JOHNSON, Circuit Judge:
Alvin Scott Loyd was convicted of first degree murder and sentenced to death by a jury for the murder of three year old Tina Giovanetti. At the state habeas level, Loyd raised a claim of ineffective assistance of counsel.
After conducting an evidentiary hearing, the state court concluded that the performance of counsel at the sentencing phase was deficient; however, because the state court concluded that such deficiency did not prejudice Loyd, the state court denied habeas relief. After exhausting all state court avenues of relief, Loyd filed a Petition for Postconviction Relief, Writ of Habeas Corpus, Request for Evidentiary Hearing and Stay of Execution with the federal district court for the Eastern District of Louisiana. The district court granted a stay of execution but denied Loyd’s requested relief without holding an eviden-tiary hearing. In regard to the ineffective assistance claim, the district court concluded that, contrary to the finding by the state court, counsel did not perform in a deficient manner.
The district court, ordering that the stay remain in effect pending appeal, issued a Certificate of Probable Cause to appeal to this Court. We conclude that the district court erred by failing to give proper deference to the factual findings of the state court pursuant to section 2254(d). Consequently, we vacate the finding of the district court on the issue of counsel’s deficient performance and remand for further proceedings consistent with this opinion.
I. BACKGROUND
A.
Chronology of the Proceedings
A grand jury in St. John the Baptist Parish returned an indictment charging Loyd with first degree murder.
After a bifurcated trial, the jury unanimously found Loyd guilty and recommended the death penalty.
The Louisiana Supreme Court affirmed Loyd’s conviction, but vacated the death sentence and remanded for a new sentencing trial because of error in the trial judge’s instructions to the jury. At the second sentencing trial, the jury again imposed the death penalty. The sentence was affirmed on appeal. State
v. Loyd,
489 So.2d 898 (La.1986),
cert. denied,
481 U.S. 1042, 107 S.Ct. 1984, 95 L.Ed.2d 823 (1987),
reh’g denied of cert. denial,
483 U.S. 1011, 107 S.Ct. 3244, 97 L.Ed.2d 749 (1987).
Approximately one week prior to the October 16, 1987, execution date, Loyd filed a Petition for Postconviction Relief in the St. John the Baptist Parish state trial court. On the same date, October 9, the state trial court denied the petition. On October 12, the Supreme Court of Louisiana granted Loyd a stay of execution and remanded the case to the state trial court for an eviden-tiary hearing on four issues including the ineffective assistance of counsel claim.
The state court conducted the evidentiary hearing and concluded that Loyd’s counsel at the sentencing trial was deficient because of counsel’s failure to pursue mitigating evidence of mental disease or defect and to retain a psychiatrist to assist in developing this clearly relevant line of evidence.
However, the state court concluded that this deficient performance was not prejudicial, and denied Loyd relief. The Louisiana Supreme Court thereafter denied Loyd’s subsequent writ of eertiorari without written decision.
State ex rel. Loyd v. Butler,
514 So.2d 446 (La.1987). On March 7, 1989, Loyd filed a Second Petition for Postconviction Relief in the state court. Loyd was denied relief, and the Louisiana Supreme Court denied certiorari.
State ex
rel. Loyd v. Butler,
539 So.2d 639 (La. 1989).
On March 29, Loyd filed a Petition for Postconviction Relief, Writ of Habeas Corpus, Evidentiary Hearing and a Stay of Execution in the United States District Court for the Eastern District of Louisiana. This petition asserted twenty-nine grounds for relief.
The district court granted Loyd a stay of execution and directed the parties to submit a brief on four of the issues. On July 27, 1989, the district court, in an order addressing all of Loyd’s claims, denied the Petition. Notably, the district court determined, in a finding contrary to the conclusion of the state court, and without holding an evidentiary hearing, that Loyd’s counsel was
not
deficient at the second sentencing trial.
The district court issued a Certificate of Probable Cause to appeal to this Court, and ordered that the stay of execution remain in effect pending appeal. Consequently, this case comes to this Court in a somewhat unusual posture; specifically, the state court, after conducting an evidentiary hearing, concluded that counsel at the sentencing phase was deficient; the district court, without conducting an evidentiary hearing, reached the opposite conclusion.
B.
State Court Proceedings Relevant to the Ineffective Assistance Claim
1. Pre-Trial
At approximately 2:30 a.m. on Monday, April 27, 1981, police officers apprehended Loyd at his home and took him to the St. John the Baptist Parish police station. Because Loyd was shaking and appeared nervous and upset, the Sheriff summoned Dr. S.J. St. Martin, the coroner for St. John the Baptist Parish and a family practitioner who had treated the Loyd family for several years. Dr. St. Martin administered a drug containing phenothiazine, an anti-psychotic medication.
On April 28, 1981, the state court appointed Donald Cicet to represent Loyd. On May 1, 1981, Hackman and Lewis were appointed to succeed Cicet as counsel. On May 13, the Louisiana court appointed Dr. St. Martin, who at this time had been treating Loyd for approximately two and one-half weeks, and Dr. Kenneth Ritter, a Board Certified Psychiatrist and a specialist in Forensic Psychiatry, to a Sanity Commission. The purpose of the Commission was to determine whether Loyd was sane or insane at the time of the offense and whether Loyd was competent to stand trial.
The Commission submitted a written report and the doctors testified at a sanity hearing as to the Commission’s conclusions that, although Loyd was sane at the time the crime was committed, he was not competent to stand trial. On the question of sanity, Louisiana courts apply the
McNaughten
Rule, codified at LSA-R.S. 14:14, which states:
If the circumstances indicate that because of a mental disease or mental defect the offender was incapable of distinguishing between right and wrong with reference to the conduct in question, the offender shall be exempt from criminal responsibility.
As to the second prong of the Commission’s task, determining competency to stand trial, the doctors found that Loyd’s depression rendered him unable to understand the charges and proceedings against him and that he could not effectively participate or assist in his defense. Based on these findings, the court committed Loyd to the Feliciana Forensic Facility (hereinafter Feliciana) for treatment of his depression. Dr. Aris Cox, a Board Certified Forensic Psychiatrist, treated Loyd while he was institutionalized at Feliciana.
At Feliciana, Loyd was also evaluated by Dr. Thomas Fain, a Board Certified Psychologist, and R. Roy Allen, M.A., a Psychological Assistant III. These men performed several diagnostic and psychological assessment tests on Loyd. The Felicia-
na report of Loyd’s psychological evaluation indicates that Loyd was abused as a child; that during Loyd’s tenure in the Navy he was treated for mental problems, used alcohol, marijuana and PCP; that Loyd reported excessive beer drinking and suspected drug ingestion on the day of the crime, that Loyd suffered a loss of memory following the crime, and that Loyd experienced headaches and difficulty sleeping while at Feliciana.
On November 19, 1981, five months after Loyd was committed to Feliciana, the Sanity Commission issued another report indicating that Loyd was competent to stand trial. Prior to the 1983 trial, Loyd’s mother gave Hackman $2000 to retain a psychiatrist to assist in the defense. Hackman used $750 of the fund to retain Dr. C.B. Scrignar to evaluate Loyd.
2. The Guilt/Innocence Phase
At trial, Loyd’s attorneys devised a defense strategy centering around Loyd’s asserted insanity at the time of the offense. Hackman called Dr. Cox, Loyd’s treating physician at Feliciana, to testify on the issue of Loyd’s sanity.
Dr. Cox indicated that it was not possible for him to state positively whether Loyd was sane or insane.
Cox did indicate that a finding of sanity did not foreclose the existence of other mental defects. “[P]eople can have significant psychiatric problems but still meet, in my opinion, the legal tests for sanity.” Dr. Cox noted that Loyd had a difficult childhood, had suffered abuse at the hand of his father, and that, on at least one occasion, his father had shot him.
The Government called Dr. Ritter to testify as to the Sanity Commission’s conclusion that Loyd was sane at the time the offense was committed. The case was submitted to the jury, which returned with a verdict of guilty.
3. The Sentencing Phase
At the first sentencing trial, Loyd’s counsel presented no evidence in rebuttal or in mitigation. The Louisiana Code of Criminal Procedure indicates that evidence in mitigation includes evidence indicating that “the offense was committed while the offender was under the influence of extreme mental or emotional disturbance,” and evidence that “at the time of the offense the capacity of the offender to appreciate the criminality of his conduct or to conform his conduct to the requirements of law was impaired as a result of mental disease or defect or intoxication.” La.Code Crim.Pro. Art. 905.5(b) and (e).
The jury returned a sentence of death. This sentence was overturned and the case remanded for a new sentencing trial due to the state court’s erroneous instruction as to unanimity of the jury.
Three weeks prior to the second sentencing trial, Hackman and Lewis sought to withdraw as counsel and to allow William Allison to take over. Because the court would only allow one of the men to withdraw, both remained on the defense team, which was headed by Allison. On the same day that Hackman and Lewis moved to withdraw, Allison moved for a continuance of trial and sought funds to hire a private investigator and a psychiatrist to assist in presenting mitigating evidence. The state court denied the continuance, denied funds for a private investigator, and granted only $600 of the $1000 requested for a psychiatrist. Hackman did not inform Allison of
the $1250 remaining from the money advanced prior to the guilt/innocence phase of the proceedings.
Allison contacted a psychiatrist who requested $2500 up front. Believing he had only $600, Allison did not engage that psychiatrist’s services. Allison testified that he was aware that there were gaps in the record that were not explained and that he was in need of psychiatric testimony to fill the gaps. Allison testified that he attempted to meet with Dr. Cox twice in order to explore these gaps, but the two men never managed to coordinate their schedules. Allison also indicated that he continued his search for an independent psychiatrist to fill the gaps in the record, but given the lack of funds and the impending trial date, he was unable to do so. The state court, following the evidentiary hearing, concluded that these attempts by Allison to consult with independent psychiatrists were halfhearted and eventually abandoned.
Without having either retained independent psychiatric counsel or explored the perceived gaps in the medical record, Allison called Dr. Cox (the treating psychiatrist at Feliciana), Dr. St. Martin (Sanity Commission member), and Dr. Ritter (Sanity Commission member) to testify at the sentencing hearing. All three doctors testified that Loyd was sane, but that his behavior was nonetheless abnormal. Dr. Ritter testified that Loyd was legally sane, but was initially incompetent to stand trial because of severe depression. Dr. Ritter acknowledged that he examined Loyd solely for the purpose of determining competency to stand trial and sanity at the time of the offense. He did not make a detailed study of personality because he was only trying to determine if Loyd was insane. Nor did Dr. Cox conduct a thorough inquiry into Loyd’s personality. Dr. Cox noted that Loyd was a child abuse victim with a history of substance abuse. The doctor also testified, however, that Loyd was atypical of the pattern violent crime offender.
[Loyd] did not have a long Rap sheet, history of previous arrests. He did not have a history of abuse of his own children, or getting in fights, or attacking people, or violent behavior. And so the fact that this occurred, this act occurred at age twenty-five, this suddenly and precipitously, to me was unusual and atypical and
I cannot explain it.
...
There had to be something going on that was unusual, what it is, I don’t know.
Testimony of Cox, Second Sentencing Hearing, Transcript Volume 5 at 215 (emphasis added). The jury heard no testimony from any source attempting to explain what was “going on.”
The Government called no psychiatric witnesses at the second sentencing hearing. The jury returned with a sentence of death.
4. The Evidentiary Hearing
After the state court denied Loyd’s First Petition for Postconviction Relief, the Louisiana Supreme Court granted Loyd a stay of execution and remanded the case to the state trial court for an evidentiary hearing on,
inter alia,
Loyd’s claim of ineffective assistance of counsel at the second sentencing hearing.
At the evidentiary hearing, Loyd’s current counsel presented the testimony of three doctors to illustrate the type of mitigating evidence which previous counsel should have explored and presented. The three doctors included Dr. Steven Honor, a clinical and neuro-psychologist from New York, Dr. Barry Scanlan, a Clinical Psychiatrist from Atlanta, Georgia, and Dr. Andrew J. Sanchez, Jr., an expert in psychiatry. Dr. Sanchez was originally retained
by the State to review the reports of the Feliciana staff as well as the reports of Drs. Honor and Scanlan. All three doctors reached the conclusion that Loyd was operating in a state of diminished capacity at the time the crime was committed. Drs. Honor and Scanlan specifically indicated that Loyd suffered from an organic brain dysfunction located in the frontal lobe.
Dr. Honor testified that he was particularly interested in exploring the possibility of an “organic or brain dysfunction[ ] since that had been raised in the original [Felicia-na] report. And there had not been any attempt at a full neuro-psychological evaluation.” State Court Evidentiary Hearing, January 22, 1988, p. 374-75. In order to isolate the possibility of a frontal lobe dysfunction, Dr. Honor performed three tests which were not performed by the Feliciana staff. Dr. Honor indicated that isolating the dysfunction was particularly relevant because “frontal lobes are primarily involved in what we would call executive functioning.[sic] Which is the ability of an individual to function rationally, to think logically, and coherently.”
Id.
Dr. Hon- or testified that the results of the test were consistent with the diagnosis of frontal lobe dysfunction.
Dr. Honor also repeated some of the tests administered at Feliciana and obtained results consistent with the Feliciana results. Dr. Honor testified that the results of these tests indicate that Loyd was prone toward alcohol and/or drug addiction and that Loyd suffered from a borderline psychotic nature with a very clear ideology and schizophrenic personality which could degenerate into paranoid schizophrenia or a temporary psychotic episode. Finally, Dr. Honor testified to his conclusion that Loyd had a diminished mental capacity at the time of the crime which rendered Loyd unable to appreciate the consequences and the significance of his actions and the requirement for him to conform to society standards.
Dr. Scanlan’s testimony at the state evi-dentiary hearing indicates that he reached conclusions similar to those reached by Dr. Honor. Specifically, Dr. Scanlan testified that Loyd’s state of mind at the time of the offense was “one of a delirium or what is non specifically referred to as an organic brain syndrome.” State Evidentiary Hearing, January 28, 1988, at 564. In light of Loyd’s “delusional ‘psychotic’ or delirious ... state,” Dr. Scanlan concluded that it was unlikely that Loyd “could have understood the moral significance of his acts.”
Id.
at 578.
In addition, Dr. Scanlan voiced his belief that members of the Feliciana staff perceived Loyd to be in a psychotic state. Specifically, Dr. Scanlan indicated that because Loyd was given phenothiazine upon his admittance to Feliciana, “there is the implication ... that there was something else going on that somebody had perceived in Loyd that they felt to move to prescribe anti-psychotic medications for him."
Id.
at 650-51.
Dr. Sanchez was the last witness called by Loyd to testify as to Loyd’s mental condition. Dr. Sanchez agreed with Dr. Scanlan’s finding that Loyd was suffering from a seriously disturbed state of mind when the crime was committed, and stated his belief that Loyd was operating under a diminished capacity to know right from wrong.
The State called Drs. Ritter, Cox, and St. Martin who had examined and/or treated
Loyd shortly after the crime. Each adhered to the Sanity Commission’s original conclusion that Loyd was sane at the time of the offense. On cross-examination, however, Dr. Ritter stated that he spent, at most, two hours with Loyd and that he did not perform a detailed personality inventory of Loyd. Dr. Cox also testified that his evaluation of Loyd was limited because Loyd was uncooperative.
Finally, a Feli-ciana report indicates that no extensive testing for organic brain damage was performed.
In summary, the evidence adduced at the state evidentiary hearing indicates that, although the indicators of organic brain dysfunction, a mental defect admissible as a mitigating circumstance under the Louisiana statutory scheme, were present, the original team of psychiatrists did not perform tests to determine the extent of the defect, and limited their analysis to the issue of whether Loyd was sane or not. Loyd’s counsel did not pursue this avenue of inquiry and did not seek, either by requesting that the Feliciana team perform additional tests or by retaining independent psychiatric counsel, to develop this avenue of mitigating evidence.
As the state court emphatically stated: [T]he actions of petitioner’s counsel fell below acceptable professional standards at the second sentencing hearing.
... [Petitioner's counsel had sufficient funds on hand to have employed a neutral psychiatrist.... [Petitioner's sanity was without a doubt a critical issue in the case. Equally evident is that counsel was aware of this fact, and indeed Mr. Allison made several half-heart-ed attempts to procure such services, but eventually abandoned those efforts. Mr. Hackman testified that he also considered the need for psychiatric testimony but did not pursue it. He stated that he did not do so because the psychiatric information he had from the Sanity Commission and the East Feliciana Hospital led him to believe that the insanity defense would not prevail. He further admitted, however, that his consideration of this issue was based only on the M’Naghten Rule, i.e., whether at the moment of the crime petitioner could distinguish right from wrong. He admitted that he did not realize that the mitigating factors of ‘mental or emotional disturbance,’ or ‘mental disease or defect’ provided in La.Code Crim.Pro. Art. 905.5(b) and (e), would be admissible at the sentencing phase, even if this evidence did not meet the M’Naghten test for insanity-
State Court Order at 4-5. After concluding that counsel at the second sentencing phase was in fact deficient, the state court turned to the issue of prejudice. Concluding that Loyd had not demonstrated prejudice, the state court denied habeas relief.
C.
In the Federal District Court
Loyd filed a habeas petition in federal district court. The district court, without holding an evidentiary hearing, denied Loyd relief. In ruling on the claim of ineffective assistance of counsel at the sentencing phase, the district court, contrary to the finding of the state court, concluded that counsel was not deficient. In reaching this conclusion, the district court wholly failed to address the factual findings on which the state court relied to reach its conclusion that counsel was deficient. The district court justified this complete absence of deference by noting that
“[i]n a federal habeas challenge to a state criminal judgment, a State Court conclusion regarding effective assistance of counsel is not a finding of fact binding on the Federal Court to the extent stated by 28 U.S.C. § 2254(d). Ineffectiveness is a mixed question of law and fact. Although the State Court findings of fact made in the course of deciding an inef
fectiveness claim are subject to the deference requirement of § 2254(d), both the performance and prejudice components of the effectiveness inquiry are mixed questions of law and fact.”
District Court Order at 40-41.
Despite the district court’s acknowledgement that “State Court findings of fact made in the course of deciding an ineffectiveness claim are subject to the deference requirement of § 2254(d),” there is no indication from the district court’s Opinion and Order that any such deference was applied.
As to the state court’s factual findings that funds were available for the procurement of independent psychiatric services, but that Allison wholly abandoned his attempt to retain such services, the district court concluded that “counsel’s failure to call psychiatric experts, such as Drs. Honor from New York, Dr. Scanlan from Georgia or Dr. Sanchez, was not deficient because ... counsel had no affirmative duty to shop all over the country for a psychiatrist who would say exactly what Loyd wished.” By so concluding, the district court has effectively ignored the state court’s factual findings that
1. Loyd’s mental condition was a critical issue in the case;
2. that Loyd’s counsel was aware of this fact;
3. that Hackman did not pursue the need for independent psychiatric services; and
4. that this failure on the part of Hackman was due to his ignorance as to the difference between legal insanity and mitigating circumstances relevant to sentencing.
After concluding that counsel’s performance at the second sentencing trial was not deficient, the district court noted it was “not required to determine if counsel’s failure to retain a psychiatrist ... ‘prejudiced’ Loyd’s sentencing.” Nevertheless, the district court addressed the issue and concluded that Loyd failed to “affirmatively prove” that he was prejudiced. Consequently, the district court denied relief.
II. DISCUSSION
A claim of ineffective assistance of counsel is reviewed under the now familiar test of
Strickland v. Washington,
466 U.S. 668, 687, 104 S.Ct. 2052, 2064, 80 L.Ed.2d 674 (1984):
First, the defendant must show that counsel’s performance was deficient. This requires showing that counsel made errors so serious that counsel was not functioning as the “counsel” guaranteed the defendant by the Sixth Amendment. Second, the defendant must show that the deficient performance prejudiced the defense. This requires showing that counsel’s errors were so serious as to deprive the defendant of a fair trial, a trial whose result is reliable. Unless a defendant makes both showings, it cannot be said that the conviction or death sentence resulted from a breakdown in the adversary process that renders the result unreliable.
The test involves two prongs: counsel’s deficient performance and prejudice to the petitioner. In reviewing Loyd’s claim of ineffective assistance of counsel at the sentencing phase, the state court, which held an evidentiary hearing, and the district court, which did not, reached opposite conclusions as to the first prong. We address each prong of the
Strickland
test in turn.
A.
Deficient Performance
Strickland
warns reviewing courts that “it is all too easy for a court, examining counsel’s defense after it has proved unsuccessful, to conclude that a particular act or omission of counsel was unreasonable.”
Id.
104 S.Ct. at 2065. Consequently, the Supreme Court warns that review of counsel’s performance must be highly deferential and that the defendant must overcome a presumption that counsel’s actions were the result of “sound trial strategy."
Id.
Specifically, in the context of a claim of failure to investigate,
Strickland
instructs that “counsel has a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary.”
Id.
at 2066.
The state court, applying the
Strickland
test in the wake of the evidentiary hearing
made in that court, made numerous explicit as well as implicit factual findings to support its ultimate conclusion that counsel was in fact deficient by not exploring the statutory mitigating circumstances, which were obviously implicated in the case, and by not making a concerted attempt to engage the services of independent psychiatrists. The district court, however, mostly ignored these factual findings, and, even to the extent the findings were addressed, the district court failed to properly defer to the state court’s findings.
Federal courts in habeas proceedings are required to grant a presumption of correctness to a state court’s explicit and implicit findings of fact if supported by the record. 28 U.S.C. § 2254(d).
See also Marshall v. Lonberger,
459 U.S. 422, 103 S.Ct. 843, 74 L.Ed.2d 646 (1983);
McCoy v. Lynaugh,
874 F.2d 954 (5th Cir.1989). Although the ultimate question of whether or not counsel’s performance was deficient is a mixed question of law and fact, state court findings of fact made in the course of deciding an ineffectiveness claim are subject to the deference requirement of section 2254(d).
Strickland,
104 S.Ct. at 2070;
Smith v. Estelle,
711 F.2d 677 (5th Cir. 1983),
cert. denied,
466 U.S. 906, 104 S.Ct. 1685, 80 L.Ed.2d 159 (1984).
The Supreme Court, in
Sumner v. Mata,
449 U.S. 539, 101 S.Ct. 764, 66 L.Ed.2d 722 (1981), has indicated that when a federal court on habeas review disregards a state court’s factual findings, the federal court must indicate which statutory circumstance enumerated in section 2254(d) is present which allows the court to abandon the presumption of correctness. The high degree of deference to state court findings required by
Sumner
mandates that a federal court must “more than simply disagree with the state court before rejecting its factual determinations. Instead, it must conclude that the state court’s findings lacked even ‘fair support’ in the record.”
Marshall,
103 S.Ct. at 850. This high standard stems from the considerations of comity that militate against federal courts reviewing fact findings with a less deferential standard.
In the instant case, the state court found that Allison made only a half-hearted attempt to engage the services of independent psychiatric expert witnesses, and eventually abandoned this attempt. The district court does not address this factual finding, which is supported by the record, but rather skirts the issue by concluding that the decision to call Drs. Cox, St. Martin, and Ritter was reasonable under the circumstances. The district court reaches this conclusion by noting that two of these men had previously treated Loyd and that Allison was under no obligation to search for experts who would testify in accordance to Loyd’s wishes. This response by the district court not only ignores the state court’s factual findings in several respects, but misconstrues the crux of Loyd’s argument.
Loyd argues that psychiatric experts should have been retained who could have expanded on the mental disease or defect evidence which was indicated in the Felicia-na reports.
Allison’s attempts to obtain funds from the state court demonstrates that he was aware that he was in need of psychiatric testimony, and the state court likewise found that Allison was aware of an unfilled gap in the medical record. Also, Allison attempted to contact Dr. Cox, but abandoned that attempt and, consequently, only spoke to this important psychiatric witness during the lunch hour on
the day of Dr. Cox’s testimony. At this testimony, Dr. Cox testified that something [was] going on,”
i.e.,
a mental disease or defect which would be admissible as a statutory mitigating circumstance, but he did not know what. In light of Dr. Cox’s ignorance as to what was “going on,” Dr. Cox cannot be considered an adequate witness to testify as to the statutory mitigating circumstances; the district court’s conclusion to the contrary is misplaced. The district court should have deferred to the state court’s findings.
The district court wholly failed to address the state court’s implicit finding that the failure to properly investigate stemmed, at least in part, from Hackman’s ignorance as to the admissibility of statutory mitigating evidence of mental disease or defect which did not reach the level of insanity under the
McNaughten
rule. Had Hackman understood the relevance of mitigating evidence, an attempt to develop such testimony could have begun at an earlier stage.
In sum, the record is replete with indications that a further investigation of Loyd’s mental state should have been undertaken. For further example, Feliciana administered an anti-psychotic drug to Loyd, Dr. Cox testified that Loyd’s profile was unusual and that something must have been going on, and the testimony at the eviden-tiary hearing indicated that the results of the Feliciana tests suggested organic brain dysfunction. This, however, was not explored by the Feliciana staff. Nor was it explored by Loyd’s attorneys despite Allison’s awareness of the need to develop this area.
The district court’s repeated assertion that Loyd’s counsel was under no duty to search for an expert who would testify according to Loyd’s wishes is not an incorrect statement of the law; it is, however, an incongruous statement in the context of this case in which at least one of Loyd’s attorneys was wholly unaware of the admissibility of statutory mitigating circumstances, such as mental disease or defect not rising to the level of insanity. We therefore vacate the holding of the district court on this issue and remand for reconsideration of the first prong of
Strickland,
paying proper deference to the state’s factual findings in light of Loyd’s assertion that statutory mitigating evidence was not properly investigated despite the obvious need to do so.
B.
Prejudice
Because we remand to the district court based on our conclusion that the district court failed to accord due deference to the state court’s factual findings, we only briefly address the issue of prejudice.
Strickland
instructs that the defendant is required to show that there is a reasonable probability that but for counsel’s unprofessional errors, the result of the proceeding would have been different. A reasonable probability is “a probability sufficient to undermine confidence in the outcome.”
Strickland
104 S.Ct. at 2068. Paraphrased, the reviewing court must be confident that at least one juror’s verdict would not have been different had the new evidence been presented.
The district court concluded that prejudice has not been demonstrated because the jury would have heard essentially the same testimony and recited:
While the two sets of psychiatrists (i.e., Dr. Cox and the Feliciana Report, Dr. St. Martin and Dr. Ritter at the sentencing phase and Dr. Honor, Dr. Scanlan and Dr. Sanchez at the Evidentiary Hearing) used different terminology and theories to label Loyd’s behavior, both sets testified that the following factors contributed to Loyd’s behavior at the time of the offense: Loyd was physically abused as a child; Loyd may have hit his head as a child; Loyd had problems in the Navy; Loyd had an absence of stable heterosexual relationships in his life; Loyd had a history of substance abuse and may have been under the influence of alcohol and/or drugs immediately prior to the offense; Loyd may have suffered from personal stresses, sleep deprivation and job-related hyperthermia immediately pri- or to the crime; and Loyd suffered loss of memory following the crime....
Therefore, to the extent that the jury-heard substantially the same testimony [from both sets of doctors], ... there is no reasonable probability that Loyd’s sentencing verdict would have been different had counsel retained and called [other psychiatric experts.]
District Court’s Opinion and Order at 47-48. While the district court’s assessment of the raw evidence is not in error, the court ignores the completely different and polarized conclusions offered by the two sets of doctors. The second set, which explored the gaps left by the initial medical record, indicated organic brain dysfunction such that Loyd could not control his impulses — a statutory mitigating circumstance. The primary diagnosis of the first set of doctors was one of depression, which is not included in the Louisiana statutory scheme of admissible mitigating circumstances.
The state court found that Loyd’s mental condition was of critical importance. The Supreme Court has indicated that, where a defendant’s mental condition is highly relevant, the risk of error from denial of such assistance is extremely high.
Ake v. Oklahoma,
470 U.S. 68, 105 S.Ct. 1087, 84 L.Ed.2d 53 (1985).
Without a psychiatrist’s assistance, the defendant cannot offer a well-informed expert’s opposing view, and thereby loses a significant opportunity to raise in the jurors’ minds questions about the State’s proof of an aggravating factor.
Id.
105 S.Ct. at 1097. On remand, the district court should give deference to the state court’s factual findings when considering the prejudice prong of
Strickland,
and should conduct an evidentiary hearing if the court determines that the record is not sufficiently developed on this issue.
III. CONCLUSION
We are reluctant to return this case to the district court, particularly given its lengthy detailed opinion. We must do so. The district court failed to give proper deference to the state court’s findings of fact as required by section 2254(d). Consequently, we vacate the district court’s determination and remand for further consideration of this issue in accordance with this opinion. The district court must afford proper deference to the state court’s factual findings when evaluating the prejudice prong of
Strickland,
and conduct an evi-dentiary hearing if the district court concludes that the record is not fully developed as to this issue.
VACATED AND REMANDED.