Although current data suggest that periodontitis may actually be multiple diseases each with varying responses to therapy, little evidence exists to support this theory in adult patients. This report describes the design and initial findings of a longitudinal field trial involving the diagnosis and treatment of adult and refractory periodontitis patients in private practice. Adult patients (N = 221) who met specific clinical criteria were selected from the private practices of 22 periodontists. Clinical characteristics were recorded and subgingival plaque samples were sent to microbiology laboratories at either UCLA or The University of Texas Health Science Center at San Antonio (UTHSCSA). Samples were processed according to protocols standardized between the 2 centers. Five different combinations of the initial clinical and microbial findings were evaluated for patterns in the data by means of cluster analysis. Plaque, bleeding on probing, bone loss scores, probing depth distributions, and microbial findings produced multiple cluster solutions. Solutions involving 6 clusters explained 39.4% to 76.4% of the variation between patients and produced ratios for variation between clusters to variation within clusters of 5.2 to 15.3. The optimal cluster solution incorporated both clinical and microbial findings, with some clusters characterized by high plaque and moderate bleeding on probing and bone loss, whereas others had low plaque but high bleeding on probing and bone loss. Microbial findings of each cluster exhibited distinct patterns with some clusters having a high prevalence (83% to 100%) of specific target bacterial species while other clusters had an absence of these species.(ABSTRACT TRUNCATED AT 250 WORDS)