Niger J Paed 2015; 42 (1): 1 –5
ORIGINAL
Umar LW
Premature discharge of children
Musa S
from hospital admission at Ahmadu
Musa A
Adeoye GO
Bello University Teaching Hospital
Zaria: A 3-year review
DOI:http://dx.doi.org/10.4314/njp.v42i1,1
Accepted: 16th March 2014
Abstract Introduction: Leaving
admissions, giving a prevalence of
hospital care prematurely could
2.0%. Under-five children consti-
Umar LW
(
)
threaten the healthy survival of
tuted 65.4%, with a male: female
Musa S, Musa A, Adeoye GO
and expose children to a risk of
ratio of 2.3:1. Thirty-one (55.3%)
Department of Paediatrics,
Ahmadu Bello University Teaching
harmful alternatives. It is also a
children were absconded
with
Hospital,
concern and a challenge to health-
while 25 (44.7%) were taken away
Shika-Zaria, Nigeria.
care providers and the health sys-
by caregivers against medical ad-
Email: umarlw@gmail.com
tem. A better understanding of its
vice. The commonest diagnoses
characteristic could help mitigate
were protein-energy malnutrition
the impact on children.
and pneumonia and majority were
Objective:
To
determine
the
from poor socio-economic family
prevalence, types of, and reasons
background. The commonest rea-
for premature hospital discharge
sons for discharge AMA were un-
amongst children.
affordable costs and perception of
Methods: We carried out a three-
improvement of child’s medical
year retrospective review of case
condition.
notes of children who were taken
Conclusion: Socio-economic fac-
away from hospital admission by
tors were significant determinants
their
parents/caregivers
before
of utilization of in-patient hospital
they were due for discharge.
services for the children studied.
Socio-demographic, clinical and
There is a need for provision of
discharge information were col-
affordable health care as well as
lected and data was entered into
efficient discharge policy to pro-
Microsoft® Excel® for Mac 2011
tect children from potential risks
(Version 14.1.0) , cleaned and ana-
associated with premature hospital
lysed. Results were presented as
discharge.
percentages, statistical means and
standard deviations, tables and
Key words: Children; admission;
charts.
premature discharge; discharge
Results: There were 56 cases of
against medical advice; abscond-
premature discharge out of 2858
ing ; elopement
Introduction
with provisions of care and reduce its occurrence 4,5 .
During the course of hospital admission children are
Discharge AMA in the case of children, occurs where a
sometimes
taken
away
against
medical
advice
child caregiver declare their intention to remove their
(discharge AMA, DAMA), or without notice of medical
ward from care, has been counselled and duly informed
staff on duty by their caregivers (absconding, elope-
about implications of their decision by an attending
ment). Such premature termination of care could pro-
healthcare provider, but still opts to leave, and sign rele-
long duration of illness and put children at risk of com-
vant documents as evidence of voluntary withdrawal of
consent . Factors that could predict the likelihood of
6
plications, death or to uncertain and precarious alterna-
tive therapies . The child right perspectives not only
1,2
leaving hospital AMA could be either facility-related or
patient-related . The former include site, structure, in-
7,8
uncover legal responsibility shortcomings of caregivers
(for decisions to withdraw consent) but also pose med-
stitutional policy, quality and cost of service as well as
ico-legal challenges for the healthcare provider and the
staff disposition (number and competence), while the
health system
2,3,4,5
. An understanding of the factors as-
latter include socio-economic status, type of illness, per-
sociated with premature termination of hospital care of
ceptions and attitudes regarding illness and treatment, as
sick children could help improve caregiver compliance
well as ability to make independent decisions regarding
2
choices of treatment . Previous studies have consis-
7,8
our study but none of the children in this report benefit-
tently shown that socio-economic status play a signifi-
ted from it.
cant role in premature discharges involving both adults
and children.
6,8,9-12
The prevalence from studies in chil-
We defined cases of premature discharge to include
dren ranges between 1-2% in most studies in developing
situations where a caregiver opts to leave hospital care
countries that reported retrospective analyses of cases of
with their ill child and declared their intentions to at-
leaving in-patient care AMA alone
13,14,15
, while higher
tending medical staff. These were documented in hospi-
prevalence was obtained in other reports that included a
tal forms that were signed by caregivers accepting all
larger proportion of neonates
6,14
. In some other circum-
responsibility for withdrawing consent and leaving hos-
stances children could be taken away from hospital by
pital care with their children (discharged AMA). A sec-
caregivers without notice of medical staff, i.e. abscond-
ond category of children whose caregivers left hospital
ing (elopement), which is considered by some as a form
care with them unnoticed (absconded) were also consid-
of discharge AMA
12,17
. The commonly associated fac-
ered as premature discharge and included in the
study
12,17
tors observed in developing countries include poverty,
.
perception of illness improvement and conditions such
as HIV, tuberculosis and protein energy malnutrition
6,
Socio-demographic, clinical data and information about
13,14,15
.
circumstances around the time of leaving were collected
from the case notes and entered into a profoma. The
Although the legal right to give consent for medical care
socio-economic status of families of children was cate-
of children lies with parents/ guardians, the unrestricted
gorised using the classification scheme of Oyedeji, using
parental educational level and family income . Those in
18
right to withdrawal of such consent puts a limit to the
professional obligation of acting in the best interest of
social classes I-III were considered to belong to high and
the child by the medical care team.
2,3,4,5,17
those in IV-V to low socio-economic status respectively.
The main objectives of this study were therefore, to de-
Data was then entered into Microsoft® Excel® for Mac
termine the prevalence of and identify the factors associ-
2011 (Version 14.1.0) , cleaned and analysed. Results
ated with premature discharge of hospitalised children.
were presented as proportions, statistical means and
standard deviation (±SD), contingency tables and charts.
Chi-square test was used to test for significance and P
values of <0.05 were considered significant.
Materials and Methods
A review of case notes of children whose caregivers
signed and left hospital admission with them against
Results
medical advice and those whose caregivers took them
away unnoticed (absconded) from the Emergency Paedi-
A total of 2858 children were admitted within the period
from 1 January 2007 to 31 December 2009, out of
st
st
atric Unit (EPU) and Paediatric Medical Ward (PMW)
of our hospital between 1 January 2007 and 31 De-
st
st
which in 56 caregivers either signed and left AMA or
cember 2009 was conducted. These are busy medical
absconded with them following a variable period of stay
wards with a total capacity of 74 beds and average bed
on hospital admission. With a mean annual 953 admis-
occupancy of 70-80% at anytime. Children presenting
sions and a mean of 18.7 (SD ±4.0) cases of premature
with emergency medical conditions and those critically
discharges this gave an annual prevalence of 2.0%. The
ill are routinely admitted for resuscitation in the EPU.
male: female ratio is 2.3:1 and children <5 years old
Neonates were not included because the newborn unit is
constituted 64.3%, with their age ranging from three
situated in another block closer to the labour ward and
months to 15 years. The gender and age distribution of
data on neonates is being analyzed separately for subse-
the study children are shown in Table 1.
quent publication. An emergency treatment voucher
allows admission and commencement of care for the
Table 1: Gender and age distribution of 56 children
first 72 hours for patients whose caregivers could not
prematurely discharged from admission
afford immediate payment of admission fees (pending
Gender
DAMA* (%)
Absconded (%)
Total (%)
when payments could be made within this period). A
system of indigent patients’ support of costs of treatment
Male
17 (30.4)
22 (38.3)
39 (69.6)
and care for HIV infected children by the US Govern-
Female
8 (14.3)
9 (16.1)
17 (30.4)
ment funded APIN (Aids Prevention Initiative in Nige-
Total
25 (44.6)
31 (55.4)
56 (100)
Age years)
DAMA* (%)
Absconded (%)
Total (%)
ria) Programme has been in existence in the hospital
< 1
5 (8.9)
7 (12.5)
12 (21.4)
from 2007. Patients are either discharged home from
1 – 5
11 (19.6)
13 (23.2)
24 (42.6)
EPU if they recovered or more often their care would be
6 – 16
9 (16.1)
11 (19.6)
20 (36.0)
transferred to the PMW under the appropriate specialist
Total
25 (44.6)
31 (55.4)
56 (100)
unit if they stayed up to 72 hours. All other children
presenting with non-emergency medical conditions get
*DAMA: Discharged against medical advice
admitted directly into the PMW during routine work
hours. The National Health Insurance Scheme (NHIS)
Thirty-one (55.4%) sick children were taken away from
has been implemented in our hospital over the period of
admission by their caregivers without notice of medical
staff on duty (absconded). For the remaining 25 (44.6%)
3
children, caregivers disclosed their intentions to leave
Corresponding data for Fig 1
despite further counselling by medical staff on duty, but
DAMA*
(%)
Absconded
(%)
agreed to sign on the appropriate documents.
Monday
5
20.0
6
19.4
The commonest diagnoses amongst all the prematurely
Tuesday
5
20.0
2
6.5
discharged children were severe protein energy malnu-
Wednesday
3
12.0
4
12.9
trition and pneumonia. Grouped together, thirty-seven
Thursday
4
16.0
3
9.7
(66.1%) children had infectious illnesses including ty-
Friday
5
20.0
6
19.4
phoid fever, severe malaria, persistent diarrhoea and
Saturday
3
12.0
5
16.1
tuberculosis. Table 2 shows the various diagnoses for all
Sunday
0
0.0
5
16.1
cases of premature discharge.
Total
25
100.0
31
100.0
Table 2: Diagnoses of 56 children prematurely discharged
*DAMA: Discharged against medical advice
from admission
Table 3: Duration of hospital stay for 56 children prematurely
Diagnosis
No.
%
discharged
Infections
Duration of
Pneumonia
11
19.6
Stay (Days)
DAMA*
(%)
Absconded
(%)
Typhoid fever
6
10.7
1-4
3
12
3
9.7
Urinary tract infection
2
3.6
5-10
17
68
12
38.7
Acute osteomyelitis
2
3.6
11-14
3
12
10
32.2
Severe and complicated malaria
4
7.1
>14
2
8
6
19.4
Total
25
100
31
100.0
Acute bacterial meningitis
3
5.4
Persistent diarrhoea
4
7.1
*DAMA: Discharged against medical advice
Disseminated tuberculosis
3
5.4
Mean duration of hospital stay = 4 days (discharged AMA)
Overwhelming sepsis
2
3.6
versus 5 days (Absconded);
X = 0.0000, P = 0.9911.
2
Others
Protein energy malnutrition
13
23.3
The social classes of families of the children are shown
Malignancies (neuroblastoma, rhabdo-
myosarcoma)
4
7.1
in Table 4. Even though in up to a quarter of the chil-
dren the family economic status was not disclosed,
Sickle cell disease vaso-occlusive crisis
2
3.6
where such information was available more children
Total
56
100
(46.4%) came from poorer socio-economic family back-
ground amongst all the categories. There was no child
Majority of discharges AMA occurred during routine
amongst those from the high socio-economic class that
work hours of weekdays (7.30 am to 4.00 pm, Monday
absconded.
to Friday), while there was no difference between week-
days and weekends for those who absconded, even
Table 4: Socio-economic classes of families of 56 children
though they did so most often outside routine working
prematurely discharged
hours on weekdays. There were relatively more dis-
charge AMA on week beginnings (Mondays and Tues-
Social Class
DAMA*
Absconded
Total (%)
days) and Fridays. The distribution of premature dis-
I
3
0
3 (5.4)
charge by days of the week is shown in Figure 1.
II
-
-
-
III
6
8
14 (25.0)
The majority of premature discharge tended to occur
IV
-
-
-
after the first 4 days of admission. Discharges that took
V
10
16
26 (46.4)
place between the first and fourth days of admission
Unspecified
6
7
13 (23.2)
were mostly of children managed in the EPU with the
Total
25
31
56 (100.0)
emergency treatment voucher provided pending the pay-
ment of hospital bills. These children were taken away
*DAMA: Discharged against medical advice
X = 0.385, P = 0.5348).
2
before they were declared to have satisfactorily recov-
ered, as shown in Table 3. There was no statistical dif-
On further analysis of the children whose family socio-
ference in the mean duration of hospital stay be tween the
economic status was specified (19 in the discharge
discharge AMA and the absconded sub-groups
(4 days versus 5 days; X = 0.0000, P = 0.9911).
2
AMA subgroup and 24 in the absconded), there was no
statistical difference between the two sub-groups (higher
social classes I-III versus lower classes IV-V; X =
2
Fig 1: Distribution of premature discharge of 56 children by days of
the week
0.385, P = 0.5348).
Reasons tendered by caregivers who indicated their in-
tention to leave with children AMA are presented in
Figure 2. The commonest reasons were financial con-
straints and perception that child has improved and was
well enough to go home. No information could be
obtained in the case records on the reasons for abscond-
ing from care for the 31 cases that absconded, but
4
majority of these had various amounts of outstanding
Majority of the diagnoses of the 56 children were pre-
bills ranging from the initial admission deposits to other
dominantly emergency infectious illnesses, initially
bills that accrued later or a combination of these.
managed in the EPU, with bronchopneumonia being the
foremost single diagnosis in this category. Amongst
Fig 2: Reasons for discharge AMA for 25 children
other children directly admitted and managed in the
PMW, protein energy malnutrition was in the lead, fol-
lowed by disseminated TB and malignancies. Various
Undisclosed reasons
studies in different parts of Nigeria also described simi-
Parental conflict
lar findings with predominance of infectious illnesses
Going for alternative care
amongst children that were discharged AMA
9,13-16
.
Dissatisfaction
While children with HIV infection alone or presenting
Financial constraints
as co-infection with TB have prominently featured in
Child is well enough
association with premature discharge in some reports
across Nigeria
12,13
, we did not find such cases in this
0
10
20
30
40
report. This we attributed to the indigent support of costs
of treatment for HIV infected children by the US Gov-
% of cases of discharge against medical advice
ernment funded APIN Programme that probably influ-
Corresponding data for Fig 2
enced caregiver decision to stay with children till com-
pletion of hospital management.
Reasons for DAMA
Frequency
%
Child is well enough
7
28
The main reason for leaving AMA tendered by caregiv-
Financial constraints
9
36
ers was financial constraints in up to 35% of cases in
Dissatisfaction
1
4
this category. It is noteworthy also that up to 46.4% of
Going for alternative care
2
8
children were from families that fell in the lower social
Parental conflict
1
4
Undisclosed reasons
4
16
classes. These finding are in keeping with what has been
observed in reports from Nigeria and other countries.
9,13-
Total
25
96
16
The findings are however in contrast with observa-
tions in developed countries where medical welfare ser-
vices like the Medicaid for the less privileged exist
1,6-
8,11,17
Discussion
. Even though the information on family socioeco-
nomic status was not available for 13 (23.2%) of chil-
This study has revealed a prevalence of 1.96% for pre-
dren discharged AMA only three (5.4%) came from the
mature discharge in children, a figure that falls within a
highest socio-economic class. For the 43 children with
range of 1-2% in similar reports from other tertiary hos-
more complete information there was no difference in
pitals in Nigeria
13,14,15
, which is also similar to what was
the family socioeconomic status between those that were
found from a General Hospital in Awka (2%) . The
16
discharged AMA and those who absconded. Although
prevalence figure is however less than what was ob-
data on reasons for absconding was unavailable for chil-
served in Port Harcourt (3.8%) , and in the Islamic Re-
6
dren that were absconded with, we presume that poor
public of Iran (5.4%) , both of which had significant
14
family socioeconomic status may have similarly influ-
neonatal sub-population in their subjects. In contrast to
enced decisions by caregivers to abscond with children.
these reports however, our data described a category of
A common reason for discharge AMA was caregiver
children that were absconded with from admission with-
perception that child has improved enough even though
out notice to medical personnel on duty, which is con-
treatment was still on going, a finding that was similarly
reported in previous studies
9,13-16
sidered as elopement, a recognised form of discharge
. This may have been
against medical advice in the literature. Two compre-
due to limited information and counselling on nature of
hensive reviews, one each from the US and the King-
the illness and need for adherence to treatment provi-
dom of Saudi Arabia have categorised elopement as a
sions made available to caregivers at admission, a strat-
12,17
egy recommended to control discharge AMA .
12
form of discharge against medical advice
. Each of
the two sub-types of premature discharge was shown to
have features that could be relevant to planning appro-
The pattern of premature discharge showed that majority
priate prevention strategy
9,12
.
Our results however,
of discharges AMA occurred during official work hours
failed to reveal significant differences in some character-
(7.30 am to 4.00 pm) on weekdays with no difference
istics of children discharged AMA and those absconded
between weekdays and weekends for those who ab-
with (mean duration of hospital stay and family socio-
sconded even though they did so outside work hours.
economic status). This could have been due to the obvi-
This might be explained by the fact that especially dur-
ous limitation in methodology, being a retrospective
ing visiting periods when relatives come around, care-
study with inadequate information and no means of
givers and their children could have conveniently left
tracking patients to ascertain reasons for absconding.
the wards in company of their relatives without notice of
Adequate information was obtained in a prospectively
attending staff. Caregivers who absconded with their
designed study that utilized a tracking method with tele-
sick children could have found it convenient to do so
phone calls to obtain responses .
6
after official work hours to evade settling hospital bills.
A case for improving the coverage of the National
5
Health Insurance Scheme (NHIS) to cater for the needs
keep children under hospital care. Improving the cover-
of children from families with poor socio-economic
age of the NHIS and developing institutional policy to
status is evident in this context. The NHIS coverage in
control premature discharge could help to protect chil-
Nigeria is still only about 3.5% as at 2010 and has been
dren as well as staff and the facility within a medico-
limited to civil service employees and their families
legal framework.
since its inception in 2005.
19
Author contributions
Umar LW: Research coordinator, theoretical concept
Conclusion
development, literature review, data analysis
and writing.
In conclusion, premature discharge from hospital goes
Musa S: Data collection, data analysis, discussion.
beyond just leaving AMA to include situations where
Musa A: Data collection, discussion.
caregivers abscond with children in our setting. Poor
Adeoye GO: Data collection.
socio-economic status and erroneous caregiver percep-
Conflict of interest: None
tion about recovery from illness influenced decisions to
Funding: None
References
1.
Hwang SW, Li J, Gupta R, Chien
8.
Aliyu Z. Discharge against medi-
15.
Eke GK, Opara PI. Discharge
V, Martin RE. What happens to
cal advice: socio-demographic,
against medical advice amongst
patients who leave hospital
clinical and financial perspec-
patients admitted into the Paediat-
against medical advice? Cana-
tives. Int J Clin Pract 2002; 56
ric wards of the University of Port
dian Med Assoc J. 2003;168 (4):
(5): 325-327.
Harcourt Teaching Hospital. Ni-
417-420.
9.
Okoromah CN, Egri-Qkwaji MT.
ger J Paediatr 2013; 40 (1): 40 –
2.
Berger JT. Discharge against
Profile of and control measures
44.
medical advice: ethical consid-
for the paediatric discharges
16.
Ejiofor OS, Uzoechina ON. Dis-
erations and professional obliga-
against medical advice. Niger
charge against medical advice
tions. J Hosp Med 2008;
Postgrad Med J 2004; 11: 21-5.
(DAMA) in children’s ward; the
3:403e8.
10.
Ibrahim SA, Kwoh CK, Krishnan
Awka, South East Nigeria experi-
3.
Al Siddique AA. Medical liabil-
E. Factors associated with pa-
ence. Trop J Med Res 2010;14:
ity: the dilemma of litigations.
tients who leave acute-care hospi-
50 -52. (2). DOI: http://
Saudi Med J 2004; 25: 901-906.
tals against medical advice. Am J
dx.doi.org/ 10.4314%
4.
Devitt PJ, Aoifinn CD, Mantosh
Public Health 2007; 97: 2204e8.
2Ftjmr.v14i2.65515.
D. An examination of whether
11.
Dershewitz RA, Paichel W. Pa-
discharging patients against
tients who leave a pediatric emer-
17.
Alfandre JD. “I’m Going Home”:
medical advice protects physi-
gency department without treat-
Discharges against medical ad-
cians from malpractice charges.
ment. Ann Emerg Med 1986;15
vice. Mayo Clin Proc 2009; 84
Psychiatr Serv 2000; 51(7):899-
(6):717-20.
(3): 255-260.
902.
12.
Al Ayyed I. What makes patients
18.
Oyedeji GA. Socioeconomic and
5.
Draper H, Sorell T. Patients'
leave against medical advice? J
cultural background of hospital-
responsibilities in medical ethics.
Taibah University Med Sci 2009;
ised children in Ilesa. Niger J
Bioethics 2002; 16(4): 335–352.
4:16-22.
Paediatr 1985; 12: 111-117.
6.
Roodpeyma S, Hoseyni SAE.
13.
Ibekwe RC, Muoneke VU, Nnebe
19.
Odeyemi IAO, Nixon J. Assess-
Discharge of children from hos-
-Agumadu UH, Amadife MA.
ing equity in health care through
pital against medical advice.
Factors influencing discharge
the national health insurance
World J Pediatr 2010; 6 (4): 353
against medical advice among
schemes of Nigeria and Ghana: a
-356.
paediatric patients in Abakaliki,
review-based comparative analy-
7.
Moyse AS. Discharges against
Southeastern Nigeria. J Trop
sis. Int J Equity Health 2013;
medical advice: a community
Pediatr 2009; 55(1): 39-41.
12:9. Available from: http://
hospital’s experience. Can J
14.
Ikefuna AN, Emodi IJ. An assess-
www.equityhealthj.com/
Rural Med 2004; 9: 148-153.
ment of factors influencing hospi-
content/12/January/2013 .
tal discharges against medical
advice of paediatric patients in
Enugu: A Review of 67 Cases.
Niger J Paediatr 2002; 29:1-4.