RESUMO
This paper reviews dilemmas and implications of erroneous data for clinical implementation of AI. It is well-known that if erroneous and biased data are used to train AI, there is a risk of systematic error. However, even perfectly trained AI applications can produce faulty outputs if fed with erroneous inputs. To counter such problems, we suggest 3 steps: (1) AI should focus on data of the highest quality, in essence paraclinical data and digital images, (2) patients should be granted simple access to the input data that feed the AI, and granted a right to request changes to erroneous data, and (3) automated high-throughput methods for error-correction should be implemented in domains with faulty data when possible. Also, we conclude that erroneous data is a reality even for highly reputable Danish data sources, and thus, legal framework for the correction of errors is universally needed.
RESUMO
BACKGROUND: A major challenge to psychological treatment for alcohol use disorder (AUD) is patient non-compliance. A promising new treatment approach that is hypothesized to increase patient compliance is blended treatment, consisting of face-to-face contact with a therapist combined with modules delivered over the internet within the same protocol. While this treatment concept has been developed and proven effective for a variety of mental disorders, it has not yet been examined for AUD. AIMS: The study described in this protocol aims to examine and evaluate patient compliance with blended AUD treatment as well as the clinical and cost effectiveness of such treatment compared to face-to-face treatment only. METHODS: The study design is a pragmatic, stepped-wedge cluster randomized controlled trial. The included outpatient institutions (planned number of patients: n = 1800) will be randomized in clusters to implement either blended AUD treatment or face-to-face treatment only, i.e. treatment as usual (TAU). Both treatment approaches consist of motivational interviewing and cognitive behavioral therapy. Data on sociodemographics, treatment (e.g. intensity, duration), type of treatment conclusion (compliance vs. dropout), alcohol consumption, addiction severity, consequences of drinking, and quality of life, will be collected at treatment entry, at treatment conclusion, and 6 months after treatment conclusion. The primary outcome is compliance at treatment conclusion, and the secondary outcomes include alcohol consumption and quality of life at six-months follow-up. Data will be analyzed with an Intention-to-treat approach by means of generalized linear mixed models with a random effect for cluster and fixed effect for each step. Also, analyses evaluating cost-effectiveness will be conducted. DISCUSSION: Blended treatment may increase treatment compliance and thus improve treatment outcomes due to increased flexibility of the treatment course. Since this study is conducted within an implementation framework it can easily be scaled up, and when successful, blended treatment has the potential to become an alternative offer in many outpatient clinics nationwide and internationally. TRIAL REGISTRATION: Clinicaltrials.gov .: NCT04535258 , retrospectively registered 01.09.20.
