Development and Validation of Automated Indicators on the Appropriateness of Oral Anticoagulant Prescriptions

Development and Validation of Indicators on the Appropriateness of Oral Anticoagulant Prescriptions in Adult Medicine Automated From the Hospital Information System (PACHA Study).

Status

Completed

Phases

Unknown

Study type

Observational

Source

ClinicalTrials.gov

Registry ID

NCT02898090

Acronym

PACHA

Enrollment

1213

Registered

2016-09-13

Start date

2018-04-28

Completion date

2022-02-01

Last updated

2024-03-27

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Anticoagulant Drugs, Atrial Fibrillation, Brain Injury, Healthcare Quality Indicators, Inappropriate Prescribing, Valvular Heart Disease, Venous Thromboembolism

Keywords

Quality indicators, Appropriateness, Oral anticoagulant prescriptions, Development and validation, Hospital information system, Information Storage and Retrieval, Natural Language Processing

Brief summary

This research focuses on the development and validation of indicators on the appropriateness of oral anticoagulant prescriptions. The investigators want to propose transferable tools to other healthcare institutions to allow automated construction of indicators as part of a structured approach to improve future practices. The main objective of the study is to develop indicators on the appropriateness of oral anticoagulant prescriptions in adult medicine automated from the hospital information system and to assess their criterion validity.

Detailed description

The appropriateness of oral anticoagulant prescriptions (Vitamin K Antagonists and Direct Oral Anticoagulants) is a major challenge for improving quality, safety and efficiency of care. The large targeted population, prescription frequency and strong risk of severe adverse events, especially in healthcare institutions, justify the importance of developing indicators on the appropriateness of oral anticoagulant prescriptions and their automated implementation from the hospital information system. There is currently a need to develop and validate such indicators for integrating them into the hospital clinical practice, as part of a structured approach to improve quality and safety of care. As each hospital information system is different, the investigators want to propose transposable tools to other healthcare institutions to allow an automated construction of these indicators. The study will consist of three main steps: 1) identification of indicators on the appropriateness of oral anticoagulant prescriptions and their conditions of appropriateness (literature review), and analysis of their potential utility and operational characteristic (Delphi consensus); 2) operational implementation of the selected indicators from the hospital information system, using tools to generalize their implementation to other hospital information systems; 3) evaluation of metrological performance and robustness of the selected indicators. This is an observation of prescriptions after it has occurred without any modification or change of treatment to patients. This study is never Interventional. Statistical analysis : for each indicator, the criterion validity will be analyzed thanks to prevalence of appropriateness (or not) of oral anticoagulant prescriptions estimated by the hospital information system, sensitivity, specificity, positive and negative predictive values, and area under the ROC curve.

Interventions

None listed

Study design

Observational model

COHORT

Time perspective

RETROSPECTIVE

Eligibility

Sex/Gender

ALL

Age

18 Years to No maximum

Healthy volunteers

Inclusion criteria

* Patients aged over 18; * Patients hospitalized at the Bordeaux University Hospital or the Georges-Pompidou European hospital for a clinical situation that could justify a prescription of oral anticoagulants or for an at risk situation in patients under oral anticoagulants, among targeted frequent clinical situations; * Patients hospitalized in medicine-surgery-obstetrics care or rehabilitation care (conventional hospitalization or outpatient) between the 1st of January 2015 and the 31th of December 2017; * Patients whose medical and administrative data are available in an electronic format in the hospital information system

Exclusion criteria

* Patients not hospitalized at the Bordeaux University Hospital or the Georges-Pompidou European hospital for a clinical situation that could justify a prescription of oral anticoagulants or for an at risk situation in patients under oral anticoagulants, among targeted frequent clinical situations; * Patients whose medical and administrative data are not available in an electronic format in the hospital information system; * Patients who do not give consent for research data use.

Design outcomes

Primary

Measure	Time frame	Description
Difference between the prevalence of appropriateness (or not) of oral anticoagulant prescriptions	between month 24 and 33	Difference between the prevalence of appropriateness (or not) of oral anticoagulant prescriptions estimated by the hospital information system in comparison to the prevalence estimated by the reference test (information manually extracted from the patient record) for each indicator, and the estimation of its confidence interval for an equivalence margin between 5% and 10%.

Secondary

Measure	Time frame	Description
The recall, precision and F-measure of tools tracking information for the construction of indicators (step 2)	between month 16 and 18	Recall, precision and F-measure are criteria for measuring performance of tools tracking information. The recall is defined by the number of relevant concepts that are correctly found by the evaluated automatic tool compared to the number of relevant concepts that are present in the text (in statistics, the recall is the sensitivity). Precision is the number of relevant concepts found on the total number of concepts proposed by the search engine for a given query. The F-measure combines recall and precision as follows: F-measure = 2 x recall x precision / (recall + precision).
The ability to use the autonomous version of automatic language processing module (Web-Service)	between month 30 and 36	to provide the same indicators with another data storage (step 2) : extraction of the concept in full text from the Rennes university hospital (yes or no)
The estimation of the metrological performances of each indicator	between month 34 and 36	sensitivity, specificity, positive and negative predictive values to identify the appropriateness (or not) of oral anticoagulant prescriptions compared to the reference test and opinion of experts with knowledge about the targeted clinical situations (step 3)
The estimation of the reliability of each indicator	between month 34 and 36	Kappa coefficients for unordered qualitative indicators, weighted Kappa coefficients for ordained qualitative indicators and Intra-Class Coefficients for quantitative indicators for the analysis of changes in the appropriateness results of indicators considering variation of the conditions and quality of encoding information in the hospital information system (step 3);
The estimation of the robustness of each indicator	between month 34 and 36	Kappa coefficients for unordered qualitative indicators, weighted Kappa coefficients for ordained qualitative indicators and intra-class coefficients for quantitative indicators for the analysis of changes in the appropriateness results of indicators considering variation of hypotheses for the construction of these indicators (step 3)

Countries

France

Outcome results