included.

Back to Top | Article Outline

Data collection

Clinical data collection was conducted by medical record review. We recorded various variables grouped into three categories: demographic variables, treatment regimen-related variables, and patient risk- and adverse event-related variables. The patient risk- and adverse event-related variables were determined after literature review and expert consultation. The following 10 variables were defined: severe coagulopathy (activated partial thromboplastin time >1.5 times basal value, prothrombin activity <35%, platelets <50,000/µl, and fibrinogen <100 mg/dl) and circuit anticoagulation; evidence of bleeding during CRRT; absence of CRRT dose prescription on medical records; absence of CRRT dose pattern on the nurses’ daily report sheets; no monitoring of phosphorus and magnesium levels; no monitoring of blood urea levels; less than 24 hours of filter survival time; circuit clotting that blocked the return of blood to the patient; mild hypothermia (32–35°C); and moderate hypothermia (less than 32°C).

Back to Top | Article Outline

Phase II

The process FMEA was structured according to the following methodology:

* Subject area of analysis (FMEA topic): implementation process of CRRT in the surgical-medical ICU of the Doce de Octubre Hospital

* Election of a team experienced in both CRRT and safety tools with decision-making capacity in the ICU organization

* Creation of a process flow diagram (Figure 1)

* Determination of all potential failure modes for each of the subprocess steps in a FMEA worksheet (Figure 2)

* Failure mode analysis that considers the probability of the event occurring (P), its severity (S), and the detectability of the event (D)

* Procurement of the risk priority number (RPN) through the multiplication of the three values (P × S × D)

* Prioritization of the failure modes based on their RPN

* Outcome measures to remove or mitigate those potential modes with higher RPNs

* Development of control indicators and evaluation of the proposed actions

* Evaluation of the effects of the proposed actions

This phase was carried out from 2011 to 2012.

Back to Top | Article Outline

Phase III

Phase III involved a cohort study after the implementation of the improvement measures adopted because of the FMEA methodology. All patients receiving CRRT admitted to our ICU from January 2013 to December 2013 were included. The variables analyzed were related to the CRRT treatment and patients’ risks and adverse effects and were the same as in the pre-FMEA period. The data collected were compared with those collected in phase I or in the pre-FMEA period.

We conducted an independent analysis of those patients who received anticoagulation with citrate/calcium because we could not use this type of anticoagulation in the pre-FMEA phase.

Back to Top | Article Outline

Ethical Considerations

In accordance with the Helsinki Declaration, no modifications were made of standard patient care or experimental procedures. No informed consent was necessary because of the approval of the Institutional Ethics Committee.

Back to Top | Article Outline

Statistical Analysis

Quantitative variables are expressed as mean with standard deviation or median and interquartile range (IQR) and qualitative variables are expressed as absolute and relative frequencies. The statistical significance of the proportion comparison was assessed by the χ2 test or Fisher’s exact tests for crosstabs. Comparisons of continuous variables were assessed by Student’s t test/ANOVA or Mann–Whitney–Wilcoxon test, as appropriate. Two-by-two multiple comparisons were performed by correcting the level of statistical significance. The correlation between continuous variables was evaluated using the Spearman correlation coefficient. SAS software (SAS Institute, Cary, NC) was used for all statistical analysis and statistical significance was set at p = 0.05.

Back to Top | Article Outline


As a result of the FMEA process, the major potential failures detected, sorted by the RPN, were as follows: lack of asepsis in the maintenance of therapies (RPN 448), inadequate treatment indication (RPN 504), improper connection to the patient (RPN 504), inadequate treatment (RPN 504), actual dosage lower than scheduled (RPN 576), and incorrect treatment withdrawal (RPN 512). All of these potential failures were associated with a lack of training and protocols. The last major potential failure detected was early circuit clotting (RPN 640), which was related to blood circuit design and catheter failure.

The different improvement actions proposed were the following: development of a CRRT protocol; performance of training sessions; increasing awareness of asepsis; and introduction of an assembly check list, including aseptic handling and improvement of equipment and consumables to prevent circuit clotting (especially changing our CRRT catheters for more appropriate ones that provide better blood flow, decreasing the mean filtration fraction [FF]).

Back to Top | Article Outline

Comparison of Pre- and Post-FMEA Results

During 2013, 536 patients were admitted for 48 hours or more to our ICU. Of these, 72 patients (13.43%) received CRRT. The mean age of the sample was 54.39 ± 13.59 years and 48 were men (66.67%). The mean Sepsis-related Organ Failure Assessment (SOFA) score at the beginning of the treatment was 12.16 ± 3.47 (median: 12; IQR: 4–14).

The data obtained were compared with those from the pre-FMEA sample of patients who received CRRT in 2010. In that year, 54 patients received CRRT (11.7% of all patients admitted to the ICU), with a mean age of 59.78 ± 14.80 years; 77.8% (42 patients) were men. The mean SOFA score and other epidemiological data were similar to those in the group of patients considered in 2013.

In 2010, the main CRRT indications were fluid overload in 45 patients (83.33%) and urea/solute removal in 43 patients (79.63%) because both conditions frequently overlap. In 17 cases (31.48%), hyperkalemia was the main indication. A similar distribution of indications was found in 2013.

In 2010, continuous venovenous hemodiafiltration (CVVHDF) was the most used mode (80.3%). After the FMEA, the most used CRRT mode was also CVVHDF (53.39%), followed by continuous venovenous hemodialysis (CVVHD; 44.25%; Figure 3). In both years, the most used vascular access was the right femoral vein: 53.7% in the pre-FMEA sample and 50% in the post-FMEA sample (followed by the left femoral and the right jugular, with similar percentages in both periods).

A comparison of the different CRRT doses is shown in Table 1. In 2013, the FF was 9.37 ± 8.86% (median: 9%; IQR: 1–16%) versus 20.7 ± 11.3% (median: 21.8%; IQR: 6–52%) in 2010 (p < 0.05). In the pre-FMEA sample, 66.7% of patients (n = 36) presented a FF greater than 20%. After the FMEA, only two patients presented a FF greater than 20%.

Regarding anticoagulation, the mean heparin dose was 4.37 ± 3.39 U/kg/hour in 2013 versus 6.82 ± 4.32 U/kg/hour in 2010. Heparin was associated with epoprostenol in 19 cases (23%) in 2013, whereas it was associated with epoprostenol in 13 cases (28.3%) in 2010.

Special mention has to be given to the group of patients that received citrate anticoagulation for CRRT. Our experience with anticoagulation with citrate/calcium began with the Multifiltrate system (Fresenius Medical Care, Schweinfurt, Bavaria, Germany). From January to June 2011, we used it in 12 patients; in four of these patients, it was used as a rescue therapy for a hypercoagulable state. The filter durability results were better than those obtained with other anticoagulants (median life of each filter, 68 hours). In four cases, who received citrate/calcium anticoagulation as a rescue therapy for a hypercoagulable state, the median filter life increased from 4.8 to 60 hours. Later, in 2013, we began to use, in a systematic and protocolized way, citrate/calcium anticoagulation with the Prismaflex system (Gambro/Baxter, Lund, Sweden). The results were similar to those obtained in the previous period but we excluded them from the analysis because we cannot compare them with those of the pre-FMEA analysis.

A comparison of the frequencies of the detection of risks and possible complications is shown in Table 2. In the pre-FMEA sample, 31 patients (57.4%) suffered circuit clotting before 24 hours of filter lifespan. This complication appeared in 21 patients (29.6%) in the post-FMEA sample (p < 0.05). In the pre-FMEA sample, there was no possibility of blood return to the patient in 64.9% of these circuit clots (63 circuits in 25 patients), whereas this situation occurred in 47.76% of the circuit clots of the post-FMEA samples (32 circuits in 16 patients; p < 0.05).

According to the filter survival life (less than 24 hours vs. greater than 24 hours), there were statistically significant differences in the fluid removal rate variable: 129.3 ± 58.7 ml/hour versus 93.6 ± 41.3 ml/hour (p = 0.03). In these circuit clots without blood return possibilities, the CRRT blood rates were lower: 148.5 ± 17.2 ml/min versus 173.9 ± 37.9 ml/min (p = 0.03). Finally, our new CRRT protocol, which uses a dynamic individualized dosage, has allowed us to reduce replacement/dialysate fluid consumption per patient by 25% versus the standard dosage of the pre-FMEA sample.

Back to Top | Article Outline


The proportion of patients receiving CRRT treatment in our ICU is significantly higher than the 5–6% rate reported in multicenter trials.1,2,9 This increased rate is directly related to the type and severity of the patients admitted to our ICU. This, in conjunction with the considerable and long-standing interest in the development of CRRT shown by the ICU staff members,10 made the surgical-medical ICU of the Doce de Octubre Hospital the ideal setting for the introduction of safety programs aimed at improving CRRT safety.

Fluid overload (83.33%) and solute removal (79.63%) were the main indications for CRRT. These two conditions are the main indications for beginning extracorporeal depuration in a critical care patient with AKI. It is also important to note that AKI usually appears in the context of multiple organ dysfunction syndrome.1

The most applied CRRT mode in both groups (pre-FMEA and post-FMEA) was CVVHDF (53.39%). However, use of CVVHD has significantly increased (44.25%). These data contrast with the prevalence of convective therapies in previous decades, with continuous venovenous hemofiltration1,9 the most used mode (79.6%).

As shown in Table 1, this approach has led to a greater mean diffusion dose (from a dialysis rate of 1,593 ml/hour in 2010 to a dialysis rate of 1,688 ml/hour in 2013) as well as a lower effluent rate in 2013 (replacement fluid rate of 1,442 ml/hour vs. 765 ml/hour) at the expense of a greater use of CVVHD. The result of this management is a decreased FF (9.37% post-FMEA vs. 20.7% pre-FMEA) and thus optimization of the treatment circuit. Other authors have shown a longer circuit duration using treatments with an increased diffusive component.11 At the same time, an increased catheter diameter (from 11.5 to 13.5 F) might explain the reported increased blood flow, which is also of paramount importance to reduce the FF.

Regarding the dose of CRRT, we have to mention two remarkable clinical trials, VA/NIH and RENAL, which appeared in 2008 and 2009, respectively.12,13 In both studies, based on a fixed depuration dose during the whole treatment, the use of greater depuration doses had no beneficial effect on mortality reduction or AKI recovery.

Nevertheless, many authors have analyzed these studies and highlighted their limitations, particularly the dynamic adjustment of the dose to the patient’s clinical condition.14–16 Adjustment of the depuration dose according to the severity of the critical care illness17,18 is used because the benefits associated with larger or smaller doses are found in different subgroups of patients.

The current CRRT protocol (post-FMEA) of the surgical-medical ICU of the Doce de Octubre Hospital is based on this dynamic management in response to changes in the clinical situation of the patient. In addition, from an economic point of view, the individualized dosage allows us to reduce replacement/dialysate fluid consumption by 25% per patient versus the standard pre-FMEA dosage.

Back to Top | Article Outline

Modification of CRRT-Associated Risks and Adverse Events After the FMEA Process

The improved safety of CRRT after the application of the FMEA is clear, as seen in Table 2. Some of the most severe adverse events, such as bleeding or anticoagulation in coagulopathic patients, have completely disappeared.

In the epidemiological, multicenter, and multinational study of the BEST Kidney group,19 which explored CRRT practices in 55 ICUs, bleeding appeared in 3.3% of patients. In relation to anticoagulation, as shown in the “Results” section, we have decreased the mean heparin and epoprostenol doses in our treatments, in line with some studies suggesting that anticoagulants are not required to maintain the circuit.20,21

Despite the significant improvement in CRRT patient care, there are two outstanding aspects requiring work: longer filter life and circuit clotting that blocks the return of blood to the patient. We have adopted some of the actions proposed to increase the mean filter survival time, such as using appropriate catheters and increasing the diffusive component to maintain the FF less than 20%, thereby reducing the number of clotted circuits that were unable to return the blood to the patient. Blood loss after circuit clotting can rise to 200 ml per circuit. Because critical care patients are characterized by low hematocrit levels,22 this blood loss is of great significance.

The use of citrate anticoagulation has also significantly improved the durability of the circuits (as mentioned in the “Results” section). Unfortunately, we cannot compare these data with those of the pre-FMEA period because this system was not in use.

Back to Top | Article Outline

Controversies Regarding FMEA

The main controversy related to this tool is its qualitative condition. In the study by Shebl et al.,23 two multidisciplinary teams each conducted FMEA of the use of vancomycin and gentamicin. In their results, they found discrepancies between the teams’ estimates and the trust incident database. They concluded that the FMEA failed in face, content, criterion, and construction validity. The same group of investigators24 recommends that healthcare organizations do not solely depend on FMEA results to prioritize patient safety issues, especially because FMEA requires considerable time, effort, and resources.

Despite its limitations, this tool gave us an appropriate methodology25 for the standardization, analysis, and outlining of the clinical process in the context of a multidisciplinary working group.

Back to Top | Article Outline


After the implementation of the proposed improvements by the process FMEA, there were several decreases in possible risks and adverse events associated with CRRT, such as longer filter survival time, greater possibility of returning blood to the patient before coagulation of the filters, better control of phosphorus monitoring, proper circuit anticoagulation prescription, and reduced development of hypothermia. The dynamic protocolized management has decreased the cost of these treatments, improving the efficiency of CRRT in our ICU.

Back to Top | Article Outline


1. Maynar J, Sánchez-Izquierdo JA, Daga D, Herrera My Grupo de Trabajo de Cuidados Intensivos Nefrológicos de la SEMICYUC. . Situación actual del reemplazo renal en las UMI de la SEMICYUC. Med Int.;112(25S):2001
2. Liaño F, Junco E, Pascual J, Madero R, Verde E. The spectrum of acute renal failure in the intensive care unit compared with that seen in other settings. The Madrid Acute Renal Failure Study Group. Kidney Int Suppl. 1998;66:S16–S24
3. Herrera-Gutiérrez ME, Seller-Pérez G, Sánchez-Izquierdo-Riera JA, Maynar-Moliner JCOFRADE investigators group. . Prevalence of acute kidney injury in intensive care units: The “COrte de prevalencia de disFunción RenAl y DEpuración en críticos” point-prevalence multicenter study. J Crit Care. 2013;28:687–694
4. Oudemans-van Straaten HM. Primum non nocere, safety of continuous renal replacement therapy. Curr Opin Crit Care. 2007;13:635–637
5. Donabedian A. The quality of care. How can it be assessed? JAMA. 1988;260:1743–1748
6. Molano Álvarez E, Guirao Moya A, Esteban Sánchez ME, Sanz Cruz P, García Delgado E, Sánchez-Izquierdo Riera JÁ. Detección de riesgos para la seguridad del paciente en la implementación de Técnicas Contínuas de Depuración Extracorpórea. Enferm Intensiva. 2011;22:39–45
7. Ruiz P, González C. El análisis modal de fallos y efectos (AMFE). Una herramienta muy útil para la seguridad del paciente. JANO. 2008;1702:45–48
8. DeRosier J, Stalhandske E, Bagian JP, Nudell T. Using health care failure mode and effect analysis: The VA National Center for Patient Safety’s prospective risk analysis system. Jt Comm J Qual Improv. 2002;28:248–67, 209
9. Uchino S, Kellum JA, Bellomo R, et al.Beginning and Ending Supportive Therapy for the Kidney (BEST Kidney) Investigators. Acute renal failure in critically ill patients: A multinational, multicenter study. JAMA. 2005;294:813–818
10. Lozano MJ, Sánchez Izquierdo JA, López MT, Cuenca M. Implantación de una nueva técnica en UCI. Estudio metodológico. Enferm Intensiva. 1993;4:S29
11. Ricci Z, Ronco C, Bachetoni A, et al. Solute removal during continuous renal replacement therapy in critically ill patients: Convection versus diffusion. Crit Care. 2006;10:R67
12. Palevsky PM, Zhang JH, et al.VA/NIH Acute Renal Failure Trial Network. Intensity of renal support in critically ill patients with acute kidney injury. N Engl J Med. 2008;359:7–20
13. Bellomo R, Cass ARENAL Replacement Therapy Study Investigators. . Intensity of continuous renal-replacement therapy in critically ill patients. N Engl J Med. 2009;361:1627–1638
14. Various authors. . Renal support in critically ill patients with acute kidney injury. N Engl J Med. 2008;359:1959–1962
15. Maynar Moliner J, Honore PM, Sánchez-Izquierdo Riera JA, Herrera Gutiérrez M, Spapen HD. Handling continuous renal replacement therapy-related adverse effects in intensive care unit patients: The dialytrauma concept. Blood Purif. 2012;34:177–185
16. Maynar J, Sánchez-Izquierdo JA, Herrera ME, Gainza FJRoglán A, Net A. Dialtrauma y otras complicaciones relacionadas con los tratamientos de depuración extracorpórea de la sangre Disfunción renal aguda en el paciente crítico. 2009 Barcelona, Spain Ars Médica:281–289
17. Luyckx VA, Bonventre JV. Dose of dialysis in acute renal failure. Semin Dial. 2004;17:30–36
18. Maynar Moliner J, Sanchez-Izquierdo Riera JA, Herrera Gutierrez M. Dosis de diálisis en la insuficiencia renal aguda. REMI. 2008;8:A92
19. Uchino S, Bellomo R, Morimatsu H, et al. Continuous renal replacement therapy: A worldwide practice survey. The beginning and ending supportive therapy for the kidney (B.E.S.T. kidney) investigators. Intensive Care Med. 2007;33:1563–1570
20. Bellomo R, Teede H, Boyce N. Anticoagulant regimens in acute continuous hemodiafiltration: A comparative study. Intensive Care Med. 1993;19:329–332
21. Martin PY, Chevrolet JC, Suter P, Favre H. Anticoagulation in patients treated by continuous venovenous hemofiltration: A retrospective study. Am J Kidney Dis. 1994;24:806–812
22. Vinagre Gaspar R, Cornejo Bauer C, Murillo Pérez A, et al. Pérdidas sanguíneas diagnósticas en pacientes con trauma. Enferm Intensiva. 2010;21:120–125
23. Shebl NA, Franklin BD, Barber N. Failure mode and effects analysis outputs: Are they valid? BMC Health Serv Res. 2012;12:150
24. Dean Franklin B, Shebl NA, Barber N. Failure mode and effects analysis: Too little for too much? BMJ Qual Saf. 2012;21:607–611
25. Ashley L, Armitage G, Neary M, Hollingsworth G. A practical guide to failure mode and effects analysis in health care: Making the most of the team and its meetings. Jt Comm J Qual Patient Saf. 2010;36:351–358

patient safety; modal analysis of failure mode and effects; FMEA; extracorporeal depuration; AKI

Copyright © 2016 by the American Society for Artificial Internal Organs