SCOW: Evidence from Fitbit device is admissible without expert testimony on foundation, reliability

State v. George Steven Burch, 2021 WI 68, 6/29/21, on certification from the court of appeals, affirming a judgment of conviction; case activity (including briefs)

The circuit court properly exercised its discretion in allowing the state to introduce evidence relating to Fitbit data without requiring expert testimony on the reliability of the device.

After Nicole VanderHeyden was found murdered, police focused their suspicion on her boyfriend, Douglass Detrie. But they shifted focus from Detrie in part because his Fitbit device showed he’d only taken 12 steps during the time period of the killing, which wasn’t consistent with other evidence regarding the perpetrator’s movements. Burch became a focus after police found his DNA on an item of VanderHeyden’s clothing and a search of his cell phone yielded other incriminating evidence. (¶¶4, 7). (We have a separate post on the case discussing the issues arising from that search.)

Burch moved pretrial to exclude evidence about Detrie’s Fitbit device step-counting data, arguing the state had to produce an expert to establish the reliability of the science underlying the Fitbit device’s technology and that the state failed to sufficiently authenticate the records. The circuit court refused to exclude the Fitbit evidence related to step-counting.  (¶¶11, 27). Applying the deferential standard of review governing evidentiary decisions, the supreme court affirms.

Regarding the need for expert testimony to establish reliability of the device’s technology, “the requirement of expert testimony is an extraordinary one” and should apply only “when the issues before the jury are ‘unusually complex or esoteric'” and “‘not with in the realm of the ordinary experience of mankind,'” a determination made by the circuit court in an exercise of its discretion “on a case-by-case basis.” State v. Kandutsch, 2011 WI 78, ¶¶28-29, 336 Wis. 2d 478, 799 N.W.2d 865 (quoted source omitted). The circuit court properly exercised its discretion here when it concluded that the “principle idea behind pedometers” has been around a long time, and Fitbit itself has been on the market for more than a decade. Though the average juror may not know “the exact mechanics” of a technology’s “internal workings,” the public may nevertheless “generally understand[] the principle of how it functions and accept[] its reliability.” The supreme court held this conclusion was reasonable and within the circuit court’s discretionary authority. (¶¶30-31).

As for authentication, under § 909.01 simply requires evidence “sufficient to support a finding that the matter in question is what its proponent claims.” See also State v. Smith, 2005 WI 104, ¶¶31-33, 283 Wis. 2d 57, 699 N.W.2d 508. In this case, that requires sufficient evidence to support a finding that the records produced by the state are in fact Fitbit’s records associated with Detrie’s Fitbit device. Burch’s authentication argument is phrased as another challenge to the reliability and accuracy of the Fitbit device, argues that “the State failed to show that the Fitbit device reliably and accurately registered Detrie’s steps that evening, which “reaches beyond the threshold question authentication presents.” Because the circuit court’s authentication obligation is simply to determine whether the records were in fact the records from Detrie’s device, not whether they accurately counted his steps, Burch’s authentication challenge misses its target. (¶¶33-34).

One justice (A.W. Bradley) dissents from the conclusion that establishing the Fitbit’s reliability doesn’t require expert testimony. As she explains in criticizing the majority’s reasoning:

¶94     An average jury member would likely know what a Fitbit is and what it does. Of course, as relevant here, it counts the wearer’s steps. But that isn’t the question. In determining whether expert testimony is required, the relevant inquiry is how a Fitbit counts the wearer’s steps and then ultimately, whether it does so with sufficient reliability.

¶95     How does it work? A Fitbit device uses a microelectronic triaxial accelerometer to capture a person’s body motion in three-dimensional space and record related data. This motion data is then analyzed by utilizing proprietary algorithms to surmise patterns and thus to identify daily steps taken.

¶96     Is it sufficiently reliable to be admitted as evidence in court? I don’t know. But, I do know that the answer does not lie in its ubiquitous use.

¶97     I also know that absent expert testimony there is insufficient foundation in this record for the majority to determine, in essence, that a presumption of accuracy and reliability attends the underlying technology of a Fitbit. The error of such a presumption is made manifest by reference to an overarching analysis of 67 studies on Fitbit accuracy disseminated by the National Center for Biotechnology Information (NCBI), under the auspices of the U.S. National  Institutes of Health (NIH). The researchers found that Fitbit devices were “likely to meet acceptable accuracy for step count approximately half the time.” Lynne M. Feehan, et al., Accuracy of Fitbit Devices: Systematic Review and Narrative Syntheses of Quantitative Data, … (2018).

¶98     In citing this study, I neither endorse nor disclaim its conclusions. It suggests, however, when a compilation of studies indicates acceptable accuracy is met only “half the time,” that something may be amiss with the majority’s presumption of accuracy and reliability.