Considering Recent Confrontation Clause Cases from the Court of Appeals

The Confrontation Clause of the Sixth Amendment bars the State from introducing testimonial hearsay statements from a witness who does not testify at trial unless the witness is unavailable and the defendant had a prior opportunity to cross-examine the witness. Because the exceptions rarely apply, the upshot of this rule is that the State typically cannot introduce testimonial hearsay statements made by a person who does not testify at trial. Most of the debate about whether the rule applies centers on whether the statement at issue is (a) testimonial (that is, was it made for the primary purpose of establishing facts relevant to a criminal proceeding) and (b) offered for its truth (in other words, hearsay).

Two years ago, the United States Supreme Court in Smith v. Arizona, 602 U.S. 779 (2024), held that statements from an absent laboratory analyst that a testifying analyst conveyed to support his expert opinion about the chemical composition of the substance seized from the defendant were hearsay. For that reason, the admission of those statements at the defendant’s trial on drug charges raised Confrontation Clause concerns. Before Smith, some states had permitted such testimony under the theory that the statements were admitted to show the basis of the testifying expert’s opinion rather than for their truth and thus were not hearsay.

The extent of the limits Smith placed on testimony by substitute analysts was – and remains – unclear. One interpretation of Smith is that a substitute analyst may almost never testify to the results of a forensic analysis done by someone else. That is because in most circumstances the testifying analyst must rely on the inadmissible (and presumably testimonial) notes or reports of an absent analyst to form an opinion. That view finds support in statements from Smith rejecting the lower court’s basis-of-opinion reasoning because it would deprive a defendant of the “right to cross-examine the testing analyst about what she did and how she did it and whether her results should be trusted.” Id. at 799-800.

Another possible interpretation of Smith would limit its reach. Smith appeared to question whether the testifying analyst had in fact reached an independent opinion, suggesting that he instead was a mouthpiece for the absent analyst. Smith criticized the reasoning of the lower court as permitting the admission of every testimonial laboratory report through a trained surrogate “however remote from the case,” perhaps signaling that testimony from an analyst closer to the case, such as a supervising or reviewing analyst, might be admissible. Finally, Smith did not address whether the statements upon which the testifying analyst relied were testimonial, thus allowing for the possibility that some statements, perhaps those related to chain of custody and adherence to laboratory protocols, might not be.

Following Smith, the North Carolina Supreme Court in State v. Lester, 387 N.C. 90 (2025), held that machine generated data generated by an automatic process is not testimonial and is not hearsay. Thus, an expert may testify to such results without implicating the Confrontation Clause. Lester reached that conclusion in the context of phone records but cited as an example of this type of data an analysis of a blood sample from a gas chromatograph. 387 N.C. at 101 (also citing the example of machine-generated analysis of the chemical composition of drugs recognized in State v. Ortiz-Zape, 367 N.C. 1, 9-10 (concluding that raw data generated by a machine may be admitted for the purpose of showing the basis of an expert’s opinion)). That analogy suggests that such machine-generated data may be relied upon by a substitute analyst who provides an expert opinion based on those results, even if the analyst only knows that the blood tested was the defendant’s and that laboratory protocols were followed based on notations made by another analyst.

Recent court of appeals cases. In three post-Lester cases, the North Carolina Court of Appeals has considered whether testimony from a substitute analyst based on part on machine-generated data collected by another analyst is admissible.

• In State v. Wise, ___ N.C. App. ___, 2026 WL 157931 (June 3, 2026), the court held that a supervising analyst properly testified to the conclusions she reached based on the results of toxicology testing of the alleged victims that another analyst performed. The court reasoned that the values reflected in the toxicology report (which the testifying analyst reviewed, approved, and certified) were not the product of human interpretation but instead were the output of the laboratory’s analytical instruments; and
• In State v. Phillips, ___ N.C. App. ___, 2026 WL 1236281 (May 6, 2026), the court held that a sexual assault nurse examiner (SANE) who reviewed and signed off on another SANE’s report that included photographs permissibly testified to her conclusion based on those photographs that the victim had been strangled.
• Yet, in State v. Holt, __ N.C. App. ___, 2026 WL 1407585 (May 20, 2026), temporary stay allowed, __ N.C. __, 929 S.E.2d 594 (June 8, 2026), the court of appeals determined that that the trial court erred in admitting testimony from an analyst regarding the results of gas chromatograph testing of the defendant’s blood where the tests were performed by another analyst, even though the testifying analyst had reviewed and signed the testing analyst’s report.

The conclusions reached by those courts arguably are inconsistent, and those inconsistencies are not easily resolved by digging into the particular circumstances of each case. Instead, they appear to reflect continued uncertainly over the proper analysis of such issues post-Smith.

In Wise, the court reasoned that although the testifying witness did not perform the toxicology testing, the values reflected in the post-mortem toxicology reports for the victims were machine-generated raw data; thus, the testifying expert could rely on that data in reaching her conclusions. ___ N.C. App. at ___, 2026 WL 157931 at *5. The court explained that the toxicology reports became human interpretations of machine-generated data when they were reviewed, approved, and certified, a process that was completed by the testifying analyst.

In Phillips, the court reasoned that because the testifying SANE concluded that the alleged victim was strangled based upon her review of machine-generated photographs without relying on any factual assertions by the absent SANE, the Confrontation Clause was not implicated. ___ N.C. App. at ___, 2026 WL 1236281 at *7. Phillips went on to emphasize that its holding was limited to cases in which the underlying data can be independently examined and its accuracy verified by the testifying expert. Phillips distinguished cases in which the testifying expert must rely on an absent expert’s statements, explaining: “In the context of drug analysis or DNA testing, for example, the accuracy of the results testified to depends on the fact that the analyst performed the tests correctly, and whether that fact is true depends on whether the absent expert properly applied her specialized skill and training.” Id. at __; 2026 WL 123628 at *7. The court opined that a photograph speaks for itself and its accuracy does not depend on the truth of any factual assertions.

In Holt, the testifying analyst reviewed the testing analyst’s reports and data, analyzed the chromatograph, and performed his own calculations – all before approving the testing analyst’s work. The testifying analyst came to the same conclusion as the testing analyst: the defendant’s blood alcohol concentration was 0.15 grams of alcohol per 100 milliliters of blood.

The appellate court noted that while the testifying analyst could testify about laboratory procedures, he lacked personal knowledge of the testing analyst’s observation and handling of the defendant’s blood sample and did not personally participate in the blood testing. The testifying analyst admitted as much at trial, acknowledging that he “‘did not perform the analysis in the case,’” and that while clotting, fermentation, packaging, leaks, volume, homogeneity, and pipetted bubbles could have presented problems in testing, he personally did not have the opportunity to check for those matters. ___ N.C. App. at ___, 2026 WL 1407585 at *2.

The court of appeals stated that because the testifying analyst did not personally participate in the actual testing of the defendant’s blood, he could not convey the testing analyst’s observations or expose his lapses. Accordingly, the court explained that the defendant was unable to question the testing analyst’s “potential human errors, including his observation of the condition of the blood sample for possible clotting, fermentation, pipetted bubbles, or homogeneity in the blood sample, confirmation of the sample sufficient blood volume, and examination of packaging mistakes or vial leaks.” Id. at __; 2026 WL 1407585 at *5. Without identifying any particular statements by the absent analyst that the testifying analyst conveyed, the court reasoned that the testifying analyst’s inability to testify as to these matters deprived the defendant of an opportunity to challenge the veracity of the absent analyst’s out of court statements “that were doing much of the work.” Id. at ___; 2026 WL 1407585 at *5.

As noted above, the North Carolina Supreme Court has granted a temporary stay in Holt, so the court of appeals may not have the last word in the matter.

Where does this leave the trial courts?  One panel of the court of appeals has found no error in the introduction of expert testimony about the presence of drugs in victims’ bodies based on the expert’s review of toxicology testing the expert did not herself perform. Another has found error in the admission of expert testimony about the concentration of alcohol in a blood sample where the expert reviewed the testing results but did not conduct the test. Yet another has said that a testifying expert may testifying to her conclusions based on her review of photographs, which are “machine-generated output of the camera,” but may not rely on the testing of DNA or drugs by machines because the accuracy of those results depends on whether the testing analyst performed those tests correctly.

My advice to a trial court confronting a Confrontation Clause objection would be to surgically analyze the objection and the proffered testimony. What are the specific statements of the testifying analyst to which the defendant objects? Is the testifying analyst serving as the mouthpiece of a person who does not appear to testify; in other words, is the testifying analyst repeating someone else’s statement? If so, the testimony is inadmissible if that repeated statement is testimonial.

On the other hand, if the testifying analyst plans to testify to her independent conclusions, which are based on her review of nontestimonial information, such as machine-generated results, and perhaps laboratory procedures and chain of custody information, that testimony arguably is admissible under Smith and Wise (though Holt reaches a different conclusion and Phillips suggests otherwise in dicta). Of course, a defense attorney may properly cross-examine such an expert about all the things the expert does not know regarding how the testing was performed and the condition of the tested sample. Moreover, as Phil Dixon noted in this earlier post on Lester, an expert’s lack of knowledge about how the laboratory data upon which she relies was generated may raise questions about the reliability of the testing and thus, whether the expert’s testimony satisfies Rule 702.

Given the current confusion in the area, trial court rulings that identify the specific statements at issue, the specific information relied upon by the expert to reach her conclusions, and whether that relied-upon information constitutes a testimonial statement are likely to contribute to future clarity in this developing area of law.