The study will be conducted in six primary care health centers of the Tramuntana area (Balearic Islands), which cater for a population of around 128,000 inhabitants, approximately 15% of them over 65 years of age.

Each cluster will comprise all the patients assigned to a primary care physician, who will be required to sign the relevant informed consent form. Only patients ≥ 65 years and on ≥ 5 chronic drugs will be eligible. In addition, patients who meet at least one of the following criteria will be excluded: regular use of private healthcare, being displaced, institutionalized, or on the home care program (Figure 1).

Figure 1.

Patient selection process. Criteria used to select participants both at cluster and at individual level.

(0,27MB).

Interventions will be carried out at cluster level and will consist in the performance of (PL-SMR) to detect instances where PIMs are used and provide recommendations to the prescribing physician about appropriate therapeutic alternatives, as explained below:

• –

  At the outset, the required baseline data will be obtained from the electronic medical record (EMR) of patients in both the IG and the CG.

  • •

    Demographic data: age, sex.

  • •

    Clinical data: chronic conditions.

  • •

    Pharmacotherapeutic data: chronic medications.

• –

  Detection of PIMs: To be carried out, both in the IG and the CG by means of a combination of explicit and implicit methods (STOPP/START criteria8, data from the medication labels [https://cima.aemps.es/cima/publico/home.html]), “Do not do” recommendations for complex chronic patients17, using CheckTheMeds® software (https://www.checkthemeds.com); and a battery of PQIs developed by the Balearic Islands Health System.

• –

  For patients in the IG, prescribing physicians will be informed, through the patients’ EMR and/or verbally if necessary, about the most appropriate strategy for managing each instance of PIM, taking the patients’ individual circumstances into consideration, e.g., whether they were treated previously with the proposed alternative, their potential comorbidities and/or other treatments). Such strategies may involve discontinuation of treatment; dose adjustments, replacement of the medication, addition of a new drug, and include specific management recommendations such as gradual withdrawal or onset of drugs when necessary. If use of a PIM is detected in the CG, prescribing physicians will be informed but no recommendations will be issued.

Given the purpose of the study, the following variables will be evaluated across all participating subjects:

Independent variables:

• –

  Demographic: age, sex.

• –

  Clinical: Charlson Comorbidity Index (CCI), percentage of patients with ≥ 2 chronic conditions.

• –

  Pharmacotherapeutic: number of chronic medications, proportion of polymedicated and hyperpolymedicated patients.

Primary variables: Difference between the proportion of patients on PIMs and the mean PIM/patient ratio over the period running from the time of intervention and 6 months post-intervention. To be considered a PIM, drug regimens must meet at least one of the following criteria: not evidence-based indication, prescribed dose higher and lower than recommended; duplication, contraindication, interaction, absence of a potentially beneficial drug. Moreover, all drugs marketed in the five years prior to the intervention that failed to demonstrate any significant therapeutic effect as compared with already available alternatives (e.g., saxagliptin, silodosin, rosuvastatin) as well as drugs other than those regarded as first-choice for the treatment of the most prevalent conditions in the outpatient setting (e.g., antiulcer agents-omeprazole, oral antidiabetics-metformin, statins-simvastatin) will also be considered PIMs, on the basis of the prescription quality criteria established by the Balearic Islands Health System. Secondary variables:

• –

  Morbidity: Difference between the mean number of (in- and outpatient) healthcare facilities visited in the 12 months prior and subsequent to the intervention. The INPATIENT variable comprises the number of hospital days (NHDs) spent (a stay of ≥ 13 hours in a hospitalization or emergency unit to receive specialized care was required18) and of visits made to the hospital's emergency department (ED). The OUTPATIENT variable comprises visits to the primary care emergency unit, the specialist care outpatient unit, and the primary care clinic19.

• –

  Mortality: Proportion of patients who died within the first 12 months following the intervention, and the number of days survived within the first 12 months following the intervention.

• –

  Cost of the medication: Difference in the estimated mean per-patient annual cost of the drugs considered PIMs (AC-PIM) at the time of intervention as compared with 6 months post-intervention (same time period as that used for the primary variable), according to the prices stated in the Balearic Islands’ formulary drugs price list.

• –

  Cost of healthcare resources: Difference in the mean per-patient cost of the health resources (INPATIENT + OUTPATIENT) used during the 12 months prior to the intervention as compared with 12 months post-intervention, according to the official rates published in the Balearic Islands’ Official Gazzette18.

The time periods corresponding to the different stages of the study were as follows:

• –

  Cluster recruitment: April-May 2014.

• –

  Cluster allocation: January 2015.

• –

  Patient recruitment: January-May 2015.

• –

  Intervention: January-May 2015.

• –

  Assessment of baseline characteristics: January-May 2015.

• –

  Assessment of primary variables: July-November 2015.

• –

  Evaluation of the cost of medications: July-November 2015.

• –

  Morbimortality evaluation: January-May 2016.

• –

  Evaluation of the cost of healthcare resources: January-May 2016.

• –

  Analysis and dissemination of results: June 2016-present.

The sample size will be calculated using proportion testing for a difference of 15% in the proportion of patients on PIMs in the IG (80%) and the CG (65%), based on the results of an unpublished pilot study. For a statistical power of 80% and assuming a loss rate of around 10%, the sample size required was set at 153. Dividing the sample size by the 21 clusters included it was established that each group should contain a mean of 8 patients. Application of an intra-cluster correlation coefficient of 0.0520 yielded an inflation factor of 1.35, which resulted in a final sample size of 207 patients per group.

• –

  Cluster level: All primary care physicians of the Tramuntana health area will be required to sign the relevant informed consent form on behalf of their respective patient practices. Subsequently, they will be recruited and randomly allocated to a GI o GC by the Balearic Islands Healthcare Research Institute (IdISBa).

• –

  Patient level: Pharmacists in charge of the PL-SMR will, on a daily basis and in consecutive order, select one IG and one CG cluster from the list resulting from the randomization process; of which, all patients with an appointment for the next day who meet eligibility criteria will be recruited. This routine will be followed until the target sample size is reached.

To prevent contamination across patients treated by the same physician, IdISBa will randomize clusters to the IG or the CG in a 1:1 ratio in four cluster blocks using a balanced randomization scheme. This will be done using the WINPEPI Computer Program for Epidemiologists (http://www.brixtonhealth.com/pepi4windows.html) ensuring that the confidentiality of the result of the randomization process is preserved to ensure proper data masking.

Blinding of participants (clusters) will not be feasible given that the PIM management strategies will be shared with the physicians allocated to the IG.

The data will be collected by the pharmacist in charge of the PL-SMR in an electronic spreadsheet, which will be duly encrypted and stored on a secure server guarded by the IT Department of the Inca County Hospital.

Although not specified in the protocol, an evaluation will be made of the effect of the intervention on hospital admissions in the different subgroups resulting from the combination of the following criteria: sex (males/females); age (below/over 85 years); number of drugs (5–9/10 or more).

It is not considered necessary to set up a formal external data monitoring committee or to conduct a preliminary analysis as no intervention-related damage is expected. Furthermore, in the event of recurrence of symptoms associated to the withdrawal of inappropriate medication, the intervention can be resumed at any time subject to clinical judgement.

The study protocol was evaluated and approved by the Balearic Islands’ Primary Care Research Ethics Committee and Research Ethics Committee (approval number IB2360/14) on 30 June 2014.

The protocol has undergone several alterations since its initial version. Such changes can be reviewed via the following link: https://clinicaltrials.gov/ct2/history/NCT02224833.

The Research Ethics Committee of the Balearic Islands agreed to ask physicians rather than patients to complete the informed consent form relative to the study. This was due to the fact that the PL-SMR is addressed to physician, who are the ones that must eventually decide whether the patient's treatment must be altered or not on the basis of the recommendations made by the pharmacist. Moreover, given that patients will be recruited the day before the medical appointment without any pre-screening being performed, and that the participating centers are widely dispersed across the Tramuntana health area, signature of the consent forms by the patients would be virtually unfeasible.

To ensure the confidentiality of information, patients will be identified by a 3-digit alphanumeric code followed by a letter.

This study is part of the PhD thesis of the principal investigator. For that reason, neither the data nor the materials will be made available to the public. They can only be transferred to third parties with the written consent of the University of Santiago de Compostela (http://www.usc.es/export9/sites/webinstitucional/gl/centros/cptf/edi/descargas_EDI/CDS_cas_06_10_16.pdf).

The results of this study will be published in peer-reviewed journals and presented at geriatrics, clinical pharmacy, and general medicine conferences, as well as in the PhD thesis of the principal investigator. The relevant author selection guidelines will be followed and no professional writers will be engaged.