To conduct studies of therapy of rheumatoid patients at the primary care level two new assessment techniques were developed. The "independent assessor" (IA) was a nonmedical person trained to gather information using a questionnaire, a guided count of tender joints, grip strength, and blood tests. The "polled index" was a statistical device for summarizing all abailable information, constructed by transformation of various clinical measures of treatment effects to a common scale, based on standard deviation units. A validation study is reported, which involved 33 rheumatoid patients, each studied by four trained observers plus the IA, before and after 10 days of hospital therapy. Major gains in sensitivity and reliability were achieved by the pooled index. The IAs total assessment provided 84% of the information available in the pooled index, and was more sensitive and reliable than any other single clinical measure. The efficacy of a brief period of hospital therapy for rheumatoid patients was also strikingly demonstrated.