Reports of Investigation
after ambulatory surgery
- a prospective study
Joanne Fortier MD FRCPC,*
Frances Chung MD FKCPC,~
Jun Su MD~f
Purpose: To determine the incidence, the reasons, and the predictive factors for unanticipated admission after
Methods: Preoperative, intraoperative, and postoperative data were collected prospectively on 15, 172 con-
secutive ambulatory surgical patients during a 32-month period. The data were built into a statistical model, and
predictive factors were identified and classified.
Results: The overall incidence of unanticipated admission was 1.42%. Admitted patients were more likely to be
older, male, and ASA status II or III. Duration of anaesthesia was longer, and surgery was more likely to be com-
pleted after 3 pm. Length of stay in the Postanaesthesia Care Unit and the Ambulatory Surgery Unit was longer.
Surgical reasons were cited in 38. 1% of admitted patients; anaesthesia-related reasons were cited in 25%; social
reasons accounted for 19.5%, and medical reasons for 17.2%. Ear, nose and throat (ENT) patients had the high-
est unanticipated admission rate (18.2%), followed by urology (4.8%) and chronic pain block (3.9%).
Gynaecological patients had the lowest rate (0.4%). Among the predictive factors found were male, ASA status II
and Ill, long duration of surgery, surgery finishing after 3 pm, postoperative bleeding, excessive pain, nausea and
vomiting, and excessive drowsiness or dizziness.
Conclusion: Earlier operating time for certain surgical procedures, screening for proper support at home, and
implementation of clinical pathways to deal aggressively with problems such as pain, nausea and vomiting should
decrease the incidence of unanticipated admission.
Objectif : DEterminer I'incidence, les raisons et les facteurs pr~dictifs d'une admission non pr~vue ~ la suite
d'une chirurgie ambulatoire.
M&hode : Les donn6es pr~op&atoires, intraop&atoires et postop&atoires ont &~ recueillies pendant 32 mois
de fa~son prospective aupr& de 15 172 patients cons&utifs, ~ la suite d'une chirurgie ambulatoire. Ces donn~es
ont 6t~ int~gr~es ~ un module statistique et les facteurs pr~dictifs ont ~t~ identifies et classifi(.~s.
R~sultats : I'incidence totale d'admission impr~vue &ait de 1,42 %. Les patients admis &aient plus susceptibles
d'&re ~g6s, de sexe m~le et d'&at ASA II ou III. La dur~e de ranesth&ie ~tait Iongue et plus susceptible de se
prolonger apr& 15 h. Les s~jours ~ I'unit~ des soins postanesth&iques et ~ I'unit~ de chirurgie ambulatoire &aient
prolong&. Les raisons chirurgicales ont ~t6 invoqu&s pour 38, 1% des patients admis ; les raisons reli~es
l'anesth~sie pour 25 % ; les raisons sociales pour 19,5 % et les raisons m~dicales pour 17,2 %. Les patients
d'oto-rhino-laryngologie (ORL) ont prEsent6 le plus haut taux d'admission non planifi6e (I 8,2 %) suivis des
patients d'urologie (4,8 %) et de ceux qui avaient recju un bloc thErapeutique pour douleurs chroniques (3,9 %).
Les patientes de gynEcologie avaient le taux le plus bas (0,4 %). Parmi les facteurs prEdictifs identifies, on a trou-
v~ le sexe m~le, l'&at ASA II et III, une chirurgie de dur~e prolong~e, une chirurgie qui se termine apt& 15 h,
les saignements postop&atoires, la douleur excessive, les naus&s et les vomissements, une grande somnolence
et des &ourdissements importants.
Conclusion : Le fait de proc~der plus t6t dans la joum6e & certaines interventions chirurgicales, le dEpistage
prEalable d'un soutien appropri~ pour le patient & domicile et la mise en application des moyens cliniques per-
mettant de s'occuper Energiquement des effets secondaires comme la douleur et, les naus~es et vomissements
devraient diminuer l'incidence de radmission non pr6vue.
From the Departments of Anaesthesia, CHUM-Hotel-Dieu Campus,* University of Montreal, and Toronto Hospital,]" University of
Toronto, Toronto, Ontario, Canada.
Address correspondence to: Dr. Frances Chung, Department of Anaesthesia, Toronto Hospital, Western Division, University of Toronto,
399 Bathurst St., Toronto, Ontario, Canada M5T 258. Phone: 416-603-5118; Fax: 416-603-6494.
Accepted for publication February 24, 1998.
CAN J ANAESTH 1998 / 45: 7 / pp 612-619
Fortier et aL: AMBULATORY SURGERY
care is a challenge. Unanticipated admission after ambu-
latory surgery is an indicator of quality care because it
concerns the basic goal of same-day discharge.
The incidence of unanticipated admission after
ambulatory surgery varies from 0.28% to 9.5%. 1-6
Predictive factors are general anaesthesia, emesis, lower
abdominal surgery, urological procedures, and anaes-
thesia duration of more than one hour; pain, bleeding,
more extensive surgery, and social reasons account for
most of the unanticipated admissions.
Because there has been no large prospective study
that examined the reasons and predictive factors for
unanticipated admission of ambulatory surgical
patients, we studied prospectively the incidence, caus-
es, and predictive factors for unanticipated admission
in our patient population.
MBULATORY surgery accounts for a large
and ever-increasing share of surgical proce-
dures. In ambulatory anaesthesia, quality
care is a primary concern, but defining good
With the approval of our institutional Ethics
Committee, data were prospectively collected on con-
secutive ambulatory surgical patients registered at the
Toronto Hospital, Western Division, a tertiary care
teaching hospital. No informed consent was required
for this study as there was no change in routine prac-
tice. The ambulatory surgery facilities consist of an
admissions area, four outpatient operating rooms, an
outpatient Postanaesthesia Care Unit (PACU), and an
Ambulatory Surgical Unit (ASU). This facility was
separate from the inpatient operating rooms. All
patients in the ambulatory surgical facility were dis-
charged on the same day as surgery. Overnight stay by
patients from the ambulatory programme was consid-
ered an unanticipated admission. Healthy ASA I
patients were not evaluated preoperatively in the
anaesthesia consultation clinic. Most patients with
ASA status II and all ASA status III were evaluated in
the anaesthesia consultation clinic preoperatively.
The following variables were studied prospectively: pre-
operative patient characteristics; intraoperative vari-
ables and adverse outcomes; and postoperative
variables and adverse outcomes. These variables were
documented in the anaesthesia, PACU, and ASU
records and on a standardized adverse outcome check-
off form. The anaesthetists completed the anaesthetic
record in a check-off format. Demographic data, pre-
operative medical illness, American Society of
Anesthesiology (ASA) status, duration of anaesthesia,
surgical procedure, and intraoperative management
(drugs, techniques, monitoring, etc.) were document-
ed in the anaesthetic record.
The surgical procedures were classified according to
the International Classification of Diseases Procedure
Code (ICD.9.CM) and recorded as one of nine groups:
ear, nose, throat and dental; general surgery; urology;
neurosurgery; gynaecology; plastic surgery; ophthal-
mology; orthopaedic; and chronic pain block. Types of
anaesthesia were classified into five groups: general
anaesthesia, monitored anaesthesia care, regional anaes-
thesia, local anaesthesia, and chronic pain block. Travel
time (reflecting the distance the patient needed to trav-
el to the hospital) was divided into areas inside or out-
side the metropolitan Toronto area, i.e., less than or
more than one hour. The time of completion of surgery
was classified as morning (8:00-11:59), afternoon
(12:00-2:59 pm), or late afternoon (> 3:00 pro).
Intraoperative and immediate postoperative adverse
outcomes, with printed concise definitions, were doc-
umented in a standardized adverse outcome check-off
form. Both anaesthetists and nurses were instructed in
the standardized definitions and recording of variables
before the start of the study. Intraoperatively, the
anaesthetists checked off adverse outcomes, and the
form accompanied the patient into the PACU and
ASU, where the nursing staff recorded adverse out-
comes in categories listed on the form. Postoperatively,
patients who had general or regional anaesthesia were
admitted to the PACU and then to the ASU. Most
patients with local anaesthesia or monitored anaesthe-
sia care went directly to the ASU. Medications given,
physiological variables, duration of stay in the PACU
and ASU, and discharge location were recorded in the
specifically designed PACU and ASU records. Patients
were discharged when they achieved a score of 9 on
the Post Anaesthesia Discharge Scoring System. 7
If the patient required admission, the nurses in the
PACU or the ASU checked off the reason and docu-
ment the details on the standardized adverse outcome
check-off form. The decision to admit the patients was
made by the surgeons or the anaesthetists in-charge.
The reason for unanticipated admission was classified
on the adverse outcome form into one of four main
groups: surgical reasons including pain, bleeding, posi-
five biopsy, misadventure, and more extensive surgery:
Anaesthesia reasons included nausea and vomiting,
dizziness, and somnolence: Medical reasons included
preexisting disease and various complications: Social
reasons included patient request, surgeon request, and
no available escort.
614 CANADIAN JOURNAL OF ANAESTHESIA
Charting was completed on discharge from the
ASU, and the data were reviewed systematically the
next day by a research assistant and an experienced
anaesthetist (FC). The data were coded and entered
into a Dbase III+ computer programme.
Data were analyzed employing the Statistical Analysis
System (SAS) version 6.11. Descriptive statistics in the
form of frequencies, means, standard deviations, and
percentages were calculated. For continuous variables,
independent t tests were used to test the differences
between the unanticipated admission group and the
same-day discharge group. For categorical variables,
the chi-square test of independence was used.
In order to identify predictive factors for unantici-
pated admission, a two-stage analysis was carried out.
First, univariable logistic regression models were used
to determine which individual factors were associated
with unanticipated admission. Second, all variables
with statistically significant association with unantici-
pated admission from the first step were simultane-
ously built into a multivariable logistic regression model.
The purpose of the second step was to get an adjust-
ed effect estimate for the different factors, controlling
for the potentially confounding effect of the other
variables in the model. ASA physical status was used as
a variable instead of the large number of preexisting
medical diseases. The type of anaesthesia was not
included in the model since in a clinical setting it is
dependent on the type of surgery. Odds ratios, their
95% confidence limits, and the corresponding P- val-
ues were reported for all significant predictive factors
from the second model. Because of the large data set,
in addition to a statistically significant P value of
<0.05, the odds ratio had to be either <0.9 or >1.1 to
be considered significant.
Of 15,179 ambulatory surgical patients, 215 (1.42%)
were admitted. Patients with unanticipated admission
were older and more likely to be male than female
(Table I). Patients of ASA status II and III were more
likely to be admitted than those of ASA status I. Body
mass index was lower in the unanticipated admission
group. In the same-day discharge and unanticipated
admission groups, there were similar numbers of
patients who lived more than one hour away.
The duration of anaesthesia was 35 min longer in
patients with unanticipated admission. The incidence
of unanticipated admission in patients whose surgical
procedures ended after 3 pm was higher (Figure 1 ). In
the unanticipated admission group, the length of stay
in the PACU was 50 min longer and in the ASU 68
min longer. Most patients (165/215) with unantici-
pated admission were discharged the next day; the
mean length of hospital stay was 1.4 days.
Reasons for unanticipated admission were divided
into four categories: surgical, anaesthesia-related,
medical, and social (Table II). Surgical reasons were
the most frequent (38.1%). Anaesthesia-related rea-
sons were cited in 25.1% of unanticipated admissions.
Social reasons accounted for 19.5% of unanticipated
admissions, and medical reasons were cited in 17.2%
of cases (Table II).
For the types of anaesthesia, the highest unantici-
pated admission rate was for chronic pain block (n =
153) at 3.9%, followed by local anaesthesia (n = 465)
at 2.8%. The unanticipated admission rate was 2.6%
for regional anaesthesia (n = 304), 1.6% for general
anaesthesia (n = 8,805), and 0.9% for monitored
anaesthesia care (n = 5,452). The incidence of unan-
8~0-1 ~:5~m 12:00.P.$g prn > 3.'00 pr
Time at completion of surgery
FIGURE 1 Incidence of unanticipated admission after various
times of completion of surgery.
FIGURE 2 Incidence of unanticipated admission after various
types of surgery. ENT/DEN, ear, nose and throat or dental;
URO, urology; Pain, chronic pain block; PLA, plastic surgery;
ORT, orthopaedic; GEN, general; NEU, neurosurgery; OPT,
ophthalmology; GYN, gynaecology.
Fortier et al.: AMBULATORY SURGERY
TABLE I Characteristics of patients studied
Variable Same-day discharge
Body mass index, kg-m 2
M 32 : F 68
46 • 21
2.5 • 0.5
ASA status I
g 1 hr
49.6 • 26
Anaesthesia duration, min
End of surgery < 3 PM
50.0 • 24
98.8 • 55
PACU duration, min
ASU duration, min
M 43.3 : F 56.7*
51 • 201"
2.4 • 0.5*
84.8 • 47*
100.4 • 68*
166.2 • 95*
Sex, ASA status, distance traveled, and end of surgery are expressed as percentages. Age, body mass index, anaesthesia duration, and
PACU and ASU duration are expressed as mean • SD. Significantly different from same-day discharge: *P < 0.001, ~'P < 0.002.
ticipated admission after chronic pain block (P <
0.001), local anaesthesia (P < 0.001), regional anaes-
thesia (P < 0.001) or general anaesthesia (P < 0.004)
was higher than that for monitored anaesthesia care.
The incidence of unanticipated admission by type
of surgical procedure is shown in Table III. The high-
est rate (18.2%) occurred with ENT surgery, followed
by urology (4.8%) and chronic pain block (3.9%).
Gynaecology had the lowest rate (0.4%) (Figure 2).
The caseload distribution (Table III) showed that
ophthalmology accounted for the most cases at 36%;
gynaecology, 34%; and orthopaedics, 17%.
First-stage analysis of predictive factors showed that
sex was a significant factor, with men more likely than
women to be admitted (Table IV). A lower body mass
index was significant. ASA physical status II and III
patients were admitted more than ASA physical status I
patients. Age and travel time of> one hour were not sig-
nificant. The presence of preoperative disease was
analysed with each factor. Heart disease, asthma, dia-
betes, hypertension, and hyperthyroidism/hypothy-
roidism were significant factors, but smoking and history
of chronic obstructive pulmonary disease were not.
When we compared surgical groups, the default used
was gynaecology, because it had the lowest admission
rate. All groups were considered significant but with
varied odds ratios. Patients with all types of anaesthe-
sia were admitted more often than patients with mon-
itored anaesthesia care. Duration of anaesthesia was a
significant factor, as was completion of surgery after 3
pm when compared with the morning (8 am - 12 pm)
or early afternoon (12 pm- 3 pm) groups. Postoperative
pain, nausea and vomiting, dizziness, drowsiness, and
especially bleeding were all significant predictors of
Further analysis and creation of statistical predictive
models for preoperative, intraoperative, and postoper-
ative periods confirmed that the following factors were
predictive (Table V). For preoperative factors, male
sex and ASA physical status II and III were found to
be significant. Intraoperative factors showed that cer-
tain surgical specialties - ENT, urology, plastic
surgery, orthopaedics, neurosurgery, and ophthalmol-
ogy - to be significant predictors. Longer duration of
anaesthesia was significant, and end of procedure after
3 pm was also a predictive factor. Postoperative pre-
dictive factors of unanticipated admission included
bleeding, severe pain, nausea and vomiting, excessive
drowsiness, and dizziness.
The overall unanticipated admission rate of 1.42% in
our study compares well with results from the other
large series of patients studied. ~,3-6,8,9 With Medicare
in Canada, admitting a patient to hospital does not
result in direct cost to the patient or potential reim-
bursement problems with insurance companies. Also,
our hospital is a large teaching centre with beds read-
ily available. These two factors might have boosted the
incidence of unanticipated admission, and the 20% of
admissions for social reasons reflect this. Therefore,
there is ample room for improvement in decreasing
the incidence of unaa~ticipated admission. Better pre-
operative screening and education of patients, family;
and surgeons can significantly reduce the incidence.
Nausea and vomiting accotmted for 14.4% of unan-
ticipated admissions, confirming the findings in previ-
ous studies, a,l~ During the study period, no routine
antiemetic prophylaxis was used.
ondansetron was not available as a routine treatment.
TABLE I I Reasons for unanticipated admision (215 patients)
CANADIAN JOURNAL OF ANAESTHESIA
SURGICAL: n ~ 82 (38.1%) ANAESTHETIC: n = 54 (25.1%)
Pain n = 26 (12.1%)
15 orthopaedic surgery
4 plastic surgery
2 peripheral nerve surgery
1 skin graft
1 ENT surgery
1 chronic pain block
Misadventure n = 13 (6.0%)
6 haemorrhage in ophthalmology
4 pneumothomx after lung biopsy
1 failed laparoscopy
1 laparotomy for bleeding D&C
1 spinal tap-dorsal column stimulator
More Extensive Surgery n = 7 (3.3%)
3 gynae-ectopic- malignancy
2 knee surgery
1 shoulder surgery
i hand surgery
Other n = 36 (16.7%)
5 postoperative care
1 unsuccessful nephrostomy tube
Nausea and Vomiting n ~ 31 (14.4%)
9 ophthalmology (1 strabismus)
8 orthopaedic surgery
6 ENT surgery
2 peripheral nerve
1 general surgery
1 plastic surgery
Somnolence n ~ 5 (2.3%)
3 post general anaesthesia
2 oversedation during local anaesthesia
Other n = 18 (8.4%)
7 block related (5 pain block, 1 eye block, 1 intravenous block)
4 possible aspiration
1 observation (malignant hyperthermia susceptible)
1 anaesthetist request
SOCIAL n = 42 (19.5%) MEDICAL n = 37 (17.2%)
Patient Request n = 13 (6.0%)
Surgeon Request n = 15 (7.0%)
No Escortn ~ 14 (6.5%)
Preexisting Disease n = 21 (9.8%)
3 low saturation and lung disease
2 sleep apnea + angina + diabetes
1 diabetes + hypertension
1 diabetes + dialysis
1 renal failure
Complications n = 14 (6.5%)
3 vagal reaction
2 low 0 2 saturation
1 myocardial infarction
1 heart failure
Other n ~ 2 (0.9%)
Fortier et al.: AMBULATORY SURGERY
TABLE III Incidence of unanticipated admission by surgical procedure
Surgical group Total no. of cases Unanticipated admission
ENT and dental n = 170 (1.1%)
Urology n = 231 (1.5%)
Chronic pain block n = 153 (1%)
General n = 431 (2.8%)
Plastic: n = 496 (3.3%)
skin graft mad other
Orthopaedic: = 2,548 (16.8%)
Neurosurgery: n = 412 (2.7%)
Ophthalmology: n = 5,510 (36.3%)
Gynaecology: n = 5,228 (34.4%)
*Other types are, for urology: cysto, turbt; general surgery: hernia, Hickmann line insertion/renaoval, lung biopsy; plastic surgery: repair
and reconstruction of skin, lipcctomy; orthopaedic: bunionectomy, bursectomy, muscle biopsy; ophthalmology: lens repositioning, eyelid
procedure, dacryocystorhinostomy, pterygium excision, keratotomy, scleral buckle, vitrectomy, conjunctival cyst excision.
Pain was also an important factor, accounting for
12.1% of the unanticipated admissions; 60% of these
were orthopaedic patients. In a previous study of 1,996
orthopaedic surgical patients, pain accounted for half
of the unanticipated admissions.14 Better management
of postoperative nausea and vomiting and pain would
dramatically decrease the incidence of unanticipated
In the category of social reasons for admission, the
differences between patient request, surgeon request,
and no escort are somewhat arbitrary, because patients
may ask the surgeon if they can stay overnight because
they are alone at home. Still, the percentage was high
and should decrease with better planning.
The single largest reason for admission was surgical,
with almost half of the patients admitted for bleeding
and almost as many for observation. Bleeding was a
predictive factor for admission, confirming results of
previous studies. 1,s,6,s,9 On our forms, one main rea-
son for unanticipated admission had to be chosen.
However, in clinical practice, a single factor may not
be important enough to warrant admission but, com-
CANADIAN JOURNAL OF ANAESTHESIA
TABLE IV Univariate logistic regression of unanticipated admission
Factor Pr (chi) OR Cl (lower) CI (upper)
Body mass index
Travel time > 1 hr
Type of surgery *
ENT and dental
Chronic pain block 0.0001
Type of anaesthesia t
Anaesthesia > 1 hr
Surgery end > 3 pm
0.1045 1.6 0.9 2.6
0.6831 1.1 0.8 1.6
Pr, probability; OK, odds ratio; CI, confidence interval.
* Type of surgery was compared with gynecological disease.
t Type of anaesthesia was compared with monitored anaesthesia
bined with other minor factors, may justify it. Future
studies should document multiple reasons for unantic-
To our knowledge, this is the first large prospective
study to determine predictive factors for unanticipated
admission and to distinguish among preoperative, intra-
operative, and postoperative factors. Rudkin et aL s did
a prospective study of 5,000 patients, but their empha-
sis was on preoperative screening. Maleness, ASA phys-
ical status, and surgery type were not found to be
significant predictive factors in previous studies. Gold et
al. 3 found general anaesthesia and age to be predictive
factors, whereas we did not. However, we found that
TABLE V Significant predictors from the multivariable logistic
regression of un~mticipated admission
ASA II and III
Surgery end > 3 pm 0.0001
Anaesthesia > 1 hr
Pr OR CI
1.5 - 2.8
0.5 - 0.9
15.1 - 58.1
2.1 - 10.5
2.5 - 7.4
1.6 - 9.6
1.1 - 3.4
1.8 - 3.7
1.6 - 3.1 0.0001
71.7 - 991.9
3.0 - 5.9
1.6 - 10.5
2.8 - 5.8
1.5 - 4.8
Pr, probability; OK, odds ratio; CI, confidence interval.
nausea and vomiting, urological procedures, and anaes-
thesia duration of more than one hour were predictive.
Twersky et al. 6 divided unanticipated admission into
three groups: avoidable, potentially avoidable, and
unavoidable. Women were found to be predictors of
avoidable unanticipated admission. In our study, male-
ness was a predictive factor of unanticipated admission.
We had a large gynaecology population, and this was
the surgical group with the lowest unanticipated admis-
sion rate; this may, in fact, have skewed the numbers of
unanticipated admission against men. The difference in
predictive factors in our study may be due to differences
in patient populations, caseloads, and the prospective
nature of the study.
Ear, nose and throat surgery was followed by many
admissions for bleeding and for social reasons. This
may reflect the reluctance of our ENT surgeons to
send their patients home after postoperative observa-
tion. Urology was almost
orthopaedics, plastic surgery, or neurosurgery to have
admissions. Other studies also found urology to have
a high unanticipated admission rate. ~,3,6,9 This is
attributed to the nature of urological procedures,
because bleeding and urinary retention are frequent
and unavoidable complications. Longer procedures (>
1 hr) reflecting more complicated procedures were
found to be twice as likely to result in admission.
Furthermore, when surgery ended after 3 pm, they
were 2.5 times more likely to result in unanticipated
admission. Freeman et aL H also found that ophthal-
mology surgery ending after 2 pm was a predictor of
twice as likely as
Fortier et al.: AMBULATORY SURGERY 619 Download full-text
Of the postoperative factors, bleeding was the
highest predictor. If bleeding is important enough to
be reported as an event, it is almost certain that cau-
tion will dictate overnight observation. Pain, drowsi-
ness, and nausea and vomiting had similar predictive
values. Because these are frequent postoperative symp-
toms, it is worthwhile to work toward preventing
them or to explore effective treatment. Implementation
of clinical pathways to deal aggressively with problems
such as pain, nausea and vomiting is necessary in each
ambulatory surgical centre. Education of anaes-
thetists, nurses and surgeons is essential in achieving
the goal of reducing the incidence of unanticipated
With the use of a multimodal approach of opioids,
NSAIDs, and local anaesthesia, the incidence of unan-
ticipated admission of patients undergoing ambulatory
laparoscopic cholecystectomy was 9.5%. is Nine of ten
patients can be discharged with considerable savings to
the cost of health care. Therefore, there is a possibility
that more extensive procedures should be considered
for ambulatory surgery. Then a higher incidence of
unanticipated admission should be accepted.
Although our overall incidence of unanticipated
admission of 1.42% is within the range of other stud-
ies, there are areas that can be targeted for improve-
ment. The goal is to ensure that resources are used in
an optimal fashion. It is less stressful for all involved if
the hospital stay is planned rather than perceived as a
complication. Education of ENT surgeons, proper
patient selection, and appropriate procedures will
reduce the incidence of admission of ENT patients.
We can use the high rates for urology to plan to have
beds available when those procedures are performed.
Prophylaxis of nausea and vomiting must be con-
sidered, at least in high-risk patients. Pain control can
certainly be improved. Scheduling should take into
consideration the complexity of the surgery, the
expected duration of surgery, and the time at the end
of surgery. When surgery is scheduled for the later
part of the day, it should be for those with the least
potential for complications or extended procedures.
There also appears to be great latitude regarding
escorts and postoperative care. Better education of
personnel involved with the postoperative phase is
needed. With proper identification of patients at risk,
better support can be arranged.
In conclusion, earlier operating times for certain
surgical procedures, screening for proper support at
home, and implementation of clinical pathways to deal
aggressively with problems such as pain, nausea and
vomiting should decrease the incidence of unantici-
1 Fancourt-Smith PF, Hornstein J, Jenkins LC. Hospital
admissions from the Surgical Day Care Centre of
Vancouver General Hospital 1977-1987. Can J"
Anaesth 1990; 37: 699-704.
2 Levin P, Stanziola A, Hand R. Postoperative hospital
retention following ambulatory surgery in a hospital-
based program. American College of Utilization
Review Physicians 1990; 5: 90--4.
3 Gold BS, Kitz DS, LeckyJI-I, NeuhausJM.
Unanticipated admission to the hospital following
ambulatory surgery. JAMA 1989; 262: 3008-10.
4 Johnson CD, Jarrett PEM. Admission to hospital after day
case surgery. Ann R Coil Surg Engl 1990; 72: 225-8.
5 Rudkin GE, Osborne GA, Doyle CE. Assessment and
selection of patients for day surgery in a public hospi-
tal. Med J Aust 1993; 158: 308-12.
6 Twersky RS, Abiona M, Thorne AC, et al. Admissions
following ambulatory surgery: outcome in seven urban
hospitals. Ambulatory Surgery 1995; 3: 141-6.
7 Chung FF, Chan VWS, Ong D. A postanaesthetic dis-
charge scoring system for home readiness after ambula-
tory surgery. Ambulatory Surgery 1993; 1: 189-93.
8 Osborne GA, Rudkin GE. Outcome after day-care
surgery in a major teaching hospital. Anaesth Intensive
Care 1993; 21: 822-7.
9 Biswas TK, Leary C. Postoperative hospital admission
from a day surgery unit: a seven-year retrospective sur-
vey. Anaesth Intensive Care 1992; 20: 147-50.
10 Meeks GR, Waller GA, Meydrech EF, Flautt FHJr.
Unscheduled hospital admission following ambulatory
gynecologic surgery. Obstet Gynecol 1992; 80: 446-50.
11 Freeman LN, Schachat AP, Manolio TA, Enger C.
Multivariate analysis of factors associated with
unplanned admission in 'outpatient' ophthalmic
surgery. Ophthalmic Surgery 1988; 19: 719-23.
12 Isenberg V, Apt L, Yamada S. Overnight admission of
outpatient strabismus patients. Ophthalmic Surgery
1990; 21: 540-3.
13 Meeks GR, Meydreeh El:, Bradford Tiff, Hollis RS.
Comparison of unscheduled hospital admission follow-
ing ambulatory operative laparoscopy at a teaching
hospital and a commtmity hospital. J Laparoendosc
Surg 1995; 5: 7-13.
14 Kinnard P, Lirette R. Outpatient orthopedic surgery: a
retrospective study of 1996 patients. Can J Surg 1991;
15 Miehaloliakou C, Chang F, Sharma S. Preoperative,
multimodal analgesia facilitates recovery after ambula-
tory laparoscopic cholecystectomy. Anesth Aa~alg 1996;