Cost analysis of the treatment of severe spinal spasticity with a continuous intrathecal baclofen infusion system

The purpose of our study was to analyse and evaluate the costs of continuous intrathecal baclofen administration as a modality in the treatment of severe spasticity in the Netherlands.

A cost analysis was conducted as part of a prospective, multicentre, multidisciplinary, randomised and placebo-controlled clinical trial. The study covered the period from December 1991 to September 1995. The data on medical consumption and costs were collected over a 3-year period from different sources: administrative databases of health insurance companies, hospital registries and a patient survey. These data were structured by means of a flowchart analysis of the medical decision-making by specialists and general practitioners (GPs). They included data on in- and outpatient care, home care and care in nursing homes. The cost analysis was conducted using data from 18 patients included in the trial and from 15 so-called 'match' patients. The latter group are patients with comparable diseases leading to spasticity and living in comparable circumstances. Next to absolute costs (direct and indirect) of care and treatment for the 2 groups of patients, cost differences between the 2 groups were considered (differential cost analysis).

Per patient cost data, collected prospectively for 2 years during the phase of clinical evaluation, and retrospectively 1 year before implantation. The data were collected on patients from in- and outpatient care, home care and care in nursing home settings.

The trial patients (8 men) had a mean age of 46 years; 11 patients had multiple sclerosis and 7 patients had spinal cord injuries. The match patients (7 men) had a mean age of 48 years; 9 patients had multiple sclerosis and 6 patients had spinal cord injuries.

Trial patients were treated with a subcutaneously implanted programmable continuous infusion pump (SynchroMed, Medtronic), filled with baclofen (a muscle relaxant) to treat patients with chronic disabling spasticity who did not respond to a maximum dose of oral baclofen, dantrolene and tizanidine.

An analysis of hospital stay between both groups showed a significant difference during the implantation year. The average number of hospital days per patient in the year in the treated group was 31.5 days and in the match group was 18.7 days. Significant cost differences between both groups in the year that started with pump implantation and the following year can be attributed mostly to the costs of implantation of the pump and related hospitalisation days. The total costs of patient selection, testing, implanting the pump and follow-up amounted to $US28,473 for the first year. Savings must be taken into consideration as well. The savings of direct costs were due to withdrawal of oral medication (estimated annual total of between $US1950 and $US2800 per patient). Indirect savings on employment and nursing home costs, amounted annually to $US1047 and $US5814, respectively. Scenarios make it possible to consider policy consequences. The case of 'extending' the indications for this treatment to a larger population has been calculated and visualised.

The costs of the therapy (continuous intrathecal infusion of baclofen) can be attributed mostly to implantation of the pump and related hospitalisation days. Savings originated from withdrawal of oral medication, job preservation and avoidance or delay of admission to a nursing home.