Estudo Prevent

8/12/2019 Estudo Prevent

1/6

Original Article

A prospective multicentre observational study of adverseiatrogenic events and substandard care preceding intensivecare unit admission (PREVENT)*D. A. Garry, 1 S. R. McKechnie,2 D. J. Culliford, 3 M. Ezra,4 P. S. Garry, 4 R. C. Loveland,5

V. V. Sharma, 6 A. P. Walden 7 and L.M. Keating, 8 on behalf of the PREVENT group

1 Specialty Trainee, 2 Consultant, Adult Intensive Care Unit, 4 Specialty Trainee, Nuf eld Department of Anaesthetics, John Radcliffe Hospital, Oxford, UK 3 Medical Statistician, Southampton General Hospital, Southampton, UK 5 Consultant in Anaesthetics and Intensive Care, Intensive Care Unit, Wexham Park Hospital, Slough, UK 6 Consultant Cardiac Anaesthetist, Anaesthesia Department, St George s Hospital, London, UK 7 Consultant in Acute Medicine and Intensive Care, 8 Consultant in Emergency Medicine and Intensive Care,Intensive Care Unit, Royal Berkshire Hospital, Reading, UK

SummaryWe examined the current incidence, type, severity and preventability of iatrogenic events associated with intensive careunit admission in ve hospitals in England. All unplanned adult admissions to intensive care units were prospectively reviewed over a continuous six-week period. In the week before admission, 76/280 patients (27%) experienced 104 iatro-genic events. The majority of iatrogenic events were categorised as medical (37%), drug (17%) or nursing events (17%).Seventy-seven per cent of the events were considered preventable and 80% caused or contributed to admission. Elevenevents were thought to have contributed to a patient s death. The mean (SD) age of patients who had an event wasgreater (63 (21) years) than those who had not (57 (19) years, p = 0.023), and they had a longer median (IQR [range])intensive care stay, 4 (1 8 [0 29]) days vs 3 (1 5 [0 20]) days, respectively, p = 0.043..................................................................................................................................................................

Correspondence to: D. A. Garry Email: [email protected]*Presented in part at the Intensive Care Society State of the Art Meeting, London, December 2011. Accepted: 3 November 2013

IntroductionIatrogenic events result from medical intervention andare not explained by underlying disease [1]. They canseverely harm patients and can lead to intensive careunit (ICU) admission. This prolongs hospital stay,places additional pressure on ICU resources andincreases the cost of hospitalisation [2]. However, thereare few data on the association of adverse iatrogenicevents and ICU admission. The only data on patients

in the UK were collected in 1992 and published in1998 by McQuillan et al. [3]. This study examined themanagement of oxygen therapy, airway, breathing,circulation and monitoring before ICU admission,along with the timeliness of admission, and found thatcare was suboptimal in 54% of the patients. Other stud-ies from the USA and France, published 10 20 yearsago, reported that between 1% and 20% of ICU admis-sions were associated with an iatrogenic event [2, 4 6].

2014 The Association of Anaesthetists of Great Britain and Ireland 137

Anaesthesia 2014, 69, 137142 doi:10.1111/anae.12535


2/6

The purpose of this study was to assess the currentincidence of iatrogenic events associated with ICUadmission in ve UK hospitals, along with their sever-ity and preventability.

MethodsThe Berkshire Research Ethics Committee determinedthat this study was not research according to theNational Research and Ethics Service (NRES) guidance.All participating hospitals registered the study: theRoyal Berkshire; Oxford University Hospital; WexhamPark; Milton Keynes; Lewisham.

We included all unscheduled adult admissions tothe general ICU between 03/05/2011 and 14/06/2011.We excluded transfers between hospitals. All hospitalsused the same data collection form that had beenpiloted in one of them. Investigators were trained touse a modi ed version of the Global Trigger Tool (seeAppendix) to screen notes in a consistent and repro-ducible way [7].

On ICU admission, we reviewed the notes for thepreceding seven days. We examined note and chartentries by medical and nursing personnel, laboratory results and radiology reports. We categorised the causefor ICU admission: for instance, an airway complica-tion in theatre was recorded as an anaesthetic event ,

whereas postoperative pneumonia unrelated to the sur-gical procedure was recorded as a medical event . Apotentially harmful act of commission or omission wasconsidered an iatrogenic event, con rmed by theinvestigators answering no to the question wouldthis standard of care be acceptable for a relative?

We reviewed cases at the end of the six-week period.Four consultant intensivists from three hospitals (LK,SM, AW, RL) reviewed every potential event. Eventswere recorded as iatrogenic if the consultants reached aconsensus decision. When consultants disagreed, the

case was discussed at a subsequent meeting, supportedby additional information from study investigators orfrom the medical notes, informed by expert opinionwhen required (e.g. radiology or vascular surgery). Con-tinued failure to reach consensus led to exclusion of theevent. We graded events by the severity of harm: A, noharm; B, contributed to or resulted in temporary harm;C, contributed to or resulted in temporary harm andprolonged ICU stay; D, contributed to or resulted in

permanent harm; E, contributed to the patient s death.We also recorded whether the event caused or contrib-uted to the ICU admission. We categorised events thatresulted from failure to follow best practice as prevent-able: for instance, repeated review of an unwell patient

with failure to escalate care; or a pulmonary embolus ina patient who did not receive appropriate thrombopro-phylaxis.

We compared subjects who did and did not expe-rience an iatrogenic event with t-tests, Mann Whitney U-tests or chi-squared tests, as appropriate. All testswere two-sided and a p value < 0.05 was consideredsigni cant. We used SAS version 9.1.3 (SAS InstituteInc., Cary, NC, USA) and R (R Foundation for Statisti-cal Computing, Vienna, Austria).

ResultsWe reviewed the notes of 280 patients. We categorisedadmissions as: 151 (54%) medical; 94 (34%) surgical;20 (7%) trauma; 8 (3%) obstetric; 4 (1%) anaesthetic;and 3 (1%) other .

We identi ed 149 potential events in 95/280patients. On review, the panel agreed that 76/280patients had experienced 104 events (Fig. 1). Eighty-three events either directly caused (28) or contributed(55) to ICU admission. Most of the events (77%) were

preventable (Tables 1 and 2). Out of the eventsexcluded by the review committee, ve caused or con-tributed to temporary harm that prolonged ICU stay.None of the excluded events caused or contributed topermanent harm or death.

The mean (SD) age of patients with a con rmedevent was greater (63 (21) years) than those without(57 (19) years, p = 0.023). They also had a longermedian (IQR [range]) ICU stay, 4 (1 8 [0 29]) days vs 3 (1 5 [0 20]) days, respectively, p = 0.043.

Ninety-five patients with 149 events

O ne Two T hre e F ou r F ive Si x

63 16 13 1 1 1

Excluded 19 patients with 45 events

Seventy-six patients with 104 events

O ne Two T hre e F ou r F ive Si x

55 14 7 0 0 0

Figure 1 Flowchart of patients with iatrogenic events(1-6 events per patient)

138 2014 The Association of Anaesthetists of Great Britain and Ireland

Anaesthesia 2014, 69, 137142 Garry et al. | Adverse iatrogenic events preceding ICU admission


3/6

Table 3 illustrates typical examples of events. Ourdata did not show any relation between the rate of iatrogenic events and timing (weekday day, weekday night or weekend), which is in contrast to the recently

published data by Dr Foster Intelligence [8].

DiscussionIn summary, we identi ed 104 events in 76/280patients (27%), higher than most published rates, butlower than 54% reported by McQuillan et al. in theUK 21 years ago [3]. We classi ed 77% of the iatro-genic events as preventable, suggesting that more work needs to be done to improve care outside ICU [9].

The majority of the medical events were delays inmanagement or seeking advice, or inadequate resusci-tation. Ten of the nursing events were misapplicationsof the early warning score, hence perhaps a new approach is required, for instance, the continuousanalysis of telemetrically monitored vital signs, whichautomatically identi es patients at high risk of anacute deterioration [10].

Other studies have reported iatrogenic events in 1.2

19.5% of patients admitted to ICU [2, 4 6]. There couldbe several reasons for these lower gures. These rateswere determined outside the UK and may not re ectcontemporary practice. The de nition of iatrogenesisused in these studies was variable. Our systematic identi-

cation of iatrogenic events, facilitated by a screening tool, may have increased the capture of events: the Glo-bal Trigger Tool may increase the rate of event detectionby up to 10-fold in hospitalised patients [7].

Table 1 Consensus classi cation of 104 iatrogenicevents in 76/280 patients admitted to intensive careunits. Values are numbers.

Event classicationMedical 39Drug 18Nursing 18Surgical 15Infective 5Procedure 4Anaesthetic 5

Outcome classicationNo harm 8Temporary harm without prolonged stay 31Temporary harm with prolonged stay 32Permanent harm 22Death 11

Association of event with ICU admissionCaused 28Contributed 55Not associated 21

PreventableYes 80No 24

Staff associated with eventConsultant 38Staff and associate specialists 3Trainee 60Nurse 16

Drug eventsPharmacist had reviewed drug chart 9Pharmacist had not reviewed drug chart 7

Medical eventsNotes written daily by medical staff 37Notes not written daily by medical staff 3Clear written management plan 27No clear written management plan 13Notes ident ied responsible consultant 35Notes did not identify responsible consultant 5Patient was on the appropriate ward 31Patient was not on the appropriate ward 9

Table 2 Subclassi cation of 104 iatrogenic events. Val-ues are numbers.

MedicalDelay/failure to manage 15

Delay/failure to escalate care/seek advice 11Delay requesting/acting on investigations 6Error reporting investigations 3Inadequate resuscitation 4

DrugPrescribing error 9Inappropriate drug 7Failure to monitor drug levels 1Drug reaction 1

NursingEarly warning score 10Inadequate monitoring 4Failure to follow instructions 2Failure to prevent avoidable fall 2

Surgical

Tissue damage 12Patient positioning 1Equipment failure 1Delay/failure to manage 1

InfectiveHospital-acquired pneumonia 3Urosepsis 1Cellulitis 1

ProcedureTissue damage 4

AnaestheticExcess opioids 2Incorrect use of cell salvage 1Airway incident 2


Garry et al. | Adverse iatrogenic events preceding ICU admission Anaesthesia 2014, 69, 137142


4/6

Our study had several limitations. As with allreviews of patient records, we relied on accurate docu-mentation. There may have been selection bias andinterobserver variability, despite the prospective collec-tion of data guided by the same screening tool. Events

that preceded ICU admission by more than one week would have been missed. The review committee was notblinded to the hospitals where the events happened,which could have introduced bias. The denominator inthis study was unplanned admissions to the ICU : thestudy was not designed to inform about the incidence of iatrogenic events in the hospital.

In summary, we found that preventable events thatcontributed to ICU admission, patient harm and death

were common. The high number of events suggests aneed for further studies on a national scale, looking for preventable events in all critically unwell patientsin hospital, including those who are not admitted tothe ICU. We hope that clinicians will use our results

to reduce avoidable critical illness.

AcknowledgementsThe PREVENT group contributors were Drs J. Kim, R.Kulanthaivelu, R. Martynoga, R. Morris-Smith, L.O Donohoe, J. Smythe, R. Stewart, N. Tam, T. Thomas,J. Thompson, A. Thomson and H. Tyler. Data analysisby a medical statistician was paid by the Royal BerkshireICU Research Fund. No competing interests declared.

Table 3 Examples of iatrogenic events.

Event classication Description Severity Preventable

MedicalDelay/failure to manage Prophylactic heparin not prescribed

in patient with suspected cancerD: cardiac arrest from PE Yes

Delay/failure to escalate Decision made to refer wardpatient to ICU A: 4-h delay Yes

Inadequate resuscitation ED admission, shocked, acuteabdomen, rapid deteriorationdespite resuscitation. Patienttransferred to CT

D: cardiac arrest in CT, emergencylaparotomy for perforatedappendix

Yes

Delay/failure to escalatecare/seek advice

72-h hypotension afterhemiarthroplasty

E: cardiac arrest, died 24 h later Yes

DrugInappropriate drug Propofol infused for 1 week,

maximum dose exceededD: rhabdomyolysis, acute kidney

injury and right heart failureYes

NursingEarly warning score Prolonged failure to act on early

warning scoresC: respiratory arrest Yes

Calculated incorrectly A: identied by rapid response

team

Yes

SurgicalTissue damage Hepatic artery rupture during

cholecystectomyC: massive transfusion,

ICU admissionNo

Elective caesarean section in patientwith broids

D: incorrect organs resected Yes

InfectiveCellulitis Insertion data and VIP score not

recordedB: hand cellulitis Yes

ProcedureTissue damage Routine colonoscopy D: splenic rupture Yes

AnaestheticAirway incident Delay securing airway after out-of-

hospital cardiac arrestC: aspiration Yes

Intra-operative airway obstruction B: pulmonary oedema,ICU admission

Yes

A, no harm caused to the patient; B, contributed to or resulted in temporary harm to the patient; C, contributed to or resulted intemporary harm to the patient and required prolonged ICU stay; D, contributed to or resulted in permanent patient harm; E, con-tributed to patient s death; PE, pulmonary embolism; ICU, intensive care unit; ED, emergency department; CT, computed tomo-graphy; VIP, visual infusion phlebitis.




5/6

References1. Brennan TA, Leape LL, Laird NM, et al. Incidence of adverse

events and negligence in hospitalized patients. Results of theHarvard Medical Practice Study I. New England Journal of Medicine 1991; 324 : 370 6.

2. Mercier E, Giraudeau B, Ginies G, Perrotin D, Dequin PF. Iatro-genic events contributing to ICU admission: a prospectivestudy. Intensive Care Medicine 2010; 36 : 1033 7.

3. McQuillan P, Pilkington S, Allan A, et al. Condential inquiryinto quality of care before admission to intensive care. BritishMedical Journal 1998; 316 : 1853 8.

4. Trunet P, Le Gall JR, Lhoste F, et al. The role of iatrogenic dis-ease in admissions to intensive care. Journal of the AmericanMedical Association 1980; 244 : 2617 20.

5. Darchy B, Le Miere E, Figueredo B, Bavoux E, Domart Y. Iatro-genic diseases as a reason for admission to the intensive careunit: incidence, causes, and consequences. Archives of Inter-nal Medicine 1999; 159 : 71 8.

6. Lehmann LS, Puopolo AL, Shaykevich S, Brennan TA. Iatrogenicevents resulting in intensive care admission: frequency, cause,and disclosure to patients and institutions. American Journal of Medicine 2005; 118 : 409 13.

7. Classen DC, Resar R, Grifn F, et al. Global Trigger Toolshows that adverse events in hospitals may be ten timesgreater than previously measured. Health Affairs 2011; 30 :

581 9.8. Stafng gures V1.7. http://www.drfosterhealth.co.uk

(accessed 12/11/2012).9. Caring to the End? A review of the care of patients who died

in hospital within four days of admission. A report by theNational Condential Enquiry into Patient Outcome and Death, 2009 . London: NCEPOD, 2009.

10. Pimentel MA, Clifton DA, Clifton L, Watkinson PJ, TarassenkoL. Modelling physiological deterioration in post-operativepatient vital-sign data. Medical and Biological Engineeringand Computing. 2013; 51 : 869 77.

Appendix

The modied Global Trigger Tool

Please check the following, for the seven-day period before the ICU admission

Domain Action

Laboratory resultsHb any sudden drop Check for cause (procedure, surgery etc)INR > 5 Check for cause

Check for evidence of bleeding

Na < 120 or > 160, K < 2.5 or > 6.5 mmol.l 1 Check for causeCheck for effect

Doubling of baseline urea/creatinine Check for cause (nephrotoxic drugs etc)Were drug doses adjusted accordingly?

Glucose < 3.0 mmol.l 1 Check for causeCheck for appropriate action

Raised troponin Check for cause (? omitted b -blocker )Positive blood culture Invasive procedures/lines/cathetersC. difcile Inappropriate antimicrobials

RadiologyCheck CXR If pneumothorax, check for procedures (CVP line etc)Check CT scans If NG tube, check for correct placementCTPA If PE, check that patient had adequate thromboprophylaxis

Drug chartAny omission in receiving medications ?If patient admitted to hospital for electivesurgery, check that appropriate chronic medicationswere taken on the day (B-blockers)

Medicine for a chronic condition (e.g. b -blocker)Anti-infectiveAnticoagulantInsulinResuscitation medicine

Any delay in receiving medications? Anti-infectives (within 6 h of diagnosis)AnticoagulantInsulinResuscitation medicine


Garry et al. | Adverse iatrogenic events preceding ICU admission Anaesthesia 2014, 69, 137142


6/6

Please check the following, for the seven-day period before the ICU admission

Domain Action

Any antagonists given? NaloxoneFlumazenilVitamin K

50% dextroseAny sedatives given? Did this precede an event (fall, respiratory arrest etc)Any mistake in drug doses?Any drugs prescribed that the patient has

a documented allergy to?Check for drug reaction

Any transfusion of blood products? Why?

Observation chartsAny gap in recording of observations? Check for lack of referral/delay in reviewWas the early warning score recorded?Was it calculated correctly?Did the EWS trigger appropriately?

Anaesthetic chartWere all non-anaesthetic drugs given in

theatre also recorded on the drug chart?Check for duplication in administration

Any unexpected blood transfusions?Any incidents recorded on the anaesthetic

chart? (give details)

Surgical chartAny removal/repair of an organ? UnplannedIs it a return to theatre? Check previous surgical charts

Medical notesWas there a cardiac arrest? Had patient been adequately monitored

Had patient been adequately reviewed?If within 24 h of surgery/anaesthetic, check relevant charts

Was there a delay? Patient review (nursing, junior doctor, consultant, outreach)If new admission, time to rst consultant review (within 12 h?)DiagnosisEssential investigationsCommencing treatmentReferralICU admission

Has the patient had a fall? Check for sedativesHad the patient been risk-assessed for falls?



Documents

Estudo Prevent