dissenting.
I respectfully dissent.
There is no overwhelming evidence of guilt in this case. The jury never heard the complete medical facts because counsel failed to interview the most important witnesses, Kevin’s treating physicians.
As of the trial, Morales’s counsel knew or should have known the following: (1) Doctors Carlstrom and Moorman treated Kevin on the night he died; (2) Doctors Carlstrom and Moorman concluded, at the time he was admitted to the hospital, that Kevin died as a result of a rebleed of a chronic subdural hematoma; (3) Kevin’s medical records, the CT Scan, and Doctor Carlstrom’s observations of Kevin’s skull during surgery showed that the blood in Kevin’s head displayed characteristics consistent with a rebleed of a chronic subdural hematoma; (4) Doctors Carlstrom and Moorman attended a meeting (along with another one of Kevin’s treating physicians) orchestrated and attended by at least four county prosecutors and medical examiner Doctor Bennett, all of whom maintained that Kevin died from shaken-slammed baby syndrome; (5) one of the county prosecutors who attended the meeting assisted in a child death review team and was romantically involved with medical examiner Doctor Bennett; (6) after the meeting, Doctors Carlstrom and Moorman changed their opinions to be consistent with those of Doctor Bennett and the government that Kevin died from shaken-slammed baby syndrome; and (7) Doctors Carlstrom and Moorman would testify for the government at trial.
Yet, despite this knowledge, the record shows that trial counsel did not personally interview Doctor Carlstrom or Doctor Moorman. It is clearly established that “[defense] counsel has a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary.” Strickland v. Washington, 466 U.S. 668, 691, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). This court has explained that under Strickland “[Reasonable performance of counsel includes an adequate investigation of facts, consideration of viable theories, and development of evidence to support those theories.” Foster v. Lockhart, 9 F.3d 722, 726 (8th Cir.1993). Morales’s counsel’s failure to interview Doctors Carlstrom and Moorman to investigate the circumstances of their changed testimony fell below constitutional standards of competence in light of the doctors’ changed opinions as to the cause of Kevin’s death.
This deficient representation undoubtedly undermines any confidence in the verdict against Morales, see Strickland, 466 U.S. at 694, 104 S.Ct. 2052, and the state court’s determination to the contrary was unreasonable. The record establishes that the important medical evidence and opinions of the attending physicians could significantly impair the government’s theory that Kevin’s death resulted from shaken-slammed baby syndrome and demonstrates that Kevin died from an old, chronic condition rather than a recent injury. Moreover, the record shows that prosecutors played a key role in presenting skewed medical opinion evidence from the physicians who treated Kevin on the day he died by hosting a meeting, not disclosed to or attended by defense counsel, and using the opinion of a possibly biased medical examiner to persuade the treating *557physicians to change their initial opinions. That meeting resulted in Doctor Carlstrom changing his opinion as to the cause of Kevin’s death. All of this would have come to light if defense counsel did what any minimally competent lawyer would do: personally interview the important witnesses.10
Armed with the information counsel would have obtained by interviewing the physicians, Morales’s trial would have been quite different. First, Doctor Carl-strom’s testimony would have been less persuasive, if not entirely different. Judge Sackett of the Iowa Court of Appeals, writing separately in this case on direct appeal, explained that in determining that Kevin suffered from shaken-slammed baby syndrome he “look[ed] particularly to the testimony of Dr. Thomas Carlstrom, the neurosurgeon who operated on the victim[J” If counsel had interviewed Doctor Carlstrom, a jury would have heard his original (and current) opinions as to the cause of Kevin’s death: the blood in Kevin’s skull was liquid, consistent with a rebleed of a chronic hematoma; the CT scan and other treating physicians confirmed this observation and diagnosis; the injury Doctor Carlstrom had observed was the type of injury that very little trauma could cause to rebleed; a fall down a flight of stairs, hitting a concrete wall at the bottom, could cause a linear fracture and, because of the presence of the chronic hematoma, could cause a rebleed and ultimately death.11
By way of example, of the information described above, the jury heard Doctor Carlstrom state, on direct examination, that the blood in Kevin’s skull was liquid and that was “a bit unusual.” On cross examination, Doctor Carlstrom surmised that Kevin’s blood was probably unable to clot. Had counsel interviewed or made a complete investigation of Doctor Carlstrom prior to trial, counsel could have confidently inquired further and a jury would have heard Doctor Carlstrom state, as he did in his post-conviction testimony:
Well, I think that the blood clot itself, when I saw it, what I saw at the time of surgery and on the CT scan, everything that I saw pointed to this blood clot being an old blood clot. I was quite certain it was. I have never seen a brand-new blood clot liquid like this one was. This would be the only case I have ever seen like that, and the only explanation for it could be that undeniable pathological identification would indicate that there was no — that this was not a *558chronic subdural hematoma. I still can’t explain how one can have a blood clot hours old that was all liquid. That’s a very difficult — a very difficult pathological — very difficult to occur.
App. at 660.
In addition, if counsel had investigated the cause of the doctors’ changed opinions, they could have presented that information in court, discrediting not only Doctor Bennett, the state medical examiner,12 but also the prosecution itself in this case. Although counsel attempted to make a professional statement about Doctor Bennett’s romantic relationship with a county prosecutor, they failed to make an adequate offer of proof and failed to link it in a material way to the case. Had counsel interviewed Doctor Carlstrom and investigated the cause of his changed testimony (the improper meeting hosted by Doctor Bennett and the prosecutor’s office), counsel would have been able to connect Doctor Bennett’s alleged bias to the case and significantly impair the credibility of the state’s witnesses in this case. Significantly, then Doctor Carlstrom himself might have begun to question the propriety of the meeting and Doctor Bennett’s conclusions.13
Both the district court and my colleagues on this court have noted that every court that has reviewed this case has been troubled by issues of fairness it presents. Those issues represent substantial flaws in Morales’s conviction. Those flaws should have bothered Morales’s trial counsel enough to prompt them to fully prepare and investigate a case calling for a possible life sentence.
Accordingly, I dissent. Morales is entitled to relief and the writ of habeas corpus should have been ordered by the district court.
. As part of the record before us, I have examined two pages (which was all that was provided to this court) of deposition testimony of Doctor Carlstrom, apparently taken by Morales's former counsel prior to trial and which were available to Morales’s trial counsel. Counsel’s reliance on this deposition alone serves as inadequate investigation of the doctors' change in testimony. The deposition reveals that Doctor Carlstrom changed his original medical opinion, which he had based on his own personal observations, upon inducement by the medical examiner and not based on any objective medical evidence that Doctor Carlstrom had observed. The information gleaned from the deposition, at a minimum, required further investigation by competent counsel in preparation for trial.
Indeed, a number of the other bases on which Morales argues ineffective assistance of counsel, such as failure to investigate and impeach Doctor Bennett and failure to pursue the slides, are derivative of counsel's failure to interview Doctors Carlstrom and Moorman. These two interviews would have resulted in counsel’s appreciation of the significance of the slides and Doctor Bennett's improper influence as avenues for Morales's defense. The slides, unfortunately, are no longer available as they have been destroyed by the State of Iowa.
. See App. at 660-663 (from Doctor Carl-strom's post-conviction testimony).
. Characterization of Doctor Bennett’s testimony by the state courts as "cumulative” belittles the weight a jury would give a state medical examiner's testimony. See Iowa v. Morales, No. 8-074/97-152, slip op. at 6 (April 24, 1998) (enbanc).
. Doctor Carlstrom has stated now that he no longer considers Doctor Bennett trustworthy, explaining: "I think that Dr. Bennett’s testimony in other child abuse cases has come into question because I think he's just a bit overzealous in his opinion giving. I have disagreed with his opinions on a number of occasions.” App. at 661.
If only Morales’s counsel had fully investigated this case, Doctor Carlstrom's skepticism of Doctor Bennett would not have come so late.