Leading At The End Of Evidence
Wonk PremiumBrief:
Leading At The End Of Evidence
Leadership Practices For Moving Through Uncertain Terrain.
By: Michael Cull, PhD, Msn
Part I of this series established a structural problem.
Child welfare's accountability structures are built around an assumption the evidence cannot support: that when the right interventions are applied correctly, safety can be reliably ensured.
The evidence says otherwise.
The treatment science is modest. The tools are contested. The workforce is depleted.
The ceiling on what child welfare can reliably deliver is lower than the systems governing it acknowledge, and it is getting lower.
The gap between what child welfare is expected to deliver and what the evidence can support is structural. It’s not a product of insufficient effort or wrong protocols.
That diagnosis resolves not into a policy fix, but an invitation into a different kind of leadership.
The strategic response isn't more rigorous application of the same frameworks, but a posture built for navigating uncertainty.
High-reliability organizations offer a starting point.
Aviation, nuclear power, and health care all manage the consequences of rare but catastrophic failures under conditions of irreducible complexity.
None of them eliminated harm by finding better protocols.
They reduced harm by building cultures capable of learning faster than they failed.
That distinction — between protocol adherence and organizational learning — is the source of questions that invite a different approach to leadership.
BUILDING A CULTURE WHERE ERRORS SURFACE
When the implicit expectation is that skilled professionals don’t make mistakes, the result is not fewer mistakes, it is better-concealed ones.
Workers don't flag near-misses. Supervisors don't surface emerging concerns. Agencies manage appearances rather than conditions.
Safety science distinguishes between cultures that respond to adverse events with punitive accountability and those that balance individual responsibility with honest inquiry into the conditions that shaped a decision.
Moving toward what safety scientists call “safety culture” is not a lowering of the bar. It is the precondition for the kind of candor that actually produces safer outcomes and a higher bar.
Consider a caseworker who visits a family and leaves with an uneasy feeling. She can’t fully articulate it to the level of documentation or escalation, but it’s off.
In a punitive culture, she files the visit and moves on.
In a safety culture, she has a supervisor she can call, a structure for flagging ambiguous concerns, and confidence that surfacing uncertainty won't be read as incompetence.
The first culture produces a clean record.
The second produces a safe one, with a chance to intervene.
TREATING ADVERSE EVENTS AS INFORMATION, NOT ONLY AS FAILURES.
Agencies that respond to child fatalities and critical incidents exclusively through the lens of individual accountability forgo the most valuable data available to them.
Structured, non-punitive critical incident review processes are how aviation and healthcare convert their hardest moments into institutional knowledge.
The question is not only who made an error, it is what organizational conditions made that error more likely, and what changes would reduce the probability of recurrence.
Child welfare has the same option. Most agencies aren't using it.
A child fatality review that asks only whether the assigned worker followed protocol will reliably find someone who didn't, and stop there, without asking why it was possible.
A review that also asks what caseload that worker was carrying, what supervision she had access to, and whether the risk assessment tool flagged anything actionable produces a different set of answers.
Those answers are the ones that change system behavior.
MAKING LEARNING AN ORGANIZATION
In organizations that sustain safety over time, teams spend time identifying what could go wrong before it does, create regular space to reflect on mistakes, and develop the communication norms that allow concerns to surface early.
These are not training programs.
They are enduring organizational habits, and they require sustained investment and visible modeling from leadership.They don't emerge from policy directives. They have to be built.
A supervisor who opens every weekly team meeting about what almost went wrong this week is doing something structurally different from one who opens with case status updates.
Over time, the first team develops a shared vocabulary for uncertainty and a habit of surfacing risk before it compounds. The second team learns what not to say out loud.
NAMING THE EVIDENTIARY LIMITS PUBLICLY
Leaders who understand the evidentiary limits of their field have an obligation to name them, to policymakers, to the public, and to the systems that set expectations for their work.
Child welfare cannot be held as the sole accountable party for outcomes that depend on housing policy, health care access, economic security, and community support.
Naming that reality is not an excuse. It is a precondition for building the cross-system accountability structures that child safety actually requires.
The lesson from substance use disorder treatment is not that child welfare should expect less of itself.
It’s that leaders must lead differently, aligning expectations, decisions, and accountability with the realities of the work.
Child safety is non-negotiable.
But achieving it requires the humility to acknowledge what we do not yet know and the discipline to build systems capable of learning their way toward safer outcomes over time.
A state child welfare director testifying before a legislative committee about a high-profile fatality faces a choice.
She can accept the frame — something went wrong, someone failed, here is what we are fixing — or she can expand it.
She can describe what the family's housing situation was, what treatment was available and when, what the caseworker's caseload looked like, and what other systems did or didn't do.
The first answer satisfies the immediate political pressure. The second one is the only answer that eventually produces a different result.
LEADING DIFFERENTLY
The gap between what child welfare is expected to deliver and what the evidence can support is structural.
It will not close through better compliance or more rigorous protocol application.
It closes (partially, incrementally) when leaders build organizations capable of learning faster than they fail, and when they tell the truth about the conditions under which they are operating.
QUESTIONS TO CULTIVATE A DIFFERENT KIND OF LEADERSHIP
Leaders across child and family policy are navigating significant uncertainty about what works. Here are questions to guide you through that terrain.
Building a Culture Where Errors Surface
What's your current posture toward errors and risk as a leader and organization, And where can you model clinical curiosity when something doesn't work?
Treating Adverse Events as Information, Not Only as Failures
Where are you digging down to the root? What's the root cause of adverse events, and how can you systematize that approach to make it routine?
Making Learning an Organization
How do you model ongoing learning, and where can you engrain it as a shared habit within your team?
Naming the Evidentiary Limits Publicly
What limit or constraint is unspoken yet central to the impact your organization can make, and who doesn’t have visibility into it but should?
