In this case we review the correctness of the trial court’s findings, based on conflicting expert testimony, that a swine flu vaccination caused plaintiff’s subsequent disease and complications. We affirm the judgment for the plaintiff.
I. Background
Plaintiff Madeline Hockett received a swine flu vaccination in October 1976. Six months later she was diagnosed as having Guillain-Barre Syndrome (GBS), a debilitating neurological disorder. She sued the government under the Federal Tort Claims Act, 28 U.S.C. § 1346(b) (1976), and the National Swine Flu Immunization Program of 1976, 42 U.S.C. § 247b (1976), alleging that she acquired GBS as a result of the vaccination. The parties stipulated that plaintiff had GBS and that she did not have to prove negligence or any other theory of liability; thus, she was required to prove only that the vaccination caused her GBS. The government contended that a flu-like illness she suffered several days prior to the onset of GBS, not the vaccination, caused the GBS.
Plaintiff’s theory of causation centered on the testimony of Dr. Eylar, a biochemistry professor. Dr. Eylar described his work with laboratory animals and experimental allergic neuritis (EAN), a demyelinating disease generally recognized by the scientific community to be an excellent animal model for human GBS. A demyelinating disease attacks myelin, the component *711of a sheath surrounding the nerve that serves a function similar to that of insulation surrounding an electrical wire. Dr. Eylar found that EAN can be produced in animals by injection of a protein designated as P2, when accompanied by a chemical adjuvant used to heighten the reaction.1 He alsovfound that experimental monkeys could develop what he termed a chronic form of the disease and that some would not have a severe clinical attack until two or three months after the injection. Dr. Eylar determined that the blood sera of all the injected monkeys contained antibodies directed against P2 protein but that none of the control monkeys had any P2 antibodies.
Dr. Eylar tested a sample from the same lot of vaccine given to plaintiff and found that it contained chicken P2 protein. He also tested plaintiffs blood serum and found that it contained antibodies that reacted with chicken P2. Dr. Eylar testified that he found P2 antibodies in the blood sera of approximately 50% of 30 people who had been vaccinated and had contracted GBS and that no P2 antibodies had ever been found in the blood sera of “conventional” GBS patients (those who had not received a swine flue vaccination).
Thus, under plaintiffs theory animal experiments suggest that P2 protein can play a causal role in GBS and that the presence of P2 in the vaccine and P2 antibodies in plaintiffs blood demonstrates that the vaccine did in fact cause plaintiffs GBS. Plaintiff contends that antibodies created by the human body in response to the foreign chicken P2 do not adequately discriminate chicken P2 protein from the human protein in the myelin sheath surrounding the peripheral nerves, and, therefore, the antibodies attack the myelin.
The government’s expert, Dr. Brostoff, testified that he tested plaintiff’s lot of swine flu vaccine and that it did not contain any P2 antibodies. Dr. Eylar disputed the reliability of Dr. Brostoff’s vaccine test, and the district court, based on the testimony of both, concluded that Eylar’s test, not Brostoff’s, was credible. The government never tested plaintiff’s blood for P2 antibodies; Eylar’s testimony that plaintiff’s blood contained these antibodies therefore remained unrefuted. Even assuming the P2 antibodies were present in plaintiff’s blood, Dr. Brostoff offered an alternative explanation for their presence. Hockett was “primed” with swine flu vaccine (i.e., chicken P2) and then suffered a viral infection that caused GBS. Her peripheral nervous system was damaged as a result of the disease process, and P2 protein (a component of the peripheral nerve) leaked into the blood stream, causing antibodies to form in response to the presence of P2. Because Hockett had chicken P2 in her body (as a result of the swine flu vaccination but having nothing to do with the GBS), antibodies also formed in response to chicken P2. Thus, when examined plaintiff’s blood contained antibodies to chicken P2. The antibodies to chicken P2, according to Dr. Brostoff, were a result of the disease process, not the cause. The district court rejected this rationalization in favor of Dr. Eylar’s testimony that his theory explained the only cause for the presence of chicken P2 antibodies in plaintiff’s blood.
The government further attempted to rebut Dr. Eylar’s evidence of causation through a statistical study tending to show that vaccination-caused GBS occurred within four to five weeks after the vaccination and that the incidence of GBS in vaccinated patients after ten weeks equaled that in the unvaccinated population. GBS occurring more than ten weeks after vaccination, the government’s experts testified, is more likely to be caused by a flu-like illness than by the vaccination.
Following a bench trial the court entered judgment for plaintiff for $324,275.61. The government asserts three errors on appeal: (1) the court erroneously shifted the burden to the government to disprove *712causation; (2) the court’s finding that the vaccination caused plaintiff’s GBS is clearly erroneous; and (3) the court’s finding that the GBS caused plaintiff’s subsequent pulmonary disease is clearly erroneous.
II. The burden of proof in the district court
The district court correctly stated that the plaintiff had the burden of proving causation by a preponderance of the evidence, i.e., that the vaccination was a substantial factor in the onset of GBS. See, e.g., Cassel v. Price, 396 So.2d 258, 266 (Fla.Dist.Ct.App.1981) (to prove causation plaintiff must show more likely than not that defendant’s conduct was a substantial factor in bringing about the result). Thus the court did not as a preliminary matter choose the wrong burden of proof.
The government contends that a portion of the district court opinion discussing the weight to be given to the government’s statistical evidence shows that the district court erroneously shifted the burden to the government to disprove that the vaccination caused plaintiff’s GBS:
It appears from the evidence that the state of the scientific and medical knowledge about GBS is that some of those people who received a swine flue vaccination later developed GBS, but most did not. Of those who did, it usually took two to three weeks for the symptoms to appear. However, the mere fact that the time interval in plaintiff’s case was significantly longer is not dispositive. The statistics reveal a skewed bell curve relationship between the vaccinations and the onset of GBS; that does not rule out the possibility of a causal relationship between the two. Dr. Daniel Bader pointed out in his deposition that the existence of a bell curve relationship doesn’t mean that things that fall outside the middle part of the curve, on either side, are not associated. After ten weeks, the incidence of GBS in the vaccinated population does not fall to zero — it drops to the background incidence rate. That background rate reflects unknown causal factors. As Dr. Barry Arnason stated in his deposition, some 60% of the cases of GBS can be associated with a history of infections [sic] illness in the prior four weeks and another 5-10% with a surgical operation in the prior month or so. Even if those were known to be causal relationships rather than mere associations, that would still leave the cause of 30-35% of the cases unexplained.
Dr. Bader’s observation that one shouldn’t throw away the outside margins of the bell curve just because the majority of things occur in the middle of the bell curve is well taken. We know that the overall risk of GBS in the vaccinated population is approximately four times greater than normal, and that some of those in the vaccinated population do in fact develop GBS more than nine or ten weeks later; we also know that not all of the cases of GBS can even be explained with associations, much less proven causes. If cases such as this were to be decided on the bell curve statistics alone, no plaintiff who contracted GBS more than nine weeks after his vaccination would ever be able to recover. While such a result might be justified in a case where there was no other evidence, in this case there is additional evidence which this Court must consider.
Manuscript op. at 6-7 (emphasis added) (footnotes omitted).
The government contends that another passage further confirms that the district court improperly placed the burden on the government to disprove that the vaccination caused the GBS:
A crucial item of additional evidence in this case is Dr. Eylar’s finding that plaintiff’s blood serum contained antibody to chicken P2 protein. There is no contrary evidence; defendant’s experts did not test plaintiffs blood. The experts agreed that if in fact plaintiff’s blood contained such antibodies, she must have somehow been exposed to chicken P2 protein, because those specific antibodies are not naturally found in the human body. Defendant has no explanation *713for such exposure, but plaintiff claims it was due to the vaccine.
Id. at 7 (emphasis added).
In arguing that the district court improperly shifted the burden of proof, the government takes portions of the district court opinion out of context and fails to analyze the opinion as a whole. First, as noted above, the district court correctly stated the burden of proof and required plaintiff to prove by a preponderance of the evidence that the vaccination caused her GBS. Manuscript op. at 6. The plaintiff proved that her GBS was caused by the vaccine and not by other causes through the testimony of Dr. Eylar, which showed that P2 antibodies were present in plaintiffs blood, P2 protein was in plaintiffs vaccine, P2 antibodies occur in animals that have been injected with P2 protein and that contract a disease similar to GBS as a result, and P2 antibodies do not appear in control animals or people who have contracted GBS without ever receiving a vaccination. The district court resolved the credibility choice between Dr. Eylar and Dr. Brostoff in Dr. Eylar’s favor.
By Dr. Eylar’s testimony the plaintiff established a prima facie case and proved causation by a preponderance of the evidence. The government then had to rebut the plaintiff’s case. It did so by relying on Dr. Brostoff’s testimony that he found no P2 protein in plaintiff’s vaccine, on his alternative explanation for the presence of P2 protein if any was in fact in plaintiff’s blood, and on the statistical study and other expert testimony to show that the vaccination-caused GBS is unlikely, though still possible, when GBS occurs more than five to ten weeks after the vaccination. Although it tested the vaccine, the government did not test plaintiff’s blood to rebut Eylar’s testimony on presence of P2 antibodies. The district court credited Dr. Eylar’s test for the presence of the P2 protein in the vaccine and his explanation for the presence of the chicken P2 antibodies in plaintiff’s blood despite the government’s theory to the contrary.
Thus the district court did not hold, as the government intimates, that proof of a mere possibility that plaintiff got GBS from the vaccine is sufficient and that the government had the burden of disproving plaintiff’s claim. The court merely looked at Dr. Eylar’s P2 testimony establishing a causal link and then at the government’s rebuttal case. The government offered statistics and testimony to show that plaintiff’s GBS was unlikely, though possibly, caused by the vaccination. The court determined that this evidence did not sufficiently rebut plaintiff’s proof of causation under Dr. Eylar’s testimony because that testimony showed (conclusively if credited) that plaintiff got GBS from the vaccination, and the government statistics, even if believed, did not exclude that theory. Had the P2 testimony not been in the case the plaintiff would have proven only a possibility. The P2 testimony carried plaintiff’s burden, and the government’s statistics and discredited attack on Dr. Eylar’s P2 testimony did not eliminate that preponderance of the evidence. The district court did not impermissibly shift the burden of proof to the government.
III. The lower court’s finding that the swine flu vaccination caused plaintiff’s GBS
The government contends that the trial court clearly erred in finding2 that the swine flu vaccination caused plaintiff’s GBS. Under the clearly erroneous standard fact findings at the trial court must stand unless the reviewing court is left with the definite and firm impression that a mistake has been made. See American National Bank v. FDIC, 710 F.2d 1528, 1534 (11th Cir.1983).
In essence the government contends that the trial court erred in crediting Dr. Eylar’s *714testimony instead of Dr. Brostoff’s, that Dr. Eylar’s testimony is speculative and conjectural, and that findings based on that testimony cannot stand. The government never objected to Dr. Eylar’s qualifications or to the admission of his testimony as incompetent or speculative. In addition, on cross-examination the government affirmatively elicited Eylar’s theory of causation.3 The government never made the district court aware by objection, motion to strike, or otherwise that the government contended that the court was precluded as a matter of law from considering Dr. Eylar’s testimony.4 At most, the court was aware of the government’s position that Dr. Eylar’s testimony, while evidently admissible, should not be believed or credited. The district court carefully considered the explanations posited by both sides, weighed the credibility and demeanor of the witnesses, and resolved those questions in favor of the plaintiff.
On appeal the government refers us to other district court cases that have resolved these credibility and factual issues differently. Our task is to review the findings in the case before us, and we must affirm them unless we are left with the definite and firm impression that they are wrong. We do not reweigh the evidence and second guess the trial court’s credibility choices. The government did not object to Dr. Eylar's qualifications as an expert. To the trial court his theory was rational and believable. His findings may be novel, but if the government contended his testimony to be so speculative, conjectural, and without foundation and general acceptance as to render it incompetent, the government was required to object to its admission or move to strike it. Absent objection the government’s concerns went to the weight of the evidence, and upon weighing the evidence the district court found for the plaintiff.
The district court acknowledged the difficulty in resolving the evidentiary conflict. After reviewing the evidence we are not left with the definite and firm conviction that a mistake has been made.
IV. The lower court’s finding that the vaccination-induced GBS caused plaintiff’s later pulmonary problems
The government also contends that the trial court clearly erred in allowing recovery for medical expenses and other damages for plaintiff’s hospitalization for chronic obstructive pulmonary disease following the GBS. The court stated that its “review of the medical records persuade[d] it that the subsequent hospitalizations were a natural and direct consequence of plaintiff’s bout with GBS.” According to the government, the medical records attribute plaintiff’s pulmonary problems to smoking *715and to “residual muscle weakness secondary to GBS.”
The government maintains that the trial court made a medical determination in the absence of expert testimony. The trial court merely adopted a medical conclusion contained in a medical record. The court did not arrive at such a conclusion on its own. Moreover, the plaintiff testified without contradiction that before the GBS she had no respiratory problems whatsoever. Her son, a licensed registered nurse (though not qualified as an expert), testified without objection that being on a respirator while paralyzed by GBS caused her respiratory problems.
The government presented contrary expert testimony that years of smoking caused plaintiffs pulmonary problems. Although plaintiff was a smoker, she and her son testified without objection that she had experienced no pulmonary problems before the GBS. The district court could reasonably have concluded that, because of plaintiffs prior freedom from pulmonary problems and occurrence of such problems only after the GBS, the GBS, not smoking, caused her pulmonary problems, especially in light of the medical records supporting that determination. In view of the causal conclusions in plaintiffs records, when coupled with her and her son’s testimony, we cannot say that the district court erred in finding that GBS caused plaintiff’s pulmonary problems.
AFFIRMED.
. Although no adjuvant was present in the swine flu vaccine, Dr. Fischer, one of the government's experts, testified that a flu-like cold or infection, such as Hockett had prior to the GBS, could act as an adjuvant to heighten the reaction to P2. Thus the lack of an adjuvant in the swine flu vaccine did not render EAN an inappropriate model for human GBS.
. The trial court stated: "[T]his Court finds that plaintiff has shown that her swine flu vaccination was a substantial factor in bringing about her Guillain-Barre Syndrome.” Manuscript op. at 9; see Cassell (to prove causation under Florida law one must show that the defendant’s action was a substantial factor in bringing about the result).
. On cross-examination of Dr. Eylar, the government elicited the following:
Q Doctor, if I understand you correctly on direct examination, it is your theory that there are three possibilities of causation in this case. Would that be fair to say?
The first one is that she responded immunologically somehow to P-2 in the vaccine. Is that correct?
A No, I wouldn’t say there are three separate theories. I think that, as I stated, or at least meant to state, that it all gels together as a very logical scenario that we might expect from many of Guillain-Barre syndrome.
Q Is it possible that she responded immunologically to P-2 in the vaccine?
A I say, we have the direct proof of that, because we found the antibody to be chicken P-2 in her serum. So there’s no question whatsoever that she responded to the P-2 protein.
And I’m saying that that offers extremely strong evidence that if she responded to the P-2 protein, the chicken P-2, and developed an antibody, she also developed immune response which led to her disease.
. The government relies upon Martinez v. Food City, Inc., 658 F.2d 369, 372 n. 1 (5th Cir.1981) (Unit A), for the proposition that a formal objection was not required. Although a formal objection may not be necessary when "a party's position has been made clear to the court, along with the action the party wishes the court to take,” the government never made the trial court aware, either in the opening statement or otherwise, that the government contended that Dr. Eylar’s testimony was incompetent as a matter of law and that the court could not as a matter of law rely upon that testimony.