In the DPP, both the metformin and lifestyle interventions were more expensive than the placebo intervention. In the metformin intervention, most of the additional cost relative to the placebo intervention was accounted for by the cost of metformin. In the lifestyle intervention, most of the additional cost relative to the placebo intervention was accounted for by staff time used for counseling and adherence monitoring. Although the lifestyle intervention cost ~37% more than the metformin intervention in year 1, the lifestyle intervention cost 12 and 7% less than the metformin intervention in years 2 and 3. Because the cost of the lifestyle intervention was greater than the cost of the metformin intervention in year 1 but less in subsequent years, the cost of the lifestyle intervention relative to the metformin intervention would decrease with follow-up beyond 3 years. To the extent that the cost of the metformin intervention can be reduced by using less expensive generic metformin and to the extent that the lifestyle intervention can be delivered with less staff time, the cost of the interventions could be substantially reduced.
The cost of identifying IGT and the cost of the interventions represented less than one-half of total direct medical costs for DPP participants over 3 years ($2,919/$7,375 = 40% in the lifestyle, and $2,681/$7,420 = 36% in the metformin group). Direct medical costs also included the costs of medical care obtained outside the DPP. The latter was influenced by both adverse health events and improvements in health related to participation in the clinical trial. There were no statistically significant differences in the incidence of serious adverse events or adverse events among the three intervention groups in the DPP. To estimate the costs of both adverse events and improved health associated with participation in the DPP interventions, we assessed hospital days, emergency room visits, urgent care visits, outpatient visits, calls to providers, and prescription medications and their differences among the intervention groups. Participation in the experimental interventions was associated with decreased direct medical costs of the care outside the DPP. Compared with the placebo intervention group, participants randomized to the metformin intervention group spent fewer days in the hospital, made fewer emergency room and urgent care visits, and took fewer prescription medications. Compared with participants in the placebo intervention group, participants in the lifestyle intervention spent fewer days in the hospital and were less likely to make urgent care visits, outpatient visits, and calls to providers. In addition, they reported taking fewer prescription medications. Taken together over 3 years, this reduced health care utilization and decreased the direct medical costs of care outside the DPP by $272 in the metformin intervention group and $423 in the lifestyle intervention group relative to the placebo intervention group. These cost savings offset a portion of the direct medical cost of the experimental interventions.
Although not generally paid by health systems, direct nonmedical costs affect the individual and society. To fully assess the impact of direct nonmedical costs, we went to great lengths to describe the resources used and to estimate their costs. Not surprisingly, participants in the lifestyle intervention group spent more time traveling to appointments, attending appointments, and exercising, and they purchased more services and products related to physical activity and diet. While the lifestyle group spent substantially more time engaged in leisure time physical activity than either metformin or placebo participants, they reported greater enjoyment of these activities. Thus, the resulting direct nonmedical cost was negligible. Both metformin and lifestyle participants reported spending less time shopping and cooking than placebo participants. Lifestyle participants also reported lower food costs than metformin and placebo participants—largely as a result of decreased cost of food consumed at restaurants. Because of more frequent DPP visits, participation in the lifestyle intervention was associated with substantially greater transportation costs. Compared to the placebo intervention group, the incremental direct nonmedical cost was $1,445 over 3 years. Over 60% of this incremental cost was related to greater participant time. In contrast, direct nonmedical costs were $9 less in the metformin intervention group than in the placebo intervention group.
When leisure-time physical activity was valued independently of the participants’ exercise preference at $0, $8, or $16 per hour, the incremental direct nonmedical cost of the lifestyle intervention compared with the placebo intervention increased substantially and ranged from $1,469 to $4,056 over 3 years. In contrast, direct nonmedical costs decreased in the metformin intervention group compared with the placebo intervention group (−$12 −$352). Thus, the cost of the lifestyle intervention relative to the placebo intervention is sensitive to the value assigned to time spent exercising. We believe that the base analysis that reflects participants’ exercise preferences is most reasonable, because individuals who enjoy exercise willingly spend their leisure time and personal resources for exercise.
Despite the greater frequency of lifestyle visits, the difference in indirect costs among intervention groups was small. The latter may reflect flexible scheduling arrangements that permit people to reduce time lost from work or usual activities. Compared with the placebo group, the indirect costs related to morbidity and mortality were lower in the lifestyle group but higher in the metformin group. Thus, compared with the placebo group, indirect costs were $174 less in the lifestyle group and $230 greater in the metformin group over 3 years.
summarizes the per capita costs of the metformin and lifestyle interventions relative to the placebo intervention over 3 years. From the perspective of a large health system, which would pay only direct medical costs, the cost of the metformin intervention relative to the placebo intervention was $2,191 over 3 years. From the perspective of society, which pays direct medical costs, direct nonmedical costs, and indirect costs, the per capita cost of the metformin intervention relative to the placebo intervention was $2,412 over 3 years. The per capita costs of the lifestyle intervention relative to the placebo intervention were $2,269 and $3,540 over 3 years from the perspective of a large health system and society, respectively.
Per capita differences in costs of the metformin and lifestyle interventions relative to the placebo intervention (year 2000, U.S. dollars) over 3 years in DPP
The DPP demonstrated that both medication and lifestyle interventions can delay or prevent progression from IGT to type 2 diabetes (1
). This analysis demonstrates that such preventive strategies are associated with modest incremental costs. From the perspective of a large health system, both the metformin and lifestyle interventions cost ~$750 per participant per year, or $2,250 per participant over 3 years. From a societal perspective, the incremental costs of both the metformin and lifestyle intervention are greater and the relative increase is greater in the lifestyle than the metformin intervention. This is not surprising in light of the greater direct nonmedical costs associated with the lifestyle intervention. Nevertheless, the incremental increases remain small. The costs of such prevention strategies must be balanced against the savings related to averted disease. It is likely that the cost of the metformin intervention will decrease substantially with the availability of less expensive generic formulations of metformin. It is also likely that the cost of the lifestyle intervention could be reduced by improving the efficiency of utilization of staff time by using group visits. Ultimate determination of the value of these interventions to health systems and society will require a formal assessment of costs relative to the health benefits achieved in the DPP.