Finnish AI tool uses six-question pulse to predict sick leave

A Finnish dissertation and a new workplace pulse tool both point to the same shift: employers can spot sick-leave risk earlier by measuring employee experience, not health data. Novogain AI says its six-question QWL pulse can forecast team-level absence risk monthly and has already shown early results in social and health care.

Why it matters: - Employers are under pressure to move from reacting to sick leave to preventing it. - The Finnish research and the Novogain AI tool both suggest that short, repeated employee surveys can surface psychosocial risk early. - The approach is designed to work without using employee health data, which can make it easier to deploy in routine management. - The timing aligns with the European Parliament’s push for stronger prevention of psychosocial risks at work.

What happened: - A Finnish dissertation found that employee experience of work conditions contains important information about future sick-leave risk. - Anniina Anttila’s 2026 dissertation at Tampere University analyzed survey responses, occupational health data and sick leave records from nearly 12,000 employees. - Novogain AI, developed in Finland, says its six-question QWL pulse can predict sick leave for each team on a monthly basis. - The company says the tool does not require employee health data. - Novogain AI is available for one euro per employee per measurement round. - The company’s developer, Marko Kesti, said employee risk can be hidden in response variation and that listening to every employee matters.

The details: - The European Parliament approved a legislative initiative on Oct. 6, 2026, to strengthen prevention of psychosocial risks. - The Parliament voted 413–210. - The resolution calls for regular risk assessment, employee participation and monitoring of preventive measures. - Committee preparations also called for scientifically validated assessment methods and early-detection mechanisms. - The Parliament’s initiative is not yet a binding directive. - Anttila’s study found that employees’ sense of influence over job content, work pace and working hours predicted recurring short sick leaves. - That single question was more informative for predicting short absences than individual depression-symptom items. - The study also combined survey and registry data in machine-learning models. - The best combined model reached an AUROC of 0.79 for predicting long sick leaves, indicating strong ability to separate higher- and lower-risk individuals. - The dissertation used an 11-page questionnaire, which Anttila said was burdensome to complete. - Anttila said shorter questionnaires could be used more often, helping organizations detect rising work-capacity risks earlier. - The dissertation did not study Novogain AI, but its findings support developing predictive measurement based on employee experience. - Novogain AI’s QWL pulse measures physical and emotional safety, belonging and identity, plus purpose and creativity. - The six-question tool showed very good internal consistency in a validation sample of 699 respondents, with Cronbach’s alpha of 0.898. - The validation dataset for QWL-based sick-leave prediction included 10,322 respondents and 85 organizational units. - The correlation between the forecast and the actual sick-leave rate was 0.427. - The model’s explained variance was 0.18, which means predictive power was statistically significant but limited at group level. - The forecast can be calibrated with an organization’s own sick-leave data if needed. - The pulse can be run monthly or every two months to give managers an advance estimate of team absence risk. - Automated analysis and AI-assisted development recommendations reduce manual analysis work.

Between the lines: - The core shift is from using long, occasional surveys to using shorter pulses that can be repeated more often. - The tool’s value depends less on perfectly predicting individual absences and more on spotting team-level risk early enough to act. - Novogain’s focus on answer dispersion, not just averages, reflects a key workplace reality: a team can look healthy on paper while one employee is struggling badly. - The company’s “Zero work ill-being” traffic-light model is meant to trigger confidential one-to-one workload discussions rather than identify individual respondents. - A strong business case runs through the release: if a low-cost pulse prevents even one sick-leave day, the measurement may pay for itself. - The reported Treili results are promising, but the release notes that observational follow-up does not prove how much of the change came from Novogain’s use.

What’s next: - Treili, a social and health care organization, adopted Novogain AI in January 2026 and started team-level sick-leave forecasting in March. - Early follow-up data showed sick-leave days per employee fell 9.2% in January–June compared with the same period in 2025. - That equals about one fewer absence day per employee. - In March–June, total sick-leave days fell 21% versus the prior-year period. - The release says the decline came despite an unusually strong flu wave in Finland in February. - Novogain says the goal is to make preventive work-capacity management continuous, effective and financially sustainable. - The company positions the tool as a way to help managers use early warning data to take the right action at the right time.

The bottom line: - Finnish research and a commercial pilot point toward the same conclusion: shorter employee pulses may help organizations spot sick-leave risk earlier, before costs and absences pile up.