Characteristics of Weight Loss Trajectories in a Comprehensive Lifestyle Intervention

Ryan J. Marek, Sandra M. Coulon, Joshua D. Brown, Janet A. Lydecker, Scott Marek, Robert Malcolm, Patrick M. O'Neil

Research output: Contribution to journalArticlepeer-review

12 Scopus citations


Objective: Focusing on average weight loss (WL) from interventions provides useful efficacy data but masks large variability across patients. In this study, parameters of weight loss trajectories were determined that differentiated individuals during a 15-week clinical lifestyle intervention. Methods: Patients (n = 595) were in a fee-for-service WL lifestyle program with a partial meal replacement diet and lifestyle change counseling. Parameters used in latent class analyses were percent WL (%WL), weight nadir, number of weekly weight gains, maximum weekly percent weight gain, standard deviation of weekly weight changes, linear slope values, and change in slope. Results: Average %WL was 9.73%. Latent class analyses revealed three groups with considerable overlap in %WL ranges but differing significantly on all trajectory parameters (Ps < 0.001). Group 1 had the most variable and least successful pattern of weight changes. Group 3 had the least variable and most successful pattern of weight changes. Group 2 fell between the others on all parameters. Conclusions: Emphasis on average WL likely obscures considerable variability in individual courses of weight change. Moreover, patients with similar %WL can have different WL trajectories. Identification of behavioral/physiological characteristics associated with different weight loss trajectories may facilitate the development of more tailored interventions, particularly for trajectories associated with less optimal outcomes.

Original languageEnglish
Pages (from-to)2062-2067
Number of pages6
Issue number12
StatePublished - Dec 2017


Dive into the research topics of 'Characteristics of Weight Loss Trajectories in a Comprehensive Lifestyle Intervention'. Together they form a unique fingerprint.

Cite this