Analysis of Deaths Related to Anesthesia in the Period 1996–2004 from Closed Claims Registered by the Danish Patient Insurance Association

The study used a retrospective design that followed claims for financial compensation as listed in the DPIA database launched in 1996. The claim from a patient comprises a description of the injury (injuries) in addition to the medical record. Each case is registered in the database under a code that identifies the patient and the medical specialty involved.

A claim for financial compensation can be made by the injured patient, the relatives, or the hospital. When a patient files a claim, the hospital is obliged to submit all medical records regarding the case to the DPIA. A lawyer evaluates the claim and may seek advice from medical specialists. Since 2002, cases regarding anesthesia have been handled by a permanently employed anesthesiologist, who provides an evaluation as to whether “best practice” has been followed. Before 2002, a medical specialist without anesthesiology training provided advice for simple claims, and an external consulting anesthesiologist (usually a professor) provided advice for more complex claims.

In general, financial compensation is granted if (1) an experienced specialist would have acted differently; (2) defects in, or failure of the technical equipment were of major concern with respect to the incident; or (3) alternative treatments, techniques, or methods were considered to be more safe and potentially offer the same benefits. At least one of the three conditions must be fulfilled before compensation is granted. In addition, an injury may lead to financial compensation if the injury is rare and more extensive than the patient would be expected to endure. It is important to note that financial compensation can be granted even in claims where no medical errors were made. The lawyer determines, in concert with the medical adviser, whether a claim qualifies for financial compensation, and the decision is forwarded to the patient. The compensation is calculated on the basis of the extent of pain and suffering, reduced income, reduced ability to work, and medical expenses, and it is considered whether the injury would be expected to be permanent.

The patient may appeal the decision of the DPIA to the Patient Damage Appeal Board and further to the courts of law.

For each claim, the DPIA creates a patient folder in which the documents of the case are kept. The most important data (sex, age, year, World Health Organization classifications of diseases [International Classification of Diseases], specialty, surgery code, procedures, complications, brief description of the circumstances, place, and so on) are, in addition, transferred to an internal electronic data system.

This review is based on information drawn from the internal data system and from the folders of all implicated patients. We searched the database for claims to injuries related to anesthesia in Denmark in the period from 1996 to 2004. We defined anesthesia related  as all procedures performed by anesthesiology staff members or patient care in one of the fields of anesthesiology (intensive care medicine, emergency medicine, perioperative care, and pain care units).

All submitted cases resulting in death were chosen for further evaluation. Two anesthesiology specialists then independently evaluated the documents of each of these cases as to whether death was caused by an anesthetic procedure or treatment. Only cases where the patient had died within a period of 3 months from the procedure or patient care were selected. Cases in which the outcome was other than death are described in a previous report.2Cases for further investigation were selected only if both anesthesiologists agreed that the case was anesthesia related. The investigation then included details as revealed in the records: age, sex, medical history, place of incident, circumstances and cause of death, and finally, judgment of preventability. A case was defined as preventable if each of the two anesthesiology specialists found that a complication from a procedure or a complication from specific patient care was considered most likely the cause of death within 3 months and that the complication could have been avoided if guidelines had been followed, or if the use of a specific technique could have prevented the complication.

We found 1,256 cases in the period from 1996 to 2004 where an injury related to the field of anesthesia resulted in a claim. Of these, 43 cases were registered as deaths. The database did not distinguish between those who died as a result of the complication and those who died from natural causes. All of the files of the 43 cases registered as deaths were thoroughly read and evaluated. By reading the files, it became clear that 13 patients only sustained minor injuries (e.g. , dental injuries, skin lesions) and died of natural causes several months after the injury. Two patients became paraplegic after epidural analgesia and died 4 and 6 months after the injury. The cause of death in both cases was considered to be cancer. Two of the claims were directed toward internal medicine and did not involve anesthetic personnel. Two patients sustained hypoxic brain damage but did not die within 3 months. Twenty-four fatal cases were considered by the two anesthesiology specialists to be a result of the anesthetic treatment. These 24 deaths involved 7 adult females, 14 adult males, and 3 children. The procedure related to the fatal outcome happened in the following locations: 13 incidents in an operating theater, 4 in a recovery room, 4 in an intensive care unit, 1 in a department of radiology, 1 in an emergency room, and 1 in the patient’s home (table 1).

The causes of the 24 deaths were as follows: 4 deaths were related to airway management (difficult intubations, aspiration, lack of observations), 2 deaths were related to ventilation management (unrecognized pneumothorax, misassembled CPAP system without expiratory valve), 4 deaths were related to administration of drugs or blood (insulin, methohexital, benzodiazepine, or incompatible blood), 4 deaths were related to infusion pump problems, and 4 deaths were related to the placement of central venous catheters. Complications related to regional blockade resulted in 4 deaths (spinal abscess, spinal lesion). Severe hemorrhage caused 1 death, and in 1 case the cause was uncertain (probably amniotic embolus or peripartum cardiomyopathy). We found that 20 of the 24 cases were potentially preventable. The average age of these 20 patients was 46 yr (range, 1–83 yr). The average age for the 4 cases categorized as not preventable was 49 yr (range, 1–79 yr). The average age of all 24 cases was 46 yr (range, 1–83 yr).

The claims for these 24 patients resulted in compensation totaling $1.1 million (range, $1,900–107,053).

In most of the 24 cases, the critical incident could be identified. We defined a critical incident as the procedure or specific patient care that eventually led to the fatal outcome.

It is important to have an incident reported in as detailed a manner as possible so that the pattern of injury may be understood, the cause of injury analyzed, and lessons from mistakes made widely known in an attempt to reduce the risk of recurrence. Because of the historic nature of the data of the current study, we were not able to perform a complete investigation of each incident, and our statements of the cause are therefore suggestive and may often seem to stop at the front-line level. A contemporary analysis would look deeper and point out causes in the systemic environment as well. In many analyses, critical events are considered preventable.3–6To reduce and avoid critical incidents, it is crucial that the anesthesiologist masters the theoretical knowledge, the practical skills, and the equipment used for the various procedures as well as coordination and communication skills.7Likewise, it is important that patient safety measures are considered in the physical and organizational structure of the working environment as well as in the apparatus used.

By examining the 24 incidents, we concluded that death could probably have been prevented in 20 of the 24 cases (table 1). The majority of these deaths can be divided into six categories: airway handling (4 deaths), ventilation handling (2 deaths), central catheter placement (4 deaths), medication errors (4 deaths), regional analgesia (4 deaths), and infusion pump problems (4 deaths). How could these deaths have been avoided?

Regarding airway and ventilation handling, in the current study, the respiratory critical incident differed in each case. Airway problems included one case where the patient developed bronchospasm after extubation probably because of aspiration. Another patient died after aspiration on a laryngeal mask where intubation should have been performed, and one patient died because an airway could not be established. One patient died unattended postoperatively, probably because of airway obstruction. The ventilation incidents included one case where the patient died because ventilation could not be established because the patient had an unrecognized pneumothorax. In another case, hypoxemia and subsequently death was caused by an incorrectly assembled ventilator. To reduce the number of injuries during airway and ventilation management, there should be more training through simulation, which is a safe means of improving anesthetic skills.8–10Finally, it is essential to have protocols and guidelines for airway and ventilation handling and to use them.

Four patients died after placement of a central venous catheter. The deaths were caused by compression of vital structures by hematomas or hemorrhage. A study from the American Society of Anesthesiologists Closed Claims Database regarding central catheter injuries showed similar serious complications due to central catheter placement.11The use of sonographic guidance in the placement of central venous catheters is a safe technique because the structures can be visualized and studies have shown very low complication rates when this technique is used.12,13Furthermore, the central venous pressure or a blood gas analysis (also compared with an arterial blood gas analysis) should be obtained before placing a large-bore catheter whenever there is any doubt about correct placement.

Medication errors and infusion pump problems caused in total eight deaths. Two of the deaths resulted when a large unintended bolus of medicine was given by an infusion pump. One of these deaths could have been avoided if an infusion pump with no free-flow possibility had been used. The other case happened because a new infusion pump was used, and the death could probably have been avoided if there had been proper education and training before the use of the infusion pump. Another two of the deaths could have been prevented if drug concentration mistakes had been avoided. Anesthetic drugs should be available in one concentration only, and if drugs are to be prepared, this should be done centrally and not locally in the acute setting. Furthermore, drug mistakes can be reduced with the use of color-coded labels for syringes and education.14 

Four patients died after regional analgesia. We concluded that one of the four deaths could probably have been avoided if guidelines for testing and epidural analgesia had been followed. One study from the American Society of Anesthesiologists Closed Claims Database regarding cardiac arrest after regional analgesia concluded that two patterns contributed to this occurrence and/or the outcome of these patients. One pattern was the use of sedative, which caused cyanosis and subsequently cardiac arrest. The other pattern was the treatment. The optimal treatment of a cardiac arrest occurring in relation to regional analgesia is positional change (Trendelenburg position) and the use of a potent α agonist.15When the combination of regional and general anesthesia is used, the patients’ risks increase, and clinicians should be sure that the benefits outweigh these risks.

In general, if the technique had been used properly, if standard procedures had been followed, and finally, if equipment had been provided with patient safety measures, many of the deaths could potentially have been avoided. Lack of knowledge of the equipment ought to be prevented by proper education. Having protocols and guidelines and using them is, in the authors’ view, essential.

Our study shows that there is a need for improvement in the safety of anesthesia in Denmark. The incidence of deaths related to anesthetic procedures is unknown. The proportion of deaths and other injuries that lead to a claim for compensation is also unknown. We found 24 deaths resulting in claims during a 9-yr period, and probably 20 of these deaths were considered preventable.