With the passage of P L 99-457 in October of 1986, the field of early intervention has been faced with the challenge of broadening its scope (Silber, 1989) This legislation expanded early intervention from a child-centered service to a serwce offered to families within a variety of contexts from a service for children with special needs to a service available to families as soon as risks are identified New programs are bemg designed and developed with an interagency focus which serve not only children with developmental disabilities but also their families in health, education and social services (Cornwell and Thurman, 1990) Current legislation suggests that services should be available to families as soon as their children are identified as being at-risk (Smith, 1987) In response to the legislation, states are developing definitions of "at-risk' based on criterie other than the child's test scores Therefore, in many cases, early mtervention services are being offered to families whilee their at-risk infants are still hospitalized and receiving intensive care Evaluation of these innovative and complex service delivery programs is bemg addressed as these services are developed With the broadening of the scope of early mtervention comes the need to rethink the tradtional methods of evaluating these programs This paper will address some of the issues involved in evaluating family-centered programs which are based in neonatal intensive care units (NICU).