Service quality and cost containment represent two critical challenges in healthcare management. Toward that end, acute appendicitis, a common surgical condition, is important and requires timely, accurate diagnosis. The diverse and atypical symptoms make such diagnoses difficult, thus resulting in increased morbidity and negative appendectomy. While prior research has recognized the use of classification analysis to support acute appendicitis diagnosis, the skewed distribution of the cases pertaining to positive or negative acute appendicitis has significantly constrained the effectiveness of the existing classification techniques. In this study, we develop a pre-clustering-based classification (PCC) technique to address the skewed distribution problem common to acute appendicitis diagnosis. We empirically evaluate the proposed PCC technique with 574 clinical cases of positive and negative acute appendicitis obtained from a tertiary medical center in Taiwan. Our evaluation includes tradition support vector machine, a prevalent resampling classification technique, Alvarado scoring system, and a multi-classifier committee for performance benchmark purposes. Our results show the PCC technique more effective and less biased than the benchmark techniques, without favoring the positive or negative class.