Staci Thomas appeals the defense verdict in her medical malpractice action against the Emory Clinic, Inc. She argues that the trial court erred in admitting hearsay evidence and abused its discretion in refusing to replace a juror who did not disclose in voir dire his niece’s relationship to several witnesses. We agree that the trial court erred in admitting hearsay. We find that the error was harmful, despite the proper admission of similar evidence. We therefore reverse. We do not reach the question whether the trial court abused its discretion in refusing to replace the juror.
Jeffrey Olson, a neurosurgeon employed by the Emory Clinic, performed surgery to remove a choroid plexus papilloma, a type of benign brain tumor, from Thomas on November 4, 2002. The gravamen of Thomas’s complaint is that foreign material was left behind.
Although the average recovery time from such surgery is six to eight weeks, Thomas’s condition did not improve. Among other problems, she vomited constantly, to the extent that she required a feeding tube; she suffered depression; and she experienced such severe headaches that she required morphine. Thomas was treated with prescription medications, underwent brain imaging, and underwent multiple examinations with other Emory specialists, including a gastroenterologist, an endocrinologist, an oncologist, a psychiatrist and a psychologist. She underwent a lumbar puncture and the insertion of a feeding tube and a shunt. Dr. Olson did not think it reasonable to perform exploratory surgery, given the images of Thomas’s brain and the risks involved.
In October 2003, when Thomas still had not improved, she sought treatment at the Mayo Clinic in Florida. There it was determined that she had a lesion in the area of her brain where the tumor had been removed. After undergoing radiation and chemotherapy at the Mayo Clinic, the lesion shrank and Thomas’s symptoms improved. But the lesion eventually started growing again, and her symptoms returned.
On March 8,2006, Kent New, a neurosurgeon at the Mayo Clinic, removed the lesion. Pathology of the specimen showed that the lesion was not a tumor but rather an inflammation caused by a reaction to the presence of cotton fibers shed from “cottonoids” or “surgical patties,” small, square or rectangular sponges made of compressed cotton and used in surgery.
Thomas filed this action against Emory, alleging that her injuries were caused by Emory employees leaving foreign material — cotton fibers — at the surgical site when her brain tumor was *458removed. She also alleged that Emory violated the standard of care by failing to determine and alleviate the cause of her symptoms. A jury returned a verdict in favor of Emory, upon which the trial court entered judgment. After the trial court denied her motion for new trial, Thomas filed this appeal.
1. The admission of testimony about the pathologists’ report,
Thomas argues that the trial court erred by admitting certain testimony of Kent New, the Mayo Clinic neurosurgeon who removed Thomas’s lesion in March 2006. New testified at his deposition that he submitted the specimen he removed for testing, and that he “had talked to the pathologists] about it and they had told [him] not only in the actual report did they not make a comment on it, but they said that from what they saw that they didn’t notice that any foreign body material was in the specimen.” He testified that the pathologists did not find “a clear portion of the [patty] in the pathologic examination. . . . [T]hey didn’t, they didn’t see a clear foreign body.” He reiterated several times that the pathologists had not seen any foreign material in Thomas’s sample. The videotape of New’s deposition was played for the jury. The court overruled Thomas’s hearsay objection to the testimony because she “did not object in a timely fashion, when this came up” at New’s deposition.
(a) The standard of review.
Generally, “[t]he admission of evidence lies in the sound discretion of the trial court.” Dept. of Transp. v. Mendel, 237 Ga. App. 900, 902 (2) (517 SE2d 365) (1999). But in this case, the court did not evaluate the merits of the evidence. Instead, it simply concluded that as a matter of law, Thomas waived her hearsay objection by failing to assert it at the deposition. We therefore review the trial court’s ruling de novo. See Knott v. Knott, 277 Ga. 380, 381 (2) (589 SE2d 99) (2003) (lower court’s conclusions with respect to matters of law are subject to de novo review).
(b) The testimony was hearsay.
Emory argues that New’s testimony about what the pathologists told him was not hearsay. We are not persuaded. Former OCGA § 24-3-1 (a) defined hearsay as evidence “which does not derive its value solely from the credit of the witness but rests mainly on the veracity and competency of other persons.” (The new Evidence Code, effective January 1,2013, defines hearsay as “a statement, other than one made by the declarant while testifying at the trial or hearing, offered in evidence to prove the truth of the matter asserted.” OCGA § 24-8-801 (c).) Hearsay is “testimony in court, or written evidence, of a statement made out of court, the statement being offered as an assertion to show the truth of the matters asserted therein, and thus resting for its value upon the credibility of the out-of-court asserter.” *459Miness v. Miness, 254 Ga. 658, 659 (1) (333 SE2d 574) (1985) (quoting McCormick on Evidence, 2nd ed., § 246, p. 584 (1972)).
Emory argues that given its concession that the specimen contained fibers, the testimony was not hearsay because it was not offered to prove the truth of the matter asserted, the absence of fibers in the specimen. Rather, Emory contends, the testimony was elicited to offer additional support for the course of treatment of Olson, the Emory neurosurgeon, as it further demonstrated the difficulty of seeing fibers left in surgical sites. But for the testimony to support such a contention, it had to be offered to show the truth of the matter asserted therein — that the pathologists “didn’t notice that any foreign body material was in the specimen.” The testimony thus rested “for its value upon the credibility of the out-of-court assert-er[s],” Miness, 254 Ga. at 659 (1), the pathologists, and the testimony was hearsay.
(c) The failure to raise the hearsay objection at the deposition did not waive the objection.
OCGA § 9-11-32 (b) provides that, with some exceptions, objections to deposition testimony “may be made at the trial or hearing to receiving in evidence any deposition or part thereof for any reason which would require the exclusion of the evidence if the witness were then present and testifying.” “Objections as to the competency and relevancy of the evidence need not be made at the taking of the depositions.” (Citation omitted.) Hamilton v. Pulaski County, 86 Ga. App. 705, 709-710 (2) (72 SE2d 487) (1952). The trial court erred by ruling that Thomas waived her hearsay objection because under OCGA § 9-11-32 (b), she had the right to object to New’s deposition testimony at trial even though she did not object at the deposition.
Emory argues that one of the exceptions to the statute applies. That exception, found at OCGA § 9-11-32 (d) (3) (A), provides: “Objections to the ... competency, relevancy, or materiality of testimony are not waived by failure to make them before or during the taking of the deposition, unless the ground of the objection is one which might have been obviated or removed if presented at that time.” We conclude that the exception does not apply because the hearsay nature of the testimony could not have been “obviated or removed” had Thomas objected at the deposition: with or without an objection, New’s testimony about what the pathologists told him was hearsay, and an objection would not have provided any way to cure that fact. Compare Andean Motor Co. v. Mulkey, 251 Ga. 32 (1) (302 SE2d 550) (1983) (if party “had obj ected during the deposition to the absence of proof of the witness’ competence to testify as an expert, defense counsel might have been able to cure this ground of objection by proof of the witness’ *460qualifications”). Accordingly, Thomas did not waive her hearsay objection by failing to assert it at the deposition.
(d) The admission of the hearsay testimony was harmful.
Because the improperly admitted evidence related to a core issue about which the evidence was conflicting, application of the “cumulative evidence” rule to find the error harmless would be inappropriate. The admission of other evidence similar in nature does not alone determine whether the error was harmful. To determine whether an error is of reversible magnitude, we must consider the nature of the improper evidence in light of the issue it tends to prove or disprove. Here, one of the core issues was whether Jeffrey Olson, the Emory neurosurgeon, committed malpractice by leaving too many cotton fibers in the plaintiff’s brain. Therefore, evidence about what was visible when the resulting lesion was later excised, such as hearsay testimony that neuropathologists saw no fibers in the lesion, is central to a critical issue. In a civil jury trial, the plaintiff need only prove her case by a preponderance of the evidence, and adding the weight of the improper evidence could easily have tipped the scales in the defendant’s favor.
“When [an appellant] brings a case here he must show error which has hurt him. This court is not an expounder of theoretical law, but it administers practical law, and corrects only such errors as have practically wronged the complaining party.” Brown v. City of Atlanta, 66 Ga. 71, 76 (1880). Accordingly, erroneously admitted evidence “will not require the granting of a new trial, it appearing that the evidence, though subject to the objection made, could not have affected the verdict rendered.” South Ga. R. Co. v. Niles, 131 Ga. 599 (62 SE 1042) (1908).
Given the existence of error, the question then becomes how to determine whether the error “affected the verdict.” “The doctrine of harmless error in civil cases is set forth in [OCGA § 9-11-61]. It is the same language found in Rule 61 of the Federal Rules of Civil Procedure. Under both rules the case should not be reversed unless the refusal would be ‘inconsistent with substantial justice.’ ” Johnson v. State, 238 Ga. 59, 61 (230 SE2d 869) (1976). Our Supreme Court in Johnson noted that it had not previously articulated an explicit standard for weighing nonconstitutional errors in criminal cases and adopted “what is known as the ‘highly probable test,’ ” i.e., whether it was “highly probable that the error did not contribute to the judgment.” Id., citing Roger Traynor, “The Riddle of Harmless Error” (1970).
If the error in a criminal case is of constitutional dimension, erroneously admitted cumulative evidence may still be harmless, but the state must demonstrate beyond a reasonable doubt that the error *461did not contribute to the verdict obtained. State v. Hightower, 236 Ga. 58, 61 (222 SE2d 333) (1976). If the erroneously admitted evidence related to a “core issue,” it is reversible even if it is cumulative. Willingham v. State, 279 Ga. 886, 887-888 (1) (622 SE2d 343) (2005); Brawner v. State, 278 Ga. 316, 319-320 (2) (602 SE2d 612) (2004); Gifford v. State, 287 Ga. App. 725, 726-727 (1) (652 SE2d 610) (2007). Cf. Williams v. State, 279 Ga. 731, 734 (5) (b) (620 SE2d 816) (2005); Watson v. State, 278 Ga. 763, 768 (2) (b) (604 SE2d 804) (2004).
We have explicated no similar standard in our case law for weighing error in civil cases. OCGA § 9-11-61 exhorts us to “disregard any error or defect in the proceeding which does not affect the substantial rights of the parties,” but how do we determine when an error affected the parties’ “substantial rights” or decide that the failure to correct the error was “inconsistent with substantial justice”? Clearly, whether an error requires reversal depends on the nature of the error and the importance of the issue to which it applies.
We have held that “error is presumed hurtful unless it appears to have had no effect upon the result of the trial.” (Punctuation and footnote omitted.) Allen v. Spiker, 301 Ga. App. 893, 897 (1) (689 SE2d 326) (2009). See also Rogers v. Johnson, 94 Ga. App. 666, 682 (6) (96 SE2d 285) (1956). In Rogers, 94 Ga. App. at 682 (6), we held that error unrelated to the plaintiff’s allegations of negligence was not reversible, and in Allen, 301 Ga. App. at 897 (1), we held that error in directing a verdict on the plaintiff’s medical bills damages claim was reversible despite the plaintiff’s verdict, because we had no way to determine how the jury had factored in the error when deciding how much money to award. Similarly, improperly admitting evidence of a disability policy was reversible error despite a plaintiff’s verdict, because we could not determine whether the jury mitigated its damages award in consideration of this improper evidence. See Foster v. Harmon, 145 Ga. App. 413, 414 (1) (243 SE2d 659) (1978). So if the erroneously admitted evidence is “remote from the real issue of the case, . . . the harmlessness of the admission of the evidence is self evident.” Pembrook Mgmt. v. Cossaboon, 157 Ga. App. 675, 678 (4) (278 SE2d 100) (1981) (no reversible error in admission of hearsay evidence that plaintiff was administered morphine when issue was whether plaintiff or defendants were “the ultimate causative force” of the accident).
In addition to holding that erroneously admitting evidence remote from the main issues of the case is not reversible, we have repeatedly held that the erroneous admission of evidence “is harmless when it is cumulative of legally admissible evidence showing the same fact.” DeGolyer v. Green Tree Servicing, 291 Ga. App. 444, 450 (5) (a) (662 *462SE2d 141) (2008) (erroneous admission of appraisal evidence harmless because appraisal value was nearly identical to value on properly admitted loan application).
In each case substantially the same evidence was admitted without objection in testimony by other witnesses or elicited on cross examination of the same witness. In such circumstances the rule is applicable that the allowance of evidence, even though erroneous as contended, does not require reversal of the judgment complained of where testimony substantially to the same effect is adduced or where counsel on cross examination of the witness and with full knowledge of the character of the evidence to be expected elicited testimony comparable in import to that to which objection was made.
Goodyear Tire & Rubber Co. v. Johnson, 120 Ga. App. 395, 397 (2) (170 SE2d 869) (1969).
The unifying link in these cumulative evidence cases is that the evidence admitted in error established a fact that was otherwise established by properly admitted evidence. See, e.g., Davis v. Reid, 272 Ga. App. 312, 318 (3) (612 SE2d 112) (2005) (no reversible error in overruling defendant’s objection to doctor testifying from pathology report because plaintiff referred to same report during direct examination of another witness and others had testified about findings from report); Stephens v. Howard, 221 Ga. App. 469, 469-470 (1) (471 SE2d 898) (1996) (no reversible error in failure to exclude hearsay physical assessment because treating doctor and plaintiff testified extensively about physical disabilities); Platt v. Nat. Gen. Ins. Co., 205 Ga. App. 705, 709-710 (1) (423 SE2d 387) (1992) (no reversible error in allowing transcripts to go out with jury because witness testified without objection to contents); Wolfson v. Rumble, 121 Ga. App. 549 (1) (174 SE2d 469) (1970) (no reversible error in allowing examining physician to testify about plaintiff’s complaints of pain because evidence was cumulative of plaintiff’s uncontradicted testimony about her pain); Savannah Transit Co. v. Williams, 107 Ga. App. 212, 215 (2) (129 SE2d 417) (1963) (no reversible error in admission of doctor’s hearsay testimony that plaintiff had been rendered unconscious in crash because plaintiff gave the same testimony without contradiction).
But “cumulative” does not automatically mean “harmless” just because facts extracted from erroneously admitted evidence are similar to facts extracted from proper evidence. It depends on the quality and quantity of both the proper and improper evidence, and the relation of that evidence to the core issues before the jury. In *463considering whether improperly admitting evidence was reversible error, this court has essentially applied the “highly probable” test espoused by Justice Traynor in “The Riddle of Harmless Error” and adopted by the Supreme Court of Georgia in considering nonconstitutional errors in criminal cases. Quoting Traynor, the Supreme Court said:
The highly probable test avoids the evils of inadequate or excessive stringency by making affirmance conditional on high probability that error did not affect the judgment. The test compels a judge to go beyond a first glance for affirmance or a fleeting glimpse for reversal. It compels him to exercise his mind in the exercise of his discretion, to go beyond the appearances of the result to an examination of what causal links there may be between error and the judgment. It keeps judicial discretion within the ample bounds of reason. It can greatly improve the net worth of the judicial process as it thus holds down excesses either of affirmance that recklessly dampens assurance of a fair day in court or of reversal that needlessly calls for still another fair day at the expense of litigants who are still awaiting their first day in court.
(Emphasis omitted.) Johnson v. State, 238 Ga. at 61.
Applying essentially the same test in the civil context, we have held the improper admission of hearsay evidence that bolstered the sufficient but slight direct evidence on a core issue constituted reversible error. See A Child’s World v. Lane, 171 Ga. App. 438, 442 (4) (319 SE2d 898) (1984). Although the trial court in that case decided that the erroneously admitted evidence was cumulative, we “[could not] say that the erroneous admission of ‘bolstering’ double hearsay testimony concerning a report. . . was harmless.” Id.
In the case before us now, one of the two core issues before the jury was whether Olson, the Emory neurosurgeon, committed malpractice by leaving too much cotton fiber in Thomas’s brain. As Emory points out in its brief, it never denied that fibers were left, only that they were difficult to see and that their presence did not indicate malpractice. That shed fibers were hard to see and that microscopic fragments were inevitably left after brain surgery of this type was not disputed. But the witnesses did dispute how much fiber had been left in Thomas’s brain. According to Emory, the fragments were few and could not even be seen with the naked eye. According to Thomas, there were so many fragments that their presence would have been obvious and a responsible surgeon would have attended to them.
*464Thomas’s expert neuropathologist testified that he examined three slides taken from fragments of the lesion removed from Thomas. The first slide contained two slices of the lesion, which were sent to neuropathology during the operation, and the expert testified that he could quite easily see threads at 40x, the lowest magnification he used, without polarization. The evidence established that the operating neurosurgeons used magnification up to 40x. The first slide also showed fibrous scarring consistent with a foreign body reaction. The second slide, taken from the lesion after it was removed, also contained an abnormal number of foreign bodies, so many they were too numerous to count, visible at 40x. The third slide, from an additional specimen the neurosurgeon submitted, was “filled with cotton threads” easily identifiable to the expert neuropathologist, who testified that while he often saw one or two fibers in freshly resected tumors, here, he saw many. Similar to Thomas’s expert neuropathologist, her expert neurosurgeon testified that the pathology slide in this case showed “a significant wad of cotton that was left behind.”
In contrast, Emory’s witnesses downplayed the amount of fibers and indicated they could not be seen with the naked eye. Olson, the Emory neurosurgeon, did not dispute that the lesion later removed from Thomas’s brain contained fiber, but testified that what he could see during the surgery in a “moving, pulsating field” was “nothing like” what a pathologist can see later in a polarized two-dimensional field. Kent New, the Mayo neurosurgeon who removed the lesion, testified that he thought the inside of the lesion looked different, “a little bit like cotton fibrils to me grossly,” so he sent a separate specimen from the inside of the lesion to the neuropathologists during the surgery. But he went on to testify, improperly, that “not only in the actual report did [the Mayo neuropathologists] not make a comment on it, but they said that from what they saw that they didn’t notice any foreign body material was in the specimen.” Upon questioning, New repeated several times that the neuropathologists told him they had not seen any foreign material and that he had told Thomas that during a telephone conversation, “trying to encourage [her] to move on.”
In light of the core issues in this case and the properly admitted evidence as outlined, New’s improperly admitted testimony — that not only did the Mayo neuropathologists’ written report not mention foreign material but also that they told him the same thing in conversation — was critical. It directly addressed the core disputed issue of whether the Emory neurosurgeon left an excessive amount of cotton in Thomas’s brain. The jurors could easily have concluded that if the Mayo neuropathologists saw no fibers, the Emory neurosurgeon would not have seen any either.
*465Decided March 26, 2013 Reconsideration denied April ll, 2013 Vincent R. Lauria, Charles M. Cork III, for appellant. Bendin, Sumrall & Ladner, Benjamin D. Ladner, Melanie S. Taylor, Brian D. Trulock, for appellee.The proper admission of evidence about the visibility or invisibility of fibers generally does not make harmless the admission of improper evidence about the amount of fiber in this case. The plaintiff need only tip the scales by a preponderance of the evidence, and two pieces of evidence favorable to the defendant is not equivalent to one. Considering the nature of the improperly admitted evidence and the core issues before the jury, its admission was reversible error.
2. The refusal to remove a juror.
In light of our decision in Division 1, we do not reach Thomas’s contention that the trial court erred in denying her motion to remove a juror.
Judgment reversed.
Barnes, P. J., concurs. McMillian, J., concurs fully in Divisions 1 (a), (b), (c) and 2, and concurs in judgment only in Division 1 (d).