- Email: [email protected]
Risk management in obstetrics
Review
Risk management in obstetrics
Raising awareness at all levels of the workforce is essential, from senior consultants to trainees, and from midwifery managers to ancillary staff (including cleaners and porters). A common misunderstanding is that risk management is the concern of service managers, not clinicians and other staff. Clinicians play an invaluable role in all aspects of risk management.
Charlotte Cottee Kate Harding
Application Increasing emphasis has been placed on risk management and clinical governance, with clinicians’ knowledge of these subjects being assessed in job interviews and included in postgraduate training and examinations. The application of risk management is seen at all levels within an NHS hospital trust. • Senior management may be concerned with the local strategy for infection control in a department, and its possible effect on unit closure if an epidemic (e.g. avian flu) arises. • A multidisciplinary team may care for a pregnant, unsupported teenager with social problems; assessing and treating the risks raised during the course of her care. • A clinician will identify a clinical risk and plan management to minimize harm. For example, she may organize an ultrasound scan to identify a placenta accreta in a woman with a placenta praevia and two previous caesarean sections. She will then organize suitable support from interventional radiologists at the time of delivery (possibly including transferring care to a tertiary centre) to prevent massive post-partum haemorrhage. After the event, there may be a meeting to reflect on care, and consider changes to guidelines to maximize future care to ensure lessons from the case are learnt.
“Those who do not remember the past are condemned to repeat it.” George Santayana (1863–1952)
Abstract Risk management is a vital tool in preventing repetition of errors which are costly to patients, staff and institutions. If risk management is used appropriately, faults in the system can be eliminated before they cause an event, or to prevent event recurrence. With detailed analysis of incidents (using tools such as root cause analysis) the fault in the system (rather that the individual) can be identified and rectified. The evolution of risk management (and its development from industry to the health sector) in the UK is described. The cost to the health service is explored and methods of analysing risk are discussed (including risk registers). There is a description of the elements that are reviewed in a Clinical Negligence Scheme for Trusts assessment. There is also a worked example of a clinical incident, its analysis and the solutions recommended.
Keywords audit; clinical governance; CNST; NPSA; risk management; root cause analysis; training
How big is the problem? Patients and staff The effect of risk on patients can range from minor effects to severe disability or even death. There are also the unquantifiable effects such as the psychological impact and loss of faith and trust in the healthcare system. Such effects may be longlasting and influence how the individual uses the healthcare system in the future. Healthcare professionals are also affected by risk. Staff can become: • disillusioned with their careers • demotivated in the workplace • demoralized and at higher risk of making an error; or more likely to leave the health service to find other jobs with less exposure to risk, leading to recruitment problems and understaffing. Morale is often severely affected and the effect on personal reputation should not be underestimated. More than one-third of doctors who are sued suffer from clinical depression.
Definition Risk management is the systematic identification, assessment and evaluation of risk. It can improve the quality of care, and develops safe systems of care that minimize adverse events for patients and staff. Risk management can be reactive (e.g. in response to a serious incident or a complaint), pro-active (e.g. establishment of a risk register or an assessment of national guidelines/reports) or preventative (e.g. ensuring adequate training and staffing levels). Healthcare professionals should understand the principles of risk management. Risk management is not only a tool to reduce litigation or a process to report incidents, it minimizes the harm that errors can cause to patients and staff. Risk management encompasses many aspects of clinical governance, from risk reporting (and response to complaints) to audit, guidelines, risk assessments (risk registers) and training.
Financial The problem of risk in obstetrics from a financial perspective is huge. Data from the NHS Litigation Authority (NHSLA) shows that, between 1995 and March 2007, obstetrics and gynaecology accounted for 21% of claims (8,532) and this was valued at 49% of the total amount claimed (£2,475 million). In total, 41% of claims are abandoned, 41% are settled out of court, and of those that go to court, ∼29% are settled in favour of the
Charlotte Cottee MRCOG is a Specialist Registrar in Obstetrics and Gynaecology at Guy’s and St Thomas’ Hospital Foundation Trust, London, UK. Kate Harding FRCOG is a Consultant Obstetrician at Guy’s and St Thomas’ Hospital Foundation Trust, London, UK.
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
155
© 2008 Elsevier Ltd. All rights reserved.
Review
claimant. This money (including legal fees) comes out of the NHS budget. About 10.8% of UK hospital patients experience an adverse event, of which 1% can lead to severe harm or death. The number of events appears to be rising (though that may be due to increased awareness), with <1,000 per quarter reported to the National Patient Safety Authority (NPSA) in 2003, and >200,000 in 2007 as the percentage of trusts reporting increased from <5% to >60%. Of these incidents, ∼35% are due to patient accidents (e.g. falling out of bed, tripping in hospital) and 60% are due to clinical events (medication, treatments, communication).
scheme and maternity (as the area of greatest financial risk) was assessed separately from the rest of the acute trust. The CNST rating reflects the breadth of risk management within the organization, and the extent to which the principles of risk management are embedded. Maternity CNST assessment level 1 (which gives a trust a 10% reduction in its annual premium) has been achieved by all maternity units in England and Wales. (Table 2). Level 3 (the highest level; it gives a 30%
Components of CNST assessment Organization Risk strategy Lead Obstetrician for labour ward Risk Manager
Evolution of risk management Several reports have been instrumental in developing risk management over the last decade. In 2001, “An organisation with a memory” highlighted the need to learn from clinical error. The NPSA was established in 2001, with a remit to develop a national approach towards reporting incidents and learning from them. This included development of a national database for: • reporting patient safety incidents • distribution of safety alerts • formulating solutions to identified risks • Internet training for root cause analysis. NPSA analysis concentrates on a systems approach rather than focusing on individuals. A document by the Department of Health entitled “Standards for better health”, published in 2004, discussed improving the quality of care by assessing it with respect to seven domains (Table 1), the first being safety. It was recognized that patients can suffer harm as a result of care that was intended to help them. Measures and strategies to reduce this happening can be implemented, and risk management is one approach.
Learning from experience Assessment on the relevance of national reports (e.g. CEMACH, NCPOD, CESDI) Assessment of attainment of recommendations of NICE guidelines Structure of incident reporting Action plans from adverse incidents and complaints Communications
Getting results into the notes Auditing action on abnormal results Handover of care between professionals Achieving an acceptable decision-to-delivery interval
Clinical care
Guidelines Audit of guidelines Action on audit recommendations
Induction, training and competence Induction of all staff Mandatory training for all staff CTG Resuscitation Drills and skills Technical competence Equipment Recovery area
Clinical Negligence Scheme for Trusts (CNST) After crown indemnity was removed (whereby the NHS centrally funded all claims), it became clear that in some financial years, some trusts (and in turn their patient population) may suffer severe financial embarrassment if there were several high-paying claims (often obstetric in origin). The CNST was set up to allow trusts to pool their claims/risks, and with each trust paying a proportion of its turnover into a common pot, high and low claims were balanced out so no single trust would suffer if they had a number of huge claims (e.g. for birth hypoxia) in one year. A trust had to show a basic level of risk management to enter the
Record keeping
Storage of CTGs Documentation audit Improvement in documentation after audit
Implementation of risk management recommendations Evidence of action plans from incidents and complaints Evidence of improvement after recommendations Up-to-date risk register Involvement of trust board involvement in risk management
Health care domains Safety Clinical and cost effectiveness Governance Patient focus Accessible and responsive care Care environment and facilities Public health
Staffing
Adequate staffing levels of all staff groups Evidence of staffing levels Midwifery supervision Actions on how to improve poor staffing Table 2
Table 1
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
156
© 2008 Elsevier Ltd. All rights reserved.
Review
is looked at by risk analysis and evaluation. When considering minimising the chances of risk happening or mitigating damage from such risk, risk treatment is required. The cost of prevention is balanced against the cost of getting it wrong. Learning from what wrong is a very important element of the process and this is encapsulated in risk control and sharing and learning from the experience Figure 1.
Identify risks
Analyse risks Communicate
Monitor
Risk identification Evaluate risks
Risk is usually identified from incident reporting (using ‘trigger lists’ (see Table 4 or unusual events), complaints, or internal investigations of major adverse incidents. These three sources are familiar to those working in clinical areas, and may be grouped as internal sources. Proactive assessment of new care pathways or guidelines should identify risks and seek to eliminate them (e.g. guidelines should be reviewed by a multidisciplinary team to avoid errors). Risk assessment in each clinical area is a trust requirement and is the responsibility of senior members of staff. The risk register is reported back to the trust clinical governance board and onto the trust board (Table 3). Complaints and claims also used to identify risk, and should feed into the risk register. The Patient Advice and Liaison service (PALS) gives impartial support to patients and relatives to voice their concerns about any aspect of care, away from the clinical area, and to help them express a complaint if that is required. Clinical audit can identify and confirm/refute risk prospectively and retrospectively. Direct consultation with staff can identify risk by surveys, interviews and departmental workshops. Risks can also be identified from external sources, such as: • national reports (e.g. CEMACH) • national guidelines (e.g. NICE intrapartum guidelines) • Government initiatives (e.g. Cleaner Hands campaign) • research (e.g. the finding that corticosteroids reduce neonatal ARDS).
Treat risks
Eliminate
Reduce
Finance
Figure 1 The risk management process.
r eduction in premium) has been achieved by 19 units, ranging from small units to the largest units in the UK. By giving trusts a financial incentive to reduce risk, it is hoped that patient safety has been improved because it ensures that at least basic risk management procedures are in place in all maternity units.
Risk management process The risk management process can be viewed as a systematic approach to dealing with a sequence of questions that aim to reduce risk. See box C. For example when looking at what could go wrong? Risk identification is required. Understanding the chances of it going wrong again and what would be the impact,
Risk register Ref
Source
Date
Description
Lead
Rating
Action
Completion
1
Trust board
1 Jan 2004
MRSA infection
Red
2
Incident
5 March 2004
Management of PPH
Head of Infection Control Head of Obstetrics
75% staff Dec 2004 Four opened Sept 2004 Nil available Purchased Nov 2004 90% trained January 2005 Completed Feb 2005
3
CEMACH report
December 2004
Puerperal suicide
Clinical Director Women’s Health
Green
Clean hands campaign Increased isolation rooms Invest in rapid assay Invest in rapid blood transfusor, Haemacue® and a TEG Mandatory training New PPH protocol Set up perinatal psychiatry team Improve links with current psychiatric teams Midwife training in booking questions
Orange
Currently no funding or staff Monthly meetings since December 2005 Introduced to new notes January 2006
Risks are described in full in the referenced appendixes. Rating depends on level of risk and probability; therefore suicide, though catastrophic, is rare and gets a low rating. MRSA is a common problem and causes significant embarrassment to the trust, so it gets a high rating.
Table 3
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
157
© 2008 Elsevier Ltd. All rights reserved.
Review
Suggested trigger list for reporting incidents in maternity units Maternal incident
Fetal/neonatal incident
Organizational incident
Maternal death Undiagnosed breech Shoulder dystocia Blood loss >1500 ml Return to theatre Eclampsia Hysterectomy/laparotomy Anaesthetic complications ITU admission Venous thromboembolism Pulmonary embolism Third/fourth-degree tears Unsuccessful forceps or ventouse Uterine rupture Readmission of mother
Stillbirth >500 g Neonatal death Apgar score <7 at 5 minutes Birth trauma Fetal laceration at caesarean section Cord pH <7.05 arterial or <7.1 venous Neonatal seizures Term baby admitted to neonatal unit Undiagnosed fetal anomaly
Unavailability of health record Delay in responding to call for assistance Unplanned home birth Faulty equipment Conflict over case management Potential service user complaint Medication error Retained swab or instrument Hospital-acquired infection Violation of local protocol
Table 4
Incident reporting system
Risk register
Clinical areas should have an anonymous system to report risk incidents. Trigger lists categorize incidents into maternal, fetal and organizational. These lists should be easily accessible in the clinical areas, and healthcare professionals should be aware of their contents. A trigger list compiled by the Royal College of Obstetricians and Gynaecologists is shown in Table 4. A multiprofessional risk management team reviews the reports and assesses if a detailed analysis is required. In a simple case, the incident will be included in a risk log; a more complex case may require a review of the notes and a discussion with staff members, and simple training measures are often all that is required. The most complex cases require thorough investigation and detailed action plan to be reviewed by the clinical directorate (and occasionally the hospital trust) to ensure that systems errors are not repeated and relevant staff learn from the event (e.g. setting up a new service or employing extra staff).
A risk register is a centrally held ‘live’ document that should contain all the risks identified by a maternity unit and the trust. The register can be populated from incidents, complaints, national alerts and proactively looking for problems. These items should be graded according to how high the effect of the risk is (financial, media exposure, patient harm) and the likelihood of them occurring. The death of a neonate from MRSA is of moderate cost, high exposure in terms of bad publicity, but relatively rare. Admission to ICU due to post partum haemorrhage is likely to occur 2–6 times a year, but is not financially costly (if well managed; Table 5). To calculate the level of risk, the likelihood of recurrence is estimated, as is incident severity (including the effect on the trust financially and in terms of public confidence). Using the matrix, a grading from green to red is allocated; in general, trusts are likely to allocate funding to all red risks. An example of a risk register is given in Table 3.
Risk matrix to assess risk severity Consequence
Insignificant
Minor
Moderate
1
2
3
Probability of recurrence
Major 4
Catastrophic 5
Certain Highly likely Likely Possible Rare
Table 5
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
158
© 2008 Elsevier Ltd. All rights reserved.
Review
Analysis and evaluation of risk Identify incident and decision to investigate Reporting an incident is the responsibility of all members of the team involved. Not all incidents will need a full systems investigation and the severity of the incident often dictates the timescale in which it needs to be investigated. The risk management team will make these decisions and they are fed back to all involved with the incident.
“What we call experience is often a dreadful list of ghastly mistakes.” J Chalmers da Costa (1863–1933). People do not intend to commit errors. The ethos and training of medicine is based on primum non nocere (firstly, do no harm). It is not productive to focus only on the individual who was involved in an incident; the entire system must be analysed. This is known as a ‘systems view’ as opposed to a ‘personcentred view’. Understanding medical error and particularly the human factor components are vital when considering what went wrong. An error cannot be analysed in isolation, and associated factors must be evaluated: • conditions in which it occurred • frequency of the activity • degree of uncertainty of clinical practice in which it took place • multiple actions leading up to the error. The different types of error must be understood. Human errors can be divided into intentional and unintentional (see Scholfield, Further reading). A systems approach involves viewing the incident in terms of failures that are active or latent. Active failures are the immediate causes of an incident. They are made by staff working in the clinical area. An example is the midwife that gives syntometrine instead of syntocinon for active management of the third stage in a woman with pre-eclampsia. Latent failures are more insidious. They are a result of actions made by the organization, often at many levels. These actions create the overall conditions in which the incidents occur, and may increase their chance of happening. In a busy understaffed unit, staff may be more likely to pick up the wrong drug or not have time to fully assess the patient (a latent failure of the system rather than an active failure of the individual). To investigate error and reduce the chance of it recurring, the conditions in which the error occurred must be analysed. Scholefield states that the commonest condition leading to error is unfamiliarity with a task, followed by shortage of time, and poor human equipment interface. These conditions can be grouped in terms of influencing factors (Table 6).
Form investigating team The team should represent all areas of the system to be investigated and should not include those who have been directly involved with the incident. For example in the case of a serious labour ward incident the team could comprise: a senior midwife, the lead Consultant for Labour Ward Data gathering Information is best collected as soon as possible after the incident occurs. Necessary documentation includes: a copy of the incident form, medical notes and detailed statements from those directly involved with the incident. Documentation related to the incident such as copies of departmental protocols and information about relevant conditions affecting the event e.g. duty rotas and reports on the level of activity in the unit at the time, are very useful. Staff interviews may be necessary and provide a good source of additional information about an incident when conducted in a structured manner by an experienced member of staff. Determine chronology of the incident A detailed chronology of the incident can be constructed using various techniques such as timelines, flow charts, time person grids and narratives. Identify care delivery problems Care delivery problems (CDPs) are often easily identified from records and interviews. It is important to recognise each problem as a specific action or omission rather than including general observations as these can lead to inclusion of contributory factors. Identify contributory factors Contributory factors are addressed for each identified CPD. Each CPD may be associated with several contributory factors. A fishbone diagram is a good way of recording contributory factors. The CDP represents the fish’s head and the contributory factors are represented by the fish bones. Another method is to tabulate the chronology of the incident and map the CDPs and contributory factors to this table.
Incident analysis
Make recommendations and deliver action plans After the above analysis of the incident a set of recommendations and improvement strategies can be formed. The contributory factors are prioritised in order of their impact on future safe healthcare delivery and action plans are made for each contributory factor. At this stage it is vital to recognise that action plans can be divided into those at a local level and those at a directorate or organisational level. To improve the implementation of recommendations it is important to identify who is responsible for implementing the actions, identify the timeframe for the implementation to be completed and a mechanism by which evidence of completion is checked. Ideally the effectiveness of the plan implemented needs to be evaluated and audit is a good tool for this.
After an incident has occurred, reproducible systems-approached analysis can be employed to investigate the event. The aim is not
Influencing factors Patient factors Task factors Individual factors Team factors Work environment Organizational and managerial factors
Figure 2 The London protocol.
Table 6
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
159
© 2008 Elsevier Ltd. All rights reserved.
Review
Organisation and management culture
Contributory factors influencing practice
Care delivery problems
Work/environment factors
Defences/barriers
Unsafe acts
Team factors Management decisions and organisational processes
Errors
Individual (staff) factors
Incident Task factors Violations Patient factors
Latent failures
Active failures
Error and violation producing conditions
Figure 3 Adapted version of Reason’s model of organisational accident.
solely to establish what went wrong, but to identify contributory factors and pathways that may be altered to minimize the chance of event reoccurrence and maximize delivery of safe healthcare. Various approaches have been used to achieve this analysis, such as root cause analysis (RCA). RCA originated in industry and uses a series of tools to investigate and analyse incidents. It has been well adapted to the UK healthcare system, and is used as a structured investigation technique for adverse outcome analysis. RCA has been criticized as being too focused on looking at the incident and what caused it, without concentrating on the chain of events and contributory factors, thereby failing to recognize factors that may have high error-producing potential. The London Protocol (Figure 2) has been devised by Taylor-Adams and Vincent to overcome this failing. It advocates a broad examination of all aspects of the healthcare system in which the incident occurred, incorporating investigation, analysis and recommendations. It is particularly focused on members of the healthcare team, from managers to those working in clinical areas, their communication and teamwork, and how they interact to create a safe environment. The London protocol approach was based on Reason’s model of organizational accident (Figure 3). This illustrates how fallible decisions at organizational levels are transmitted down departmental pathways to clinical areas, creating task and environmental factors that lead to errors. Defence barriers are in place to protect against human or equipment errors (e.g. training and checking) but an incident may occur if these are permeated at multiple levels. This is also known as the ‘Swiss cheese effect’. Figure 4 shows an example of management of clinical risk.
carrying it out. We have moved from the concept of ‘see one, do one, teach one’ where individuals were not formally trained and their competence was not formally assessed. It was possible for a doctor to be expected to do a breech vaginal delivery despite having never seen one (or practised on a simulator). In the twenty-first century, middle-grade obstetricians are expected to have attended a skills course and to have had hands-on training for skills such as managing shoulder dystocia, breech delivery and postpartum haemorrhage. These skills can be assessed and competence tested before the trainee can pass on to the next training level. Junior doctors move between hospitals regularly. Induction to their new workplace ensures that they know: • how to use the emergency call system • the protocols (and how to access them) • who to call for help and how • the layout of their work place, and location of emergency trolleys and theatres. This induction process applies to all staff, including new midwives and nurses, new consultants, and locum and agency staff. Without an adequate induction process, the time of new staff joining the team is one of increased risk. Induction and basic skills training are important parts of the CNST assessment, with emphasis on ensuring that >90% of staff have annual skills training (including CTG analysis, maternal and neonatal resuscitation).
Conclusions Risk management is an approach to clinical care that actively searches for weaknesses in the system and attempts to correct them before harm comes to patients or staff. Risk management is often reactive rather that proactive, but at least this prevents errors being repeated. Done well, risk management involves all members of staff, is seen as a learning process rather than a
Training, induction and competence A prerequisite of good medical care is that the clinician is trained in the procedure he is undertaking and competent at
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
160
© 2008 Elsevier Ltd. All rights reserved.
Review
Risk management: case study A term pregnant female presented to labour ward on a particularly busy evening. She appeared to be in the early stages of labour and had been aware of reduced fetal movements all day. The unit did not have vacant rooms, so she was put in the waiting area. The attending midwife explained the situation and informed the woman that she would be moved if a room became available. The midwife did not take a full history or examine the patient because she was in the waiting room. Two hours later, the woman was moved into a room; difficulty was experienced with fetal heart auscultation. Ultrasound confirmed intrauterine death. The care delivery problem is delayed diagnosis of intrauterine death. There are several contributory factors, and are depicted in a fishbone analysis diagram. Fishbone diagram depicting contributory factors for a case of intrauterine death Environment
Individual
Patient
No es en te d es in el gu
In
La
su
ck
ffi
of
cie
nt
Bu
id
sy
ro
un
om
it
s
te
e bl
t en
la
m
ila
ss
va
se
ya
as
ac
or
Pr
Po
iv
pr Resources
Intrauterine death
Working conditions
Organisational and management
Making recommendations, developing action plans and auditing outcome The fundamental risk management issue is that the fetal heart was not auscultated on arrival. The possibility for intervention was missed. Each contributory factor can be addressed and action plans formed, as shown below. Contributory factor
Action
Late presentation Poor assessment of patient on arrival due to lack of privacy
Educate patients about self-referral system to triage, and opening times Extra triage rooms provide privacy for historytaking and examination
Lack of rooms
Triage suite
Level of activity within the unit
Escalation policy: guideline detailing actions to be taken if beds are not available Guideline detailing timeframe from arrival to full heart auscultation
Delay in auscultating fetal heart
The major action was to establish a triage facility to enable women to be seen soon after arrival in a private area to assess their clinical needs. The development of the triage facility was achieved by a multidisciplinary team that included representatives from the resource directorates, organizational and management teams, through to obstetricians, senior midwives and administrative staff. An audit six months after establishment of the triage showed that all women were cared for according to the timeframes stated in the guidelines, and that the average time from arrival to assessment was under five minutes. Figure 4
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
161
© 2008 Elsevier Ltd. All rights reserved.
Review
means of looking for fault, and improves outcomes for everyone involved in health care at every level. ◆
Scholfield H. Risk management in obstetrics. Curr Obstetr Gynecol 2005; 15: 237–43. Taylor-Adams S. Vincent. Systems analysis of clinical incidents: the London protocol. Clin Risk 2004; 10: 211–20.
Further reading Department of Health. An organisation with a memory: report of an Expert Advisory Group on learning from adverse events in the NHS. London: Stationery Office, 2000. Department of Health. Standards for better health. London: Stationery Office, 2004. Reason JT. Understanding adverse events: the human factor. In: Vincent C, ed. Clinical risk management: enhancing patient safety, 2nd edn. London: BMJ Books, 2001, p. 9–30. Royal College of Obstetricians and Gynaecologists. Clinical governance advice number 2. Improving patient safety: risk management for maternity and gynaecology. London: RCOG, 2005.
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
Practice points • Risk management is vital to prevent the recurrence of error • Root cause analysis identifies faults in the system • Errors cost the NHS millions of pounds every year (>2,000 million pounds since 1995 in obstetrics alone) • More than 10% of inpatients experience harm (1% die) due to an adverse incident.
162
© 2008 Elsevier Ltd. All rights reserved.
Risk management in obstetrics
Raising awareness at all levels of the workforce is essential, from senior consultants to trainees, and from midwifery managers to ancillary staff (including cleaners and porters). A common misunderstanding is that risk management is the concern of service managers, not clinicians and other staff. Clinicians play an invaluable role in all aspects of risk management.
Charlotte Cottee Kate Harding
Application Increasing emphasis has been placed on risk management and clinical governance, with clinicians’ knowledge of these subjects being assessed in job interviews and included in postgraduate training and examinations. The application of risk management is seen at all levels within an NHS hospital trust. • Senior management may be concerned with the local strategy for infection control in a department, and its possible effect on unit closure if an epidemic (e.g. avian flu) arises. • A multidisciplinary team may care for a pregnant, unsupported teenager with social problems; assessing and treating the risks raised during the course of her care. • A clinician will identify a clinical risk and plan management to minimize harm. For example, she may organize an ultrasound scan to identify a placenta accreta in a woman with a placenta praevia and two previous caesarean sections. She will then organize suitable support from interventional radiologists at the time of delivery (possibly including transferring care to a tertiary centre) to prevent massive post-partum haemorrhage. After the event, there may be a meeting to reflect on care, and consider changes to guidelines to maximize future care to ensure lessons from the case are learnt.
“Those who do not remember the past are condemned to repeat it.” George Santayana (1863–1952)
Abstract Risk management is a vital tool in preventing repetition of errors which are costly to patients, staff and institutions. If risk management is used appropriately, faults in the system can be eliminated before they cause an event, or to prevent event recurrence. With detailed analysis of incidents (using tools such as root cause analysis) the fault in the system (rather that the individual) can be identified and rectified. The evolution of risk management (and its development from industry to the health sector) in the UK is described. The cost to the health service is explored and methods of analysing risk are discussed (including risk registers). There is a description of the elements that are reviewed in a Clinical Negligence Scheme for Trusts assessment. There is also a worked example of a clinical incident, its analysis and the solutions recommended.
Keywords audit; clinical governance; CNST; NPSA; risk management; root cause analysis; training
How big is the problem? Patients and staff The effect of risk on patients can range from minor effects to severe disability or even death. There are also the unquantifiable effects such as the psychological impact and loss of faith and trust in the healthcare system. Such effects may be longlasting and influence how the individual uses the healthcare system in the future. Healthcare professionals are also affected by risk. Staff can become: • disillusioned with their careers • demotivated in the workplace • demoralized and at higher risk of making an error; or more likely to leave the health service to find other jobs with less exposure to risk, leading to recruitment problems and understaffing. Morale is often severely affected and the effect on personal reputation should not be underestimated. More than one-third of doctors who are sued suffer from clinical depression.
Definition Risk management is the systematic identification, assessment and evaluation of risk. It can improve the quality of care, and develops safe systems of care that minimize adverse events for patients and staff. Risk management can be reactive (e.g. in response to a serious incident or a complaint), pro-active (e.g. establishment of a risk register or an assessment of national guidelines/reports) or preventative (e.g. ensuring adequate training and staffing levels). Healthcare professionals should understand the principles of risk management. Risk management is not only a tool to reduce litigation or a process to report incidents, it minimizes the harm that errors can cause to patients and staff. Risk management encompasses many aspects of clinical governance, from risk reporting (and response to complaints) to audit, guidelines, risk assessments (risk registers) and training.
Financial The problem of risk in obstetrics from a financial perspective is huge. Data from the NHS Litigation Authority (NHSLA) shows that, between 1995 and March 2007, obstetrics and gynaecology accounted for 21% of claims (8,532) and this was valued at 49% of the total amount claimed (£2,475 million). In total, 41% of claims are abandoned, 41% are settled out of court, and of those that go to court, ∼29% are settled in favour of the
Charlotte Cottee MRCOG is a Specialist Registrar in Obstetrics and Gynaecology at Guy’s and St Thomas’ Hospital Foundation Trust, London, UK. Kate Harding FRCOG is a Consultant Obstetrician at Guy’s and St Thomas’ Hospital Foundation Trust, London, UK.
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
155
© 2008 Elsevier Ltd. All rights reserved.
Review
claimant. This money (including legal fees) comes out of the NHS budget. About 10.8% of UK hospital patients experience an adverse event, of which 1% can lead to severe harm or death. The number of events appears to be rising (though that may be due to increased awareness), with <1,000 per quarter reported to the National Patient Safety Authority (NPSA) in 2003, and >200,000 in 2007 as the percentage of trusts reporting increased from <5% to >60%. Of these incidents, ∼35% are due to patient accidents (e.g. falling out of bed, tripping in hospital) and 60% are due to clinical events (medication, treatments, communication).
scheme and maternity (as the area of greatest financial risk) was assessed separately from the rest of the acute trust. The CNST rating reflects the breadth of risk management within the organization, and the extent to which the principles of risk management are embedded. Maternity CNST assessment level 1 (which gives a trust a 10% reduction in its annual premium) has been achieved by all maternity units in England and Wales. (Table 2). Level 3 (the highest level; it gives a 30%
Components of CNST assessment Organization Risk strategy Lead Obstetrician for labour ward Risk Manager
Evolution of risk management Several reports have been instrumental in developing risk management over the last decade. In 2001, “An organisation with a memory” highlighted the need to learn from clinical error. The NPSA was established in 2001, with a remit to develop a national approach towards reporting incidents and learning from them. This included development of a national database for: • reporting patient safety incidents • distribution of safety alerts • formulating solutions to identified risks • Internet training for root cause analysis. NPSA analysis concentrates on a systems approach rather than focusing on individuals. A document by the Department of Health entitled “Standards for better health”, published in 2004, discussed improving the quality of care by assessing it with respect to seven domains (Table 1), the first being safety. It was recognized that patients can suffer harm as a result of care that was intended to help them. Measures and strategies to reduce this happening can be implemented, and risk management is one approach.
Learning from experience Assessment on the relevance of national reports (e.g. CEMACH, NCPOD, CESDI) Assessment of attainment of recommendations of NICE guidelines Structure of incident reporting Action plans from adverse incidents and complaints Communications
Getting results into the notes Auditing action on abnormal results Handover of care between professionals Achieving an acceptable decision-to-delivery interval
Clinical care
Guidelines Audit of guidelines Action on audit recommendations
Induction, training and competence Induction of all staff Mandatory training for all staff CTG Resuscitation Drills and skills Technical competence Equipment Recovery area
Clinical Negligence Scheme for Trusts (CNST) After crown indemnity was removed (whereby the NHS centrally funded all claims), it became clear that in some financial years, some trusts (and in turn their patient population) may suffer severe financial embarrassment if there were several high-paying claims (often obstetric in origin). The CNST was set up to allow trusts to pool their claims/risks, and with each trust paying a proportion of its turnover into a common pot, high and low claims were balanced out so no single trust would suffer if they had a number of huge claims (e.g. for birth hypoxia) in one year. A trust had to show a basic level of risk management to enter the
Record keeping
Storage of CTGs Documentation audit Improvement in documentation after audit
Implementation of risk management recommendations Evidence of action plans from incidents and complaints Evidence of improvement after recommendations Up-to-date risk register Involvement of trust board involvement in risk management
Health care domains Safety Clinical and cost effectiveness Governance Patient focus Accessible and responsive care Care environment and facilities Public health
Staffing
Adequate staffing levels of all staff groups Evidence of staffing levels Midwifery supervision Actions on how to improve poor staffing Table 2
Table 1
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
156
© 2008 Elsevier Ltd. All rights reserved.
Review
is looked at by risk analysis and evaluation. When considering minimising the chances of risk happening or mitigating damage from such risk, risk treatment is required. The cost of prevention is balanced against the cost of getting it wrong. Learning from what wrong is a very important element of the process and this is encapsulated in risk control and sharing and learning from the experience Figure 1.
Identify risks
Analyse risks Communicate
Monitor
Risk identification Evaluate risks
Risk is usually identified from incident reporting (using ‘trigger lists’ (see Table 4 or unusual events), complaints, or internal investigations of major adverse incidents. These three sources are familiar to those working in clinical areas, and may be grouped as internal sources. Proactive assessment of new care pathways or guidelines should identify risks and seek to eliminate them (e.g. guidelines should be reviewed by a multidisciplinary team to avoid errors). Risk assessment in each clinical area is a trust requirement and is the responsibility of senior members of staff. The risk register is reported back to the trust clinical governance board and onto the trust board (Table 3). Complaints and claims also used to identify risk, and should feed into the risk register. The Patient Advice and Liaison service (PALS) gives impartial support to patients and relatives to voice their concerns about any aspect of care, away from the clinical area, and to help them express a complaint if that is required. Clinical audit can identify and confirm/refute risk prospectively and retrospectively. Direct consultation with staff can identify risk by surveys, interviews and departmental workshops. Risks can also be identified from external sources, such as: • national reports (e.g. CEMACH) • national guidelines (e.g. NICE intrapartum guidelines) • Government initiatives (e.g. Cleaner Hands campaign) • research (e.g. the finding that corticosteroids reduce neonatal ARDS).
Treat risks
Eliminate
Reduce
Finance
Figure 1 The risk management process.
r eduction in premium) has been achieved by 19 units, ranging from small units to the largest units in the UK. By giving trusts a financial incentive to reduce risk, it is hoped that patient safety has been improved because it ensures that at least basic risk management procedures are in place in all maternity units.
Risk management process The risk management process can be viewed as a systematic approach to dealing with a sequence of questions that aim to reduce risk. See box C. For example when looking at what could go wrong? Risk identification is required. Understanding the chances of it going wrong again and what would be the impact,
Risk register Ref
Source
Date
Description
Lead
Rating
Action
Completion
1
Trust board
1 Jan 2004
MRSA infection
Red
2
Incident
5 March 2004
Management of PPH
Head of Infection Control Head of Obstetrics
75% staff Dec 2004 Four opened Sept 2004 Nil available Purchased Nov 2004 90% trained January 2005 Completed Feb 2005
3
CEMACH report
December 2004
Puerperal suicide
Clinical Director Women’s Health
Green
Clean hands campaign Increased isolation rooms Invest in rapid assay Invest in rapid blood transfusor, Haemacue® and a TEG Mandatory training New PPH protocol Set up perinatal psychiatry team Improve links with current psychiatric teams Midwife training in booking questions
Orange
Currently no funding or staff Monthly meetings since December 2005 Introduced to new notes January 2006
Risks are described in full in the referenced appendixes. Rating depends on level of risk and probability; therefore suicide, though catastrophic, is rare and gets a low rating. MRSA is a common problem and causes significant embarrassment to the trust, so it gets a high rating.
Table 3
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
157
© 2008 Elsevier Ltd. All rights reserved.
Review
Suggested trigger list for reporting incidents in maternity units Maternal incident
Fetal/neonatal incident
Organizational incident
Maternal death Undiagnosed breech Shoulder dystocia Blood loss >1500 ml Return to theatre Eclampsia Hysterectomy/laparotomy Anaesthetic complications ITU admission Venous thromboembolism Pulmonary embolism Third/fourth-degree tears Unsuccessful forceps or ventouse Uterine rupture Readmission of mother
Stillbirth >500 g Neonatal death Apgar score <7 at 5 minutes Birth trauma Fetal laceration at caesarean section Cord pH <7.05 arterial or <7.1 venous Neonatal seizures Term baby admitted to neonatal unit Undiagnosed fetal anomaly
Unavailability of health record Delay in responding to call for assistance Unplanned home birth Faulty equipment Conflict over case management Potential service user complaint Medication error Retained swab or instrument Hospital-acquired infection Violation of local protocol
Table 4
Incident reporting system
Risk register
Clinical areas should have an anonymous system to report risk incidents. Trigger lists categorize incidents into maternal, fetal and organizational. These lists should be easily accessible in the clinical areas, and healthcare professionals should be aware of their contents. A trigger list compiled by the Royal College of Obstetricians and Gynaecologists is shown in Table 4. A multiprofessional risk management team reviews the reports and assesses if a detailed analysis is required. In a simple case, the incident will be included in a risk log; a more complex case may require a review of the notes and a discussion with staff members, and simple training measures are often all that is required. The most complex cases require thorough investigation and detailed action plan to be reviewed by the clinical directorate (and occasionally the hospital trust) to ensure that systems errors are not repeated and relevant staff learn from the event (e.g. setting up a new service or employing extra staff).
A risk register is a centrally held ‘live’ document that should contain all the risks identified by a maternity unit and the trust. The register can be populated from incidents, complaints, national alerts and proactively looking for problems. These items should be graded according to how high the effect of the risk is (financial, media exposure, patient harm) and the likelihood of them occurring. The death of a neonate from MRSA is of moderate cost, high exposure in terms of bad publicity, but relatively rare. Admission to ICU due to post partum haemorrhage is likely to occur 2–6 times a year, but is not financially costly (if well managed; Table 5). To calculate the level of risk, the likelihood of recurrence is estimated, as is incident severity (including the effect on the trust financially and in terms of public confidence). Using the matrix, a grading from green to red is allocated; in general, trusts are likely to allocate funding to all red risks. An example of a risk register is given in Table 3.
Risk matrix to assess risk severity Consequence
Insignificant
Minor
Moderate
1
2
3
Probability of recurrence
Major 4
Catastrophic 5
Certain Highly likely Likely Possible Rare
Table 5
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
158
© 2008 Elsevier Ltd. All rights reserved.
Review
Analysis and evaluation of risk Identify incident and decision to investigate Reporting an incident is the responsibility of all members of the team involved. Not all incidents will need a full systems investigation and the severity of the incident often dictates the timescale in which it needs to be investigated. The risk management team will make these decisions and they are fed back to all involved with the incident.
“What we call experience is often a dreadful list of ghastly mistakes.” J Chalmers da Costa (1863–1933). People do not intend to commit errors. The ethos and training of medicine is based on primum non nocere (firstly, do no harm). It is not productive to focus only on the individual who was involved in an incident; the entire system must be analysed. This is known as a ‘systems view’ as opposed to a ‘personcentred view’. Understanding medical error and particularly the human factor components are vital when considering what went wrong. An error cannot be analysed in isolation, and associated factors must be evaluated: • conditions in which it occurred • frequency of the activity • degree of uncertainty of clinical practice in which it took place • multiple actions leading up to the error. The different types of error must be understood. Human errors can be divided into intentional and unintentional (see Scholfield, Further reading). A systems approach involves viewing the incident in terms of failures that are active or latent. Active failures are the immediate causes of an incident. They are made by staff working in the clinical area. An example is the midwife that gives syntometrine instead of syntocinon for active management of the third stage in a woman with pre-eclampsia. Latent failures are more insidious. They are a result of actions made by the organization, often at many levels. These actions create the overall conditions in which the incidents occur, and may increase their chance of happening. In a busy understaffed unit, staff may be more likely to pick up the wrong drug or not have time to fully assess the patient (a latent failure of the system rather than an active failure of the individual). To investigate error and reduce the chance of it recurring, the conditions in which the error occurred must be analysed. Scholefield states that the commonest condition leading to error is unfamiliarity with a task, followed by shortage of time, and poor human equipment interface. These conditions can be grouped in terms of influencing factors (Table 6).
Form investigating team The team should represent all areas of the system to be investigated and should not include those who have been directly involved with the incident. For example in the case of a serious labour ward incident the team could comprise: a senior midwife, the lead Consultant for Labour Ward Data gathering Information is best collected as soon as possible after the incident occurs. Necessary documentation includes: a copy of the incident form, medical notes and detailed statements from those directly involved with the incident. Documentation related to the incident such as copies of departmental protocols and information about relevant conditions affecting the event e.g. duty rotas and reports on the level of activity in the unit at the time, are very useful. Staff interviews may be necessary and provide a good source of additional information about an incident when conducted in a structured manner by an experienced member of staff. Determine chronology of the incident A detailed chronology of the incident can be constructed using various techniques such as timelines, flow charts, time person grids and narratives. Identify care delivery problems Care delivery problems (CDPs) are often easily identified from records and interviews. It is important to recognise each problem as a specific action or omission rather than including general observations as these can lead to inclusion of contributory factors. Identify contributory factors Contributory factors are addressed for each identified CPD. Each CPD may be associated with several contributory factors. A fishbone diagram is a good way of recording contributory factors. The CDP represents the fish’s head and the contributory factors are represented by the fish bones. Another method is to tabulate the chronology of the incident and map the CDPs and contributory factors to this table.
Incident analysis
Make recommendations and deliver action plans After the above analysis of the incident a set of recommendations and improvement strategies can be formed. The contributory factors are prioritised in order of their impact on future safe healthcare delivery and action plans are made for each contributory factor. At this stage it is vital to recognise that action plans can be divided into those at a local level and those at a directorate or organisational level. To improve the implementation of recommendations it is important to identify who is responsible for implementing the actions, identify the timeframe for the implementation to be completed and a mechanism by which evidence of completion is checked. Ideally the effectiveness of the plan implemented needs to be evaluated and audit is a good tool for this.
After an incident has occurred, reproducible systems-approached analysis can be employed to investigate the event. The aim is not
Influencing factors Patient factors Task factors Individual factors Team factors Work environment Organizational and managerial factors
Figure 2 The London protocol.
Table 6
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
159
© 2008 Elsevier Ltd. All rights reserved.
Review
Organisation and management culture
Contributory factors influencing practice
Care delivery problems
Work/environment factors
Defences/barriers
Unsafe acts
Team factors Management decisions and organisational processes
Errors
Individual (staff) factors
Incident Task factors Violations Patient factors
Latent failures
Active failures
Error and violation producing conditions
Figure 3 Adapted version of Reason’s model of organisational accident.
solely to establish what went wrong, but to identify contributory factors and pathways that may be altered to minimize the chance of event reoccurrence and maximize delivery of safe healthcare. Various approaches have been used to achieve this analysis, such as root cause analysis (RCA). RCA originated in industry and uses a series of tools to investigate and analyse incidents. It has been well adapted to the UK healthcare system, and is used as a structured investigation technique for adverse outcome analysis. RCA has been criticized as being too focused on looking at the incident and what caused it, without concentrating on the chain of events and contributory factors, thereby failing to recognize factors that may have high error-producing potential. The London Protocol (Figure 2) has been devised by Taylor-Adams and Vincent to overcome this failing. It advocates a broad examination of all aspects of the healthcare system in which the incident occurred, incorporating investigation, analysis and recommendations. It is particularly focused on members of the healthcare team, from managers to those working in clinical areas, their communication and teamwork, and how they interact to create a safe environment. The London protocol approach was based on Reason’s model of organizational accident (Figure 3). This illustrates how fallible decisions at organizational levels are transmitted down departmental pathways to clinical areas, creating task and environmental factors that lead to errors. Defence barriers are in place to protect against human or equipment errors (e.g. training and checking) but an incident may occur if these are permeated at multiple levels. This is also known as the ‘Swiss cheese effect’. Figure 4 shows an example of management of clinical risk.
carrying it out. We have moved from the concept of ‘see one, do one, teach one’ where individuals were not formally trained and their competence was not formally assessed. It was possible for a doctor to be expected to do a breech vaginal delivery despite having never seen one (or practised on a simulator). In the twenty-first century, middle-grade obstetricians are expected to have attended a skills course and to have had hands-on training for skills such as managing shoulder dystocia, breech delivery and postpartum haemorrhage. These skills can be assessed and competence tested before the trainee can pass on to the next training level. Junior doctors move between hospitals regularly. Induction to their new workplace ensures that they know: • how to use the emergency call system • the protocols (and how to access them) • who to call for help and how • the layout of their work place, and location of emergency trolleys and theatres. This induction process applies to all staff, including new midwives and nurses, new consultants, and locum and agency staff. Without an adequate induction process, the time of new staff joining the team is one of increased risk. Induction and basic skills training are important parts of the CNST assessment, with emphasis on ensuring that >90% of staff have annual skills training (including CTG analysis, maternal and neonatal resuscitation).
Conclusions Risk management is an approach to clinical care that actively searches for weaknesses in the system and attempts to correct them before harm comes to patients or staff. Risk management is often reactive rather that proactive, but at least this prevents errors being repeated. Done well, risk management involves all members of staff, is seen as a learning process rather than a
Training, induction and competence A prerequisite of good medical care is that the clinician is trained in the procedure he is undertaking and competent at
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
160
© 2008 Elsevier Ltd. All rights reserved.
Review
Risk management: case study A term pregnant female presented to labour ward on a particularly busy evening. She appeared to be in the early stages of labour and had been aware of reduced fetal movements all day. The unit did not have vacant rooms, so she was put in the waiting area. The attending midwife explained the situation and informed the woman that she would be moved if a room became available. The midwife did not take a full history or examine the patient because she was in the waiting room. Two hours later, the woman was moved into a room; difficulty was experienced with fetal heart auscultation. Ultrasound confirmed intrauterine death. The care delivery problem is delayed diagnosis of intrauterine death. There are several contributory factors, and are depicted in a fishbone analysis diagram. Fishbone diagram depicting contributory factors for a case of intrauterine death Environment
Individual
Patient
No es en te d es in el gu
In
La
su
ck
ffi
of
cie
nt
Bu
id
sy
ro
un
om
it
s
te
e bl
t en
la
m
ila
ss
va
se
ya
as
ac
or
Pr
Po
iv
pr Resources
Intrauterine death
Working conditions
Organisational and management
Making recommendations, developing action plans and auditing outcome The fundamental risk management issue is that the fetal heart was not auscultated on arrival. The possibility for intervention was missed. Each contributory factor can be addressed and action plans formed, as shown below. Contributory factor
Action
Late presentation Poor assessment of patient on arrival due to lack of privacy
Educate patients about self-referral system to triage, and opening times Extra triage rooms provide privacy for historytaking and examination
Lack of rooms
Triage suite
Level of activity within the unit
Escalation policy: guideline detailing actions to be taken if beds are not available Guideline detailing timeframe from arrival to full heart auscultation
Delay in auscultating fetal heart
The major action was to establish a triage facility to enable women to be seen soon after arrival in a private area to assess their clinical needs. The development of the triage facility was achieved by a multidisciplinary team that included representatives from the resource directorates, organizational and management teams, through to obstetricians, senior midwives and administrative staff. An audit six months after establishment of the triage showed that all women were cared for according to the timeframes stated in the guidelines, and that the average time from arrival to assessment was under five minutes. Figure 4
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
161
© 2008 Elsevier Ltd. All rights reserved.
Review
means of looking for fault, and improves outcomes for everyone involved in health care at every level. ◆
Scholfield H. Risk management in obstetrics. Curr Obstetr Gynecol 2005; 15: 237–43. Taylor-Adams S. Vincent. Systems analysis of clinical incidents: the London protocol. Clin Risk 2004; 10: 211–20.
Further reading Department of Health. An organisation with a memory: report of an Expert Advisory Group on learning from adverse events in the NHS. London: Stationery Office, 2000. Department of Health. Standards for better health. London: Stationery Office, 2004. Reason JT. Understanding adverse events: the human factor. In: Vincent C, ed. Clinical risk management: enhancing patient safety, 2nd edn. London: BMJ Books, 2001, p. 9–30. Royal College of Obstetricians and Gynaecologists. Clinical governance advice number 2. Improving patient safety: risk management for maternity and gynaecology. London: RCOG, 2005.
OBSTETRICS, GYNAECOLOGY AND REPRODUCTIVE MEDICINE 18:6
Practice points • Risk management is vital to prevent the recurrence of error • Root cause analysis identifies faults in the system • Errors cost the NHS millions of pounds every year (>2,000 million pounds since 1995 in obstetrics alone) • More than 10% of inpatients experience harm (1% die) due to an adverse incident.
162
© 2008 Elsevier Ltd. All rights reserved.