A quality assurance trial which dealt with tuberculin testing and isoniazid prophylaxis for tuberculosis infection among hospitalized patients was undertaken at a short-term, general hospital of the Indian Health Service. Standards for care were developed by the hospital staff using questionnaire and Delphi process techniques. Following the demonstration of baseline deficiencies in clinical performance, tuberculin testing became a standing admission order. Subsequent changes in the completeness of tuberculosis screening and in physician response to a positive tuberculin skin test were demonstrated. Among hospitalized patients of unknown tuberculin reactivity, PPD screening coverage increased from 45 per cent to 87 per cent (p<0.005). Improvement in the appropriate initiation of isoniazid prophylaxis was demonstrated. The absence of simultaneous changes in other health surveillance activities suggests that the improvement in tuberculosis screening was attributable to the quality assurance intervention and not to other intervening variables. An approach to selecting alternatives for remedial action when quality assessment has documented a deficiency is presented and discussed.