delivered the opinion of the court:
Defendant, Pamela Jacobazzi, appeals from the order of the circuit court of Du Page County denying her postconviction petition following an evidentiary hearing. On appeal, defendant argues that the trial court’s determination was manifestly erroneous. We affirm but remand for a further evidentiary hearing.
BACKGROUND
This being the third appeal in this case, the background facts are well known by the parties and this court, and there is no need for a full recapitulation. Instead, we present a brief summary of the background and will include where appropriate other facts relevant to the disposition of this appeal.
In 1999, defendant was tried for the first-degree murder (720 ILCS 5/9 — 1(a)(2) (West 1994)) of the victim, Matthew C. The State’s evidence showed that the victim was essentially comatose when his mother picked him up on August 11, 1994, from the home of defendant, who was the victim’s day care provider. The victim was hospitalized and found to have injuries including a subdural hema-toma, subarachnoid bleeding, and retinal hemorrhages. He never regained consciousness and died several months later. The State’s theory was that the victim’s injuries were so severe that they could not have been inflicted accidentally. The State further argued that the injuries would have been immediately disabling and that since there was no evidence that the victim was acting abnormally before he was placed in defendant’s care on the morning of August 11, 1994, they must have been inflicted while the victim was in defendant’s charge that day. The State noted that in the medical literature subdural *893hematoma, subarachnoid bleeding, and retinal hemorrhages in infants are often collectively referred to as “shaken baby syndrome” because they are typically associated with a violent shaking of a child. The State presented the foregoing theory through 10 expert witnesses.
The defense challenged the strength of the State’s theory that the victim suffered a violent shaking by defendant on August 11, 1994. The defense suggested that the victim’s condition on August 11 was from an injury on a prior date. Defendant testified that, on August 8, the victim was sitting on the kitchen floor when he lost his balance and fell forward from a sitting position, striking his head on the tile and sustaining a bruise to which defendant applied ice. Dr. Kenneth Sullivan, the neuroradiologist who read the victim’s CT scans on the night he was hospitalized, and Dr. Jan Leestma, a neuropathologist and defendant’s sole retained expert, testified that the CT scans showed “old” blood, indicative of a hematoma that originated days before and rebled on August 11, mimicking the indicia of shaken baby syndrome. Dr. Leestma also found old blood in the pathology slide of the hematoma. Dr. Leestma opined that the oldest of the blood in the hematoma was 10 or 11 days old and that defendant’s accident three days before August 11 might have aggravated the hematoma.
Defendant was convicted and sentenced to 32 years of imprisonment. She filed a timely appeal raising 47 separate claims of ineffective assistance of trial counsel. See Strickland v. Washington, 466 U.S. 668, 80 L. Ed. 2d 674, 104 S. Ct. 2052 (1984). We rejected these contentions and affirmed the conviction. See People v. Jacobazzi, No. 2 — 00— 0523 (2001) (unpublished order under Supreme Court Rule 23) (Jacobazzi I). One of defendant’s arguments was that trial counsel was ineffective for choosing Dr. Leestma as an expert, because (1) his testimony “was not consistent with facts developed at trial”1; and (2) he was extensively impeached with his past professional writings in which he warned against many of the angles that defense counsel themselves were using at trial to discredit shaken baby syndrome as a viable diagnosis both in general and in the case at hand. Our discussion of that issue spanned just these few sentences:
“Other allegations of error simply employ hindsight to question defense counsel’s performance, without considering the circum*894stances as known to defense counsel at the time. For example, defendant alleges that defense counsel ineffectively selected Leestma as an expert witness because he was impeached with his prior writings and his opinion was inconsistent with the facts developed at trial. Defendant fails to identify the expert witness, if one exists, who was available to testify on her behalf and would have provided more favorable testimony. Nor does defendant identify anything in the record to suggest that defense counsel knew of such a witness and failed to call her or him. We believe that this argument is little more than an attempt to construe the evidentiary weakness of defendant’s case as the ineffectiveness of defense counsel.” (Emphasis added.) Jacobazzi I, slip op. at 95-96.
We immediately followed with these comments:
“We find that the weaknesses we have identified above are common to many of the remaining allegations of error in defendant’s ineffective assistance argument. However, there is no need to inquire further into the individual allegations of error, because we may proceed directly to the second prong of the Strickland analysis. [Citation.] Therefore, the question before us is whether, in light of the evidence presented, the outcome of defendant’s trial would have been different absent the alleged errors.
The State presented overwhelming evidence of defendant’s guilt. *** [T]he State presented one medical expert after another who testified that the victim’s injuries were the result of shaken baby syndrome and could not have resulted from dancing, tumbling, or other accidental causes, except an automobile accident or a fall from a height of two or more stories. The experts also substantially agreed that the injuries occurred on August 11, 1994, during the time period while the victim was in defendant’s care. Other witnesses testified regarding the victim’s condition, narrowing the possible time frame for the victim’s injuries to a small period on the afternoon of August 11. The only medical expert who testified that the victim’s injuries could have occurred outside this time frame, [Dr. Leestma,] was thoroughly impeached with prior inconsistent statements that supported the State’s theory of the case. *** We do not find that absent the errors defendant identifies, that [sic] a substantial probability exists that the jury would have evaluate [sic] the expert medical testimony or [the victim’s mother’s] credibility differently.” (Emphasis added.) Jacobazzi I, slip op. at 96-97.2
*895On May 17, 2002, defendant filed a petition for relief pursuant to the Post-Conviction Hearing Act (the Act) (725 ILCS 5/122 — 1 et seq. (West 2002)), raising two new claims of ineffective assistance of trial counsel. First, defendant argued that counsel was ineffective for failing to call certain witnesses on her behalf and to cross-examine the victim’s mother regarding the victim’s medical history. Second, defendant contended that counsel was ineffective for failing to review, and have Dr. Leestma review, the records of Dr. David Nadelman concerning his treatment of the victim in the years before the injury. Defendant argued that counsel was ineffective for failing to present evidence at trial that the victim’s injuries may have resulted from one or more of the following preexisting medical conditions that were either expressly diagnosed in the Nadelman records or inferable from them: sickle cell trait, external hydrocephalus (enlarged head), hemophilia, fever, dehydration, and anemia. To support her claim that counsel could have presented a meritorious defense based on the Nadelman records, defendant attached affidavits from Dr. Claus Speth and Dr. Urna Subramanian.3 Both challenged the conclusiveness of the State’s claim that the various pathologies found in the victim on August 11, 1994, were the result of a violent shaking that day.
Dr. Speth averred that he is board-certified in anatomic, clinical, and forensic pathology and is a former state medical examiner. Dr. Speth noted that the victim’s subdural hematoma was “unilateral” and that “in shaken baby syndrome, subdural hematoma in the majority of cases is just a thin layer of blood that does not compress the brain and is more often bilateral.” Dr. Speth also noted: “[S]ubarach-noid hemorrhage is not a general finding characteristic of shaken baby, although it is frequently present, but then in just small amounts and not necessarily in relationship with subdural hematoma.” Dr. Speth further remarked that the victim also suffered “cerebral infarction,” which is “not generally a characteristic of shaken baby.”
Dr. Speth also opined that, assuming the conditions found in the victim have some general association with shaken baby syndrome, in *896this particular case those conditions are more likely attributable to a preexisting pathology known as sickle cell trait. Dr. Speth described sickle cell trait and distinguished it from sickle cell disease:
“Hemoglobin molecules found within the round red blood cells of the circulating blood[ ] take up oxygen in the lungs and release the oxygen out at the ends of the blood stream to the cells. Hemoglobin molecules are made up of four protein molecules called globins that chemically lend the characteristics of hemoglobin-F in the newborn, gradually changing over to adult hemoglobin-A by 4 to 6 months of age. Sickle cell disease involves the inheritance of mutated globins in the hemoglobin molecules (designated hemoglobin-S). If a child inherits hemoglobin-S from both parents (called homozygous), the child has sickle cell disease; if from only one parent (called heterozygous), the child has sickle cell trait. The hemoglobin-S tends to polymerize (form chains that lend it a viscous gelatin-like character), damaging the red blood cell membranes, causing the red blood cells to adhere to vessel walls and to eventually take on a sickle shape. In those with homozygous sickle cell disease, these abnormal red blood cells clog small vessels with multiorgan damage and infections, and resulting in vast destruction of the sickled red blood cells in the spleen resulting in anemia. In the child with only the trait, there is enough adult hemoglobin-A in the red blood cells to compensate and these complications occur only under severe hypoxia (lack of oxygen), severe exertion or dehydration. The abnormal hemoglobin can be identified by a laboratory process called hemoglobin electrophoresis. However, in some persons with apparent sickle cell trait there can be what is called co-dominance (as if they were homozygous) and there are also other variants of hemoglobin-S, and then these persons will have variable degrees of complications like those with the full sickle-cell disease. These abnormal forms of hemoglobin-S require special testing methods including DNA analysis.” (Emphases in original.)
Dr. Speth then summarized the portions of the Nadelman records that mention sickle cell trait, fever, and anemia4:
“The [victim] was found to have what was characterized as being ‘consistent with sickle cell trait’ (001096) by a method called iso-electric focusing on 10/7/93, to be repeated (001113). On 11/29/93 (001079) and on 12/6/93 (001080 & 001081) it is reported that blood drawn on 10/12/93 showed sickle cell trait (SAS trait). Because of continuing anemia, and possibly also because of recurrent upper respiratory infections, even though on 12/6/93 it stated ‘no further specimen needed’ (001081), another one was drawn on 12/7/93 and reported on 12/15/93 reported [sic] as: ‘Possible *897abnormal hemoglobin present. Please retest at 2 years of age’ (001101)[.] On 3/16/94 the [victim] was with ‘mom’s boyfriend’s sisters’ while the mother works, had an upper respiratory infection, was not sleeping and not eating (001083). On 3/21/94 the [victim] was better (incidentally the [victim] was found to have bilateral metatarsal adductens, an orthopedic deformity) (001084). On 5/9/94 it was determined again that the [victim] was still anemic (001084), indicating yet again the need to test for sickle cell hemoglobin. This anemia was again apparent on 5/31/94 (001085) and again with recurrent infections and fever. On 6/2/94 hemoglobin electrophoresis was again ordered (001085) — the results were reported as 40% hemoglobin-S, as well as continuing anemia (001086). On 7/18/94 the [victim] again had a febrile illness and dehydration, with continuing anemia (001086). Finally on 8/09 to 8/10/94 the [victim] had a febrile illness (temperature 102), one day before the incident (000001).”
Dr. Speth opined that the victim had “a variant of sickle cell trait that was causing continuing anemia and recurrent infections, and threatening to also cause the various unpredictable complications normally seen in persons with the homozygous sickle cell disease.” (Emphases in original.) Dr. Speth noted that “[t]esting to identify the exact mutation of [sickle cell trait] was never conducted.” Dr. Speth opined that sickle cell trait could have caused the conditions seen in the victim on August 11, 1994, including “a unilateral subdural hematoma compressing the brain, significant subarachnoid hemorrhage beneath the area of subdural hematoma!,] and infarction of the brain in the same region.” He explained:
“The complications that might arise under these circumstances (especially under the effects of dehydration associated with a just remitting febrile illness) include: [i] unilateral subdural hematoma (a collection of blood creating a space within the inner-most cell layers of the thick lining dura) compressing the brain, [ii] subarachnoid hemorrhage (blood spreading out through the cerebrospinal fluid over a surface of the brain under that portion of the arachnoid membrane) causing chemical irritation and spasms of the adjacent blood vessels, [iii] ischemic and hemorrhagic cerebral infarctions (destruction of brain tissue usually due to lack of blood supply and or hemorrhage into the brain tissue from vessels) and [iv] a high incidence of cerebral aneurysms (outpouching of an artery supplying blood to the brain) which can rupture into the subarachnoid and subdural spaces.”
Dr. Speth noted that he possessed “a collection of peer-reviewed articles dealing with these complications of sickle cell anemia from recognized medical journals.”
*898Dr. Speth explained that a subdural hematoma or subdural hemorrhage caused by sickle cell trait may be mistakenly attributed to shaken baby syndrome:
“It is vitally important to note that, in the context of shaken baby, an important differential diagnosis is the rupture of a cerebral aneurysm (an outpouching of an artery supplying blood to the brain) or a ruptured arteriovenous [AV] malformation (a congenital knot of arteries joined directly with veins on or in the brain). These do produce unilateral subdural hematoma and/or subarachnoid hemorrhage in the area of rupture, will cause sudden severe rise in intracranial pressure and can also cause secondary infarction of brain tissue — all found in [the victim]. It should be recalled! ] that ruptured aneurysms have been found in association with sickle cell anemia. Of note, though, is that the CT scans, performed on [the victim], are not designed to visualize aneurysms or AV malformations, and the blood flow scans that followed could not visualize the left side of the brain because of the severe swelling, where an aneurysm or malformation would have been located. The surgeon did not visualize the brain and, finally, the autopsy report of the brain does not mention examination of the arteries in the area of subdural hematoma/subarachnoid hemorrhage and brain infarction.”
Dr. Speth went on:
“Subdural hematomas in shaking do not compress the brain in the absence of beatings and therefore do not cause serious injury or death. They are simply indicators of some type of to-and-fro acceleration/deceleration, but in which all other possible causes (such as ruptured aneurysms, AV malformations, sickle cell disease and a myriad of other causes) have been fully explored and ruled out. The subdural hematomas in shaking do not in any way infer or lend information about the degree of force applied! And, as stated above, generally are not massive and unilateral.”
Dr. Speth observed that retinal hemorrhages, too, may result from sickle cell trait:
“Sickle cell disease, already beginning in infants, and progressing in early childhood, causes very significant changes in the eyes. Sickle cell disease characteristically causes retinal neovascularization (growth of new small vessels in the retina) and vitreous hemorrhages (hemorrhages in the glass body of the eye itself). But it also afflicts these infants with retinal hemorrhages and retinal detachment, albeit somewhat different in character and distribution than that which was identified in [the victim]. But the infants afflicted with sickle cell disease would obviously be vulnerable and susceptible to the rapid development of the type of findings *899described in [the victim’s] retina on August 12, 1994, when such a predisposed infant is subjected to very sudden increases in intracra-nial pressure (as occurred in [the victim]) or even minor trauma or relatively minor accelerations/decelerations. For example, it is known that the rupture of aneurysms (a high incidence of which is found in association with sickle cell disease) — with their attendant sudden increased intracranial pressure and massive subarachnoid hemorrhage, not infrequently also breaching the arachnoid causing associated subdural hematoma — can cause retinal folds and vitreous hemorrhage in non-predisposed persons. It is no great step to realize and predict that this could then be found in an infant under the circumstances when that infant is predisposed by sickle cell disease and its attendant retinal changes. Finally, oft-quoted peer-reviewed literature emphasizes that in the majority of cases, the type of unilateral retinal findings in [the victim] (but in the absence of the prediposing sickle cell disease) would generally be found in infants subjected to, not only severe shaking, but severe beating— [and] there was virtually no external evidence of a severe beating on [the victim].” (Emphases in original.)
Dr. Speth noted that for these claims as well he had “a collection of peer-reviewed articles *** from recognized medical journals.”
Dr. Speth also averred that the Nadelman records show that the victim had a “disproportionately enlarging head during his infancy,” which might indicate external hydrocephalus. Dr. Speth quoted at length from a 1999 article by Dr. Joseph H. Platt noting that, in infants with external hydrocephalus, subdural hematomas have been found “ ‘to be secondary to minor trauma or have even been described as spontaneous.’ ” (Emphasis in original.) Dr. Platt also noted that “ ‘although retinal hemorrhages in infancy has been considered virtually pathognomonic of child abuse, in the setting of external hydrocephalus a more cautious interpretation may be appropriate.’ ” Dr. Platt elaborated:
“ ‘External hydrocephalus is a common condition characterized by diffuse enlargement of the subarachnoid spaces. ***
*** Whereas enlargement of the subarachnoid spaces persists during infancy, the brain does not fill the space available within the cranial cavity, and a state of “craniocerebral disproportion” exists. The presence of craniocerebral disproportion that develops from any case — external hydrocephalus, internal hydrocephalus, arachnoid cyst, or chronic subdural hematoma — makes the patient exceptionally susceptible to subdural hematoma after what would otherwise be inconsequential trauma. In particular, the prevalence of subdural hematoma in case series of patients with external hydrocephalus seems significant. In a study reported *900by Azais and Echenne, “spontaneous” subdural hemorrhage was demonstrated in five of 41 patients with “benign enlargement of the subarachnoid spaces” on ultrasonography and computerized tomography scanning. Laubacher, et al., have described a “non-battered” infant among their 22 patients with pericerebral subarachnoid space enlargement, in whom a “spontaneous” sub-dural “hygroma” developed that the authors attributed to the vulnerability of the bridging subarachnoid veins. Mori, et al., have reported three infants with subdural hemorrhage [out] of 20 patients with infantile subarachnoid fluid collection. Although they fell within the shadow of publication bias, these series suggest that the prevalence of subdural hematoma in infants with external hydrocephalus may be as high as 11%.
Caffey, who first described the shaken infant syndrome, recognized that infants with hydrocephalus were prone to the development of subdural hematomas, but neither he nor any subsequent author has arrived at the following logical conclusion: the development of a subdural hematoma after minor trauma in an infant with craniocerebral disproportion might be the occasion for unjustified accusations of child abuse. The concomitant presence of retinal hemorrhages in such an instance would reinforce that the diagnosis of inflicted injury is incorrect in such cases.’ ” (Emphases in original.)
Dr. Speth noted that no imaging studies were performed on the victim’s head prior to August 11, 1994, and that subsequent imaging would not have detected external hydrocephalus, if any, because of the severe brain swelling.
Dr. Speth concluded:
“With regard to this case, based on all of the above, there is no compelling evidence that [the victim] was shaken on August 11, 1994, and no evidence that he was beaten or subjected to blunt injury. Whether or not there may have been [an] old injury is not at issue in this [affidavit] and was examined by the experts at trial. What was not explored by [the] Defense in Direct Examination were the more likely causes of [the victim’s] malady. High on the list would be a ruptured aneurysm (particularly in association with sickle cell anemia), a ruptured AV malformation, a so-called ‘spontaneous’ subdural hematoma due to a minor trivia energy-loading event in the presence of external hydrocephalus, dehydration[,] and sickle cell anemia, or subarachnoid infarction in conjunction with sickle cell anemia.”
In her affidavit, Dr. Subramanian averred that she is a medical professor at the University of Chicago and is board-certified in pediatrics and pediatric hematology and oncology. She opined that the *901victim’s conditions could have been caused by hemophilia, a stroke or infarction induced by sickle cell trait, or a ruptured aneurysm. Dr. Subramanian elaborated:
“Hemophilia
*** [T]he cause of [the victim’s] August 11, 1994 trauma[ ] could have been *** an undiagnosed bleeding disorder — hemophilia. Hemophilia is an X-linked bleeding disorder (mother is carrier of gene and male offspring manifests the disease). Hemophilia is due to decreased clotting factors VIII or IX. When the hemophilia is mild (level of clotting factor is between 5% and 40%), bleeding can be brought about by minor trauma. [The victim] had a history of at least one fall, hitting his head, three days prior to onset of signs and symptoms. With hemophilia the onset of bleeding can be delayed. There have been reported cases in which parents/ caregivers have been accused of child abuse and later[ ] the infant is found to be afflicted with a bleeding disorder. The absence of bleeding history in the family or absence of a carrier gene in the mother cannot rule out the diagnosis of hemophilia because gene mutation can occur. In light of the history that [the victim] fell and hit his head, on at least one occasion, within a few days prior to August 11, 1994, it can not [sic] be ruled out that the cause of the subdural bleeding that presented on August 11, 1994, could have been hemophilia.
Sickle Cell Trait
Though Sickle Cell trait per se should not cause neurological problems; however, there have been reported and documented cases in literature in which strokes as well as infarction has [sic] been known to occur as a result of Sickle Cell trait. (Handler & Perkin 1982; Reyes 1989; Greenberg T, Massey E.W 1985; and Radhakris-han K. et al. 1990).
Ruptured Aneurysm or A-V Malformation
Ruptured aneurysm or A-V malformation causing a spontaneous bleed and sudden deterioration. Since the CT Scan showed the combination of old and new blood, the possibility of an intracranial bleed due to organic causes existed. Especially[ ] with the presence of hemophilia, severe, (factor VIII or IX level is less than 1%) spontaneous intracranial bleed[ing] can occur and can be repetitive. Children less than 1 year of age can be at risk for such a bleed.
Based on my review of [the victim’s] records received from Dr. Nadelman, the existence of at least one of the above three organic conditions in [the victim] on August 11, 1994, is a strong possibility. It is impossible to rule out that the bleeding was not due to *902some underlying organic cause[ ] as opposed to Shaken Baby Syndrome because the above and other bleeding disorders were never ruled out by [the victim’s] treating physicians subsequent to August 11, 1994. Furthermore, the presence of one of the organic causes would also explain the continued bleeding suffered by [the victim] *** during his initial stay at Lutheran General Hospital subsequent to his initial surgery on August 11, 1994; and provide a contrary medical explanation to Shaken Baby Syndrome for the cause of the affliction suffered by [the victim] on August 11, 1994.”
Defendant also attached a May 16, 2002, letter from Dr. Leestma to Dr. Speth. In the letter, Dr. Leestma stated that none of the materials he was provided for his preparation in the case showed that the victim had sickle cell trait. Dr. Leestma had no memory of being told that the victim had that condition. Had he known of the victim’s sickle cell trait, he would have “incorporated [it] into [his] analysis and the formulation of [his opinion] in the case with respect to the range of medical conditions that might have or could have caused, or contributed to[,] the child’s intracranial pathology that ultimately brought about his death.”
After a preliminary review, the trial court found that defendant’s petition alleged facts sufficient to state the gist of a constitutional claim. The State later filed a motion to dismiss. Following a hearing, the trial court dismissed the petition. The court found that defendant’s arguments were waived due to her failure to raise them on direct appeal and, moreover, were not sufficient under Strickland.
Defendant filed a timely appeal challenging the dismissal of her postconviction petition. On November 24, 2003, this court affirmed in part, reversed in part, and remanded for additional proceedings. See Jacobazzi II, No. 2 — 02—1292. Specifically, this court determined that defendant’s first claim, that trial counsel was ineffective for failing to cross-examine the victim’s mother and call certain witnesses to testify regarding the victim’s medical history prior to the date of the victim’s fatal injuries, was barred by res judicata because it had been considered in defendant’s direct appeal. Jacobazzi II, slip op. at 6.
As for defendant’s second claim, that trial counsel was ineffective for failing to review the Nadelman records and provide them to Dr. Leestma, this court held that, although defendant failed to raise the contention in her direct appeal, the fundamental fairness exception to the waiver doctrine was applicable. Jacobazzi II, slip op. at 6. Thus, we went on to determine whether the claim warranted an evidentiary hearing. See People v. Edwards, 197 Ill. 2d 239, 246 (2001) (petition advances to third-stage evidentiary hearing if, together with any supporting affidavits, it makes a substantial showing of a constitutional *903violation). We defined the issue strictly as whether trial counsel and Dr. Leestma had reviewed the Nadelman records and, if not, whether Dr. Leestma would have testified differently had he known of the victim’s preexisting medical conditions as described in the records. We found that the affidavits of Drs. Subramanian and Speth cast no light on either aspect of this issue:
“[The affidavits] do not satisfy the first prong of the Strickland test, that the defendant’s trial counsel’s performance fell below an objective standard of reasonableness. [The affidavits] do not show that the defendant’s attorney did anything wrong. They certainly do not show that the defendant’s attorney failed to review and forward the victim’s medical records. They simply establish that there are other experts out there who would have testified more favorably to the defendant than Dr. Leestma. Dr. Leestma’s alleged deficient performance does not render the defendant’s attorney’s performance deficient. The fact that the [defendant] has now found two new experts who will testify differently and more favorably than Dr. Leestma also does not render the defendant’s attorney’s performance deficient.
Furthermore, [the affidavits] do not satisfy the prejudice prong of the Strickland test, that but for trial counsel’s alleged error, the result of the trial would have been different. It is important not to loose [sic] sight of the fact that Dr. Leestma was the defendant’s trial expert. The fact that Dr. Subramanian and Dr. Speth would have testified a certain way had they been given Dr. Nadelman’s records does not necessarily mean that Dr. Leestma would have testified that same way had he been given Dr. Nadelman’s records. Indeed, the question at hand is whether Dr. Leestma would have rendered a more favorable expert opinion had he been given Dr. Nadelman’s records. In other words, what is relevant in the present case is whether Dr. Leestma would have opined that the victim’s preexisting maladies were the cause of the victim’s fatal injury. Drs. Subramanian and Speth do not provide the answer to this issue.” (Emphases added.) Jacobazzi II, slip op. at 9-10.
We went on to find the insufficiency of the affidavits “inconsequential because sufficient fact questions [were] raised by the record itself.” Jacobazzi II, slip op. at 10. We found that the record disclosed that: (1) “defendant’s trial attorney may not have reviewed Dr. Nadelman’s records concerning the victim”; and (2) “defendant’s trial attorney may not have forwarded Dr. Nadelman’s records to Dr. Leestma.” Jacobazzi II, slip op. at 10. Accordingly, we concluded that “the portion of the defendant’s petition alleging that her trial attorney was ineffective for failing to review and forward the victim’s medical records should have survived second stage dismissal.” Jacobazzi II, *904slip op. at 11. We remanded for an evidentiary hearing at which the court was to address issues including, but not limited to: “(1) whether defendant’s trial counsel reviewed Dr. Nadelman’s medical records concerning the victim; (2) whether defendant’s trial counsel forwarded Dr. Nadelman’s records to Dr. Leestma; (3) whether but for defendant’s trial counsel’s alleged errors the outcome of the defendant’s trial would have been different.” Jacobazzi II, slip op. at 11.
On remand, the trial court ordered that the evidentiary hearing be bifurcated. The court would initially hear evidence relating only to issues under the first (i.e., performance) prong of Strickland, specifically, whether defendant’s trial counsel reviewed the Nadelman records and forwarded them to Dr. Leestma for review. Only if defendant proved counsel’s performance deficient under the first prong would the court take evidence relating to Strickland’s second (i.e., prejudice) prong.5 Of the three issues we expressly identified, the trial court took issues (1) and (2) to relate to the performance prong and issue (3) to the prejudice prong.
On the performance prong, the court heard testimony from Dr. Leestma and from defendant’s trial counsel, attorneys Anthony Montemurro and Richard Butera.
Dr. Leestma testified that, in November 1997, he was retained by Montemurro and Butera as an expert witness in defendant’s case. Dr. Leestma’s billing statements for his trial work, running from November 5, 1997, to May 13, 1999, were admitted into evidence. Dr. Leestma testified that, according to his billing statements, he first met with Montemurro and Butera on November 5, 1997, to discuss defendant’s case. Dr. Leestma recalled receiving the victim’s medical records for the first time at this meeting. The documents he received were photocopies. He was not given an inventory of the medical records and did not create one himself. He simply maintained a folder with all the records. Dr. Leestma brought the folder with him to the hearing, and it was entered into evidence. The Nadelman records were also admitted into evidence. Dr. Leestma testified that he first saw the Nadelman records after trial, in May or April 2002. His folder did not contain copies of the Nadelman records or give any indication that he ever reviewed the records. Dr. Leestma testified that his billing statements also contain no suggestion that he reviewed the Nadelman *905records before trial. He acknowledged that he had moved his office since trial and may have disposed of some records in his folder. He also noted that his folder was thinner than he remembered it. Dr. Leestma was sure, however, that he would have remembered that the victim had sickle cell trait had he been so informed while preparing for trial. He also would have remembered that the victim had anemia, which he noted is “not terribly unexpected” with sickle cell trait. Instead of taking account of these conditions, Dr. Leestma’s preparation for trial was premised on the victim having been a “normal[,] healthy” child prior to August 11, 1994. Dr. Leestma noted that he always asks defense attorneys in cases of this manner whether the victim had a prior illness, and in this case the answer he received was negative. Consistent with his May 2002 letter to Dr. Speth, Dr. Leestma testified that, had he known of the victim’s sickle cell trait, he would have examined the victim’s hematoma more closely.
Dr. Leestma was then asked about the victim’s medical records from Marianjoy Rehabilitation Hospital (Marianjoy), one of the facilities where the victim was treated after he was hospitalized on August 11, 1994. The Marianjoy records reported in relevant part:
“Past medical history includes sickle [cell] trait. The patient has a history of anemia which was worked up by an outside pediatrician. Laboratory values are unknown but the patient was treated with iron. The patient also has history of left metatarsus adductus.”
Dr. Leestma noted that, though his folder did not contain the Marian-joy records, he remembered reviewing them in preparing for trial.6 Dr. Leestma acknowledged testifying at trial that he received “ ‘all medical records in relation to this case’ ” and that he “ ‘paged through them.’ ”
Butera testified next. He stated that Montemurro was lead counsel on defendant’s case and had asked Butera to assist with the medical aspect of the defense. During the case, Butera never received a complete set of medical records; rather, he borrowed portions of the *906records when necessary from Montemurro. Butera testified that his own case file did not suggest, nor did he independently recall, that he ever received, reviewed, or discussed the Nadelman records with Dr. Leestma or Montemurro prior to trial. Butera did not become aware of the Nadelman records until after trial. Butera recalled at least two meetings with Dr. Leestma prior to trial but denied discussing any of the following preexisting conditions in the victim: sickle cell trait, persistent fever, anemia, possible AV malformation, bilateral metatarsal adduct, dehydration, or low hemoglobin and hematocrit readings. Butera noted that the defense theory at trial was that the victim had a prior injury or other medical problem. Butera recalled that the only information the defense team had as to a possible prior problem was Dr. Sullivan’s report that the CT scan revealed a subdural hematoma with both old and new blood. Butera did not remember reviewing any medical records that would have indicated that the victim had any blood disorder or other abnormality prior to August 1994.
Butera testified that he did not know which records Dr. Leestma reviewed, because it was Montemurro who kept the records and gave Dr. Leestma, and Butera, access to them. Butera acknowledged that, at trial, he asked Dr. Leestma if he received “ ‘all medical records in relation to this case’ ” and that Dr. Leestma replied in the affirmative. Butera also acknowledged that Dr. Nadelman testified at trial one week prior to Dr. Leestma and that Dr. Nadelman made references to the records of his treatment of the victim. Butera testified that he did not recall that, following Dr. Nadelman’s testimony, the Nadelman records were made an exhibit and ordered to stay in court until the completion of trial.
Montemurro testified that he was lead counsel at defendant’s trial but was not the original attorney to represent her in the case. Monte-murro took over the defense from attorney Louis Elovitz, who provided Montemurro his complete file, including all of the victim’s medical records. Montemurro did not recall obtaining any additional medical records after he began representing defendant. Although he could not remember where all of the medical records were subpoenaed from, he remembered that there were records from Dr. Nadelman. At some point in his preparation for trial, Montemurro had all the medical records assembled into binders. In April 1996, he retained a nurse, Joey Asher-Tamler, to prepare an abstract of the progress notes in the records. Montemurro explained that he did not have Asher-Tamler abstract the Nadelman records because (as he explained later in his testimony) Dr. Leestma “attached no significance to [the records] to indicate that we would use [them] at trial or [they] would be part of our trial strategy.”
*907Montemurro testified that he reviewed the Nadelman records “many times” in preparation for trial. Montemurro gave varying answers as to what he remembered from the Nadelman records. Asked at one point if he “recall[ed] reviewing the records which indicate that [the victim] had a fever on March 14, 1994,” Montemurro said, “Sure I do.” Elsewhere, however, he testified that “all [he] remember[ed] about [the Nadelman records]” was that the victim had sickle cell trait. When asked whether he had the Nadelman records with him at defendant’s trial, Montemurro testified that he assumed he had them but acknowledged that his recollection could be refreshed by reviewing the trial transcript. Defense counsel then introduced into evidence a portion of the trial transcript of Dr. Nadelman’s testimony. The transcript showed that when Montemurro questioned Dr. Nadelman about when he had treated the victim for illness, Dr. Nadelman stated that he would need to look at his chart to give specific dates. Monte-murro then stated, “I don’t have the chart.” At the evidentiary hearing, Montemurro explained that when he made this statement he was referring to the doctor’s original chart. At trial, he did not have the original chart but had a copy. He assumed the State had the original. Montemurro also explained that, at trial, he had wanted the original chart, rather than a copy, marked as an exhibit and saved for later reference.
Montemurro then testified about the medical records he provided to Dr. Leestma for his trial preparation. Montemurro testified that Dr. Leestma had “access” to all medical records Montemurro had for the case. Dr. Leestma “went through the medical records himself and chose what was significant and what wasn’t significant to pull out and make copies for himself.”
Montemurro described his discussions with Dr. Leestma about the victim’s prior medical conditions:
“Q. Now, did you ever discuss sickle cell — this sickle cell trait with Dr. Leetsma — 7
A. Yes.
Q. —prior to trial?
A. Yes.
Q. And he had the report at his—
A. Yes.
Q. And he had access to that report prior to trial; is that correct?
A. Yes.
Q. Do you know approximately how many times you discussed this sickle cell trait with Dr. Leetsma?
*908A. No, but I specifically recall discussing the sickle cell trait with him.
Q. And did he assign any significance to the pre-existing sickle
cell trait in the victim?
***
A. No.
# * #
Q. *** [N]ow you indicated also that you had a nurse abstract certain medical records; is that correct?
A. Correct.
Q. And you did not have the Nadelman records abstracted; is that correct?
A. Correct.
Q. And would you explain why?
A. Well, she just basically did the surgeries and the admittance to Lutheran General Hospital, plus when we ran all of this by Dr. Leetsma, he attached no significance to this to indicate that we would use this at trial or it would be part of our strategy.
* * *
Q. Now, do you recall specifically what you discussed about the child’s prior medical history with Dr. Leetsma?
A. No.
Q. You said you mentioned sickle cell trait; is that—
A. For some reason I remember that conversation.
Q. Now prior — earlier you testified that you don’t remember that the child was also diagnosed with being anemic prior to August—
A. That’s correct. I don’t remember talking about that.
Q. And you didn’t talk about the persistent fevers with Dr.—
A. I — I don’t recall speaking about that.
iji * #
Q. When you had these — was it just one occasion you discussed this prior history with Dr. Leetsma?
A. I can’t recall.
Q. You don’t — you don’t know?
A. As I sit here today I can’t recall.
Q. Okay.
Q. Do you recall who was present at the meeting when you had this—
A. It would be myself and Mr. Butera.
Q. Okay. And Dr. Leetsma?
*909A. And Dr. Leetsma.
Q. All right. Now, when you discussed the sickle cell trait with Dr. Leetsma, was it in a conversation you had or do you specifically recall pulling out some medical records to show him?
A. I don’t recall pulling out specific medical records.
Q. Okay. So you might have just mentioned to him that hey, the child had sickle cell traits. Does this mean anything to you.
A. Well, I think we read through the — the records.
Q. Okay.
A. I mean, let me say that I’m an attorney not a doctor.”
Montemurro testified that the defense theory he and Dr. Leestma eventually settled on was that the victim died of a prior injury that caused a slow bleed. Dr. Leestma supported this theory with the report of Dr. Sullivan indicating that there was both acute and chronic bleeding on the victim’s brain.
Joey Asher-Tamler testified that Montemurro hired her in April 1996 to transcribe a set of medical records. Asher-Tamler testified that she finished the transcription in June 1996 and has not done any work for Montemurro since.
Following this testimony, defendant filed a motion requesting that Butera’s case file be entered into evidence. At a hearing on that motion, Butera argued that his case file contained, in part, privileged information and requested that the whole case file not be entered as evidence. The trial court decided to conduct an in camera inspection of the case file. As a result of that inspection, the court allowed certain portions of Butera’s case file to be admitted into evidence.
Thereafter, defense counsel recalled Butera to answer questions relating to his case file. Butera testified that his case file had remained intact in his office since trial. Butera identified various portions of his case file. Defendant’s exhibit No. 14 was three pages of the victim’s mother’s discovery deposition that was taken in the related civil case for wrongful death. Defendant’s exhibit No. 15 was one page of a transcript from the victim’s father’s discovery deposition in the wrongful death case.8 The testimony on these pages contained no references to the Nadelman records but did mention that the victim suffered from sickle cell trait and an iron deficiency. Butera did not attend these depositions but copies of the transcripts were in his file. He acknowledged that one of the pages of exhibit No. 14 had a handwritten note in the margin stating “sickle cell, anemic.” Butera acknowledged that the handwriting “may” be his.
*910Defendant’s exhibit No. 16 was six pages of handwritten notes titled “General Notes — People v. Jacobazzi.” Butera explained that these were notes he made near the beginning of defendant’s case, before he had reviewed all the medical records in Montemurro’s office. Butera was becoming acquainted with the case and focused on records relating to the treatment of the victim after his hospitalization on August 11, 1994.
Defendant’s exhibit No. 17 was a one-page document titled “Questions of Dr. Leestma.” Butera identified it as notes he made to himself relating to questions for Dr. Leestma. One note fragment read: “Pin Point: Where or when would be exactly based on slide did this happen. How old.” Another fragment read: “How does diabetes and sickle cell trait [affect].” Butera believed that he created this document as well near the beginning of the case. Butera believed that the information in the notes, such as the victim’s sickle cell trait, came from conversations with Montemurro, not from Butera’s own review of the victim’s medical records.
Butera identified defendant’s exhibit No. 19 as more notes he had taken about the theories of the defense. He acknowledged that one note read: “normal baby, something happened, go along with prosecution on normal [and] flip on them.” Butera noted that, by “normal baby,” he meant “a baby that didn’t have any medical condition to explain his injury.” Butera said that the note was in reference to Dr. Leestma’s theory that the victim “seemed to be normal but there might be something else going on to cause” the bleeding. Butera testified that exhibit No. 19 was prepared “closer to trial.”
Butera identified defendant’s exhibit No. 21 as notes he made based on discussions with the attorney handling the civil case. Butera generated the notes “[s]omewhere between the beginning and the middle portion of the case.” The purpose of the meeting was to discuss the case and develop a theory. However, Butera did not recall ruling out any defense theories during or after the meeting with the civil attorney. The only documents discussed at the meeting were the discovery depositions of the victim’s parents. Butera acknowledged that there were notes stating “Pre-Existing Conditions” and “Look to pre-existing chronfic] symptomology” on the exhibit. Butera explained that these entries were intended as a reminder to look at these areas in conjunction with Dr. Leestma. Butera testified that these entries led to no exchange of records between the civil attorneys and Butera’s team.
Butera identified defendant’s exhibit Nos. 18, 20, 22, and 23 as more of his notes. He acknowledged that the handwriting on exhibit Nos. 16 through 23 was his handwriting. Butera noted that, though *911there were references in exhibit Nos. 16 through 23 to the victim’s preexisting conditions, there was no specific mention of the Nadelman records. Butera further testified that he “did not know if he saw the Nadelman records or not” prior to trial.
Following this testimony, the trial court denied the defendant’s postconviction petition. The trial court first determined that trial counsel reviewed the Nadelman records prior to trial. The court found the evidence “overwhelming” that both Montemurro and Butera reviewed the Nadelman records. The court found “clear, credible and convincing” Montemurro’s testimony that he reviewed the Nadelman records and discussed them with Dr. Leestma. The court noted that although Butera “did not recall, frankly, simply wasn’t sure” whether he reviewed the Nadelman records, Butera’s pretrial notes made it “abundantly clear” that Butera reviewed the Nadelman records. The notes were “replete” with references to the victim’s preexisting conditions. The court noted specifically (1) defendant’s exhibit No. 15, the deposition testimony of the victim’s father, which was in Butera’s file and indicated that the victim had sickle cell trait and an iron deficiency; (2) defendant’s exhibit No. 17, which contained a notation from Butera referencing sickle cell trait; and (3) Butera’s references in defendant’s exhibit No. 21 to preexisting conditions and preexisting chronic symptomology. The court also determined that the reference to “sickle cell, anemic” on page 3 of defendant’s exhibit No. 14 was in Butera’s handwriting.
The court then found that the Nadelman records had been provided to Dr. Leestma. According to the court, Dr. Leestma acknowledged that “he had access to all the records that [Monte-murro] had.” As for Dr. Leestma’s assertion that he would have remembered the victim’s sickle cell trait if he had read the Nadelman records, the court found it “destroyed” by his admission that he reviewed the Marianjoy records, which themselves referenced sickle cell trait. The court thus found “incredible” Dr. Leestma’s claim that he never saw the Nadelman records. The evidence was clear, the court found, that Dr. Leestma was given access to and provided with the Nadelman records. Accordingly, the trial court determined that there was no ineffective assistance of counsel and denied defendant’s post-conviction petition. Defendant filed a timely notice of appeal.
ANALYSIS
I. The Factual Findings on Remand
Defendant’s main argument on appeal is a substantive attack on the trial court’s factual findings that defendant’s trial attorneys reviewed the Nadelman records and provided them to Dr. Leestma. *912She prefaces it, however, with a challenge to the procedures the trial court employed on remand. She argues that the court erred in bifurcating the evidentiary hearing. In this part of our analysis, we affirm the trial court’s factual findings concerning the Nadelman records. In Part II, we hold that the bifurcation was erroneous and remand for a further evidentiary hearing.
“Following an evidentiary hearing where fact-finding and credibility determinations are involved, the trial court’s decision will not be reversed unless it is manifestly erroneous.” People v. Beaman, 229 Ill. 2d 56, 72 (2008). This deferential standard is appropriate because the trial court is in the best position to observe and weigh the credibility of the witnesses. People v. Ortiz, 385 Ill. App. 3d 1, 6 (2008).
First, we hold that there was no manifest error in the trial court’s determination that Montemurro and Butera reviewed the Nadelman records. First, we note Montemurro’s testimony. Montemurro testified that the Nadelman records were part of the original medical records that he had obtained in the case from Elovitz. He also testified that he reviewed the Nadelman records “many times” in preparation for trial. The trial court specifically found that Montemurro was clear, credible, and convincing in his testimony that he discussed the issue of sickle cell trait with Dr. Leestma before trial.
Defendant argues that Montemurro’s testimony was not credible. First, defendant claims that Montemurro’s testimony was inconsistent. Defendant points to Montemurro’s testimony that the reason he did not have Asher-Tamler include the Nadelman records in her abstract of progress notes was that, “well, she just basically did the surgeries and the admittance to Lutheran General Hospital, plus when we ran all of this by Dr. Leetsma [sic], he attached no significance to this to indicate that we would use this at trial or it would be a part of our strategy.” Defendant notes that Dr. Leestma did not become involved in this case until November 1997 and that Asher-Tamler completed her work in June 1996. We do not see the inconsistency. Defendant appears to believe that Montemurro’s statement implied that Asher-Tamler and Dr. Leestma were working contemporaneously. In fact, Montemurro spoke of two different time periods. Montemurro indicated that he did not have Asher-Tamler transcribe the Nadelman records and that later, when Dr. Leestma came on the case (after Asher-Tamler had finished), Dr. Leestma assigned no significance to the records. Defendant confusedly sees a temporal overlap and an implication that, based on Dr. Leestma’s assessment, Montemurro decided not to recall Asher-Tamler in 1997 to have the Nadelman records transcribed.
*913Second, defendant argues that Montemurro’s testimony was deceptive. Defendant claims that, when asked whether he had the Nadelman records in court when he examined Dr. Nadelman, Montemurro “attempt[ed] to confuse the issue and try to distinguish whether he had the original or a copy of the Nadelman records during the trial.” It was indeed Montemurro’s remark to Dr. Nadelman, “I don’t have the chart,” as well as his later comment to the trial court that the defense team “may” have had a copy of the chart, that led us in Jacobazzi II to remand for an evidentiary hearing on whether trial counsel had the Nadelman records prior to trial. See Jacobazzi II, slip op. at 10. At the evidentiary hearing, Montemurro explained that, in stating, “I don’t have the chart,” he meant that he did not have Dr. Nadelman’s original chart but had only a copy. Defendant’s point is that, if Montemurro had a copy of the Nadelman chart, he could have, and indeed almost certainly would have, provided it to Dr. Nadelman if he wanted him to answer questions based on the chart. Defendant also argues that Montemurro’s further explanation, that he wanted the original chart marked as an exhibit and left in court for future reference after Dr. Nadelman completed his testimony, is beside the point. We understand defendant’s concerns, but we are equally mindful that the trial court was better positioned to judge the credibility of the witnesses. It was the province of the trial court, as fact finder, to resolve tensions and even outright conflicts in Montemurro’s testimony. See People v. Evans, 209 Ill. 2d 194, 211 (2004). This included, of course, assessing the credibility of Montemurro’s explanations for his remarks at trial. We see no ground for upsetting the court’s determination.
Third, defendant notes that Montemurro admitted that the victim’s sickle cell trait was the only condition he remembered from the Nadelman records. Montemurro’s testimony was not unequivocal here, for at one point he testified that he remembered that the victim had a fever in March 1994. Again, the trial court is the superior judge of the credibility of the witnesses. See Evans, 209 Ill. 2d at 211. In conclusion, we find nothing in Montemurro’s testimony on which we can override the trial court’s express credibility determination.
We also find no ground for disturbing the trial court’s finding that Butera, too, reviewed the Nadelman records. Although Butera testified that he was not aware of the Nadelman records prior to trial, Butera’s case notes, as pointed out by the trial court, made multiple references to the victim’s sickle cell trait and anemia. Butera’s case file also contained transcripts of the civil discovery depositions of the victim’s parents. Their testimony referenced the victim’s sickle cell trait and anemia. The transcript of the father’s deposition contained a margin note, “sickle cell, anemic,” which the court found was in *914Butera’s handwriting. We recognize that nothing in the record eliminates the possibility that Butera learned of the sickle cell trait and anemia from the civil depositions or from conversations with Montemurro rather than from his own review of the Nadelman records. Neither, however, is the latter scenario foreclosed by the record. Montemurro testified that, at one point in their preparations for trial, he and Butera met with Dr. Leestma. Asked for the particulars of this meeting, Montemurro testified, “I think we read through the — the records” (emphasis added). Given this testimony, we find no manifest error in the trial court’s conclusion that Butera himself reviewed the records.
We affirm as well the finding that Dr. Leestma was provided with the Nadelman records. Dr. Leestma testified that he must not have seen the Nadelman records prior to trial because if he had seen them, he would have remembered the sickle cell trait and the anemia. The trial court found this testimony incredible because Dr. Leestma acknowledged that, prior to trial, he reviewed the Marianjoy records, which themselves indicated that the victim had been diagnosed with sickle cell trait and anemia. The trial court also found Montemurro’s testimony, that he had provided the Nadelman records to Dr. Leestma and that they had discussed the sickle cell trait, to be particularly credible. Montemurro testified that the Nadelman records were contained in the medical records he received at the beginning of the case from Elovitz. Dr. Leestma acknowledged that he testified at trial that he was provided with and reviewed “ ‘all medical records in relation to this case.’ ”
Defendant argues that the trial court improperly focused only on whether the attorneys and Dr. Leestma had “access” to the Nadelman records. Defendant misconstrues the court’s findings. The court spoke of Butera and Dr. Leestma having “access” to the records because, according to Montemurro and Butera, the records were stored at Monte-murro’s office and Butera and Dr. Leestma borrowed and made copies of them. The court found not just that counsel and Dr. Leestma had “access” to the records but that they reviewed them as well.
Defendant also argues that the trial court erred in focusing only on whether defense counsel and Dr. Leestma were aware of the sickle cell trait, since the Nadelman records showed that the victim suffered from a plethora of other medical conditions. The trial court did not mention only sickle cell trait. Rather, the court relied on (1) Butera’s case notes that referenced both the victim’s sickle cell trait and anemia; (2) Montemurro’s testimony that he reviewed the Nadelman records and discussed the victim’s medical history with Dr. Leestma; (3) Montemurro’s testimony that the Nadelman records were included *915in the victim’s medical records at the time he took over the case from Elovitz; and (4) Dr. Leestma’s acknowledgment that, at trial, he testified that he had paged through all of the victim’s medical records.
For the foregoing reasons, we affirm the trial court’s determination that trial counsel and Dr. Leestma reviewed the Nadelman records. As we explain in Part II, we remand for an evidentiary hearing on additional issues. We do want to point out that, in our analysis above, we noted several conflicts and tensions within the testimony below. While they do not suffice for us to upset the trial court’s factual determinations, nothing in our analysis should be construed as a bar to the court revisiting those findings on remand if it sees fit.
Next, we address defendant’s argument that the trial court’s bifurcation of the Strickland hearing was erroneous.
II. The Bifurcation on Remand
Defendant argues that the trial court erred by barring her from producing evidence, consistent with the affidavits of Drs. Speth and Subramanian, that the preexisting conditions identified in the Nadelman records would have been beneficial to the defense. Defendant contends that the court operated on the false assumption that the bare fact that trial counsel and Dr. Leestma reviewed the Nadelman records would be evidence enough that “trial counsel employed a trial strategy not to raise any pre-existing conditions as part of their defense at trial.” Defendant argues that the only way she could have proven that the failure to use the Nadelman records was not sound strategy was
“to demonstrate the importance of the information contained in the Nadelman Records. This is exactly the type of testimony and evidence contained in the Affidavits of [Drs. Speth and Subrama-nian] that [defendant] was not allowed to introduce because [the trial court] held that information to be relevant only if this matter proceeded into an evidentiary hearing concerning the second prong of Strickland.”
Oddly, the dissent believes this issue was forfeited because defendant “did not develop [her] contention in the body of her argument” in her opening brief. 398 Ill. App. 3d at 935. The above passage, however, comes from none other than “the body of [the] argument” in defendant’s opening brief (page 63, to be exact), where it was accompanied by citations to the record and to legal authorities.
We agree that the bifurcation was erroneous, but we do not fault the trial court. The decision to bifurcate was a reasonable application of our directions on remand, which can be fairly read as reserving these issues alone for the first-prong Strickland inquiry: “(1) whether defendant’s trial counsel reviewed Dr. Nadelman’s medical records *916concerning the victim; [and] (2) whether defendant’s trial counsel forwarded Dr. Nadelman’s records to Dr. Leestma.” Jacobazzi II, slip op. at 11. Having in essence been asked to revisit that decision, we take that invitation. We conclude that the remand directions posed too narrow a question for the first-prong Strickland hearing. See People v. Sutton, 375 Ill. App. 3d 889, 894 (2007) (reviewing court may depart from decision made in prior appeal if “palpably erroneous”).
We first note that the remand directions were insufficient even for their own purposes. Under Strickland, “[t]rial counsel has the right to make ultimate decisions with respect to trial strategy and tactics and those decisions are ordinarily not reviewable.” (Emphases added.) People v. McCullum, 386 Ill. App. 3d 495, 514 (2008). On the duty to investigate (the facet of trial counsel’s performance assailed here), Strickland said:
“[Strategic choices made after thorough investigation of law and facts relevant to plausible options are virtually unchallengeable; and strategic choices made after less than complete investigation are reasonable precisely to the extent that reasonable professional judgments support the limitations on investigation. In other words, counsel has a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary. In any ineffectiveness case, a particular decision not to investigate must be directly assessed for reasonableness in all the circumstances, applying a heavy measure of deference to counsel’s judgments.” (Emphases added.) Strickland, 466 U.S. at 690-91, 80 L. Ed. 2d at 695, 104 S. Ct. at 2066.
Strickland thus protects the “decisions” and “judgments” of counsel. Our remand directions skirted the true issue. The fact that trial counsel “reviewed” the Nadelman records and “forwarded” them to Dr. Leestma would not of itself tell us whether there was a conscious decision by trial counsel not to present evidence of the medical conditions indicated in the records. As it happened, the testimony on remand exceeded the scope of questions (1) and (2). There was testimony that trial counsel not only reviewed the Nadelman records and forwarded them to Dr. Leestma, but that counsel also discussed the records with him and, based on his assessment, elected not to incorporate them into the defense at trial.
Even with this evidence, the scope of the first-stage hearing was too narrow. Certainly, “[c]ourts will far more readily find incompetency where there has been ‘an abdication — not an exercise — of professional judgment.’ ” 3 W. LaFave, J. Israel, N. King & O. Kerr, Criminal Procedure §11.10(c), at 956 (3d ed. 2007), quoting McQueen v. Swen-son, 498 F.2d 207, 216 (8th Cir. 1974). This is not to say, however, that *917the performance inquiry has ended where it is shown that counsel’s failure to pursue a certain avenue of defense was the result of conscious judgment rather than inattention. Some avenues have such an assurance of success that the failure to pursue them is per se deficient performance. See, e.g., People v. Staton, 154 Ill. App. 3d 230, 231-32 (1987) (failure to raise statute of limitations defense); State v. Stacey, 482 So. 2d 1350, 1351 (Fla. 1985) (failure to challenge sentence on ex post facto grounds). In other cases, as where defense counsel omits certain evidence, the decisiveness of the omission may not be so readily apparent. In any scenario, however, where deficient performance is claimed, the materiality of the omission is relevant in judging counsel’s competency:
“[Wjhile the issues of prejudicial impact and incompetency are separate prongs of *** Strickland, the potential prejudicial impact of the subject dealt with by counsel reaches over into the competency determination, as that impact obviously relates to the care and effort expected from a competent adversary.” 3 W LaFave, J. Israel, N. King & O. Kerr, Criminal Procedure §11.10(c), at 970 (3d ed. 2007).
The performance and prejudice prongs overlap in the sense that the performance prong deals with the “potential impact” of an error while the prejudice prong deals with “actual (not potential) impact.” 3 W. LaFave, J. Israel, N. King & O. Kerr, Criminal Procedure §11.10(c), at 970 (3d ed. 2007).
It seems this court implicitly recognized this truth when we determined in Jacobazzi II that under Strickland’s first prong a question of fact existed “relating to serious alleged deficiencies in the defendant’s trial counsel’s performance.” Jacobazzi II, slip op. at 11. We could not have found a “serious” alleged deficiency in counsel’s performance unless we had determined that the Nadelman records contained potentially exculpatory material. Yet the questions we framed for remand relegated to the prejudice prong all consideration of the potential force of the omitted evidence. This was not proper under Strickland.
People v. Popoca, 245 Ill. App. 3d 948 (1993), cited by defendant, typifies the calculus that must underlie a first-prong Strickland analysis. In Popoca, the defendant was convicted of the attempted murder of his wife and child. At trial, the defense claimed voluntary intoxication. The defendant testified that, in the two hours before the alleged incident, he drank three or four six-inch glasses of vodka and orange juice and did not remember what happened afterward. The defendant’s wife testified that his eyes were red and glassy and his speech incoherent at the time of the incident. Other witnesses, *918however, including the arresting officers and the defendant’s friend Hernandez, testified that the defendant had control of his motor skills when they observed him shortly before and after the incident. The only scientific evidence defense counsel presented was from a lab technician who testified that the blood sample he drew after the defendant’s arrest showed a count of 201.8 milligrams of alcohol per deciliter of blood (the technician did not translate this into the more familiar blood-alcohol concentration (BAG)). The technician testified that since he was not a physician, he was not qualified to give an opinion on how that alcohol level would have affected a person. The technician noted, however, that based on the standards physicians use to treat patients (according to which a level of 100 milligrams per deciliter of blood means intoxication, 350 to 450 milligrams means severe intoxication, and 500 milligrams means a potentially fatal level of alcohol), the defendant’s blood alcohol level indicated that he was intoxicated at the time of the incident. Popoca, 245 Ill. App. 3d at 950-51.
After trial, the defendant claimed ineffective assistance, alleging that counsel presented inadequate evidence to support the defense of voluntary intoxication. The claim proceeded to a hearing at which the defendant presented two expert witnesses who described the evidence the defendant claimed was wrongfully omitted from the defense. The first, a clinical psychologist, testified that the defendant was alcohol dependent at the time of the incident and that blackouts are a common symptom of alcohol dependency. The psychologist also testified that the defendant’s BAG of 0.20 (rendered from the technician’s calculation of 201.8 milligrams of alcohol per deciliter of blood), which was twice the legal limit and four times the level of clinical intoxication, would have made him incapable of forming the intent necessary to commit murder. The second expert, a substance abuse counselor, testified that the defendant’s BAG would have affected his judgment, emotions, perception, and motor coordination. Trial counsel also testified at the hearing. Her testimony revealed that she had made a conscious choice not to present additional scientific testimony on how the defendant’s blood alcohol level would have affected his mental state. She had decided not to call an expert witness because she believed the defendant’s actions at the time of the incident illustrated his mental state “better than an expert could speculate about it.” Popoca, 245 Ill. App. 3d at 954. Counsel believed that, while further scientific evidence might not have harmed the defendant’s case, it would not have significantly helped it either. Popoca, 245 Ill. App. 3d at 952-54.
*919The appellate court held that the defendant proved his ineffectiveness claim. The court first set forth the standards for applying Strickland's first prong:
“The failure to present particular evidence does not by itself prove that a defense counsel was ineffective, because the evidence may be flawed or damaging to the defendant so that competent counsel must make a strategic choice not to present it. [Citations.] For example, where the evidence of intoxication is not strong enough to support the defense, and counsel wanted to ‘play down’ the influence of alcohol, counsel made a tactical decision. [Citation.] However, where the failure is the result of a lack of diligence in investigating the facts and law, rather than a drawback or strategy, counsel may be ineffective. [Citations.]” Popoca, 245 Ill. App. 3d at 955.
The court then noted in detail the potential strength of the omitted expert testimony. The court’s discussion here appeared also to subsume the Strickland second-prong prejudice analysis:
“Here, the State tried to disprove defendant’s intoxication was so extreme as to negate the specific intent to kill. The prosecutor relied on the testimony of Hernandez and the officers that defendant had control of his motor skills. The prosecutor argued defendant’s wife was biased in his favor by testifying he was drunk and that defendant was not credible because his alleged blackout was really selective memory loss.
However, motor skills are not directly related to the thought process. A psychologist could testify that an alcoholic may appear normal while his thinking could be totally awry and may commit irrational acts without full control of his reason; an intoxicated alcoholic would not necessarily stagger or slur his speech. [Citations.] An expert would have made defendant’s testimony that he could not remember the incident less unreasonable and less incredible or could have shown that defendant’s failure to remember the details of the offense was symptomatic of extreme intoxication. [Citations.] Moreover, defense counsel did not investigate calling the paramedic or hospital personnel to testify regarding the extent of defendant’s condition or to rebut the opinion of the officers.
While defendant’s trial counsel did introduce the laboratory technician, his testimony caused more harm than benefit to defendant’s cause. Counsel never translated the 201.8 milligrams of alcohol per deciliter of blood to the familiar 0.20 BAG or tied the reading to the statutory presumptions [citations]. Counsel did not give the jury any relevant standard to gauge the BAG. The 100/350/ 550-milligram gauge is relevant only to pathological treatment, not the impairment of mental abilities. The trial testimony implied *920that the pathological level was not severe, but a psychological expert could have testified that the level of mental impairment was extreme.” Popoca, 245 Ill. App. 3d at 955-56.
Having found that “defendant *** established a prima facie case for the ineffective assistance of counsel inasmuch as the evidence showing defendant’s mental impairment could have been far stronger,” the trial court returned to Strickland’s first prong, to “inquire further into whether trial counsel’s failure to call the experts was a matter of trial strategy.” Popoca, 245 Ill. App. 3d at 958. The court said:
“The circuit court mentioned that if defense counsel had called an expert the State would have called one as well. ***
However, just as a reviewing court should not second-guess the strategic decisions of counsel with the benefit of hindsight, it should also not construct strategic defenses which counsel does not offer. [Citation.] Trial counsel did not mention the fear of a State’s expert witness as her reason for not calling her own, and the State did not offer any opposing witness at the post-conviction hearing. Trial counsel testified that she did not think an expert would hurt the cause but rather would not significantly help it.” Popoca, 245 Ill. App. 3d at 958-59.
In Popoca, it was precisely on the issue of how the omitted evidence might have helped the defendant that the performance prong overlapped with the prejudice prong in the court’s analysis. The court’s scrutiny of counsel’s rationale for failing to introduce the evidence was not divorced from its observations on the materiality of that evidence; rather, the court’s analysis of counsel’s purported strategy was preceded by, and premised on, the court’s finding that “the evidence showing defendant’s mental impairment could have been far stronger.” Popoca, 245 Ill. App. 3d at 958.
Here, the trial court (again, reasonably applying our remand directions) scrupulously excluded from the first-prong Strickland hearing evidence of how the Nadelman records might have benefitted the defense at trial.9 The finding that trial counsel reviewed the Nadel-man records and passed them on to Dr. Leestma was but a threshold *921finding. It might have cleared counsel of a charge of inattention, but it begged the question of whether even a conscious decision not to employ the Nadelman records was competent representation under Strickland. Evaluating counsel’s decision would involve weighing their proffered rationale against the potential impact of a defense based on the Nadelman records.
Not only were our remand directions erroneous in their own right, they were in tension with a portion of our decision in Jacobazzi I. There (as recounted above), we rejected defendant’s claim that trial counsel was ineffective for choosing Dr. Leestma when his past writings made him vulnerable to heavy impeachment by the State. We faulted defendant for not identifying a witness “who was available to testify on her behalf and would have provided more favorable testimony” than Dr. Leestma. Jacobazzi I, slip op. at 95. Defendant, having learned her lesson in pleading, supported her postconviction petition with affidavits from experts whose testimony she claimed would have strengthened the defense at trial. We dismissed the relevance of that showing because it “simply establishes that there are other experts out there who would have testified more favorably to the defendant than Dr. Leestma.” Jacobazzi II, slip op. at 9. Thus, in one place, defendant was admonished to adduce a witness who “would have provided more favorable testimony” than Dr. Leestma; in another place, defendant was told that identifying such a witness simply showed that “there are other experts out there who would have testified more favorably to the defendant than Dr. Leestma.” Speaking colloquially, this was a bait and switch or, perhaps more precisely, a moving of the goal posts.
In fairness to ourselves, we did offer a rationale (yet no acknowledgment of our apparently contradictory position in Jacobazzi I), but it was misconceived. We remarked in Jacobazzi II that “the question at hand is whether Dr. Leestma would have rendered a more favorable opinion had he been given Dr. Nadelman’s records.” Jacobazzi II, slip op. at 10. We found that “Drs. Subramanian and Speth do not provide an answer to this issue.” Jacobazzi II, slip op. at 10. We now see we were in error to think that the performance question reduced to whether Dr. Leestma would have found the Nadelman records material. First, that is not the way defendant framed her ineffectiveness claim. Counsel was ineffective, she claimed, for failing to adduce “an *922 expert witness to testify consistently with the report provided by Dr. Speth and[ ] Dr. Subramanian.” (Emphasis added.) Defendant believed that counsel should have presented expert testimony based on the Nadelman records; she did not suggest that the testimony would have had to come from Dr. Leestma or not come at all. Defendant evidently assumed that, though Dr. Leestma was counsel’s chosen expert, Dr. Leestma’s recommendation not to rely on the preexisting conditions identified in the Nadelman records would not have necessarily absolved counsel of the duty of further inquiry, even to the point of seeking an opinion from another expert.
Defendant had the right view of the law. Strickland, as we emphasized above, protects strategic or tactical decisions. “[J]ust as a reviewing court should not second-guess the strategic decisions of counsel with the benefit of hindsight, it should also not construct strategic defenses which counsel does not offer.” Popoca, 245 Ill. App. 3d at 959. “A lawyer cannot be deemed effective when he hires an expert consultant and then either wilfully or negligently keeps himself in the dark about what that expert is doing, and what the basis for the expert’s opinion is.” Richey v. Bradshaw, 498 F.3d 344, 362-63 (6th Cir. 2007).
In Richey, the petitioner, who had been convicted of arson in Ohio, brought a habeas corpus action alleging that his trial counsel was ineffective for failing to challenge the State’s solely circumstantial case linking him to the crime. Trial counsel Kluge hired an expert, DuBois, who, after reviewing the reports of the State witnesses, agreed with them that the fire was caused by arson. Kluge did not question DuBois or ask him to explain why he agreed with the State. Kluge had disclosed DuBois as a witness prior to learning his conclusions. When the State learned that Kluge did not intend to use DuBois at trial, the State itself subpoenaed him, and he testified against the petitioner that the fire was caused by arson. Kluge did not cross-examine DuBois and only minimally cross-examined the State’s other experts. Kluge also did not present any expert testimony to rebut the State’s evidence that the fire was caused by arson. Richey, 498 F.3d at 347-48.
The appellate court found glaring fault in Kluge’s handling of DuBois. The court’s discussion merits a lengthy quote:
“The scientific evidence of arson was *** fundamental to the State’s case. Yet Richey’s counsel did next to nothing to determine if the State’s arson conclusion was impervious to attack. True, Richey’s counsel retained DuBois to review the State’s arson evidence, so this case does not exemplify that most egregious type, wherein lawyers altogether fail to hire an expert. But the mere hiring of an expert is meaningless if counsel does not consult with *923that expert to make an informed decision about whether a particular defense is viable. [Citation.] At bottom, the record shows Richey’s counsel did not conduct the investigation that a reasonably competent lawyer would have conducted into an available defense — that the fire was not caused by arson — before deciding not to mount that defense.
‘A lawyer who fails adequately to investigate, and to introduce into evidence, information that demonstrates his client’s factual innocence, or that raises sufficient doubts as to that question to undermine confidence in the verdict, renders deficient performance.’ [Citation.] Here, Richey’s counsel appeared unconcerned about the State’s scientific evidence from the very beginning. Although he was aware of the State’s arson theory in July, Kluge did not contact DuBois until September and did not authorize him to begin work on the case until mid-November, just two months before the start of trial. By the time DuBois advised Kluge in December that he agreed entirely with the State, Kluge might have been prevented by the imminence of trial from obtaining another expert, even had he formed misgivings about DuBois. Even more importantly, it is inconceivable that a reasonably competent attorney could have failed to know what his expert was doing to test the State’s arson conclusion, [citation], would have failed to work with the expert to understand the basics of the science involved, at least for purposes of cross-examining the State’s experts, and would have failed to inquire about why his expert agreed with the State. *** Even DuBois expressed astonishment at Kluge’s ready acquiescence to his conclusion that the State was right. ***
The point is not that Kluge had a duty to shop around for another expert who would refute the conclusions of DuBois and the State’s experts. The point is that Kluge had a duty to know enough to make a reasoned determination about whether he should abandon a possible defense based on his expert’s opinion. [Citation.] Having simply been served up with DuBois’ flat agreement with the State, and not having known either what DuBois did to arrive at his conclusion or why he came out where he did, Kluge was in no position to make this determination.” Richey, 498 F.3d at 362-63.
Given how Strickland and subsequent cases define counsel’s duty of reasonable representation, we were wrong to believe that Dr. Leest-ma’s opinion on the materiality of the Nadelman records would have been necessarily definitive no matter the context in which it was rendered. Counsel’s election not to craft a defense based on the Nadel-man records would have had to be evaluated in light of “all the circumstances” (Strickland, 466 U.S. at 691, 80 L. Ed. 2d at 695, 104 S. Ct. at 2066). Figuring significantly in this evaluation, of course, *924would be any opinion counsel received from Dr. Leestma as to the significance of the Nadelman records, for counsel’s consultation with an expert would be evidence of proper diligence. Even the recommendation of a retained expert, however, would not elevate beyond challenge counsel’s decision to forgo use of the Nadelman records. Richey illustrates how an attorney can be too accepting of the opinions of an expert. Where counsel relies on an expert’s opinion in declining to present a defense based on certain evidence, it is a genuine question whether counsel properly scrutinized the expert’s opinion before adopting it. That question involves a sliding scale determination in which the intensity of due scrutiny bears some relation to the value that the evidence in question would have for the defense, as perceived by a reasonably competent attorney. Of course, in the case of medical evidence, an attorney’s perception of what the evidence means may be strongly dependent on the explanations of medical professionals. An attorney cannot, however, escape the responsibility of exercising autonomous judgment, and, as Richey teaches, the attorney may exercise that judgment only by grasping the underpinnings of the expert’s opinion. Only with that knowledge, combined with the attorney’s legal expertise, can the attorney intelligently assess the need to verify the expert’s conclusions through consultation with another expert. See Richey, 498 F.3d at 363 (“The point is that Kluge had a duty to know enough to make a reasoned determination about whether he should abandon a possible defense based on his expert’s opinion”); see also Marcrum v. Luebbers, 509 F.3d 489, 511 (8th Cir. 2007) (“Where counsel has obtained the assistance of a qualified expert on the issue of the defendant’s sanity and nothing has happened that should have alerted counsel to any reason why the expert’s advice was inadequate, counsel has no obligation to shop for a better opinion” (emphasis added)). As the Richey court said in a prior decision in the same proceeding:
“Even though trial counsel was not a scientist, this should not relieve him of his responsibility to understand the evidence being used to convict and execute his client. *** That counsel was a lawyer, rather than scientist, should not have left him unable to comprehend even the basics of the science at issue.” Richey v. Mitchell, 395 F.3d 660, 684 (6th Cir. 2005).
Since the question in these kinds of matters is whether counsel adequately exercised his own judgment with respect to potential medical evidence, not only are the opinions of the experts consulted by counsel relevant, but so, too, is the nature of the evidence itself, for it casts light on what counsel should have gathered for himself from the evidence and should have used in assessing the opinions of the experts. *925The nature of the medical evidence in question may be elucidated through the competent opinions of experts retained by the defendant for the Strickland proceeding. In this case, defendant adduced the opinions of Drs. Speth and Subramanian. Their opinions were relevant to the competency determination, and it was error for us to implicitly direct the trial court to exclude such evidence from the first-stage Strickland hearing.
We note, of course, that there is already evidence in the record as to Dr. Leestma’s opinion of the Nadelman records.10 Because, however, the potential impact of the records was excluded from the hearing, context was lacking in which to judge Dr. Leestma’s opinion.
The dissent recognizes that “there may be cases where the performance and prejudice prongs of the Strickland test overlap,” but insists that “this is not such a case.” 398 Ill. App. 3d at 936. The dissent attempts to show this in part by distinguishing Popoca and Richey and comparing this case favorably to Marcrum — all on the merits. The dissent misconceives the true relevance of these cases. In each of the three cases the Strickland claim went to a full evidentiary hearing at which the defense was permitted to adduce the full nature of the omission. At this stage of the proceeding, we take nothing more from Popoca, Richey, and Marcrum than that defendant also should have been given an opportunity to show the import of the omission, that is, to elucidate the nature of the Nadelman records. The cases are not comparable on the merits with the present case because there has yet to be a full and proper Strickland hearing. There has yet to be a full presentation of how the Nadelman records might have benefitted the defense at trial. It is too soon for the dissent to say that “[a]ny inadequacies in Dr. Leestma’s expert assistance cannot be the basis for a meritorious ineffective assistance claim.” (Emphasis added.) 398 Ill. App. 3d at 939. The hearing below was incomplete due to the wrongful exclusion of competent evidence — such as the opinions of Dr. Speth and Dr. Subramanian — as to the potential impact of the Nadelman records. The dissent provides citation after citation to the principle that trial counsel is presumed competent in his decision to rely on the opinion of an expert. This rule might well ultimately carry the day here, but for now our concern is to ensure that defendant receives the whole inquiry that Strickland requires, which is precisely what happened in Richey, Popoca, and Marcrum.
The dissent does cite one case, Fautenberry v. Mitchell, 515 F.3d 614 (6th Cir. 2008), in which the Strickland claim was dismissed prior *926to reaching an evidentiary hearing. In Fautenberry, the defendant argued that defense counsel failed to retain “ ‘reasonable and necessary’ ” experts at the penalty phase of his proceedings. Fautenberry, 515 F.3d at 625. The defendant argued that one of his expert witnesses at the penalty phase “misdiagnosed [the defendant’s] mental condition when she concluded that he did not suffer from organic brain damage.” Fautenberry, 515 F.3d at 625. The court held that, since defense counsel had no “ ‘good reason’ [citation]” to suspect that the expert was “incompetent,” “any inadequacies in [the expert’s] assistance — assuming there were any — cannot be the basis for a meritorious ineffective-assistance claim.” Fautenberry, 515 F.3d at 625-26.
The difficulty in comparing this case to Fautenberry is that, in her postconviction petition, defendant did not question Dr. Leestma’s opinion with regard to the Nadelman records, for she insisted that counsel never gave Dr. Leestma the Nadelman records. Her attack was directed solely at counsel, whom she broadly claimed was ineffective for failing to present “an expert witness to testify consistently with the report provided by Dr. Speth and[ ] Dr. Subramanian.” Since defendant’s pleading assumed that counsel received no pretrial expert opinion on the Nadelman records, the criteria applied in Fautenberry are inapposite.
Of course, it has since come to light that Dr. Leestma might have shared his opinion on the Nadelman records with trial counsel (although the trial court made no finding on whether that opinion was indeed given). Although we are convinced that, at the time we remanded this case in Jacobazzi II, there was need enough to allow defendant to introduce evidence of the import of the Nadelman records, we think it even more imperative now that there is some indication that counsel may have lacked a proper understanding of Dr. Leestma’s opinion — assuming such was given in accord with Monte-murro’s testimony. Montemurro testified that he not only forwarded the Nadelman records to Dr. Leestma but also discussed aspects of the records with him. Above, we quoted Montemurro’s testimony at length. Of the Nadelman records, he stated that he and Butera “ran all of this by Dr. Leetsma [sic] [and] he attached no significance to this to indicate that we would use this at trial or it would be part of our strategy.” When pressed for details of that discussion, Montemurro said, “let me say that I’m an attorney, not a doctor.” There is no suggestion from Montemurro that he ever understood why Dr. Leestma found no significance in any of the preexisting conditions identified in the Nadelman records. Though Montemurro claimed it was strategy not to present evidence from the Nadelman records, it appears from *927the record that his strategy was to accept Dr. Leestma’s assessment without question. As we saw from Richey, a lawyer who forgoes a defense theory on his expert’s recommendation before attempting to understand the basis for that opinion has exercised not tactical or strategic judgment but rather obeisance unbecoming a careful and critically minded professional. If Montemurro indeed never learned why Dr. Leestma found the victim’s preexisting medical conditions not “significant,” he could not have made “a reasoned determination” (Richey, 498 F.3d at 363) not to present evidence of those conditions.
We recognize that, where counsel in Richey “did next to nothing” to challenge the State’s scientific evidence (Richey, 498 F.3d at 362), counsel here did present expert testimony questioning the State’s scientific evidence that the victim had shaken baby syndrome caused by injuries inflicted while he was in defendant’s care. Specifically, counsel suggested that the victim might have suffered a prior brain injury that rebled and caused symptoms that were mistaken for evidence of shaking. What a defense lawyer has done to craft a defense, however, does not necessarily absolve him of failing to do more. There is no “merit bank” from which an otherwise competent attorney may draw to ameliorate a critical error:
“Very often *** a single error which was both glaring and related to a matter of obvious significance has been held sufficient to establish incompetency. *** Although the court must look to the level of counsel’s overall performance, clearly negligent treatment of a crucial deficiency in the prosecution’s case or an obvious strength of the defense will outweigh the adequate handling of series of minor matters.” 3 W LaFave, J. Israel, N. King & O. Kerr, Criminal Procedure §11.10(c), at 970 (3d ed. 2007).
“The right to effective assistance of counsel *** may in a particular case be violated by even an isolated error *** if that error is sufficiently egregious and prejudicial.” Murray v. Carrier, 477 U.S. 478, 496, 91 L. Ed. 2d 397, 413, 106 S. Ct. 2639, 2649 (1986); see Kimmel-man v. Morrison, 477 U.S. 365, 385-86, 91 L. Ed. 2d 305, 326, 106 S. Ct. 2574, 2588 (1986) (failure, in rape case, to request pretrial discovery, which would have revealed that a bed sheet containing physical evidence was seized by police without a warrant, was deficient performance despite the facts that the State’s case at trial hinged less on physical evidence than on witness credibility and that defense counsel vigorously cross-examined the State’s witnesses: “While the relative importance of witness credibility vis a vis the bedsheet and related expert testimony is pertinent to the determination whether respondent was prejudiced by his attorney’s incompetence, it sheds no light on the reasonableness of counsel’s decision not to request any discovery”).
*928 Popoca is again illustrative. Many of the other cases defendant cites, People v. Morgan, 187 Ill. 2d 500 (1999), People v. Baldwin, 185 Ill. App. 3d 1079 (1989), and People v. Murphy, 160 Ill. App. 3d 781 (1987), concern basic failures of investigation. See, e.g., Morgan, 187 Ill. 2d at 545 (at sentencing stage, defense counsel “totally failed to investigate or present evidence of defendant’s medical history or social background from readily available sources,” and hence “counsel’s failure to investigate was not a strategic choice among available options”). In contrast, Popoca shows that ineffectiveness may be found even where defense counsel does not completely abdicate her duty but rather presents a coherent defense theory that, though supported by lay and scientific testimony, nonetheless lacks an essential element that counsel consciously, and unreasonably, omitted. We emphasize that we are not holding today that counsel was ineffective for failing to include the Nadelman records as part of the defense. Our holding is only that the Strickland inquiry so far has been incomplete and that further evidence must be allowed.
Applying these principles from Strickland and subsequent cases, we hold that the opinions of Drs. Speth and Subramanian are probative as to whether counsel had a duty to incorporate the Nadelman records into the defense theory despite any contrary recommendation by Dr. Leestma. Drs. Speth and Subramanian mention several preexisting conditions that might have accounted for the victim’s conditions, particularly the subdural hematoma and subarachnoid bleeding. Some of the conditions, like sickle cell trait, anemia, and fevers, are expressly diagnosed in the Nadelman records. Others, like external hydrocephalus and hemophilia, are inferable or at least surmisable from the records. What these conditions all have in common, according to Drs. Speth and Subramanian, is that they suggest a predisposition to bleeding in the victim. Such evidence might have bolstered the defense that the victim’s massive intracranial hemorrhaging could have been precipitated by even minor trauma, such as the fall described by defendant. Moreover, because these preexisting conditions also suggest the possibility of entirely spontaneous bleeding, the defense might have been able to avoid altogether the “prior trauma” theory, which had the drawback of positing substantial bleeding that was not immediately symptomatic. “The dismissal of a postconviction petition is warranted at the second stage of the proceedings only when the allegations in the petition, liberally construed in light of the trial record, fail to make a substantial showing of a constitutional violation.” People v. Alberts, 383 Ill. App. 3d 374, 376 (2008). At the second stage, all factual allegations that are not positively rebutted by the record are accepted as true. Alberts, 383 Ill. App. 3d at 376. Defendant’s postcon-*929viction petition raises a question whether the Nadelman records contained evidence of such potentially exculpatory force that counsel had a duty to incorporate them into the defense at trial. We erred in Jacobazzi II by fashioning remand directions that reasonably could be, and were, read to exclude evidence of the potential impact of the Nadelman records.
As noted in Part I, we uphold the trial court’s findings that trial counsel and Dr. Leestma reviewed the Nadelman records. We remand for a further evidentiary hearing to supplement those findings. The trial court is to address two additional questions: (1) whether, given the potential benefit, if any, that the Nadelman records might have had for the defense at trial, counsel exercised reasonable professional judgment in declining to present a defense based on the records; and (2) whether there is a reasonable probability that, but for trial counsel’s omission, the result of the proceeding would have been different. When considering either prong, the trial court must allow competent evidence from Dr. Speth, Dr. Subramanian, or other sources, as to how evidence of the preexisting conditions identified in the Nadelman records might have benefitted the defense at trial. As we noted in Part I, our affirming the trial court’s findings that counsel and Dr. Leestma reviewed the Nadelman records prior to trial does not preclude the trial court from revisiting those findings on remand should it see fit.
For the foregoing reasons, the judgment of the circuit court of Du Page County is affirmed and the cause is remanded with directions.
Affirmed and remanded with directions.
BOWMAN, J., concurs.