The Tao of Bao : A Randomised Controlled Trial Examining the Effect of Steamed Bun Consumption on Night-Call Inpatient Course and Mortality

The Tao of Bao: a Randomised Controlled Trial—Min Han Tan et al
255
The Tao of Bao: A Randomised Controlled Trial Examining the Effect of Steamed
Bun Consumption on Night-Call Inpatient Course and Mortality
Min-Han Tan,1MBBS, MRCP(UK), Ziying Lee,1 Beatrice Ng1, Eng Swen Sim1, Ying Ying Chua,1MBBS (Melb),
Mark Tien,1MBBS (Ire), Choon Jin Ooi,2,3MBBS, FRCP (Edin), FAMS
Abstract
Background: Medical superstitions remain prevalent in today’s stressful and technology
driven healthcare environment. These irrational beliefs commonly involve night calls, which are
periods of volatile workload. In Singapore and Hong Kong, it is commonly held that consumption
of steamed buns (“bao”) by on-call physicians is associated with increased patient admissions and
mortality, due to a homonymous interpretation of the word “bao” in dialect. Materials and
Methods: A prospective unblinded randomised controlled trial with a permuted block
randomisation design was performed on weekdays over 6 weeks. Steamed buns or control food
were offered to the internal medicine night-call team of a tertiary-care hospital on a nightly basis.
Information on admissions and mortality was collected from the hospital electronic database.
Data on sleep patterns and shift duration were obtained by interview. Results: There were no
significant differences in the median number of hours slept on days on “bao” administration
versus “control” intervention (2 ± median absolute variation of 1.5 h vs 2 ± 1.5 h, P = 0.30) or in
the number of hours spent in the hospital (30.8 ± 1.9 h vs 30.5 ± 2.2 h, P = 0.09). There were no
significant differences in the median number of general ward admissions per night (n = 73 ± 6
versus 71 ± 7 admissions, P = 0.35), monitored care unit admissions (4 ± 1.5 vs 4 ± 1.5 admissions,
P = 0.65) or inpatient mortality (2 ± 1.5 vs 2 ± 1.5 deaths per night, P = 0.47). Conclusion: The
consumption of steamed buns (“bao”) has no effect on inpatient admissions, mortality, or sleep
duration on call. Regardless, our results indicate that the night call in Singapore remains a
challenge in terms of workload and shift duration.
Ann Acad Med Singapore 2008;37:255-7
Key words: Medical superstition, Night duty, Occupational medicine, Sleep deprivation
Introduction
In Singapore and Hong Kong, the best known
The modern world remains full of superstition. Often,
superstitition in the healthcare community is the admonition
despite their better judgement, intelligent rational people
against consumption of steamed buns (“bao”) during a
adopt rituals and strategies to cope with anxiety and
night call, lest the night-call staff be occupied with morbidity
uncertainty. The medical profession is not exempt.
and mortality. Such a belief has been commonly accounted
Several studies have investigated the prevalence and
for by the homonymous interpretations of the term “bao”,
nature of such beliefs in the medical community. Doctors
which may be interpreted as “to have the lion’s share” of
and nurses in the Western world own to beliefs that the full
admissions or as a “shroud” for corpses. A recent survey of
moon affects workload, trauma and emergency room
healthcare personnel in Singapore indicated that
admissions.1-4 A majority of staff surveyed in some studies
46% of responders actively avoided these steamed buns
report that certain individuals are associated with “black
during night calls.7
clouds”, attracting an excess of work.5 The belief that
This study was undertaken to determine the effect of
stating that a night call was quiet (“the Q word”) would
steamed bun (“bao”) consumption by night-call staff on
result in a supply of additional work is also common,
patient admissions and mortality, as well as total sleep
though no difference has been demonstrated in a randomised
duration each night.
controlled study in the United States.6
1 Department of Medical Oncology, National Cancer Centre, Singapore
2 Department of Medical Manpower, Singapore Health Services, Singapore
3 Department of Gastroenterology, Singapore General Hospital, Singapore
Address for Correspondence: Dr Tan Min-Han, Department of Medical Oncology, National Cancer Centre, 11 Hospital Drive, Singapore 169610.
Email: tan.min.han@nccs.com.sg
March 2008, Vol. 37 No. 3

---

256
The Tao of Bao: a Randomised Controlled Trial—Min Han Tan et al
Materials and Methods
Consent was sought from a total number of 144 physicians,
A prospective randomised controlled unblinded trial was
and 137 provided informed consent. Of the 7 physicians
conducted in the internal medicine service of a 1600-bed
who declined participation, 5 cited concerns about a
tertiary care hospital. The study extended over a period of
potentially heavier on-call workload, 1 claimed food allergy,
30 weekdays (excluding Saturdays and Sundays) extending
and the last physician cited an unspecified religious
over 6 weeks from February to April 2007. There was no
injunction against participation. Median compliance
public holiday during this period. Sample size calculations
recorded on a daily basis was good and noted to be
were based on a power of 0.8, ? = 0.05, and a projected
equivalent in both arms of the study (86 ± 10% and 86% ±
absolute difference of 20 admissions and sigma of 20, with
2% on both A and B calls, P = 0.7).
double-sided testing and continuity correction. Permuted
There was no statistical difference between the median
block randomisation with a 1:1 ratio, using the day of the
number of total admissions on A calls (n = 73 ± 6) and B
week, was used to assign each day randomly to consumption
calls (n = 71 ± 7) (P = 0.35). Similarly, no difference was
of steamed buns (A calls) or control food (B calls), to
observed between the number of admissions to the monitored
ensure group and periodic balance due to the expected
units on A and B calls (4 ± 1.5 versus 4 ± 1.5 admissions
variation of admission load over different weekdays. Primary
respectively, P = 0.55). No difference in the number of
end points were night-call inpatient admissions, monitored
deaths was observed, with a median of 2 ± 1.5 deaths (A
unit (ICU/high dependency) admissions and overall
calls) and 2 ± 1.5 deaths (B calls) (P = 0.54).
inpatient mortality. In view of the multiracial and multi-
The distribution of the number of hours slept per call, as
religious nature of the night-call staff, who hail from
well as the call shift duration is presented in Table 1. There
countries as diverse as Malaysia, Myanmar, India,
is no significant difference observed between A and B
Philippines, China, Pakistan, and the United Kingdom, the
calls. All subset analyses were pre-planned.
choice of filling for the steamed bun was minced vegetables,
and therefore suitably halal and vegetarian. Water was
10
supplied.
Fourteen night-call staff were on call each night with
8
responsibility for inpatient care, inclusive of the medical
intensive care unit: 5 house officers, 7 night call medical
6
officers, and 2 registrars responsible for inpatient wards.
Informed consent was obtained from each subject. Food
4
was administered by individual delivery to the subject
eekday night-calls (n)
concerned in the afternoon preceding the call. Data on
W
2
ward admissions and mortality between 4 pm and 8 am of
the following day were collected from the hospital electronic
medical record database. Data from each physician about
0
compliance and on-call sleep duration were also collected
55
60
65
70
75
80
85
90
the day following the call. All data were analysed in R
Night-call admissions (n)
version 2.2.1. Values in brackets represent the median
Fig. 1. Histogram of weekday night-calls by admission number.
absolute deviation. Non-parametric testing was performed
in all situations (Wilcoxon’s test). A P value of <0.05 was
100
regarded as significant. Missing observations were excluded
from the analysis.
80
The study was approved by the hospital institutional
review board (IRB).
60
Results
The admission load over the period of the study is
40
presented in Figure 1. Variation between the admissions by
day of the week was observed (P = 0.04 by one-way
20
Number of night-call admissions
analysis of variance), with heaviest patient loads seen on
Mondays (Fig. 2). As mentioned, permuted block
0
randomisation design was used to account for this anticipated
Monday
Tuesday Wednesday Thursday
Friday
variation.
Fig. 2. Mean number of night-call admissions by day of the week.
Annals Academy of Medicine

---

The Tao of Bao: a Randomised Controlled Trial—Min Han Tan et al
257
Table 1. The distribution of the Number of Hours Slept Per Call and the Call
Interestingly, it was noted that the highest ranked
Shift Duration
physicians with the heaviest responsibilities both on- and
A calls
B calls
P value
off-call, the registrars, spent the longest total duration per
call in the hospital, extending to a mean of 33 hours. The
Median number of
All doctors
2 ± 1.5
2 ± 1.5
0.30
underlying reason is multifactorial, but presumably related
hours slept (h)
Registrars
3.0 ± 2.2
3.0 ± 2.9
0.69
primarily to service demands and training requirements.
Medical Officers
2.5 ± 1.1
2.5 ± 1.5
0.71
House Officers
1.5 ± 1.5
1.5 ± 1.1
0.62
Conclusions
Median call shift
All doctors
30.8 ± 1.9
30.5 ± 2.2
0.09
This study demonstrates that steamed buns (“bao”) have
duration (h)
Registrars
33.8 ± 1.7
34 ± 2.2
0.60
no effect on night-call admissions or mortality, contrary to
Medical Officers
30.5 ± 1.6
29.9 ± 1.4 0.053
medical superstition in Singapore and Hong Kong. At the
House Officers
30.7 ± 1.6
30.5 ± 1.5
0.79
same time, this study points to the need to address the issue
Variation is represented with the median absolute deviation.
of workload and sleep deprivation among physicians.
Acknowledgements:
Discussion
The authors would like to thank the 46% of night-call physicians of the
This study conclusively demonstrates that steamed buns
Department of Internal Medicine at the Singapore General Hospital, who
despite their better judgment, agreed to participate in this study. This study
(“bao”) do not have an effect on inpatient admissions or
was sponsored by contributions by the Junior Doctor Welfare Committee
mortality, contrary to medical and nursing superstition in
and call allowances of the corresponding author.
Singapore and Hong Kong. The sadly rational result of this
Conflict of Interest Statement:
study is consistent with the conclusions of most studies
No financial interests in any food or beverage companies by the authors are
designed to examine the basis of irrational beliefs within
declared. Though eagerly sought, no pharmaceutical sponsorship was
and without the medical and nursing community.
forthcoming.
The weaknesses of this study are unlikely to overtly
compromise the results. The study was designed with
intervention based on days rather than individuals, resulting
in a reduced granularity in analysis. A short response
duration limited to several hours could not be excluded,
given the study design. The choice of spiced vegetable
REFERENCES
1. Mathew VM, Lindesay J, Shanmuganathan N, Eapen V. Attempted
filling rather than the more typical barbecued pork does
suicide and the lunar cycle. Psychol Rep 1991;68(3 Pt 1):927-30.
mean that a more inimical effect of the latter cannot be
2. Nunez S, Perez Mendez L, Aguirre-Jaime A. Moon cycles and violent
excluded.
behaviours: myth or fact? Eur J Emerg Med 2002;9:127-30.
3. Stair T. Lunar cycles and emergency-room visits. N Engl J Med
The results of this study need not be limited to an
1978;298:1318-9.
interpretation of medical superstition. Indeed, what may
4. Allen JL. Emergency admissions and lunar cycles. J Emerg
potentially be more important is the insight provided into
Nurs 1986;12:85-8.
the current state of physician welfare in Singapore. Whether
5. Tanz RR, Charrow J. Black clouds. Work load, sleep, and
resident reputation. Am J Dis Child 1993;147:579-84.
medical staff may continue to function optimally and care
6. Ahn A, Nallamothu BK, Saint S. “We’re jinxed” – are residents’ fears of
for patients after a potentially uninterrupted 30 hours of
being jinxed during an on-call day founded? Am J Med 2002;112:504.
work is a difficult question for doctors, healthcare
7. Lim EC, VM Oh, Quek AM, Seet RC. Take a bao if you are not
administrators and patients. There are data suggesting an
superstitious. Ann Acad Med Singapore 2007;36:217-20.
increase in errors, percutaneous needle injuries, adverse
8. Barger LK, Ayas NT, Cade BE, Cronin JW, Rosner B, Speizer FE, et al.
Impact of extended-duration shifts on medical errors, adverse events,
events and even driving accidents among sleep-deprived
and attentional failures. PLoS Med 2006;3:e487.
junior doctors.8-12 A questionnaire-based study in 2004
9. Lockley SW, Landrigan CP, Barger LK, Czeisler CA. When policy meets
surveying sleep patterns of house officers in Singapore has
physiology: the challenge of reducing resident work hours. Clin Orthop
raised similar concerns as we do here.13
Relat Res 2006;449:116-27.
10. Ayas NT, Barger LK, Cade BE, Hashimoto DM, Rosner B, Cronin JW,
The compromise between training, clinical service,
et al., Extended work duration and the risk of self-reported percutaneous
manpower and error rates is a very real issue that should be
injuries in interns. JAMA 2006;296:1055-62.
examined in a rational fashion driven by data. While
11. Chow KM, Szeto CC, Chan MH, Lui SF. Near-miss errors in laboratory
blood test requests by interns. QJM 2005;98:753-6.
continuous on-call shifts in the Western world in excess of
12. Barger LK, Cade BE, Ayas NT, Cronin JW, Rosner B, Speizer FE, et al.
24 hours and even 36 hours remain common and are
Extended work shifts and the risk of motor vehicle crashes among
justified by many senior physicians on the grounds of
interns. N Engl J Med 2005;352:125-34.
training, there is a growing voice that the likely resulting
13. Puvanendran K, Venkatramani J, Jain A, Farid M. Sleep deprivation in
junior doctors – house officers in Singapore. Ind Health 2005;43:129-32.
error rates are unacceptable to society at large.
March 2008, Vol. 37 No. 3

---