Clinical failings leading to avoidable harm
Patient harm is a global phenomenon that happens everywhere patient care takes place. Over 1.25 million staff in the NHS have 1.5 million patient interactions every day, so it is inevitable that mistakes will happen. But, despite improvements over the last ten years, the estimated number of 11,000 avoidable deaths every year is shocking.
After analysing the 22 cases where we found that a death was avoidable, we identified four broad themes of clinical failings leading to avoidable harm:
- failure to make the right diagnosis
- delays in providing treatment
- poor handovers between clinicians
- failure to listen to the concerns of patients or their families.
Failure to make the right diagnosis
In more than half of the cases we looked at, we found that a failure to diagnose directly led to or significantly contributed to the avoidable death of a patient. In these cases, the right diagnosis could have been made at the right time. If this had happened and the correct treatment had been provided, it is likely the patient would not have died.
A range of factors and circumstances can lead to clinicians not making the right diagnosis at the right time, especially when a patient’s condition is complex or critical. In our casework analysis, failures to make the right diagnosis were mainly the result of:
- not seeking more senior or specialist input where that would be appropriate
- not observing or monitoring someone at regular enough intervals to recognise deteriorating health or new issues
- failings in imaging, such as failing to follow up on an unexpected finding.
In one case we looked at, the patient’s National Early Warning Score (NEWS score) showed he was very unwell and should have been reviewed by a more senior clinician with the knowledge and experience to diagnose and treat him while his condition was getting worse. The Trust should have asked for an urgent review by a more senior clinician, but this did not happen. When the patient’s condition got worse, the Trust delayed asking for an urgent review, and there were further delays in reviewing the patient. This meant the Trust failed to recognise the patient had sepsis and did not give him the right treatment. The patient suffered a cardiac arrest (when the heart stops pumping blood around the body) and died.
We identified the importance of appropriate and timely senior input in our report on sepsis, ‘Time to Act’, noting that ‘the skills necessary for early identification of patients at risk of severe sepsis are high level and develop with long experience’. It is concerning that a decade after this report, and following huge efforts to improve the early recognition and treatment of sepsis, we still see these kinds of failings.
Case study: failure to diagnose pulmonary embolism
A man contacted an out-of-hours GP service because he had been suffering from shortness of breath for over a month and had a The GP diagnosed a chest infection and prescribed steroid medication.
The patient’s family called for an ambulance the same evening because his breathlessness had started to get worse. The ambulance crew arrived and took him to hospital.
The patient’s observations showed a fast heart rate, raised rate of breathing and a low level of oxygen in his Staff gave him a nebuliser and his oxygen levels improved.
A doctor reviewed the patient. An X-ray did not show any concerns. The doctor diagnosed pneumonia, prescribed antibiotics and decided not to admit the patient.
The following day, a GP visited the patient at home and prescribed an inhaler. They suggested the patient should return to the hospital.
Later that day, the patient’s family called an ambulance because they were concerned about his condition. When the ambulance arrived, the patient was conscious but short of breath.
The patient suffered a cardiac arrest and the paramedics tried to resuscitate. Unfortunately, they could not do so and the patient died. The cause of death was a pulmonary embolism (a blocked blood vessel in the lungs).
We found the doctors failed to follow the relevant guidance when they reviewed the patient at the hospital.
The clinical signs the patient presented suggested a pulmonary embolism was the likely cause of their breathlessness, high heart rate, high respiratory rate and low oxygen. Doctors should have arranged a specific blood test, but this did not happen. Instead, the patient was sent home without the right treatment.
The doctors made a diagnosis of pneumonia (inflammation of the lungs) but the chest X-ray results showed the patient’s lungs were clear and there was no evidence of blood tests to confirm a diagnosis of infection.
We found it likely the patient would have survived had he received the right treatment after the doctor reviewed him. This would likely have led to the diagnosis of pulmonary embolism and successful treatment with the right medication.
We spoke to the patient’s mother who described her and her family’s experience of seeing the patient die at home:
‘My son came home with absolutely no information whatsoever … He should have been in hospital, he should have been on a ward. He died in his bedroom and I had to hear it … My grandson was in the room next door with me and he heard him die too. It’s not something that you would want for anybody.’
Failing to observe or monitor someone at regular enough intervals can result in missed opportunities to recognise deteriorating health or new issues. In one case we looked at, a man was admitted to the Trust’s emergency department with chest pains. His condition got worse in hospital but staff did not carry out observations to monitor him as frequently as they should have done. This resulted in a missed opportunity to notice the patient’s health getting worse due to sepsis and to involve critical care. It delayed the diagnosis and treatment of sepsis for over ten hours. We found that if the patient had been monitored at appropriate intervals, it is likely his worsening health would have been identified, he would have received treatment for sepsis and, more likely than not, he would have survived.
In another case, a mental health Trust did not observe a long-term inpatient as frequently and as closely as it should have done. The woman attempted to harm herself in a way that could have led to her death on two occasions. Staff managed to stop her from harming herself both times. The Trust should have updated the risk assessment to include actions to reduce any risks and minimise harm, including risks posed by ligature points (anything that could be used to attach a cord, rope or other material for the purpose of self-harm). The following day, the patient was found unresponsive and died because of a hypoxic brain injury (when the brain does not get enough oxygen). We found that if the Trust had observed the patient more closely and frequently and taken action to address ligature points, it was unlikely the suicide attempt would have been completed. If the patient had been monitored more frequently, it is likely there would have been the opportunity to intervene in time and prevent the woman’s death.
We spoke with the patient’s sister who told us the Trust did not manage to keep her sister safe:
‘It was supposed to be a place of safety and it clearly wasn’t. My sister shouldn’t have died in the hospital.’
Case study: failure to carry out repeat risk assessments
A patient went to a Trust’s emergency department with a swollen left hand. The patient had cellulitis (a bacterial skin infection that causes swelling) and gout (a build-up of uric acid causing swelling in the joints).
The Trust admitted the patient to the hospital and considered their risk of developing venous thromboembolism (VTE, a condition that happens when a blood clot forms in a vein).
The Trust decided there was no risk and did not carry out any further risk assessments. Staff treated the patient, which included elevating their arm and giving them antibiotics.
A couple of days later, the patient started experiencing chest pains and tightness in their The Trust carried out an electrocardiogram (ECG, a test to check the heart’s rhythm and electrical activity) and provided medication. Later that day, the patient became sick and collapsed.
The doctor planned to move the patient to a bed with more monitoring. But, before this could happen, the patient suffered a cardiac arrest. The Trust resuscitated the patient and gave them medication to break down blood clots.
After this, the patient remained stable for a short time during which they were moved to the intensive care unit. Sadly, their condition did not improve and they died that evening. The Trust carried out an autopsy which confirmed the cause of death as a pulmonary embolism (blockage of a lung artery).
We found the Trust failed to carry out a full assessment of the patient’s risk of developing VTE, and then failed to reassess them 24 hours later as outlined in NICE guidance. It missed two opportunities to provide appropriate and timely treatment to prevent a pulmonary embolism.
Had the Trust given the right treatment on time, it is more likely it could have prevented the pulmonary embolism and avoided the patient’s death.
Delays in the treatment response
Our casework shows that serious harm can be caused by delays in providing treatment. In our analysis, we found that these delays often relate to:
- a diagnosis being made but then not being acted on properly or quickly enough
- not acting quickly enough on observations.
In one of the complaints we looked at, the Trust appropriately and quickly diagnosed a woman with vasculitis (inflammation of the blood vessels). Vasculitis is a condition that needs urgent treatment. But the Trust did not prescribe the medication needed for another 16 hours after the diagnosis.
After prescribing the medication, the Trust did not give it to the patient for another 11 hours. The Trust significantly delayed giving the patient the life-saving treatment she needed for 27 hours. Our investigation found the patient could have survived if she been given the correct medication in time. It is more likely than not that her death could have been avoided.
When there is evidence that a patient’s health is getting worse it is vital to respond quickly, but we found instances where this did not happen. For example, in a case where the on-call doctor did not attend when requested to, we found that the way they were contacted (via a ‘bleep’) did not include information about the urgency of the request. The Trust explained the system does not give any sign of urgency unless the doctor is being called to attend a cardiac arrest. Our investigation report highlighted that this is a problematic system because it does not allow clinicians to prioritise their patients’ needs. We argued that it undermines the benefits of a warning system like NEWS if clinicians are not meaningfully empowered to act quickly in response.
This is a well-known issue. It is highlighted, for example, in a recent national investigation report by HSIB on recognising and responding to critically unwell patients. The report notes that early warning scores ‘can place a high demand on medical staff and the current escalation protocols may not be achievable owing to a task versus resource mismatch’. Clearly, monitoring is only useful if there are enough staff to respond when deterioration is detected.
Case study: failure to provide timely sepsis treatment
A patient went to hospital for a hysteroscopy (a procedure to examine the inside of the womb).
The procedure was completed without incident and the plan was for her to stay in hospital to be monitored.
After being transferred to the ward, she felt Staff carried out assessments to examine her symptoms. The results showed abnormal blood tests, including a very low white blood cell count.
The patient was referred to and reviewed by the emergency medical response team. The doctors considered sepsis as one of several possibilities for explaining the clinical findings. But they thought this was unlikely so they did not prescribe antibiotics.
The patient was transferred to the acute medical unit for closer look. The doctor who reviewed her recommended an immediate dose of antibiotics.
Unfortunately, she did not receive the antibiotics until four hours later. Her health deteriorated significantly in that time.
This delay in receiving antibiotics meant the patient did not receive timely treatment for sepsis in line with the NICE sepsis guidance.
Staff performed a CT scan (a computerised tomography scan, which creates detailed images of the inside of the body) the following morning. They prescribed another dose of antibiotics, reviewed the patient and transferred her to the intensive care unit to receive medication to support circulation.
The patient’s condition did not Doctors explained to her family they could give no further treatment and started end of life care. Sadly, she died that evening.
We found that the delays in starting antibiotic treatment and admitting the patient to the intensive care unit (to support her circulation) significantly affected her chances of survival.
If the Trust had done what it should have done, it was more likely that the patient would not have died.
Poor handovers
Good communication between clinicians and appropriate handover between teams is an essential part of patient safety. General Medical Council guidance to doctors, for example, stresses the importance of working collaboratively, communicating effectively and sharing information between teams. Equally, the Nursing and Midwifery Council Code of Practice highlights the need to maintain effective communication and keep health and care professionals informed when sharing the care of individuals, to ‘preserve the safety of those receiving care’. This includes the need for timely and appropriate handovers between teams. Research has identified that transfers into or out of intensive care units and when information has to be communicated to other teams during a critical care stay can be risk points for patient safety incidents.
In almost half of the cases we analysed, we found failings in communication between different clinicians and teams. In one case, for example, we could not find any evidence that the Trust followed a clear and coordinated clinical management plan when a man in his seventies attended the Trust’s emergency department with a twisted bowel. We found that poor communication between teams contributed significantly to the Trust failing to take urgent action to reduce the risk of bowel perforation. When investigating another case, we could not find evidence of a documented handover between teams. This meant we could not understand what actions the Trust had taken at the time to make sure the patient continued to receive adequate care.
Case study: failure to diagnose and lack of handover between teams
A patient went to a Trust’s emergency department after an out-of-hours GP told them to call 999. The patient said they had a racing heart, felt generally unwell and had pain in their forearms.
The Trust carried out a series of checks and found them to be within normal limits. The doctor prescribed an antibiotic and discharged the patient.
A few days later, the patient was admitted to a medical assessment unit after a GP referral for a respiratory opinion.
A consultant reviewed them and noted they might have a blood clot and fluid around the lung.
The consultant requested a troponin test (a blood test that can help assess heart damage), echocardiogram (a scan to look at the heart and nearby blood vessels) and repeat ECG, and prescribed a blood-thinning injection to treat the suspected clot.
Another doctor reviewed the patient and considered a pleural effusion (when excess fluid builds up in the space between the lungs and the chest wall) rather than a blood clot. Staff gave the patient antibiotics and drained the fluid around the lung. The following day, the doctor noted the patient’s symptoms were improving.
A couple of days later, a consultant respiratory physician noted that the patient had aching pain in both arms, breathlessness and difficulty breathing when lying down. They reviewed the test results and considered the patient had congestive heart failure (when the heart cannot pump blood around the body properly), which was probably secondary to a recent heart attack.
The consultant asked for an ECG and troponin test and prescribed medication. They left instructions for staff to triage the patient as a cardiology patient and arrange an ECG to check the heart’s rhythm and electrical activity.
The next day, the patient was moved to a ward and nurses noted that a doctor needed to do a medical review, including reviewing the ECG and troponin test and prescribed medication. They left instructions for staff to triage the patient as a cardiology patient and arrange an ECG to check the heart’s rhythm and electrical activity.
Sadly, the patient died the next day after suffering a cardiac arrest.
We found multiple failings in relation to the patient’s care.
The Trust had not carried out the troponin test and echochadiogram the first time a consultant requested them, and staff did not act on early warning observations when they should have alerted a doctor.
There was evidence of a heart attack in the ECG results, but this did not get an urgent medical. And no medical review was carried out after the patient was transferred to the ward.
If the Trust had made a timely diagnosis and given appropriate treatment, there would have been an improved chance of the patient surviving. It missed an opportunity to provide potentially lifesaving It was more likely than not that the patient’s death was avoidable.
We found that communication with the patient’s next of kin was flawed and incomplete, and there were failings in the Trust’s complaint handling that caused them distress and frustration.
Failure to listen to the concerns of patients and their families
Listening to patients and their families is a vital part of providing good quality care and treatment that puts the patient front and centre. In our casework, we identified a small number of cases where failures to listen to the concerns of patients and their families had a clinical impact.
For example, a man who was admitted to hospital with abdominal pain raised concerns with clinical staff on multiple occasions about his care and treatment. During our investigation we found evidence that he was concerned he did not feel well enough to leave the hospital. Despite this, the Trust still considered him well enough to go home and discharged him on the same day. Two days later, he returned to hospital and died shortly after being admitted due to gastric aspiration (where vomit enters the larynx and lungs). Our investigation found that the Trust should have carried out an urgent CT scan, which would have led to the patient having lifesaving surgery.
In another example, a complainant described how he had to persuade staff to admit his mother after she had attended the Trust’s emergency department multiple times. During our investigation, we found evidence that staff doubted the patient’s symptoms and failed to appreciate the serious nature of her condition. Understandably, this was upsetting for the complainant who could see that his mother was very unwell. The complainant’s mother died after having a stroke in hospital. This could have been avoided if she had been diagnosed correctly and treated during an earlier admission.
Summary
The themes we have identified around clinical failings are wide-ranging but will be very familiar to anyone with experience in patient safety. The caseworkers who worked on these cases observed that they reflect the types of issues we see in most of our health casework, not just those that resulted in the most serious possible outcome of a patient’s avoidable death.
Although a lot can be done to reduce clinical harm and improve patient safety, and certain incidents can be prevented entirely with the right processes, systems and cultures in place, it is unrealistic to expect that all clinical errors and omissions can be eliminated. But it should always be possible to respond well, with compassion, and in a way that shows learning and accountability when avoidable harm has happened. In the next section we look at the evidence from our casework about the impact on families when this does not happen.