Outcomes During a One-Year Multidisciplinary Lifestyle Program in Japanese Specialty Care: A Retrospective...

Multidisciplinary lifestyle care in routine practice occurs alongside medication changes and irregular testing. We described outcomes during a one-year high-contact specialty-clinic program, separating average post-baseline differences from approximately 12-month change.

This retrospective single-center study included 170 adults with hypertension, dyslipidemia, or diabetes. The primary estimand was the adjusted average difference between baseline and any available post-baseline measurement. Separate linear mixed models included a participant random intercept and fixed terms for follow-up status, age at measurement, sex, and time-varying antidiabetic, antihypertensive, and lipid-lowering medication classes. A complete-pair analysis used the measurement nearest 12 months within a 10-14-month window. Outcome-specific missingness was reported. A Bonferroni benchmark of P<0.0028 was applied across 18 mixed-model outcomes; paired P values were adjusted across all 18 prespecified outcomes.

Mean age was 63.0 years (SD 12.6), and 92 participants (54.1%) were women. Body-composition data were nearly complete, whereas physical-function data were available for a selected subset. Differences meeting the multiplicity benchmark were body weight (-2.11 kg, 95% CI -2.68 to -1.54), body fat percentage (-2.04 percentage points, -2.53 to -1.56), waist circumference (-2.72 cm, -3.43 to -2.01), body mass index (-0.90 kg/m², -1.13 to -0.68), 1-min walk (+7.50 steps, 4.51 to 10.49), chair stand (+2.27 repetitions, 1.47 to 3.07), one-leg stand (+7.19 s, 3.37 to 11.00), and low-density lipoprotein cholesterol (-6.44 mg/dL, -10.49 to -2.39). The 10-14-month complete-pair analysis was directionally consistent.

During this single-arm program, adiposity measures changed modestly and lower-extremity performance showed larger standardized differences. No comparator, outcome-specific missingness, unknown attendance, and medication-measurement limitations preclude causal attribution.