In a prospective cohort study of 583 older adults hospitalized with acute myocardial infarction, machine learning identified 4 frailty trajectories over 1 year — including a rapidly worsening group affecting 13.4%. Older age and lower functional status were the strongest predictors of rapidly worsening frailty, while anxiety and Depression predicted a slowly progressive course.

Why It Matters To Your Practice

• Frailty after acute myocardial infarction is not static: patients followed markedly different recovery or decline paths over time.

• The highest-risk group — rapid worsening frailty — may be identifiable early, creating an opportunity for closer monitoring and targeted intervention.

• Psychological symptoms were not peripheral findings: anxiety and Depression were linked to slowly progressive frailty.

Clinical Implications

• Assess baseline function early, especially activities of daily living, because lower Modified Barthel Index scores strongly predicted rapid frailty worsening.

• Use age, comorbidity burden, nutrition, cardiac function, and psychological status to risk-stratify older adults after acute myocardial infarction.

• Consider follow-up pathways that integrate rehabilitation, nutrition support, and mental health screening rather than focusing only on cardiac recovery.

Insights

• The study enrolled 583 older adults between March 2023 and March 2024 and collected multidimensional data at 6 time points across 1 year.

• A patient similarity network plus Structural Entropy Clustering separated patients into 4 groups: Stable Non-Frail (44.7%), Slowly Progressive Frailty (37.4%), Rapidly Worsening Frailty (13.4%), and Improving Frailty (4.5%).

• Group differences were seen in functional status, psychological scores, nutritional status, left ventricular ejection fraction, and Charlson Comorbidity Index.

The Bottom Line

• AI-style clustering can help clinicians move beyond one-time frailty labels and identify which older post-MI patients are most likely to decline.

• For practice now: prioritize early functional assessment and watch for psychological distress, especially when planning post-discharge care.

• The findings support differentiated nursing and follow-up strategies, but multicenter validation is still needed.