Feature Christmas 2011: Professional Matters

When is a “free” registrar in clinic not free?

BMJ 2011; 343 doi: http://dx.doi.org/10.1136/bmj.d7869 (Published 20 December 2011) Cite this as: BMJ 2011;343:d7869
  1. Amrit S Lota, specialist trainee in cardiology,
  2. Charlotte H Manisty, specialist registrar in cardiology,
  3. Richard Sutton, emeritus professor of cardiology,
  4. Darrel P Francis, reader in clinical cardiology
  1. 1International Centre for Circulatory Health, National Heart and Lung Institute, Imperial College London, London W2 1LA, UK
  1. Correspondence to: A S Lota amrit.lota{at}doctors.org.uk


Objective To investigate whether using registrars (doctors undergoing higher specialist training, whose salary is reimbursed) rather than consultants in outpatient clinics saves money

Design Development of a formula calculating the economic breakeven point and application to retrospective audit data from 273 outpatient consultations.

Setting General cardiology outpatient clinic in a secondary and tertiary referral NHS hospital.

Outcomes Difference in probability of a registrar and a consultant making a diagnostic decision that completes a clinical episode. Use of UK costings for consultant salaries and outpatient attendances to determine the economic breakeven point.

Results The formula showed that if a registrar’s episode completing probability is 12 percentage points lower than that of a consultant, then using a registrar costs the hospital more. Real life data showed that episode completion probabilities are 43 percentage points lower for registrars than for consultants (26% versus 69%, 95% CI 32% to 54%, P<0.0001).

Conclusion It is wrong to assume that external reimbursement of registrar salaries makes them a money saving option for staffing clinics. The apparent service role of a registrar can be a disservice.


The Trojan people were warned by their high priest to “fear the Greeks, even those bearing gifts,” but without hard data he could not prevent them giving in to the temptation of a free gift and suffering calamity.1 Are UK hospitals relying on apparently free registrars in outpatient clinics heading for the same fate? Patients attending outpatient clinics in NHS hospitals are often seen by higher specialist trainees (registrars) rather than consultants. Registrars are appointed by regional training bodies (deaneries), who allocate them to individual hospitals for specialist training and reimburse the hospitals (NHS trusts) for their salaries. Trusts may therefore view the use of registrars in clinics as saving money.

Numbers of registrars provided by deaneries are set to fall. The first reaction in most hospitals to this prospect is to ask, “How will our clinics cope?” However, rather than struggling to explain to hospitals that registrars are allocated to receive training and not primarily to deliver clinical services, it may be wiser for deaneries to appeal to hospital’s baser instincts: finance.

After an outpatient visit patients are either discharged back to primary care (at which point the hospital receives funding for the completed episode) or offered further follow-up, often with intervening clinical investigations. In addition to the costs of an additional outpatient clinic consultation, the true cost of a follow-up appointment includes innumerable costs of clerical, nursing, and technical staff to arrange and document both that next appointment and any intervening tests. As registrars have less experience than consultants, they may be less likely to reach a diagnosis and complete the treatment episode. We determined the level of deficit in episode completion rates of registrars at which it makes no financial difference to the hospital whether a patient is seen by a registrar or a consultant. We then used data from our cardiology unit to find out whether the nominally “zero cost” registrar is a net financial benefit or burden.


Developing the formula

Let the probability of consultants and registrars concluding the episode by discharging the patient be PC and PR respectively. The costs of an outpatient appointment to the hospital comprise the consultant salary (C), which applies only if the patient sees a consultant, and all other costs associated with a consultation such as clinic nurses, secretaries, medical records, porters, and unreimbursed outpatient test costs, which we will call A. These apply regardless of whom the patient sees. For the purpose of this analysis, we consider the net salary cost to the hospital of the registrar during clinic to be zero as it is paid by the deanery.

We can now consider the alternatives quantitatively. A convenient base state for comparison is that of a patient seeing a registrar who successfully completes the episode (although any other base state can be used). If a consultation does not complete the episode, then the immediate cost consequence is A, composed of the administrative elements of a future appointment plus the typical intervening tests. It is not necessary to consider the doctor’s salary of the next consultation because with entry into the consultation room on subsequent consultations, the patient re-enters the analysis again.

Suppose the consultation is done by a registrar. There is a probability PR that the patient completes his or her episode (zero incremental cost). There is also a probability 1-PR that the episode is not completed and therefore the patient consumes further resources A (before re-entering the analysis). Therefore, as a whole, seeing a registrar has a net financial effect compared with the base state of PR×0+(1–PR)×–A or –A(1–PR).

The alternative is that the consultation is done by a consultant. In this case there will definitely be a salary cost C and, with probability (1–PC), the cost of an additional future consultation A. Having a patient seen by a consultant therefore has a net financial effect compared with the base state of –C A(1–PC).

Break-even point

The costs of non-completion between consultant and registrar are balanced when

CA(1–PC) = –A(1–PR) or, more concisely, C/A =PC PR. When PCPR >C/A, having the patient seen by a registrar instead of a consultant is a net drain on hospital resources. All that need be done to complete the case is to calculate actual values for C/A, and PCPR.

Audit of discharge probabilities of consultants and registrars

We conducted an audit of episode completion probabilities for consultants and registrars in several cardiology clinics at our trust during April–June 2010. One observer (ASL) reviewed individual clinic letters to determine the nature of the appointment (new versus follow-up), principal diagnosis, type of doctor (registrar or consultant), and whether the episode was concluded. Proportions are reported as percentages, with 95% confidence intervals calculated by using the binomial theorem. We compared proportions using Fisher’s exact test, two tailed. A P value <0.05 was considered significant.


Break-even point in UK

From NHS billing rates, a follow-up cardiology consultation including all bundled tests is currently reimbursed at £108 (€126; $170).2 We used this as an estimate of A. If we assume a consultation is about 20 minutes, the average consultant salary cost (C) is £13.3 Thus C/A is about 0.12.

This means that if a registrar’s probability of completing a patient episode is more than 12 absolute percentage points lower than that of a consultant, it is financially damaging to use registrars in clinic, even when their salaries are fully reimbursed.

Audit of discharge probabilities

We audited 273 consultations for 273 unique patients in the general cardiology clinic; 149 were new referrals and 124 were follow-up appointments. Table 1 shows the spectrum of diagnoses.

Table 1

 Principal diagnoses in cardiology outpatient audit

View this table:

The overall completion probabilities are 69% (95% confidence interval 61% to 77%) for consultants and 26% (19% to 33%) for registrars, a difference of 43 percentage points (32% to 54%, P<0.0001). For new patients, the completion probabilities are 77% (68% to 86%) for consultants and 31% (20% to 42%) for registrars, a difference of 46% (31% to 60%, P<0.0001). For follow-up patients, the completion probabilities are 53% (37% to 69%) and 22% (13% to 31%) respectively, a difference of 31% (12% to 49%, P=0.0014). All these figures (table 2) far exceed C/A, indicating that the net effect of using a nominally free registrar in outpatient clinics is financially adverse.

Table 2

 Number of patients seen and episodes completed by consultants and registrars for new referrals and follow-up patients

View this table:


The assumption in many NHS hospital trusts that registrars whose salaries are externally reimbursed are always financially favourable for hospital clinics may be incorrect. This assumption is far from true in our cardiology department. Our figures indicate that the extra consultants necessary to replace registrars in clinics could be paid for 3-4 times over with the money saved on administrative costs generated by unnecessary follow-up investigations and appointment administration.

Implications for trusts

Trusts worried about the financial damage of losing registrars from clinics will be relieved to learn that the converse may be true. It is wrong to look in isolation at the salary cost of an employee when financially pressed; resource drainage by some employees may be many times greater than their salary. Using our formula, trusts can identify situations where it is more cost effective to pay for an additional consultant rather than rely on an apparently free pair of hands.4

Hospitals currently think they do not have funds to employ more consultants. Our study shows that additional consultants could be readily funded from the cost savings recouped by not using registrars for outpatient clinics.

Training perspective

Registrars are supplied to trusts to receive training, but in times of extreme pressure trusts may think it is rational to treat all employees as tools to deliver service. Educational and clinical supervisors have a professional responsibility to prioritise training, and the knowledge that using registrars in clinics may be more expensive will assist them in protecting registrars from being used by trusts as service providers.5

Training is changing.6 The modern trainee has many fewer years and many fewer hours a year in which to gain the skills necessary to become a consultant.7 Our results support rearranging timetables to favour training.

Trainees should not be excluded from outpatient clinics because they do need to learn to manage clinic consultations independently once they become consultants. Our analysis shows merely that registrars should not be asked to make decisions beyond their expertise. Paradoxically, trusts would save money by having the clinic registrars sit in with consultants and receiving one to one training, as long as this did not decrease the productivity of the consultant.

Study limitations

We have developed a new formula and applied it to a single specialty in a single institution. We do not know if the absolute difference in episode completion rate (PCPR) is similarly large in other institutions or other specialties. However, it would have to be about three times smaller for the free gift registrar to be truly free of financial harm to the trust.

This was not an interventional study, and we did not set out to investigate clinical outcomes or patient outcomes and did not assess or incorporate patients’ preferences regarding type of physician. Nor can it confirm that there are suitable candidates for new consultant posts in every specialty. However, in cardiology they are plentiful, with trainees nationwide now as numerous as consultants.

New patients had a higher probability of completion than those attending follow-up appointments. Patients already under follow-up tended to continue to need further follow-up regardless of whom they saw. Nevertheless, the probability of completion was still lower for registrars than consultants for these types of appointment and use of a registrar was a net financial harm.


The true reason that registrars should not be filling consultant roles in outpatient clinics is that they are supposed to be receiving training instead. However, trainers who want to achieve good training, but face resistance from trusts, can also argue using allegory,1 algebra, or audit.

What is already known on this topic?

  • The salaries of doctors in higher specialist training posts in NHS hospitals are reimbursed by the regional deaneries

  • Some trusts assume they are a zero cost gift that can help with outpatient service provision

What this study adds

  • A simple formula can be used to determine the costs or savings from using registrars instead of employing additional consultants

  • Audit data from our hospital’s cardiology clinics show that the costs of extra follow-up appointments and tests ordered by registrars far exceeds the savings from reimbursed salaries

  • The Trojans were wrong to assume a free gift can have no cost; trusts should not make the same miscalculation


Cite this as: BMJ 2011;343:d7869


  • Contributors: RS recognised that trusts erroneously assume a registrar is financially beneficial. DPF (guarantor) developed the three pronged explanation for trusts, ASL derived the formula and conducted the audit, CHM critically reviewed and edited the manuscript.

  • Competing interests: All authors have completed the ICJME unified declaration form (available on request from the corresponding author) and declare: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous 3 years; no other relationships or activities that could appear to have influenced the submitted work.

  • Provenance and peer review: Not commissioned; externally peer reviewed.