Collaborative medication management services: improving patient care

We used a participatory action research design,5 in which researchers worked with participants to design, implement and evaluate the service, allowing researchers and participants to solve problems that arose as the research progressed. The participatory action research process involved general practitioners, pharmacists and consumers in a series of workshops, focus groups and feedback sessions.

A call for expressions of interest in participating in the project was sent to all Divisions of General Practice in South Australia. Eight of the 15 Divisions responded and six participated (three rural and three urban Divisions). A Divisional Liaison Officer (DLO) was employed in each Division to facilitate the interaction between GPs and pharmacists and to assist in the local implementation of the HMR service.

A collaborative service delivery model (Box 1) was agreed on, and standard forms for documenting relevant patient information, reports and action plans were developed in consultation with participating GPs and pharmacists.

Two researchers coded the medical conditions, medications taken and medication-related problems from patient case notes. Medical conditions were categorised according to the International classification of diseases (ICD-9-CM),6 medications according to the Anatomical and therapeutic classification7 and medication-related problems according to previously employed criteria (Box 3).1,3,8 From patient case notes, the researchers noted actions taken to address the problems (Box 4) and outcomes of those actions (Box 5). For some patients, the GP and pharmacist identified more than one problem.

Pharmacists identified 2764 medication-related problems, the most common (17.5%) being the need for additional tests (Box 3). On average, 2.5 problems were identified per person.

Thirty-seven per cent of all problems related to medicine selection, 20% to patient knowledge and skills, and 17% to the medication regimen. Only 1% of problems related to drug–drug interactions, 2% to contraindicated therapy, 2% to the use of medicine without indication, and 1% to duplication of therapy.

In response to the identified problems, pharmacists recommended 2764 actions to GPs, of which 1163 (42%) were documented in patient case notes as having been implemented. No information was available on implementation of the remaining actions. This may mean that no action was taken, that the action was rejected by the GP, that follow-up had not occurred at the time of the study's completion, or that the action had been previously trialled, was still under consideration or was implemented but poorly documented. The types of actions most commonly implemented included changes to medicine selection; changes to the dose, frequency or duration of therapy; and patient education (Box 4).

Follow-up data, as recorded in the patient's case notes, were available for 978 (84%) of the problems for which an action was documented as having been implemented. Problems were documented as being "resolved" or "well managed" in 61% of these cases, and "improving" in a further 20% (Box 5). Outcome data were available for 85 of the problems for which no actions were implemented. In 61 of these cases the problem was unresolved or worse at follow-up, with only 13 of the problems resolved or well managed at follow-up.

The model required GPs and pharmacists to negotiate actions based on the pharmacist's report. In 23 instances the GP refused to implement the action(s); in 31 instances, although the GP and pharmacist agreed on the action, the patient refused.