676
50/mm3 (50 000/ml) in urine samples obtained by either technique, counts/mm3 (bag vs nappy) were 750 vs 0, 90 vs 10, 500 vs 220, 500 vs 20, and 70 vs 0. To investigate this difference further, urine that contained large numbers of
erythrocytes and white blood cells was poured onto a nappy and left at room temperature for 3 h. Microscopic examination of the extracted urine showed an 8-fold drop in both cell counts, which indicates retention or lysis of the cells by the nappy material. The results of biochemical analysis of urinary constituents from midstream urine specimens before and after retrieval from disposable nappies are shown in the figures. Fig 1 shows linear regression plots: correlation coefficients (r) were 0-999 for sodium, 0-997 for potassium, 0-999 for creatinine, 0-999 for osmolality, 0-999 for urea, 0-997 for phosphate, and 0-978 for calcium (all p < 0-001). Fig 2 shows Bland-Altman2 plots, which compare the differences between the methods (y axis) against the mean of the two methods (x axis), with results expressed as bias (mean difference) and limits of agreement (mean difference ± 2 SD). In general there was good agreement between concentrations measured in fresh and recovered urine, with values of bias from the Bland-Altman plots of approximately 0; the largest variations were seen for calcium and phosphate concentrations, and urine from nappies may be unsuitable for measurement of these constituents. Discussion Culture results from urine specimens obtained from
padded disposable nappies showed that children with urinary tract infections were correctly identified, and that most children without infections had sterile samples. Contaminated samples that gave a mixed growth of organisms or less than 10s CFU/ml of one species were seen with both types of urine collection, but were almost twice as
bag collections compared with nappy samples (17/45 vs 10/45, respectively). As urine passed into a bag will wash repeatedly over the perineum, whereas urine voided into a nappy rapidly passes through the surface layer to remain in deeper fibres which are not in contact with the skin, this finding may be less surprising than it first appears. On light microscopy of urine obtained by both techniques, erythrocyte and leucocyte counts were much lower, and therefore unreliable, in nappy samples, but bacterial counts, which may be more helpful for diagnosis of urine infections in children,3were not altered. Urine specimens collected from disposable nappies are also usable for the measurement of most commonly analysed biochemical constituents (with the probable exception of calcium and phosphate) and of osmolality. Our findings indicate that urine collection from a padded (not ultra-absorbent) disposable nappy which has been common
from
for less than 4 h and is not soiled with faeces is a convenient and, for most urinary constituents, an accurate alternative to collection in urine bags. This novel technique is simple, non-invasive, rapid, and inexpensive (in the UK the cost of a 20 ml syringe is less than a tenth of the cost of a urine bag); it should be particularly useful in hospital and outpatient departments and general practitioners’ surgeries, and for domiciliary visits-indeed, the specimen may often worn
arrive with the
patient. REFERENCES
SB, Lucas A. A nappy collection method for measuring urinary constituents and 24-hour urine output in infants. Arch Dis Child 1985; 60: 1018-20. 2. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; i: 307-10. 3. Vickers D, Ahmad T, Coulthard MG. Diagnosis of urinary tract infection in children: fresh urine microscopy or culture? Lancet (in 1. Roberts
press).
Critical incidents in the intensive therapy unit
Preventable mishaps in an intensive therapy unit were studied over 12 months by the critical incident technique. Staff were encouraged to complete confidential questionnaires describing incidents in which they had participated or had observed. This allowed classification of the events and examination of the views of staff on causes, detection, and prevention. 110 (80%) of 137 events were felt to have been due to human error; the remainder were due to equipment failure. Inexperience with equipment and shortage of trained staff were the factors most often felt to contribute to incidents. The critical incident technique is a useful way of improving standards of clinical care.
Introduction The critical incident technique was developed by Flanagan and others in the 1940s and was shown to be useful in the investigation of aircraft accidents. By examining factors responsible for an accident, the technique
draws attention to certain critical incidents that have an observable effect on the course of events. The technique has been widely appliedand we wondered whether it could be helpful in investigating mishaps in an intensive therapy unit
(ITU). We based our study on investigations of anaesthetic mishaps.3A critical incident was defined as an occurrence that could have led (if not discovered or corrected in time), or did lead, to an undesirable outcome. The incident had to involve an error by a member of staff or a failure of equipment. It had to have occurred while a patient was under the care of ITU staff but not necessarily within the ITU. The incident had to be describable in detail by a participant or observer and it had to be clearly preventable.
Methods We used a questionnaire based on one developed by Williamson and colleagues.’ The questionnaire asked staff to describe the ADDRESS: Intensive Therapy Unit, Western General Hospital, Edinburgh EH4 2XU, UK (D. Wright, FCAnaes, S. J. Mackenzie, FCAnaes, I. Buchan, RGN, Dip N, C. S. Cairns, RGN, L. E. Price, RGN) Correspondence to Dr David Wright.
677
critical incident and its consequences and also what was believed to have contributed to the occurrence, detection, and correction of the incident. It also asked for suggestions on how to avoid future incidents. Blank questionnaires were left in the staff lounge. They were completed anonymously and posted in a letter box that was
regularly emptied by
one
of
The questionnaires were group of senior doctors and prepared and circulated to all staff. us.
summarised before consideration by nurses.
A monthly report was
TABLE II-FACTORS FELT TO CONTRIBUTE TO INCIDENTS
a
Results 145 questionnaires describing 137 events were completed 1 year (table I). Error by staff was felt to have contributed to 80% of events, while only 20% could be ascribed to equipment failure. None of the incidents had a serious long-term outcome. The "human error" events fell into four main groups: those related to ventilation, to vascular lines, to drug administration, and a miscellaneous group. The following are examples of the events reported. over
Example 1-A 68-year-old man with chronic obstructive airways disease and respiratory failure was on a ventilator. The ventilator prescription and the patient’s 24 h recording chart described the ventilator setting in SIMV mode (synchronous intermittent mandatory ventilation) as 10 breaths per min. However, when the patient was checked during a ward round, the ventilator setting was SIMV 8 breaths per min. About 1 h before the ward round the ventilator setting had been changed for good clinical reasons but this change had not been documented. The new setting
then correctly prescribed. Undocumented changes in ventilator settings are common and usually of little consequence, but they are mistakes and may have more was
serious results.
Example 2-A 70-year-old patient returned from theatre seen to be restless, agitated, and hypotensive. The patient was supposed to be receiving a paraveretum infusion, but, on checking, a three-way tap in the intravenous line was turned off. The intravenous syringe driver had not sounded an alarm. The tap was opened, a bolus of paraveretum given, the patient settled and the infusion continued. Three-way taps are often turned off for good reason-eg, to give a bolus of one drug when two drugs are normally being given. Mistakes occur when staff forget to return the tap to the position that allows both drugs to run. Example 3-A three-way tap in an intravenous catheter inserted into a patient’s femoral vein was accidentally knocked open. Because a bung was not fitted, blood leaked out under the sheets. The patient was lying on a bed made and was
from several air-filled cushions rather than on a conventional mattress. Blood passed between the cushions and fell on to the floor where it was noticed more rapidly than if bleeding had been on to a mattress. Two factors were important in this incident: most intravenous lines are not TABLE I-CAUSES OF 137 CRITICAL INCIDENTS
TABLE III-FACTORS FELT TO CONTRIBUTE TO DETECTION OF INCIDENTS
under the sheets-this is a problem with femoral lines; and a bung on the three-way tap would have prevented the leak. Example 4-Before a ward round, haematology results for several patients were telephoned to the ITU. A doctor noted them on a piece of paper and wrote patients’ initials against each set of results. Instead of transferring results to the notes, he asked another doctor to do so. This action led to juxtaposition of patients’ results in the notes by the second doctor. On the ward round, one set of results for haemoglobin, white blood cell count, and platelets gave a misleading impression of a patient’s progress. The mistake was discovered quickly but inappropriate treatment could have been given. Example —A 64-year-old man with a recent myocardial infarction was being treated for acute left ventricular failure. Intravenous hydration fluid was prescribed as 50 ml per h via an infusion pump. However, the pump was inadvertently set at 500 ml per h and ran for an hour before the mistake was noticed. No adverse effects resulted but hydration was stopped for some hours. A factor contributing to the incident was felt to be placement of the pump below the level of the bed, because several pumps were in use on the same patient, and so the setting was not easily visible. The factors most often felt by staff to contribute to critical incidents were inexperience with equipment and shortage of trained staff (table 11). Factors contributing to detection of critical incidents included repeated regular checking, presence of alarms on equipment, and presence of experienced staff (table III). Suggestions for avoiding future incidents included repeated checking, particularly at handovers; written protocol or check lists; "don’t ignore alarms"; better supervision of inexperienced staff; "suspect more than one cause"; and "if you don’t know, ask".
Discussion That the majority of incidents reported in a "high tech" speciality were due to human error may seem surprising, but other studies have shown similar results.3.4 We have no reason to suppose that our staff are particularly error prone and suspect that such incidents are widespread. We have chosen our examples to typify events that occurred more than once. Studies such as this can identify which incidents are common and why they occurred, and suggest ways in which they can be prevented.
678
Unfamiliarlity with equipment is important, particularly new staff, and this led us to increase tuition and supervision. Errors in drug administration stress the importance of having sufficient intravenous lines to avoid multiple use of a line and the placing of pumps at a height where settings are easily seen. Regular checking by experienced staff is critical in detecting errors, but this may be adversely affected by nurse staffing policies where agency staff are commonly used or where little time is available for handovers. It is important to remember that while human error is reponsible for most critical incidents, it is staff, usually experienced staff, who correct the incidents. It has been suggested5 that systematic study of medical accidents and their causes has been neglected. Mortality and morbidity meetings are of help in analysing particular incidents, but they may not be systematic and they can produce a public apportioning of blame that does not help understanding and acceptance of responsibility for error. The critical incident technique looks systematically for areas of weakness that may have a serious outcome, identifying errors that are "both clinically important and highly prevalent’? The incidents analysed need not have serious outcomes, indeed in this study no incident had a serious long-term effect. Nevertheless, analysis of incidents produced many changes in our practice. The anonymity of the questionnaires encouraged participation and an acknowledgement by staff of their responsibility in causing with
error.
Several events were reported by more than one person, and this was easily noted because of the detailed reporting that the questionnaires encouraged. Concern over the medicolegal implications of dissemination of knowledge about a serious mishap is lessened by having only one person reading the completed questionnaires who summarises details for others. Questionnaires are then destroyed while summaries are retained. Studies such as this do not tell us how common critical incidents are, since under-reporting is likely. Williamson et al pointed out that "the technique makes no claims whatever for being able to detect the absolute incidence of anything"."4 They are, however, a way of improving clinical standards that are acceptable to staff and do not require additional resources. Once instituted, the continued use of critical incident reporting allows the impact of corrective measures or other changes to be assessed. The critical incident technique should be used more widely in medical practice. REFERENCES 1.
Flanagan JC. The critical incident technique. Psychol Bull 1954; 51:
327-58. 2. Anon. Critical questions, critical incidents; critical answers. Lancet 1988; i: 1373-74. 3. Cooper JB, Newbower RS, Long CD, McPeek B. Preventable anesthesia mishaps: a study of human factors. Anesthesiology 1978; 49: 399-406. 4. Williamson JA, Webb RK, Pryor GL. Anaesthesia safety and the "critical incident" technique. Aust Clin Rev 1985; 5: 57-61. 5. Vincent CA. Research into medical accidents: a case of negligence. Br Med J 1989; 299: 1150-53.
VIEWPOINT Ethics of lung transplantation with live donors
Lung transplantation has become a viable treatment option for many patients with terminal pulmonary disease but the scarcity of donor lungs is a severe constraint. Hence many patients accepted into lung transplant programmes die while awaiting transplantation-as have nearly a quarter of patients admitted to the lung transplant programme in Toronto since 1984. The use of live donors might relieve donor lung scarcity. Live donors have been used in kidney transplantation since 1954 and in liver transplantation since 1989. In October, 1990, the Stanford group did the world’s first live-donor lung transplant.l,2 Ethical questions have been raised about this procedure.3 Recent experience with live-donor liver transplantation has demonstrated the potential benefit of analysing ethical considerations before embarking on innovative transplant procedures," which is what we plan to do here for lung transplantation. Risks and benefits
Recipients For the recipient the main disadvantage of a live donor is that only a lobe rather than the whole lung will be transplanted. Survival rates with transplants from live donors are unknown, but they are likely to be similar to those
obtained with cadaveric-donor single-lung transplantation, which are 76% at 6 months, 65% at 1 year, 58% after 2 years, and 53% after 3 years7 (the longest single-lung transplant survivor died of renal failure in March, 1990,6-25 years after the operation). Live-donor lung transplant also carries the risk that as the recipient grows the transplanted lobe will no longer be sufficient to support increasing functional requirements; however, the extent of this risk is likely to be similar for a cadaveric-lung transplant in a
growing recipient. The first of the four potential benefits to recipients is a reduction in numbers dying while waiting for a transplant since, unlike the case with kidney failure, where continued dialysis is an alternative to transplantation, there is no means of replacement therapy for end-stage lung disease except for transplantation. Secondly, by boosting donor numbers the ADDRESSES: Centre for Bioethics (L. R. Shaw, MA, P. A. Singer, MD), and Departments of Medicine (A. S. Slutsky, FRCPC, J. R. Maurer, FRCPC) and Surgery (J D. Miller, FRCSC, J. D. Puskas, MD, G. A. Patterson, FRCSC), University of Toronto, Toronto, Canada. Correspondence to Dr P. A. Singer, Centre for Bioethics, University of Toronto, Tanz Building, Room 303, 6 Queen’s Park Crescent West, Toronto, Canada M5S-1A8.
50/mm3 (50 000/ml) in urine samples obtained by either technique, counts/mm3 (bag vs nappy) were 750 vs 0, 90 vs 10, 500 vs 220, 500 vs 20, and 70 vs 0. To investigate this difference further, urine that contained large numbers of
erythrocytes and white blood cells was poured onto a nappy and left at room temperature for 3 h. Microscopic examination of the extracted urine showed an 8-fold drop in both cell counts, which indicates retention or lysis of the cells by the nappy material. The results of biochemical analysis of urinary constituents from midstream urine specimens before and after retrieval from disposable nappies are shown in the figures. Fig 1 shows linear regression plots: correlation coefficients (r) were 0-999 for sodium, 0-997 for potassium, 0-999 for creatinine, 0-999 for osmolality, 0-999 for urea, 0-997 for phosphate, and 0-978 for calcium (all p < 0-001). Fig 2 shows Bland-Altman2 plots, which compare the differences between the methods (y axis) against the mean of the two methods (x axis), with results expressed as bias (mean difference) and limits of agreement (mean difference ± 2 SD). In general there was good agreement between concentrations measured in fresh and recovered urine, with values of bias from the Bland-Altman plots of approximately 0; the largest variations were seen for calcium and phosphate concentrations, and urine from nappies may be unsuitable for measurement of these constituents. Discussion Culture results from urine specimens obtained from
padded disposable nappies showed that children with urinary tract infections were correctly identified, and that most children without infections had sterile samples. Contaminated samples that gave a mixed growth of organisms or less than 10s CFU/ml of one species were seen with both types of urine collection, but were almost twice as
bag collections compared with nappy samples (17/45 vs 10/45, respectively). As urine passed into a bag will wash repeatedly over the perineum, whereas urine voided into a nappy rapidly passes through the surface layer to remain in deeper fibres which are not in contact with the skin, this finding may be less surprising than it first appears. On light microscopy of urine obtained by both techniques, erythrocyte and leucocyte counts were much lower, and therefore unreliable, in nappy samples, but bacterial counts, which may be more helpful for diagnosis of urine infections in children,3were not altered. Urine specimens collected from disposable nappies are also usable for the measurement of most commonly analysed biochemical constituents (with the probable exception of calcium and phosphate) and of osmolality. Our findings indicate that urine collection from a padded (not ultra-absorbent) disposable nappy which has been common
from
for less than 4 h and is not soiled with faeces is a convenient and, for most urinary constituents, an accurate alternative to collection in urine bags. This novel technique is simple, non-invasive, rapid, and inexpensive (in the UK the cost of a 20 ml syringe is less than a tenth of the cost of a urine bag); it should be particularly useful in hospital and outpatient departments and general practitioners’ surgeries, and for domiciliary visits-indeed, the specimen may often worn
arrive with the
patient. REFERENCES
SB, Lucas A. A nappy collection method for measuring urinary constituents and 24-hour urine output in infants. Arch Dis Child 1985; 60: 1018-20. 2. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; i: 307-10. 3. Vickers D, Ahmad T, Coulthard MG. Diagnosis of urinary tract infection in children: fresh urine microscopy or culture? Lancet (in 1. Roberts
press).
Critical incidents in the intensive therapy unit
Preventable mishaps in an intensive therapy unit were studied over 12 months by the critical incident technique. Staff were encouraged to complete confidential questionnaires describing incidents in which they had participated or had observed. This allowed classification of the events and examination of the views of staff on causes, detection, and prevention. 110 (80%) of 137 events were felt to have been due to human error; the remainder were due to equipment failure. Inexperience with equipment and shortage of trained staff were the factors most often felt to contribute to incidents. The critical incident technique is a useful way of improving standards of clinical care.
Introduction The critical incident technique was developed by Flanagan and others in the 1940s and was shown to be useful in the investigation of aircraft accidents. By examining factors responsible for an accident, the technique
draws attention to certain critical incidents that have an observable effect on the course of events. The technique has been widely appliedand we wondered whether it could be helpful in investigating mishaps in an intensive therapy unit
(ITU). We based our study on investigations of anaesthetic mishaps.3A critical incident was defined as an occurrence that could have led (if not discovered or corrected in time), or did lead, to an undesirable outcome. The incident had to involve an error by a member of staff or a failure of equipment. It had to have occurred while a patient was under the care of ITU staff but not necessarily within the ITU. The incident had to be describable in detail by a participant or observer and it had to be clearly preventable.
Methods We used a questionnaire based on one developed by Williamson and colleagues.’ The questionnaire asked staff to describe the ADDRESS: Intensive Therapy Unit, Western General Hospital, Edinburgh EH4 2XU, UK (D. Wright, FCAnaes, S. J. Mackenzie, FCAnaes, I. Buchan, RGN, Dip N, C. S. Cairns, RGN, L. E. Price, RGN) Correspondence to Dr David Wright.
677
critical incident and its consequences and also what was believed to have contributed to the occurrence, detection, and correction of the incident. It also asked for suggestions on how to avoid future incidents. Blank questionnaires were left in the staff lounge. They were completed anonymously and posted in a letter box that was
regularly emptied by
one
of
The questionnaires were group of senior doctors and prepared and circulated to all staff. us.
summarised before consideration by nurses.
A monthly report was
TABLE II-FACTORS FELT TO CONTRIBUTE TO INCIDENTS
a
Results 145 questionnaires describing 137 events were completed 1 year (table I). Error by staff was felt to have contributed to 80% of events, while only 20% could be ascribed to equipment failure. None of the incidents had a serious long-term outcome. The "human error" events fell into four main groups: those related to ventilation, to vascular lines, to drug administration, and a miscellaneous group. The following are examples of the events reported. over
Example 1-A 68-year-old man with chronic obstructive airways disease and respiratory failure was on a ventilator. The ventilator prescription and the patient’s 24 h recording chart described the ventilator setting in SIMV mode (synchronous intermittent mandatory ventilation) as 10 breaths per min. However, when the patient was checked during a ward round, the ventilator setting was SIMV 8 breaths per min. About 1 h before the ward round the ventilator setting had been changed for good clinical reasons but this change had not been documented. The new setting
then correctly prescribed. Undocumented changes in ventilator settings are common and usually of little consequence, but they are mistakes and may have more was
serious results.
Example 2-A 70-year-old patient returned from theatre seen to be restless, agitated, and hypotensive. The patient was supposed to be receiving a paraveretum infusion, but, on checking, a three-way tap in the intravenous line was turned off. The intravenous syringe driver had not sounded an alarm. The tap was opened, a bolus of paraveretum given, the patient settled and the infusion continued. Three-way taps are often turned off for good reason-eg, to give a bolus of one drug when two drugs are normally being given. Mistakes occur when staff forget to return the tap to the position that allows both drugs to run. Example 3-A three-way tap in an intravenous catheter inserted into a patient’s femoral vein was accidentally knocked open. Because a bung was not fitted, blood leaked out under the sheets. The patient was lying on a bed made and was
from several air-filled cushions rather than on a conventional mattress. Blood passed between the cushions and fell on to the floor where it was noticed more rapidly than if bleeding had been on to a mattress. Two factors were important in this incident: most intravenous lines are not TABLE I-CAUSES OF 137 CRITICAL INCIDENTS
TABLE III-FACTORS FELT TO CONTRIBUTE TO DETECTION OF INCIDENTS
under the sheets-this is a problem with femoral lines; and a bung on the three-way tap would have prevented the leak. Example 4-Before a ward round, haematology results for several patients were telephoned to the ITU. A doctor noted them on a piece of paper and wrote patients’ initials against each set of results. Instead of transferring results to the notes, he asked another doctor to do so. This action led to juxtaposition of patients’ results in the notes by the second doctor. On the ward round, one set of results for haemoglobin, white blood cell count, and platelets gave a misleading impression of a patient’s progress. The mistake was discovered quickly but inappropriate treatment could have been given. Example —A 64-year-old man with a recent myocardial infarction was being treated for acute left ventricular failure. Intravenous hydration fluid was prescribed as 50 ml per h via an infusion pump. However, the pump was inadvertently set at 500 ml per h and ran for an hour before the mistake was noticed. No adverse effects resulted but hydration was stopped for some hours. A factor contributing to the incident was felt to be placement of the pump below the level of the bed, because several pumps were in use on the same patient, and so the setting was not easily visible. The factors most often felt by staff to contribute to critical incidents were inexperience with equipment and shortage of trained staff (table 11). Factors contributing to detection of critical incidents included repeated regular checking, presence of alarms on equipment, and presence of experienced staff (table III). Suggestions for avoiding future incidents included repeated checking, particularly at handovers; written protocol or check lists; "don’t ignore alarms"; better supervision of inexperienced staff; "suspect more than one cause"; and "if you don’t know, ask".
Discussion That the majority of incidents reported in a "high tech" speciality were due to human error may seem surprising, but other studies have shown similar results.3.4 We have no reason to suppose that our staff are particularly error prone and suspect that such incidents are widespread. We have chosen our examples to typify events that occurred more than once. Studies such as this can identify which incidents are common and why they occurred, and suggest ways in which they can be prevented.
678
Unfamiliarlity with equipment is important, particularly new staff, and this led us to increase tuition and supervision. Errors in drug administration stress the importance of having sufficient intravenous lines to avoid multiple use of a line and the placing of pumps at a height where settings are easily seen. Regular checking by experienced staff is critical in detecting errors, but this may be adversely affected by nurse staffing policies where agency staff are commonly used or where little time is available for handovers. It is important to remember that while human error is reponsible for most critical incidents, it is staff, usually experienced staff, who correct the incidents. It has been suggested5 that systematic study of medical accidents and their causes has been neglected. Mortality and morbidity meetings are of help in analysing particular incidents, but they may not be systematic and they can produce a public apportioning of blame that does not help understanding and acceptance of responsibility for error. The critical incident technique looks systematically for areas of weakness that may have a serious outcome, identifying errors that are "both clinically important and highly prevalent’? The incidents analysed need not have serious outcomes, indeed in this study no incident had a serious long-term effect. Nevertheless, analysis of incidents produced many changes in our practice. The anonymity of the questionnaires encouraged participation and an acknowledgement by staff of their responsibility in causing with
error.
Several events were reported by more than one person, and this was easily noted because of the detailed reporting that the questionnaires encouraged. Concern over the medicolegal implications of dissemination of knowledge about a serious mishap is lessened by having only one person reading the completed questionnaires who summarises details for others. Questionnaires are then destroyed while summaries are retained. Studies such as this do not tell us how common critical incidents are, since under-reporting is likely. Williamson et al pointed out that "the technique makes no claims whatever for being able to detect the absolute incidence of anything"."4 They are, however, a way of improving clinical standards that are acceptable to staff and do not require additional resources. Once instituted, the continued use of critical incident reporting allows the impact of corrective measures or other changes to be assessed. The critical incident technique should be used more widely in medical practice. REFERENCES 1.
Flanagan JC. The critical incident technique. Psychol Bull 1954; 51:
327-58. 2. Anon. Critical questions, critical incidents; critical answers. Lancet 1988; i: 1373-74. 3. Cooper JB, Newbower RS, Long CD, McPeek B. Preventable anesthesia mishaps: a study of human factors. Anesthesiology 1978; 49: 399-406. 4. Williamson JA, Webb RK, Pryor GL. Anaesthesia safety and the "critical incident" technique. Aust Clin Rev 1985; 5: 57-61. 5. Vincent CA. Research into medical accidents: a case of negligence. Br Med J 1989; 299: 1150-53.
VIEWPOINT Ethics of lung transplantation with live donors
Lung transplantation has become a viable treatment option for many patients with terminal pulmonary disease but the scarcity of donor lungs is a severe constraint. Hence many patients accepted into lung transplant programmes die while awaiting transplantation-as have nearly a quarter of patients admitted to the lung transplant programme in Toronto since 1984. The use of live donors might relieve donor lung scarcity. Live donors have been used in kidney transplantation since 1954 and in liver transplantation since 1989. In October, 1990, the Stanford group did the world’s first live-donor lung transplant.l,2 Ethical questions have been raised about this procedure.3 Recent experience with live-donor liver transplantation has demonstrated the potential benefit of analysing ethical considerations before embarking on innovative transplant procedures," which is what we plan to do here for lung transplantation. Risks and benefits
Recipients For the recipient the main disadvantage of a live donor is that only a lobe rather than the whole lung will be transplanted. Survival rates with transplants from live donors are unknown, but they are likely to be similar to those
obtained with cadaveric-donor single-lung transplantation, which are 76% at 6 months, 65% at 1 year, 58% after 2 years, and 53% after 3 years7 (the longest single-lung transplant survivor died of renal failure in March, 1990,6-25 years after the operation). Live-donor lung transplant also carries the risk that as the recipient grows the transplanted lobe will no longer be sufficient to support increasing functional requirements; however, the extent of this risk is likely to be similar for a cadaveric-lung transplant in a
growing recipient. The first of the four potential benefits to recipients is a reduction in numbers dying while waiting for a transplant since, unlike the case with kidney failure, where continued dialysis is an alternative to transplantation, there is no means of replacement therapy for end-stage lung disease except for transplantation. Secondly, by boosting donor numbers the ADDRESSES: Centre for Bioethics (L. R. Shaw, MA, P. A. Singer, MD), and Departments of Medicine (A. S. Slutsky, FRCPC, J. R. Maurer, FRCPC) and Surgery (J D. Miller, FRCSC, J. D. Puskas, MD, G. A. Patterson, FRCSC), University of Toronto, Toronto, Canada. Correspondence to Dr P. A. Singer, Centre for Bioethics, University of Toronto, Tanz Building, Room 303, 6 Queen’s Park Crescent West, Toronto, Canada M5S-1A8.
