You have developed a logistic regression model to classify email spam and non-spam messages. The model achieves a high accuracy, but you suspect that accuracy alone may not be a sufficient metric to evaluate the model's performance. Explain why accuracy might be misleading in the context of email spam classification.