Table 2.
The Ottawa weight loss prediction model (OWL-PM).
| Variable | Parameter estimate (standard error) | P-value | Independent mean change in % relative weight difference† at program's end due to factor |
|---|---|---|---|
| Intercept | 1.848 (0.17) | < .0001 | – |
| Program week increased by 1 | − 0.762 (0.04) | < .0001 | − 9.65 |
| (Program week increased by 1)2 | 0.02 (0.003) | < .0001 | |
| Female | − 0.224 (0.116) | 0.0531 | +1.71 |
| Female*Week | 0.302 (0.027) | < .0001 | |
| Female*Week 2 | − 0.012 (0.002) | < .0001 | |
| Baseline WeightA increased by 25lbs | − 0.007 (0.001) | < .0001 | +0.41 |
| (Baseline WeightA increased by 25lbs)*Week | 0.001 (0) | < .0001 | |
| (Baseline WeightA increased by 25lbs)*Week2 | − 0.00003 (0) | 0.0594 | |
| AgeB increased by 10 years | − 0.011 (0.004) | 0.0089 | +0.15 |
| (AgeB increased by 10 years)*Week | 0.002 (0.001) | 0.0021 | |
| TSHC increased by 1 IU | 0.058 (0.042) | 0.1645 | +0.08 |
| (TSHC increased by 1 IU)*Week | 0.024 (0.01) | 0.0156 | |
| (TSHC increased by 1 IU)*Week2 | − 0.002 (0.001) | 0.0074 | |
| # of ATPIII MS Criteria increased by 1 | − 0.027 (0.045) | 0.5566 | − 0.28 |
| (# of ATPIII MS Criteria increased by 1)*Week | − 0.039 (0.011) | 0.0003 | |
| (# of ATPIII MS Criteria increased by 1)*Week2 | 0.001 (0.001) | 0.0292 | |
| Smoking status | |||
| Never | REF | – | – |
| Never*Week | REF | – | |
| Never*Week2 | REF | – | |
| Previous | 0.006 (0.093) | 0.9871 | − 0.68 |
| Previous*Week | − 0.042 (0.022) | 0.0718 | |
| Previous*Week2 | − 0.001 (0.001) | 0.0525 | |
| Current | − 0.02 (0.159) | 0.9871 | − 0.04 |
| Current*Week | − 0.066 (0.038) | 0.0718 | |
| vCurrent*Week2 | 0.005 (0.002) | 0.0525 | |
| Diabetes status | |||
| None | REF | – | – |
| None*Week | REF | – | |
| None*Week2 | REF | – | |
| Diet controlled | − 0.038 (0.122) | 0.6608 | +0.16 |
| Diet controlled*Week | 0.044 (0.029) | 0.1545 | |
| Diet controlled*Week2 | − 0.002 (0.002) | 0.0161 | |
| Weight neutral meds only | 0.199 (0.212) | 0.6608 | +1.86 |
| Weight neutral meds only*Week | 0.101 (0.05) | 0.1545 | |
| Weight neutral meds only*Week2 | 0.002 (0.003) | 0.0161 | |
| Non-Weight neutral meds | 0.111 (0.167) | 0.6608 | +1.43 |
| Non-Weight neutral meds*Week | 0.022 (0.039) | 0.1545 | |
| Non-Weight neutral meds*Week2 | 0.006 (0.002) | 0.0161 | |
| Obstructive sleep apnea status | |||
| None | REF | – | – |
| None*Week | REF | – | |
| OSA, on CPAP | − 0.045 (0.125) | 0.9353 | − 0.80 |
| OSA, on CPAP*Week | − 0.058 (0.019) | 0.0052 | |
| OSA, not on CPAP | − 0.019 (0.143) | 0.9353 | − 0.58 |
| OSA, not on CPAP*Week | − 0.043 (0.022) | 0.0052 | |
†Compared to relative weight difference predicted by NIH-BWP.
The OWL-PM is a longitudinal analysis model that predicts the relative difference between observed weights and those predicted by the NIH-BWP. The expected percent relative weight difference (Eq. 1) from the OWL-PM is calculated by multiplying patient-specific variable values by the corresponding parameter estimate. This value can be back-transformed to a predicted weight using Eq. 2
ACentered on 270lbs BCentered on 45 years CCentered on 2.2 IU.
(lbs = pounds; TSH = thyroid stimulating hormone; IU = international units; ATPIII = 3rd adult treatment panel; MS = metabolic syndrome; OSA = obstructive sleep apnea).