Making it safe for acutely ill patients - a whistlestop tour of medical error & patient harm Sara Barton Acute Physician Salford Royal NHS Foundation Trust
What is medical error? Medical errors can be defined as the failure of a planned action to be completed as intended or the use of a wrong plan to achieve an aim. (IOM To Err is Human Nov 1999) Mortality Morbidity Financial Reputational second victim
Error vs Harm Not all error leads to harm Not all harm includes an error
Hospital associated harm Unintended, physical injury resulting from or contributed to by medical care that requires additional monitoring, treatment or hospitalization, or that results in death. Category E: Temporary harm to the patient and required intervention Category F: Temporary harm to the patient and required initial or prolonged hospitalization Category G: Permanent patient harm Category H: Intervention required to sustain life Category I: Patient death IHI Global Trigger Tool
Types of error Diagnostic i. Error or delay in diagnosis ii. Failure to employ indicated tests iii. Use of outmoded tests or therapy iv. Failure to act on results of monitoring or testing Treatment i. Error in the performance of an operation, procedure, or test ii. Error in administering the treatment iii. Error in the dose or method of using a drug iv. Avoidable delay in treatment or in responding to an abnormal test v. Inappropriate (not indicated) care Preventive i. Failure to provide prophylactic treatment ii. Inadequate monitoring or follow-up of treatment Other i. Failure of communication ii. Equipment failure iii. Other system failure
Reducing error to reduce harm Multiple errors occur every day Not all errors reach the patient Some errors more likely than others Difficult to predict in a complex environment How do you know where to start?
How Safe is Healthcare? Health Care
Where Should We Be? Health Care Blood Transfusion Anesthesia
Case 1 - Sepsis First weekend in august Elderly woman presents to ED with urinary sepsis Seen by ED doctor within 30 minutes dose of antibiotics prescribed Seen by EAU consultant & junior within 80 minutes (still in ED) agreed sepsis Consultant asked junior to prescribe meds
Case 1 - sepsis Patient transferred to EAU (9pm) EAU nurse did not check medicines chart Patient transferred to post acute ward late due to bed pressures (11pm) Ward nurse noted no regular antibiotics prescribed Put on routine doctors job list for review as patient EWS ok
Case 1 - sepsis Junior doctor picks up task next day ensures antibiotics dosed and given (but 24 hours since first dose) Patient reviewed EWS 1 & seems ok Patient deteriorated acutely 6 hours later Cardiac arrest unsuccessful CPR
Case 1 - Sepsis August & weekend precondition to unsafe acts Atypical method of prescribing in ED precondition to unsafe act, unsafe supervision Seen by EAU consultant & junior together - organisational influences, precondition to unsafe act Consultant asked junior to prescribe meds (antibiotics omitted) unsafe act BUT unsafe supervision, organisational influences,
Case 1 - sepsis EAU nurse didn t check meds chart unsafe act organisational, precondition Patient transferred to post acute ward 2 hours later for bed pressures (11pm) organisational, preconditional Ward team noted failure, didn t act unsafe act - preconditions Put on routine doctors job list unsafe act, BUT precondition, organisational
Case 1 - sepsis 4 active failures omission, prescription, communication, escalation Multiple latent failures Series of reliability issues
Reliability Measured as the inverse of the system s failure rate Failure free operation over time Chaotic: failure in greater than 20% of events 10-1 : 1 or 2 failures out of 10 10-2 : <5 failures per 100 10-3 : <5 failures per 1000 10-4 : <5 failures per 10000 Chapter 4 Reliability & Resilience Roger Resar & Frank Federico The Essential Guide for Patient Safety Officers, June 13
Doing your best, with the best possible resources achieves reliability around 80% of the time This includes: Standard order sheets, multiple choice protocols written policies/procedures Personal check lists Feedback of compliance Working harder Awareness & training Posters in the loo
Reliability First process: clerking document = 80% reliable Second process: PTWR check = 80% reliable Second process should pick up 80% of slips in first process (i.e. 16%) Overall reliability = 96%
Step 1 simplify & standardise Step 2 controls to prevent error Step 3 errors that slip through studied & redesigned if needed. Reliability
Back to Case 1 Standardise: i. antibiotic prescribing across ED & EAU ii. ED EAU handover, EAU to ward handover iii. sepsis education for all clinical areas iv. communication & escalation Control: i. check of drugs chart at each handoff ii. checklist for identifying patients to move by EAU coordinator iii. Second level check for prescribing PTWR vs Meds rec vs Board round
Case 2 elective vs non-elective 22 year old patient presented with viral illness and headache PTWR recommended lumbar puncture to investigate? Viral meningitis LP done patient feeling much better On call junior doctor saw at about 7pm results not available Sent patient home we will write with results Patient returned next day with seizure
Elective vs non elective Junior doctor working on Neuro ward different population Essentially unsafe act No controls in system to prevent it However LP results were normal Independent medical opinion viral illness with lowered seizure threshold not preventable
Actions Standardise: 1 standard for documentation of consultant review Clear documentation of discharge/ transfer plans Control: Discharge by registrar or consultant input only No acute patients discharged without full LP results
Case 3 communication Patient with acute on chronic renal failure K + result 6.2 HAN allocated to on call FY1 Phoned registrar for advice Reg advised 15 units actrapid FY1 heard 50 units actrapid Nurse thought big dose but reassured that reg advised it Patient suffered hypoglycaemic fit (no long term sequelae)
Case 3 communication Active failure lack of knowledge Predisposition to unsafe act no available policy Latent failures unsafe supervision (did reg need to see patient or review notes) Organisational nurse knew dose wrong but felt falsely reassured Human Factors Just a routine operation: http://www.youtube.com/watch?v=jzlvgtpiof4
Case 3 communication Simplify: Written guidelines for hyperkalaemia Made guidelines findable on intranet Control: Work with nurses around what constitutes a large dose (not just for insulin) Worked to provide psychological safety for check and challenge Junior doctor shared story for medicines safety work
Case 4 unexplained drowsiness 34 year old prisoner 2 officers in situ Presented Tuesday with haematemesis All medicines prescribed as per prison chart OGD Wednesday gastritis Reviewed Thursday unexpectedly drowsy Patient admitted to concealing & selling some of his diazepam in prison diazepam stopped Friday less drowsy not fully recovered Discharged to prison hospital ward Saturday readmitted large SDH
A pattern of deviation in judgement whereby inferences may be drawn in an illogical pattern Confirmation bias seeking information to confirm a conclusion Anchoring bias the act of relying too heavily to one piece of information Cognitive bias We all have cognitive bias & need to be aware of them
Bias 52 year old Polish man, poor spoken english Complains of unsteadiness, weakness & blurred vision Haemoglobin 52 Previous admission 6/12 ago with bleeding duodenal ulcer GI Bleed (alcohol?) Waldenstroms Macroglobulinaemia retinal haemorrhages, bone marrow failure
Bias 73 year old lady, acute episodic confusion 3 rd admission in 3 months Smells a bit.. urine dip pro and leuk?delirium & UTI Liver Abscess None of her previous urine samples showed any growth
Cognitive Bias Pattern recognition helps in medicine Acute medicine relies on balancing risks Personal/ reflective learning Discussion and challenge Structured approach to diagnosis Reassess diagnosis if ongoing issues Second opinion
A Perfect EAU Simple pathways with minimal variation Accessible information and protocols Standardisation of documentation & practices Well trained and supervised staff Early senior review Built in checks, with scheduling & role allocation Culture of openness and check/challenge Minimal handoffs Long term stable work force
Sound familiar?
Recommendations Learn (& teach) about reliability science, human factors & quality improvement Consider predictive tools when planning changes e.g. failure mode event analysis Embrace systems thinking, lose blame Reports incidents, learn from them, use the whole team. Consider mortality & morbidity reviews, global trigger tool etc to identify patterns
In our defence. 2500 patients a month 55 beds 85% of medical take managed wholly in EAU EAU are high incident reporters & investigators Considered Outstanding by CQC High scores for patient and staff satisfaction
Questions? Sara.barton@srft.nhs.uk