A review into the quality of NHS complaints investigations where serious or avoidable harm has been alleged

Annex B: The review - summary

We undertook this review because our casework tells us that there is a wide variation in the quality of NHS investigations into complaints that patients have suffered serious avoidable harm.  We completed this in January 2015 and the aim was to establish whether the NHS complaints process is acting adequately as a safety net to identify and deal with failings in care and patient safety incidents. We also looked for features of good practice. 


We identified and considered 288 cases about the NHS in England that we investigated in 2014.  In each of the 288 cases a patient, or relative, alleged that they had suffered avoidable harm because of NHS treatment.  Out of those cases, we identified 150 that raised issues of serious avoidable harm or death at acute trusts. The focus of our review was to look at the features and quality of the NHS investigation into the allegation, rather than the result of our subsequent investigation.  We therefore did not discriminate between cases that we had upheld or not upheld. 

Our investigators reviewed the case file for each of the 150 cases.  They answered a series of questions3 about the quality of the trust’s original investigation into the complaint and the evidence that the trusts had relied on in coming to their decisions.  

The questions were: 

  • Was the allegation of avoidable harm or avoidable death?
  • What was the nature of the alleged avoidable harm?
  • What was the main alleged clinical failing leading to avoidable harm or avoidable death?
  • Which specialism was complained about?
  • Was a serious incident investigation carried out?
  • Do you consider that it should have been? 
  • Did the organisation understand and investigate the complaint put to it? 
  • Was the complaints investigation carried out by appropriate staff?
  • Did the organisation communicate adequately with the complainant?
  • Did the organisation have access to the relevant clinical records?
  • Was there a review of the care and treatment by appropriate clinical staff?
  • If yes, was the review done by a clinician not involved in the patient’s care?
  • Were key staff interviewed? 
  • Were key staff asked to provide a written statement?
  • Was any relevant evidence missing or not considered? 
  • Were the investigation findings reasonable and based on evidence?
  • Did the organisation give the complainant an adequate explanation of what happened and why?
  • Did the organisation find failings relating to avoidable harm or death?
  • If yes, did the organisation find out why things went wrong?
  • If failings were found, did the organisation take action to ensure patient safety?
  • How long did the investigation take?
  • Was the investigation adequate or inadequate?
  • Was the complaint upheld or not upheld by us?

What we found

Our initial review bore out our premise that the NHS complaints process does not adequately address complaints about avoidable harm.  Out of the cases we reviewed, over one third of investigations into allegations by patients, or their relatives, were not good enough to identify if something had gone seriously wrong.  

We found that one third of investigations did not have reasonable conclusions that were based on evidence, and did not reliably identify when something had gone wrong.  

Equally we found that, even when investigations did identify failings, the trusts did not always try to find out why something had gone wrong, or take remedial action.

In our review, 14 investigations (9%) found failings relating to avoidable harm; however, our subsequent investigations identified failings relating to avoidable harm in 52 cases (35%). Furthermore, in only 9 of the 14 cases did the trust try to find out why something had gone wrong, and in only 10 of the cases did the trust take action to try to make sure patients were safe in the future.

In the majority of cases the trusts had access to the relevant clinical records, and in 56% of investigations written statements were obtained and 38% involved interviewing key staff. In 90% of cases a review of the clinical care was carried out, but only 52% of cases involved an independent clinical review.  In almost a fifth of cases we found that relevant evidence was missing from the trust’s investigation. Some of the reasons that our investigators gave for this included that evidence had been given orally, and not documented; interviews or written statements, although considered necessary, were not obtained, and some clinical records could not be obtained. 

We looked at the features of the investigations that we considered adequate, and those we considered inadequate.  There was no significant difference in the adequate or inadequate groups in how frequently the trusts obtained written statements, interviewed staff, or obtained independent clinical reviews.  

However, 71% of complaints that should have triggered a serious incident investigation were not identified as such.  The 20 cases that should have been classified as a serious incident included: complaints about missed opportunities to survive; delays in providing medication and fluids that could have contributed to death;  problems administering blood transfusions, leading to adverse consequences, including brain damage; and unexpected deaths.  We found that for these 20 cases:

  • 9 did not obtain written statements;
  • 9 did not interview key staff;
  • 7 did not either obtain written statements or interview key staff;
  • 4 had evidence missing;
  • 4 did not obtain a clinical review; and
  • 6 of the 16 clinical reviews carried out were not independent.

Given the seriousness of these complaints, we considered that, even if the trusts did not recognise that these cases should have been classified as a serious incident, they should have followed a more thorough investigation process.

In addition to how trusts investigated the complaints, we also looked at how they communicated with complainants.  Having reviewed the complaints files, we considered that in 27% of cases the trusts did not communicate adequately with the complainants. The reasons they gave for this include: delays in the complaints process; infrequent contact with complainants; and not keeping complainants updated about the progress of the investigation.  We also found that in 41% of cases the trusts did not provide complainants with an adequate explanation of what happened and why.

3The criteria for the questions were informed by, but not confined to, the requirements of the Serious Incident Framework.