# How does the implementation of AI-based automation of administrative tasks affect healthcare professionals’ work? Study protocol for a longitudinal embedded case study in Swedish primary and specialist care

## TL;DR

This study explores how AI-based automation of administrative tasks in healthcare affects professionals' work in Swedish care facilities over 18 months.

## Contribution

The study introduces a longitudinal embedded case design to examine AI's impact on professional boundaries and work environments in healthcare.

## Key findings

- AI implementation may alter professional boundaries and work identities in healthcare.
- Longitudinal data will reveal how AI adoption affects administrative efficiency and job satisfaction.
- Mixed-methods analysis will identify barriers and facilitators to AI integration in clinical settings.

## Abstract

Artificial intelligence (AI) is transforming healthcare through enhanced computational capabilities that process vast amounts of data with unprecedented speed and precision. AI-powered administrative systems hold significant promise for reducing the substantial documentation burden on healthcare professionals, potentially improving operational efficiency and job satisfaction. However, critical knowledge gaps exist regarding how AI implementation affects professional boundaries, work identities and the healthcare work environment. This study protocol describes the Project EFAAI (Efficient and Flexible Working Life in Healthcare with the Support of Digital Tools and AI), which aims to explore how the implementation of AI-based automation of administrative tasks affects healthcare professionals’ work.

This research employs a single-embedded case study with a longitudinal approach, following the implementation of an AI-powered healthcare administration system across 12 primary care facilities and 20 specialist care facilities of a major private healthcare provider in Sweden. Data collection spans 18 months through three sequential waves at 6-month intervals. The project consists of three work packages (WP): WP1 explores changes in professional boundaries and AI literacy requirements using exploratory meetings with healthcare leaders and in-depth semistructured interviews with 20 physicians from primary care and 20 from specialist care (n=40); WP2 investigates implementation barriers and facilitators through in-depth semistructured interviews with the same 40 physicians as WP1 for longitudinal follow-up; and WP3 examines the impact on efficient and flexible work conditions using a mixed-methods approach integrating quantitative operational and administrative data from the healthcare provider’s information systems and administrative databases (eg, documentation completion times, consultation duration, workload distribution metrics, operational efficiency ratios) with qualitative data from four focus groups and workshops with physicians, nurses, healthcare leaders and administrative staff.

The study adheres to the ethical guidelines of the Swedish Research Council and complies with the regulations of the Swedish Ethical Review Authority. Data will be collected with informed consent from all participants, anonymised and stored in accordance with General Data Protection Regulation (GDPR) requirements. Results will be disseminated through international publications, conferences, teaching activities and stakeholder engagement.