Justice GUZMAN
delivered the opinion of the Court,
in which Justice HECHT, Justice WAINWRIGHT, Justice MEDINA, Justice JOHNSON, and Justice WILLETT joined, and in which Chief Justice JEFFERSON, Justice GREEN, and Justice LEHRMANN joined as to Parts I and II.A.
When circumstantial evidence is consistent with several possible medical conclusions, only one of which establishes that the defendant’s negligence caused the plaintiffs injury, an expert witness must explain why, based on the particular facts of the case, that conclusion is medically superior to the others. If the expert fails to give any reason beyond an unsupported opinion, the expert’s testimony is legally insufficient evidence of causation. In this case, we determine whether legally sufficient evidence supports the jury’s verdict in favor of the estate of Eloísa Casas1 against Rio Grande Regional Hospital (the Hospital).2 Following her admission to the Hospital with abdominal pain, doctors placed Casas on antibiotics used to treat and prevent certain intra-abdominal infections. Two days later she underwent major abdominal surgery and continued on the antibiotics for another five days, but the Hospital allowed the prescriptions to lapse for four-and-a-half days. The Hospital admits it should have continued the antibiotics but denies that the lapse caused Casas any additional pain. We hold that the Casases failed to present legally sufficient evidence that Casas suffered from an infection the omitted antibiotics would have treated. Accordingly, we reverse the court of appeals’ judgment and render judgment that the Casases take nothing.3
In a separate petition, Dr. Michael Jeli-nek, one of Casas’s treating physicians sued by the Casases, argues that the trial court should have granted his motion for sanctions and dismissal because the Casas-es’ expert report was deficient. We agree and hold that an award of attorney’s fees is proper. Therefore, we reverse and remand to the trial court for an award of attorney’s fees and costs.
[530]*530I. Background
In 2000, Eloísa Casas was diagnosed with colon cancer and underwent surgery, radiation, and chemotherapy. A year later, doctors told her that the cancer appeared to be in remission, and she thought she was cured. But on July 10, 2001, she was admitted to the Hospital with abdominal pains; she also had a fever and a mildly elevated white-blood-cell count, potentially indicating an infection. To treat this possible infection, her surgeon and primary physician, Dr. Carlos Garcia-Cantu, consulted with an infectious disease specialist at the Hospital, Dr. Michael Jeli-nek, who on July 11 prescribed two medications, Maxipime (a broad-spectrum antibiotic), and Flagyl (an antibiotic used to treat anaerobic bacteria).
The Hospital performed several diagnostic tests, which revealed abnormal collections of fluid in Casas’s abdomen. On July 13, she underwent major abdominal surgery during which Dr. Garcia-Cantu discovered that “fairly extensive” metastatic cancer had perforated Casas’s colon and allowed material to leak into her abdominal cavity, causing an intra-abdominal abscess. Dr. Garcia-Cantu drained the abscess, repaired Casas’s colon, and inserted a Jackson-Pratt drain to prevent further problems. Following the surgery, Dr. Garcia-Cantu continued the Maxipime and Flagyl prescriptions, and a culture of the removed abscess revealed an E. coli infection, which is effectively treated with Max-ipime. Casas received Maxipime and Fla-gyl for another five days, but hospital staff inadvertently failed to place a prescription renewal form on Casas’s chart, resulting in a four-and-a-half-day period between July 18 and 23 during which Casas did not receive either medication. Even so, Casas never tested positive for E. coli again and a culture of the incision site on July 18 instead grew Candida (a fungus) for which Diflucan (an antifungal) was prescribed. Then, on July 21, a second culture from a blood sample grew coagulase-negative staph, for which Vancomycin was prescribed.4 Neither Maxipime nor Flagyl would have treated the Candida or coagu-lase-negative staph infection.
On July 23, Dr. Garcia-Cantu noted an abscess in the wound, which he drained by removing the staples and opening the wound. The next day, records indicate that a foul smell was emanating from the wound site, and hospital staff brought fans into the room to dissipate the odor. When Dr. Jelinek learned of the lapsed prescription on July 23, he informed Casas and then prescribed different antibiotics, Leva-quin and Vancomycin. On July 25, after a CAT scan showed no abscess, Dr. Garcia-Cantu removed the drain. Casas left the Hospital on August 23, but she returned in early September and died two months later.
In May 2003, several members of Ca-sas’s family, including her husband and son, filed suit against the Hospital, Dr. Garcia-Cantu, and Dr. Jelinek. The plaintiffs claimed that the defendants’ negligence caused Eloísa Casas to “suffer grievous embarrassment and humiliation, as well as excruciating pain the remainder of her life which she would not have suffered to such degree or extent if properly diagnosed, treated and cared for.” The plaintiffs sought to recover damages for Casas’s injuries and mental anguish. They twice amended their petition, ultimately leaving the Casases as the sole plaintiffs.
[531]*531As required by former article 4590i § 13.01 of the Medical Liability and Insurance Improvement Act, see Tex.Rev.Civ. Stat. art. 4590i § 13.01,5 the Casases filed an expert report within 180 days of filing the original petition. In the report, Dr. John Daller opined that Dr. Garcia-Cantu and Dr. Jelinek were negligent in failing to discover that the antibiotics were not being given to Casas and that within “reasonable medical probability” this negligence resulted in a prolonged hospital stay and increased pain and suffering. Dr. Jelinek later filed a motion for sanctions and dismissal under article 4590i § 13.01(e), alleging that the expert report was deficient because, among other things, it failed to explain any causal connection between the negligence and the purported injury. The trial court denied the motion. Before trial began, however, the Casases nonsuited Dr. Jelinek and Dr. Garcia-Cantu.
At trial, Dr. Daller testified as the Ca-sases’ medical expert. During direct examination, he analyzed the Hospital’s daily patient notes regarding Casas and identified the significant events. He noted changes in Casas’s vital signs on July 21 and 22, such as increased heart rate and temperature, inflammation, and tenderness of the surgery site. Dr. Daller stated that “in medical probability” there was an infection in the abdomen, but on cross-examination he admitted that “there was no objective evidence present to demonstrate that intra-abdominal infection.” When reviewing the patient notes for July 24, which noted the presence of a foul smell, he suggested that the smell was consistent with an anaerobic infection that would be difficult to culture because anaerobic bacteria die when exposed to air. Dr. Carl Berkowitz, the Hospital’s expert, offered several other explanations for the smell, such as the Candida infection or dying tissue.
The Casases also called Casas’s relatives to testify about her condition. Consistent with Dr. Daller’s testimony, Casas’s son linked the smell with the opening of the wound to drain the abscess: “The odor that I noticed was after they had taken out the staples on her incision, and one day that I went to see her as soon as they opened the door the whiff of this putrid smell just engulfed me.” He also testified that Casas was upset upon learning that she had not received the antibiotics but was even more upset when the incision had to be opened and drained: “Well, after she was told and I was told that she wasn’t getting antibiotics, like I said, she was upset. What really upset her more was when they had to — they had to take out the staples out of her incision, and they had to open her incision up again.” Ca-sas’s husband testified that, while she was upset and did not trust the nurses or doctors after learning of the lapsed prescription, “she was still fighting. She ... wanted to beat this cancer she had.” The son testified that Casas did not lose hope until she witnessed the events of September 11, 2001, following her re-admission to the Hospital: “That’s why I remember that day so vividly in my mind because that was the turning point in my mom. She seemed to just give up, not fight, not want to fight anymore like she used to. And that was a very, very sad day.”
[532]*532The jury found that the negligence of the Hospital, Dr. Jelinek, and Dr. Garcia-Cantu proximately caused Casas’s injury. The jury apportioned ninety percent of the negligence to the Hospital, five percent to Dr. Jelinek, and five percent to Dr. Garcia-Cantu. It awarded $250,000 in damages to the Casases as compensation for Casas’s pain and mental anguish.
The Hospital appealed, arguing that the evidence was legally and factually insufficient to prove causation or damages for mental anguish. Dr. Jelinek also appealed, challenging the trial court’s denial of his motion for sanctions and dismissal. The court of appeals affirmed on all issues. — S.W.3d-.
II. Analysis
We address in turn the two issues raised in this appeal: the legal sufficiency of the causation evidence and the sufficiency of the Casases’ expert report.
A. Sufficiency of the Evidence
The facts of this case are unfortunate: a woman with advanced colon cancer underwent surgery to repair her cancer-perforated and infected colon, and in the course of treatment for her many symptoms the Hospital failed to renew her antibiotic prescriptions for a four-and-a-half-day period. The Hospital admits it should have continued the antibiotics. Even so, the plaintiff bears the burden to prove that the negligence caused an injury: “[A]t trial the plaintiff must establish two causal nexuses in order to be entitled to recovery: (a) a causal nexus between the defendant’s conduct and the event sued upon; and (b) a causal nexus between the event sued upon and the plaintiffs injuries.” Morgan v. Compugraphic Corp., 675 S.W.2d 729, 731 (Tex.1984). Only the second nexus is at issue here.
In City of Keller v. Wilson, we considered at length the parameters of legal sufficiency review, quoting with approval Chief Justice Calvert’s seminal article on the topic:
“No evidence” points must, and may only, be sustained when the record discloses one of the following situations: (a) a complete absence of evidence of a vital fact; (b) the court is barred by rules of law or of evidence from giving weight to the only evidence offered to prove a vital fact; (c) the evidence offered to prove a vital fact is no more than a mere scintilla; (d) the evidence establishes conclusively the opposite of the vital fact.
168 S.W.3d 802, 810 (Tex.2005) (quoting Robert W. Calvert, “No Evidence’’ and “Insufficient Evidence’’ Points of Error, 38 Tex. L.Rev. 361, 362-63 (I960)). ‘When the evidence offered to prove a vital fact is so weak as to do no more than create a mere surmise or suspicion of its existence, the evidence is no more than a scintilla and, in legal effect, is no evidence.” Kindred v. Con/Chem, Inc., 650 S.W.2d 61, 63 (Tex.1983). The same is true when the vidence equally supports two alternatives: “ “When the circumstances are equally consistent with either of two facts, neither fact may be inferred.’ ” City of Keller, 168 S.W.3d at 813 (quoting Tubelite, a Div. of Indal, Inc. v. Risica & Sons, Inc., 819 S.W.2d 801, 805 (Tex.1991)). When considering such cases, “we must ‘view each piece of circumstantial evidence, not in isolation, but in light of all the known circumstances,’ ” id. at 813-14 (quoting Lozano v. Lozano, 52 S.W.3d 141, 167 (Tex.2001) (per curiam)), and we “must consider not just favorable but all the circumstantial evidence, and competing inferences as well.” Id. at 814.
To meet the legal sufficiency standard in medical malpractice cases “plaintiffs are required to adduce evidence of a [533]*533‘reasonable medical probability’ or ‘reasonable probability’ that their injuries were caused by the negligence of one or more defendants, meaning simply that it is ‘more likely than not’ that the ultimate harm or condition resulted from such negligence.” Kramer v. Lewisville Mem’l Hosp., 858 S.W.2d 397, 399-100 (Tex.1993) (citations omitted). Thus, we examine the record to determine if the Casases presented legally sufficient evidence that “in reasonable medical probability” the Hospital’s negligence caused Casas additional pain and suffering.
When distilled to its essence, the Casas-es’ claim is predicated on the presence of an infection — treatable by the lapsed antibiotics — that caused Casas pain and mental anguish above and beyond that caused by the cancer, the surgery, and the other known infections. The absence of an infection treatable by Maxipime and Flagyl would undermine the Casases’ claim, for then the prescription lapse would amount to an unfortunate, but harmless, occurrence. The Hospital argues that the Ca-sases presented no evidence that the Hospital’s negligence caused such an infection. The Casases’ expert admitted there is no direct evidence of an anaerobic infection, leaving the jury to consider the circumstantial evidence and make proper inferences from it. In reviewing the record, we initially decide if jurors can determine causation under these facts unaided, by expert testimony — that is, whether lay testimony regarding causation is legally sufficient.
1. Lay Testimony of Causation
Lay testimony may be used as evidence of causation in certain circumstances, but “[w]hen expert testimony is required, lay evidence supporting liability is legally insufficient.” City of Keller, 168 S.W.3d at 812. In medical malpractice cases, expert testimony regarding causation is the norm: “The general rule has long been that expert testimony is necessary to establish causation as to medical conditions outside the common knowledge and experience of jurors.” Guevara v. Ferrer, 247 S.W.3d 662, 665 (Tex.2007); see also Bowles v. Bourdon, 148 Tex. 1, 219 S.W.2d 779, 782 (1949) (“It is definitely settled with us that a patient has no cause of action against his doctor for malpractice, either in diagnosis or recognized treatment, unless he proves by a doctor of the same school of practice as the defendant: (1) that the diagnosis or treatment complained of was such as to constitute negligence and (2) that it was a proximate cause of the patient’s injuries.”). We have allowed lay evidence to establish causation “in those cases in which general experience and common sense will enable a layman to determine, with reasonable probability, the causal relationship between the event and the condition.” Morgan, 675 S.W.2d at 733 (citing Lenger v. Physician’s Gen. Hosp., Inc., 455 S.W.2d 703, 706 (Tex.1970)). Care must be taken to avoid the post hoc ergo propter hoc fallacy, that is, finding an earlier event caused a later event merely because it occurred first. Stated simply, correlation does not necessarily imply causation. As we noted in Guevara, “[ejvidence of an event followed closely by manifestation of or treatment for conditions which did not appear before the event raises suspicion that the event at issue caused the conditions. But suspicion has not been and is not legally sufficient to support a finding of legal causation.” 247 S.W.3d at 668.
When lay testimony is credited as evidence of causation, it usually highlights a connection between two events that is apparent to a casual observer. In Morgan, for example, a previously healthy employee, upon exposure to leaking chemicals, suffered watering of the eyes, blurred [534]*534vision, headaches, and swelling of the breathing passages. 675 S.W.2d at 733. In such a circumstance, lay testimony sufficed to connect the specific injury to the negligence with no evidence of causation beyond the leaking chemicals. Id. Likewise in Guevara, we stated that determining causation of “certain types of pain, bone fractures, and similar basic conditions” following an automobile accident was within the competence of lay jurors. 247 S.W.3d at 668. But we held that expert testimony was required to prove that a patient’s medical expenses resulted from the accident, noting that “[pjatients in hospitals are often treated for more than one condition brought on by causes independent of each other.” Id. at 669. These cases illustrate this basic premise: “[N]on-expert evidence alone is sufficient to support a finding of causation in limited circumstances where both the occurrence and conditions complained of are such that the general experience and common sense of laypersons are sufficient to evaluate the conditions and whether they were probably caused by the occurrence.” Id. at 668.
The present case does not fall within this rule. Unlike in Morgan, an otherwise healthy person did not suddenly experience health difficulties following the defendant’s negligent conduct when the plaintiffs symptoms were reasonably attributable to the negligence and to nothing else. Rather, a patient with terminal colon cancer did not receive antibiotics for four-and-a-half days following major abdominal surgery and after having received the medications for eight days. There is no direct evidence that she suffered from an infection treatable by the omitted antibiotics, but there is evidence that she had two other infections that accounted for all of her symptoms during that time. Given Casas’s medical condition, expert testimony was crucial to link the prescription lapse to an infection causing additional pain and suffering beyond what she would otherwise have experienced. See Kaster v. Woodson, 123 S.W.2d 981, 983 (Tex.Civ. App.-Austin 1938, writ refd) (“What is an infection and from whence did it come are matters determinable only by medical experts.”); see also Hart v. Van Zandt, 399 S.W.2d 791, 792 (Tex.1966) (“In determining negligence in a case such as this, which concerns the highly specialized art of treating disease, the court and jury must be dependent on expert testimony. There can be no other guide, and where want of skill and attention is not thus shown by expert evidence applied to the facts, there is no evidence of it proper to be submitted to the jury.”).
The Casases point to testimony by Ca-sas’s husband and son to support their argument that she deteriorated rapidly after discovering she did not receive the antibiotics. But this characterization overstates the evidence. While Casas’s husband testified she was upset and did not trust her doctors following the discovery, she was still determined to fight her cancer. The son also observed Casas’s anger and lack of trust but testified that the opening of her wound, which occurred the same day she learned of the lapse, upset her even more. As Dr. Daller admitted, Candida likely caused the abscess that required Dr. Garcia-Cantu to drain the wound. Further, based on his experience at Casas’s bedside, her son pinpointed the tragic events of September 11, 2001, and their effect on his mother as the turning point in her mental state. The latter event was some seven weeks after discovery of the lapsed prescriptions and after Casas’s discharge from and re-admission to the Hospital. This evidence does not bear out the Casases’ claim of a marked shift in Casas’s mental resilience following the omission of the medications.
[535]*535More importantly, Casas’s husband and son were unable to precisely identify the cause of her suffering. While they could accurately describe her discomfort, they were unable to say if it was the cancer, the surgery, the other infections, or the lapse that caused it. Even testimony that Casas suffered after learning of the omission raises no more than a mere suspicion of causation, and that is not enough, see Guevara, 247 S.W.3d at 668, particularly in light of the evidence that Casas thought she was cured of cancer before the surgery and then learned that not only was it “back with a vengeance,” it was terminal. The testimony of Casas’s husband and son is evidence of her suffering, but not of its cause. Thus, we hold that the lay testimony presented by the Casases is legally insufficient to establish that the Hospital’s negligence caused Casas additional pain and suffering.
2. Expert Testimony
The Casases also presented expert testimony regarding causation. The Casases’ expert, Dr. Daller, testified that the Hospital’s negligence “in medical probability” caused Casas additional pain and suffering. He based this opinion on the presence of an intra-abdominal infection that could have been treated using Maxi-pime and Flagyl. Admitting that no direct evidence indicated such an infection, Dr. Daller pointed to various circumstantial indicators that suggested an infection. These indicators were primarily Casas’s changed vital signs, such as fever and increased heart rate: “Well, given the fact that two to three days after the antibiotics had been mistakingly [sic] stopped her fever curve went up and her heart rate went up, to me that suggests the presence of on going [sic] infection.”6 But on cross-examination, he conceded these data were equally consistent with two other infections cultured from Casas’s incision and blood — Candida and coagulase — negative staph — neither of which is treatable by Maxipime or Flagyl:
Q. Now, Candida, infection of a wound like this, they can cause high temperatures. Correct?
A. Fungal infections can cause a high temperature, yes.
Q. It can cause increased heart rate?
A. That is correct.
Q. And inflammation?
A. That is correct.
Q. Pain?
A. That is correct.
Q. How about an abscess?
A. It caused or is part of the abscess in that wound that was present, that wound infection that needed to be opened.
Q. So when Doctor Garcia went in on 7/23 ... and drained that wound at bedside that abscess was within a reasonable degree of medical probability caused by the Candida?
[536]*536A. That was one of the organisms that was there. It was the organism that was cultured. That is correct.
[[Image here]]
Q. ... This coagulase negative staph causes fever?
A. Correct.
Q. Increased heart rate?
A. The fever will cause increased heart rate.
[[Image here]]
Q. It can cause pain?
A. Depending upon the site. Correct,
Q. Okay. All of these things can be caused by coagulase negative staph and Candida, which we know were present 7/18 through 7/23, the time period she did not get antibiotics?
A. That’s correct.
Q. Neither one would have been killed by Maxipime or Flagyl?
A. That’s correct.
It is not enough for an expert simply to opine that the defendant’s negligence caused the plaintiffs injury. The expert must also, to a reasonable degree of medical probability, explain how and why the negligence caused the injury. We have rejected expert opinions not grounded in a sound evidentiary basis: “[I]f no basis for the opinion is offered, or the basis offered provides no support, the opinion is merely a conclusory statement and cannot be considered probative evidence, regardless of whether there is no objection. ‘[A] claim will not stand or fall on the mere ipse dixit of a credentialed witness.’ ” City of San Antonio v. Pollock, 284 S.W.3d 809, 818 (Tex.2009) (quoting Burrow v. Arce, 997 S.W.2d 229, 235 (Tex.1999)); see also Whirlpool Corp. v. Camacho, 298 S.W.3d 631, 637 (Tex.2009) (“Con-clusory or speculative opinion testimony is not relevant evidence because it does not tend to make the existence of material facts more probable or less probable.”). When the only evidence of a -vital fact is circumstantial, the expert cannot merely draw possible inferences from the evidence and state that “in medical probability” the injury was caused by the defendant’s negligence. The expert must explain why the inferences drawn are medically preferable to competing inferences that are equally consistent with the known facts. Thus, when the facts support several possible conclusions, only some of which establish that the defendant’s negligence caused the plaintiffs injury, the expert must explain to the fact finder why those conclusions are superior based on verifiable medical evidence, not simply the expert’s opinion. See Lenger, 455 S.W.2d at 707 (“[E]xpert testimony that the event is a possible cause of the condition cannot ordinarily be treated as evidence of reasonable medical probability except when, in the absence of other reasonable causal explanations, it becomes more likely than not that the condition did result from the event.”); Hart, 399 S.W.2d at 792 (“The burden of proof is on the plaintiff to show that the injury was negligently caused by the defendant and it is not enough to show the injury together with the expert opinion that it might have occurred from the doctor’s negligence and from other causes not the fault of the doctor. Such evidence has no tendency to show that negligence did cause the injury.”).
By conceding that Casas’s symptoms were consistent with infections not treatable by Maxipime or Flagyl, Dr. Daller undermined his conclusion that an undetected infection was also present. While it is possible that Casas did have such an infection, its presence can only be inferred from facts that are equally consistent with the Candida and coagulase-negative staph infections. “ ‘When the circumstances are [537]*537equally consistent with either of two facts, neither fact may be inferred.’ ” City of Keller, 168 S.W.3d at 813 (quoting Tubelite, 819 S.W.2d at 805). Here, objective data — the cultures — support the Candida and staph infections but not the supposed anaerobic infection.7
Based on the record evidence, an anaerobic infection cannot be proved or disproved. It is equally plausible that Ca-sas had such an infection or that she did not. Dr. Daller opined that she did, but he did not explain why that opinion was superior to the opposite view. Such evidence raises no more than a possibility of causation, which is insufficient. As we said in Bowles v. Bourdon, “ ‘[t]he proof must establish causal connection beyond the point of conjecture. It must show more than a possibility. Verdicts must rest upon reasonable certainty of proof. Where the proof discloses that a given result may have occurred by reason of more than one proximate cause, and the jury can do no more than guess or speculate as to which was, in fact, the efficient cause, the submission of such choice to the jury has been consistently condemned by this court and by other courts.’” 219 S.W.2d at 785 (quoting Ramberg v. Morgan, 209 Iowa 474, 218 N.W. 492, 498-99 (1928)).
The Casases argue that the foul smell, which is consistent with an anaerobic infection, is strong evidence of such an infection. Looking at the patient notes for July 24, Dr. Daller commented on the smell:
A. The text says something about drainage to the abdomen with moderate amount of drainage. And it says that it is foul smelling.
[[Image here]]
Q. The [previous notes] that I remember that we have gone over didn’t say anything about foul smelling?
A. That’s correct. They were just described as I recall as being purulent and looking like puss [sic].
Q. What does that mean when it says “foul smelling”?
A. When you have foul smelling, it suggests that the organism is an anaerobe. In other words, one of those bacteria that didn’t need oxygen in order to grow that, for example, Flagyl would treat.
Q. Okay. Does that give you clinical evidence that had she been continued on Maxipime and Flagyl that they would have had some effect with regards to the condition as we see it on the 24th?
A. Well, like I said, most anaerobes are sensitive or susceptible to Flagyl. And she had previously been on Flagyl and at this time she is not. So I would have expected that that would be an appropriate antibiotic that would have covered the organism that’s causing that foul smell.
Dr. Berkowitz, the Hospital’s expert, offered several other explanations for the smell, including necrotic tissue, dead cancer tissue, and the Candida infection.8 As [538]*538noted, Casas’s son noticed the smell after the incision was opened to drain the abscess, which Dr. Daller admitted was likely-caused by Candida.
Here again, there are competing explanations for the smell, which amounts to no more than circumstantial evidence of some kind of infection or possibly dying tissue. Because there is no direct evidence of the infection and the circumstantial evidence is meager, we “must consider not just favorable but all the circumstantial evidence, and competing inferences as well.” City of Keller, 168 S.W.3d at 814. Courts should not usurp the jury’s role as fact finder, nor should they question the jury’s right to believe one witness over another. But when reviewing a verdict for sufficiency of the evidence, courts need not — indeed, must not — defer to the jury’s findings when those findings are not supported by credible evidence. When the evidence compels the jury to guess if a vital fact exists, a reviewing court does not undermine the jury’s role by sustaining a no-evidence challenge. The evidence in this case — being consistent with an anaerobic infection that was treatable by Flagyl, a fungal infection that was not, or even with dying tissue, cancerous or otherwise — did not provide the jury a reasoned basis from which to infer the presence of a negligence — induced infection. Because the jury could not reasonably infer an infection caused by the Hospital’s negligence, we agree with the Hospital that no evidence supports the jury’s verdict.
We understand the Casas family’s predicament and frustration at the Hospital’s conduct, and we recognize the difficulty of proving that the lapsed prescriptions caused a painful infection. But the Casas-es shouldered that burden and must prove the causal link with reasonable certainty. In that quest, the Casases offered the testimony of Dr. Daller, but he did not explain why an undetected, anaerobic infection is medically more probable than one based on the known infections and the dying tissue, leaving the jury to guess if the lapsed prescriptions caused additional pain and suffering. Without probative medical testimony that the lapse caused— by means of an infection treatable by Max-ipime and Flagyl — more pain than the cancer, the surgery, and the other infections already inflicted, there is no legally sufficient evidence of causation. Dr. Daller did not provide that causal link; accordingly, we hold that his testimony is legally insufficient to support the jury’s verdict. Because the Casases failed to prove causation, we reverse the judgment of the court of appeals and render judgment that the Casases take nothing.
B. Adequacy of the Expert Report
In his petition, Dr. Jelinek raises a single issue: whether the trial court abused its discretion by denying his motion for sanctions and dismissal because the Casases’ expert report was deficient under former article 4590i § 13.01, the statute in effect at the time. See Tex.Rev. Civ. Stat. art. 4590i § 13.01. Article 4590i required the report to provide “a fair summary of the expert’s opinions as of the date of the report regarding applicable standards of care, the manner in which the care rendered by the physician or health care provider failed to meet the standards, and the causal relationship between that [539]*539failure and the injury, harm, or damages claimed.” Id. § 13.01(r)(6). “If a plaintiff timely files an expert report and the defendant moves to dismiss because of the report’s inadequacy, the trial court must grant the motion ‘only if it appears to the court, after hearing, that the report does not represent a good faith effort to comply with the definition of an expert report in Subsection (r)(6) of this section.’ ” Bowie Mem’l Hosp. v. Wright, 79 S.W.3d 48, 51-52 (Tex.2002) (per curiam) (quoting § 13.01(i)). Dismissal for failure to serve an adequate expert report also carried mandatory sanctions, requiring an award to the defendant of his costs and attorney’s fees against the plaintiff or the plaintiffs attorney. See Am. Transitional Care Ctrs. of Tex., Inc. v. Palacios, 46 S.W.3d 873, 877 (Tex.2001) (citing § 13.01(e)).
We have defined a “good-faith effort” as one that provides information sufficient to (1) “inform the defendant of the specific conduct the plaintiff has called into question,” and (2) “provide a basis for the trial court to conclude that the claims have merit.” Wright, 79 S.W.3d at 52 (citing Palacios, 46 S.W.3d at 879). All information needed for this inquiry is found within the four corners of the expert report, which need not “marshal all the plaintiffs proof’ but must include the expert’s opinion on each of the three main elements: standard of care, breach, and causation. Id. Importantly for this case, the “report cannot merely state the expert’s conclusions about these elements,” but “ ‘the expert must explain the basis of his statements to link his conclusions to the facts.’ ” Id. (quoting Earle v. Ratliff, 998 S.W.2d 882, 890 (Tex.1999)). “A report that merely states the expert’s conclusions about the standard of care, breach, and causation” does not fulfill the two purposes of a good-faith effort. Palacios, 46 S.W.3d at 879.
We review the trial court’s grant or denial of a motion for sanctions and dismissal under the abuse-of-discretion standard. Palacios, 46 S.W.3d at 877-78. A district court “abuses its discretion if it acts in an arbitrary or unreasonable manner without reference to any guiding rules or principles.” Wright, 79 S.W.3d at 52 (citing Downer v. Aquamarine Operators, Inc., 701 S.W.2d 238, 241-42 (Tex.1985)).
Dr. Jelinek argues that the Casases’ report is deficient in two ways, failing (1) to state the applicable standard of care, and (2) to provide more than conclusory statements of causation. We focus on the latter. Dr. Daller’s report concluded that Dr. Jelinek’s breach of the appropriate standard of care in “reasonable medical probability, resulted in a prolonged hospital course and increased pain and suffering being experienced by Ms. Casas.” Aside from repeating essentially the same phrase twice more, the report says nothing more regarding causation. The Casases argue this statement is sufficient to meet the good-faith requirement. We disagree.
An expert cannot simply opine that the breach caused the injury. Stated so briefly, the report fails the second Palacios element — it does not give the trial court any reasonable basis for concluding that the lawsuit has merit. See 46 S.W.3d at 879. An expert’s conclusion that “in medical probability” one event caused another differs little, without an explanation tying the conclusion to the facts, from an ipse dixit, which we have consistently criticized. See Pollock, 284 S.W.3d at 818 (citing Burrow, 997 S.W.2d at 235); Earle, 998 S.W.2d at 890 (“An expert’s simple ipse dixit is insufficient to establish a matter; rather, the expert must explain the basis of his statements to link his conclusions to the facts.”). Instead, the expert must go further and explain, to a reasonable de[540]*540gree, how and why the breach caused the injury based on the facts presented. While we have said that no “magical words” need be used to meet the good-faith requirement, mere invocation of the phrase “medical probability” is likewise no guarantee that the report will be found adequate. See Wright, 79 S.W.3d at 53.
Under these standards, the Casases’ report is conclusory on causation. It offers no more than a bare assertion that Dr. Jelinek’s breach resulted in increased pain and suffering and a prolonged hospital stay. Beyond that statement, the report offers no explanation of how the breach caused the injury. Again, the plaintiff need not marshal all of his proof in the report, but he must include sufficient detail to allow the trial court to determine if the claim has merit. Because the Casases’ report lacks any explanation linking the expert’s conclusion to the relevant facts, we hold that the trial court abused its discretion by denying Dr. Jelinek’s motion and the court of appeals erred by affirming that ruling.9 See id. at 52. Accordingly, we remand the case to the trial court for an award of attorney’s fees and costs10 under former article 4590i § 13.01(e) against the Casases and their counsel.11
[541]*541III. Conclusion
For the foregoing reasons, we reverse the court of appeals’ judgment, render judgment that the Casases take nothing, and remand to the trial court for an award of Dr. Jelinek’s attorney’s fees and costs consistent with this opinion.
Chief Justice JEFFERSON filed an opinion, dissenting in part, in which Justice GREEN and Justice LEHRMANN joined.
Justice LEHRMANN filed an opinion, dissenting in part.