What CQC inspectors actually look for

For most small independent providers, the inspection is unannounced. The inspector arrives on the day, identifies themselves at reception, and the visit begins. There is no notice period and no advance information request to prepare for.

Announced inspections still happen, but they tend to be reserved for larger or multi-service providers where the inspection covers several core services or domains and needs to be scheduled. If you are a single-site independent provider, assume unannounced as the default.

Whether announced or unannounced, the inspector starts work before they get to your door by reading everything CQC already has on you. That includes previous inspection reports, any provider information you have submitted, any notifications you have filed in the last twelve months, any concerns received from members of the public, any safeguarding or police involvement that has been notified to CQC, and the contents of your provider page on the CQC website.

Most of an inspector's view of your service is formed before they meet you. If your previous report flagged a should-do item, they will arrive expecting to see the action taken. If you filed three serious incident notifications in the last six months, they will arrive with specific questions about each.

For an announced inspection, an additional information request usually arrives ahead of the visit. It typically asks for the training matrix, the policies relevant to the focus areas, recent governance meeting minutes, complaint logs, safeguarding case logs, and a list of staff with roles. The deadline is usually short. The format the inspector wants matters less than the substance being legible and current. For an unannounced inspection, the equivalent request will usually arrive after the visit rather than before; the inspector picks up what they need during the day and follows up in writing afterwards.

Either way, the worst use of any preparation time is rewriting policies or backfilling minutes. The best use is making sure that what you would hand over already says what it claims to say: that the training matrix matches the staff actually on duty, that the incident logs cite the right notification numbers, that the actions in your last governance minutes have visible follow-through.

Inspectors are professionals, and the first half-hour is a calibration period. They are forming an initial read on three things.

Atmosphere. The waiting area, the corridor, the reception interaction. This sounds like soft data; it is hard data for inspectors. A chaotic reception, staff who do not know what is going on, patients waiting without explanation, these read as Well-led concerns immediately, before any record is opened. A calm, professional first impression buys you genuine credit.

Whether the records match the reality. The inspector has read your records and arrived with a mental picture of your service. The first interactions test whether that picture is accurate. A receptionist who does not know who the registered manager is, when the manager is clearly listed in your records, is a signal. A clinical area that looks different from how it was described in the pre-inspection materials, the same. The inspector is checking that what you claimed about yourself is true.

The registered manager's presence. Not whether the manager is good (the inspector cannot assess that in thirty minutes); whether the manager is present, knows what is going on, and can answer specific questions about specific records. A registered manager who has to look up basic facts about their own service is a Well-led concern. A registered manager who can speak fluently to the records the inspector has read about is not.

After calibration, the inspector works through a structured plan covering the five key questions: safe, effective, caring, responsive, well-led. CQC's methodology evolves over time, but the substance of what an inspector does is roughly the same as it always was: read the evidence across multiple categories, weigh what corroborates and what conflicts, and reach a judgement.

Records review. The inspector reads actual records (incidents, complaints, safeguarding, governance meetings, training matrix, staff supervision, patient records sampled at random) and tests whether what is recorded matches what is being said in conversation. Inconsistencies between records and conversation are the single most consequential signal in any inspection. Records that match what the team describes is the strongest possible signal in your favour.

Interviews. The inspector talks to staff, the registered manager, the nominated individual, and patients where appropriate. The interviews are not interrogations; they are structured conversations to triangulate what the records say. Staff who do not feel comfortable speaking honestly are a signal. Staff who all give exactly the same scripted answer are a different but equally concerning signal.

Observation. The inspector walks the service, watches care being delivered, observes how the team interacts with each other, watches how decisions get made on the day. This category cannot be evidenced in advance; the only preparation is to keep the actual service consistent with the recorded service.

This is the question I get asked most often. The honest answer is that the difference is rarely about the substance of the care. The substance, in most small independent services, is fine. The difference is about evidence trail integrity.

A Good service has records that:

• Show actions following from decisions, with closure dates
• Show notifications filed promptly and matching the underlying incidents
• Show the training matrix matching the staff actually working
• Show governance meetings producing actions that get done
• Show response to previous inspection findings with the work visible

A Requires Improvement service usually has the same underlying clinical work but records that:

• List actions that have been open for months with no progress
• Have notifications that do not quite match the incident records
• Have a training matrix that is six months stale
• Have governance meetings with thin minutes and no action follow-through
• Cannot trace what was done about the previous inspection's findings

The work between Requires Improvement and Good, in most cases, is not changing what the service does. It is making the trail of what the service does legible.

Three things that, after years inside the regulator, I would say the framework under-weighs. They are useful to know about because they tell you what to do beyond what an inspection asks for.

The genuine quality of team relationships. Inspectors can read records and interview staff, but they cannot in two days assess whether the team genuinely respects each other. Strong team relationships are the underlying cause of most well-led services I saw; the framework cannot really measure them.

Long-term outcome trends. The framework can look at outcomes in the last twelve months but cannot easily assess whether a service is getting steadily better over five years. Small independent services often have this signal, they are quietly improving year over year, and the inspection cannot really see it.

The hidden work of running a small service. The registered manager who covers staff sickness, rewrites the rota at 6am, takes the safeguarding referral call on a Sunday, drives a patient's prescription across town because the courier failed, none of this shows up in the inspection. Inspectors know it happens; they cannot inspect it.

The implication is that an inspection is a partial picture. A Good rating tells you the records support the service. It does not tell you the service is genuinely well-run. A Requires Improvement rating tells you the records do not support the service. It does not tell you the service is bad.

The inspector leaves the site, writes the report, and sends a draft. You have a factual-accuracy review window, usually around ten working days, to challenge specific factual errors. The review is for facts, not for the inspector's judgment on those facts. This distinction matters: trying to argue the rating is a category error and rarely productive; pointing out that a specific finding cites the wrong date, the wrong staff member, or a record that says something different from what the report claims, is exactly the kind of correction the process is designed for.

Once the final report publishes, it goes on your CQC provider page and shapes the next inspection cycle. The inspector's findings become the prior-inspection findings the next inspector reads first. Closing out those findings cleanly, with visible evidence, is the cheapest possible piece of work for your next inspection rating.

Three things, in order.

Stop preparing in the last week. Whatever you would do in the week before an inspection should already have been done. The work of inspection preparation is the work of running a service well, distributed across the months between inspections.

Read your last inspection report end to end, with a pen. For each finding, write what you did about it, when, with what evidence. If you can do this for every finding, you are ready. If you cannot, that is your work.

Make the records match the reality, both directions. If the records claim something the service does not do, fix the records. If the service does something the records do not show, fix the records the other way. The two should be the same document by the time the inspector arrives.