Introduction First-term attrition (FTA), or failure of a military service member to complete their initial service contract, is a major financial burden and source of lost manpower in the US Navy. The objective of the present study was to examine medical correlates of FTA using healthcare and disability rating data.

Methods In this retrospective cohort study, all US Navy-enlisted personnel between the years 2003 and 2018 with FTA (n=58 777) and regular discharge (n=203 084) were identified for analysis from accession dates in the Career History Archival Medical and Personnel System. Medical diagnoses from outpatient and inpatient records were abstracted from the Military Health System Data Repository. For a subgroup of the study population discharged with a disability rating (n=12 880), diagnoses were identified from the Integrated Disability Evaluation System. The FTA and regular discharge groups were compared using relative risks (RRs) and 95% CIs, and per cent differences for the disability subgroup analysis.

Results Compared with regular discharges, those with FTA were more likely to have outpatient and inpatient diagnoses for mental health disorders. Personality disorder yielded the strongest association with FTA in both outpatient (RR=10.45, 95% CI 9.79 to 11.16) and inpatient settings (RR=18.97, 95% CI 14.16 to 25.42). Other disorders associated with FTA included schizophrenia, substance-related disorders, poisoning by psychotropic agents and adjustment disorders. In the disability analysis, the FTA group relative to regular discharges had the largest per cent differences for ‘arthritis, degenerative (hypertrophic or osteoarthritis)’ (10.8% vs 2.5%) and ‘tibia and fibula, impairment’ (3.0% vs 0.4%).

Conclusions This study provides evidence that FTA is associated with both mental and physical health conditions. Mental and physical factors related to FTA require further examination, particularly whether pre-enlistment screening or early career intervention could lead to mitigation strategies. Future research should extend this analysis to other services and population subgroups.