Before booking again, read every station against the three scored domains and check the limits that apply to you. The GMC requires PLAB 2 to be passed within 2 years of the PLAB 1 pass, normally allows 4 attempts, permits a fifth only by application, and ends eligibility after an unsuccessful fifth attempt. The GMC's 2025 table records a 59.8% pass rate, which means about 4 in 10 people sitting that year did not pass. Your own result is the starting point for deciding what to change.
Start with the result letter
Start with the result letter, not your memory of the day. Memory gives too much weight to the station that felt humiliating and often misses smaller losses elsewhere. The GMC result guidance provides quantitative and qualitative feedback, including scores across 3 domains and information about station performance. Give those records more weight than a reconstruction made several days later.
Put the stations into a simple table. Record each domain score, the station result, examiner feedback and the station type. The scoring guide explains why station cut scores vary by diet and why a raw total cannot diagnose performance on its own. Once the information sits side by side, repeated problems are easier to see. Several low interpersonal scores point towards a different repair from severe losses in a few unfamiliar station types.
Numbers and comments answer different questions. The score shows where performance fell, while a specific examiner comment can suggest why. Read a station total beside that station's cut score, because the GMC calculates station standards for each diet. A raw score below a percentage remembered from elsewhere can send preparation in the wrong direction. One acceptable station total can also conceal a weak domain score that returns elsewhere.
The GMC's recent table does not publish pass rates separated by first, second or later attempt. A headline second-attempt percentage cannot tell you what went wrong in your sitting.
A weakness repeated across different stations
When one domain is repeatedly weaker across otherwise different stations, focus on the scored behaviour rather than restarting the whole case bank. Short drills can isolate the problem before you put it back into an 8-minute consultation. More cases will not help if the same error goes uncorrected.
For an interpersonal deficit, a drill can present one patient cue every 60 seconds. You must acknowledge the cue, explore it without assumption, replace jargon with plain language, check understanding and involve the patient in the next decision. The GMC describes interpersonal skills through observable communication and professional behaviours. Accent is not a fourth scoring domain. You do not need to sound British, but speech must be understandable, responsive and appropriate to the patient. Record 5 cue drills and count how many cues received an acknowledgement in the next response, how many explanations contained unexplained jargon, and how many plans included a genuine choice or check of understanding. A study partner can score the same behaviours without judging fluency or accent. If the count improves in drills but falls again inside an 8-minute case, the remaining problem is integration under time rather than basic communication ability.
For a management deficit, each drill begins after the diagnosis has already been supplied. In 90 seconds, the candidate states the safe first step, immediate treatment or escalation, relevant follow-up and a specific safety net. In the next 60 seconds, the examiner or partner changes one fact, such as pregnancy, allergy, haemodynamic instability or refusal, and the plan must change with it. After 2 weeks, the structured preparation plan can reintroduce full cases while preserving a separate domain score after every attempt.
Deep losses in a few station types
If the 3 domains are generally sound but a few stations collapse, look for a shared task that your preparation rarely covered. It may be an angry patient, medical error, safeguarding concern or breaking bad news rather than a broad medical knowledge problem.
In GK's PLAB 2 Study Centre's bank as of August 2026, Medical Errors and Abuse & Safeguarding are much smaller categories than the largest clinical systems. The guide to PLAB 2 cases and station types shows that distribution. It does not predict a particular diet, but it explains how random selection can leave communication and professional scenarios under-practised.
Targeted exposure is faster than restarting the whole syllabus. If 4 station types caused deep losses, practise several versions of each. Change the emotional or ethical pressure as well as the diagnosis. A medical-error station might test disclosure, apology and escalation; a safeguarding station might test privacy, immediate risk and senior help. Then retest the same skills in an unfamiliar version under the 8-minute limit.
Catastrophic stations cannot safely be dismissed as bad luck. The GMC scoring system applies both an overall requirement and a station-level requirement for the diet, so several severe losses can matter even when stronger stations preserve a respectable total. The candidate should therefore retest all 3 or 4 failed types in a mixed circuit, without being told which type comes next. Surprise is part of the diagnostic test.
Small losses across most of the circuit
When no domain or station looks disastrous, check whether time is being lost in the same place. The opening may run long, data gathering may fail to narrow, or the closing phase may repeatedly disappear. Another month of reading is unlikely to repair a loss that appears at minute 7.
Practice should be divided by time, not by pages. One workable training rule is to reach a focused assessment by about minute 4, protect the next 2 minutes for explanation and management, then reserve the final 2 minutes for patient involvement, safety-netting and closure. That is a practice scaffold, not a GMC marking formula. The exact split will change for an examination, procedure or breaking-bad-news station, but every 8-minute attempt needs a deliberate endpoint.
A closing protocol can use 4 checks: the plan is clear, the patient's concern has been answered, worsening symptoms have a specific action, and follow-up has an owner and timescale. Then confirm the timing across a complete circuit. A full mock exam tests whether pacing survives 16 stations, fatigue and recovery after a poor performance. If the same late closure appears across several mock stations, you have found a repeated timing problem. Recovery between stations also needs rehearsal. A 30-second reset can name the previous error once, release it, read the next task and identify the first clinical priority. Carrying one poor station into the next 3 can create a cluster of avoidable losses. A 16-station mock reveals that spillover in a way that isolated practice cannot.
Rebooking logistics and money
Rebook after you understand the result. The GMC process requires the existing place to be cancelled before another available place is booked through GMC Online. Its fees and cancellation table uses a sliding deduction: cancellation more than 42 days before PLAB 2 costs 10% of the test fee, and the deduction rises closer to the date. A preferred replacement date is not guaranteed. Check eligibility, available places and the 2-year deadline before cancelling.
From 1 April 2026, the GMC charges £1,036 for each PLAB 2 booking.
That makes an emotional booking expensive. The time needed depends on the result, your progress in timed practice, available places and the two-year deadline. Broad unsafe management may take longer to repair than a narrow timing problem. Book against evidence of improvement, not embarrassment.
The 2-year clock needs date arithmetic. Suppose a candidate passed PLAB 1 on 15 January 2025 and received a failed PLAB 2 result after a September 2026 sitting. PLAB 2 must be passed by 15 January 2027. A retake on 10 December 2026 would leave only 36 days before the deadline, which gives almost no room for another result, a cancelled sitting or an exceptional-circumstances application. In that example, a 10-week rebuild may be educationally sound but operationally unsafe.
Attempt arithmetic matters too. After a first failure, a candidate normally has 3 standard attempts left. After a third failure, only one standard attempt remains before any application for a fifth and final attempt. The GMC booking rules make the fifth conditional, not automatic. Each new £1,036 booking should therefore have a written repair target, a timed test of that target and a stop rule if the target is still failing in mocks.
When to consider stopping
After a third failed attempt, continuing should require a formal route review before another £1,036 fee is paid. The candidate has one normal attempt left, a fifth is available only by application, and the 2-year PLAB 1 clock may remove choices sooner than the headline attempt count suggests. The review should compare the diagnosed gap, evidence of improvement across at least one 16-station mock, total travel and preparation cost, and the career return. Since the Medical Training (Prioritisation) Act took effect on 6 March 2026, that return is harder for many IMGs; the guide to UK training prioritisation sets out the present rules. A different registration route, another country's exam or a home-country career can be the rational decision when the remaining attempt cannot be prepared safely or no longer supports the candidate's actual career goal.
Questions candidates actually ask
How soon can you retake PLAB 2 after failing?
The GMC does not publish a mandatory waiting period in its booking guidance. You can rebook when eligible and a place is available, but the date should leave enough time to correct the problems shown in your result and still meet the two-year PLAB 1 validity rule.
How many attempts are allowed for PLAB 2?
A candidate normally has four PLAB 2 attempts. A fifth and final attempt is available only by application, and failure at that attempt ends eligibility to take PLAB again.
Should you change your preparation after failing PLAB 2?
Yes. The result breakdown should determine whether the next block targets a repeated domain deficit, several catastrophic station types or a uniform timing problem. Repeating the same cases for more hours leaves the original cause untreated.