Determining the follow-up rates for shoulder arthroplasty: a retrospective review of a multicenter study of 4700 shoulders

Abstract

BackgroundPostoperative follow-up is a routine component of the surgical care of a patient undergoing shoulder arthroplasty. Orthopedic journals require a high follow-up percentage in prospective studies to ensure reproducibility, generalizability of the data, and to limit bias. This study’s primary aim was to report the true follow-up rates after total shoulder arthroplasty among a multicenter cohort at 1, 2, 5, and 10 years and to explore various demographic factors associated with follow-up rates. MethodsA retrospective review of patients undergoing anatomic and reverse total shoulder arthroplasties at five U.S. centers from 2000 to 2020 was performed. The presence or absence of follow-up at 1-, 2-, 5-, and 10-year follow-up intervals was recorded as well as associated patient demographics and history variables. Univariate and multivariable logistic regression models were employed to examine the factors associated with patient follow-up. Results4700 shoulder arthroplasties were identified across five centers for review. Eligible follow-up rates were 66% at 1 year, 47% at 2 years, 26% at 5 years, and 9% at 10 years. When compared with the majority baseline groups at 1 year in multivariable analysis, the follow-up rate was significantly less for male patients (odds ratio [OR] = 0.83, P = .004), those over age 80 (OR = 0.66; P < .001), black patients (OR = 0.76, P = .035), Hispanic patients (OR = 0.54, P = .041), self-paying patients (OR = 0.5, P < .001), disabled patients (OR = 0.69, P = .021), those with psychiatric history (OR = 0.76, P < .001), and those with preoperative opioid use (OR = 0.81, P = .016). At 2 years, significantly lower follow-up rates were identified with patients ages 35 to 49.9 years (OR = 0.62, P = .013), over age 80 (OR = 0.67, P < .001), Hispanic (OR = 0.42, P = .010), unemployed (OR = 0.65, P = .014), psychiatric history (OR = 0.82, P = .017), and smoking history (OR = 0.71, P = .010). ConclusionTrue clinical follow-up rates following shoulder arthroplasty are much lower than the 80% 2-year follow-up rate expected by most orthopedic journals. Many factors contribute to differences among follow-up rates. Promoting follow-up at all time points would limit bias in orthopedic literature and potentially allow for early intervention of complications, but would need to be balanced with the additional burden and resources needed from patients and providers.