Table 3.
Progression of mixed‐effects models predicting global EDE‐Q scores over treatment by sexual/gender minority status
| Model 1 (N = 2818) | Model 2 (2818) | Model 3 (N = 2813) | Model 4a(N = 2797) | |||||
|---|---|---|---|---|---|---|---|---|
| γ (SE) | df | γ (SE) | df | γ (SE) | df | γ (SE) | df | |
| Fixed effects | ||||||||
| Model for initial status (intake EDE‐Q score) | ||||||||
| Intercept b | 2.887 (0.027) *** | 2597 | 3.756 (0.037) *** | 2556 | 3.684 (0.041) *** | 2554 | 4.026 (0.086) *** | 2527 |
| Initial level of care—PHP | −0.333 (0.097) *** | 2345 | −0.290 (0.086) ** | 2564 | −0.309 (0.086) *** | 2560 | −0.439 (0.084) *** | 2522 |
| Initial level of care—IOP | −0.341 (0.073) ** | 3171 | −0.576 (0.106) *** | 2560 | −0.555 (0.106) *** | 2556 | −0.783 (0.109) *** | 2529 |
| Initial level of care—RTC (reference) | ‐ | ‐ | ||||||
| Sexual/gender minority | 0.422 (0.085) *** | 2563 | ||||||
| Unsure of sexual orientation | 0.002 (0.127) | 2532 | ||||||
| Cisgender heterosexual (reference) | ‐ | ‐ | ||||||
| Covariates (and gender moderation) | ||||||||
| Sexual/gender minority male | 0.195 (0.302) | 2506 | ||||||
| Sexual/gender minority female | 0.223 (0.086) ** | 2526 | ||||||
| Unsure of sexual orientation male | −0.456 (0.398) | 2467 | ||||||
| Unsure of sexual orientation female | 0.020 (0.129) | 2499 | ||||||
| Cisgender heterosexual male | −1.156 (0.161) *** | 2507 | ||||||
| Cisgender heterosexual female (reference) | ‐ | ‐ | ||||||
| No history of sexual abuse | −0.275 (0.077) *** | 2515 | ||||||
| No history of other trauma | −0.052 (0.071) | 2522 | ||||||
| No history of bullying | −0.262 (0.065) *** | 2516 | ||||||
| Age 19 and older | 0.238 (0.081) ** | 2586 | ||||||
| Months since eating disorder onsetc | 0.073 (0.041) | 2740 | ||||||
| No prior eating disorder treatment | 0.121 (0.078) | 2513 | ||||||
| Intake body mass indexc | 0.130 (0.033) *** | 2609 | ||||||
| Model for rate of acceleration (initial slope) | df | df | df | |||||
| Intercept d | 1.387 (0.053) *** | 1303 | 1.344 (0.059) *** | 1329 | 1.603 (0.126) *** | 1383 | ||
| Initial level of care—PHP | ‐0.401 (0.113) *** | 1457 | −0.417 (0.113) *** | 1452 | −0.494 (0.114) *** | 1430 | ||
| Initial level of care—IOP | −0.519 (0.236) * | 2294 | −0.522 (0.237) * | 2281 | −0.581 (0.239) * | 2194 | ||
| Initial level of care—RTC (reference) | ‐ | ‐ | ‐ | ‐ | ||||
| Sexual/gender minority | 0.316 (0.118) ** | 1492 | ||||||
| Unsure of sexual orientation | −0.174 (0.182) | 1249 | ||||||
| Cisgender heterosexual (reference) | ‐ | ‐ | ||||||
| Covariates (and gender moderation) | ||||||||
| Sexual/gender minority male | −0.206 (0.435) | 1423 | ||||||
| Sexual/gender minority female | 0.200 (0.124) | 1466 | ||||||
| Unsure of sexual orientation male | −0.100 (0.591) | 1168 | ||||||
| Unsure of sexual orientation female | −0.187 (0.193) | 1223 | ||||||
| Cisgender heterosexual male | −0.538 (0.248) * | 1344 | ||||||
| Cisgender heterosexual female (reference) | ‐ | ‐ | ||||||
| No history of sexual abuse | −0.160 (0.112) | 1412 | ||||||
| No history of other trauma | −0.234 (0.104) * | 1403 | ||||||
| No history of bullying | −0.102 (0.095) | 1389 | ||||||
| Age 19 and older | 0.091 (0.118) | 1498 | ||||||
| Months since eating disorder onsetc | 0.007 (0.061) | 1779 | ||||||
| No prior eating disorder treatment | 0.241 (0.113) * | 1427 | ||||||
| Intake body mass indexc | 0.067 (0.051) | 1943 | ||||||
| Model for rate of change (phase 2 slope) | df | df | df | |||||
| Intercept e | −0.004 (0.001) *** | 523 | −0.004 (0.001) *** | 545 | −0.005 (0.002) ** | 583 | ||
| Initial level of care—PHP | −0.005 (0.001) *** | 633 | −0.005 (0.001) *** | 633 | −0.005 (0.001) *** | 615 | ||
| Initial level of care—IOP | −0.004 (0.003) | 1092 | −0.005 (0.003) *** | 1092 | −0.003 (0.003) | 1021 | ||
| Initial level of care—RTC (Reference) | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ||
| Sexual/gender minority | 0.000 (0.002) | 641 | ||||||
| Unsure of sexual orientation | −0.003 (0.003)* | 476 | ||||||
| Cisgender heterosexual (reference) | ‐ | ‐ | ||||||
| Covariates (and gender moderation) | ||||||||
| Sexual/gender minority male | −0.006 (0.006) | 761 | ||||||
| Sexual/gender minority female | 0.001 (0.002) | 610 | ||||||
| Unsure of sexual orientation male | −0.002 (0.007) | 397 | ||||||
| Unsure of sexual orientation female | −0.001 (0.003) | 466 | ||||||
| Cisgender heterosexual male | 0.001 (0.004) | 578 | ||||||
| Cisgender heterosexual female (reference) | ‐ | ‐ | ||||||
| No history of sexual abuse | 0.002 (0.002) | 599 | ||||||
| No history of other trauma | −0.001 (0.002) | 590 | ||||||
| No history of bullying | 0.001 (0.001) | 572 | ||||||
| Age 19 and older | 0.002 (0.002) | 618 | ||||||
| Months since eating disorder onsetc | 0.000 (0.001) | 722 | ||||||
| No prior eating disorder treatment | −0.004 (0.002) ** | 604 | ||||||
| Intake body mass indexc | 0.000 (0.001) | 1346 | ||||||
| Random effects | ||||||||
| Level 1 | σ 2 (SE) | σ 2 (SE) | σ 2 (SE) | σ 2 (SE) | ||||
| Within‐person | 1.833(0.043) *** | 0.213 (0.012) *** | 0.213 (0.012) *** | 0.214 (0.012) *** | ||||
| Level 2 | τ (SE) | τ (SE) | τ (SE) | τ (SE) | ||||
| Initial status | 1.121 (0.057) *** | 2.344 (0.073) *** | 2.322 (0.073) *** | 2.126 (0.068) *** | ||||
| Rate of acceleration (slope 1) | 2.303 (0.133) *** | 2.289 (0.133) *** | 2.239 (0.132) *** | |||||
| Rate of change (slope 2) | 0.0002 (0.000) *** | 0.0002 (0.000) *** | 0.0002 (0.000) *** | |||||
| Goodness‐of‐fit statistics | ||||||||
| −2 Log likelihood | 23185.8 | 20145.9 | 20136.0 | 19876.7 | ||||
| Akaike's information criterion | 23189.8 | 20159.9 | 20150.0 | 19890.7 | ||||
Note: Model 1 represents the unconditional means model (with entry level‐of‐care controlled to account for sample structure), Model 2 represents the unconditional growth model (with entry level‐of‐care controlled to account for sample structure), Model 3 tests hypotheses 2a, 2b, and 3, Model 4 is confounder‐adjusted.
Abbreviations. ED, eating disorder; EDE‐Q, Eating Disorder Examination Questionnaire; IOP, intensive outpatient program; PHP, partial hospitalization program; Ref, reference; RTC, residential treatment center.
Model 4 intercepts represent an individual who scored at the mean on the continuous variables and had the following specifications on the categorical variables: cisgender heterosexual, female, age < 19 years, had prior ED treatment, “yes” to history of sexual abuse, trauma, and bullying.
Interpretation of Intercept for Initial Status Model is the mean admission EDE‐Q score for the patient who entered at the residential level‐of‐care.
Standardized variables.
Interpretation of Intercept for Rate of Acceleration Model is the average initial improvement for the patient entering the residential level‐of‐care.
Interpretation of Intercept for Rate of Change Model is the average improvement by discharge from final level‐of‐care for the patient entering residential treatment.
p < .001.
p < .01.
p < .05.