We examined the consistency of pain reporting by patients in a community pain management practice in Michigan. We compared pain levels (range 0-10) entered by patients in questionnaires versus those provided during their face-to-face physician encounter on the same day. Both of these values were available for approximately 10,000 encounters during the study period (2010-2014). Two subpopulations of patients were identifiable. One was consistent in reporting worst or least pain levels on the questionnaire and during the provider encounter. The other was discordant. Factor analysis had previously identified severity scales for patient biopsychosocial characteristics derived from the full questionnaire. The two subpopulations differed in their factor profiles even though they had similar demographics. In general, pain reported directly to physicians was more correlated to biopsychosocial indicators. Pain self-reporting using questionnaires has often been assumed to be ground truth, but those obtained during the physician encounter may be more reliable.