Acute appendicitis in patients over the age of 65 years; comparison of clinical and computer based decision making

The role of clinical and computer based decision in the diagnosis of acute appendicitis in the elderly was studied in connection with the Research Committee of the World Organization of Gastroenterology (OMGE) survey of acute abdominal pain. A total of 220 patients over the age of 65 years presenting with acute abdominal pain were included in the study at the Central Hospital of Savonlinna and at the University Hospital of Tampere. Twenty-two preoperative clinical history variables, 14 clinical signs and three tests were evaluated in a single variable and multivariate analysis to find the best combination of predictors of acute appendicitis in the aged. In order to sum up the contributions of independent diagnostic factors, a diagnostic score (DS) was built: DS = 2.81 x (rectal digital tenderness; 1 = yes, 0 = no) + 2.54 x (rigidity; 1 = yes, 0 = no) + 2.06 x (renal tenderness; 1 = no, 0 = yes) + 2.33 x (bowel sounds; 1 = normal, 2 = absent/abnormal) - 8.13. The sensitivity of preoperative clinical decision in detecting acute appendicitis in the aged was 0.79 with a specificity of 0.92, an efficiency of 0.90 and usefulness index (UI) of 0.56. At a cut-off level of -2.78 the DS reached a sensitivity of 0.84 in detecting acute appendicitis with a specificity of 0.87, an efficiency of 0.87 and UI of 0.68. When the patients with a DS value between -2.78 and -0.45 were considered as non-defined (n = 28, follow-up required before the decision to operate), the sensitivity of the computer-aided diagnosis in detecting acute appendicitis in the elderly was 0.77 with a specificity of 0.97, an efficiency of 0.96 and UI of 0.57. In the elderly patients where a leucocyte count was available (n = 157), location of pain, rectal digital tenderness and leucocyte count predicted significantly acute appendicitis. At a cut-off level of -2.62 the DS reached a sensitivity of 0.81 in detecting acute appendicitis with a specificity of 0.92, an efficiency of 0.91 and UI of 0.59. When the patients with a DS value between -2.62 and 0.06 were considered as nondefined (n = 12, follow-up required before the decision to operate), the sensitivity of the computer-aided diagnosis (leucocyte count available) in detecting acute appendicitis in the elderly improved to 0.86 with a specificity of 0.94, an efficiency of 0.93 and UI of 0.69. In our study the diagnostic scoring system for the elderly performed well considering the simple nature of its structure.(ABSTRACT TRUNCATED AT 400 WORDS)