*691 OPINION

G. BARRY ANDERSON, Judge.

In this medical-malpracticé action, the district court held a Frye-Mack hearing and excluded appellants’ experts’ opinions because their opinions lacked a reliable and generally accepted scientific basis. Because appellants did not have expert testimony to support the medical-negligence claim, the district court granted summary judgment. Appellants argue that the district court erred. We affirm.

FACTS

In April 2000, Dr. Susan Evans diagnosed appellant Kari McDonough with multiple sclerosis. Dr. Evans is a neurologist and is McDonough’s treating physician. On October 16, 2001, Dr. Evans sent an order to Allina Infusion Therapy Services to administer an intravenous infusion of 200 mg/kg per day of immunoglobulin (“IVIg”) to McDonough for two days. Dr. Evans’s order did not specify the infusion rate at which the IVIg was to be infused.

Pharmacist Ceci Do, an employee of respondent Allina Health Systems, prepared a plan for the administration of the IVIg. Based on McDonough’s weight of approximately 200 pounds, Do calculated a plan for the administration of Baxter Gamma-gard 5% IVIg, which resulted in the infusion of a total of 400 ml over two and a half hours, starting at an infusion rate of 25 ml/hr and increasing at 15 minute intervals to 50 ml/hr, 75 ml/hr, 100 ml/hr, 150 ml/hr, 200 ml/hr, and then 250 ml/hr until the infusion was completed.

On October 17, 2001, Allina Home infusion nurse Patricia Beaver administered 400 ml of IVIg over approximately three and a half hours to McDonough. McDon-ough developed chills while the infusion rate was running at 150 ml/hr. Beaver stopped the infusion and called Do. Do advised Beaver to reduce the infusion rate to 125 ml/hr and see if the chills were still present at the lower rate. The infusion rate was slowed to 75 ml/hr and thereafter increased to 100 ml/hr until the end of the infusion. McDonough’s chills disappeared after the reduction of the infusion rate. At the end of the infusion, McDonough experienced a fever. Beaver instructed McDonough to obtain over-the-counter Tylenol and Benadryl and to take them 30 minutes before the infusion scheduled for the next day.

On October 18, 2001, Beaver administered the second day of infusion to McDon-ough. McDonough took Tylenol and Be-nadryl before the infusion and did not experience chills or fever during the infusion. Do’s plan for October 18 was for an infusion of 400 ml over four hours and 37 minutes. Do prepared the bag of IVIg to be infused on October 18 and placed a label on the bag stating, “DO NOT EXCEED MAX RATE OF 100 ML/HR DUE TO PATIENT’S INTOLERANCE TO HIGHER RATE.” Do testified that this schedule was meant as a suggested infusion rate for Beaver and that, in her opinion, it would have been appropriate for Beaver to use any infusion rate up to the maximum recommended by the manufacturer of 400 ml/hr as long as the patient was comfortable. Baxter’s recommended rate of infusion as set out in the package information sheet is as follows:

It is recommended that initially a 5% solution be infused at a rate of 0.5 mL/ *692 kg/Hr. If infusion at this rate and concentration causes the patient no distress, the administration rate may be gradually increased to a maximum rate of 4 mL/kg/Hr. Patients who tolerate the 5% concentration at 4 mL/kg/Hr can be infused with the 10% concentration at 0.5 mL/kg/Hr. If no adverse effects occur, the rate can be increased gradually up to a maximum of 8 mL/kg/Hr.

The parties dispute how long the October 18 infusion actually lasted. Appellants contend that the infusion lasted approximately one hour and 45 minutes, with half infused during the last 15 minutes at a rate equal to or greater than 800 ml/hr. Beaver contends the infusion lasted two and a half hours. McDonough testified that she felt fine following the infusion and did not experience any adverse effects that day.

On October 19, 2001, McDonough awoke with a severe headache in the right temple area and was nauseated. An MRI showed that McDonough had suffered a stroke, causing an acute non-hemorrhagic infarction of the distribution of the right middle cerebral artery.

Appellants filed suit against respondents in August 2002, alleging that Beaver was negligent in performing the IVIg infusion by administering the IVIg at a rate beyond that authorized and ordered, and, as a result of this negligent infusion, McDon-ough suffered a stroke. Appellants claim that had Beaver followed the plan for infusion set out by Do, McDonough’s stroke would not have occurred. In support of appellants’ claims, appellants offered expert opinions from Dr. Evans and Rodney Richmond, a pharmacist retained by appellants.

Richmond opines that McDonough’s stroke was probably related to an infusion rate that exceeded the patient’s tolerance. Richmond bases his opinion on his use of an adverse-drug-reaction-probability-assessment tool, his interpretation of reports in the literature, the manufacturer’s recommendations, and the care plan prepared by the physician and pharmacist.

Dr. Evans is a board-certified neurologist and treats patients with IVIg, and she treats stroke victims. Dr. Evans has not published any articles related to the use of IVIg and its effects. Dr. Evans opines that McDonough’s stroke was directly caused by a rapid rate of infusion of IVIg on October 18, 2001. The primary basis for Dr. Evans’s opinion is that the chills and fever experienced by McDonough on October 17 were a warning of McDon-ough’s intolerance to the IVIg and an indication that McDonough might have a stroke if the rate of infusion on October 18 exceeded 100 ml/hr. Dr. Evans also bases her opinion on her own knowledge of abrupt-increased and rapid-infusion rates, on the medical tests performed on McDon-ough, and on her collaboration with colleagues. Some of the testing that Dr. Evans performed on McDonough to rule out other causes of the stroke included checking blood work, homocystine levels, and performing a spinal tap and MRI scan of the brain to check for vasculitis or sten-osis. An electrocardiogram showed a patent foramen ovale, 1 but Dr. Evans conferred with the cardiologist to eliminate this as a cause. The medical literature that Dr. Evans primarily relies on is the Grillo study and the Baxter Alert.

After the incident at issue, a study by J.A. Grillo and others was published in the November 2001 issue of Neurology concerning use of rapid infusion of IVIg at *693 rates reaching 800 ml/hr. The Grillo report notes 89 adverse events in 341 rapid infusions in 50 patients, 3.5% of which were considered “major.” Thereafter, Baxter sent a letter dated March 25, 2002, to health-care providers concerning the subject of “Thrombotic Events and Immune Globulin Intravenous.” The Baxter Alert states the following:

There is clinical evidence of a possible association between Immune Globulin Intravenous (Human) (IGIV) administration and thrombotic events.[ 2 ] The exact cause of this is unknown; therefore, caution should be exercised in the prescribing and infusion of IGIV in patients with a history of cardiovascular disease or thrombotic episodes.

From both the medical literature and our internal pharmacovigilanee/quality assurance program, we continue to receive reports describing serious thrombotic (vascular occlusive) events possibly associated with the infusion of immune globulin intravenous (IGIV). Analysis of these events indicates that the etiology is complex and multi-variant and the cause of this association is not clearly understood.

However, our own recent analysis of serious adverse events reported via pharmacovigilance, has identified rapid infusion of immune globulin intravenous as a possible risk factor.

[[Image here]]

In the event where there is a possible risk of a thrombotic event, we strongly recommend that the infusion concentration be no more than 5%, and the infusion rate should be initiated no faster than 0.5 milliliter per kilogram body weight per hour and advanced slowly only if well tolerated to a maximum rate of 4 milliliter per kilogram body weight per hour. In other words, the rate of infusion and percent of the solution concentration should be flexible and targeted to the • safety of the patient rather than convenience.

The Baxter Alert also states that “chest pain, myocardial infarction, congestive cardiac failure, severe headache requiring hospitalization, pulmonary embolism and ‘transfusion related acute lung injury’ ... are serious events almost certainly directly related to the rapid infusion protocol (reaching as high as 800 ml/hr) in what is essentially an at-risk population.”

Respondents challenged the admissibility of the opinions of Dr. Evans and Richmond and moved to exclude their expert testimony. In support of respondents’ motion, they offered the testimony of neurologists Dr. Thomas Brannagan and Dr. Kenneth Gorson. Both doctors have published numerous articles on the use of IVIg and are nationally recognized as experts on IVIg and its adverse affects.

Dr. Brannagan testified that Beaver complied with the applicable standards of care and infused the IVIg at a rate that was reasonable and appropriate. Dr. Brannagan also opined that the chills and fever experienced by McDonough had no relation to her stroke. He explained that the mechanism of how IVIg causes stroke or other thromboembolic events is unknown, and medical science has only postulated a number of hypotheses concerning the mechanism.

*694 Dr. Gorson also testified that Beaver’s administration of the IVIg was within the accepted standards of care. Dr. Gorson, a co-author of the Grillo study, explained that in his study involving 341 rapid infusions, there were no resulting strokes. Dr. Gorson also explained that the findings in the study do not provide evidence that rapid infusion up to 800 ml/hr is related to stroke or thromboembolic events, as distinguished from complications such as congestive heart failure related to the volume of IVIg. Dr. Gorson concluded that the chills and fever McDonough experienced probably represented an infusion-related side effect but that such side effects are not predictive of stroke or other thrombo-embolic events.

Both doctors testified that there is no reliable scientific evidence that proves with reasonable medical certainty that McDon-ough’s stroke was caused by the rate at which the IVIg was infused on October 18 or that infusion at a lower rate would not have caused McDonough’s stroke. They also testified that there is no generally reliable scientific evidence that the rate of infusion of IVIg causes stroke or other thromboembolic events and that it is not generally accepted by experts in the field that the rate of infusion causes stroke or other thromboembolic events. Both doctors testified that McDonough had a number of risk factors predisposing her to stroke,' including obesity, high blood pressure, a patent foramen ovale, migraine headaches, and a history of possible transient ischemic attack. The doctors explained that there is no way of knowing to what extent any of these risk factors may have contributed to McDonough’s stroke.

Respondents filed a motion to exclude appellants’ expert testimony and for summary judgment. After a Frye-Mack hearing, the district court determined that appellants’ experts’ opinions were not sufficiently established or generally accepted and did not have foundational reliability, and it granted respondents’ motion to exclude this testimony. The district court also concluded that without expert opinion, appellants failed to establish a prima facie case of medical negligence and granted respondents’ motion for summary judgment. This appeal followed.

ISSUES

I. Did the district court err or abuse its discretion in excluding appellants’ experts’ opinions?

II. Did the district court err in granting respondents’ motion for summary judgment?

ANALYSIS

A. Frye-Mack Standard

Minnesota adheres to the Frye-Mack standard to determine the admission of expert testimony based on scientific techniques and principles. Goeb v. Tharaldson, 615 N.W.2d 800, 814 (Minn.2000). Under this two-pronged standard, the proponent of scientific evidence must establish the proper foundation for admissibility by showing that the scientific theory is generally accepted in the applicable medical or scientific community, Frye v. United States, 293 F. 1013, 1014 (D.C.Cir.1923), and that the principles and methodology used are reliable. State v. Mack, 292 N.W.2d 764, 768 (Minn.1980). Whether a particular theory or technique satisfies the first prong as generally accepted in the relevant scientific or medical community is a question of law, which we review de novo. Goeb, 615 N.W.2d at 815. The second prong, that the theory is reliable and trustworthy based upon well-recognized scientific principles and independent vali *695 dation, is reviewed under an abuse-of-discretion standard. Id.

The district court determined that appellants’ experts’ opinions are neither generally accepted nor reliable, and do not satisfy either prong of the Frye-Mack standard. As to reliability, appellants argue that their experts’ opinions should not have been excluded because medical opinion regarding causation, based on proper differential diagnosis, should be sufficiently reliable to satisfy Goeb, even without citing published studies. See, e.g., Kudobeck v. Kroger Co., 338 F.3d 856, 862 (8th Cir.2003) (stating that “there is no requirement ‘that a medical expert must always cite published studies on general causation in order to reliably conclude that a particular object caused a particular illness’ ”) (quoting Heller v. Shaw Indus., Inc., 167 F.3d 146, 155 (3d Cir.1999)); see also Westberry v. Gislaved Gummi AB, 178 F.3d 257, 262-63 (4th Cir.1999) (stating that a reliable differential diagnosis alone provides valid foundation for causation opinion). 3

But if respondents “point[] to a plausible alternative cause and the doctor offers no explanation for why he or she has concluded that was not the sole cause, that doctor’s methodology is unreliable.” Heller, 167 F.3d at 156 (quotation omitted); see also Turner v. Iowa Fire Equip. Co., 229 F.3d 1202, 1209 (8th Cir.2000) (excluding treating physician’s causation opinion because differential diagnosis did not properly rule out all other possible causes).

Here, Dr. Brannagan explained numerous hypotheses concerning the mechanism by which IVIg may cause stroke or other thromboembolic events, other than the rate of infusion, and noted certain risk factors associated with McDonough that could have caused her stroke. When asked why certain of these hypotheses were not factors in McDonough’s case, Dr. Evans explained why some were excluda-ble, did not address others, and only stated, “I still don’t feel it’s a factor” and a conclusory “no” when asked whether mi-crolevel spasms or a transient ischemic event or accident could have caused Mc-Donough’s stroke, with no further explanation. Because the differential diagnosis of Dr. Evans did not rule out all other hypotheses, or at least explain why the other conceivable causes are excludable, her differential diagnosis is not sufficiently reliable to be used for the purpose of proving causation. See Heller, 167 F.3d at 156; see also Lauzon v. Senco Prods., Inc., 270 F.3d 681, 694 (8th Cir.2001) (“After discounting obvious alternatives through scientific testing ... [appellant] need only be able to explain why other conceivable causes are excludable.”). 4

In Kudabeck, the Eighth Circuit allowed a treating chiropractor’s opinion to be admitted based on a differential diagnosis. 338 F.3d at 861-63. But, significantly, there was no independent analysis challenging the chiropractor’s conclusions as suspect, and no medical experts testified disputing the chiropractor’s theory. Id. at *696 859, 861. In contrast, here, while Dr. Evans did perform various medical tests to attempt to rule out causes of McDonough’s stroke other than the infusion rate, respondents disputed the differential diagnosis by calling medical experts to testify challenging Dr. Evans’s theory.

Appellants’ experts further rely on the Grillo report and the Baxter Alert to show that there is a direct relationship between infusion rates and strokes. But out of the 50 patients infused at high rates in the Grillo study, not one suffered a stroke. Also, we note that the Baxter Alert was not distributed until well after the infusion at issue in this case. Even so, these two pieces of literature are at best equivocal as to a direct relationship between infusion rates and strokes and therefore at most demonstrate that there is not yet a general consensus in the scientific community whether the infusion of IVIg is related to stroke or other thromboembolic events. Coincidence is not causation, and the testimony from both of respondents’ experts confirms that there is no reliable scientific evidence that the rate of infusion of IVIg causes stroke or other thromboembolic events and that Dr. Evans’s theory is not generally accepted by experts in the field.

Furthermore, Dr. Evans testified that the exact mechanism by which IVIg causes major side effects such as stroke is not known with scientific certainty and that medical science is attempting to understand what causes minor side effects such as chills and fever. Dr. Evans also admitted that the “viscosity” theory that she relies on is only one “postulate” or “hypothesis” and that her hypotheses have not yet been either scientifically established or generally accepted by experts in the field. In addition, Dr. Evans acknowledges that because it has not been scientifically established what causes either minor side effects such as chills or major side effects such as stroke, it has not been scientifically established that the cause of minor side effects is also a cause of major side effects. Because the record reflects that appellants’ experts’ opinions are not supported by generally accepted research and objective data, or are generally accepted by the medical or scientific communities, appellants did not meet their burden under the Frye-Mack standard. Therefore, we conclude that the district court did not err in granting respondents’ motion to exclude appellants’ experts’ opinions.

B. Rate-of-Infusion

Appellants also argue that the district court erred in precluding the continued questioning of Dr. Brannagan regarding the assumption that 148.75 ml of IVIg was infused in the last 15 minutes on October 18, which would have established that the rate of infusion was 875 ml/hr at that time-not a generally accepted rate. “[Ejvidentiary rulings, including a decision to exclude expert testimony, lie within the sound discretion of the [district] court.” Benson v. N. Gopher Enters., 455 N.W.2d 444, 445 (Minn.1990). Absent a clear abuse of discretion, this court will not reverse a district court’s evidentiary ruling. Id. at 446.

Even if appellants’ representation of the facts is true, appellants’ evidentiary argument is irrelevant. Assuming that the IVIg was infused at a rate greater than 800 ml/hr, appellants still did not produce any evidence showing that it is generally accepted by the medical or science communities that such a high rate of infusion causes strokes. Instead, Dr. Evans conceded that experts in the field do not generally accept her theory. Therefore, even assuming that the IVIg was infused at an extremely high rate, appellants still did not meet their burden under the Frye-Mack standard, and the district court did *697 not abuse its discretion in making the decision to sustain respondents’ objection to appellants’ line of questioning.

II.

“On an appeal from summary judgment, we ask two questions: (1) whether there are any genuine issues of material fact and (2) whether the [district] court[] erred in [its] application of the law.” State by Cooper v. French, 460 N.W.2d 2, 4 (Minn.1990). A motion for summary judgment shall be granted when the “pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that either party is entitled to a judgment as a matter of law.” Minn. R. Civ. P. 56.03. Whether the district court properly granted summary judgment on the issue of medical causation is determined by a review of the evidence in the light most favorable to the party against whom summary judgment was granted. Goeb, 615 N.W.2d at 816-17.

In actions against health-care providers, a prima facie case of malpractice is established by showing “(1) the standard of care recognized by the medical community as applicable to the particular defendant, (2) that the defendant departed from that standard, and (3) that the defendant’s departure was a direct cause of the plaintiffs injuries.” Fabio v. Bellomo, 504 N.W.2d 758, 762 (Minn.1993). As to the third element, a plaintiff must present competent expert testimony showing that the defendant’s action or inaction was a direct cause of the injury. Teffeteller v. Univ. of Minn., 645 N.W.2d 420, 428 (Minn.2002). A mere possibility of causation is not enough to sustain a plaintiffs burden of proof. Walton v. Jones, 286 N.W.2d 710, 715 (Minn.1979). The failure to provide such admissible expert testimony results in the failure to establish an essential element of that party’s case, and the moving party is entitled to summary judgment as a matter of law. Goeb, 615 N.W.2d at 817.

Here, appellants cannot establish a pri-ma facie case of medical negligence without competent -expert testimony as to the cause of McDonough’s stroke. Because we conclude that -the district court did not err in excluding the opinions of appellants’ experts, summary judgment was also appropriate.

DECISION

The district court did not err or abuse its discretion in excluding appellants’ experts’ opinions and did not err in granting respondents’ motion for summary judgment.

Affirmed.

. A patent foramen ovale is an opening between the upper two chambers of the heart that has failed to close after birth. Stedman’s Medical Dictionary 606 (25th ed.1990).

. A thrombus is ''[a] fibrinous clot formed in a blood vessel or in a chamber of the heart.” The American Heritage Dictionary of the English Language 1802 (4th ed.2000). Embolic means "caused by [a mass, such as an air bubble or a blood clot, that travels through the bloodstream and lodges so as to obstruct a blood vessel].” Id. at 583. Therefore, thrombotic or thromboembolic events are processes that can cause clots that travel and can cause damage to tissues (i.e., strokes or heart attacks).

. A differential diagnosis "eliminates the possibility of competing causes or confounding factors.” Goeb, 615 N.W.2d at 815. "In performing a differential diagnosis, a physician begins by 'ruling in’ all scientifically plausible causes of the [patient’s] injury. The physician then 'rules out’ the least plausible causes of injury until the most likely cause remains.” Glastetter v. Novartis Pharms. Corp., 252 F.3d 986, 989 (8th Cir.2001).

. Because we conclude that the differential diagnosis of Dr. Evans is inadequate to establish causation here, and thus appellant is unable to meet the requirements of the Frye-Mack test as set out by the Minnesota Supreme Court, we do not reach appellants’ argument that a differential diagnosis, standing alone, is sufficient to meet the Frye-Mack standard.