Prevalence and predictors of ambulatory care physicians' documentation of mobility limitations in older adults.

OBJECTIVE: To determine how often physicians document mobility limitations in visits with older adults, and which patient, physician, and practice characteristics associate with documented mobility limitations.

DESIGN: We completed a cross-sectional analysis of National Ambulatory Medical Care Surveys, years 2012-2016. Multivariate analyses were conducted to identify patient, physician, and practice-level factors associated with mobility limitation documentation.

SETTING: Ambulatory care visits.

PARTICIPANTS: We analyzed visits with adults 65 years and older. Final sample size represented 1.3 billion weighted visits.

INTERVENTION: N/A MAIN OUTCOME MEASURE: We defined the presence/absence of a mobility limitation by whether any ICD-9 or ICD-10 code related to mobility limitations were documented in the visits.

RESULTS: The overall prevalence of mobility limitation documentation was 2.4%. The most common codes were falls-related. Patient-level factors more likely to be associated with mobility limitation documentation were visits by individuals over 85 years of age, relative to 65-69 years, (OR 2.32, 95% CI 1.76-3.07]; with a comorbid diagnosis of arthritis (OR 1.35, 1.18-2.01); and with a comorbid diagnosis of cerebrovascular disease (OR 1.60, 1.13-2.26). Patient-level factors less likely to be associated with mobility limitation documentation were visits by males (OR 0.80, 0.64-0.99); individuals with a cancer diagnosis (OR 0.76, 0.58-0.99); and by individuals seeking care for a chronic problem (relative to a new problem [OR 0.36, 0.29-0.44]). Physician-level factors associated with an increased likelihood of mobility limitation documentation were visits to neurologists (ORs 4.48, 2.41-8.32) and orthopedists (OR 2.67, 1.49-4.79) compared to primary care physicians. At the practice-level, mobility documentation varied based on the percentage of practice revenue from Medicare.

CONCLUSIONS: Mobility limitations are under-documented and may be primarily captured when changes in function are overt.