Cost-of-illness (COI) studies often include the 'indirect' cost of lost production resulting from disease, disability, and premature death, which is an important component of the economic burden of chronic conditions assessed from the societal perspective. In most COI studies, productivity costs are estimated primarily as the economic value of production forgone associated with loss of paid employment (foregone gross earnings); some studies include the imputed value of lost unpaid work as well. This approach is commonly but imprecisely referred to as the human capital approach (HCA). However, there is a lack of consensus among health economists as to how to quantify loss of economic productivity. Some experts argue that the HCA overstates productivity losses and propose use of the friction cost approach (FCA) that estimates societal productivity loss as the short-term costs incurred by employers in replacing a lost worker. This review sought to identify COI studies published during 1995-2017 that used the FCA, with or without comparison to the HCA, and to compare FCA and HCA estimates from those studies that used both approaches. We identified 80 full COI studies (of which 75% focused on chronic conditions), roughly 5-8% of all COI studies. The majority of those studies came from three countries, Canada, Germany, and the Netherlands, that have officially endorsed use of the FCA. The FCA results in smaller productivity loss estimates than the HCA, although the differential varied widely across studies. Lack of standardization of HCA and FCA methods makes productivity cost estimates difficult to compare across studies.