OBJECTIVES

Structured treatment programs have been recommended for management of patients with chronic conditions to overcome ill-coordinated care. We aimed to evaluate a disease management program (DMP) for diabetes mellitus patients in Switzerland.

METHODS

We performed a prospective observational study with a propensity score-matched usual care control group from a claims database. We included type-1 and type-2 diabetes patients from a primary care setting. The DMP (intervention) comprised a structured treatment approach with an individual treatment plan, treatment goals and an interprofessional team approach. Our comprehensive outcome measures included quality of life (QOL: EQ-5D-5L), pre-defined indicators for diabetes guideline adherence, number of used services and direct medical costs. We applied a difference-in-difference (DID) approach to compare DMP with usual care (follow-up 1 year). Costs were calculated with non-parametric bootstrapping (2017 Swiss Francs, CHF; conversion rate to Euros: 0.85) from a third-party payer perspective (Swiss health care insurance).

RESULTS

QOL in a sub-sample of 80 patients did not change during follow-up (mean utility 0.89 at baseline and follow-up; p=0.94). Guideline adherence showed slight improvements for DMP. For example, “non-adherence” (baseline DMP: 19%) decreased in the DMP group by -3 %-points (DID; 95%-CI: -0.07 to 0.01) but not in the control group. A general trend emerged, though mostly not statistically significant, with less used services in the DMP group compared to the control group. Costs increased in both groups during follow-up, but the increase was higher in the control group (DID, mean total costs per patient per year: CHF -950.00 [95%-CI: -1959.53 to 59.56]). Such a negative difference-in-difference estimate in favor of DMP also emerged for cost sub-categories (e.g. costs for inpatient and outpatient care).

CONCLUSIONS

The structured treatment program under evaluation is a promising approach to improve diabetes care in a Swiss primary care setting but more follow-up data are needed.