In 1996, China adopted a virological classification of acute flaccid paralysis (AFP) cases for its surveillance system. Only AFP cases with wild poliovirus in stool specimens are confirmed as polio. Cases with adequate stool specimens that are negative for wild poliovirus are not counted. This paper describes a methodology to rule out poliomyelitis in AFP cases with inadequate stool specimens. National surveillance data were analysed using dot maps to detect clusters of AFP cases with high-risk factors for poliomyelitis. The surveillance system and vaccine coverage were assessed during field investigations. Four clusters of AFP cases were identified, but no poliomyelitis cases. Programmatic failures in the identified high-risk areas included low vaccination rates, poor stool specimen collection and inadequate AFP surveillance. Programme strategies were implemented to correct the identified failures. Use of this methodology provides strong evidence consistent with the absence of wild poliovirus in China.