Skip to main content
. 2018 Apr 23;15(4):828. doi: 10.3390/ijerph15040828

Table 3.

Phase 2: Results of conditional logistic regression predicting career-related outcomes among treatment-seeking and non-treatment-seeking U.S. Marine Corps service members (N = 178) *.

Career-Related Variables Total Treatment-Seeking Control Conditional Model 2
(N = 178) (n = 40) (n = 138) OR (95% CI) p
Record status
  Active record 47 (26.4) 2 (5.0) 45 (32.6) 1.00 -
  Separated 125 (70.2) 38 (95.0) 87 (63.0) 10.16 (2.30–44.91) 0.002
  On orders 5 (2.8) 0 5 (3.6) - 0.986
Separation code 1
  Voluntary 26 (20.8) 13 (34.2) 13 (14.9) 1.00 -
  Involuntary 44 (35.2) 15 (39.5) 29 (33.3) 0.42 (0.13–1.41) 0.159
  Completion of service 55 (44.0) 10 (26.3) 45 (51.7) 0.19 (0.05–0.66) 0.010
Security classification
  Secret/top secret/SCI 131 (73.6) 26 (65.0) 105 (76.1) 1.00 -
  Denied 2 (1.1) 1 (2.5) 1 (0.7) 5.11 (0.30–87.42) 0.261
  No status 39 (21.9) 10 (25.0) 29 (21.0) 1.50 (0.61–3.70) 0.374
  Revoked/loss of jurisdiction 6 (3.4) 3 (7.5) 3 (2.2) 3.79 (0.71–20.29) 0.120
Legal action
  No 127 (71.4) 22 (55.0) 105 (76.1) 1.00 -
  Yes 51 (58.7) 18 (45.0) 33 (23.9) 2.84 (1.32–6.24) 0.008
  Average time to separation, mean (SD), years 3.3 (2.2) 1.5 (1.5) 3.8 (2.1) - <0.001

* Data presented as No. (%), except as noted. 1 Due to unreported data, some variables will not have the full sample of 178 participants. 2 Logistic regression results where a cluster variable was created to account for the matched case: control study design. Note. SCI = sensitive compartmented information.