Detail
Článek
Článek online
FT
Medvik - BMČ
  • Je něco špatně v tomto záznamu ?

Clinical decision-making in benzodiazepine deprescribing by healthcare providers vs. AI-assisted approach

I. Bužančić, D. Belec, M. Držaić, I. Kummer, J. Brkić, D. Fialová, M. Ortner Hadžiabdić

. 2024 ; 90 (3) : 662-674. [pub] 20231203

Jazyk angličtina Země Anglie, Velká Británie

Typ dokumentu časopisecké články, práce podpořená grantem

Perzistentní odkaz   https://www.medvik.cz/link/bmc24006987

Grantová podpora
MSCF-ITN-764632 Marie Skłodowska-Curie Foundation
CZ.02.1.01/0.0/0.0/18_069/0010046 InoMed
SVV 260 551 European Horizon 2020 I-CARE4OLD
START/MED/093 EN.02.2.69/0.0/0.0/19_073/0016935 European Horizon 2020 I-CARE4OLD
965341 European Horizon 2020 I-CARE4OLD
Faculty of Pharmacy, Charles University

AIMS: The aim of this study was to compare the clinical decision-making for benzodiazepine deprescribing between a healthcare provider (HCP) and an artificial intelligence (AI) chatbot GPT4 (ChatGPT-4). METHODS: We analysed real-world data from a Croatian cohort of community-dwelling benzodiazepine patients (n = 154) within the EuroAgeism H2020 ESR 7 project. HCPs evaluated the data using pre-established deprescribing criteria to assess benzodiazepine discontinuation potential. The research team devised and tested AI prompts to ensure consistency with HCP judgements. An independent researcher employed ChatGPT-4 with predetermined prompts to simulate clinical decisions for each patient case. Data derived from human-HCP and ChatGPT-4 decisions were compared for agreement rates and Cohen's kappa. RESULTS: Both HPC and ChatGPT identified patients for benzodiazepine deprescribing (96.1% and 89.6%, respectively), showing an agreement rate of 95% (κ = .200, P = .012). Agreement on four deprescribing criteria ranged from 74.7% to 91.3% (lack of indication κ = .352, P < .001; prolonged use κ = .088, P = .280; safety concerns κ = .123, P = .006; incorrect dosage κ = .264, P = .001). Important limitations of GPT-4 responses were identified, including 22.1% ambiguous outputs, generic answers and inaccuracies, posing inappropriate decision-making risks. CONCLUSIONS: While AI-HCP agreement is substantial, sole AI reliance poses a risk for unsuitable clinical decision-making. This study's findings reveal both strengths and areas for enhancement of ChatGPT-4 in the deprescribing recommendations within a real-world sample. Our study underscores the need for additional research on chatbot functionality in patient therapy decision-making, further fostering the advancement of AI for optimal performance.

Citace poskytuje Crossref.org

000      
00000naa a2200000 a 4500
001      
bmc24006987
003      
CZ-PrNML
005      
20240423155629.0
007      
ta
008      
240412s2024 enk f 000 0|eng||
009      
AR
024    7_
$a 10.1111/bcp.15963 $2 doi
035    __
$a (PubMed)37949663
040    __
$a ABA008 $b cze $d ABA008 $e AACR2
041    0_
$a eng
044    __
$a enk
100    1_
$a Bužančić, Iva $u Center for Applied Pharmacy, Faculty of Pharmacy and Biochemistry, University of Zagreb, Zagreb, Croatia $u City Pharmacy Zagreb, Zagreb, Croatia
245    10
$a Clinical decision-making in benzodiazepine deprescribing by healthcare providers vs. AI-assisted approach / $c I. Bužančić, D. Belec, M. Držaić, I. Kummer, J. Brkić, D. Fialová, M. Ortner Hadžiabdić
520    9_
$a AIMS: The aim of this study was to compare the clinical decision-making for benzodiazepine deprescribing between a healthcare provider (HCP) and an artificial intelligence (AI) chatbot GPT4 (ChatGPT-4). METHODS: We analysed real-world data from a Croatian cohort of community-dwelling benzodiazepine patients (n = 154) within the EuroAgeism H2020 ESR 7 project. HCPs evaluated the data using pre-established deprescribing criteria to assess benzodiazepine discontinuation potential. The research team devised and tested AI prompts to ensure consistency with HCP judgements. An independent researcher employed ChatGPT-4 with predetermined prompts to simulate clinical decisions for each patient case. Data derived from human-HCP and ChatGPT-4 decisions were compared for agreement rates and Cohen's kappa. RESULTS: Both HPC and ChatGPT identified patients for benzodiazepine deprescribing (96.1% and 89.6%, respectively), showing an agreement rate of 95% (κ = .200, P = .012). Agreement on four deprescribing criteria ranged from 74.7% to 91.3% (lack of indication κ = .352, P < .001; prolonged use κ = .088, P = .280; safety concerns κ = .123, P = .006; incorrect dosage κ = .264, P = .001). Important limitations of GPT-4 responses were identified, including 22.1% ambiguous outputs, generic answers and inaccuracies, posing inappropriate decision-making risks. CONCLUSIONS: While AI-HCP agreement is substantial, sole AI reliance poses a risk for unsuitable clinical decision-making. This study's findings reveal both strengths and areas for enhancement of ChatGPT-4 in the deprescribing recommendations within a real-world sample. Our study underscores the need for additional research on chatbot functionality in patient therapy decision-making, further fostering the advancement of AI for optimal performance.
650    _2
$a lidé $7 D006801
650    12
$a umělá inteligence $7 D001185
650    _2
$a benzodiazepiny $x škodlivé účinky $7 D001569
650    12
$a depreskripce $7 D000069340
650    _2
$a klinické rozhodování $7 D000066491
650    _2
$a zdravotnický personál $7 D006282
655    _2
$a časopisecké články $7 D016428
655    _2
$a práce podpořená grantem $7 D013485
700    1_
$a Belec, Dora $u Center for Applied Pharmacy, Faculty of Pharmacy and Biochemistry, University of Zagreb, Zagreb, Croatia
700    1_
$a Držaić, Margita $u Center for Applied Pharmacy, Faculty of Pharmacy and Biochemistry, University of Zagreb, Zagreb, Croatia $u City Pharmacy Zagreb, Zagreb, Croatia
700    1_
$a Kummer, Ingrid $u Department of Social and Clinical Pharmacy, Faculty of Pharmacy in Hradec Králové, Charles University, Hradec Králové, Czech Republic
700    1_
$a Brkić, Jovana $u Department of Social and Clinical Pharmacy, Faculty of Pharmacy in Hradec Králové, Charles University, Hradec Králové, Czech Republic $u Department of Social Pharmacy and Pharmaceutical Legislation, Faculty of Pharmacy, University of Belgrade, Belgrade, Serbia
700    1_
$a Fialová, Daniela $u Department of Social and Clinical Pharmacy, Faculty of Pharmacy in Hradec Králové, Charles University, Hradec Králové, Czech Republic $u Department of Geriatrics and Gerontology, 1st Faculty of Medicine in Prague, Charles University, Prague, Czech Republic
700    1_
$a Ortner Hadžiabdić, Maja $u Center for Applied Pharmacy, Faculty of Pharmacy and Biochemistry, University of Zagreb, Zagreb, Croatia $1 https://orcid.org/0000000315789764
773    0_
$w MED00000858 $t British journal of clinical pharmacology $x 1365-2125 $g Roč. 90, č. 3 (2024), s. 662-674
856    41
$u https://pubmed.ncbi.nlm.nih.gov/37949663 $y Pubmed
910    __
$a ABA008 $b sig $c sign $y - $z 0
990    __
$a 20240412 $b ABA008
991    __
$a 20240423155626 $b ABA008
999    __
$a ok $b bmc $g 2081151 $s 1216754
BAS    __
$a 3
BAS    __
$a PreBMC-MEDLINE
BMC    __
$a 2024 $b 90 $c 3 $d 662-674 $e 20231203 $i 1365-2125 $m British journal of clinical pharmacology $n Br J Clin Pharmacol $x MED00000858
GRA    __
$a MSCF-ITN-764632 $p Marie Skłodowska-Curie Foundation
GRA    __
$a CZ.02.1.01/0.0/0.0/18_069/0010046 $p InoMed
GRA    __
$a SVV 260 551 $p European Horizon 2020 I-CARE4OLD
GRA    __
$a START/MED/093 EN.02.2.69/0.0/0.0/19_073/0016935 $p European Horizon 2020 I-CARE4OLD
GRA    __
$a 965341 $p European Horizon 2020 I-CARE4OLD
GRA    __
$p Faculty of Pharmacy, Charles University
LZP    __
$a Pubmed-20240412

Najít záznam

Citační ukazatele

Nahrávání dat ...

Možnosti archivace

Nahrávání dat ...