[Prognostic factors from a long-term follow-up of heroin-assisted treatment in Switzerland 1994-2007].

OBJECTIVE

To identify prognostic factors for a positive or negative termination of heroin-assisted treatment (HAT) in Switzerland.

METHOD

A complete census of all 3155 patients ever admitted was analysed using the proportional hazard model (including time dependent covariates).

RESULTS

Median length of stay was 11.4 years; the maximal length of stay was 13.9 years. 299 positive and 463 negative terminations were registered. Terminations clustered in the first year. Both time to positive and negative termination was significantly dependent on historical treatment cohorts since 1994. Positive termination was negatively associated with treatment in larger treatment centres (OR: 0.77, CI: 0.61-0.97) and positively with income from the social system (OR: 1.33; CI: 1.03-1.72). Negative terminations were positively associated with HIV infection before treatment (OR: 1.74; CI: 1.40-2.16), delinquence (OR 1.36; CI: 1.09-1.69), and higher levels of distrust (OR: 1.18 per scoring point; CI = 1.05-1.31).

CONCLUSIONS

Length of stay in Swiss HAT is considerable. The proportion of positive terminations did not increase with longer stays, indicating that the majority of patients are in chronic palliative care. Negative terminations outweighed positive terminations, with a low predictive power from co-variates. The routine assessment and analysis of different covariates, such as indicators of treatment process, has the potential to improve the therapeutic outcomes of HAT.