MILKEY, J.

In this care and protection case, a Juvenile Court judge found the mother of two children unfit and terminated her parental rights as to them. The judge's decision was based in critical part on her assessment that the mother was unable to appreciate or address both children's extensive special needs. The mother and one of the children have appealed. For the reasons that follow, although we agree with the judge that serious issues regarding the mother's fitness have been raised, we nevertheless conclude that various shortcomings in the proceedings necessitate that the matter be remanded.2

Background.3 1. The mother. Born in 1980, the mother moved to Massachusetts at age seventeen with her parents. In January of 1999, the Department of Children and Families (DCF) removed the mother from her home after reports that she was being physically abused, and she remained in DCF's care and protection until August of 2002 (the month she turned twenty-two). At that point, the mother was placed with the Department of Developmental Services (DDS) because of her mental disabilities.

The judge did not make findings about the degree of the mother's mental disabilities, nor is that clear from the trial record. On one hand, there is a reference in one of the exhibits to the mother being "very cognitively challenged," and the extent of the services that she has received from DDS suggests a substantial mental disability. On the other hand, there are other suggestions in the record that her disability is only "moderate" or even "mild," with one person describing her as being "smart as a whip and doing fine." No expert testified as to her disabilities; in fact, no one from DDS testified at all.4 There was evidence that the mother's therapist had assigned to her a "global assessment functioning [rating] of 51 out of 100," without any explanation of what that meant or how it mattered. Although the judge appears to have accepted that assessment, no evidence of how that particular level of functioning affected the mother's parenting skills was presented.

In 2016, the mother was diagnosed as also suffering from a moderate degree of major depressive disorder.5 In addition, it is uncontested that the mother is morbidly obese (weighing over 500 pounds at the time of trial), and that this condition at least somewhat affects her mobility.

2. The nature of the trial evidence. Before turning to a summary of DCF's involvement with the family, we highlight the nature of the evidence adduced at trial. Although the mother testified at length, her testimony mainly addressed the period after the children were removed in September of 2014. There were three other witnesses at the trial: the woman who had served as the foster mother of the children for a period of time after their removal, the DCF adoption worker assigned in 2015, and the DCF social worker assigned in 2015. Thus, there was almost no live testimony that addressed the mother's parenting of the children while they were in her care.

Instead, DCF relied on the documentary record with regard to what happened prior to the children's removal. That record, consisting of thirty-six trial exhibits, included reports produced pursuant to G. L. c. 119, §§ 51A and 51B (51A reports and 51B reports), and the report of a court-appointed investigator appointed pursuant to G. L. c. 119, § 24. Those reports were submitted with agreed-upon redactions, and the parties and the judge shared a common understanding that the 51A reports could be used only to "set the stage," and that the 51B reports were "admitted to the extent that they contain[ed] primary fact and statements of the mother." See Custody of Michel, 28 Mass.App.Ct. 260, 266-267 (1990). See also Adoption of Luc, 94 Mass.App.Ct. 565, 566-569 (2018). We turn next to what the documentary record established with regard to DCF's involvement with the family prior to the removal of the children.

3. The children. Chad was born in November of 2006. The mother and Chad lived in a DDS-funded group home in the Dorchester section of Boston, operated by Dare Family Services (Dare), where the mother had been placed just before Chad was born. At the group home, the mother had twenty-four hour assistance from a full-time staff supplied by DDS. The record reflects that during this period, the mother was able to attend to Chad's everyday needs with the help of the Dare staff, which led DCF to close a case it had opened before newborn Chad had been discharged from the hospital. In June of 2008, a 51A report was filed expressing concern over the mother's yelling at Chad and lack of attentiveness toward him, but DCF closed the matter after concluding that the mother showed apparent improvement.

By February of 2011, the mother and Chad had moved to a "shared living" home in the Roxbury section of Boston, still with round-the-clock supervision, where they remained under the care of DDS. That same month, the mother gave birth to Anne, and the family moved back to Dare's Dorchester group home. During this period, DCF looked into allegations that the mother was neglecting newborn Anne and physically abusing Chad. However, further investigations "indicated that the mother and children were doing fairly well," and DCF ultimately concluded that the allegations were unsupported.

4. The family moves to Brockton. In December of 2011, DDS moved the mother and her children to a foster home in Brockton. At the Brockton home, also run by Dare, the proprietor, Betsy Goodacre (a pseudonym), looked after the family. In addition, the mother was later provided the weekly services of an aide to assist her in staying on top of appointments and such. The family remained at the Brockton home until the summer of 2014.

In March of 2012, Chad, then five years old, began to exhibit sexualized behavior at school, such as telling a girl that he wanted "to lick her between her legs." Concerned that Chad might have been subjected to sexual abuse, DCF investigated. The mother and Chad denied any allegations of sexual abuse, and DCF did not find such allegations supported.6 DCF did document a number of ways that Chad could have been exposed to inappropriate sexual images or activity. These included his having observed instances of nudity and sexual behavior involving third parties at the Dorchester group home, and his having observed pornography on the mother's tablet computer or cellular telephone. Notably, the clinical supervisor at Dare — one of two parties who reported Chad's sexualized behavior — nevertheless concluded that Chad could remain in the home (so long as he had a separate bedroom), and she expressed her view that the mother "ensures that the basic needs of her children are met at all times."

In August of 2012, while the family was still living in Brockton, the mother had to be hospitalized for a physical ailment, and she left her children in the care of Goodacre. This led to a 51A report expressing concern about the mother's ability to care for the children, especially if her absence became prolonged.7 Apparently after the mother was discharged from the hospital, DCF screened out the concerns based on existing supports.

In February of 2014, an early intervention provider for one of the children expressed concerns to DCF about the mother's ability to care for the children. The concerns were over whether the mother's cognitive issues and weight-related immobility were causing the mother not to meet the children's basic needs or to follow up on recommended services for them. At this time, Goodacre stated her view that the mother "is a good mother and does the best that she can." The Dare clinical supervisor shared that view and requested that an aide be arranged for the mother; one was ultimately provided to her for four months.

5. The family's stay in hotels. In July of 2014, Goodacre went on vacation, and the mother and children were placed in respite care. During this period, the mother decided that lingering problems with her Brockton shared living placement were sufficiently serious that she could not go back.8 This led to something of a crisis, because a suitable substitute placement could not be found. As a result, the mother and the children — together with a full-time aide — had to be housed in hotels.

During the period that the family was placed in hotels, various individuals filed a series of 51A reports alleging neglect or abuse of the children by the mother. The documentary record includes contradictory "evidence" about these allegations and suggests unanswered questions. For example, based on a bungee cord being attached to Anne's crib and some marks on Anne's ankle, a DDS clinical supervisor had concern that the mother may have tied Anne to the crib to keep her there. Given that the mother had around-the-clock supervision, the record reflects some perplexity on the part of DCF about how — if the allegations were true — someone could have missed the mother's alleged abuse or neglect. In addition, the record reflects that one of the mother's aides told DCF that she did "not have any concerns for [the] mother's parenting since she was placed at the hotel." Of potential significance, DCF's evaluation of the living situation at this time includes the following statement:

In any event, matters came to a head on September 8, 2014, when one of the mother's aides reported to DCF two significant concerns. The first was that the mother was lying in bed all day and not responding in a timely manner to prompts that she feed the children. The second was that Chad allegedly had told her that he had seen the mother masturbating (describing this in detail), and that both children were exhibiting sexualized behavior, including Chad "french kiss[ing]" his three year old sister.10 DCF conducted an emergency removal of the children the following day.

6. The placement of the children postremoval. Anne was placed at an intensive foster home operated by a foster parent, Susan Johnson (a pseudonym). Johnson had specialized training in addressing children with special needs, and she herself had a developmentally disabled child. After a two-month temporary placement elsewhere, in November of 2014, Chad also was placed with Johnson, who retained physical custody of both children until December of 2015. However, Johnson ultimately realized that she could not handle both children in her home, in great part because Chad needed full-time supervision, and the children continued to exhibit sexualized behavior with each other if left unattended. Chad therefore was sent to the Bridge Home, a facility that could provide "a higher level of care and supervision." From there, he was transferred to the Stetson School in Barre, which had a specialized residential treatment program.11 It is well established that Chad has learning disabilities and attention deficit problems (ADHD) in addition to his other special needs. He has also been diagnosed with posttraumatic stress disorder.

Meanwhile, Anne stayed at Johnson's home until June of 2016. At that time, Johnson moved to Florida for reasons not explained in the record, and Anne was placed in a series of temporary foster homes. Johnson eventually moved back to Massachusetts and sought to regain custody of Anne; however, Johnson was unable to obtain suitable housing. At the time of trial, DCF viewed Johnson as a potential adoption resource for Anne if Johnson could solve her housing needs.12 In the time between Anne's placement at Johnson's home and the trial, her sexualized behavior improved significantly to the point where she no longer presented such behavior. However, Anne did exhibit other problematic behaviors such as throwing extreme tantrums, and she has some degree of developmental disabilities.

7. The mother postremoval. After the children were removed, the mother was placed in a DDS group home in Chelsea until November of 2015. Then, she moved to a DDS group home in Dorchester. As the judge found, the group home was "more of an independent living program" that had a separate bedroom for each of the four residents, and a communal kitchen, living room, and dining room. By this time (and continuing through the trial), the mother participated in a day program that extended from 9 A.M. to 3 P.M. each weekday in Dorchester. That program "offered education, community support, fitness[,] clinical rehabilitation," and assistance in seeking employment. As of the date of trial, the mother — who once had been employed at Goodwill — had participated in one job interview, but had received no offers of outside employment. She did do custodial-type work at the day program, for which she received a small amount of income. Otherwise, her income consisted of Social Security payments based on her disability, from which amounts were withheld for her share of housing payments and for access to the Greater Boston public transportation service known as "The RIDE."

The mother also participated in other self-improvement efforts including individual therapy, which, at the urging of DCF, she began in May of 2015. She met regularly with her first therapist until that therapist went on maternity leave in October of 2015. The mother then met with a different therapist, but that therapist left the agency in January of 2016. When DCF discovered the lapse in therapy, it urged the mother to contact a certain health center to obtain a new therapist. The mother stated that she would do so, but this never came to fruition.

Again at DCF's urging, the mother enrolled in classes at a parenting program in March of 2016. That program, which encompassed forty-five hours of training, met every Thursday. The mother completed this program in June of 2016, and presented her certificate of completion to DCF prior to the termination trial. Although the record indicates that the mother agreed to, and participated in, a formal evaluation of her parenting skills, that parenting evaluation was never admitted at trial,13 and the judge therefore did not have the benefit of it.

Much of the live testimony went to the face-to-face visits, and other contact, between the mother and the children after they had been removed. The first-hand accounts of the various visits between the mother and the children generally are consistent: the children hugged and kissed the mother, interacted with her, enjoyed the gifts or appropriate snacks that she brought, and said things such as "[b]ye Mom" when they left.14 There was also uncontested testimony that the mother expressed concern for their well-being, for example, by checking for dirt under the children's fingernails, and on one occasion questioning whether Chad had warm enough clothing for a visit to a park. The third-party accounts of the visits did tend to emphasize the mother's relative immobility. For example, the DCF adoption worker who supervised a visit at a restaurant testified that the mother "remained in a seat the entire time during the visit." The judge herself repeatedly highlighted the mother's sedentary nature in her subsidiary findings, for example, by noting that during a visit with the children at the Dorchester group home, the mother spent most of the time sitting on the bed.15

A combination of factors placed some constraints on when and where face-to-face visits could take place. For example, the mother's Chelsea group home did not allow visits there. The mother frequently canceled visits, based on asserted reasons such as physical ailments (e.g., sore feet) or the lack of proper outdoor clothing. Once DCF placed Chad in the Stetson School in Barre, transportation there became an obvious problem. DCF offered to bring Chad to Worcester for visits and suggested that the mother travel from Dorchester to Worcester on public transportation, which would be a four-hour round trip.16 After Chad was moved to Barre, only one face-to-face visit occurred between him and the mother, and that was when DCF social workers drove Chad from Barre to Dorchester, and back.17 In the face of the travel difficulties, the mother sought to have remote visual contact with Chad through the Internet-based video-chat service known as Skype. DCF left it to the mother to make the necessary arrangements with the Stetson School, and this did not happen. However, the mother did maintain frequent contact with Chad by telephone, speaking with him at a regular time each Monday, "like clockwork."

8. The judge's rulings. At trial, the mother and both children supported the family being reunited. Nevertheless, the judge found the mother unfit and ruled that her parental rights should be terminated. While the judge's findings raise a number of concerns about the mother's fitness, she rested her ruling on one ground, namely, that the mother was incapable of addressing the children's special needs. Although the findings and rulings that the judge issued to explain her decision total fifty-six pages,18 her reasoning is capsulized in the following key passages:

Then, with regard to the efforts the mother made to improve her parenting abilities, the judged added the following:

Although the judge terminated the mother's parental rights, she ordered regular posttermination and postadoption visitation between Chad and Anne, and between the children and the mother. The mother and both children appealed, but eventually Anne participated as an appellee in support of the decree terminating the mother's parental rights as to her.19

Discussion. A finding of parental unfitness must be supported by "clear and convincing evidence." Adoption of Paula, 420 Mass. 716, 729 (1995). That means that "[t]he requisite proof must be strong and positive; it must be `full, clear and decisive.'" Adoption of Iris, 43 Mass.App.Ct. 95, 105 (1997), quoting Callahan v. Westinghouse Broadcasting Co., 372 Mass. 582, 584 (1977). "Parental unfitness, as developed in the case law, means more than ineptitude, handicap, character flaw, conviction of a crime, unusual life style, or inability to do as good a job as the child's foster parent" (footnotes omitted). Adoption of Katharine, 42 Mass.App.Ct. 25, 28 (1997). "[T]he issue is not `whether the parent is a good one, let alone an ideal one; rather, the inquiry is whether the parent is so bad as to place the child at serious risk of peril from abuse, neglect, or other activity harmful to the child'" (citation omitted). Adoption of Zoltan, 71 Mass.App.Ct. 185, 188 (2008).

A parent may be found unfit because of mental deficiencies, but only where it is shown that such "deficiencies impaired her ability to protect and care for the children." Adoption of Quentin, 424 Mass. 882, 888-889 (1997). "Where a parent, as here, has cognitive or other limitations that affect the receipt of services, [DCF's] duty to make reasonable efforts to preserve the natural family includes a requirement that [DCF] provide services that accommodate the special needs of a parent." Adoption of Ilona, 459 Mass. 53, 61 (2011).

Having stated these background principles of law, we turn now to examining the ground on which the judge relied, the mother's inability to address the children's special needs. Certain aspects of this issue are not subject to reasonable dispute. First, the judge's finding that both children have serious special needs (not the least of which relate to the sexualized behavior they both have exhibited) is unassailable. Second, it is indisputable that the mother lacks the capacity to address those special needs on her own, whether due to her cognitive limitations, depression, or weight-related immobility. In fact, the judge's assessment that the mother is unable even to understand what the children's special needs were is well supported by the record. Thus, the concerns that animated the termination decrees are both serious and well substantiated.

At the same time, while it is undisputed that the mother could not address the children's special needs on her own, it is also undisputed that their special needs could not be managed by other individuals either. For example, as the judge herself recognized, "[Chad] has required the highest level of intensive care in a residential therapeutic program which is particularly able to address his sexualized behavior and his trauma history." In addition, the judge's analysis does not take into account the availability of support resources to help the mother manage her life, including her role as a parent.20 The judge did not speak directly to the nuanced question that the mother's situation posed: whether, with available assistance, the mother would be able to leverage the outside support that both children plainly need. To be clear, we note that we do not presume that the answer to that question is "yes"; in the end, it may well be that the mother's demonstrated problems with completing tasks even with some assistance prove too profound. Our point is simply that before we can countenance the "extreme step" of terminating a parent's rights (citation omitted), Adoption of Ilona, 459 Mass. at 59, further proof is warranted as to how the mother's mental disability and other issues affect her ability to serve the children's best interests. See Adoption of Quentin, 424 Mass. at 888.21

In addition, we note that the judge did not squarely address the separate question whether termination of the mother's parental rights was warranted even if she was not fit to assume custody herself. See Adoption of Imelda, 72 Mass.App.Ct. 354, 360 (2008) ("Unfitness does not mandate a decree of termination"). This is not a case where the facts dictated that the relationship between the parent and children be severed; to the contrary, in recognition of the bond and positive relationship between the mother and children, the judge mandated posttermination and postadoption visitation. Nor is this a case where the children were well-situated for adoption or other stable, long-term placements. We observe, for example, that in the mere five months between when Johnson moved to Florida and the trial, Anne — the child that all parties appear to consider the easier one to place — went through approximately eight placement transitions. Again, none of this is to say that the judge could not terminate the mother's parental rights, and we recognize that such a step may be necessary as a precursor to a recruited adoption. However, the question of how termination would serve the children's best interests deserves to be addressed directly.22

A few additional observations are in order. Although the judge ultimately focused exclusively on whether the mother could meet the children's special needs, the record reveals significant additional concerns regarding her fitness. First among these is whether the mother was causing the children's sexualized behavior, not merely failing to address it adequately. Unless such allegations were substantiated at trial, however, the mother's parental rights could not be terminated based on them. See Adoption of Eden, 88 Mass.App.Ct. 293, 296 (2015) ("It is a bedrock principle that parental rights may not be terminated on the basis of an unproven allegation, even one as grave as [the sexual abuse allegation there]"). See also Custody of Eleanor, 414 Mass. 795, 800-801 (1993) (where allegation of sexual assault was later withdrawn, that allegation "in the absence of any corroboration or physical evidence of sexual abuse ... cannot be said [to establish] parental unfitness ... by clear and convincing evidence").

On this issue, the largely documentary record included signposts that pointed in differing directions, and the strongest evidence that the mother was somehow the source of the problem was embedded hearsay (the statements attributed to Chad). DCF did not press the judge to resolve the question of the cause of the children's sexualized behavior, and the judge neither did so nor improperly rested on unproved allegations. We make these observations not to fault the judge — who we acknowledge was hamstrung by the far from perfect evidentiary record put before her — but to highlight the anomaly that the most serious concerns about the mother's fitness lay unresolved.23 We offer these observations as potential assistance to the judge and parties as they formulate the proceedings on remand.

The concerns just voiced apply as well to a second serious concern about the mother's fitness on which the judge did not rely: whether the mother is able, even with some assistance, to meet the children's basic needs, not just their special needs. Again, there were conflicting indications about this in the largely documentary record. On one hand, that record appears to indicate that DCF largely was satisfied that the mother historically was able to meet the children's basic needs so long as she had sufficient prompts from an aide supplied to her by DDS or otherwise. On the other hand, the extent to which appropriate supports were available to the mother prospectively was never directly explored at trial; indeed, no one from DDS even testified. As a result, the fundamental question whether the mother was in a position to meet the children's basic needs was never resolved.

Our concern about the state of the record and the limited nature of the judge's rulings are amplified by arguments the mother has raised about the motives of DDS and its contractor, Dare. Although no one from those entities testified, their observations or views were reflected in the reports that were included in the documentary record. Much of that evidence was positive toward the mother, but it became increasingly negative after the mother refused to return to the Brockton shared living placement and therefore had to be housed in hotels. The mother has raised nontrivial arguments that the views or observations of those at DDS or Dare at that point may have been colored by institutional bias. Those entities faced an incentive to justify DCF's removal of the children, the argument goes, because the presence of the children was the reason for the expensive hotel placements and prevented DDS and Dare from finding a suitable alternative placement for the mother (DDS's only client in the family). DCF properly notes that courts are to apply a presumption that public officials have acted in an honest and impartial manner. See, e.g., Konover Mgt. Corp. v. Planning Bd. of Auburn, 32 Mass.App.Ct. 319, 326 (1992). However, as noted above, one of DCF's own reports reflects the view that DDS and Dare sought to have the children removed for these very reasons. Moreover, the mother's argument is not so much that DDS and Dare personnel were acting in bad faith, but rather that their narrow institutional mission created incentives regarding how they observed and portrayed the interests of the children. The absence of live witnesses directly speaking to the mother's parental deficiencies deprived the judge of the ability to assess the extent to which the concerns expressed by DDS and Dare were supported by objective fact.

While the judge's subsidiary findings regarding the mother's compliance with her service plans are not inaccurate, they nevertheless inaccurately leave the impression of wholesale noncompliance. In fact, the mother made significant efforts to improve her skills despite her disabilities and mobility challenges. The most prominent example is her completion of the forty-five hour parenting course, which required her to travel to regularly scheduled weekly classes.24 Of course, whether completion of that course actually improved her parenting skills is a different matter.25 To be sure, there were service plan tasks that the mother did not complete. However, such noncompliance must be viewed in light of the limited efforts that DCF and DDS made to assist the mother in overcoming her demonstrated problems in completing tasks on her own once the children had been removed. The record contains several examples of unexplained failures by the assigned officials to provide support to help the mother succeed in keeping the family together.

We additionally note some concern about the degree of emphasis that the author of the reports, DCF witnesses at trial, and the judge herself appear to have placed on the mother's weight-related mobility issues. Indeed, the reference to these issues appear to outnumber references to concerns about the mother's mental disabilities. Yet mostly absent from the judge's findings and rulings is an analysis of how those mobility issues in fact help render the mother unfit. Certainly, the judge did observe that, in light of the mobility issues, "the services which are necessary to meet the children's specialized needs must be convenient for the mother in order for them to be utilized." But especially with the mother living in Boston and having access to The RIDE, it is not clear whether such theoretical constraints will matter in practice. Of course, it is indisputable that Chad did for a period require services located outside of Greater Boston, but the fact remains that he was able to take advantage of those services without the mother's rights being terminated.

In sum, although we agree with the judge that the record raises serious concerns about the mother's capacity to address the children's needs, we conclude that, at a minimum, further exploration and explication is necessary before the mother's parental rights may be terminated.26 We therefore vacate the decrees terminating the mother's parental rights and remand this case to the Juvenile Court for further proceedings consistent with this opinion. We affirm the decrees terminating the parental rights of the children's fathers.

So ordered.