|Home | About | Journals | Submit | Contact Us | Français|
Surgeons will sometimes advise against an operation because the patient is ‘old and frail’. A simple starring system (one to five), based on performance and lifestyle, has been devised to assess the biological age of elderly patients. 10 consultant surgeons and 10 trainees answered questions about their treatment recommendations for hypothetical patients of standard age and medical history but with various star ratings and surgical conditions. 1000 decisions were available for analysis.
The four and five star patients (those leading an independent existence) were recommended 266 interventions, the one and two star patients 55. Trainees were more inclined to intervene than consultants, recommending operations in half the patients rather than one-third.
These results indicate that decisions on surgical management are strongly influenced by the patient's star rating or biological age. If the starring method proves reproducible in other patient groups and settings, it could allow better communication on an important factor in clinical decisions.
Improvements in perioperative care have enabled surgeons to undertake major surgery in the elderly, and excellent results have been published1,2. There is, however, always a degree of selection, either before referral or by the surgeon, and the mortality figures in published series do not reflect the true risk of surgery in advanced age.
Patient selection may occur after thorough physiological investigation and the prediction of risk with P-POSSUM or a similar tool3. A risk—benefit analysis can then be used to guide decisions4. The consultation between surgeon and patient, at which a decision must be made regarding intervention, is vitally important. In theory, the surgeon can provide all the available figures on risks and benefits and abdicate the decision entirely. Most patients, however, wish to know what the surgeon believes is right for them. They have difficulty weighing up the figures and their implications, particularly in the very elderly. Others sense that the surgeon has some further information on which to base his advice. They understand that figures drawn from a large population may not reflect the risk—benefit equation for an individual. For example, the national 30-day mortality for a standard operation on an 85-year-old with no other significant medical conditions takes no account of the variation between individual surgeons' results, the difficulty of anatomical access in different patients or the size of a tumour. More importantly, a chronological age since birth gives little indication of how biologically ‘old’ a patient really is. Turner et al.5 have expressed concern that cancer treatment may sometimes be inappropriately denied to individuals classified as ‘old and frail.’ They state that no validated scale for frailty exists for general use and ‘we do not know what variables influence decision making’.
We, however, believe that the apparent or biological age of the patient is a critical factor in the initial decision which a surgeon makes regarding advisability of intervention. Clinicians will convey this information to a colleague with phrases such as ‘remarkable for 88’ or ‘a rather old 74-year-old’. Various scales for global physical function and multidimensional assessment have proved useful research tools to measure the impact of disease or treatment on a patient's life and can also give a measure of biological age. Their complexity, however, renders them unsuitable for use in ordinary clinical circumstances.
In 1999 the first author therefore devised a simple grading system for the elderly. It is biased towards performance and lifestyle and gives a crude indication of biological age6. It requires no physiological measurements or lifestyle questionnaires and merely formalizes the assessment every surgeon makes of an elderly patient during a consultation. The details are outlined in Box 1.
A study was designed to test the hypothesis that an elderly patient's ‘star rating’ or biological age greatly influences the clinical decisions taken by surgeons.
10 consultant surgeons and 10 trainees (senior house officers and specialist registrars) were interviewed. A hypothetical referral letter was shown to them. In the first instance the patient was described as an 83-year-old widowed retired accountant. He was on atenolol for hypertension and there was a history of a myocardial infarct 8 years ago from which he had made a good recovery. He had had a hip replacement. The chronological age and medical history was thus standardized.
Five very different elderly men, representing each of the star ratings, were then the hypothetical patients. The referral letter could have been applicable to any of them. The brief descriptions of these five patients (Box 2) were then given in turn to the participating surgeons.
Each surgeon was asked to consider the correct management for 10 common surgical problems, first in the 2-star patient and then in the 5, 3, 1 and 4-star patients. The surgical conditions are indicated in Box 3. Decisions were recorded as: Y=‘Yes, I would recommend the operation or intervention suggested’; N=‘No, I feel it would be contraindicated’ or DK=‘I don't know. I would need more information to make a decision’.
Each patient had 10 interventions considered by 20 surgeons. 1000 decisions were available for analysis. The ‘don't knows’ were under 10% (93). Table 1 shows the total number of recommendations for intervention for each hypothetical patient. The 4-star and 5-star patients (those maintaining an independent existence) were recommended 266 interventions compared with the 55 interventions recommended for the 1-star and 2-star patients (P<0.001, Wilcoxon rank-sum test). Trainees were more inclined to intervene than consultants, recommending operation in almost half the patients (243) compared with the consultants' one-third (162) (P<0.01, Wilcoxon ranksum test).
Advancing biological age was perceived by the surgeons in this study as a major risk factor separate from chronological age and other medical conditions. It seemed to weigh particularly with the more experienced clinicians, although the difference from trainees was not significant. Other risk assessment tools such as P-POSSUM use chronological age which, as already discussed, is of limited value in the elderly as natural lifespans vary. The physiological variants measured often do not highlight the physiological frailty of extreme age. These can be defined as the lack of capacity to adapt to stress, in this case surgical7. Japanese workers report that in octogenarians total lymphocyte count and performance status correlate positively with survival after abdominal surgery8. A questionnaire survey of UK vascular surgeons showed that the level of independence of a patient was a greater influence on management than cardiac, respiratory or renal function9. The relevance of performance status to surgical decisions is underlined by the findings in the present study: independent 4-star and 5-star patients had intervention recommended in 66% of the decisions, the dependent 1-star and 2-star patients in only 13%.
Not every aspect of risk can be easily measured by mortality and morbidity statistics. The elderly patient leaving hospital within 24 hours of a herniorrhaphy and experiencing no surgical or medical complication may still have his life ruined by the operation. After 2 weeks' convalescence he realizes he has lost confidence on his bicycle and can no longer cycle daily to the village for provisions. A move to sheltered accommodation follows. Many such factors are informally assessed in all clinical decisions and the surgeon's ‘gut feeling’ should not be dismissed10. Subjective opinions may not always be inferior to objective measures: complex computer systems are still inferior to the average man in determining the gender of a face11. It is, however, important to ask whether the surgeon is making the right decision and whether the level of risk acceptable to a patient is similar to that acceptable to the surgeon. In the 10 scenarios outlined in Box 3 the implications of advising no intervention are various. Without surgery the patient with the symptomatic aortic aneurysm will die within a few days. Patients with known malignancy are likely to die of the disease if it is left untreated, but this may be some months or even years hence. In the remaining six scenarios the advantages of intervention are more tenuous, but in most of them inaction could still result in life-threatening complications.
Superficially it might seem that there was nothing to lose by attempting surgery on every patient with a symptomatic aortic aneurysm. This might be so if those who died succumbed during the operation or immediately afterwards. Unfortunately, if operative and postoperative intensive management is offered to all, many who finally die will have spent several weeks in intensive care with needless additional suffering both to themselves and to close family members. In addition, amongst the elderly survivors there will be others who have deteriorated so profoundly that they are unable to regain a life-style they would have regarded as worthwhile before their operation. Such factors are difficult to quantify but patients and relatives often expect the surgeon to take them into account. They will, however, wish to be sure he is not influenced by other concerns. The surgeon may be worried over falling staff morale in the face of multiple postoperative deaths. Intensive care is expensive, and if measured against an infrequent successful outcome becomes progressively more expensive. Intensive care beds are precious, and elective cases are cancelled or postponed when the intensive care unit is full. All these considerations are reasonable in a rationed health service but the surgeon who takes these factors into account can no longer act as the individual patient's advocate.
The decisions in the known malignancies are also a complex balance of risk. The patient may initially only see the advantage of a possible cure. If, however, an operation offers only a 25% chance of a final cure then a 15% risk of postoperative death becomes more telling. The length and quality of survival without treatment is also important. A mean 2-year survival without severe symptoms is a devastating prognosis at 30 years but not at 80 years. The 2-year survival of extreme biological old age in the absence of disease is not even known. The ill, elderly patient cannot be expected to weigh all these risks and make decisions without guidance from the surgeon. A new and pernicious influence on the surgeon could be the pressures of national audit and hospital league tables. Good figures can always be produced by denying intervention to the highest risk patients.
An encouraging observation emerged from this study. When the data showed the consultants overall less keen than the trainees to operate on the elderly, scenarios 1, 2, 4 and 10 were analysed separately: these all represented cases in which intervention would carry a high risk but in which the patient was likely to die without operation. The consultants recommended surgery in 37% of these scenarios, the trainees in only 33%. The small difference here contrasts with the trainees' greater enthusiasm for intervention in general. This is the group of conditions in which concerns over resources and audit would be higher and such concerns, if they were unduly influencing decisions, would have been more likely to have impinged on the decisions of the consultants. The consultants were only more cautious in recommending non-essential surgery.
This study suggests that an elderly person's biological age or ‘star rating’ is highly influential in surgical management. However, future studies may be required to test the general assumption that biological age, or increasing frailty, is a good predictor of surgical outcome. ‘Star rating’ is little more than formalization of the assessment a surgeon makes during every consultation. If it proves reproducible in other patient groups and by other clinicians, it could prove a valuable aid to communication between clinicians.
We thank all surgeons who took part in this study.