Silvia Calvino Gu¨nther Carole Schwebel Aure´lien Ve´sin Judith Remy Geraldine Dessertaine Jean-Franc¸ois Timsit
Interventions to decrease tube, line, and drain
removals in intensive care units: the FRATER
study
Received: 5 February 2009 Accepted: 4 June 2009
Ó Springer-Verlag 2009
The article was written on behalf of the FRATER study group. FRATER is the acronym for French Risk Assessment for TubE Removal, whose participants are listed in the appendix.
Electronic supplementary material The online version of this article
(doi:10.1007/s00134-009-1555-8) contains supplementary material, which is available to authorized users.
A. Ve´sin J.-F. Timsit (
)
)INSERM U823, University Joseph Fourier, Albert Bonniot Institute,
38706 La Tronche Cedex, France e-mail: jftimsit@chu-grenoble.fr Tel.: ?33-476-768779
Fax: ?33-476-765569
S. Calvino Gu¨nther C. Schwebel J. Remy G. Dessertaine J.-F. Timsit Medical ICU, Michallon Teaching Hospital, BP 217, 38043 Grenoble Cedex 9, France
Abstract Purposes: To evaluate the incidence of unintended tube, line, and drain removals (UTRs) in our ICU, to identify system factors associated with UTRs, and to evalu-ate interventions designed to decrease UTR occurrence. Methods: Inter-ventional study in the 18-bed medical ICU of a French general university hospital. We prospectively deter-mined the incidence and
circumstances of UTRs in our ICU over a 2-year period. Demographic and clinical data were collected for consecutively admitted patients, and additional information was recorded about patients experiencing UTRs. Investigators analyzed UTR data twice a month to identify possible causes and developed interventions to decrease UTRs (mainly securing tubes and sedation protocol). Condi-tional logistic regression stratified on length of stay was used to identify risk factors for UTRs and segmented linear regression analysis to test the effects of interventions. Results: Of 2,007 admitted patients (12,256 patient days), 193 (9.6%) experienced
270 UTRs (22/1,000 patient days). Clinical or therapeutic consequences occurred for 17% of UTRs. Three factors were independently associated with UTR; two were risk factors, namely, admission for coma [OR, 2.68; 95% CI (1.87; 3.84);
P\ 0.0001] and mechanical ventila-tion in over 65% of all ICU patients [OR = 1.65 (1.19; 2.29); P = 0.003], and one was protective, namely, mean SAPS II [45 in all ICU patients [OR, 0.54; 95% CI (0.39; 0.75);
P = 0.0003]. Segmented regression analysis showed a 67.4% drop [95% CI (17.2%; 117.3%); P = 0.009] in the UTR rate after the first interven-tion was introduced. System factors played a major role in UTR occur-rence. Conclusion: UTRs are common. A continuous quality-improvement program can reduce UTR rates in the ICU.
Keywords Tube removal Intensive care Intervention Adverse event Quality of care
Introduction
Problems with line, tube, and drain placement, mainte-nance, or removal are among the most common adverse events in the intensive care unit (ICU) [1]. UTRs adversely affect morbidity and mortality [2–6], length of ICU stay, and hospitalization costs [4–6]. Moreover,
UTRs increase the workload and generate stress among ICU staff [7,8]. Studies suggest that at least one-third of all UTRs may be preventable and that most UTRs are ascribable to human error [4,9–12]. Decreasing the UTR rate may improve patient outcomes [13].
Among published studies, most were designed to identify risk factors or to evaluate early detection methods
Intensive Care Med
[5,10,14,15]. Furthermore, many studies were confined to a single type of UTR, usually unintended endotracheal tube removal [9,15,16].
The aims of this study were to assess the incidence of UTRs in an ICU, to look for system factors associated with UTRs, and to evaluate the effects of a continuous quality-improvement program developed by the ICU staff.
Materials and methods
This study was conducted at the medical ICU of the university hospital in Grenoble, France from January 2005 to January 2007. The ICU has 18 beds in four geographically separate sections. Eight nurses and six nursing assistants work the morning shift, eight nurses and four nursing assistants the afternoon shift, and seven nurses and two nursing assistants the night shift.
Data collection
The ICU nurses were asked to participate in a quality-improvement program designed to evaluate and reduce the incidence of UTRs. UTRs were defined as all unplanned withdrawals of lines, tubes, or drains that might cause patient harm, including accidental removals by patients or staff and intentional self-extubation. Patient characteristics and information on each UTR were col-lected daily on standardized forms and entered into a computer after validation by the research nurse (Elec-tronic Supplement E1). UTRs were classified as severe if they had at least one clinical consequence that required specific treatment (Electronic Supplement E2).
We recorded at each patient admission the nurse-to-patient ratio, proportion of nurse-to-patients receiving mechanical ventilation (MV), and mean SAPS II in the ICU to control for severity and workload. We used the system factors approach to analyze UTR occurrence in relation to six potential system factors (patient, provider, team, training, task, and management characteristics) [4].
The FRATER study was approved by the institutional review board, which waived the need for informed consent.
Interventions to decrease unintended tube removals A committee of nursing assistants, nurses, head nurses, and intensivists met twice a month to review UTRs and to devise means of reducing UTR occurrence. The three major interventions were changing the type of suture used
to secure drains and catheters and reinforcing standard procedures regarding dressings (1 June 2005), a new method for securing endotracheal tubes (1 July 2005), and introducing a sedation protocol (1 October 2006). Other minor changes were implemented (Electronic Supplement E3).
Once a new intervention was implemented, adherence was checked every morning by the steering committee and staff nurses. Deviations were corrected immediately. No other interventions took place during the FRATER study period. Feedback was given to the ICU team twice a month.
Statistical analysis
Differences in patient characteristics across study periods were assessed using the chi-square or Kruskal–Wallis test. Univariate and multivariate conditional logistic regression stratified by length of stay was used to identify UTR risk factors. Variable values for each patient were converted into weekly values, and an interrupted time-series analy-sis was done using segmented linear regression to assess the effect of interventions on the weekly UTR incidence [17] with adjustment of confounders. P values\0.05 were considered significant (see Electronic Supplement E4 for details). All statistical analyses were performed using SAS 9.1 (SAS Institute; Cary, NC).
Results
Patient characteristics
We included 2,007 patients over 12,256 ICU days. One or more UTRs occurred in 193 (9.6%) patients (Table1). Patients with UTRs had a lower ICU mortality rate than other patients (9.8 vs. 18.0%) but had longer ICU stays: median (IQR), 7 [14] vs. 3 [4].
Risk factors for having at least one UTR are listed in the Electronic Supplement E5. Three factors were inde-pendently associated with UTR occurrence: admission for coma [OR, 2.68; 95% CI (1.87; 3.84); P \ 0.0001], mean SAPS II in the ICU [45 [OR, 0.54; 95% CI (0.39; 0.75); P = 0.0003]; and proportion of patients on MV [65% [OR = 1.65 (1.19; 2.29); P = 0.003].
Description of unintended tube removals
During the study, 270 UTRs occurred (22/1,000 patient days), of which 38 (17%) were severe. They chiefly involved nasogastric tubes (19.9/1,000 procedure days),
endotracheal tubes (13.7/1,000 procedure days), periphe-ral vein catheters (7.4/1,000 procedure days), and chest tubes (5.1/1,000 procedure days). Overall, 47.4% of removed tubes were reinserted.
UTRs were not more common on weekends. However, 31.5% of UTRs occurred during nurse shift changes at 7 a.m., 2 p.m., and 9 p.m. Of patients with UTRs, 44% had no physical restraints.
The system factors analysis identified at least two contributing factors for 65% of UTRs. Management (175) and teamwork (118) factors were the most common, whereas no task factors were identified (Electronic Sup-plement E6).
Effects of interventions designed to prevent unintended tube removal
Patients admitted after implementation of the first inter-vention were less likely to experience UTRs (Electronic Supplement E5), but the difference was no longer significant after adjustment for independent factors asso-ciated with UTRs [OR = 0.76 (0.50; 1.13); P = 0.18]. However, interrupted time-series analysis showed that introduction of the first intervention was followed by a 67.4% decrease in UTRs/1,000 procedure days [95% CI (17.2%; 117.3%); P = 0.009] (see Fig.1 and Electronic Supplements E7 and E8). Specifically, UTRs decreased
Table 1 Patient characteristics at admission for each study period Study period at patient admission
Parameters Control period N = 338 Intervention 1a N = 59 Interventions and 2a N = 242 Interventions 1, 2 and 3a N = 1,368 P valueb All N = 2,007 Patient characteristics Age, mean ± SD 59.6 ± 17.9 59.6 ± 18.4 59.6 ± 17.7 60.3 ± 17.7 0.86 60.1 ± 17.7 Female, n (%) 141 (42) 21 (35.6) 111 (45.9) 529 (38.7) 0.15 802 (40)
Main reason for ICU admission, n (%)
Scheduled surgery/monitoring 94 (28) 14 (23.7) 64 (26.5) 389 (28.4) 0.036 561 (28)
Multiple organ failures and shock 92 (27) 15 (25.4) 71 (29.3) 322 (23.5) 500 (25)
Drug overdose/alcoholic coma 68 (20) 13 (22.0) 54 (22.3) 261 (19.1) 396 (20)
Respiratory failure 72 (21) 13 (22.0) 49 (20.3) 292 (21.4) 426 (21) Other 12 (4) 4 (6.8) 4 (1.7) 104 (7.6) 124 (6.2) Type of patient, n (%) Scheduled surgery 21 (6) 4 (6.8) 14 (5.8) 96 (7.0) 0.11 135(7) Emergency surgery 56 (17) 9 (15.3) 35 (14.5) 148 (10.8) 248 (12) Medical 261 (78) 46 (78.0) 193 (79.8) 1,124 (82.2) 1,624 (81) ICU sectionc, n (%) A 38 (11) – 11 (5) 207 (15) \0.0001 256 (13) B 48 (14) 6 (10) 33 (14) 266 (14) 353 (18) C 61 (18) 14 (24) 36 (15) 193 (14) 304 (15) D 190 (56) 39 (66) 162 (67) 702 (51) 1,093 (54)
Length of ICU stay in days, median (IQR)
4 (6) 3 (5) 3 (4) 3 (4) 0.07 3 (4)
SAPS II, mean ± SD 39 ± 19 40 ± 21 38 ± 22 43 ± 22 \0.0001 42 ± 22
Died in the ICU, n (%) 50 (15) 8 (14) 37 (15) 251 (18) 0.28 346 (17)
ICU characteristics at patient admission Mean SAPS II in the ICU
mean ± SD
41 ± 3 43 ± 4 43 ± 4 47 ± 5 \0.0001 46 ± 5
Proportion of patients on MV, mean ± SD
66.9 ± 10.3 77.1 ± 12 61.2 ± 10.8 63.2 ± 13 \0.0001 64 ± 12.6
Nurse/bed ratio mean ± SD 0.40 ± 0.02 0.43 ± 0 0.41 ± 0.02 0.42 ± 0.02 \0.0001 0.42 ± 0.02
Nursing assistant/bed ratio, mean ± SD
0.17 ± 0.03 0.18 ± 0.04 0.19 ± 0.04 0.20 ± 0.04 \0.0001 0.19 ± 0.04
ICU intensive care unit, SAPS II simplified acute physiology score version II, COPD chronic obstructive pulmonary disease, IQR interquartile range
a
First intervention: reinforcement of chest-tube and abdominal-drain dressings, with a change in the type of suture (1 June 2005); second intervention: new method for securing endotracheal tubes (1 July 2005); third intervention: routine use of pain and comfort scales (1 October 2005)
b
P value comparing the study periods: control period, intervention 1, intervention 1 and 2, intervention 1, 2, and 3
c
Sections A, B, and C each have four one-bed rooms. Section D has two two-bed rooms and two one-bed rooms. The section was unknown for one patient. Section A was closed during the inter-vention 1 period
from 6.9 to 3/1,000 procedure days for drains and cath-eters after intervention 1, from 23.7 to 10.7/1,000 pro-cedure days for endotracheal tubes after intervention 2, and from 8.0 to 5.2/1,000 procedure days for all tubes after intervention 3.
Discussion
Our large single-center study provided detailed informa-tion about UTRs and the patients who experienced them. Admission for coma and admission when over 65% of ICU patients were on MV were associated with higher UTR rates, whereas a mean SAPS II greater than 45 in the ICU was associated with fewer UTRs. The adjusted effect of the interventions was not significant at the patient level. However, the interrupted time-series analysis showed a significant 67.4% decrease in the estimated UTRs rate after the first intervention (Fig.1). The system factors analysis (Electronic Supplement E6) identified team and management factors amenable to improvement; few pro-vider, training, and task factors were incriminated. The magnitude of the risk decrease was similar for all tube types, indicating that system factor correction was effec-tive in decreasing both targeted and non-targeted UTRs (Electronic Supplement E8).
Our study has several limitations. Because the data-collection forms were anonymous, no data were collected on the training and skills of staff involved in UTRs. Second, data on pain and comfort were collected retro-spectively and only in patients with UTRs, precluding comparisons of patients with and without UTRs. Third, the impact of the interventions may have been overesti-mated due to regression toward the mean and the
Hawthorne effect. Indeed, the segmented regression showed that the overall UTR rate fell with implementa-tion of the first intervenimplementa-tion, but remained unchanged by subsequent interventions.
Our finding that greater MV use in the ICU was associated with more UTRs and higher SAPS II scores with fewer UTRs indicated that, for a similar intensity of care (reflected by the percentage of MV patients), UTRs were more common in patients with less severe disease. One possible explanation is that patients who were doing better were less likely to be sedated.
Our finding that 44% of UTRs occurred in patients without physical restraints prompted a review of our restraint policy. Studies found that restraints either decreased UTR rates [11] or had no effect [18]. We developed a new written policy for using restraints, and we introduced the routine use of bed barriers (see Elec-tronic Supplement E3). The most common risk factors listed in the UTR reports were patient related. Whether these factors are amenable to interventions is unclear.
Workload and care organization may affect UTR occurrence. Although the staff reported a high workload for 40% of UTRs, neither nurse workload nor under-staffing was associated with the incidence of UTRs. A single study [10] found that high nurse workload was associated with an increased risk of major iatrogenic complications.
Mortality was not increased in patients with UTRs, in keeping with other studies [5,9,14]. Early UTR detection followed by appropriate care may have avoided an adverse impact on survival in most patients. Moreover, most of the UTRs consisted in removal of nasogastric tubes or peripheral vascular catheters without adverse consequences. This fact may have masked a possible effect on mortality of the more worrisome events. 0 2 4 6 8 10 12 14 16 18 20 9 19 29 39 49 59 69 79 89 99 Week
Rate of tube removals for 1000 procedure days
Tube removal rate (for 1000 procedure days) Predicted tube removal rate without intervention 1 Predicted tube removal rate with Intervention 1
Before Intervention In te rv ent ion 1 In te rv ent ion 2 In te rv ent ion 3
Fig. 1 Number of observed and predicted tube removals per 1,000 procedure days, before and after the interventions. Diamonds represent the observed rate of unintended tube removals per 1,000 procedure days, reported each week. The black line indicates the predicted rate of unintended tube removals per 1,000 procedure days with the intervention; the dashed line shows the predicted rate without any intervention. Predictions were obtained using the most parsimonious model 1b (Table 5) obtained after stepwise elimination of nonsignificant terms
Our study shows that the UTR rate can be decreased by implementing a continuous prevention program. Efforts are needed to identify the types of training that would help ICU staff members to decrease the rate of UTRs and to remain involved in quality-improvement programs.
Appendix
Members of the FRATER study group Scientific committee
Rebecca Hamidfar-Roy MD, Medical ICU, Michallon Teaching Hospital, Grenoble.
Patricia Peurois, Medical ICU, Michallon Teaching Hospital, Grenoble.
Alexis Tabah, Medical ICU, Michallon Teaching Hospital, Grenoble.
Alexandra Michel-Herzberg RN, Medical ICU, Michallon Teaching Hospital, Grenoble.
Laure Hammer MD, Medical ICU, Michallon Teach-ing Hospital, Grenoble.
Nade`ge Oudart RN, Medical ICU, Michallon Teach-ing Hospital, Grenoble.
Agne`s Bonadona MD, Medical ICU, Michallon Teaching Hospital, Grenoble.
Maria Dacosta RN, Medical ICU, Michallon Teaching Hospital, Grenoble.
Biostatistical and informatics expertise
Adrien Francais M.Sc, Outcomerea, Inserm U823, Team Outcome of Cancer and Critical Illness, Grenoble, France. Aure´lien Vesin M.Sc, Outcomerea, Inserm U823, Team Outcome of Cancer and Critical Illness, Grenoble, France.
Timsit Jean-Francois, Outcomerea, Inserm U823; University Joseph Fourier (Grenoble 1); Grenoble, France.
References
1. Buckley TA, Short TG, Rowbottom YM, Oh TE (1997) Critical incident reporting in the intensive care unit. Anaesthesia 52:403–409
2. De Lassence A, Alberti C, Azoulay E, Le Miere E, Cheval C, Vincent F, Cohen Y, Garrouste-Orgeas M, Adrie C, Troche G (2002) Impact of unplanned extubation and reintubation after weaning on nosocomial
pneumonia risk in the intensive care unit: a prospective multicenter study. Anesthesiology 97:148–156
3. Garrouste Orgeas M, Timsit JF, Soufir L, Tafflet M, Adrie C, Philippart F, Zahar JR, Clec’h C, Goldran-Toledano D, Jamali S, Dumenil A-S, Azoulay E, Carlet J (2008) Impact of adverse events on outcomes in intensive care unit patients. Crit Care Med 36:2041– 2047
4. Needham DM, Sinopoli DJ, Thompson DA, Holzmueller CG, Dorman T, Lubomski LH, Wu AW, Morlock LL, Makary MA, Pronovost PJ (2005) A system factors analysis of ‘‘line, tube, and drain’’ incidents in the intensive care unit. Crit Care Med 33:1701–1707 5. Rose DK, Cohen MM, Wigglesworth
DF, DeBoer DP (1994) Critical respiratory events in the postanesthesia care unit, patient, surgical, and anesthetic factors. Anesthesiology 81:410–418
6. Sinopoli DJ, Needham DM, Thompson DA, Holzmueller CG, Dorman T, Lubomski LH, Wu AW, Morlock LL, Makary MA, Pronovost PJ (2007) Intensive care unit safety incidents for medical versus surgical patients: a prospective multicenter study. J Crit Care 22:177–183
7. Vincent C (2003) Understanding and responding to adverse events. N Engl J Med 348:1051–1056
8. Vincent C, Taylor-Adams S, Chapman EJ, Hewett D, Prior S, Strange P, Tizzard A (2000) How to investigate and analyse clinical incidents: clinical risk unit and association of litigation and risk management protocol. BMJ 320:777–781
9. Auriant I, Reignier J, Pibarot M, Bachat S, Tenaillon A, Raphael J (2002) Critical incidents related to invasive mechanical ventilation in the ICU: preliminary descriptive study. Intensive Care Med 28:452–458
10. Giraud T, Dhainaut JF, Vaxelaire JF, Joseph T, Journois D, Bleichner G, Sollet JP, Chevret S, Monsallier JF (1993) Iatrogenic complications in adult intensive care units: a prospective two-center study. Crit Care Med 21:40–51 11. Tominaga GT, Rudzwick H, Scannell
G, Waxman K (1995) Decreasing unplanned extubations in the surgical intensive care unit. Am J Surg 170:586– 589 discussion 589-590
12. Ursprung R, Gray JE, Edwards WH, Horbar JD, Nickerson J, Plsek P, Shiono PH, Suresh GK, Goldmann DA (2005) Real time patient safety audits: improving safety every day. Qual Saf Health Care 14:284–289
13. Ferraris VA, Propp ME (1992) Outcome in critical care patients: a multivariate study. Criti Care Med 20:967–976
14. Betbese AJ, Perez M, Bak E, Rialp G, Mancebo J (1998) A prospective study of unplanned endotracheal extubation in intensive care unit patients. Crit Care Med 26:1180
15. Epstein SK, Nevins ML, Chung J (2000) Effect of unplanned extubation on outcome of mechanical ventilation. Am J Respir Crit Care Med 161:1912– 1916
16. Pesiri AJ (1994) Two-year study of the prevention of unintentional extubation. Crit Care Nurs Q 17:35–39
17. Wagner AK, Soumerai SB, Zhang F, Ross-Degnan D (2002) Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther 27:299– 309
18. Mion LC, Minnick AF, Leipzig R, Catrambone CD, Johnson ME (2007) Patient-initiated device removal in intensive care units: a national prevalence study. Crit Care Med 35:2714–2720 quiz 2725
