Discussion
Acute abdominal pain is the cardinal symptom behind a vast number of abdominal conditions including several that require immediate surgical treatment. The ambition of the surgeon responsible is therefore to decide, as soon as possible, whether the underlying condition requires urgent or sub-acute surgical intervention. However, harmless or non-urgent problems may lie behind the same cardinal symptom. By employing cost- effective diagnostic measures avoiding unnecessary exposure of the patient to radiation, the challenge remains to identify those patients requiring emergency surgery from those who suffer from a less serious condition that may be treated conservatively and without time limitation. Dealing with such a highly complex decision-making process calls for a logically coordinated and systematic overall process. It is understood that the diagnosis of appendicitis is based on a balanced evaluation of signs, symptoms and tests, though, how one arrives at this balanced judgement can be discussed. We have tried to see this from the individual surgeon's point of view, i.e., we have analysed not how decision-making should be done, but rather how it is done.
The emergency department is a unique clinical milieu of inconstancy, uncertainty, variety, and complexity. In the emergency room setting the management of trauma and illness has a limited time perspective and is often carried out under pressure. This situation forces physicians to adopt a distinctive way of thinking. The aim of this study was to describe how decisions are taken under these circumstances.
When analyzing the various signs and symptoms of appendicitis we found that nausea and vomiting to some extent, were present in many cases, but had no impact on decision-making. The same was seen regarding loss of appetite. These symptoms were obviously thought of as being a sign that the patient was "ill" but not that the patient had an "appendicitis requiring surgery". The only pain characteristics taken into account by the surgeon when deciding to operate were pain in the right lower fossa and pain migration (highest scored). Pain in the right fossa and indirect tenderness were the only signs that caused the surgeon to think of surgery, and rigidity of the abdominal wall – indicating a more severe peritonitis, but only found in 7 % of cases – was obviously overshadowed by pain in the right fossa (impact on decision to operate in 38 % versus 76 % resp.). Imaging studies (ultrasonography, computed tomography), however, had the greatest impact on the surgeons' decision to operate. Even though these were performed in only two thirds of cases, results were an important basis for the decision to operate in most. In this study the findings of imaging studies were not registered, but we speculate that in most of the 70 cases the radiologic verdict was "appendicitis". This would confirm that it is mentally difficult not to operate on a radiologically or ultrasonographically demonstrated sick appendix. This presents a problem since imaging does not always describe the truth and furthermore not all confirmed cases of appendicitis require surgery. It is also interesting to note that a raised white cell count and a raised and increasing CRP– were chosen by the surgeons as factors with high impact in only a third of cases when only the option was to choose three. This is interesting since all three are well-known to correlate with the degree of inflammation and therefore likelihood of acute appendicitis.
All factors included in the study were, to varying extent, factors that strengthened the decision to perform appendicitis. However, we did not include factors that decreased the probability of appendicitis. Such factors, e.g., gastrointestinal bleeding, gynecological symptoms and decreasing CRP and leukocytes, may also have a great impact on the decision process, although in the negative direction.
This study did not aim to define those factors of most importance for decision-making in cases of suspected appendicitis, but to gain insight into what makes the surgeon decide to operate. It must also be understood the surgeons in this study were not experienced experts with a certain interest in appendicitis and appendectomy. They were, in all but a few cases, surgeons under training (½ to 5 years prior surgery) working alone but with the back-up of a resident if needed (decision of the intern) not present at the Emergency Department, usually during on-call hours. This study thus describes reality in a Scandinavian surgical department, and not an ideal situation with highly experienced surgeons. Despite the presence of senior colleague, residents adhere to a hierarchy when seeking advice in clinical matters. Furthermore, the cognitive processes employed by residents experienced in critical care are quantitatively and qualitatively different from those used by their junior counterparts; this is why the setting of our study is of importance.
The role of computed tomography and ultrasonography in cases of suspected appendicitis has recently been the subject of intensive discussion. To diagnose appendicitis, CT has a greater sensitivity and negative predictive value in older than in younger patients. CT is also associated with less negative appendectomy- rates for all female patients regardless age. In this study two thirds of the patients underwent diagnostic radiography when appendicitis was suspected, but almost all surgeons considered that this was one of the three most important factors in decision-making. Nowadays imaging is performed in more than half of patients with suspected appendicitis, but despite this, surgeons continue to rank signs, symptoms, and laboratory results as the key factors leading to appendectomy. In this study, however, we cannot say if the radiological results (positive, negative, or equivocal) influenced the process of decision-making.
The decision to rely on image diagnostics as well as performing surgery based on data available at the first examination is, to a great extent, dependent on the age of the patient. With increasing age, the prevalence of pathological conditions (e.g., diverticulitis and colon tumours) mimicking appendicitis increases. This may have had an impact on the impact of image diagnostics on the decision to perform surgery.
The surgeons were requested to state what had the greatest impact on the treatment to perform surgery when the procedure was already completed. There is, however, a natural course in acute appendicitis. All exams are not always performed at the same time as blood samples are taken or CT scan. This may have had an impact on the registration, since diagnostic measures late in the course, when the diagnosis had become more obvious, may have been attributed a greater impact than those registered immediately after admission.
An interesting conclusion from this study is that the differences in frequency of symptoms, rather than the three symptoms without ranking were considered most important by the surgeon making the decision to take the patient to the surgery. Insidious occurrence of pain (50 %), pain provoked by movement (47 %) and elevated leukocyte count (76 %) are also high on the list of symptoms not being assigned an important predictive value by the surgeon. This is probably because these symptoms only indicate abdominal disease and are thus not specific for acute appendicitis.
It is also interesting to note that in only 25 % of the 110 cases with a history of pain in the right lower quadrant was included among the three most important signs and symptoms in these cases. On the other hand, 66 % (only!) of the 106 patients with pain on palpation in the right lower quadrant were considered to have appendicitis, i.e., this was among the three signs and symptoms the surgeons ranked highest. It is possible that the surgeon considered the results of his/her investigation more important than the patient history. It may also be that when choosing from the list of alternatives, patient history gave similar information but was omitted since the surgeon put more trust in his/her own clinical investigation.
It is also noteworthy that only half of the patients (n = 66) noted pain migration (from the umbilical areal to right lower fossa), but even more astonishing is that in only half of these cases the surgeon ranked that among the three most important signs and symptoms. In textbooks this is often portrayed as being a pathognomonic sign of appendicitis. The insidious occurrence of pain and pain provoked by movement also made little impression on the surgeons.
Fever and indirect tenderness were fairly uncommon signs among our patients, and were also given low diagnostic value by the attending surgeon. Moreover raised leukocyte count, raised CRP, and increasing CRP were only seen in 76, 89, and 41 % of patients respectively. In these cases these three signs were ranked by the surgeon to be amongst the three most valuable signs in only 38, 34, and 32 % of cases. Maybe it would have been more suitable to ask for normal CRP and leukocyte counts, that usually may exclude appendicitis, (if not measured too early) in the cause of the disease.
The present study reveals that decision-making in patients with appendicitis is largely based on "hard" data such as lab results and the results of radiographic investigations. It seems that bedside clinical skill has come under pressure, be it right or wrong. There is evidence that computed tomography, for example, has a higher accuracy than the best clinical scores for diagnosing acute appendicitis. Perhaps our dependence on signs and symptoms – once the gold standard – should be re-evaluated. There are many publications that have scrutinized the various aspects of initial assessment and emergency management of acute abdominal pain. The large body of evidence, however, seems to miss articles that describe a formally correct priority- and problem-based approach. Considerable evidence suggests that wide regional variation exists in the service received by patients. Evidence-based guidelines that incorporate quality-of-life and patient preference may help address this problem. Systematic cost-effectiveness analyses may be used to improve resource allocation decisions. However, clinical decision-making has, until now, always been the cornerstone of high-quality care in emergency medicine. The intensity of decision- making in this unique milieu is unusually high, and a combination of strategies has, of necessity, evolved to cope with the load. Cognitive short-cutting strategies may be especially adaptive in situations with time and resource limitations that prevail in many emergency departments, but occasionally these fail. Detection and recognition of these cognitive phenomena must be a first step in achieving cognitive de-biasing to improve clinical decision-making in the Emergency Department.
The study has some important limitations. The most important is that patients not operated upon are excluded. When doing the next study the unoperated, should be studied as well. The other factor that should be handled differently is the result of the imaging; the way a positive finding influence, if the patient should be or not, might be quite different from a negative one.
The present study was not aimed at providing any definite confirmation regarding the correctness of the decision. The endpoint was the decision to perform surgery, not the outcome of the procedure. Accordingly, we have not considered the final outcome in terms of clinical parameters or histopathologic examination, Although this information may have served as a corroboration of the clinical decision, the purpose of the study was to study the decision process, not what it finally lead to.