*1252 Opinion

I. The Case

When Eileen, then a little more than a year old, was made a dependent of the juvenile court in the fall of 1999, the trial court denied any reunification services to her mother, Olga, because of the severity of physical abuse perpetrated by the father, Jose. (See Welf. & Inst. Code, § 361.5. subd. (b) (5). 1 ) Eileen’s injuries were basically internal. Olga did not appreciate the significance of a lump on the baby’s back and did not tell the child’s doctor about it (though she did take Eileen to a sobadora). In consideration of a writ petition, this court upheld the denial of reunification services, though we noted in our opinion that Olga, whose only sin was ignorance and lack of vigilance, had been under the stresses of her own mother’s recent death from cancer, a move to a newly purchased home, and her husband Jose’s unemployment and drug addiction. We also noted that the social worker actually recommended reunification services for the mother (something fairly unexpected in such a case). Indeed, the same social worker also recommended that an evaluation be made under Evidence Code 730 to ascertain the bond between mother and child, though Olga’s trial counsel did not ask for such an evaluation.

Things changed in the interim between the November denial of reunification services and the .26 hearing held in March. Jose pled guilty to felony child abuse and was sentenced to state prison for five years. Despite the denial of reunification services, Olga attended parenting classes once a week and went to Al-Anon. She began seeing an individual counselor and never missed a session. She attended all of Eileen’s medical appointments, which sometimes entailed taking time off work. She visited Eileen as much as she was allowed, and on days when she could not visit she called and asked to have the infant put up to the phone. She consulted an attorney to complete a divorce from Jose, and Jose stated an intention to obtain a divorce himself. 2

Meanwhile, Eileen’s injuries healed. By the time of the .26 hearing she was “cleared medically” (in the social worker’s phrase) and we are informed *1253 by the minor’s brief on appeal (filed about five months later, in August) that Eileen “has now made almost a complete recovery.” Moreover, as minor’s appellate counsel notes, Eileen’s foster parents (the maternal aunt and uncle) are not willing to adopt. 3

Despite these positive changes, Olga’s trial counsel failed to file a petition for modification under section 388 when the .26 hearing came up in early March 2000. In her appeal from the termination of her rights at the .26 hearing, she now contends, among other things, that the failure constituted ineffective assistance of counsel. As we now explain, she is correct.

II. Ineffective Assistance of Counsel in the Juvenile Dependency Context

A. Why The Ineffective Assistance Claim Can Be Raised Directly on Appeal in This Case

In her opening brief filed by appointed appellate counsel, Olga addresses the merits of her ineffective assistance claim, arguing that her trial counsel’s failure to file a section 388 modification petition resulted in the termination of her parental rights. A section 388 petition would, Olga contends, have apprised the court “in a procedurally correct fashion” of the positive changes in her life and warranted the initiation of reunification services.

Speaking of procedural correctness, however, requires a preliminary detour through the question of whether Olga’s appropriate remedy is by writ of habeas corpus rather than by appeal, a subject not addressed by either of the parties. Olga’s appellate counsel has not filed any petition for habeas corpus raising the ineffective assistance claim—the claim is confined to her opening brief on appeal. For its part, county counsel has not raised the issue either. Rather, it has simply argued the merits of the ineffective assistance claim.

Despite the fact that the county has not raised the issue, we must make the detour anyway, to show that addressing the merits of the ineffective assistance claim in this appeal is consistent with a body of Supreme Court jurisprudence which indicates most of the time the proper way to raise an ineffective assistance claim is by writ of habeas corpus, not appeal. (E.g., People v. Mendoza Tello (1997) 15 Cal.4th 264, 267 [ 62 Cal.Rptr.2d 437 , 933 P.2d 1134 ] [“claims of ineffective assistance are often more appropriately litigated in a habeas corpus proceeding”]; People v. Pope (1979) 23 *1254 Cal.3d 412, 426 [ 152 Cal.Rptr. 732 , 590 P.2d 859 , 2 A.L.R.4th 1 ] [“Where the record does not illuminate the basis for the challenged acts or omissions, a claim of ineffective assistance is more appropriately made in a petition for habeas corpus.”].) In Mendoza Tello , for example, the Supreme Court said that the Court of Appeal should not have reversed a conviction of cocaine possession in an appeal on the theory that trial counsel was not effective in failing to bring a suppression motion based on a possibly illegal patdown search. The high court reversed the appellate court because of the possibility that trial counsel might have had “ ‘[additional facts, irrelevant to the issues at the trial and possibly prejudicial to appellant, [which] may very well have justified’ ” trial counsel’s decision not to file a suppression motion. (Mendoza Tello, supra, 15 Cal.4th at p. 267 , quoting Justice Rylaarsdam’s dissent at the appellate level.) The Supreme Court said the proper way to have reached the issue of ineffective assistance of counsel was on a petition for writ of habeas corpus.

The basic theory behind the general need to file a writ petition in criminal cases, rather than merely be content to raise the matter in an appellate brief, is that resources would be wasted if appellate courts reversed judgments for ineffective assistance, only to have “new defense counsel on retrial do exactly what the original counsel did” based on some “informed tactical choice” not apparent in the appellate record. (See People v. Pope, supra, 23 Cal.3d at pp. 425-426.) 4

But the rule requiring a writ is not absolute. The formulation “more often appropriately” leaves a little wiggle room. The court in Pope mentioned an exception in cases where “there simply could be no satisfactory explanation” for trial counsel’s action or inaction (People v. Pope, supra, 23 Cal.3d at p. 426 ), language which was later picked up in both People v. Wilson (1992) 3 Cal.4th 926, 936 [ 13 Cal.Rptr.2d 259 , 838 P.2d 1212 ] and People v. Mendoza Tello, supra, 15 Cal.4th at page 266 .

Mendoza Tello itself cited People v. Camilleri (1990) 220 Cal.App.3d 1199 [ 269 Cal.Rptr. 862 ] as an example of a proper consideration of an ineffective assistance claim on appeal (as distinct from on petition for writ of habeas corpus), because there the record of a “full hearing” on a suppression motion was adequate to conclude that, while trial counsel had indeed made a suppression motion, it wasn’t quite up to snuff as a “proper” suppression *1255 motion. (See People v. Mendoza Tello, supra, 15 Cal.4th at pp. 267-268.) Camilleri itself reached the ineffective assistance claim on appeal by citing the “ ‘no satisfactory explanation’ ” language from Pope . (See Camilleri, supra, 220 Cal.App.3d at p. 1203 .)

In the present case, as in Camilleri , the “no satisfactory explanation” exception to a general mandate requiring ineffective assistance claims be raised by habeas corpus also applies. There simply could be no satisfactory explanation for Olga’s trial counsel’s failure to file a request for a section 388 modification under the circumstances of this case: Olga clearly wanted reunification services (hence the earlier writ proceeding), and she had nothing to lose by filing the section 388 petition. Nothing. Without reunification services, her chances of avoiding a termination of parental rights at the .26 hearing based on the (c)(1)(A) “benefit exception” 5 were effectively nil. Her only hope to maintain her parental rights was a section 388 modification petition. It is the extremely rare case where parents who are offered reunification services and then have them validly terminated at a referral hearing can successfully assert the (c)(1)(A) benefit exception, it is virtually impossible for a parent who was denied reunification services at the very beginning of the case to successfully assert the exception. (See In re Arturo A. (1992) 8 Cal.App.4th 229, 239 [ 10 Cal.Rptr.2d 131 ] [quoting Senate Committee Report to the effect that decision to terminate parent rights will be “ ‘relatively automatic’ ” when court determines that minor cannot be returned home and then terminates reunification services].) Indeed, there was not even a downside risk of any embarrassing information coming out at a modification hearing that might prompt social workers to try to remove any other children who would then have been in her custody, because Eileen was Olga’s only child.

Nor could one say that counsel was somehow trying to stay on the judge's good side by avoiding a “frivolous” section 388 petition. (As we explain below, such a petition would have been the opposite of frivolous.) In any *1256 case where a parent goes out of her way to see a counselor and pay for it out of her own pocket, take classes and enter a rehabilitation program, and never miss a visit, and begin what for her are costly divorce proceedings, we may safely assume that a section 388 petition would not be frivolous. In short, Olga had no place to go but up by filing a section 388 modification petition. (See In re Marilyn H. (1993) 5 Cal.4th 295, 309 [ 19 Cal.Rptr.2d 544 , 851 P.2d 826 ] [the “escape mechanism” provided by § 388 is vital to constitutionality of California’s dependency scheme].) 6

B. Why the Ineffective Assistance Claim Can Be Raised in an Appeal from the .26 Hearing in This Case

Another preliminary problem is whether ineffective assistance claims are, as intimated by In re Meranda P. (1997) 56 Cal.App.4th 1143 [ 65 Cal.Rptr.2d 913 ], per se barred after a .26 hearing has taken place at the trial level. In Meranda P. the court refused to issue an order to show cause in response to a parent’s petition for habeas corpus, holding that the mother’s claim of ineffective assistance of counsel at a variety of stages in the dependency process had been waived for purposes of a habeas petition filed after a .26 hearing. (See 56 Cal.App.4th at pp. 1160-1166.)

Meranda P. is, at root, an exposition of the implications of subdivision (i) of section 366.26, which provides that orders permanently terminating parental rights are “conclusive and binding,” albeit subject to the right to appeal the order. (See In re Meranda P., supra, 56 Cal.App.4th at p. 1161 .) Because, as the Meranda P. court put it, the “statute forbids alteration or revocation of an order terminating parental rights except by means of a direct appeal from the order” (ibid., italics added), the court reasoned that a mother who had not appealed or challenged by writ a series of errors that occurred at various prior proceedings, including even the detention hearing, could not raise her ineffective assistance challenge in a petition for habeas corpus to the appellate court filed contemporaneously with her appeal. (Id. at pp. 1152, 1160-1161.)

Without writing a law review article on the Meranda P. analysis, 7 it is enough to note here that the case, like Mendoza Tello in the context of the *1257 presumption in favor of habeas corpus in criminal cases, still leaves the door open a crack for ineffective assistance challenges in statutorily provided-for appeals from .26 terminations. The central thesis of Meranda P. is that by statute the only challenges allowed to .26 terminations are by appeal, and therefore the failure to use the appellate remedy to challenge ineffective assistance earlier was, in effect, a waiver.

It must also be remembered that Meranda P. involved a number of meaningful opportunities for the parent to challenge the efficacy of the assistance she was receiving, none of which she availed herself. The ineffective assistance of counsel allegation there was based on an alleged “collection of erroneous orders” (including even a detention order) which were the result of lack of counsel or incompetent representation by counsel. (See 56 Cal.App.4th at pp. 1146, 1152.) The court noted the mother had the right to an appeal of an allegedly erroneous jurisdictional and disposition order long before her parental rights were terminated at a permanency hearing. She also had the right to appeal from the interim 6- and 12-month review orders. Further, she was clearly present at the 12- and 18-month review hearings (and would have had no excuse for not being present at the other hearings). (See id. at pp. 1146-1148.) Clearly, if the mother’s lawyers *1258 had adopted a do-nothing posture and she were dissatisfied with it, she could have spoken up. Under such circumstances, the Meranda P. court could accurately say that the mother had “squandered” at least four opportunities to complain to the appellate court about ineffective representation. (Id. at p. 1158.)

In the present case, by contrast, the initial denial of reunification services resulted in the quick acceleration of the case, disposition to termination within five months. Moreover, it is safe to say that a woman in Olga’s position would not have known that her only real chance of derailing the .26 hearing was a section 388 modification petition. She would reasonably rely on the professional expertise of her counsel to know about that procedural “escape mechanism.” But how many lawyers who practice outside the dependency system have ever heard of section 388 modification orders? Counsel’s failure to file such a petition is clearly not something she could have ever known about. 8 In short, unlike Meranda P., a waiver of the ineffective assistance claim could not reasonably be found. Olga’s only real chance to raise an ineffective assistance of counsel claim as to the failure to file a section 388 modification petition is now. 9

C. The Standard in Juvenile Dependency Cases Is Whether the Parent Has Made a Prima Facie Showing of Prejudicial Ineffective Assistance

Now that we have shown that reaching the merits of Olga’s ineffective assistance claim is consistent with existing law, logically, the next questions are: What exactly must she show to prevail on her claim in this appeal, and what happens if she does prevail?

Preliminarily, one distinction must be drawn between juvenile dependency cases and criminal cases as far as ineffective assistance claims are concerned. Unlike the situation in criminal cases, a reversal in juvenile dependency case for ineffective assistance by the trial counsel will not put the case *1259 back into the same procedural condition as it was at the time the trial court made its supposed error. (See In re Arturo A., supra, 8 Cal.App.4th 229, 243-244 .) The simple fact is that appellate courts cannot, in dependency cases, mechanistically make determinations without regard to the passage of time and intervening events in a child’s life. (See generally In re Jasmon O. (1994) 8 Cal.4th 398, 414-422 [ 33 Cal.Rptr.2d 85 , 878 P.2d 1297 ] [court on remand obligated to consider additional facts concerning the child such as the nature of the child’s bonds to his or her current caretaker]; Arturo A., supra, 8 Cal.App.4th at p. 243 [very nature of juvenile dependency law does not allow the appellate court the luxury of “static conditions”].) As the Arturo A. court explained, a reversal of an order freeing a child for adoption because of ineffective assistance necessarily entails what would be, in effect, an “updated review hearing” in which the court would consider the child’s “current status.” (Id. at p. 245.)

The need to consider a child’s current status on any remand will, however, effectively act as a fail-safe mechanism operating in a child’s best interest when there is a successful ineffective assistance claim. 10 Strong attachments to current caretakers may still justify a decision to preclude the parent from receiving further reunification services, or preclude direct placement in some other context. (Cf. In re Jasmon O., supra, 8 Cal.4th 398, 418-419 [in § 388 hearing trial court could place “great weight” on the effect of severing bond with foster parents after “lengthy period of foster care”].) On the other hand, where there genuinely has been a miscarriage of justice because of counsel’s inaction, and the child will benefit from reuniting with the parent whose previous assistance was ineffective, a family unit will not have been unnecessarily rent forever asunder.

There is, however, a flip side to this fail-safe mechanism. A parent need not convince the appellate court that the ineffective assistance merits a reversal in which the trial court is directed to make things the way they would have been had the assistance been effective. For example, in this case we need not direct that the trial court grant a hypothetical section 388 petition that should have been made but never was. Rather, a prima facie *1260 showing of prejudicial ineffective assistance is all that is required for the updated review hearing that reversal brings. (See In re Kristin H. (1996) 46 Cal.App.4th 1635, 1671-1672 [ 54 Cal.Rptr.2d 722 ].) Ineffective assistance itself is defined as the reasonably probability “ ‘that a result more favorable to the appealing party would have been reached in the absence of the error.’ ” (Id. at p. 1668.) 11

D. Olga Has Made a Prima Facie Showing of Prejudicial Ineffective Assistance Here

We have already recounted the dramatic changes that Olga made in the short period between the dispositional order denying her any reunification services at all and the .26 hearing. What is most remarkable is that she did not abandon hope when the trial court denied her any reunification services, but initiated what was, in effect, her own reunification plan, including counseling, parenting classes, Al-Anon, and steps to keep Jose out of her life. Jose—the actual perpetrator of the original abuse—left the picture completely. There clearly was a change of circumstances—the question is whether she has made a prima facie showing that it was “reasonably probable that a result more favorable” to her would have occurred had her trial counsel filed a section 388 modification petition?

The section 388 modification petition plays a particularly important role in juvenile dependency cases, because it is the mechanism by which the changing conditions of a family and a child’s life may be taken into account. (See generally In re Marilyn H., supra, 5 Cal.4th at p. 309 .) Indeed, the mechanism is vital to the constitutionality of our basic dependency law. (Ibid.)

*1261 While pegged to a “best interests” standard, rather than the usual binary “detriment, no detriment” determination (e.g., § 366.22, subd. (a) [18-month review standard]), there is a structure for trial courts to use in analyzing modification petitions. At least three factors should be evaluated in any modification petition: (1) the seriousness of the problem which led to the dependency and the reason for any continuation of that problem; (2) the strength of the relative bonds between the child and his or her parents and caretakers; and (3) the degree to which the problem leading to the dependency may be easily removed or ameliorated, and the degree to which it actually has been. (In re Kimberly F. (1997) 56 Cal.App.4th 519, 532 [ 65 Cal.Rptr.2d 495 ].) Application of these factors to the present case shows that it was certainly “probable” that a result more favorable to Olga would have occurred if trial counsel had filed a modification petition.

While the physical abuse that led to the dependency was very serious, the precise nature of the “problem” leading to the dependency as it impacted Olga was her inability to be present when Jose physically abused the child for crying (she was working, he was unemployed and at home). Olga also failed to tell Eileen’s doctor about a lump (the outward sign of the internal injuries), an omission that is not going to happen again given the experience she has now had in dependency system. It would be different, of course, if Olga were the perpetrator of the abuse, or even if there was evidence of a pattern of forming relationships with abusive men and failing to protect her child from them, but the county points to no evidence in that direction. Here, the “problem”—from Olga’s point of view—was one of omission and ignorance. These are not in the same category as drug abuse, crime, or being the “offending parent” in a case of physical abuse.

Even more significant is the relative strength of the bonds. At this point we must note that minor’s counsel, appointed to represent her in this appeal, has taken a different position from that taken by the child’s counsel at the trial level. Minor’s appointed appellate counsel now sides with the mother, while at trial the child’s counsel was deadset against any reunification, even though' the social worker favored reunification services. 12 Minor’s appellate counsel cites certain guidelines for appointed minor’s counsel on appeal which appear to require a significant change of circumstances before appointed appellate counsel may depart from the position taken by the trial counsel; he believes that this is a such a case because he has ascertained that the current caretakers do not want to adopt Eileen. They were pressured into agreeing to adopt on two hours’ notice under the threat that if they did not do so, Eileen would be placed outside the extended family. (In reality, the *1262 guidelines for appointed minor’s counsel should not be so conservatively read as minor’s counsel apparently has in this case. Appointed appellate counsel should always use his or her independent judgment regardless of the position taken by trial counsel. 13 )

With the revelation that the current caretakers did not really want to adopt Eileen (and the pressure to say otherwise which would presumably have been brought out by competent trial counsel prosecuting the § 388 petition), the bond factor overwhelmingly favors Olga, who, in the face of the denial of reunification services, assiduously visited Eileen every chance she got, took classes and entered counseling sessions all in the hopes of maintaining her ties to her child.

The third factor also overwhelmingly favors Olga. The cause of the injuries—and the occasion of Olga’s sin of omission—is gone. Jose is in prison. Restraining orders or other orders within the context of the dependency can be “easily” obtained to limit his contact with his child. And Olga’s parenting classes have presumably taught her the need to insure that any caregiver with whom she leaves her child must be trustworthy.

In short, a section 388 modification petition was a clear winner. Olga has demonstrated a prima facie case of prejudicial ineffective assistance.

*1263 Disposition

In light of our determination, Olga’s other claims of error need not be addressed. The order terminating Olga’s parental rights at the .26 hearing is reversed. The juvenile court is directed to conduct a review hearing in which it will consider, under section 388, whether changed circumstances, including circumstances which may have arisen during the pendency of this appeal, mean that Eileen’s best interests will be promoted by offering reunification services to Olga.

For his part, Jose filed a letter pursuant to In re Sade C. (1996) 13 Cal.4th 952 [ 55 Cal.Rptr.2d 771 , 920 P.2d 716 ] declining to present an opening brief on the merits. From what we have already indicated about Jose’s role in this case, that was clearly the wisest course. In light of the reversal of the termination order as to Olga, .however, we are required by rule 1463(g) of the California Rules of Court (“the court shall not terminate the rights of only one parent unless that parent is the only surviving parent”) to reverse the termination of Jose’s rights as well. We trust that any reunification services offered to Olga will make adequate provision for Eileen’s protection from Jose.

Bedsworth, J., and O’Leary, J., concurred.

On December 18, 2000, the opinion was modified to read as printed above.

All further statutory references are to the Welfare and Institutions Code. All references to “.26” are to section 366.26 of that code, and all references to “(c)(1)(A)” are to subdivision (c)(1)(A) of section 366.26.

After the detention, Olga contacted a lawyer about obtaining a divorce, and was quoted the price of $2,500, which she didn’t have. The county counsel makes no argument that Olga’s inability to complete dissolution proceedings somehow evidences a lingering attachment to Jose. The uncontroverted evidence is that Olga does not want anything to do with him anymore.

According to minor’s brief on appeal, a social worker gave them two hours to agree to adopt Eileen on threat of placement outside the family.

Justice Mosk dissented in Pope , making pretty much the same point in the opposite direction. That is, where the record on appeal “includes sufficient information” to consider the ineffective assistance claim, it serves judicial economy to do so, lest there be “potentially duplicative hearings and appeals.” (People v. Pope, supra, 23 Cal.3d at p. 440 (dis. opn. of Mosk, J.).)

The “benefit exception” found in section 366.26, subdivision (c)(1)(A) may be the most unsuccessfully litigated issue in the history of law. In the wake of the Legislature’s attempt to annul our Supreme Court’s decision in In re Matthew C. (1993) 6 Cal.4th 386 [ 24 Cal.Rptr.2d 765 , 862 P.2d 765 ] (error in terminating reunification services at 12- or 18-month review could be raised on appeal from .26 termination) by adding subdivision (l) to section 366.26 (error at referral hearing must be raised in writ petition), the benefit exception is one of the few issues appointed appellate counsel have left. And it is almost always a loser. The real problem is that the single most important aspect of appellate representation undertaken on behalf of indigent parents is not the appeal from the .26 hearing, but the preparation of the writ petition challenging the termination of reunification services, and that task has been shifted to typically overworked trial counsel, who spend most of their days covering trial courtrooms, and do not have the time for the reflection typically afforded appointed appellate counsel.

Our remarks on the point should not be taken out of context. We are not saying that trial counsel should always ñle a section 388 modification petition because a parent would ordinarily have nothing to lose. It is enough to note that this parent had nothing to lose by filing such a petition.

The Meranda P. court gave two other reasons, besides a strict analysis of the language of section 366.26, subdivision (i). First, it relied on an adoption case, Adoption of Alexander S. (1988) 44 Cal.3d 857 [ 245 Cal.Rptr. 1 , 750 P.2d 778 ], extracting from that decision a common law principle that habeas corpus was not available to attack “a final, statutorily *1257 (§ 366.26, subd. (i)) nonmodifiable order, issued by a court with subject jurisdiction over the cause.” (In re Meranda P., supra, 56 Cal.App.4th at p. 1162.) Second, as a matter of public policy, the Meranda P. court indicated that the “paramount interests of the child in a stable, secure, long-term, continuous home environment and the associated interest of the state in reasonable expedition and finality” overcome the “parent’s interests in maintaining the family relationship.” (Meranda P., supra, 56 Cal.App.4th at p. 1163.)

Because Olga’s case deals with an ineffective assistance claim raised directly on appeal from a .26 hearing, Alexander S. is clearly distinguishable. In Alexander S., a biological mother from Romania failed to challenge by appeal an order denying her petition to withdraw her consent for a voluntary adoption. Then, when that order was final and the time for appeal had lapsed, she took an appeal from another order, this time one denying a petition to establish a father-child relationship (even though the father remained in Romania). The appellate court reached out to treat her appeal from the denial of the father-child petition as a petition for habeas corpus challenging the order on the petition to withdraw. That “creative use of habeas corpus” was rejected by the Supreme Court. (See Adoption of Alexander S., supra, 44 Cal.3d at pp. 862, 864.) The present case involves no such “creative use of habeas corpus.”

As to the general public policy rationale in Meranda P., suffice to say that the Legislature itself contemplated the possibility that the interest of the state and the child in the stability provided by an adoptive home after termination of reunification services could still he overcome by the rights of the parent in a successful appeal from a .26 hearing—lest the Legislature would never have provided for appeals from .26 hearings, which it has (under § 395). Despite our earlier comments about the likelihood of success in an appeal based on the (c)(1)(A) benefit exception, we decline to be so cynical as to presume that the Legislature really never expected any parent to fit within the exception. And since the Legislature itself allows for the possibility of reversals of .26 termination orders, we can hardly say that the interest in stability is so “paramount” that it absolutely precludes any ineffective assistance challenge.

The importance of a parent’s having actual notice of the opportunity to take the action omitted by trial counsel is illustrated by In re Cathina W. (1998) 68 Cal.App.4th 716 [ 80 Cal.Rptr.2d 480 ], issued by the same court (though different panel) that filed Meranda P. Cathina W. involved the failure to file a writ petition challenging the termination of services at a referral hearing. In an appeal from a termination order at the .26 hearing, Cathina W. allowed a review of prior findings made at the referral hearing because the trial court had failed to comply with a rule of court that requires it to send a first class letter to any party parent not present at a 12-month review or an 18-month review advising him or her of the need to file a writ petition. (See Cathina W., supra, 68 Cal.App.4th at pp. 721-724.)

Because this case properly involves an ineffective assistance claim raised by appeal rather than by writ petition, we are spared the need to consider the effect of Meranda P. had such a writ petition been filed. Whether, consistent with federal due process, the juvenile dependency statutes could absolutely preclude a parent from ever bringing an ineffective assistance challenge after a .26 hearing in any context is a matter which may be left for another day.

True, any reversal of a .26 termination order does cast at least some doubt on whatever stability a child may have in terms of an adoptive placement. But if avoidance of that instability were of automatically overriding importance, the Legislature should never have provided for appeals from .26 hearings in the first place. If the sole consideration was the absolute minimization of any disruption to the placement of the dependent child, it would make far more sense to provide for direct, expedited appeals from hearings where reunification services were either terminated (usually 12- or 18-month reviews) or denied in the first place. But that solution would require additional resources be committed to the appellate courts, because currently there is a cost savings to having trial counsel do appellate work in what too often amounts to their spare time.

In re Kristin H., supra, 46 Cal.App.4th 1635 is itself generally illustrative of the process. There, trial counsel failed to call to the stand, in a jurisdictional hearing, a psychiatrist who had prepared a report rebutting the notion that the mother was mentally ill. (See id. at p. 1672.) The court noted that the record contained no “practical or tactical reason” to explain what appeared to be the attorney’s failure to act in a reasonably competent manner, and said that the mother made a “prima facie showing” that the failure was prejudicial. (Id. at pp. 1671-1672.) But instead of granting direct relief, the appellate court returned the matter to the trial court to determine whether an evidentiary hearing was needed and, if so, to identify the issues to be decided. The trial court was then to “make findings and grant or deny appropriate relief.” (Id. at p. 1673.)

Given what the Kristin H. court said generally about the merits of the parent’s right to effective counsel and the showing that the parent made that she had been subjected to prejudicial ineffective assistance (see In re Kristin H., supra, 46 Cal.App.4th at pp. 1661-1672), it is not entirely clear why the court left to the trial court’s discretion the question of whether “an evidentiary hearing [was] needed.” (See id. at p. 1673.) We would surmise that the answer is that the record on the habeas corpus petition—which, after all,.went to whether the trial court might have come to a different result at the jurisdictional hearing—was sufficient to allow the trial court to make the ultimate determination of prejudice without an evidentiary hearing.

Indeed, a major portion of our earlier opinion was taken up explaining how the trial court could deny reunification services despite the social worker’s recommendation.

“Guidelines for Minor’s Counsel on Appeal” promulgated by a committee appointed by Division One of this district and adopted by Appellate Defenders, Inc. for the Fourth District Court of Appeal state, in pertinent part: “Deference to trial counsel. In deciding what position to take, the rebuttable presumption is that appellate counsel should defer to trial counsel and take the same position on appeal as was taken at trial, unless appellate counsel believes trial counsel was clearly wrong or unless circumstances have changed significantly.”

The above language appears to be somewhat at odds with the balance of the guidelines, which emphasize the importance of an independent evaluation by minor’s counsel. (E.g., the guidelines also state: “It is important to emphasize that every case is unique, and the overriding principle always is that counsel must be diligent in protecting the minor’s interests, as determined by the circumstances of the particular case.”)

This case affords us the opportunity to reemphasize the importance of appellate counsel’s independent evaluation. It is fundamental that the duty of an independent minor’s appellate counsel is always to his or her client, not the client’s trial counsel. Presumptions as to what might be best for the child cannot take the place of counsel’s own duty to make an independent evaluation based on the circumstances in each case. The position taken by independent minor’s counsel on appeal can make a particular difference in dependency cases, where evidence as to how well the child is doing in a placement during the pendency of the appeal may be taken into account by the appellate court. (See In re Jonathan M. (1997) 53 Cal.App.4th 1234, 1236, fn. 2 [ 62 Cal.Rptr.2d 208 ].) In such cases, with the stability of a child’s life and his or her well-being at stake, it is especially important that independent minor’s appellate counsel evaluate what is in the best interests of their clients without any presumptions. (Of course, in most cases, there will have been a substantive reason trial counsel will have taken a position at the trial level that will still be valid during the pendency of the appeal, but that is most cases, not all of them.)