Assessment
16. The Trust should have assessed Mrs I in line with the GMC’s Good Medical Practice guidance. In line with section 15 of this guidance, the doctors should have: • adequately assessed her conditions, taking account of her history, and her views and values. If necessary, the doctors should have examined her • promptly provided or arranged suitable advice, investigations or treatment where necessary • referred her to another practitioner where this served her needs.
17. Mrs I’s medical records show that the ED doctor took a relevant history, examined her, arranged appropriate investigations, and referred her to a more specialised doctor for further advice. At this stage, the assessment of her needs was in line with the GMC’s Good Medical Practice guidelines.
18. As part of the initial assessment, Mrs I had a chest X-ray in the ED which indicated she had pneumonia. Once the Trust knew she had pneumonia, the ED doctor used the CURB65 tool to inform decision-making in her care. This is recommended by Section 1.2.3 of NICE guideline CG191 (pneumonia in adults), which also sets out what action should be taken based on the CURB65 score.
19. The CURB65 tool uses five parameters to estimate the risk of death arising from pneumonia in adults. For each parameter met by the patient, the patient scores one point: • confusion • blood urea nitrogen (BUN) over 7mmol per litre • respiration rate equal to or over 30 breaths per minute • systolic blood pressure under 90 or diastolic blood pressure under 60 • aged over 65.
20. The ED clinician scored Mrs I as two on the CURB65 scale, based on her BUN/urea levels and her age. However, her diastolic blood pressure was also consistently under 60 in the ED, which meant she should have also scored a point for this parameter. This would have made her CURB65 score three. Because of this error we cannot reasonably conclude that the Trust appropriately assessed Mrs I in the ED.
21. We have found that although the initial tests were undertaken in line with the GMC’s Good Medical Practice guidelines, the Trust made an error in using the results of these investigations to inform her care. Miscalculating Mrs I’s CURB65 was a failing in the assessment of her needs in the ED.
Treatment plan and discharge
22. The treatment plan for Mrs I’s pneumonia should have been guided by the CURB65 tool, as outlined in NICE Guideline CG191. In line with this guidance, the clinicians should have used their clinical judgement alongside the CURB65 score. The scores of the CURB65 assessment should have indicated the following treatment: • for scores of 0-1, consider home-based care • for a score of 2 or more, consider hospital-based care • for a score of 3 or more, consider an intensive care assessment for the patient.
23. Our emergency medicine adviser noted that the clinical narrative for the decision-making around Mrs I’s treatment plan does not reference the CURB65 score. Her score of two meant hospital-based care should have been considered. This does not necessarily mean she should have been admitted to hospital, but a clear rationale of the clinical judgement used to make this decision should have been documented. This did not happen.
24. The CURB65 score was also incorrect, as Mrs I should have had a total score of three. This score made her a higher mortality risk and the doctors should not only have considered hospital treatment, but also an assessment for intensive care.
25. As outlined in CG191, clinicians should use their clinical judgement alongside the CURB65 score. This means that if a clinician clearly documents why they have departed from the guidance, and this appears to be clinically sound, then this would not necessarily indicate a failing. In this case, however, the rationale for discharge into the community was simply stated to be to ‘avoid [an] admission’, without any indication of why this was necessary or any reference to her CURB65 score. This was not a sufficient documentation of the clinical decision-making in Mrs I’s care.
26. Our emergency medicine adviser explained that based on Mrs I’s physical observations it was not appropriate for the Trust to have discharged her home with antibiotics and a follow-up appointment. They explained that, at the very least, a senior clinician should have assessed her disease severity and considered the need for high dependency care. This course of action is outlined in guidance published by the British Thoracic Society on the management of pneumonia. Section 49 of this guideline states that ‘patients who have a CURB85 score of three or above should be reviewed by a senior clinician at the earliest opportunity to refine disease severity assessment and should usually be managed as having high severity pneumonia’.
27. We have found that the decision to discharge Mrs I without a review of her disease severity by a senior clinician was not appropriate and was not in line with the guidelines from the British Thoracic society. The decision was made based on the outcome of a standardised tool score that had been miscalculated by the ED doctor. This was a serious oversight by the Trust and was a failing.
Impact 28. We asked our physician adviser whether the failings in the ED could have caused Mrs I to deteriorate and, ultimately, whether this led to her death.
29. Research undertaken by Carlos et al. (2023) indicates that for a patient of Mrs I’s age with a CURB65 score of three, she would have had between a 78% and 84% chance of survival with the correct treatment. This means that whilst there would still have been a risk of death, it would have been more likely than not that she could have survived with the correct treatment.
30. What we must consider, however, is how the failings we have identified impacted on her chances of survival. It is still possible that Mrs I could have died if these mistakes had not happened. Therefore, we must consider what happened in the three days between her discharge from the Trust’s ED and approaching a different ED, and whether what happened over these three days more likely than not caused her death. We must also consider what could have been different had she received treatment consistent with a CURB65 score of three.
31. Our physician adviser noted that when Mrs I attended the second ED her clinical condition appeared to be very similar to her presentation three days prior. Her kidney function had not deteriorated over these three days, and she had not become severely hypoxic (low levels of oxygen in the blood). Her CRP (an important indicator of inflammation) and white cell count were both high on admission to the second ED, but we do not have a record of her CRP or white cell count from the first ED. This means we cannot know if these had increased during these three days.
32. Overall, our physician adviser noted no significant clinical deterioration during these three days; however, we also know that Mrs I was feeling worse, which was the reason why she went back to another ED. In addition, the impression of the doctors at the second ED was that Mrs I’s clinical presentation indicated ‘severe community acquired pneumonia’. This was different to the conclusion reached by the physicians three days prior, and this could indicate there had been a deterioration in her clinical presentation.
33. We have also considered the treatment provided to Mrs I and whether this could have been different had the Trust correctly calculated her CURB65 score. Mrs I was prescribed oral amoxycillin upon discharge from hospital. NICE guideline CG138 (pneumonia: antimicrobial prescribing) recommends amoxycillin for individuals with a CURB65 score of two or below; however, for those with a CURB65 score of three to five, patients should be prescribed co-amoxiclav, alongside either clarithromycin or erythromycin. This indicates, on balance, that the type of treatment Mrs I was prescribed would more likely than not have been different had the Trust not miscalculated the CURB65 score.
34. Research undertaken by Wei et al (2024) indicates that for patients who are in hospital, there is no difference in mortality outcomes for patients who are prescribed co-amoxiclav and those who are prescribed amoxicillin, when controlling for disease severity. This does not mean these two drugs are equivalent, but it does mean that there is evidence indicating little difference between mortality rates in patients who are prescribed these drugs.
35. This indicates that although the type of antibiotic prescribed to Mrs I did not align with NICE guideline CG138 for patients with a CURB65 score of three, this is unlikely to have significantly affected her risk of mortality. That said, this study focused on patients who were in hospital. Therefore, it is possible that there could be a difference in mortality outcomes for patients who are discharged into the community. Unfortunately, we have not been able to find any evidence of outcomes for patients being treated in the community.
36. The main difference between treatment in the community and treatment in a hospital would be the provision of supportive measures, for example fluids and supplemental oxygen therapy. Mantero et al. (2017) found good evidence that ‘supportive care is essential to ensure stability of vital functions altered by the acute condition and to prevent complications related to loss of function’ in patients with pneumonia.
37. We cannot know precisely what impact the supportive treatment provided by a hospital admission could have had on Mrs I’s chance of survival, though this evidence suggests that it can be an important component of patient care and it can support a better clinical outcome.
38. We must also consider that hospital admissions can also negatively impact health outcomes for older adults. Brooker and Purdy (2022) outline that admission to hospital itself is also associated with an increased risk of mortality in older patients. Therefore, we must balance the possible benefits of the supportive treatments that hospitals can provide with the increased risks associated with hospital admissions.
39. When we balance the evidence, we know that there was still a risk of death even if Mrs I had received the correct assessment and treatment at the first ED. Though her chances of survival would have been statistically high, there was still an approximately one in five chance of mortality even if nothing went wrong. This means whether or not her death could have been avoided hinges on whether the difference in treatment over those three days led to a clinical deterioration that could have contributed to her death.
40. On balance, we cannot robustly conclude that Mrs I would more likely than not have survived had her CURB65 score been correctly calculated and appropriate treatment provided.
41. We have found that the errors made by the Trust resulted in a missed opportunity for a better clinical outcome. This means that whilst we cannot say Mrs I’s death could have been avoided, we can say she lost the opportunity to have a more specialised clinical assessment. This assessment could have identified that different antibiotics and hospital-based supportive interventions may have been advantageous in improving her prognosis.
42. Mrs I’s family will never know whether this missed opportunity could have led to a different clinical outcome, and this is a serious injustice to them. It is understandable that this would cause the family distress.
43. When Miss U complained to the Trust about her cousin’s care, it failed to identify the error we have found. This is concerning as, according to the Trust’s responses to Miss U, two different clinicians reviewed Mrs I’s medical records and did not realise her CURB65 score had been miscalculated. This indicates the error may not be a one-off instance of human error, and that there may be a broader issue with the assessment of pneumonia in older adults at this Trust.
44. The Trust has not yet acknowledged what went wrong or taken steps to put this right. Therefore, we have outlined the actions that we recommend the Trust take to put things right below.