Abstract
Subarachnoid haemorrhage (SAH) causes high mortality and morbidity. Learning lessons from litigation are a standard aspect of risk management in many industries. This study, therefore, examined data from the NHS Litigation Authority (NHSLA) on SAH to determine its usefulness in informing clinical risk management.
Under the Freedom of Information Act, the NHSLA database was interrogated for cases of SAH. Data were retrieved from 1995 to 2008.
Over this period, 126 cases relating to SAH were reported, 102 of which were closed. Damages were awarded in 64 cases, totalling £13.5 million. Seven claimants, associated with a high cost of continuing care, were awarded more than £0.5 million. Mean damages awarded were circa £211k. The majority of successful claims related to failure/delay to diagnose or treat.
On average, 10 cases of SAH were reported annually to the NHSLA. The rate of successful litigation was circa 1:769 presenting cases of SAH. Using lessons learned from litigation to manage risk requires good quality-supporting data. Information in the NHSLA database is limited and precludes detailed root-cause analysis. The initial concept of the NHSLA database as a claims management tool perhaps represents a lost opportunity.
Introduction
Subarachnoid haemorrhage (SAH) causes high mortality and morbidity. Many patients die prior to receiving medical attention. The case fatality rate is around 50%, though has improved modestly compared with studies from around 40 years ago. 1,2 Fifteen percent of patients will re-bleed within 2 hours of the initial bleed. 3 Misdiagnosis of SAH is thus highly likely to lead to significant neurological disability and death.
A Swedish survey of formal complaints in the management of stroke, showed that nearly all disciplinary actions against physicians concerned misdiagnoses of SAH. 4 An American study, of a large cohort of SAH patients found failure to obtain a computed tomography (CT) scan was the commonest cause of diagnostic error (73%). 5
Through this study, we hoped to (i) quantify the extent of litigation in cases of SAH and (ii) determine whether the National Health Service Litigation Authority (NHSLA) database could be used to help identify potential pitfalls in the diagnosis and treatment of SAH and the data used to inform the risk management process. A review of such causative factors may enable us to design and implement appropriate, cost-effective treatment protocols to help avoid unnecessary mortality and morbidity and avoid future litigation.
Methods
The study was designed as a retrospective review of the NHSLA database, which holds data on negligence claims against the NHS from 1995 to 2008. This database was searched to retrieve cases which related only to SAH claims. The database, from which this information is taken, was designed primarily as a claims management tool rather than for clinical risk management or research purposes; therefore, the details available are often relatively limited.
A request for information on all litigation claims relating to SAH was made, under the Freedom of Information Act. This yielded 134 cases. Three cases were excluded as they related to delivery problems in newborns. Five further cases were excluded as these related to incident reports and not actual claims.
Results
The available case data, recorded over the 13 years, since the inception of the NHSLA, can be summarized as follows:
One hundred and twenty-six cases related to SAH were recorded. One hundred and two cases were closed. Twenty-four cases were open. Damages were awarded in 64 cases out of 102 (63%). Eighteen cases (18%) were successfully defended, with defence costs only, totalling £115k, awarded. Twenty cases (19%) were successfully defended, with no costs awarded.
Failure/delay in diagnosis — 43 cases out of 64 (67%).
Failure/delay in treatment — seven cases out of 64 (11%).
Delay in performing operation – three cases out of 64 (5%). Distribution of causes by claim value
The three most common causes in successful claims were:
Figure 1 shows the distribution of causes related to claim value.
In successful claims, the top three adverse patient outcomes were:
thirty-six fatalities thirteen cases of brain damage seven cases of unnecessary pain. Accident & Emergency (32 cases). Clinical Areas (27 cases). Outpatient Department (3 cases). Casualty/Accident & Emergency (25 cases). General Medicine (20 cases). Neurosurgery (6 cases). Neurology (5 cases).
The three most common incident locations, in successful claims, were:
The four most common specialties implicated in successful claims were:
Over the 13-year period, the total cost of damages awarded to patients in closed cases was £13.5 million. Seven claimants, associated with permanent brain damage, were awarded more than £0.5 million each, presumably due to the extent of disability and high cost of continuing care for those affected. The mean damages awarded were £211k. The maximum damages awarded in any single claim were £2.1 million. The total cost of successful claims to the NHS was £17.1 million.
Discussion
Anecdotal evidence suggests that ‘missing’ or failing to promptly treat SAH is a common fear among medical professionals involved in acute neurological emergencies. Subsequent litigation claims can also be very costly to employing organizations.
Between 1990 and 1998, the rate of litigation against the NHS increased from 0.46 closed claims per 1000 finished consultant episodes to 0.81 per 1000. The rate of closed claims increased during the 1990s by about 7% per annum. 6
Within the NHS, between 1995 and 2005, the cost of all closed claims for neurological illnesses was £37 million (2% of expenditure on claims for medical and surgical specialties over that period). Nine percent of the total neurological litigation was due to misdiagnosis of SAH. 7
The data regarding population rates of misdiagnosis of SAH are sparse. One Canadian study reported about one in 20 SAH patients was ‘missed’ during an Emergency Department (ED) visit. 4 The diagnosis can be challenging because headache is often the only manifestation. Important risk factors for misdiagnosis include the degree of physician experience and whether diagnostic resources are used appropriately. The risk was greater in non-teaching hospitals (adjusted for annual volume of SAHs seen in the ED) and access to CT scanners. Approximately 2% of SAH are not identified on CT scan, and are picked up on lumbar puncture. 1
Our analysis correlates with one meta-analysis, which found three preventable errors: 8 firstly, a lack of awareness of the spectrum of presentation, secondly a failure to obtain a CT scan and thirdly failure to correctly perform/analyse cerebrospinal fluid results.
The concept of informing risk management through lessons learned from experience is a standard part of the business operating system of most large private enterprises. Processes in such industries are formally documented and reviewed for their efficacy using tools such as Potential Failure Mode and Effects Analysis (PFMEA), a tool developed by the National Aeronautics and Space Administration to improve reliability and adopted by the United States Department of Defense. 9 The use of PFMEA spread through the Western automotive industry in the 1990s, in response to the superior quality and reliability of products from the Far-East.
PFMEA considers the potential failures associated with each step in a process and their effects on the client. Each potential failure is allocated three rankings (on a scale of 10):
Severity: based on severity of its effect (10 = serious disability or death).
Occurrence: based on field data on complaints and litigation.
Detection: based on current controls and their reliability.
These are multiplied together to give a Risk Priority Number (RPN). Efforts can then be focused on implementing additional controls or permanent corrective actions for the potential failures with the highest RPN.
The recent, widespread introduction of corporate responsibility laws has driven the introduction of PFMEA and similar tools across most industrial sectors. Such formal processes to identify, stratify and manage risks could be applied in the clinical setting, making good use of appropriate supporting data.
To enable adoption of such processes in the clinical setting, it would be beneficial if the NHSLA would routinely record more detailed clinical data in greater depth, especially root-cause information, as this could be very useful in informing clinical risk management. The initial concept of the NHSLA database as purely a claims management tool does seem to represent a lost opportunity.
Our analysis indicates that diagnostic and treatment delays are the main cause of errors leading to successful claims. The database does not provide sufficient information to establish the detailed root cause(s) in each case.
A recent review of a range of litigation databases from medical defence organizations and the NHSLA also suggested that the National Patient Safety Agency should play an important role in putting systems in place to make more clinical use of the available data. However, a recent study of a range of litigation databases in the UK, found the overall reliability of the NHSLA database to be 82%. 10 On a random sample of 435 entries, 18% had errors, mostly related to inadequate or incomplete data. The reliability of the information coding in our study is also somewhat doubtful. In some successful cases citing delay in performing operation as the cause, the clinical vignettes suggest the real cause may have been a delay in diagnosis. In one unsuccessful claim, the cause was recorded as application of excess force, where the case vignette had recorded paralysis as a result of hydrocephalus.
The potential for litigation is a concern for many clinicians. One UK study found that over one-third of senior clinicians had at some point been involved in litigation, which had caused them personal distress. 11 Schattner 12 suggests the response to the fear of litigation is often to cast aside evidence-based medicine and replace it with batteries of tests (which include expensive imaging) and unnecessary follow-ups, referrals and admissions.
This study suggests that the absolute risk of litigation is low. Only 102 cases were closed, over 13 years, and damages were awarded in only 63% of those. Given the incidence of SAH of 1:10 000 per annum, 13 this represents a risk of successful litigation against a clinician (in England and Wales) in the order of 1:1098 cases arising. Assuming 30% pre-intervention mortality, the approximate risk of successful litigation would appear to be in the region of 1:769.
However, while the risk of litigation may be relatively low in SAH, where failure does occur, the physical cost to the patient and the financial cost to the NHSLA are often high. The PFMEA model, if applied, would be more likely to identify and prioritize situations where intervention or proactive risk management is necessary. The Chair of the NHSLA Professor Dame Higgins recently stated that collaboration between the NHSLA and NHS Trusts was important to identify key risk areas and to learn from mistakes. 14 However, we are not aware of a mechanism that easily facilitates such discussions.
In June 2012 the Health Sector Regulator, Monitor, launched an independent review into the ability of providers of NHS services to participate fully in improving patient care. 15 A readiness to make available data from past mistakes, and to implement systems in which clinicians and managers can use it to learn to avoid future incidents, might be one way of achieving this.