This study is being done to find out if a multi-faceted intervention designed to optimize medication in hospitalized older people with multiple chronic medical conditions exposed to multiple medications can reduce unplanned hospital readmission and emergency department attendance compared to current usual medication management. The study intervention aims to minimize potentially inappropriate medications in a structured way and involves follow up with patients and GPs. Patients will be allocated equally to (i) standard medication management (control arm) or (ii) trained physician-delivered intervention or (iii) clinical pharmacist-delivered intervention.

Background:

Recurrent hospitalization and unplanned emergency department (ED) attendance resulting from potentially inappropriate medication is an increasingly common phenomenon in older people with multi-morbidity and associated polypharmacy. With a growing older population with multi-morbidity/polypharmacy, there is a pressing need to address the increasing challenge of potentially inappropriate medications (PIMs) and potential prescribing omissions (PPOs) and associated problems that accentuate adverse drug reactions/events and avoidable excess morbidity. There is also an imperative to curb excess healthcare expenditure related to preventable medication-related problems.

Objective:

To test the clinical and economic impact of a multi-faceted medication optimization definitive intervention on avoidable rehospitalization and unscheduled ED attendance in multi-morbid patients aged ≥ 70 years hospitalized with acute illness.To test the clinical and economic impact of a multi-faceted medication optimization definitive intervention (DI) on avoidable rehospitalization and unscheduled ED attendance in multi-morbid patients aged ≥ 70 years hospitalized with acute illness.

Design:

A randomized controlled clinical trial is proposed in which it is anticipated to randomize 3 x 463 patients to one of 3 groups: (a) standard pharmaceutical care, or (b) trained physician-implemented DI, or (c) clinical pharmacist-implemented DI.

Setting: Acute care environment in 3 large tertiary referral teaching hospitals with similar custom and practice relating to management of older people with acute illness.

Participants: Patients aged ≥ 70 years with multi-morbid illness i.e. ≥ 3 chronic medical conditions and associated polypharmacy i.e. ≥ 5 daily prescription medications admitted with acute unselected illness.

Intervention:

The definitive multi-faceted intervention will consist of the following components:modified structured history of medication (SHiM), Screening Tool of Older Persons' Prescriptions (STOPP) and Screening Tool to Alert to Right Treatment (START) screening for PIMs and PPOs using STOPP/START version 3.0, drug-drug interaction screening using Stockley's Drug Interaction Checker, face-to-face consultation with attending hospital physicians to discuss PIMs, PPOs, interactions and other issues, pre-discharge medication review and adjustment, detailed medication adjustment discharge report to patients' general practitioners (GPs), post-discharge follow-up contact with patients' GPs and community pharmacists at 1 week and 1 and 3-6 months post-discharge.

Study Population:

OPTIMATE will focus on patients admitted with acute medical or surgical illness to Cork University Hospital, University Hospital Waterford and Ghent University Hospital. Patients to be screened for enrolment are those admitted primarily under the care of specialist departments other than geriatric medicine.

Data Collection:

All OPTIMATE trial data will be collected electronically and entered on a bespoke trial proforma. Once verified as fully correct and complete, all individual participant trial data will be stored on a fully secure clinical trial database.

Statistical Analysis:

Treatment effects for each of the two active intervention arms (versus control) in terms of binary outcomes (e.g. readmission) will be estimated using logistic regression, while Health Related QoL (EQ-5D 5L) will be analyzed using ordinal regression with a logit link function (i.e. a proportional odds model). We will report two models for each outcome: one adjusted for centre and admitting service type; and another further adjusted for centre, age (years), sex, number of comorbidities at baseline, and number of prescribed medications at baseline. Effect estimates will be reported as odds-ratios with 95% confidence intervals and exact p-values. There will be no correction for multiplicity, but we will report results for all outcomes regardless of the result and provide enough information for the reader to make whatever corrections they may consider appropriate. Equivalence tests comparing the two active arms (for each outcome) will be done using a "two one-sided tests" procedure based on a 90% confidence interval (which equates to a 5% type 1 error rate). All analyses will be conducted on an intention-to-treat basis and will be conducted under the quality system and SOPs of the HRB CRF-C Statistics and Data Analysis Unit. A complete statistical analysis plan (SAP) will be pre-registered on the Open Science Framework prior to patient recruitment database lock. Any necessary deviations from this SAP will be documented and explained in the trial report. There will be only one single trial data analysis at the end of the project. All reporting will be carried out in accordance with CONSORT guidelines for clinical trials.

Inclusion Criteria:

1. Age ≥ 70years
2. 3 or more chronic conditions.
3. ≥ 5 daily medications pre-admission, all medications taken for at least 4 weeks continuously.
4. Can speak and understand English (in the two Irish medical centres), and Dutch or French in Ghent University Hospital (Ghent is predominantly Dutch-speaking).
5. Can give informed consent or give witnessed verbal consent or have a suitable proxy who can give informed assent on the patient's behalf.
6. Agrees to follow-up contact post-discharge up to 180 days post-randomization.
7. Agrees to primary researcher contacting the GP and community pharmacist post-discharge.

Exclusion Criteria:

1. Terminal illness.
2. Severe dementia and clearly unable to understand the purpose of the trial or give consent to participation.
3. Severe communication disorder, making informed consent impossible.
4. Likely to be discharged from hospital within 48 hours of arrival.
5. Intensive Care Unit (ICU) admission.
6. Primary psychiatric presenting illness.
7. Unavailable for post-discharge follow-up for any reason.
8. Non-accidental poisoning.
9. Previous participation in medication optimization trials.
10. Active participation in another clinical trial
11. Infectious illness requiring strict isolation (including COVID-19 infection) blocking access of the primary researcher to the patient for enrolment.
12. End-stage renal, liver or lung disease requiring organ replacement therapy.
13. Admitted under the care of specialists in Clinical Pharmacology, Palliative Medicine, Clinical Oncology or Haematology.
14. Admitted under the care of specialists in Geriatric Medicine in Ghent University Hospital.
15. Trial participation refusal.

Consent from patients participating in OPTIMATE will be obtained for all data to be shared publicly, such as data used in the generation of publications arising from the study. Data sharing repositories will be formerly identified via careful alignment of the expected data object outputs and evaluated using the re3data resource (re3data.org). This is to ensure maximum utility and interoperability of the final data package(s) and assignment of a persistent digital object identifier (DOI). Additional post-study data provenance will be enacted through sharing of analysis scripts and study protocols via Open Science Framework projects with an accompanying DOI(s) and/or through the Health Research Board (HRB) Open Research publishing platform.

Data will be made available six months after publication.

Open access

To mimic the intervention for blinding purposes of the participants, both control and intervention arm patients will receive a sham intervention.While the intervention will be known to clinical staff, the trial statistician will remain blinded until the results are finalized.