Impact of a National Multimodal Intervention to Prevent Catheter-Related Bloodstream Infection in the ICU: The Spanish Experience Mercedes Palomar, MD, PhD1; Francisco Álvarez-Lerma, MD, PhD2; Alba Riera, RN3; María Teresa Díaz, RN4†; Ferrán Torres, MD, PhD5; Yolanda Agra, MD, PhD6; Itziar Larizgoitia, MD, MPH, PhD4; Christine A. Goeschel, ScD, MPA, MPS, RN7; Peter J. Pronovost, MD, PhD7; on behalf of the Bacteremia Zero Working Group

†Deceased. 1 Intensive Care Unit, Hospital Universitari Arnau de Vilanova, Lleida, Universitat Autònoma de Barcelona, Spanish Society of Intensive and Critical Care Medicine and Coronary Units (SEMICYUC), Barcelona, Spain. 2 Intensive Care Unit, Parc de Salut Mar, Universitat Autònoma de Barcelona, SEMICYUC, Barcelona, Spain. 3 Intensive Care Unit, Hospital Universitari Vall d´Hebron, Barcelona, Spain. 4 Patient Safety Programme, World Health Organization, Geneva, Switzerland. 5 Statistics and Methodology Support Unit (USEM), Hospital Clínic, Universitat de Barcelona, Barcelona, Spain. 6 Quality Agency of the National Health System, Spanish Ministry of Health, Social Policy and Equality, Madrid, Spain. 7 Armstrong Institute for Patient Safety and Quality at Johns Hopkins Medicine, Johns Hopkins University School of Medicine, Bloomberg School of Public Health, and School of Nursing, Baltimore, MA. Authors belonging to the Bacteremia Zero Working Group: Mercedes Palomar, Francisco Álvarez-Lerma, Juan José Otal, Cristobal León, and Alba Riera (SEMICYUC); Yolanda Agra, Eduardo Sierra, Mar Fernández, and Enrique Terol (now with the European Commission) (Quality Agency of the National Health System, Spanish Ministry of Health, Social Policy and Equality); Itziar Larizgoitia and Maria Teresa Díez (Patient Safety Programme, World Health Organization); Rafael Sierra, Maria Aranzazu Isastola, Lourdes Benítez, Carmen Díaz Marisa Dotor, Concepción Barrera, and Enrique Fernández (Andalucía); Isabel Gutiérrez, Sonia Delgado, Teresa Giménez, and Carlos Aibar (Aragón); Armando Blanco, Manuel Valledor, Carmen Martínez, and Belén Suárez (Asturias); Marcio Sa-Gorges and Carlos Campillo (Baleares); Manuel Sánchez, Pablo Jadraque, Carlos Agustín Cruz, and Paloma Garcia de Carlos (Canarias); Jose Luis Teja, Lourdes Gómez, Ana Rosa Díaz, Francisco Antolín, Henar Rebollo, and Trinidad Valle (Cantabria); Enriqueta Muñoz, Llanos Soriano, María Ángeles Arrese, and Sonia Cercenado (Castilla-La Mancha); Maria Jesus López Pueyo, Celia Díaz Tobajas, Sonsoles Paniagua Tejo, and María José Pérez Boíllos (Castilla-León); Mercedes Palomar, Francisco Álvarez-Lerma, Alba Riera, Rosa María Balaguer, Milagros Calizaya, Elisabet Gallart, Pilar Latre, María Isabel Quintana, Jordi Vallés, Juan Carlos Yebenes, Magdalena Campins, Joan María Ferrer, and Josep Davins (Catalunya); Marta Montans, Isabel María Román, Nieves García, and Carmen Gómez (Extremadura); Ana María Díez, Concepción Miralles, Dolores Martin, and Mercedes Carreras (Galicia); Ángel Caballero and Mª Jesús Navaridas (La Rioja); Joaquín Álvarez, Mercedes Nieto, Susana Sánchez, María Jesús Luengo, María del Mar García, Francisco Javier Morón, and Juan Carlos Ansande Copyright © 2013 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/CCM.0b013e3182923622

Critical Care Medicine

(Madrid); Juan Antonio Martínez and Susana Valbuena (Murcia); Eduardo Layana (Navarra); Pedro Olaechea, Ana María Sánchez, David Cantero, and Marbella García (País Vasco); and Paula Ramírez, Josefina Monzón, Rafael Zaragoza, Roberto Reig, Cristina Huertas, and Ricard Meneu (Valencia), Health Region Coordinators, Spanish National Health System. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal). Supported, in part, by the Spanish Ministry of Health, Social Policy and Equality through a contract with SEMICYUC (Number 2008/0273−227−06) and specific contracts with the Health Regions during 2009 and 2010. The authors have disclosed that they do not have any potential conflicts of interest. Address requests for reprints to: Dr. Mercedes Palomar, MD, PhD, Intensive Care Unit, Hospital Universitari Arnau de Vilanova, Av Alcalde Rovira Roure 80, E-250080 Lleida, Spain. E-mail: [email protected]

Objective: Prevention of catheter-related bloodstream infection is a basic objective to optimize patient safety in the ICU. Building on the early success of a patient safety unit-based comprehensive intervention (the Keystone ICU project in Michigan), the Bacteremia Zero project aimed to assess its effectiveness after contextual adaptation at large-scale implementation in Spanish ICUs. Design: Prospective time series. Setting: A total of 192 ICUs throughout Spain. Patients: All patients admitted to the participating ICUs during the study period (baseline April 1 to June 30, 2008; intervention period from January 1, 2009, to June 30, 2010). Intervention: Engagement, education, execution, and evaluation were key program features. Main components of the intervention included a bundle of evidence-based clinical practices during insertion and maintenance of catheters and a unit-based safety program (including patient safety training and identification and analysis of errors through patient safety rounds) to improve the safety culture. Measurements and Main Results: The number of catheter-related bloodstream infections was expressed as median and interquartile range. Poisson distribution was used to calculate incidence www.ccmjournal.org

1

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Palomar et al rates and risk estimates. The participating ICUs accounted for 68% of all ICUs in Spain. Catheter-related bloodstream infection was reduced after 16–18 months of participation (median 3.07 vs 1.12 episodes per 1,000 catheter-days, p < 0.001). The adjusted incidence rate of bacteremia showed a 50% risk reduction (95% CI, 0.39–0.63) at the end of the follow-up period compared with baseline. The reduction was independent of hospital size and type. Conclusions: Results of the Bacteremia Zero project confirmed that the intervention significantly reduced catheter-related bloodstream infection after large-scale implementation in Spanish ICUs. This study suggests that the intervention can also be effective in different socioeconomic contexts even with decentralized health systems. (Crit Care Med 2013; 41:00–00) Key Words: catheter-related bacteremia; intensive care unit; multimodal intervention; patient safety; prevention; safety indicators

C

atheter-related bloodstream infection (CRBSI) among ICU patients is a global problem (1, 2) and responsible for increased mortality, length of ICU stay, and healthcare costs (3–5). Prevention of this condition is crucial to optimize patient safety. Although recommendations have been made to reduce CRBSI (6, 7), adherence is low (8, 9) and rates of CRBSI remain high in many countries (10, 11). Multimodal programs aimed at facilitating compliance with known evidence-based measures while improving teamwork and safety culture have shown to reduce CRBSI rates (12, 13). Among these, the program pioneered by the Johns Hopkins Quality and Safety Research Group (Keystone ICU project) was associated with a dramatic and sustained reduction of CRBSIs (14) and a reduction in mortality for elderly ICU patients (15). In Spain, data from the National Surveillance Program of ICU-Acquired Infection (Estudio Nacional de Vigilancia de Infección Nosocomial-Hospital In Europe Link for Infection Control through Surveillance) (2), active since 1994, showed stable rates of 6–8 CRBSIs per 1,000 central venous catheter (CVC)-days until 2007. In 2008, a collaborative agreement of the Spanish Ministry of Health, Social Policy and Equality, the Patient Safety Programme of the World Health Organization (WHO), the Spanish Society of Intensive and Critical Care Medicine and Coronary Units (Spanish Society), and the Johns Hopkins Quality and Safety Research Group (now the Armstrong Institute for Patient Safety and Quality) was established to implement the Keystone project in Spanish ICUs. A multifaceted intervention, called Bacteremia Zero, was adapted and implemented nationwide over an 18-month period. It was hypothesized that implementation of such intervention would result in a significant reduction in the rates of CRBSI. The main objective of this study was to assess the impact of the Bacteremia Zero project in reducing the rates of CRBSI in Spanish ICUs.

METHODS Design and Setting The Bacteremia Zero study was conducted between April 2008 and June 2010 using a prospective cohort study design. There 2

www.ccmjournal.org

were three phases during the study. The preparatory phase (e.g., activity planning, engagement of ICUs, and baseline data collection) took place from April to December 2008. The implementation phase started in January 2009 and was defined flexibly for each ICU as the first 3 months of participation in the study. Finally, the postintervention phase lasted from month fourth of implementation. The study ended in June 2010. All adult ICUs from the 17 Health Regions in the Spanish National Health System were invited to participate (see Appendix for participating hospitals). WHO and the Spanish Ministry of Health provided supervision and program coordination across the Health Regions, with funding from the Spanish Ministry of Health. The Armstrong Institute provided technical support and advice, whereas the Spanish Society led and coordinated scientific and technical aspects in the ICUs. Details of the general organization of the Spanish healthcare system, intensive care medicine, and description of the ENVIN-HELICS registry are provided in the supplemental data (Supplemental Digital Content 1, http://links.lww.com/ CCM/A683). The study protocol was approved by the Ethics Committee of hospital Parc de Salut Mar, Barcelona (Spain). Also, the institutional review board waived the need for informed consent. Intervention The original Keystone ICU project (14) was adapted to fit the organizational and cultural characteristics of the Spanish healthcare system. Briefly, the Keystone project (14) developed an intervention that targeted clinician’s use of five evidencebased procedures recommended by the Centers for Disease Control and Prevention and identified as having the greatest effect on the rate of CRBSI and the lowest barriers to implementation. These procedures were hand washing, using of full-barrier precautions during the insertion of CVCs, cleaning the skin with chlorhexidine, avoiding the femoral site if possible, and removing unnecessary catheters. In addition to these measures, the ICUs implemented a comprehensive unit-based safety program to improve the safety culture. The intervention was based on the principles of engage, educate, execute, and evaluate as proposed by the Keystone ICU project (14). Engagement was sought through regular and structured awareness raising and problem-solving meetings at the regional, hospital, and ICU levels. An electronic interactive web-based platform (EZcollab site) and direct and timely access to ICU results through http://hws.vhebron.net/bacteriemia-zero/ favored engagement and commitment. ICU staff took a 6-hour online training course (2) about the prevention and control of CRBSI, as well as on patient safety concepts. Staff in each ICU implemented the intervention consisting of a bundle of clinical evidence-based practices intended to prevent the transmission of bacteria during catheter insertion and maintenance, and selected tools drawn from the Keystone comprehensive unit-based safety program aimed at strengthening the safety culture in the ICU. The Spanish technical bundle recommended the subclavian vein as the preferred catheter insertion site (instead of avoiding the femoral site as in the /CTOBERs6OLUMEs.UMBER

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Clinical Investigation

original intervention) (14) and added an item about proper catheter maintenance. In this respect, staff was instructed to reduce the handling of hubs and clean the injection ports with isopropyl alcohol 70° before accessing the system. Patient safety culture evaluation was carried out through the Spanish version of the Hospital Survey on Patient Safety (16). Other items of the comprehensive unit-based program included patient safety training (adapted tools), identification and analysis of errors in daily practice through patient safety rounds and other activities, and alliance with hospital managers. Evaluation was facilitated through regular and timely access to ICU data. Teams involving ICU physicians and nurses coordinated and supervised operations at every unit, and across units in the health region, with the participation of at least one staff intensivist and one staff nurse for each unit. They interacted with hospital authorities and provided coordination across units, as well as leadership, coaching, and supervision. At the local level, data were reviewed by the responsible person at each ICU. The web-based program provided real-time information of data entered, including both infection-related and patient-related information. The program has the appropriate mechanisms that prevented to enter illogical data and advises if data are incomplete. In the context of a safety strategy to guarantee feedback, all healthcare personnel of each ICU had online access to the evolution of infection rates, which represented a supplementary control measure. Also, the coordinators in each health region and at national level reviewed the data every 3 months. Two preselected participants from the Bacteremia Zero Working Group (A.R., J.J.O.) were responsible to solve problems in case of discordant data by contacting the corresponding ICU and checking the information of each particular case. Finally, the database administrator reviewed data in the search of discordant results. Although, it is not possible

to make a 100% exclusion of duplicate data, this possibility is largely improbable. Definitions We defined bloodstream infections (BSIs) as those acquired in the ICU and those diagnosed within 48 hours after patient transfer to another hospital unit. We used definitions from the European surveillance program for nosocomial infections (Table 1) (17). Rates of CRBSI included BSI with origin catheter and BSI with unknown origin. Data Collection and Measures Physicians and/or nurses responsible for surveillance collected and entered the number of patients, patient-days, and catheter-days (patients with one or more CVC) per month at each ICU (2). The device use ratio was the number of catheter-days divided by the number of patient-days. The catheterrelated BSI (including CVC, arterial, and peripheral catheters) and BSI secondary to infections in other foci (e.g., respiratory and abdominal) were recorded, as well as patient’s age, gender, Acute Physiology and Chronic Health Evaluation (APACHE) II score, length of ICU stay, and mortality. Other variables were ICU characteristics, public or private hospital, teaching status, and hospital size (< 200, 200–500, and > 500 beds). A seasonality variable to adjust for the rotation of temporary staff in May and June (which is the period of starting Summer holidays shifts) was also introduced. Furthermore, the number of staff completing the 6-hour online training course, as well as the hospital survey on patient safety, was recorded. We measured adherence to each element of the Spanish BSI bundle (supplemental data, Supplemental Digital Content 1, http://links.lww. com/CCM/A683), as well as the length of follow-up (duration

Bloodstream Infections as Defined by National Surveillance Program of ICU-Acquired Infection TABLE 1.

Infection

Description and Criteria

BSI

A positive BC with recognized pathogen or two BCs with skin contaminant including clinical symptoms

CRBSI

BSI with origin “Catheter” (BSI-C) + BSI with origin “Unknown” (BSI-U) BSI-C

Isolated of the same microorganism (species and identical antibiotic susceptibility testing) in the BC from a peripheral vein and a) A quantitative CVC tip culture ≥ 103 cfu/mL or semiquantitative CVC tip culture > 15 CFU b) A quantitative blood culture ratio CVC blood sample/peripheral blood sample > 5 c) Differential delay of positivity of blood cultures: CVC blood sample culture positive 2 hr or less before peripherals blood culture (blood sample drawn at the same time) d) Positive culture with the same microorganism from pus from insertion site

BSI-U

BSI without apparent focus of infection a) Positive BC, negative CVC tip culture, and disappearance of symptoms within 48 hr after removal of the venous catheterb)Positive BC without identification of the responsible focus

BSI-S

BSI secondary to other infection site. BC positive for the same pathogen isolated from samples from secondary sites (wound site, urine, respiratory tract, other sterile site, and other nonsterile site)

BSI = bloodstream infection, BC = blood culture, CRBSI = catheter-related bloodstream infection, CVC = central venous catheter, CFU = colony-forming unit.

Critical Care Medicine

www.ccmjournal.org

3

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Palomar et al

TABLE 2.

Characteristics of the Study Sample Period

Variable

Baseline

Implementation (0–3 mo)

4–6 mo

7–9 mo

10–12 mo

13–15 mo

16–18 mo

Postimplementation (4–18 mo)

ICUs, na

106

192

192

181

162

147

100

192

ICU admissions, n

14,970

26,160

26,165

24,097

21,777

19,631

10,975

102,645

Patient-days, total n

95,000

164,204

160,983

154,916

140,797

118,870

65,678

641,244

Catheter-days, total n

78,092

128,060

125,834

117,526

112,719

93,234

51,983

501,296

CRBSI, total n

334

379

454

336

308

191

110

1,399

Bloodstream infection origin catheter, total n

148

153

194

144

139

102

52

631

Bloodstream infection origin unknown, total n

186

226

260

192

169

89

58

768

Patients with CRBSI (%)

292 (1.95)

345 (1.32)

Patients with > 1 CRBSI

36 (12.3%)

29 (8.41)

416 (1.59) 313 (1.30) 283 (1.30) 183 (0.93) 106 (0.97) 33 (7.93)

20 (6.39)

21 (7.42)

6 (3.28)

4 (3.77)

1,279 (1.25) 103 (8.05)

CRBSI = catheter-related bloodstream infection (includes catheter-related BSI origin catheter + origin unknown). a Number of units reporting data was not uniform over time.

of participation) for each participating ICU. Patients were followed up to 48 hours after discharge from the ICU. Exposure and Outcomes We defined exposure as the postimplementation period after full implementation of the study intervention, which involved six temporal variables, and compared these values with the baseline value. The primary outcome was the quarterly rate of CRBSI. Secondary outcomes were the rates of CRBSI according to characteristics of the hospital (size and public/private status). Statistical Analysis The number of CRBSI, catheter-days, and incidence rates were expressed as medians and interquartile ranges (25th–75th percentile) and mean incident rates. Monthly data were aggregated into 3-month periods (quarters) to coincide with the implementation periods. The quarterly infection rate was calculated as the number of infections per 1,000 catheter-days for each 3-month period. To explore the exposure-outcome relationship, we used generalized linear mixed regression models with a Poisson distribution (18) to calculate the incidence rates, incidence rate ratio, and 95% CI, considering the ICU unit as random and the other factors as fixed effects. In the final regression analysis, period estimates were adjusted by hospital size, teaching status, and seasonality. All tests of significance were two sided and set at p < 0.05. We used SAS version 9.2 software (Cary, NC) for the analysis.

RESULTS A total of 192 ICUs (68% of ICUs in Spain) participated in the study. Seventy-three ICUs (38%) corresponded to large 4

www.ccmjournal.org

hospitals (> 500 beds), 86 ICUs (44.8%) to medium-sized hospitals, and 33 ICUs (17.2%) to small hospitals (< 200 beds). Ninety-three ICUs (48.4%) belonged to teaching hospitals, and 178 ICUs (92.7%) were found in public hospitals. Overall, 14,879 healthcare professionals (1,616 physicians, 8,598 registered nurses, 4,331 clinical assistants, and 334 other professionals) followed the online training course, and 6,629 of them completed the hospital survey on patient safety culture. Postintervention data correspond to 102,645 ICU admissions, 641,244 patient-days, and 501,296 catheter-days (Table 2). The device use was 82.2% in the preintervention period and 78.2% in the postintervention period (p < 0.001). Patients with at least one CRBSI decreased from 1.95% at baseline to 0.97% at 16–18 months of implementation (p < 0.001). Patients with more than one CRBSI decreased from 12.3% at baseline to 3.77% at 16–18 months after implementing the intervention (p = 0.013). There were no differences across patients developing CRBSI during the preintervention (n = 292) and postintervention (n =1,279) periods in terms of age (mean [SD], 58.6 [15.9] vs 58.9 [16.3] yr; p = 0.766), sex distribution (69.2% vs 69.0% male; p = 1.00), and APACHE II score at ICU admission (mean [SD], 18.9 [7.9] vs 19.7 [8.5]; p = 0.154). The overall median rate of CRBSI decreased from 3.07 infections per 1,000 catheter-days at baseline to 1.12 at 16–18 months after implementing the intervention (p < 0.001) (Table 3). The adjusted incidence rate demonstrated a 50% risk reduction (95% CI, 0.39–0.63) in CRBSI from baseline 3.64 (3.04–4.35) to the end of the 18-month follow-up period 1.80 (1.40–2.32) (p < 0.001) (Table 4). The rate reduction seemed consistent across all hospitals regardless of size (p = 0.383) and public /CTOBERs6OLUMEs.UMBER

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Clinical Investigation

TABLE 3.

Incidence of Catheter-Related Bloodstream Infection Events

Median Incidence Ratea

Catheter-days

Period

ICUs

n

Median (IQR)b

Baseline

106

334

2 (0–5)

78,092

Implementation (0–3 mo)

192

379

1 (0–3)

4–6 mo

192

454

7–9 mo

181

10–12 mo

n

Median (IQR)

Median (IQR)

582 (308–1,022)

3.07 (0.00–5.35)

128,060

529 (280–930)

2.08 (0.00–3.71)

1 (0–3)

125,834

492 (282–903)

2.50 (0.00–4.71)

336

1 (0–3)

117,526

514 (277–894)

2.06 (0.00–4.07)

162

308

1 (0–3)

112,719

519 (323–961)

2.10 (0.00–3.80)

13–15 mo

147

191

1 (0–2)

93,234

473 (253–876)

0.86 (0.00–2.58)

16–18 mo

100

110

1 (0–2)

51,983

370 (222–702)

1.12 (0.00–3.24)

IQR = interquartile range. a Per 1,000 catheter-days. b IQR = 25th–75th percentile.

versus private status (p = 0.176), although it seemed more pronounced in nonteaching hospitals (p = 0.006). An increase in the rates of CRBSI was observed during the months of May and June (p < 0.001) corresponding to a greater number of ICUs entering the program and turnover period of temporary staff, which was adjusted for in the final analysis. The rates of CRBSI seemed to decline during the first 12 months of follow-up but slightly increased thereafter. The group of ICUs for which data were available from the basal period, which accounted for 73% of the total participating units, showed a similar risk reduction of 48% (95% CI, 0.37–0.64) in CRBSI from baseline 3.74 (3.11–4.51) to the end of the 18-month follow-up period 1.81 (1.37–2.40) (p < 0.001). Figure 1 illustrates the crude and adjusted incidence of CRBSI over the study period and across teaching status and hospital size. The different etiologies of CRBSI during the 18-month period after implementation of the Bacteremia Zero project are shown in Supplemental Table 1 (Supplemental Digital Content 1, http://links.lww.com/CCM/A683).

DISCUSSION The Bacteremia Zero study attempted to reduce CRBSIs throughout Spanish ICUs and demonstrates the impact of a multifaceted intervention known to prevent CRBSI in large scale up in Spain. Despite necessary modifications of the intervention, due to cultural and organizational factors, nearly 70% of ICUs across all regions of Spain significantly reduced CRBSI, impacting over 100,000 ICU admissions during the study period, suggesting the cross-cultural effectiveness of the multifaceted intervention developed by Pronovost et al (14). The results suggest that the intervention was effective in reducing the incidence of CRBSI in all types of hospitals. Although nonteaching hospitals showed higher rate reductions than university-affiliated centers, all participating hospitals saw reductions Critical Care Medicine

in infections, suggesting the robustness of the intervention despite the variations adopted in Spain. The strength of the intervention seems thus to lie in its ability to standardize measures and evidence-based practices, while the program encourages local variation in how these practices are implemented. The results are also consistent with others reporting about 60% reduction in the rate of CRBSI, with final rates close to 1 per 1,000 CVC-days (14–19), although some other studies in developing countries still showed rates at more than 7 per 1,000 CVC-days (10). The overall infection rate in the Bacteremia Zero study did drop more slowly than in the original project (14). Furthermore, the median rate of zero achieved in the third month of implementation in Michigan was never reached in Spain. One important reason may be due to the lesser uptake of the full intervention. Few ICUs implemented the entire safety intervention, notably the comprehensive Unit-based safety component as was proposed in Michigan (14, 20, 21) (most ICUs performed safety rounds and partnered with executives, only occasionally or not at all). Difficulties in the follow-up and acceptability of some of these elements were observed among participants since the start of the program. Although the reasons behind this observation need to be further studied, this experience suggests the need for strengthening the patient safety culture and the cross-cultural adaptation of interventions to ensure their uptake and effectiveness. Possible differences in the healthcare system across the United States and Spain and in the organization across and within Spain ICU may also contribute to differences in the overall result. Whereas the Keystone project involved one state only, whereas the Spanish program involved the whole National Health System, with 17 Health Regions and high variability in implementation, leadership, and patient safety culture. For example, there is great variability across the different health regions regarding incentives for health professionals. www.ccmjournal.org

5

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Palomar et al

TABLE 4. Univariate and Adjusted Multivariate Analyses for Catheter-Related Bloodstream Infection Univariate Analysisa Incidencec Estimate (95% CI)

Incidence Rate Ratio Estimate (95% CI)

Incidencec Estimate (95% CI)

Incidence Rate Ratio Estimate (95% CI)

Baseline

4.27 (3.75–4.86)

Ref (p < 0.001)

3.64 (3.04–4.35)

Ref (p < 0.001)

Implementation (0–3 mo)

3.00 (2.64–3.39)

0.70 (0.60–0.82)

2.83 (2.38–3.37)

0.78 (0.66–0.92)

4–6 mo

3.66 (3.26–4.10)

0.86 (0.74–1.00)

3.31 (2.81–3.89)

0.91 (0.78–1.07)

7–9 mo

2.92 (2.56–3.33)

0.68 (0.58–0.80)

2.87 (2.40–3.45)

0.79 (0.66–0.95)

10–12 mo

2.79 (2.43–3.20)

0.65 (0.55–0.77)

2.74 (2.28–3.30)

0.75 (0.63–0.91)

13–15 mo

2.06 (1.74–2.45)

0.48 (0.40–0.59)

1.94 (1.57–2.39)

0.53 (0.43–0.65)

16–18 mo

2.19 (1.77–2.71)

0.51 (0.41–0.65)

1.80 (1.40–2.32)

0.50 (0.39–0.63)

> 500 beds

3.31 (2.99–3.67)

Ref (p = 0.026)

2.94 (2.58–3.34)

Ref (p = 0.383)

200–500 beds

2.69 (2.33–3.12)

0.81 (0.68–0.97)

2.61 (2.26–3.01)

0.89 (0.74–1.07)

< 200 beds

2.28 (1.53–3.40)

0.69 (0.45–1.04)

2.45 (1.65–3.63)

0.83 (0.55–1.27)

Yes

3.37 (3.07–3.69)

Ref (p < 0.001)

3.06 (2.58–3.64)

Ref (p = 0.006)

No

2.33 (1.99–2.74)

0.69 (0.58–0.83)

2.31 (1.93–2.76)

0.75 (0.62–0.92)

May–June

3.58 (3.24–3.96)

Ref (p < 0.001)

2.97 (2.53–3.48)

Ref (p < 0.001)

Other months

2.82 (2.57–3.09)

0.79 (0.71–0.87)

2.38 (2.05–2.77)

0.80 (0.71–0.90)

Public

3.08 (2.83–3.35)

Ref (p = 0.176)

Private

2.28 (1.48–3.50)

0.74 (0.48–1.15)

5–6 mo

4.54 (3.25–6.34)

Ref (p = 0.190)

7–9 mo

3.16 (2.27–4.40)

0.70 (0.43–1.11)

10–12 mo

2.80 (2.09–3.77)

0.62 (0.40–0.96)

13–15 mo

2.94 (2.46–3.52)

0.65 (0.44–0.94)

16–18 mo

2.99 (2.69–3.34)

0.66 (0.46–0.94)

Variable

Effect

Period

Hospital size

Teaching status Seasonality Hospital type Study follow-up

Adjusted Multivariate Analysisb

Ref = reference category for the risk estimates. a One separate model was fitted for each variable. b A generalized linear mixed regression model with a Poisson distribution was done, adding hospital size, university unit, and seasonality to adjust period estimates. c Per 1,000 catheter-days.

Furthermore, the Spanish Health system is for the most part public, providing free access to care; thus, there were no restrictions in access according to ability to pay. At the same time, in Spain, there are only 8.2 ICU beds per 100,000 inhabitants as compared with 20 beds per 100,000 inhabitants in the United States (22), but patients in Spanish ICUs seem to have more CVC (78% of Spanish patients (2) compared with 50% in the U.S. patients (23). Multiple catheterizations may magnify the rates of CRBSI (24), and in the Bacteremia Zero project, patients showed a mean of 1.3 CVC and 0.5 arterial catheters inserted, while many of the CRBSI occurred in patients having concurrent CVC, arterial catheters, and/or hemodialysis catheters. Overall, rates for CRBSI in Europe seem, in general, consistently higher than those in the United States (25), although differences in the systems and definitions of 6

www.ccmjournal.org

surveillance of ICU-acquired infections make comparisons difficult (26). Interestingly, we observed an unexpected peak in the rate of CRBSI during the months of May and June corresponding with the arrival of temporary staff to cover summer leaves. As a result, specific training for temporary staff was reinforced in the second year of the intervention, although the peak remained still observable. This finding was an important lesson learnt from the study, which will presumably lead to further action in the healthcare system. There are limitations to our study. Baseline data were only available for 106 ICUs, and the number of ICUs participating postintervention varied over time. This is typical for large-scale quality improvement studies having voluntary data collection (15). However, improvement was comparable among ICUs with data available for the basal period and those joining the program /CTOBERs6OLUMEs.UMBER

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Clinical Investigation

The Bacteremia Zero study showed that a bundle of evidence-based infection prevention practices can be systematically implemented across Spain and significantly reduce CRBSI among ICU patients. This study demonstrated that large-scale reductions in preventable harm are possible if guided by science, monitored by rigorous measurement, and sensitive to contextual differences among countries and hospitals. Mechanisms to strengthen patient safety culture and thus sustain more effectively the intervention are also probably needed. The Bacteremia Zero project suggests that the original intervention pioneered in Michigan can be effective in relatively different healthcare systems, despite Figure 1. Adjusted period incidence rates estimations for catheter-related infection. Dashed line = all: varying levels of organizational univariate, black continuous line = all: adjusted multivariate, dotted line = university unit: no, blue continuous line = university unit: yes, red continuous line = hospital size: > 500 beds, dotted and dashed line = hospital structure and safety culture. size: 200–500 beds, dotted line and red diamond = hospital size: < 200 beds. The further understandings of its effectiveness and the requirelater, although the fact to remain in the program for more ments for its successful adaptation in other more challenging than 12 months appeared to optimize the results. Although environments remain a priority for other studies. a validation study was not performed, the annual Spanish national incidence study (Estudio Nacional de Prevalencia de ACKNOWLEDGMENTS la Infección Nosocomial en los hospitales españoles) showed a We are grateful for the collaboration of Alberto Infante former reduction in both the incidence of infected patients and infec- General Director of the Quality Agency of the National Health tions, particularly the incidence of bacteremia decreased proSystem of Spain (currently at the National School of Health) in gressively from 0.25% in 2008 to 0.20% in 2009 and 0.19 in ensuring the project could get off the ground; we thank Abig2010 (27). Nonetheless, most participating ICUs had an extenuei Torrents, Hospital Clinic Barcelona, in conducting the data sive experience in nosocomial infection surveillance, which analysis; Cyrus Engineer, Nittita Prasopa-Plaizier, Katthyana would minimize BSI identification and notification errors as Aparicio, and Edward Kelley from the World Health Organidemonstrated in other registries (28). zation in supporting the preparation of the intervention and This study has important public health implications. On its study proposal; María Santaolaya from the Quality Agency extrapolating the study results to the baseline rates according to of the National Health System and José Rodríguez Paz from the reduction in the rates of CRBSI obtained in the study and the Johns Hopkins Hospital in supporting the cultural adaptaconsidering the number of catheter-days in the postimplemention of the materials and the preparation of the intervention; tation period, approximately 742 CRBSIs were prevented with Elizabeth Colantuoni, statistician at Johns Hopkins University, the Bacteremia Zero intervention. With an attributable mortalfor providing valuable statistical advice; Martin Fletcher, now ity of 9% and a prolonged ICU stay of 12 days per bacteremia with the Australian Health Practitioner Regulation Agency, was (29), this decrease may have save 66 lives, 8,904 ICU days, and instrumental in the setting up of the intervention; and Sir Liam approximately 27,629,112€ (for an average cost of 3,103€/ICU Donaldson, now WHO Envoy for Patient Safety, provided the day) (30). The cost of the project for the Ministry of Health vision. We are indebted to managers and healthcare personincluding contracts with Sociedad Española de Medicina nel of the participating hospitals, as well as staff members of Críticay Unidades Coronarias and Health Regions, meetings, the ICUs for their valuable cooperation. We also thank Marta and statistical analysis was 2.340.000€. In the present era of Pulido, MD, PhD, freelance author’s editors for editing the budget and economic constraints, the repercussion of this strat- manuscript and editorial assistance. The fees of medical editegy to reduce CRBSI in ICU patients is very relevant. ing were supported by the Spanish Society of Intensive Care Critical Care Medicine

www.ccmjournal.org

7

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Palomar et al

Medicine (SEMICYUC). We also thank Christine G. Holzmueller, BLA, for her assistance in editing the manuscript; she was not compensated for her work.

REFERENCES

1. Edwards JR, Peterson KD, Yi Mu Y, et al; National Healthcare Safety Network (NHSN) Report: Data summary for 2006 through 2008, issued December 2009 Am J Infect Control 2009; 37:783–805 2. Sociedad Española de Medicina Intensiva, Grupo de Trabajo de Enfermedades Infecciosas (SEMICYUC-GTEI): Estudio Nacional de Vigilancia de Infección Nosocomial en UCI (ENVIN-UCI). Informes de los años 2001–2009. Available at: http://hws.vhebron.net/envinhelics/. Accessed July 25, 2013 3. European Centre for Disease Prevention and Control: Annual Epidemiological Report on Communicable Diseases in Europe 2009. Stockholm, Sweden, European Centre for Disease Prevention and Control, 2009 4. Lambert ML, Suetens C, Savey A, et al: Clinical outcomes of healthcare-associated infections and antimicrobial resistance in patients admitted to European intensive-care units: A cohort study. Lancet Infect Dis 2011; 11:30–38 5. Olaechea PM, Ulibarrena MA, Alvarez-Lerma F, et al; ENVIN-UCI Study Group: Factors related to hospital stay among patients with nosocomial infection acquired in the intensive care unit. Infect Control Hosp Epidemiol 2003; 24:207–213 6. Mermel LA, Allon M, Bouza E, et al: Clinical practice guidelines for the diagnosis and management of intravascular catheter-related infection: 2009 update by the Infectious Diseases Society of America. Clin Infect Dis 2009; 49:1–45 7. Boyce JM, Pittet D: Guideline for hand hygiene in health-care settings: Recommendations of the Healthcare Infection Control Practices Advisory Committee and the HICPAC/SHEA/APIC/IDSA Hand Hygiene Task Force. Society for Healthcare Epidemiology of America/Association for Professionals in Infection Control/Infectious Diseases Society of America. MMWR Recomm Rep 2002; 51:1–45 8. Rubinson L, Wu AW, Haponik AE, et al:Why is it that internists do not follow guidelines for preventing intravascular catheter infections? Infect Control Hosp Epidemiol 2005; 26:525–33 9. Krein SL, Hofer TP, Kowalski CP, et al: Use of central venous catheterrelated bloodstream infection prevention practices by US hospitals. Mayo Clin Proc 2007; 82:672–678 10. Rosenthal VD, Maki DG, Rodrigues C, et al; International Nosocomial Infection Control Consortium Investigators: Impact of International Nosocomial Infection Control Consortium (INICC) strategy on central line-associated bloodstream infection rates in the intensive care units of 15 developing countries. Infect Control Hosp Epidemiol 2010; 31:1264–1272 11. Mehta A, Rosenthal VD, Mehta Y, et al: Device-associated nosocomial infection rates in intensive care units of seven Indian cities. Findings of the International Nosocomial Infection Control Consortium (INICC). J Hosp Infect 2007; 67:168–174 12. Institute for Healthcare Improvement. Available at: http://www.ihi.org. Accessed December 22, 2010 13. Miller MR, Griswold M, Harris JM 2nd, et al: Decreasing PICU catheter-associated bloodstream infections: NACHRI’s quality transformation efforts. Pediatrics 2010; 125:206–213

APPENDIX: HOSPITALS PARTICIPATING IN THE PROJECT ANDALUCIA: Hospital de Valme, Hospital Virgen de la Macarena, Hospital de la Merced, Hospital San Juan de Dios del Aljarafe, Hospital Comarcal de la Axarquía, Hospital General Básico de la Serranía de Ronda, Hospital Universitario Virgen de la Victoria, Hospital Costa del Sol, Hospital Univer, Médico Quirúrgico (Compl. Hosp. de Jaén), Hospital San Juan 8

www.ccmjournal.org

14. Pronovost P, Needham D, Berenholtz S, et al: An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med 2006; 355:2725–2732 15. Pronovost PJ, Goeschel CA, Colantuoni E, et al: Sustaining reductions in catheter related bloodstream infections in Michigan intensive care units: Observational study. BMJ 2010; 340:c309 16. Saturno PJ, Da Silva ZA, Oliveira-Sousa SL, et al; Grupo Proyecto ISEP: Análisis de la cultura sobre seguridad del paciente en los hospitales del Sistema Nacional de Salud Español. Med Clin Monogr (Barc) 2008; 131(Suppl 3):18–25 17. HELICS-ICU: Surveillance of nosocomial infections in intensive care units. Protocol v.6.1. Available at: http://helices.univ-lyon1.fr/protocols/icu_protocol.pdf. Accessed July 25, 2013 18. Molenberghs G, Verbeke G: Models for Discrete Longitudinal Data. New York, Springer Science+Business Media, 2005 19. Peredo R, Sabatier C, Villagrá A, et al: Reduction in catheter-related bloodstream infections in critically ill patients through a multiple system intervention. Eur J Clin Microbiol Infect Dis 2010; 29:1173–1177 20. Sexton JB, Berenholtz SM, Goeschel CA, et al: Assessing and improving safety climate in a large cohort of intensive care units. Crit Care Med 2011; 39:934–939 21. Romig M, Goeschel C, Pronovost P, et al: Integrating CUSP and TRIP to improve patient safety. Hosp Pract (Minneap) 2010; 38:114–121 22. Wunsch H, Angus DC, Harrison DA, et al: Variation in critical care services across North America and Western Europe. Crit Care Med 2008; 36:2787–2793, e1 23. Centers for Disease Control and Prevention (CDC): Vital signs: Central line-associated blood stream infections - United States, 2001, 2008 and 2009. MMWR Morb Mortal Wkly Rep 2011; 60:243–248 24. Aslakson RA, Romig M, Galvagno SM, et al: Effect of accounting for multiple concurrent catheters on central line-associated bloodstream infection rates: Practical data supporting a theoretical concern. Infect Control Hosp Epidemiol 2011; 32:121–124 25. Antimicrobial resistance and healthcare-associated infections, Annual Epidemiological Report 2010. Available at: http://www.ecdc.europa. eu/en/activities/surveillance/hai/pages/default.aspx. Accesssed July 4, 2011 26. Hansen S, Sohr D, Geffers C, et al: The concordance of European and US definitions for healthcare-associated infections (HAI). BMC Proc 2011; 5(Suppl 6): O2 27. EPINE 1990–2011: 22 AÑOS. Available at: http://www.sempsph.com/ sempsph/.../378_9-epine_1990-2011.pdf. Accessed July 25, 2013 28. Backman LA, Melchreit R, Rodriguez R: Validation of the surveillance and reporting of central line-associated bloodstream infection data to a state health department. Am J Infect Control 2010; 38:832–838 29. Olaechea PM, Álvarez-Lerma F, Palomar M, et al; ENVIN Study Group “Mortality attributable to primary and catheter-related nosocomial bacteremia. A case control study.” 22nd Annual Congress of European Society of Intensive Care Medicine. Vienna 11–14 October 2009. Intensive Care Med 2009; S1–269 30. Ministerio de Sanidad, Política Social e Igualdad [Internet]. Madrid: Estadística de Establecimientos Sanitarios con Régimen de Internado. Indicadores Hospitalarios. Evolución 2000–2008. Available at: http:// www.mspsi.gob.es/estadEstudios/estadisticas/estHospiInternado/ inforAnual/homeESCRI.htm. Accessed July 4, 2011

de la Cruz, Hospital Neurotraumatológico de Jaén, Hospital Alto Guadalquivir, Hospital Médico Quirúrgico Virgen de las Nieves (UCI), Hospital General Básico de Baza, Hospital General Básico Santa Ana, Hospital Médico Quirúrgico Vírgen de las Nieves (UC y UCC), Hospital Universitario Puerta del Mar, Hospital del SAS de Jerez, Hospital Universitario de Puerto Real, Hospital Punta Europa, Hospital de Antequera, Hospital de Montilla, Hospital Reina Sofía, Hospital Valle de los Pedroches, /CTOBERs6OLUMEs.UMBER

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Clinical Investigation

Hospital Infanta Margarita, Hospital General de Huelva Juan Ramón Jiménez, Hospital General Básico de Riotinto, Hospital Torrecárdenas, Hospital de Poniente, Hospital Comarcal La Inmaculada, Hospital de La Línea. ARAGON: Hospital Clínico Universitario Lozano Blesa, Hospital Royo Villanova, Hospital General San Jorge, Hospital Obispo Polanco. ASTURIAS: Hospital Central de Asturias (UCI 1), Hospital Central de Asturias (UCI pediátrica), Hospital Central de Asturias (UCI HGA), Hospital Central de Asturias (UCI INS), Hospital de Cabueñes, Hospital de San Agustín, Hospital Valle del Nalón. BALEARES: Hospital Son Dureta, USP Clínica Palmaplanas, Hospital Son Llàtzer, Fundación Hospital Manacor, Clínica Rotger Hospita, Hospital Mateu Orfila, l Can Misses, Eivissa. CANARIAS: Hospital Universitario Insular de Gran Canaria, Hospital Universitario de Gran Canaria Dr. Negrín (U. Cardiología), Hospital Universitario de Gran Canaria Dr. Negrín (U. Neurotrauma), Hospital General de Lanzarote, Hospital General de Fuerteventura, Hospital Universitario de Canarias, Hospital Ntra. Sra. de Candelaria, Hospital General de La Palma. CANTABRIA: Hospital Marqués de Valdecilla (UCI 1), Hospital Marqués de Valdecilla (UCI 2 Politrauma), Hospital Marqués de Valdecilla (UCI 3). CASTILLA-LA MANCHA: Hospital Sierrallana (URCE), Hospital Virgen de la Salud, Hospital Nacional de Parapléjicos, Hospital Virgen de la Salud (U.C. Críticos Cirugía Cardíaca), Hospital Provincial de la Misericordia, Hospital Nuestra Señora del Prado, Hospital General Universitario de Albacete, Hospital General Universitario de Albacete, Hospital General de Ciudad Real, Hospital Universitario de Guadalajara, Hospital Virgen de la Luz. CASTILLA LEON: Hospital Universitario Río Hortega, Hospital Clínico Universitario de Valladolid, Hospital Virgen de la Vega, Hospital Clínico de Salamanca, Complejo Hospitalario de Soria, Hospital General de Segovia, Hospital Ntra. Sra. de Sonsoles, Hospital General Yagüe, Hospital de León (UCI polivalente), Complejo Hospitalario de Palencia, Hospital Virgen de la Concha. CATALUNYA: Hospital General Hospitalet (Unidad de semi-intensivos), Capio Hospital General de Catalunya, Hospital General Vall Hebron (UCI), Hospital de Traumatología Vall Hebron, Hospital General Vall Hebron (UPCC), Hospital Clínic i Provincial (UCI Quirúrgica), Hospital Asepeyo. Sant Cugat del Vallés, Hospital de Sant Pau (UCI Polivalente), Centro Médico Delfos. Barcelona, Hospital del Mar, Hospital Dos de Maig, Hospital Plató, Hospital Universitari Sagrat Cor, Hospital de Barcelona (SCIAS), Hospital General Hospitalet, Hospital de Traumatologia Vall Hebron (U. Quemados), Clínica Corachán, Hospital Universitario Mutua Terrassa, Hospital de Terrassa, Hospital Parc Tauli, Hospital de Mataró. Consorci Sanitari del Maresme, Hospital Sant Joan de Deu de Manresa, Hospital de Igualada, Hospital

Critical Care Medicine

General de Granollers, Clínica Girona, Hospital Universitari de Girona Doctor Josep Trueta, Hospital Universitari Arnau de Vilanova de Lleida (UCI), Hospital de Santa María de Lleida, Hospital Universitari Joan XXIII, Hospital Universitari de Sant Joan, Hospital Verge de la Cinta. EUSKADI: Hospital Santiago de Vitoria, Hospital de Txagorritxu, Vitoria, Hospital Donostia (Ntra. Sra. de Aranzazu), Hospital de Basurto, Hospital de Galdakao. EXTREMADURA: Hospital de Mérida, Hospital San Pedro de Alcántara, Hospital San Pedro. GALICIA: Complexo Hospitalario Universitario Juan Canalejo (Reanimación), Complexo Hospitalario Universitario Juan Canalejo (Quemados), Complexo Hospitalario Universitario A Coruña (UCI 6), Complexo Hospitalario Universitario de A Coruña (UCI 5ª), Hospital Clínico Universitario de Santiago (UCI), Hospital Arquitecto Marcide, Complexo Hospitalario de Ourense, Hospital Montecelo, Hospital Povisa, Hospital Meixoeiro—C. Hosp. Universitario de Vigo (UCI Médica). MADRID: Hospital Clínico Univer. San Carlos (U. Médico-Quirúrgica), Hospital Clínico Univer. San Carlos (U. Neuro-Politrauma), Hospital La Paz (Unidad de Quemados), Hospital La Paz (UCI polivalente), Fundación Jiménez Díaz, Clínica Puerta de Hierro, Hospital Ramón y Cajal, Hospital Clínico Universitario San Carlos (U. Cardiovascular), Hospital Universitario Fundación Alcorcón, Hospital Gregorio Marañón, Clínica Moncloa, Hospital Universitario 12 de Octubre, Hospital Universitario 12 de Octubre (UCI Traumatología), Hospital Príncipe de Asturias, Hospital General. Móstoles, Hospital Severo Ochoa. Leganés, Hospital del Henares, Hospital de Getafe (UCI Polivalente), Hospital Infanta Cristina, Hospital Universitario 12 de Octubre (UCP), Hospital de la Princesa, Hospital Universitario de Fuenlabrada, Hospital del Sureste, Hospital del Tajo, Hospital Infanta Sofía, Hospital Infanta Leonor. MURCIA: Hospital Virgen de la Arrixaca, Hospital Universitario J. Mª Morales Meseguer, Hospital Santa María del Rosell, Hospital General Universitario Reina Sofía. NAVARRA: Hospital de Navarra, Hospital Virgen del Camino, Hospital García Orcoyen, Clínica San Miguel. VALENCIA: Hospital Universitario La Fe (U. Reanimación), Hospital Universitario La Fe, Hospital Arnau de Vilanova de Valencia, Hospital Clínico Universitario de Valencia (U. Polivalente), Hospital General Universitario de Valencia (UPCC), Hospital General de Requena, Hospital Doctor Peset, Hospital de Sagunto, Valencia, Hospital de la Ribera, Hospital Francesc de Borja de Gandía, Hospital Marina Baixa de Villajoyosa—departamento 16 AVS, Hospital Lluís Alcanyís de Xàtiva, Hospital de Torrevieja Salud, Hospital General Universitario de Alicante (UCI), Hospital General Universitario de Elche, Hospital de Sant Joan, Hospital General de Castellón, Consorcio Hospitalario Provincial de Castellón, Hospital Comarcal de Vinaròs.

www.ccmjournal.org

9

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited