A SERVICE DELIVERY PROGRAM MUST SEEK ITS OBJECTIVES AND BASIC DEFINITION WITHIN THE CONTEXT OF CONSUMER IMPACT AND GOALS. TWO CATEGORIES OF CLIENTS ARE INVOLVED: THE DYSFUNCTIONAL INDIVIDUAL AND HIS COMMUNITY. THE GOAL WILL BE TO DECREASE THE COMMUNITY'S RESPONSIBILITY, EXPENSE OR CONTROL BY INCREASING THE INDIVIDUAL'S INDEPENDENCE AND ABILITY TO FUNCTION IN THE COMMUNITY THROUGH TREATMENT, INTERVENTION, AND PREVENTION ACTIVITIES. A CLIENT-ORIENTED COST OUTCOME SYSTEM REQUIRES FIRST THE DEVELOPMENT OF A GLOBAL LEVEL OF FUNCTIONING SCALE. ONCE THIS HAS BEEN DONE, SERVICE PLANNING FOR EACH TARGET POPULATION CAN BE DEVELOPED IN WHICH EACH SERVICE ACTIVITY IS DESCRIBED BY THE RANGE OF FUNCTIONING LEVELS FOR WHICH IT IS DESIGNED TO HAVE IMPACT. DESCRIPTION OF ALL SERVICES IN THIS MANNER PERMITS THE IDENTIFICATION OF SERVICE GAPS OR OVERLAPS. THEN, CLINICAL AND ADMINISTRATIVE PLANNING AND DECISIONMAKING IS UNDERTAKEN BASED ON THE EMPIRICAL RELATIONSHIPS AMONG CONSUMER PROBLEMS, PERSONAL AND COMMUNITY RESOURCES, OUTCOMES, AND RELATED SERVICE EFFORTS. FINALLY, THE BASIC ELEMENTS OF THE CLINICAL DECISION PROCESS SHOULD BE CLEARLY IDENTIFIED IN CLINICAL RECORDS AND IN THE INFORMATION SYSTEM USED TO SUPPORT THE PROGRAM'S MANAGEMENT. THERE ARE SEVEN DOCUMENTATION AND SUPPORT COMPONENTS OF THE CLIENT-ORIENTED COST OUTCOME SYSTEM: PROBLEM OR GOAL-ORIENTED RECORDS, GOAL-ORIENTED SERVICE PLANS, CLIENT PARTICIPATION AND UNDERSTANDING, SERVICE EFFORT AND COST ACCOUNTING, ONGOING OUTCOME MEASURES, STRATEGIC PROGRESS REVIEW AND QUALITY CONTROL PROCEDURES, AND AN INTEGRATED MANAGEMENT INFORMATION SYSTEM. GIVEN THESE SEVEN COMPONENTS, MANY FORMS OF EVALUATION RESEARCH CAN FOLLOW BOTH AT THE CLINICAL AND AT THE ADMINISTRATIVE POLICY LEVEL. A COST OUTCOME MATRIX PROVIDES INFORMATION ON THE CLIENT'S LEVEL OF FUNCTIONING ON A DYSFUNCTIONAL/FUNCTIONAL CONTINUUM FROM ADMISSION TO CASE TERMINATION, AND INCLUDES THE COSTS PER TREATMENT EPISODE PER CLIENT. THUS, THE SINGLE MATRIX IS CAPABLE OF DESCRIBING GLOBAL CHANGES IN FUNCTIONING LEVEL AND GROSS DIFFERENCES IN SERVICE HISTORY, VIA AVERAGE COSTS, FOR A SPECIFIED GROUP OF CLIENTS. A REVIEW OF THE CELL ENTRIES IN THE MATRIX WILL RECOMMEND A MYRIAD OF PROGRAM-ANALYTIC QUESTIONS. APPLICATION OF THIS SYSTEM TO 118 CLIENTS AT A PENNSYLVANIA MENTAL HEALTH CLINIC IS BRIEFLY ILLUSTRATED, AND STEPS TAKEN IN INTERPRETATION OF COST OUTCOME DATA ARE REVIEWED. REFERENCES DIAGRAMS AND GRAPHS ARE PROVIDED. (JAP)