METHODS
All inpatients whose medical prescription was analysed by clinical pharmacists (1 senior pharmacist or 1 resident) in the geriatric acute care unit between January 27 and April 30, 2020 were included in this single-centre, observational, retrospective and comparative cohort study. The characteristics of those patients collected for the study were the following: demographical measures (age and gender), autonomy score (AGGIR scale - Autonomy Gerontology Iso-Resources Groups), length of stay and in-unit mortality. Those data were retrieved by consulting the hospital medical records.
All PIs carried out by the clinical pharmacy team during the patients’ prescriptions analysis were gathered. They were distinguished according to whether they were conducted prior to COVID-19 (between January 27 and March 18) or during the first wave of COVID-19 (between March 19 and April 30).
The endpoints to answer our main objective were the rate of PIs per patient and per line of treatment analysed. To answer our secondary objectives we collected the following data: drug class managed by the PI, DRP identified and nature of advice given. The PIs data were collected for the study via Business Object© (v12.1.0, SAP Walldorf, Germany).
The study was conducted in accordance with the ethical standards set forth in the Helsinki Declaration (1983). The Angers ethical committee approved the study protocol under number 2020/140. The study protocol was declared to the National Commission for Information Technology and civil Liberties (CNIL) under number ar20-0058v1.
Computerized medication orders and pharmaceutical analysis were allowed with three interfaced software packages (v8.2.6, Maincare, Cestas, France): Crossway® and Horizon Expert Order® for the medical prescription and M-Pharmacie® for the pharmaceutical analysis.
The prescriptions were analysed on working days (Monday to Friday) by the senior clinical pharmacist or their resident according to the standards set out by the French Society of Clinical Pharmacy (SFPC) [13] based on the patient’s medical records, test results, medication records and with the Hospital’s therapeutic booklet taken into account.
The tools used for analysing were: i) the French drug compendium (Vidal Hoptimal® database), ii) the kidney adapted prescription guide website (GPR®) which provides dosing adjustments according to renal clearance, iii) the screening tool to detect potentially inappropriate prescribing in persons aged 65 or older (Laroche list [14] and STOPP/START list [15]) and iiii) the Geriatric Dosage Handbook 14th Edition (Semla T., Beizer J., Higbee M.). The latest internal clinical guidelines for COVID-19 were used during the pandemic.
PIs can be carried out by treatment lines. A treatment line corresponds to the prescription of a new drug (original prescription or addition during hospitalisation), a discontinuation or suspension of a drug or a dosage adjustment.
The standardisation of the PIs was proposed by the SFPC [16]. Their tool for the documentation of PIs includes the identification of the drug related problem and the therapeutic advice given. The detailed categories are presented in Appendix 1.
A PI is notified in both the analysis software and the prescription software. PIs are discussed orally with the medical team and considered as accepted if they lead to a change in the prescription. The acceptance rate of our PIs was assessed in this study.
The clinical impact of a PI was evaluated with the Clinical, Economic and Organizational (CLEO) tool v3 [17] after a consultation between the senior pharmacist and a geriatrician from the unit. This consultation was made retroactively while analysing the data and the assessment of the problem wasn’t patient-specific. The clinical impact scale ranges from −1C (harmful) to 4C (vital). The different scores are presented in Appendix 2.
All statistics were performed using SAS© (v9.4, SAS Institute Inc., Cary, NC, USA). The Chi-squared test or Fisher test were used for assessing differences in proportions and the Student test or the Mann–Whitney–Wilcoxon test were used to compare the distribution of ranks between the groups. P-values<0.05 were considered significant.