A jury convicted the defendant of one count of motor vehicle homicide by negligent operation under G. L. c. 90, § 24G(h). On appeal, she contends that (1) a medical record, which contained only a vague reference to an unquantified amount of amphetamine in her system, was unfairly transformed, during trial, into a surprise theory of criminal liability; there was no foundation for the amphetamine evidence used to sup*71port this new theory; and, in arguing that theory in his closing, the prosecutor committed error by arguing that “speed kills”; (2) the instructions on causation were inadequate; and (3) a lay opinion as to her vehicle’s speed should not have been admitted. We conclude that the amphetamine evidence, which became a mainstay of the Commonwealth’s case, was not adequately supported in the record and was unfairly portrayed by the Commonwealth. Consequently, the jury were permitted to speculate upon important questions, resulting in a substantial risk of a miscarriage of justice. We reverse.
1. Background. The defendant was employed by a private company that transported criminal fugitives throughout the United States. On May 24, 2000, at about 6:00 a.m., she was driving a van with an inmate in the back, and her coworker, Scott Bellon, asleep in the front passenger seat. She had been driving since 9:00 p.m. on the previous evening, the road was wet and slippery, and it was raining lightly. The van collided with a stone abutment (part of a railroad overpass) on Route 41 in Great Barrington, and Bellon was killed.
The principal defense theory was that the accident resulted from poor road design and ineffective warning signs.
The Commonwealth’s theory of the case began as one of speeding, resulting in death. However, this theory was supplemented by the theory that the death was the result of the defendant’s driving while under the influence of amphetamines, the amount of which was never defined. That second theory held center stage for most of the trial and prosecutorial closing despite the fact that there was an insufficient evidentiary foundation for it.
The complaint contained one count for the civil infraction of speeding (G. L. c. 90, § 17), and a second count for motor vehicle homicide by negligent operation, tracking a portion of the language of G. L. c. 90, § 24G(¿>). The complaint did not allege impaired operation while under the influence of proscribed chemicals or substances, which is an alternative offense under the statute.1 At trial, the prosecutor, in his opening statement, repeatedly emphasized “reckless driving” and exces*72sive speed as causes of the accident. The opening made no mention of amphetamines or other chemical substances.2
The Commonwealth’s first three witnesses testified to the progress of the van as the defendant drove south on Route 41, prior to the accident. These percipient witnesses testified as to the vehicle’s excessive speed, and two witnesses (Jesse Crane and Robert Walder) also testified that the van had passed other vehicles in a no-passing zone. Their testimony provided evidence that the defendant was the driver of the van and operated it at an excessive speed, on a road that was both wet and irregular in its curves, with some signage pertaining to curves and speed limits.
The Commonwealth also called two Great Barrington police officers, David Cobb and Kevin Larkin, who had responded to the accident site. Officer Larkin, qualified as an accident reconstruction expert, gave opinion evidence to the effect that excessive speed caused the accident that resulted in Bellon’s death.3
The Commonwealth’s case-in-chief turned next to the appearance of Dr. Robert Brown, a pathologist. Although Dr. Brown was called to testify as to the victim’s injuries and the cause of *73his death, the doctor’s testimony shifted, in response to an apparently unexpected sequence of questioning, to an analysis of certain of the defendant’s medical records from Fairview Hospital, where she had been taken for the treatment of her own serious injuries.4 5The records admitted as exhibits included the defendant’s emergency room medical summary record, accompanied by a lab report containing a “toxicology” section which listed “AMPHETAMINES, urine, quai POS” among the test results. The report also recites “DRUG CONFIRMATION BY TLC[:] Amphetamines, confirmation [ — ] POS.”
As elucidated through Dr. Brown’s testimony, the defendant’s medical records revealed only that her urine screen “tested positive for one drug, amphetamines . . . [described as] a drug which is a central nervous system stimulant . . . the purpose [of which] is to stimulate the brain or the central nervous system.” Notwithstanding that the records contained no evidence whatsoever of any concentration of amphetamines or other substances in the defendant’s system, Dr. Brown was asked by the prosecutor for his opinion about the uses and effects of amphetamines. The doctor noted that the medical use of amphetamines is indicated for children with attention deficit disorder and hyperactivity, and in a rare adult disease called narcolepsy. From this vantage point — and again without any evidence as to the concentration of amphetamines in the defendant’s system — the prosecutor next sought to develop this line of inquiry in even greater detail by attempting to elicit testimony that other uses for the drug were illegal.5 The doctor also referenced a second, “confirmation test,” which he said had confirmed the presence in the defendant’s blood of amphetamines, as well as caffeine and meprobamate — the lat*74ter being described by the doctor as “a sedative for anti-anxiety and for muscle relaxation.”
Even putting to one side that Dr. Brown had been called as a pathologist to opine on the cause of death, there was no foundation evidence in the record concerning any of these three substances — amphetamine, meprobamate or caffeine — and no evidence concerning concentration of the substances in the defendant’s system and the effect upon her ability to operate the van.
To establish the defense that poor road design and faulty signage were a superseding cause of the accident, the defense called Roland Proulx as an expert witness concerning appropriate signage for Route 41, given its layout and the surrounding terrain at the scene of the accident, and approaching the scene.
The defendant then testified and stated that she reduced her vehicle’s speed in accordance with the existing warning signs. However, during her cross-examination, the prosecution concentrated upon its new theory, the amphetamine cause-of-death theory. The defendant was repeatedly and forcefully questioned concerning her use of amphetamines. She denied, generally, ever having used the drug, and denied its use specifically at the time of the accident. In light of continued questioning by the prosecutor, the defendant repeated the denial several times. Notwithstanding the defendant’s denial of drug use and the lack of any evidence as to the level of amphetamines found in her system, this repeated questioning produced an admission by the defendant that she knew that amphetamines were used to stay awake, that the drug was also called “speed,” that professional drivers (such as truck drivers) used amphetamines to stay awake on the road, and that amphetamines can impair one’s ability to drive.
The problematic nature of the amphetamine evidence and questioning was enhanced in the closing arguments. The prosecutor argued vigorously that the defendant had amphetamines in her system, had lied about her illegal drug use, was negligent in her driving as a consequence of the drug use, and that, standing alone, her use of the drug at the time of the ac*75cident could be sufficient evidence for the jury to find the defendant guilty.6
2. Discussion. The defendant claims that admission of the amphetamine evidence, coupled with the Commonwealth’s shifting theory of proof — without evidence of the concentration of amphetamines in her system and the amount required to impair her driving — created a substantial risk of a miscarriage of justice, because, among other reasons, she had no notice that the Commonwealth would offer and then use it as evidence of her criminal liability, and that she was deprived of a fair opportunity to contradict the evidence, e.g., with expert testimony.
Commonwealth counters that the prosecutor properly utilized the toxicology report to argue that the defendant’s use of amphetamines was (1) one of several factors, and (2) a sufficient independent basis, to prove the defendant’s criminal negligence. The Commonwealth also argues that the report was properly used to impeach the defendant’s credibility by showing that she made inconsistent statements regarding drag ingestion, and to anticipate and rebut her testimony that she had taken no drugs besides caffeine.
Based on our review of the record, we conclude that the amphetamine theory of criminal liability, which became the mainstay of the Commonwealth’s case in mid-trial and was used to prove that the defendant was driving under the influence of drags, lacked evidentiary foundation — both by way of any evidence concerning the concentration of amphetamines in the defendant’s system as well as any expert evidence concem*76ing the effect of such substances on her ability to drive. These errors, in our view of the case, created a substantial risk of a miscarriage of justice. See Commonwealth v. Miranda, 22 Mass. App. Ct. 10, 16, 21 (1986).
There was no foundation to support the theory that the amphetamine use was the cause of the accident. To be properly admitted, the evidence of amphetamines in the defendant’s system in this case required, at a minimum, (1) reliable evidence as to the amount or concentration of the dmg in the defendant’s system; and (2) expert testimony indicating that the concentration of the drug in the defendant’s system would impair her ability to operate a motor vehicle. See Commonwealth v. Dube, 413 Mass. 570, 571 n.2 (1992) (noting that Commonwealth would presumably offer expert testimony explaining how serum test converts to blood alcohol scale); Commonwealth v. Johnson, 59 Mass. App. Ct. 164, 169 (2003) (record of urine screen showing presence of drug, without explanation, left jury to “speculate as to the record’s significance upon the important questions of whether, and when, the defendant had ingested a controlled substance, and whether such a controlled substance was in the defendant’s system in concentrations sufficient to affeet his ability to operate a motor vehicle”). See also Commonwealth v. Kirkpatrick, 423 Mass. 436, 447 n.8, cert. denied, 519 U.S. 1015 (1996) (medical records were “difficult to decipher” and, without “testimony from a medically knowledgeable witness,” invited jury to speculate).
Because neither of these prerequisites was met, the evidence of the mere presence of amphetamines cannot justify the verdict and the supplemental theory of proof advanced by the Commonwealth. As presented and argued, such evidence invited unacceptable speculation by the jury. See Commonwealth v. Johnson, 59 Mass. App. Ct. at 169.
The defendant argues on appeal that the amphetamine evidence was introduced either by accident or ambush. In this case, such distinctions are irrelevant.7 The point is that toxic *77evidence ended up before the jury unfairly and was presented in a manner that undermines the validity of the verdict.8
Evidence of guilt must not require a jury to speculate. See Commonwealth v. Kirpatrick, supra; Commonwealth v. Martin, 427 Mass. 816, 821-822 (1998) (distinguishing probative value of preliminary urine screen and subsequent test confirming presence of drug); Commonwealth v. Johnson, supra.
The judgment is reversed.9
So ordered.
General Laws c. 90, § 24G(i>), as amended by St. 1982, c. 376, § 2, provides, in pertinent part:
*72“Whoever, upon any way or in any place to which the public has a right of access or upon any way or in any place to which members of the public have access as invitees or licensees, operates a motor vehicle . . . while under the influence of intoxicating liquor, or of marihuana, narcotic drags, depressants or stimulant substances, all as defined in section one of chapter ninety-four C, or the vapors of glue, or whoever operates a motor vehicle recklessly or negligently so that the lives or safety of the public might be endangered and by any such operation causes the death of another person, shall be guilty of homicide by a motor vehicle and shall be punished by imprisonment in a jail or house of correction for not less than thirty days nor more than two and one-half years, or by a fine of not less than three hundred nor more than three thousand dollars, or both” (emphasis added).
When the judge addressed the venire, he informed the juiy (without comment or correction by the prosecutor) that the Commonwealth’s case was one of motor vehicle homicide by reason of negligent operation. He made no mention of impaired driving while under the influence of proscribed substances.
Nothing in the officers’ testimony suggested alcohol or other chemical intoxication; nor was there any reference to impaired driving because of the influence of drags or alcohol.
The defendant’s medical records were admitted without objection on the condition that they be redacted as to questions of liability.
At this point, a defense objection to the form of these questions was sustained. The record, however, contains only a partially intelligible sidebar conference following the prosecutor’s question concerning other uses for amphetamines. The defendant’s attorney complained that Dr. Brown had been listed and was disclosed as an autopsy witness, not “as an expert to testify as to the drugs of the defendant. . . .” Taken as a whole, we do not believe that this undifferentiated objection as to form preserved the issue. Further, as noted infra, there was no objection to the prosecutor’s closing argument; hence we apply the substantial risk of a miscarriage of justice standard.
The prosecutor’s closing argument included the following:
“And members of the jury, speed can kill. Speed kills, not only motor vehicle speed but drug speed, members of the jury. And in this case, sadly it did. And for this alone, members of the jury, just the amphetamines, I suggest to you, just on that one point the defendant was negligent — negligent driving, and the defendant is guilty of this crime.”
The defendant’s only objection to the prosecutor’s closing was directed to his assertion that the presence of amphetamines in the defendant’s system was shown conclusively by the testimony of Dr. Brown and the lab reports. While the sidebar conference was, in part, inaudible, the judge, implicitly affirming that the jury were entitled to evaluate the amphetamine evidence, instructed the jury that Dr. Brown had testified that it was possible to have false positive results on such lab tests.
“A defendant has a right to be definitely informed of the charges against [her] in order to facilitate her ability to prepare a defense and in order to prevent [her] being taken by surprise at trial.” Commonwealth v. Miranda, *77415 Mass. 1, 6 (1993). Where defense counsel is surprised through no fault of her own, and the unexpected evidence is important, there may be sufficient prejudice to warrant some form of relief. See Commonwealth v. Nester, 32 Mass. App. Ct. 983, 985-986 (1992).
At the same time, a defendant’s attorney has a responsibility to conduct a thorough investigation which, in many cases, involves retrieval and review of the defendant’s medical records, Commonwealth v. Alvarez, 433 Mass. 93 (2000), and may require examination of those records by an expert, as the data in medical records is usually not self-explanatory in the absence of accompanying medical testimony. See Commonwealth v. Kirkpatrick, 423 Mass. at 447-448. Competent defense counsel is also obligated, when confronted with surprise, to take steps to mitigate or eliminate the potential for prejudice. See Commonwealth v. Martin, 424 Mass. 301, 306 n.1 (1997) (a continuance might be sought to mitigate surprise in the midst of a trial); Commonwealth v. Giontzis, 47 Mass. App. Ct. 450, 460 (1999). Compare Commonwealth v. Sitko, 372 Mass. 305, 308 (1977). In evaluating prejudice, we determine what the defendant could or should have done to deal with it and “how the defense would have altered any tactic if it had been informed earlier.” Commonwealth v. Pizzotti, 27 Mass. App. Ct. 376, 381 (1989).
We grant relief for the reasons stated, without implying that the defendant’s attorney performed without contribution to the client’s prejudice.
It is unnecessary to address the defendant’s other claims of error.