METHODOLOGICAL ISSUES IN NURSING RESEARCH
A method comparison study to assess the reliability of urine
collection pads as a means of obtaining urine specimens
from non-toilet-trained children for microbiological examination
Michael Farrell BSc RGN RSCN
Lecturer/Practitioner, Department of Nursing, University of Liverpool/Alder Hey ± Royal Liverpool Children's NHS Trust,
Liverpool, UK
Karen Devine RGN RSCN
School Nurse (formerly Staff Nurse Ward D2), Alder Hey ± Royal Liverpool Children's NHS Trust, Liverpool, UK
Gillian Lancaster BSc MSc PhD FRSS CSat
Medical Statistician, University Of Liverpool, Liverpool, UK
and B. Judd FRCPCH
Paediatric Consultant Nephrologist, Alder Hey ± Royal Liverpool Children's NHS Trust, Liverpool, UK
Submitted for publication 9 January 2001
Accepted for publication 19 November 2001
Ó 2002 Blackwell Science Ltd 387
Correspondence:
Michael Farrell,
Room 274 Mulberry House,
Royal Liverpool Children's NHS Trust,
Eaton Road,
Liverpool L12 2AP,
UK.
E-mail: [email protected]
F A R R E L LF A R R E L L MM., D E V I N ED E V IN E KK., L A N C A S T E RL A N C A S T E R GG. & J U DD& J U DD B . ( 2 00 2 )B . ( 2 00 2 ) Journal of
Advanced Nursing 37(4), 387±393
A method comparison study to assess the reliability of urine collection pads as a
means of obtaining urine specimens from non-toilet-trained children for
microbiological examination
Background. In young, non-toilet-trained children, the collection of a urine sample
for microbiology can be challenging, with the application of a urine bag being the
main method of collection. However, recent research has shown that absorbent pads
can be used yielding similar results to bag specimens. However, weaknesses in study
design erode con®dence in research ®ndings. Therefore, improvements in research
design are required to fully evaluate the reliability of pad collection.
Aims of the study. This pilot study sought to test the feasibility of a technique for
the collection of concurrent bag/pad urine samples from non-toilet-trained children,
and to assess the reliability of urine pads over bags as a collection method for urine
specimens for microbiological evaluation.
Design. A pilot, method comparison study.
Methods. Twenty concurrent bag and pad specimens were collected from non-
toilet-trained children, following parental consent. Urine specimens were analysed
for presence or absence of white cell count (WBC), and bacterial growth, using
standard laboratory methods.
Data analysis. The Kappa (j) statistics and con®dence interval (CI) estimation were
used to assess agreement between the two collection methods.
Results. Despite concurrent samples there was a lack of agreement between bag and
pad specimens on both main outcome measures. Agreement between bag and pad
specimens for the presence of WBC yielded a j� 0á10 (95% CI: 0á19, 0á39),
indicating poor agreement, while a j of 0á5 (95% CI: 0á12, 0á88) was calculated for
Introduction
Urinary tract infections (UTI) are common in children
(Woodward & Grif®ths 1993, Davies et al. 1996). The need
for early detection of urine infections in childhood to prevent
long-term urological sequelae, and promote the comfort of
the child, has been emphasized (Royal College of Physicians
1991, Kramer et al. 1993, Pead & Maskell 1994, Turner &
Coulthard 1995).
To facilitate prompt diagnosis of UTI, a specimen of urine
is collected and tested for microbiological culture. Usually a
mid-stream (or clean catch) urine specimen is preferred.
However, this method of collection is dif®cult in non-toilet-
trained children. For this group of children, a urine collection
bag device is usually ®tted. However, given that these bags
are prone to displacement, leakage and cause discomfort, and
are associated with a high contamination rate there has been
interest in developing other methods of obtaining urine
samples for microscopy, with signi®cant interest in the
development of urine collection pads (Roberts & Lucas
1985, Suri 1988, Ahmad et al. 1991, Pierro et al. 1993,
Vernon 1995, Liaw et al. 2000), and nurses taking a
particular lead in several studies (Suri 1988, Vernon et al.
1994, Lewis 1998, Feasey 1999).
Several studies have been undertaken to test the reliability
of retrieving urine from nappies for laboratory investigation
(Roberts & Lucas 1985, Suri 1988, Ahmad et al. 1991,
Vernon et al. 1994). These studies, although based on
relatively small samples (range 6±45), have demonstrated,
in part, that retrieval of urine for microbiological and
biochemical investigation is reliable. However, due to the
increased absorbency of current makes of nappies, it is now
more dif®cult to recover suf®cient amounts of urine. There-
fore, Vernon et al. (1994) tested the use of sanitary pads as a
way of collecting urine for microbiological investigation and
found that urine recovered from the pads yielded identical
culture results to urine collected from urine bags. As a result,
these pads are now commercially available in the United
Kingdom (UK).
The pad technique seems to have obvious bene®ts as a
method of collecting urine for laboratory investigation in that
pads are more comfortable and less expensive than bag-
collection methods. However, several methodological prob-
lems weaken the reliability of the available evidence. For
example, previous studies do not indicate whether the samples
compared were concurrent. For those samples taken in imme-
diate sequence there is often a lack of con®rmation that no
intervention occurred between sample collections. In addition,
operational de®nitions for con®rmation of either positive or
negative results have not always been speci®ed; sample sizes
have not been statistically determined and there have been no
indications as to whether children might have been receiving
antibiotics near or during the time that the sample was taken.
The impact of these de®cits means that it is not possible to fully
evaluate potential sources of bias and variation, and so
jeopardises the reliability of research ®ndings.
Reports in the literature have suggested that white blood
cell (WBC) count might be lower from urine collected from a
pad than collected from a urine bag (Ahmad et al. 1991,
Vernon et al. 1994, Feasey 1999). Again this is of critical
importance given that diagnosis of UTI is con®rmed when
there is a signi®cantly high urinary white blood count
(usually >100 WBC/mL). Given these limitations further
research is required to assess the reliability of the pad system.
Therefore, this pilot study sought to test the reliability of
pads compared with bags, which were used as the reference
method, as a means of collecting urine samples for micro-
biological examination. In this pilot study we particularly
wanted to establish the feasibility of using a concurrent
sample urine-collection technique as previously reported by
Suri (1988). In addition, we sought to gain preliminary data
to assess any degree in variation between pad and bag results,
the degree of agreement in bacterial growth re¯ecting moderate agreement.
Differences in proportions of the presence of WBC between bag and pad did not
quite reach signi®cance at the 5% level 0á2 (95% CI: 0á00, 0á42, P� 0á062). For
cultures the difference was calculated as 0á15 (95% CI: 0á05, 0á35, P� 0á125).
Conclusion. The pilot study demonstrates that concurrent urine samples can be
obtained without dif®culty. Despite poor to moderate agreement on outcome
measures the level of agreement is greater than reported in those other studies, that
use non-current methods of urine collection, suggesting an advantage of the
concurrent technique. It is recommended that larger scale studies be undertaken
using the concurrent collection technique to assess reliability of these ®ndings.
Keywords: method comparison, urine collection methods, pad, bag, children,
microbiology investigation
M. Farrell et al.
388 Ó 2002 Blackwell Science Ltd, Journal of Advanced Nursing, 37(4), 387±393
with the aim that such data could provide information for
planning a future large-scale study.
The study
Design
A pilot, method comparison study design was used.
Ethical considerations
The study received formal research approval from the Trust's
internal Research Review Committee. Ethical approval was
obtained from the Local Paediatric Ethics Committee.
Sample
A prospective, convenience, non-consecutive sampling tech-
nique was used. All samples were obtained from children
attending a large children's hospital, situated in the North-
west of England, UK. Parent(s) of 28, non-toilet-trained
children, who required urine microscopy as part of their
clinical care, were approached to participate in the study.
Most of the sample population were recruited from within the
outpatient setting (12/60%), where children were attending a
urology clinic for follow-up care, usually following a previous
incidence or suspected UTI. The remainder (8/40%) were
recruited from the inpatient setting. In 12 (60%) samples
urine microscopy was required as part of routine surveillance
care, whilst in eight (40%) microscopy was requested to
con®rm either the existence or absence of UTI in the presence
of signs of acute illness. The majority of samples were
collected from females (16/80%), whilst four (20%) were
collected from males. The children participating in this study
had an average age of 16 months (SDSD, 8á17, range 2±27).
Children were not recruited where it was known that they
had commenced antibiotic treatment within 7 days of the
proposed date of urine specimen collection as we were unsure
whether variability of differences in resolution of infection,
and type of antibiotics might affect results. However, chil-
dren who were receiving long-term prophylactic antibiotic
cover for UTI management, and had been receiving this
treatment in excess of 14 days, were recruited as we thought
this period of administration allowed for some stability of
antibiotic effects. In addition, we were aiming to recruit some
children with known urine infection to demonstrate validity
of the pad system.
Procedure
Specimens were collected by two members of the research
team (Farrell and Devine), having agreed and tested the
technique with the creative use of training manikins. This
meant that consistency in technique could be maintained,
thus satisfying criteria considered important for method
comparison studies (Craig et al. 2000).
In all cases, detailed explanation of the purpose of the study
was given and written consent for participation was obtained
from the child's parents. Following consent the child was
prepared for the procedure using appropriate comforting
measures. Collection of the specimen was undertaken in
accordance with universal precautions. In readiness for urine
collection the perineum was cleansed with tap water and
bacterial wipes, the skin was then dried using paper towelling.
A urine specimen bag (Biotrol Urinocol, B.B. Braun Medical
Ltd, Shef®eld, Great Britain), suitable for the size of the child,
was applied, having ®rst made a large bore perforation in the
bag, with a sterile 14 FG hypodermic needle, to allow some
drainage of urine into the pad. A pad (Euron Urine Collection
Pad, Ontex, Buggenhout) was then placed over the bag and
then the child's nappy secured. The Euron pad is non-sterile,
contains cotton wool-like material, measures 20 ´ 7 cm and
has an adhesive strip to assist in securing the pad in the nappy.
The child was then monitored every 30 min for evidence of
micturition following application of the bag and pad.
Following micturition, the nappy area was exposed and
then using a clean non-touch technique, urine was retrieved
from the pad by pressing the tip of a sterile syringe into the
pad. The plunger was then withdrawn, drawing urine into the
syringe barrel. Urine from this sample was then decanted into
a sterile specimen container, labelled with the child's details
and marked `Pad' specimen. The urine bag was then removed
and, using a similar clean non-touch technique, urine was
retrieved from the bag and decanted into a sterile specimen
container. The child's details were recorded and this container
labelled as the `Bag' specimen. Volumes of at least 0á5 mLs of
urine from both the pad and bag were retrieved to enable
analysis.
Within 30 min of collection all specimens were sent to the
laboratory, refrigerated and processed according to standard
hospital microbiological procedures. Laboratory staff was
not blinded to the urinary collection technique.
Given that we were using a concurrent collection method
we anticipated that the bag and pad would show exactly the
same microbiological results. To guide analysis and inter-
pretation of results, diagnostic de®nitions for con®rmation of
positive urinary tract infection, colonization and contamin-
ated specimen were agreed with a consultant microbiologist
(see Table 1).
Statistical methods
Statistical analysis was performed using the StatsDirect
statistical package (Ver 1.7.4, Buchan 2001). Because this
Methodological issues in nursing research The reliability of urine collection pads
Ó 2002 Blackwell Science Ltd, Journal of Advanced Nursing, 37(4), 387±393 389
was a pilot study the focus of the analysis was on levels of
agreement between bag and pad results and con®dence
interval (CI) estimation, rather than hypothesis testing.
Con®dence interval estimation provides an interval range
within which a true population value is likely to lie, within a
certain level of con®dence. It is usual for con®dence limits to
be set at 95% or higher (Gardner & Altman 1989). Thus a
95% CI indicates that under repeated sampling that there
would be a 95% con®dence level that the true value, of the
parameter being measured, would lie between the upper and
lower limits of the interval range.
The two key microbiological outcome measures analysed
were WBC count and bacteria growth pro®le. Given the small
sample size, we were not able to use the absolute data values
of WBC count or bacteria growth level to compute 95%
limits of agreement (Bland & Altman 1986). Instead, we
classi®ed urine results into whether WBCs were present or
absent, and likewise for bacteria growth. We used the Kappa
statistic to assess agreement between the two collection
methods. On computation of the Kappa statistic, a 95% CI
was generated to show the range of values for the true popu-
lation Kappa with 95% certainty. For WBC count 0á5 was
added to each cell in order to calculate the Kappa statistic. To
interpret the Kappa we used the levels of agreement as
suggested by Brennan and Silman (1992), as shown in
Table 2. In addition, we used Newcombe's (1998) method to
calculate the differences and 95% CIs for the population
paired differences in proportions between the two collection
methods in the proportion of samples with WBCs and with
bacterial growth.
Results
Concurrent bag and pad urine samples were eventually
collected from 20 non-toilet-trained children, aged less than
30 months. Lost samples were due to displacement of the urine
bag leading to inability to retrieve urine (two), and lengthy
micturition times for some children, resulting in the families
wishing to discontinue participation in the study (three), and
effects of faecal contamination (two). For the collected samples
median time from placement of bag and pad to micturition was
62á5 minutes (Interquartile range 49±119 minutes).
In the context of the operational de®nitions used to guide
this study no proven UTIs or colonizations were seen in this
series of results. However, microbiological analysis demon-
strated a lack of agreement between pad and urine specimens
in several specimens, as indicated in Tables 3 and 4.
There were four (20%) disagreements between the two
methods in terms of WBC count, and in the majority of cases
(80%) no WBCs were found (Table 3). Analysis yielded a
Kappa of 0á10 (95% CI: ÿ0á19, 0á39), suggesting poor
agreement between the two methods in terms of WBCs. The
difference in proportions of the presence of WBC between bag
and pad was found to be 0á2 (95% CI: 0á00, 0á42, P� 0á0625).
With respect to bacterial growth there were three (15%)
disagreements between the two methods (Table 4). A Kappa
Table 1 Operational de®nitions for the con®rmation of urinary tract
infection and other microbiological states
Infection: A microbiologically proven clinical diagnosis of
generalized and/or local in¯ammation. In case of
urinary tract infection, minimally 100 white
blood cells/mL combined with a pure culture
growth >105 bacteria/mL.
Colonization: The presence of one bacterium in a
concentration of 105/mL in general, with
WBC < 100/mL.
In¯ammation: The urine is sterile, i.e. no bacterium but there
are >100 WBC/mL.
Contamination: There is no clinical evidence of infection
(WBC < 100/mL) in the presence of two or
more microorganisms.
Table 2 Interpretation of Kappa Statistic as suggested by Brennan
and Silman (1992)
Kappa Statistic Strength of agreement
<0á21 Poor
0á21±0á40 Fair
0á41±0á60 Moderate
0á61±0á80 Good
0á81±1á00 Very good
Table 3 Agreement in white blood cells (WBC) count results between
pad and bag urine specimens
Bag
Pad WBC present WBC absent Total
WBC present 0 0 0
WBC absent 4 16 20
Total 4 16 20
Table 4 Agreement in culture result between pad and bag urine
specimens
Bag
Pad Bacteria present Bacteria absent Total
Bacteria present 2 0 2
Bacteria absent 3 15 18
Total 5 15 20
M. Farrell et al.
390 Ó 2002 Blackwell Science Ltd, Journal of Advanced Nursing, 37(4), 387±393
of 0á5 (95% CI: 0á12, 0á88) was calculated for the degree of
agreement in bacterial growth between bag and pad speci-
mens, suggesting moderate agreement. The difference in
proportions of bacterial growth between the two groups
was calculated as 0á15 (95% CI: 0á05, 0á35, P� 0á125).
Discussion
The use of urine pads for collection of urine samples for
microbiological analysis has gained popularity given that the
child's comfort is enhanced and that it is possible to extract
an appropriate amount of urine without dif®culty. An
additional bene®t is that the pad collection technique is less
expensive than collecting urine through the application of a
urine bag. Therefore, potential cost savings might be realized
should the pad method of collection be adopted. Previous
research has identi®ed the bene®ts of pad collection for home
collection of urine samples by parents (Liaw et al. 2000).
Our research supports these advantages, ®nding pads easy
to use and generally enabling easy retrieval of urine.
However, concerns regarding reduced WBC counts and
altered bacterial pro®les from pad samples remain signi®cant.
This study has shown that, despite concurrent paired samp-
ling there is a possibility of poor agreement between the pad
and bag systems. The danger is that a relatively large degree
of difference in WBC values between the two systems could
cause the recording of a false-negative result.
It is interesting to note that other studies have shown
differences in proportions of WBC between bag and pad
collection. Feasey (1999) compared WBC counts and bacterial
results in 48-paired samples (although not concurrent)
obtained from children. In that data set three positive urine
tract infections were microscopically con®rmed and the
researcher indicates that, despite variance, in the magnitude
of the differences obtained between bag and pad specimens,
that the pad specimen yielded suf®cient information to con®rm
a positive UTI. Accordingly, Feasey concluded that urine pads
were reliable. Yet, while the researcher rightly cautions that
WBC counts are lower she did not indicate the magnitude of
the change. Thus we reanalysed the raw data for agreement
and for differences in WBC proportions using CI estimation.
The results con®rm poor agreement between methods, giving a
Kappa 0á19 (95% CI: ÿ0á04, 0á43), and a difference between
proportions of children in whom WBCs were detected between
bag and pad as 0á31 (95% CI: 0á13, 0á46, P� 0á0009). This
difference between WBC count proportions is greater than in
our study, which used concurrent samples, thus suggesting
some advantage in the concurrent sampling method.
The difference in WBC and bacterial counts between bag
and pad specimens suggests that the pad system has some
deleterious impact on bacteria, although it is not possible to
indicate why this occurs. Perhaps cells become embedded in
the fabric, which prevents their reclaim, and thus distorts the
counts obtained from urine retrieved from pads.
There is some evidence from earlier studies, which have
shown that changes in other urine constituents can also be
affected when trying to retrieve urine from absorbent mater-
ials, con®rming the suspicion of some impairment in the
retrieval process. For example Roberts and Lucas (1985),
undertaking an laboratory based experimental study, found
that while urine reclaimed from commercially available brands
of nappies, yielded the same results against a known control
concentration for a range of urinary constituents, including
sodium, urea, creatinine levels, that the levels of some constit-
uents increased (phosphorous and calcium). While the
researchers note the stability of levels of the key urinary
constituents between nappy and usual collection methods is
equivalent, they postulate that the increase in some levels is due
to a leeching action, activated when urine contacts the nappy
material. In another study researchers retrieving urine from
absorbent cotton wool, placed in the nappy area to aid
collection of urine from 14 infants for the investigation of
proteinuria, found that urinary albumin and retinal binding
protein levels could be signi®cantly decreased, in comparison
to usual collection method (Smith & Taylor 1992). Interest-
ingly they noted that while albumin levels were decreased, the
level of decrease was not uniform, ranging from a small loss in
some samples up to 35±40% in others. These researchers also
found that in respect of retinal binding protein that most of the
loss occurred within 15 minutes of contact with the absorbent
material. They conclude that caution is needed when using
such materials for the retrieval of urine specimens required for
urinary analysis. The implications of such ®ndings for the
current study is that further research work is needed to con®rm
the mechanisms involved which precipitate reduction in
overall urinary pro®le when urine comes in contact with
absorbent material used for the collection of urine specimens.
Thus further research is required to explore further these
®ndings as there is a danger that any change in urine pro®le
may lead to errors in con®rming UTIs, by microscopic means.
Macfarlane et al. (1999) report their concerns that urine
pads have a high association with contamination and worry
that such contamination can hide infection. This must be a
considerable concern, given the signi®cant morbidity associ-
ated with UTIs in childhood. Therefore, Macfarlane and
colleagues consider that heavily contaminated urine pad
samples will often need to be repeated, thus eroding the
bene®ts of the pad collection system. However it should be
noted that urine bags also have a high contamination rate
(Ramage et al. 1999). Thus, given the limitations of current
Methodological issues in nursing research The reliability of urine collection pads
Ó 2002 Blackwell Science Ltd, Journal of Advanced Nursing, 37(4), 387±393 391
collection methods in children the clean-catch technique
should remain the preferred way to collect urine from young
non-toilet-trained children (Macfarlane et al. 1999, Ramage
et al. 1999).
Limitations of the study
Several limitations impact upon the ®ndings of the present
study. This was a pilot study and thus sample size has not
been statistically determined. The small sample size requires
that a larger study be undertaken, with the sample size
determined by power calculation, to study the reliability and
validity of ®ndings. Furthermore, in a larger study, it may be
possible to use the methods of Bland and Altman (1986) to
analyse data in their original form, rather than having to
dichotomise the results into presence or absence of WBCs and
bacterial growth. Using these methods it would be possible to
determine the magnitude of differences between the two
methods using 95% limits of agreement.
In this study no children were con®rmed as having a UTI.
Therefore, it is not possible to indicate the degree of variance in
results between pads and bags in the presence of signs of urinary
tract infection. There is a particular need to study the magni-
tude of difference in WBC counts and culture growths between
bag and pad specimens in children known to have a UTI.
Conclusion
Notwithstanding its limitations, the study has shown that it is
possible to obtain concurrent paired samples without dif®-
culty. It has also shown that, despite concurrent sampling,
there is a likelihood of poor agreement between microbio-
logical results, suggesting that pads have a deleterious impact
on WBC and bacterial counts. Although the concurrent
method shows differences in key microbiological results it
appears that these differences are less than would be found in
studies that use non-concurrent urine samples. Therefore, it is
suggested that the concurrent technique of bag/pad urine
collection should be used in larger studies to enhance study
design in ascertaining the reliability of urine pads for the
collection of urine.
Acknowledgements
We appreciate the comments made by peer reviewers and
thank Dr C Watkins, University of Manchester, for helpful
additional comments.
This study was supported with a research grant made by
the Endowment Committee of the Royal Liverpool Childrens
NHS Trust.
References
Ahmad T., Vickers D., Campbell S., Coulthard M.G. & Pedler S.
(1991) Urine collection from disposable nappies. Lancet 338,
674±676.
Bland J.M. & Altman D.G. (1986) Statistical methods for assessing
agreement between two measures of clinical measurement. Lancet
I, 307±310.
Brennan P. & Silman A. (1992) Statistical methods for assessing
observer variability in clinical measures. British Medical Journal
304, 1491±1494.
Buchan I. (2001) http://www.StatsDirect.com.
Craig J.V., Lancaster G.A., Williamson P.R. & Smyth R.L. (2000)
Temperature measured at the axilla compared with rectum in
children and young people: a systematic review. British Medical
Journal 320, 1174±1178.
Davies E.G., Elliman D.A.C., Hart C.A., Nicoll A. & Rudd P.T.
(1996) Manual of Childhood Infections. British Paediatric Associ-
ation, W.B. Saunders Co, London.
Feasey S. (1999) Are Newcastle urine collection pads suitable as a
means of collecting specimens from infants? Paediatric Nursing 11,
17±18, 20±21.
Gardner M.J. & Altman D.G. (1989) Statistics with Con®d-
ence: Con®dence Intervals and Statistical Guidelines. BMJ,
London.
Kramer M.S., Tange S.M., Drummond K.N. & Mills E.L. (1993)
Urine testing in young febrile children: a risk bene®t analysis.
Journal of Pediatrics 125, 6±12.
Lewis J. (1998) Clean-catch versus urine collection pads: a pros-
pective trail. Paediatric Nursing 10, 15±16.
Liaw L.C., Nayar D.M., Pedler S.J. & Coulthard M.G. (2000) Home
collection of urine for culture from infants by three methods:
survey of parents' preferences and bacterial contamination rates.
British Medical Journal 320, 1312±1313.
Macfarlane P.I., HoughtonC. & HughesC. (1999) Pad urine collection
for early childhood urinary-tract infections. Lancet 354, 571.
Newcombe R. (1998) Improved con®dence intervals for the differ-
ence between binomial proportions based on paired data. Statistics
in Medicine 17, 2635±2650.
Pead L. & Maskell R. (1994) Study of urinary tract infections in
children in one health district. British Medical Journal 309,
631±634.
Pierro A., Jones M.O. & Lloyd D.A. (1993) A method for urine
collection in infants. Archives of Disease in Childhood 69(1 spec
no), 85±86.
Ramage I.J., Chapman J.P., Hollman A.S., Elabassi M., McColl J.H.
& Beattie T.J. (1999) Accuracy of clean-catch urine collection in
infancy. Journal of Pediatrics 135, 664±666.
Roberts S.B. & Lucas A. (1985) Measurement of urinary constituents
and output using disposable napkins. Archives of Disease in
Childhood 60, 1021±1024.
Royal College of Physicians (1991) Guidelines for the acute manage-
ment of urinary tract infection in childhood. Report of a working
group of the Research Unit, Royal College of Physicians. Journal of
the Royal College of Physicians 25, 36±42.
Smith G.C. & Taylor C.M. (1992) Recovery of protein from urine
specimens collected in cotton wool. Archives of Disease in
Childhood 67, 1486±1487.
M. Farrell et al.
392 Ó 2002 Blackwell Science Ltd, Journal of Advanced Nursing, 37(4), 387±393
Suri S. (1988) Simplifying urine collection from infants and children
without losing accuracy. Journal of Maternal and Child Nursing
13, 438±441.
Turner G.M. & Coulthard M.G. (1995) Fever can cause pyuria in
children. British Medical Journal 311, 924.
Vernon S. (1995) Urine collection from infants: a reliable method.
Paediatric Nursing 7, 26±27.
Vernon S., Refearn A., Pedler S., Lambert H.J. & Coulthard M.G.
(1994) Urine collection on sanitary towels. Lancet 344, 612.
Woodward M.N. & Grif®ths D.M. (1993) Use of dipstick for routine
analysis of urine from children with acute abdominal pain. British
Medical Journal 306, 1512.
Methodological issues in nursing research The reliability of urine collection pads
Ó 2002 Blackwell Science Ltd, Journal of Advanced Nursing, 37(4), 387±393 393
Top Related