The Upper Tribunal has ruled that the Home Office unlawfully operated a secret, unpublished practice of limiting many recognised victims of trafficking to 12 months’ permission to stay in the UK while they recovered from exploitation. The judgment also explains why the now-published version of that practice survived challenge. The case is R (TK) v Secretary of State for the Home Department, JR-2025-LON-001083.
The tribunal found that the practice was inconsistent with the Home Office’s published policy, unlawfully fettered discretion and was irrational. The tribunal did not, however, strike down the Home Office’s current approach. Since December 2025, the practice has been published and presented as a general starting point rather than an inflexible rule. That distinction lies at the heart of the judgment.
Background
TK is a recognised victim of trafficking who suffered severe physical and psychological abuse. Following a positive conclusive grounds decision, he applied for temporary permission to stay on the basis that continued treatment was necessary for his recovery.
His medical evidence confirmed long-term treatment needs but could not identify when treatment would end. The Home Office granted 12 months’ leave and maintained that decision on reconsideration.
During the judicial review proceedings, the Home Office disclosed that, for around nine years, officials had followed an unwritten practice of granting 12 months’ leave where treatment was open-ended. It accepted that the existence of the practice should have been disclosed earlier.
Having operated for years as an unwritten verbal instruction, the practice was finally published in the current circumstances questionnaire issued to applicants on 17 December 2025.
The legal framework
Section 65 of the Nationality and Borders Act 2022 requires the Home Secretary to grant leave where this is necessary to assist a confirmed victim’s recovery from physical or psychological harm caused by trafficking.
Appendix Temporary Permission to Stay for Victims of Human Trafficking or Slavery allows grants of up to 30 months. Home Office guidance adopts a victim-centred approach and requires decision-makers to consider matters such as prognosis, the stage of treatment and its expected duration. The published policy states that ‘if treatment does not have an end date, then the decision on the length of the VTS to grant must be based on all available information’.
The unpublished practice was unlawful
TK argued that applicants could not make meaningful representations unless they knew that the absence of a treatment end date would normally result in only 12 months’ leave. The Home Office maintained that publication was unnecessary because its existing guidance already explained the relevant factors and allowed grants of up to 30 months.
The tribunal disagreed. Applying the principles in R (Lumba) v Secretary of State for the Home Department [2011] UKSC 12 it held that fairness required the practice to be published.
Knowing that leave may be granted for up to 30 months is materially different from knowing that, in practice, the Home Office would default to 12 months where clinicians could not specify an end date. Without that knowledge, applicants could not explain why a longer grant was justified or obtain evidence directed at the real issue.
The tribunal noted that TK’s own representations changed significantly once the practice was disclosed, illustrating why publication mattered.
The Home Office unlawfully fettered discretion
The tribunal also found that the unpublished practice conflicted with the Home Office’s published policy. The policy required decisions to be based on all available information where treatment had no end date. In practice, however, as the Home Office’s evidence showed, officials treated 12 months as ‘the only option’ if clinicians could not provide a timeframe.
That approach unlawfully narrowed the discretion conferred by the policy.
A clinician may be unable to predict precisely when treatment will finish while still providing valuable evidence about its likely duration, the stages of recovery or the consequences of interrupting treatment. Applicants may also provide evidence about how repeated immigration applications could hinder recovery.
The tribunal held that decision-makers were required to consider that broader evidence rather than simply asking whether a medical end date had been identified.
For the same reasons, it concluded that the unpublished practice was irrational.
Why the current practice survived
The tribunal reached a different conclusion about the practice after it was published in December 2025. The current circumstances questionnaire informed applicants that, where medical evidence does not identify a treatment end date, 12 months will ‘generally’ be regarded as appropriate. Applicants receive that information before a decision is made and that gives them an opportunity to submit further evidence and argue for a longer grant, satisfying the requirement that they be able to make meaningful representations. That satisfied the publication requirement identified in Lumba.
The tribunal held that no further detail needed to be published because the wider statutory framework, immigration rules and Home Office guidance already explain the factors relevant to the decision; the questionnaire fills the only remaining gap by explaining what will normally happen where treatment has no fixed end date.
Although the tribunal questioned why this guidance appeared only in the questionnaire rather than the main policy document, it did not regard that as unlawful.
Other grounds
The tribunal rejected the argument that a 12-month grant breached Article 14 of the Council of Europe Convention on Action against Trafficking in Human Beings. The Convention requires residence permits to be renewable where necessary but does not prescribe their duration.
The tribunal also found that the Home Office had failed to demonstrate that it had given due regard to the public sector equality duty when adopting the unpublished practice. Although the practice was capable of producing differential treatment between applicants who could, and those who could not, provide a treatment end date, the judge considered that insufficient to justify a separate declaration, particularly as relief had already been granted on the public law grounds and the practice had since been published.
Implications of the judgment
The judgment does not establish that recognised victims of trafficking are entitled to 30 months’ leave, nor does it make every 12-month grant unlawful. Instead, it reinforces two important public law principles.
First, practices affecting people’s rights cannot remain hidden where knowledge of those practices is necessary to make meaningful representations. Transparency and fairness (principles derived from Lumba) are essential parts of decision-making.
Second, decision-makers cannot replace an individual assessment with a rigid rule. The Home Office’s own published policy required consideration of all available information where treatment had no fixed end date. An unwritten rule treating 12 months as automatic was incompatible with that obligation.
For practitioners, the decision underlines the importance of addressing the appropriate duration of leave directly. Where clinicians cannot provide a treatment end date, evidence should explain the likely course of recovery, why treatment is expected to continue and why a longer grant is necessary. Representatives should also explain why the general 12-month starting point should not apply in the individual case.
The judgment may also prompt challenges by applicants whose cases were decided before the practice was published. Whether any individual decision should be revisited will depend on its own facts.
Going forward, the key question will be how the published practice operates in reality. If officials genuinely treat 12 months as a starting point while considering all relevant evidence, the policy is likely to withstand challenge. If, however, 12 months continues to be applied automatically, similar public law challenges are likely to arise again.
