BRENNEN

Industries / Government, education and community / Social services / Foster care and adoption

Approving foster carers

This is the work of finding people willing to foster, checking their background and home, training them, approving them and keeping them fostering.

7

Milestones achieved, from 1945 to 2026.

2

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1945

    A death exposes thin checks

    In January 1945 Dennis O'Neill died at Bank Farm, Minsterley, in Shropshire, England, killed by the foster father with whom he and his brother had been boarded out. Sir Walter Monckton's public inquiry, presented to Parliament in May 1945, found that the foster mother had been approved after one visit by a Poor Law officer, with no enquiries made, and that the boys had one official visit and were never seen by a medical officer, against the 1933 boarding-out rules. New boarding-out rules were issued in 1946, and the Curtis Committee was set up to review the care of all children in public care.

    LessonApproval rules protect a child only when someone checks the home, makes enquiries about the family and keeps visiting.

    api.parliament.ukthetcj.orgsw100.ed.ac.uk

    Credit

    Sir Walter Monckton (report, Cmd. 6636)

  2. 2006

    Fingerprints before approval

    The US Adam Walsh Child Protection and Safety Act, signed on 27 July 2006, required states to run fingerprint-based checks of national crime databases on every prospective foster or adoptive parent before approval, and to check the child abuse registries of any other state where a prospective parent or another adult in the home had lived in the previous 5 years. It also ended states' option to opt out of these checks. By September 2018 the laws of every state, the District of Columbia, Guam and Puerto Rico required background checks before a child is placed, by Child Welfare Information Gateway's review.

    govinfo.govartifacts.childwelfare.gov

    Credit

    US Congress (Public Law 109-248, section 152)

  3. 2008

    A standard home study tested

    A home study is the written assessment of a family who wants to foster or adopt. The Structured Analysis Family Evaluation (SAFE), from the nonprofit Consortium for Children, gives every worker the same questionnaires and rating tools. In 2008 researchers reported what they called the first empirical study of SAFE: 209 staff in Colorado, New Jersey, Nevada and Utah rated it more comprehensive, fairer and more accurate than their old methods. SAFE then ran in 8 US states; Wisconsin now requires it for every foster home license, and the consortium says it is used across the United States and Canada.

    LessonThe study measured what staff preferred, not whether children were safer in the homes it approved.

    sswr.confex.comfcc.wcwpds.wisc.edusafehomestudy.org

    Credit

    Consortium for Children; study by Thomas Crea, Laura Chintapalli, Rachel Buchanan and Richard Barth

  4. 2020

    Hub homes keep carers

    The Mockingbird model, created by the Mockingbird Society in the United States and run in the UK by the Fostering Network under licence since 2015, groups 6 to 10 foster families around one experienced hub carer who offers sleepovers, advice and peer support. An independent evaluation for England's Department for Education, led by the Rees Centre at the University of Oxford, found in 2016 to 2019 data from 11 sites that participating carers were less likely to stop fostering. It found no improvement in placement stability in the same data, and put the return at 99 pence saved for each £1 spent. By March 2020 the 12 sites studied had 41 such groups with 320 member homes.

    LessonSupport from other carers kept carers fostering, but did not by itself show fewer moves for children.

    assets.publishing.service.gov.uk

    Credit

    The Mockingbird Society; The Fostering Network; evaluation by Ellie Ott and colleagues, Rees Centre

  5. 2023

    Kin get their own standards

    Since 2000, US federal rules had required states to apply the same licensing standards to relatives as to unrelated foster families. A final rule from the Administration for Children and Families, published on 28 September 2023 and in force from 27 November 2023, lets states and tribes use separate standards for relative or kin homes and requires them to pay licensed kin the same foster care payment as other foster homes. By the rule's own figures, 27.6% of relative foster home placements were claimed for federal Title IV-E funds in fiscal 2019, against 47.7% of other placements. Generations United reports that more than 20 states and tribes were approved for kin-specific licensing within two years, and that 39% of US children in foster care, about 127,449, lived with relatives in 2024, an all-time high.

    federalregister.govgrandfamilies.org

    Credit

    Administration for Children and Families; tracking by Generations United

  6. 2024

    Registered carers wait unused

    A June 2024 review of 29 child guidance centers by Japan's Ministry of Internal Affairs and Communications found many registered foster carers with no child, because children's needs did not match what carers had said they could take, and because carers who both worked could not get nursery places. At the end of fiscal 2021, 10,763 of 15,607 registered foster households, 69.0%, had no child placed; among carers registered to adopt, 5.0% had a child. The ministry recommended more short-term placements and priority nursery admission.

    LessonRecruiting more carers does not place more children unless the carers fit the children who need homes and get the support to take them.

    soumu.go.jpisvd.or.jp

    Credit

    Administrative Evaluation Bureau, Ministry of Internal Affairs and Communications, Japan

  7. 2026

    England's carer count steadies

    Ofsted counted 33,450 mainstream (not kin) fostering households in England on 31 March 2026, up from 33,435 a year earlier, after four years of decline. Approved mainstream carers numbered 56,145, against 63,890 in 2021. Of 5,345 applications completed in 2025–26, 2,080 were approved, 3,195 were withdrawn and 75 were rejected. About a third of carers are 60 or older, and that group made up 35% of carers who left.

    LessonMost applications that end without approval are withdrawn, not refused, so keeping applicants matters as much as finding them.

    gov.ukgov.uk

    Credit

    Ofsted, Fostering in England statistics, published 18 November 2026

Watch for

  1. A national homes-per-child goal

    The US Administration for Children and Families (ACF) says there are 57 licensed foster homes for every 100 children in foster care. Under its "A Home for Every Child" initiative, which all 50 states, the District of Columbia and Puerto Rico had joined, it plans a public scorecard of each state's ratio, updated monthly, and USD 7 million in prizes, from October 2026 to September 2027, for the states with the highest and most improved ratios.

    Achieved whenthe ratios, checked against AFCARS (the federal foster care data system) or by an outside body, show more licensed homes per child in care at the end of the challenge than at the start.

    acf.gov

    Claimed by

    ACF, July 2026

  2. Most kin carers licensed

    Maryland's Department of Human Services says that after it introduced kin-specific licensing, the share of kin caregivers who were licensed rose from 25% in December 2024 to 86% in December 2025. Unlicensed kin often cannot get the foster care payments and services that licensed carers receive.

    Achieved whenfederal data or an independent evaluation confirms the licensed share and shows whether kin placements lasted longer.

    dhs.maryland.govbinti.com

    Claimed by

    Maryland Department of Human Services, February 2026

Open

  1. 01

    Licensing rules as open data

    A machine-readable table of what a foster home must meet in each US state, the District of Columbia and the four UK nations: minimum age, training hours, bedroom and space rules, income, health checks, background checks, renewal period and any separate rules for kin. Each cell cites the regulation, and each US state is compared with the Model Family Foster Home Licensing Standards.

    The brief
    Credit · unclaimed
  2. 02

    Where applicants give up

    A public table, for each English council, of fostering applications, approvals, withdrawals by applicants and by services, and the share of approvals reached within 180 and 240 days of starting assessment, for every year Ofsted has published, with each council's regional recruitment hub marked. Then compare how these figures changed in councils that joined a hub with how they changed in councils that did not.

    The brief
    Credit · unclaimed

Open 01 · the brief

Licensing rules as open data

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A machine-readable table of what a foster home must meet in each US state, the District of Columbia and the four UK nations: minimum age, training hours, bedroom and space rules, income, health checks, background checks, renewal period and any separate rules for kin. Each cell cites the regulation, and each US state is compared with the Model Family Foster Home Licensing Standards.

Why it is still open

The model standards, written by the ABA Center on Children and the Law, the Annie E. Casey Foundation, Generations United and the National Association for Regulatory Administration, come with a crosswalk tool that each state fills in for itself. Child Welfare Information Gateway's state review covers background checks only and is current through September 2018. Rules are changing fast: more than 20 states and tribes were approved for kin-specific licensing in two years, and Colorado's 2024 kinship law cut training for kin carers from 27 hours to 6, by Binti's account. A search found no cross-state table of the full requirements.

gu.orgartifacts.childwelfare.govgrandfamilies.orgbinti.com

Built with what exists

Public administrative codes and legislation.gov.uk; a model that pulls rules out of long documents, such as Meta's Muse Ember (see the Note "The jobs agents repeat"); Python; a public Git repository.

  1. First step

    Fix the field list from the model standards and fill it for three states.

  2. Done looks like

    All 50 states, the District of Columbia and the four UK nations, at least 30 fields each, every cell cited, with a 10% sample checked by a second person and the error rate published.

Open 02 · the brief

Where applicants give up

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public table, for each English council, of fostering applications, approvals, withdrawals by applicants and by services, and the share of approvals reached within 180 and 240 days of starting assessment, for every year Ofsted has published, with each council's regional recruitment hub marked. Then compare how these figures changed in councils that joined a hub with how they changed in councils that did not.

Why it is still open

Ofsted's 2025–26 figures show 3,195 completed applications withdrawn and 2,080 approved, and its data file gives the counts council by council. Research by Ipsos for the Department for Education, published in July 2023, was qualitative, based on 10 case studies; it found that services count enquiries in different ways, so enquiry figures cannot be fairly compared, and it called for clearer measures. A search found no published council-by-council comparison of withdrawals and approval times.

gov.ukassets.publishing.service.gov.ukassets.publishing.service.gov.uk

Built with what exists

Ofsted's "Fostering in England" underlying data spreadsheets for 2021 to 2026, which hold counts per council, not records of people; Python with pandas; council and hub web pages for hub membership; a public dashboard.

  1. First step

    Load the council table from the 2026 file and compute withdrawal and approval-time rates for all councils.

  2. Done looks like

    Every council for five years, with withdrawal rates and approval times, and a published estimate, with its margin of error, of how much hub membership changed them.