Provider Attitudes About Gaining Consent for
T. YEE KHONG, MD, DEBORAH TURNBULL, PhD, AND ALAN STAPLES, MEc
Objective: To examine the attitudes of neonatologists, obste-
tricians, midwives, and neonatal nurses toward perinatal
autopsy and survey physicians about whom they perceive
influence women’s decisions on autopsy consent.
Methods: A postal survey that incorporated a question-
naire of eight fictitious case scenarios and combined three
factors (confidence of antemortem diagnosis, intention to
have future pregnancy, and parental attitude toward au-
topsy) in various permutations was sent to various Austra-
lian physicians and nurses (all consultant neonatologists
working in neonatal intensive care units and a sample of
consultant obstetricians, midwives, and neonatal nurses in
level III maternity hospitals). Respondents were asked to
rate how likely they were to seek consent for or suggest
autopsies on a seven-point Likert scale (1 ? ? ? certainly will
not, 7 ? ? ? certainly will). Interactions between factors and
respondents were measured by analysis of variance, and
differences were compared using Mann-Whitney U, ?2, and
generalized estimating equation tests.
Results: The overall response rate was 70% (neonatologists
57%, obstetricians 62%, midwives 77%, and neonatal nurses
75%). Neonatologists (median score 7, interquartile range 7,
7) were more likely to ask for autopsies than neonatal nurses
(5; 2, 6) (P < < < .001), as were obstetricians (7; 7, 7) compared
with midwives (6; 3, 7) (P < < < .001). Physicians rated midwives
and neonatal nurses as having some to substantial influence
on mothers’ decisions about consent for autopsy.
Conclusion: Physicians are not averse to seeking consent
for perinatal autopsies. Midwives and nurses are influenced
by the three factors studied, which might negatively influ-
ence the consent rate for perinatal autopsies. Intervention
strategies aimed at changing nurses’ attitudes should be
considered. (Obstet Gynecol 2001;97:994–8.
American College of Obstetricians and Gynecologists.)
© 2001 by The
Perinatal autopsy rates are slowly decreasing and, in
some centers, have fallen below the recommended 75%
level.1–4This decline is disturbing because examination
after neonatal death, stillbirth, or termination of preg-
nancy because of antenatally diagnosed conditions is
important for determining cause of death and effects of
treatment, auditing diagnoses, and understanding dis-
ease mechanisms. Such examinations also might help
with counseling for future pregnancies and with par-
ents’ grieving process.1,2,5–7Epidemiologic data and
public health policies can be compromised by flawed
data from death certification without autopsies,2,8,9and
it has been suggested that high numbers of perinatal
autopsies are necessary for maintenance of skills of
Studies of adult autopsies indicate that failure to seek
consent or to counsel appropriately is a major factor in
the decline of adult autopsy rates from 70% to 80% to
less than 20%.11Mothers and their partners are as likely
to seek advice about autopsies after perinatal losses
from medical as from nursing staff because of the
rapport between mothers and nursing attendants in
delivery wards or neonatal intensive care units.12,13
Thus, our aim was to determine whether midwives,
neonatal nurses, neonatologists, and obstetricians dif-
fered in their attitudes toward factors that might influ-
ence autopsy consent. This information is necessary for
identification of target groups should intervention strat-
egies be considered. We also surveyed physicians to see
how much influence they thought other health profes-
sionals had on women and partners in their decisions
on autopsy consent.
Materials and Methods
From the Australian and New Zealand Neonatal Net-
work, a registry of all accredited neonatologists, we
identified consultant neonatologists from Australia’s 23
NICUs. We used a cluster sample method to identify
obstetricians, neonatal nurses, and midwives. To gather
From the Departments of Obstetrics and Gynaecology and General
Practice and Psychology, University of Adelaide; and Department of
Histopathology, and Division of Paediatric Medicine, Women’s and
Children’s Hospital, Adelaide, Australia.
Supported by the Adelaide Women’s and Children’s Hospital Perina-
tal Pathology Trust Fund. The authors thank Nicole Pratt for additional
statistical assistance. Copies of the questionnaires and the analysis of
variance tables are available from the corresponding author on request.
Obstetrics & Gynecology
respondents from other health professional categories,
we selected one hospital with a NICU from each state or
territory. There was only one such hospital in two states
or one territory (Tasmania, Western Australia, and
Australian Capital Territory), whereas in the remaining
four states we selected two with the most (New South
Wales and Victoria) and two with the second-most
(Queensland and South Australia) deliveries for each
state. We invited all consultant obstetricians in those
hospitals to participate, as well as all midwives in the
delivery suite and birthing centers, and all neonatal
nurses in the NICUs on particular days (24-hour peri-
ods). A different day was allocated for sampling each of
the seven hospitals to reduce potential bias. Individual
covering letters and questionnaires were mailed or
distributed through heads of departments between Au-
gust and October 1998.
The covering letter to physicians explained that the
survey was intended “to study the factors influencing
clinicians in seeking or not seeking consent for an
autopsy” and asked the respondents’ “likelihood of
seeking consent for an autopsy” for each of eight case
scenarios. The covering letter to neonatal nurses and
midwives explained that the survey was intended “to
study your reaction . . . if the woman or her partner
asked for your opinion of whether a postmortem exam-
ination should be performed” and asked the respon-
dents’ “likelihood of suggesting an autopsy.”
Questionnaires incorporated eight fictitious case sce-
narios that combined positive and negative values of
three factors (confidence of antemortem diagnosis, pa-
rental attitude toward autopsy, and parental intention
to have future pregnancy). Those factors were selected
by a panel of obstetricians, a neonatologist, a perinatal
pathologist, and a psychologist. We field-tested the
questionnaires on a neonatologist, a perinatal obstetri-
cian, a general practitioner, neonatal nurses, and mid-
wives who did not participate further in the study. We
designed the scenarios so that practitioners were likely
to have encountered them clinically. Accordingly, sce-
narios were identical for obstetricians and midwives
but different from those for neonatologists and neonatal
nurses, who shared common scenarios. Below is an
example of a scenario for obstetricians or midwives that
incorporates negative value for confidence of antemor-
tem diagnosis, negative value for parental attitude
toward autopsy, and positive value for parental inten-
tion to have future pregnancy:
A 26-year-old gravida 2, para 1 nonsmoker for whom
induced labor is scheduled for the next day presents
after term with absent fetal movements for the previous
24 hours. After induced labor, she is delivered of a
slightly macerated, normally formed male stillbirth
weighing 3336 g and a normal placenta with some gritty
calcification weighing 527 g. Cause of death is unclear.
The couple are keen to have another child. They do not
seem sure an autopsy will be helpful and appear not
keen to request one.
Below is an example of a scenario for neonatologists or
for neonatal nurses, showing positive value for confi-
dence of antemortem diagnosis, positive value for pa-
rental attitude toward autopsy, and positive value for
parental intention to have future pregnancy:
A 26-year-old gravida 2, para 1 woman ruptured her
membranes at 18 weeks’ gestation. Oligohydramnios
resulted, but the couple opted to continue the preg-
nancy. Spontaneous labor occurred at 24 weeks. The
prognosis was discussed with the parents, and it was
decided that the degree of resuscitation would depend
on the physical state of the infant at delivery. After a
short labor, a female infant weighing 465 g was deliv-
ered. Apgar scores were 1 and 3 at 1 and 5 minutes,
respectively, with poor attempts at respiration. In view
of the obvious pulmonary hypoplasia, active resuscita-
tion was not attempted, and the infant died peacefully
in the parents’ arms. The parents are keen to have
another pregnancy. They appear keen to have an au-
For each of the eight scenarios, we asked the respon-
dents how likely they were to seek, in the case of
physicians, or to suggest, in the case of nurses, autopsy
consent by rating a seven-point Likert scale (1 ? cer-
tainly will not, 7 ? certainly will). We also collected
additional personal data related to status, years since
professional qualification, and gender. We also asked
consultant neonatologists and obstetricians to rate on a
four-point Likert scale, 1 being none and 4 being much,
how likely various health professionals were to influ-
ence parents’ decision making on consent for autopsy.
A reply-paid envelope was attached to each question-
naire. All questionnaires were anonymous and, accord-
ingly, there was no follow-up of nonrespondents.
We analyzed interactions of the three factors and
respondents by a nonparametric analysis of variance for
neonatologists, obstetricians, and nurses (midwives and
neonatal nurses). Where appropriate, we used ?2and
Mann-Whitney tests for associations between personal
demographic details and responses. To compare re-
sponses between neonatologists and neonatal nurses
and between obstetricians and midwives, we did the
sum of each subject’s responses to the eight questions,
calculated to yield a total score with a maximum of 56
points. The resulting scores were log-transformed so
distribution of scores was normalized. A generalized
estimating equation with robust standard errors was
used to account for the clustering effect on hospitals to
VOL. 97, NO. 6, JUNE 2001
Khong et al
Perinatal Autopsy Attitudes
obtain estimates of the average total score.14Data were
analyzed using the GENMOD procedure in SAS (SAS
Inc., Cary, NC).
The overall response rate to the questionnaire was 70%
(neonatologists 65 of 114, 57%; obstetricians 56 of 90,
62%; midwives 133 of 172, 77%; and neonatal nurses 194
of 258, 75%). Table 1 gives the median and mean scores
for all scenarios. The response scores were different
between obstetricians and midwives (P ? .001) and
between neonatologists and neonatal nurses (P ? .001),
even after adjusting for cluster effects using the gener-
alized estimating equation model.
For one scenario that had a positive value for confi-
dence of diagnosis and negative values for parental
attitude toward autopsy and intention to have a future
pregnancy, additional information about pregnant
women could have influenced responses of midwives.
However, the median score of midwives for the remain-
ing seven scenarios was 6 (interquartile range 4–7),
which indicated that the erroneously presented sce-
nario did not bias the midwives’ attitudes toward
perinatal autopsy. That scenario had the least support
from health providers, which strengthened the argu-
ment that the error in the scenario did not alter the
overall outcomes for all eight scenarios for midwives.
Confidence of diagnosis was the most influential
single factor for neonatologists, neonatal nurses, and
midwives, but irrelevant on its own to obstetricians.
Parental desire for autopsy, in isolation, was influential
to all four groups. Parental intention for future preg-
nancy, in isolation, was of no consequence to neonatol-
ogists, but important to obstetricians and a minor factor
for neonatal nurses and midwives. Although intention
for future pregnancy and parental desire for autopsy
were important for obstetricians, no response to any
factor in isolation differed from the mean obstetrician
response (analysis of variance tables not shown but
available from the authors). The scenario that elicited
the least support for an autopsy from all four profes-
sional groups was that in which the diagnosis was clear,
the parents were not keen for an autopsy, and no future
pregnancy was planned. The obverse scenario, in which
there was uncertain diagnosis, desire for autopsy, and
desire for future pregnancy, prompted the highest
mean response from all groups.
Neonatal nurses with more than 10 years’ experience
(median 5, interquartile range 2–7) were more inclined
to suggest autopsy than those with less experience
(median 4, interquartile range 2–6, Mann Whitney U
test z ? 2.53, P ? .01), but no effect of experience was
found with the other three professional groups. No
effect of gender of the health professional on their
attitude was found. Obstetricians and neonatologists
rated other health professionals, with the exception of
interns and residents in neonatal units, as somewhat or
substantially influential in parents’ decisions about con-
senting to autopsy (Table 2).
Seeking permission for a perinatal postmortem exami-
nation is difficult, and many factors influence the atten-
dant’s decision to ask.13We did not examine some
factors, such as parental ethnicity, parity, gestational
age at delivery, length of stay in NICU, possible litiga-
tion, availability of perinatal pathologist expertise, and
reimbursement costs of autopsy because they have been
the subject of previous research.15–18Perinatal autopsies
also are free to parents, and perinatal pathologic exper-
tise is available in all hospitals with neonatal intensive
care units in Australia. Instead, we studied attitudes of
hospital personnel because of a dearth of research in
that area, and we used a novel approach to study
attitudes of physicians, midwives, and neonatal nurses
toward perinatal autopsy. That method of fictitious
histories to investigate influences of various decision
factors has two advantages: by allowing respondents to
show their attitudes in the form of role enactment, the
Table 1. Likelihood of Seeking Consent (Physicians) or
Suggesting (Nurses) Autopsy
* Statistically different (P ? .001).
†Statistically different (P ? .001).
Table 2. Perceived Influence of Trainee Physicians and
Paramedical Staff on Women and Partners
Midwife or nurse 52
* n ? Number of obstetricians or neonatologists who rated for each
of the perceived source of influence: 0 ? none, 1 ? not much, 2 ?
some, and 3 ? substantial influence.
996Khong et al
Perinatal Autopsy AttitudesObstetrics & Gynecology
gap between self-described and actual behavior is re-
duced; and the interactions between those factors can be
studied.19–21A lengthy questionnaire to accommodate
the other additional factors in various permutations of
their positive and negative values could have resulted
in a lower reply rate. The resultant overall response rate
was satisfactory for postal questionnaires of that nature
without follow-up of nonrespondents and provides a
valid idea of the prevalent attitudes among Australian
Contrary to published data that indicated negative
attitudes of clinicians toward perinatal autopsies,2,15,16
this survey showed that neonatologists and obstetri-
cians were not averse to seeking consent for autopsy,
and generally were not influenced by the factors stud-
ied. It is possible that social desirability bias might exert
a small but pervasive effect and lead to a response set,22
as shown by physicians’ scores, but we negated that
potential bias by seeking anonymous replies to ques-
tionnaires and examining attitudes of physicians, mid-
wives, and nurses at the same time. Thus, the determi-
nant for the perinatal autopsy rate would appear to be
failure to grant consent by the parents rather than
failure to seek consent by the medical staff. Neonatal
care is centralized in specialist centers in Australia, but
it would be of interest to conduct a similar study of the
attitudes of obstetricians in nonspecialist centers.
The nurses’ lower mean scores indicate that they
were influenced by the factors studied, which might
affect consent giving negatively. In the delivery ward or
the NICU, women and their partners often form a
rapport with nursing staff because of extended continu-
ity of care and might seek advice or counsel regarding
autopsy.12,13That observation is corroborated by the
high rating given the influence of midwives and neo-
natal nurses by the consultant obstetricians and neona-
tologists. Although midwives and neonatal nurses
might not see their roles as seeking consent, they might
nevertheless face negligence claims when they do not
communicate patient concerns to primary providers,
and they must be knowledgeable about their duties in
the informed consent process.23Indeed, with changes in
the physician–nurse relationships,24nurses and mid-
wives increasingly might obtain explicit verbal and
Junior medical staff was perceived by physicians as
having some influence on women and their partners,
but only the senior registrar of the junior physician
grades achieved a mean rating higher than the nurses.
We did not study the attitudes of junior medical staff;
therefore cannot say whether their influence likely was
positive or negative.
This study suggests that strategies should be imple-
mented to emphasize the value of perinatal postmortem
examination to nursing staff. Preferably, such strategies
would be broad-based; for example, the difference be-
tween neonatal nurses with more than 10 years’ expe-
rience and those with less is probably too small in real
terms to warrant targeting specific sociodemographic
groups. Interventions that are implemented need to
acknowledge the growing consensus that diverse strat-
egies are necessary and that the strategies should rely
on more than passive dissemination of information.26,27
1. Landers S, MacPherson T. Prevalence of the neonatal autopsy: A
report of the study group for complications of perinatal care.
Pediatr Pathol Lab Med 1995;15:539–45.
2. Cartlidge PHT, Dawson AT, Stewart JH, Vujanic GM. Value and
quality of perinatal and infant postmortem examinations: Cohort
analysis of 400 consecutive deaths. BMJ 1995;310:155–8.
3. Khong TY. A review of perinatal autopsy rates worldwide, 1960s
to 1990s. Paediatr Perinat Epidemiol 1996;10:97–105.
4. Joint Working Party of the Royal College of Obstetricians and
Gynaecologists and the Royal College of Pathologists. Report on
fetal and perinatal pathology. London: Royal College of Patholo-
5. Saller DN, Lesser KB, Harrel U, Rogers BB, Oyer CE. The clinical
utility of the perinatal autopsy. JAMA 1995;273:663–5.
6. Faye-Peterson OA, Guinn DA, Wenstrom KD. Value of perinatal
autopsy. Obstet Gynecol 1999;94:915–20.
7. Rushton DI. Should perinatal post mortems be carried out by
specialist pathologists? Br J Obstet Gynaecol 1995;102:182–5.
8. Kirby RS. The coding of underlying cause of death from fetal death
certificates: Issues and policy considerations. Am J Public Health
9. Duley LMM. A validation of underlying cause of death, as
recorded by clinicians on stillbirth and neonatal death certificates.
Br J Obstet Gynaecol 1986;93:1233–5.
10. Doyle LW. Editorial: Effects of perinatal necropsy on counselling.
11. Joint Working Party of the Royal College of Pathologists, Royal
College of Physicians, and Royal College of Surgeons. Autopsy
and audit. London: Royal College of Pathologists, 1991.
12. Khong TY. Improving perinatal autopsy rates: Who is counseling
bereaved parents for autopsy consent? Birth 1997;24:55–7.
13. Chiswick M. Perinatal and infant postmortem examination: Diffi-
cult to ask but potentially valuable. BMJ 1995;310:141–2.
14. Liang KY, Zeger SL. Longitudinal data analysis using generalized
linear models. Biometrika 1986;73:13–22.
15. VanMarter LJ, Taylor F, Epstein MF. Parental and physician-
related determinants of consent for neonatal autopsy. Am J Dis
16. Cottreau C, McIntyre L, Favara B. Professional attitudes toward
the autopsy. A survey of clinicians and pathologists. Am J Clin
17. Maniscalco WM, Clarke TA. Factors influencing neonatal autopsy
rate. Am J Dis Child 1982;136:781–4.
18. Landers S, Kirby R, Harvey B, Langston C. Characteristics of
infants who undergo neonatal autopsy. J Perinatol 1994;14:204–7.
19. Oppenheim AN. Questionnaire design and attitude measurement.
London: Heinemann, 1984.
VOL. 97, NO. 6, JUNE 2001
Khong et al
Perinatal Autopsy Attitudes
20. Elstein A, Shulman LS, Sprafka SA. Medical problem solving, an Download full-text
analysis of clinical reasoning. Cambridge, MA: Harvard University
21. Start RD, Hector-Taylor MJ, Cotton DWK, Startup M, Parsons MA,
Kennedy A. Factors which influence necropsy requests: A psycho-
logical approach. J Clin Pathol 1992;45:254–7.
22. Bowling A. Research methods in health. Investigating health and
health services. Buckingham: Open University Press, 1997.
23. Klepatsky A, Mahlmeister L. Consent and informed consent in
perinatal and neonatal settings. J Perinat Neonatal Nurs 1997;11:
24. Salvage J, Smith R. Doctors and nurses: Doing it differently. BMJ
25. Scholefield HA, Viney C, Evans J. Expanding practice and obtain-
ing consent. Prof Nurse 1997;13:12–6.
26. Rubin GL, Frommer MS, Vincent NC, Phillips PA, Leeder SR.
Getting new evidence into medicine. Med J Aust 2000;172:180–3.
27. Oxman AD, Thomson MA, Davis DA, Haynes RB. No magic
bullets: A systematic review of 102 trials of interventions to
improve professional practice. CMAJ 1995;153:1423–31.
Address reprint requests to:
T. Yee Khong, MD
Department of Histopathology
University of Adelaide
Women’s and Children’s Hospital
72 King William Road
5006 North Adelaide, South Australia
Received July 6, 2000.
Received in revised form December 11, 2000.
Accepted January 12, 2001.
Copyright © 2001 by The American College of Obstetricians and
Gynecologists. Published by Elsevier Science Inc.
998 Khong et al
Perinatal Autopsy Attitudes Obstetrics & Gynecology