Letter to the Editor: Individual Patient-reported Activity Levels Before and After Joint Arthroplasty Are Neither Accurate nor Reproducible

To the Editor,

We read with interest the study by Vaughn et al. [1] which questions the accuracy of patients' estimates of post-surgery walking distance compared to measurements from accelerometers. We share the authors’ worries about the inaccuracies of patient-reported measures of postoperative activity levels and believe that accelerometers may provide a useful adjuvant.

However, we do have some concerns regarding their statement that one should use “… caution in the adoption of accelerometers as adjuncts to outcome measurements …” as this was based on their patients’ roughly 50% noncompliance with wearing the clip-on accelerometer used by the study.

We have observed in our own ongoing study that a wrist-worn accelerometer, which may have more consumer appeal, may result in patients being more willing to use the devices as advised by a physician (we have observed that approximately 90% of patients do so). An added benefit of accelerometers is the ability to assess the impact of external events on recovery, such as weather patterns or the stay-at-home orders and suspension of nonurgent services during the current coronavirus disease 2019 (COVID-19) pandemic. Although the latter may have profoundly impacted the recovery of patients who underwent elective surgery before the pandemic, the impact of this will be difficult to estimate using only patient-reported estimates of functional limitations.

In the context of our own ongoing study, we have been measuring patients’ (n = 5) daily activity before and after lumbar/cervical spinal fusion surgery in an IRB-approved study at our institution in New York City, one of the COVID-19 hotspots in the United States, using the Fitbit Flex 2 (Fitbit, San Francisco, USA) accelerometer. While few of our study participants’ recoveries have been affected by statewide stay-at-home orders, these data provide an important window into how patients’ recoveries may have been impacted.

We tracked median daily steps of the five included patients and compared three time periods: 6 weeks before the statewide stay-at-home order in New York, NY, USA (March 22), the 3 weeks after that, and the subsequent 3 weeks. At the time of these measurements, patients were a median of 4 months (2.5–5.5 months) into their surgical recoveries.

Results from segmented regression analyses using these time-series data (Fig. 1) show an acute decrease in daily steps in the 3 weeks after the stay-at-home order (a decrease of 1561 steps/day, p = 0.02) while in the subsequent 3 weeks daily steps increased (an increase of 184 steps/day, p = 0.02). When compared qualitatively to the activity profiles of similar patients’ recoveries that had been studied prior to the pandemic as part of our ongoing study, it seems that although patients’ activity levels initially were restricted by the stay-at-home order, they returned to a more-typical post-operative trajectory by 6 weeks after it. However, the extent to which this transient decrease in activity levels will affect long-term outcomes is unclear. While these data are influenced by several potential sources of heterogeneity (types of surgery, potential seasonality effects), they provide a unique window into the potential effects of the pandemic on patient mobility during recovery.

Fig. 1.

The physical activity levels of patients active during the pandemic decreased significantly immediately following the issue of the stay-at-home order but began to increase and grossly return towards typical postoperative behavior by 6 weeks after the order was issued.

The true impact of the COVID-19 pandemic on patients’ recoveries will become clear only after “return-to-normal” operations. These impacts may be more pronounced in other orthopaedic populations, like total joint replacement, where physical function—specifically ambulation in the postoperative period—is more directly related to biomechanical changes after surgery. Regardless, as also noted by Vaughn et al. [1], we have observed that accelerometers may be useful tools in accurately estimating patient mobility, while also allowing for estimations of the impact of external factors like the COVID-19 pandemic on patterns in recovery.