Reading Case Files Like a Professional
I spent about eight years doing supervision reviews for local authority children's services before moving into independent consulting. What I am about to describe is not hypothetical. I have sat in rooms where managers signed off on assessments that were clearly wrong, and I have written my own share of reports that would not survive scrutiny today. This is the thing nobody likes to talk about openly. The most basic failure I encountered repeatedly was assessment drift. A social worker starts with a focused plan to investigate a specific safeguarding concern, and somewhere around week three the original purpose gets buried under a stack of paperwork, home visits, and meetings that have stopped connecting back to the initial threshold criteria. I worked on a case once where a child had been referred for possible neglect related to educational attendance. Six months later, the assessors were still visiting weekly but had no clear written rationale for those visits. The father was not neglectful. The mother was hiding a psychiatric episode. The original reason for the referral had evaporated from the file entirely, and because nobody had challenged the drift, the child ended up on a child protection plan for the wrong reason. That was 2019. It happened on my watch. Poor practice typically clusters around recording, timing, and reflection rather than any single dramatic error. The failures are rarely catastrophic in real time. They look like acceptable work until you look at them sideways. A common pattern I see is when a practitioner completes a significant risk assessment but files it without cross-referencing the conclusions in the rest of the chronology. Another case involved a worker who identified a parent's substance misuse in January, documented the referral to specialist drug services, and then failed to record that the referral was declined or that the parent withdrew consent. Two years later, during an inspection, that gap looked like either ignorance or deliberate concealment. Both are defensible in court. Neither reflects careful practice.
I have noticed that practitioners who produce thorough documentation often still fail on what I call contemporaneous reasoning. The dates line up, the signatures are there, the timelines are chronological, but the thread of why decisions were made at each point disappears between the entries. A reader cannot reconstruct the decision-making process. That is a structural failure, not a clerical one, and it is far more common than people admit.
How I Approach Reviewing Poor Practice
When I am brought in to review a case file, I do not start by reading chronologically. I read the chronology backward, starting from the most recent decision and working toward the original referral. This method exposes gaps immediately because you can see whether each earlier entry actually supports the later action. If the final plan was placed on a child protection register in March, the January assessment should already contain the threshold criteria language that justifies that outcome. When it does not, you have your problem. I then check three specific things in order of importance: whether the child's voice is reflected with sufficient directness, whether alternative explanations for the presenting concerns were actively considered and ruled out, and whether the recording format matches the gravity of what was recorded. The third point sounds trivial but it is not. I have seen detailed forensic evaluations of domestic abuse recorded in bullet points that the same worker would never have used for a routine well-child visit. The inconsistency itself becomes evidence of poor practice when it surfaces during a serious case review. The workaround I use when I encounter a file with chronic reasoning gaps is to map it visually. I draw the timeline on paper and connect each decision node to its supporting evidence with lines. Where the lines are sparse or missing, the file is weak. This takes about forty-five minutes for a typical complex file and usually reveals three or four substantive gaps that a linear read would miss entirely.
Get the Full Details
Specific Failure Modes That Regularly Surface
Unverified assumptions recorded as fact is probably the single most damaging pattern I encounter. A worker hears a neighbour say the household is chaotic and records it as known fact rather than as an allegation needing verification. When a subsequent review challenges the entry, the worker has no way to demonstrate they tested the claim. The file then looks competent on the surface but collapses under any real scrutiny. Another pattern involves over-reliance on single-source information. I reviewed a file recently where every key finding came from one school referral. The social worker had never spoken to the GP, never contacted the youth offending team, and never conducted an unannounced home observation because they felt the school referral was sufficiently corroborating. It was not. The child was being sexually exploited by a peer group the school had not detected. The case is now a national learning publication, and the worker in question passed their fitness to practise hearing but with conditions attached. That was the right outcome for them personally but it does not excuse the structural failure that produced it. Delayed recording is frequently dismissed as a minor administrative issue. It is not. When a decision is made in an emergency strategy meeting and the full record appears three weeks later, the delay itself becomes part of the legal story. Courts have ruled that late entries undermine the credibility of the entire file even when the content is accurate. I have advised multiple agencies to adopt same-day recording for any intervention involving a child under Section 47 enquiries. It sounds extreme until you see what happens in cross-examination.
What Actually Works Instead
There is no published manual that replaces competent supervision. I have seen agencies invest heavily in template-based recording systems and training modules while leaving supervision sessions under-resourced and often cancelled. Templates standardise format but they do not challenge thinking. A worker using a perfect template can still produce a file that fails on substance if their analysis is shallow or incomplete. Supervision that actually interrogates the worker's reasoning is what prevents drift and unverified assumptions from entering the record. The practice model I recommend to any team is the decision audit trail. For every recorded decision, the worker answers three questions in the file: what was the evidence at that moment, what alternative explanations were considered and rejected, and what would have changed the decision. This takes approximately two extra minutes per entry and eliminates roughly seventy percent of the gaps that appear during serious case reviews. It is not elegant but it is effective. I should be honest about where this approach breaks down. It assumes the worker has the analytical capacity to identify genuine alternative explanations, not just ones they list perfunctorily. It also depends on supervisors who will push back on weak answers rather than accepting them to keep the file moving. In high-volume teams with constant staffing turnover, both conditions are rarely met simultaneously. Under those circumstances, even the best template system will not prevent poor practice from accumulating.
The bottom line is that poor practice in social work is rarely the result of malice or incompetence in the way people imagine. It is the product of small omissions that compound over time, of files that drift from their original purpose, and of systems that reward completion over quality. I have spent more of my career cleaning up those compounding errors than I have on anything that feels like heroic intervention.
