Table 2.

Fit indices for various growth mixture models. The number of classes denotes the number of fatigue trajectory subgroups being estimated based on the individual trajectories of patients on the SF-36 VT. For example, the 1-class solution implies that a single mean growth trajectory fits all individual trajectories, whereas the 2-class solution assumes that patients with early OA can be divided into 2 subgroups based on their fatigue development. For each model, all classes contained linear and quadratic slopes to assess nonlinear fatigue development, if present.

ModelsGroup Size, nBICSSABICLMR LRTBLRTEntropy
1 class100048,555.9648,505.14———
2 class517/48348,143.1448,057.39493.82**498.93**0.534
3 class212/335/45348,110.3347,980.11128.20*129.52**0.492
3 class revised213/330/45748,090.2147,975.87133.81*135.19**0.492

  • ↵* p < 0.05.

  • ↵** p < 0.01. SF-36 VT: Medical Outcomes Study Short Form-36 Vitality scale; OA: osteoarthritis; BIC: Bayesian Information Criterion; SSABIC: sample size-adjusted BIC; LMR LRT: Lo-Mendell-Rubin adjusted likelihood ratio test; BLRT: bootstrapped likelihood ratio test.