Neurourology and Urodynamics 25:674^683 (2006)

Systematic Review and Meta-Analysis of Methods of Diagnostic

Assessment for Urinary Incontinence
J.L. Martin,1*y K.S. Williams,2{ A.J. Sutton,2‰ K.R. Abrams,2‰‰ and R.P. Assassa3{
1
School of Electrical and Electronic Engineering, The University of Nottingham, Nottingham, United Kingdom
2
Department of Health Sciences, University of Leicester, Leicester, United Kingdom
3
Pontefract General In¢rmary, Mid Yorkshire Hospitals NHS Trust, Pontefract, United Kingdom
Aims: To evaluate the performance of all tests proposed for the diagnosis of urinary incontinence.
Methods: A systematic review and meta-analyses of the published literature of methods for diag-
nostic assessment of urinary incontinence. Results: One hundred twenty-one papers were included
in the full review [Martin et al., 2006]. The quality of reporting in the primary studies was poor
which reduced the number of studies that could be included in the data analysis. The literature
suggests that women with urodynamic stress incontinence (USI) can be correctly identi¢ed in pri-
mary care from clinical history alone with a sensitivity of 0.92 (95% C.I.: 0.91^ 0.93) and speci¢city of
0.56 (0.53 ^ 0.60). A clinical history for the diagnosis of detrusor overactivity (DO) was found to be
0.61 (0.57^ 0.65) sensitive and 0.87 (0.85 ^ 0.89) speci¢c. Within secondary care imaging of leakage by
ultrasound was found to be e¡ective in the diagnosis of USI in women with a sensitivity of 0.89
(0.84 ^ 0.93) and speci¢city of 0.82 (0.73 ^ 0.89). Conclusions: Clinical interpretation of the results
of the review is di⁄cult because few studies could be synthesized and conclusions made. The pub-
lished evidence suggests that a large proportion of women with USI can be correctly identi¢ed in
primary care from history alone. Ultrasound o¡ers a useful diagnostic tool which could be used
prior to, and possibly instead of, multi-channel urodynamics in some circumstances. If a patient is
to undergo urodynamic testing, multi-channel urodynamics is likely to give the most accurate
result. Further primary studies adhering to STARD guidelines are required on commonly used
tests. Neurourol. Urodynam. 25:674 ^683, 2006. ß 2006 Wiley-Liss, Inc.

Key words: DO; history; multi-channel urodynamics; pad-test; sensitivity; speci¢city; USI

INTRODUCTION It is particularly important when implementing certain

Whilst urinary incontinence (UI) is not life threatening, it
can have enormous costs to individuals and the health service
in terms of expenditure and impact on quality of life. Epide-
miological studies have demonstrated that UI is a very com-
mon urinary symptom, with reported prevalence of any UI
(in those aged 40 and over) of 34% for women and 14% for
men [McGrother et al., 2003]. The proportion of people
¢nding these symptoms impact on their lives is estimated to
be around 29% for women and 14% for men [McGrother
et al., 2003]. A UK study reported the total cost of treating
urinary storage disorders in community dwelling adults
over the age of 40 was »536 million in 1999/2000 prices. In
addition there is an estimated cost of »207 million that is
borne by the individual for managing their symptoms [Turner
et al., 2004].
Pathways to diagnostic assessment are inconsistent, with
some individuals being assessed and treated in primary care
settings by general practitioners and nurses, and others being
referred directly to a variety of specialists in secondary care
without any assessment or treatment. Assessment can be
undertaken at a number of levels employing di¡erent combi-
nations of screening tests.
treatment interventions that a diagnosis is made in order to
determine the most e¡ective treatment intervention and of
course it is imperative before surgical intervention. If a diag-
nosis is not made, then inappropriate and unnecessary inter-
ventions may be implemented. Two types of diagnosis can be
made: symptomatic diagnosis and condition-speci¢c diagno-
sis. Generally symptomatic diagnoses are made in primary
No con£ict of interest reported by the author(s).
Work undertaken at: Department of Health Sciences,University of Leicester,
Leicester, UK.
y
Research Fellow.
{
Senior Research Fellow in Nursing.
‰
Senior Lecturer in Medical Statistics.
‰‰
Professor of Medical Statistics.
{
Consultant Urogynaecologist.
Grant sponsor: UK National Health Service Health Technology Assessment
Programme.
*Correspondence to: J.L. Martin, School of Electronic Engineering, The
University of Nottingham, Nottingham, UK.
E-mail: Jennifer.martin@nottingham.ac.uk
Received 28 September 2005; Accepted 20 July 2006
Published online 2 October 2006 in Wiley InterScience
(www.interscience.wiley.com)
DOI 10.1002/nau.20340

ß 2006 Wiley-Liss, Inc.

 Systematic Review of Diagnostic Assessments for UI 675

care using clinical history taking, urinary diaries, pad-tests, the main report, were identi¢ed based on possible diagnostic
and validated symptom scales. Condition-speci¢c diagnoses are tests and possible permutations of their names.
made in secondary care using urodynamic techniques. The use
of diagnostic assessment methods is in£uenced by the clinical set-
Selection and data extraction. In order to be included, a
ting and the expertise of the individual undertaking the assess-
paper had to provide a quantitative comparison between two or
ment. The evidence available on the accuracy and acceptability
moredi¡erentmethodsofdiagnosing urinary incontinenceand
of these diagnostic processes is inconsistent and variable.
meet all of the following criteria: were published in English,
Clinical decisions and practice are made on the basis of the
comprised adult populations and were not investigating
best available evidence, it is seldom possible for practitioners
interventions where diagnostic tests were used as outcome
to access, read, and evaluate all the available evidence on a par-
measures. All papers were assessed for relevance by the ¢rst
ticular intervention or diagnostic test. Systematic reviews
author and approximately 15% by the second author, any
appraise and summarize the e⁄cacy results of individual stu-
discrepancies were discussed.
dies to provide health care providers and others with a picture
The QUADAS tool was used to assess the included studies for
of the overall performance of an intervention or diagnostic
quality (Table I). The tool consists of 14 questions regarding the
test. Resources exist to help ensure that appropriate methods
quality of the study and quality of reporting [Whiting et al.,
are used to conduct and report systematic reviews.
2004]. Summary data were extracted from papers that provided
The objective of this systematic review was to evaluate the
data in a suitable way to allow a cross tabulation of the results or
performance of methods used as diagnostic tests for urinary
sensitivity and speci¢city to be calculated. Details on the size, sex,
incontinence in primary and secondary care; to evaluate the
and age of the sample and the care setting and country where the
quality of the evidence and appraise the clinical e⁄ciency of
study was performed were also collected.
alternative diagnostic assessment methods and also to identify
areas for future research on diagnostic methods for urinary
Study Characteristics
incontinence [Martin et al., 2006].
Included studies were categorized according to the diag-
nostic tests they compared. A high number of studies used
multi-channel urodynamics as the reference test. We used the
MATERIALS AND METHODS
International Continence Society (ICS) de¢ned criteria that
multi-channel urodynamics is the gold standard test for diag-
Searching
nosing USI or DO [Abrams et al., 2003]. Decisions on the
The online bibliographic databases Medline (January appropriateness of combining quantitative data, for example,
1966 ^December 2002), Cinahl (January 1982 ^December from similar but not identical tests, were judged clinically.
2002), and Embase (January 1980 ^December 2002) were used A total of 13 di¡erent diagnostic tests were identi¢ed with
to obtain the literature.The search strategies were based on the 22 di¡erent comparisons. Table II displays the number of studies
Cochrane and NHS CRD strategies for identifying studies of that were identi¢ed and which diagnostic tests they compared.
diagnostic performance [NHS Centre for Reviews and Disse- Only the areas of greatest clinical interest will be presented in
mination, 2001]. Keywords, a full list of which can be found in this paper: namely the most commonly used simple tests (clinical

TABLE I. Results of Quality Assessment of Included Studies

Item Yes No Unclear

1 Was the spectrum of patients representative of the patients who will receive the test in practice? 78 12 31
2 Were selection criteria clearly described? 65 47 9
3 Is the reference standard likely to correctly classify the target condition? 102 1 18
4 Is the time period between reference standard and index test short enough to be reasonably sure that the target
condition did not change between the two tests? 43 0 78
5 Did the whole sample or a random selection of the sample, receive veri¢cation using a reference standard of diagnosis? 110 6 5
6 Did patients receive the same reference standard regardless of the index test result? 104 2 15
7 Was the reference standard independent of the index test (i.e., the index test did not form part of the reference standard)? 93 2 26
8a Was the execution of the index test described in su⁄cient detail to permit replication of the test? 78 25 18
8b Was the execution of the reference standard described in su⁄cient detail to permit its replication? 71 33 17
9a Were the index test results interpreted without knowledge of the results of the reference standard? 23 3 95
9b Were the reference standard results interpreted without knowledge of the results of the index test? 10 11 100
10 Were the same clinical data available when test results were interpreted as would be available when the test is
used in practice? 95 26 0
11 Were uninterpretable/intermediate test results reported? 22 25 74
12 Were withdrawals from the study explained? 24 15 82

Neurourology and Urodynamics DOI 10.1002/nau

676 Martin et al.
TABLE II. Matrix Illustrating Distribution of Literature

Urodynamic
investigations History Scales Pads Diary Battery sEMG

History 42 1a 6a 3a 1a
Scales 8 1a
Pads 7 4a
Diary 4 2a
Ultrasound 9
Urodynamic investigations 37 b
Paper towel test 1a
Physical examination 1a 1a 2a
a
Q-Tip test 4
Algorithm 3a
Battery 2a 1a
Conductance 1a
a
These comparisons are not discussed here but are presented and discussed in full in the main report [Martin et al., 2006].
b
Studies that compared multi-channel urodynamics with other urodynamic tests: cystometry, stress test, ambulatory urodynamics, urethral pressure
pro¢ling.

history, pad-test, urinary diary, validated scales) and advanced
tests (X-ray, ultrasound, and urodynamic investigations such as
the clinical stress test, single-channel urodynamics, urethral pres-
sure pro¢ling (UPP), ambulatory urodynamics) compared with
multi-channel urodynamics. A number of less common tests
were identi¢ed including sEMG and the Q-tip test and the results
of all studies included in the review can be found in the main
systematic review report [Martin et al., 2006].
Due to the di¡erent underlying conditions that cause urin-
ary incontinence in men and women, it is not appropriate to
generalize ¢ndings on the accuracy of diagnostic tests for urin-
ary incontinence from men to women and vice versa. In the
systematic review data from men and women were analyzed
separately to provide measures of diagnostic accuracy for
either men or women.
Quantitative Data Analysis
Studies which reported the results of applying the same
diagnostic procedure using the same threshold value (cut-o¡)
were pooled using random e¡ects meta-analysis models in
order to produce pooled estimates of sensitivity and speci¢-
city together with associated 95% con¢dence intervals. Sensi-
tivity refers to the proportion of people with disease who have
a positive test result, speci¢city to the proportion of people
without disease who have a negative test result.
For a study to be included in a meta-analysis the results had
to be presented in a way that allowed the number of false posi-
tives, true positives, false negatives, and true negatives to be
calculated. Therefore, either individual patient data or a
cross-tabulation of the results of the index test by the results
of the reference standard was required.
RESULTS
A total of 6,099 papers were identi¢ed by the electronic
searches. After exclusion processes involving reading
Neurourology and Urodynamics DOI 10.1002/nau
abstracts and full-text papers, 121 di¡erent, original studies
met the inclusion criteria of the review: papers that provided
a quantitative comparison of two or more diagnostic tests used
for the detection of urinary incontinence (Fig. 1). Forty-six of
these papers (38%) did not present data in a way that allowed a
full contingency table to be constructed and, therefore, could
not be included in any quantitative data analysis. Some studies
reported just the correlation between the two diagnostic tests;
others reported the percentage of agreement between the two
diagnoses.
Quality Assessment
Table I presents the results from the quality assessment of
the papers included in the review. There was a lot of variation
in terms of the quality of the reporting of the studies. The
items that resulted in the most favorable ratings were ques-
tions 3, 5, 6, and 7, which were all concerned with the quality
of study design. A number of items were poorly described in
the papers: 39% of the papers did not clearly describe the selec-
tion criteria used in the study, which made it impossible to
judge the appropriateness of the sample. It was unclear in
79% and 83% of papers whether the reference or index tests
were interpreted without knowledge of the other test. Sixty-
four percent of papers did not report the time period between
the two diagnostic tests: this information is required to deter-
mine whether a patient’s condition could have altered between
the two tests, although it is likely that the tests were performed
either on the same day or within a few days this could not be
assumed.
Following a pilot of the QUADAS tool a number of clari¢-
cations were added to the instructions to make the questions
more objective. This included stating that the maximum time
period between the two tests for question 4 to be answered
positively was 24 hr and that no assumptions should be made,
for example, when judging the period between the two tests. It

Fig. 1. Flow diagram for studies.
can be seen fromTable I that all studies were rated either ‘‘yes’’
or ‘‘unclear’’ on question 4: the studies either reported that
both tests were performed on the same day or they did not
state what the gap between the tests was.
Diagnostic Accuracy
Evidence on a large number of tests for diagnosis of urinary
incontinence was identi¢ed; Table II presents the literature
according to the two diagnostic tests compared. Table III pro-
vides data on characteristics and performance of the included
studies broken down by each diagnostic test. In addition,
results of each meta-analysis are also provided.
The diagnosis of urinary incontinence in men will not be
discussed in this paper: little evidence was found on the diag-
nosis of men and what exists is disparate and fragmented,
therefore it is di⁄cult to draw any ¢rm conclusions about test
accuracy. A full discussion of all the papers included in the
systematic review can be found in the original report [Martin
et al., 2006].
Neurourology and Urodynamics DOI 10.1002/nau
Systematic Review of Diagnostic Assessments for UI 677
Primary Care Tests
Clinical history. Fifteen studies compared diagnosis made
by clinical history with thatby multi-channelurodynamics and
presented results in a form that allowed data to be combined to
provide apooled sensitivityof 0.92 (95% C.I.: 0.91^ 0.93) and spe-
ci¢city of 0.56 (0.53 ^ 0.60) for the diagnosis of USI based on the
presence ofstress incontinence symptoms.
Eight papers studied the use of history of urge symptoms
for the diagnosis of DO. Symptoms of urge incontinence were
found to be 0.61 (0.57^ 0.65) sensitive and 0.87 (0.85 ^ 0.89) spe-
ci¢c for the diagnosis of DO in women.
Validated scales. Eight studies compared the use of vali-
dated scales with multi-channel urodynamics. A number of
di¡erent scales were investigated, therefore it was not possible
to combine these studies. Two studies looked speci¢cally at
question 3 of the UDI (short form) scale. Due to the homoge-
neity of these papers it was possible to combine the data to
produce a pooled sensitivity of 0.87 (0.82 ^ 0.92) and speci¢city
of 0.60 (0.51^ 0.69) for the diagnosis of USI from question 3 of
the UDI-6.
Pad-test. Seven studies investigated the diagnostic accu-
racy of a number of di¡erent pad-tests, however only two stu-
dies provided su⁄cient data to allow calculation of diagnostic
performance. As these papers studied two di¡erent pad-tests
the data could not be pooled. One paper found the ICS 1-hr
pad-test to be 0.94 sensitive (0.73 ^ 0.99) and 0.44 speci¢c
(0.28 ^ 0.61) for diagnosing any leakage compared with multi-
channel urodynamics [Jorgensen et al., 1987].Versi et al. [1996]
found the 48-hr pad-test to be 0.92 sensitive (0.82 ^ 0.97) and
0.72 speci¢c (0.57^ 0.83) for the diagnosis of USI.
Urinary diary. Out of four studies on the diagnostic accu-
racy of urinary diaries only one provided cross-tabulated
results. This showed that a scale derived from a 7-day diary
was 0.88 sensitive (0.71^ 0.96) and 0.83 speci¢c (0.77^ 0.87) for
diagnosis of DO in women [Contreras Ortiz et al., 1993].
Secondary Care Tests
Imaging ultrasound. A total of seven studies investigated
the use of ultrasound for the diagnosis of urinary incontinence
and provided enough data to allow cross-tabulation of results.
Four studies used the imaging of leakage from the bladder
during stress as an indicator of USI, which was found to be 0.89
(0.84 ^ 0.93) sensitive and 0.82 (0.73 ^ 0.89) speci¢c.
Three studies used ultrasound imagining of bladder neck
descent during stress for the diagnosis of USI in women
compared with multi-channel urodynamics. The data from
these studies were combined to provide a pooled sensitivity
of 0.84 (0.76 ^ 0.90) and speci¢city of 0.86 (0.79 ^ 0.91).
X-ray. Four studies investigated the use of X-ray imaging of
thelower urinarytractfor thediagnosisofurinaryincontinence
in women. Two studies used observance of leakage from the
bladder under stress as the indicator for presence of USI and
678 Martin et al.
TABLE III. Raw Data and Meta-Analysis Data From Included Studies

True False False True

Reference positives positives negatives negatives Sensitivity 95% CI Speci¢city 95% CI

Clinical history for USI in womena

Cundi¡ et al. [1997] 416 60 17 42 0.96 0.94 ^ 0.98 0.41 0.32 ^ 0.51
De Muylder et al. [1992] 228 58 14 108 0.94 0.91^ 0.97 0.65 0.57^ 0.72
Diokno et al. [1990] 65 14 30 52 0.68 0.58 ^ 0.78 0.79 0.67^ 0.88
Diokno et al., [1987] 145 40 9 6 0.94 0.89 ^ 0.97 0.13 0.05 ^ 0.26
FitzGerald and Brubaker [2002] 187 51 22 33 0.90 0.85 ^ 0.93 0.39 0.29 ^ 0.51
Ishiko et al. [2002] 152 4 14 28 0.92 0.86 ^ 0.95 0.88 0.71^ 0.97
Korda et al. [1987] 451 39 52 24 0.90 0.87^ 0.92 0.38 0.26 ^ 0.51
Kujansuu and Kauppila [1982] 46 20 11 43 0.81 0.68 ^ 0.90 0.68 0.55 ^ 0.79
Lagro-Janssen et al. [1991] 76 9 3 15 0.96 0.89 ^ 0.99 0.63 0.41^ 0.81
Niecestro et al. [1992] 13 17 3 32 0.81 0.54 ^ 0.96 0.65 0.50 ^ 0.78
Ouslander et al. [1987] 82 31 5 17 0.94 0.87^ 0.98 0.35 0.22 ^ 0.51
Ramsay et al. [1993] 72 28 19 81 0.79 0.69 ^ 0.87 0.74 0.65 ^ 0.82
Sand et al. [1988] 114 43 0 66 1.00 0.97^1.00 0.61 0.51^ 0.70
Sandvik et al. [1995] 179 26 4 27 0.98 0.95 ^ 0.99 0.51 0.37^ 0.65
Sunshine and Glowacki [1989] 73 14 0 15 1.00 0.95 ^1.00 0.52 0.33 ^ 0.71
Pooled (RE) 0.92 0.91^ 0.93 0.56 0.53 ^ 0.60
LRþ 2.09 [95% CI: 1.83 ^ 2.35]
Clinical history for DO in women
Ishiko et al. [2002] 25 6 4 154 0.86 0.68 ^ 0.96 0.96 0.92 ^ 0.99
Cundi¡ et al. [1997] 42 17 60 416 0.41 0.32 ^ 0.51 0.96 0.94 ^ 0.98
Sandvik et al. [1995] 23 8 18 187 0.56 0.40 ^ 0.72 0.96 0.92 ^ 0.98
De Muylder et al. [1992] 147 91 89 81 0.62 0.56 ^ 0.69 0.47 0.40 ^ 0.55
Lagro-Janssen et al. [1991] 11 4 7 81 0.61 0.36 ^ 0.83 0.95 0.88 ^ 0.99
Sand et al. [1988] 10 3 20 185 0.33 0.17^ 0.53 0.98 0.95 ^1,00
FitzGerald and Brubaker [2002] 10 21 27 235 0.27 0.14 ^ 0.44 0.92 0.88 ^ 0.95
Cantor and Bates [1980] 107 53 11 43 0.91 0.84 ^ 0.95 0.45 0.35 ^ 0.55
Pooled (RE) 0.61 0.57^ 0.65 0.87 0.85 ^ 0.89
LRþ 4.69 [95% CI: 4.05 ^5.33]
Clinical history for DO in elderly womenb
Diokno et al. [1987] 2 6 12 180 0.14 0.02 ^ 0.43 0.97 0.93 ^ 0.99
Ouslander et al. [1987] 12 10 25 88 0.32 0.18 ^ 0.50 0.90 0.82 ^ 0.95
Pooled (RE)
Validated scale
UDI-6 scale for USI in women
Lemack and Zimmern [1999] 39 30 7 52 0.85 0.71^ 0.94 0.63 0.52 ^ 0.74
FitzGerald and Brubaker [2002] 135 22 18 27 0.88 0.82 ^ 0.93 0.55 0.40 ^ 0.69
Pooled (RE) 0.87 0.82 ^ 0.92 0.60 0.51^ 0.69
LRþ 2.18 [95% CI: 1.49 ^ 2.86]
DIS scale for USI in womenc
Klovning et al. [1996] 92 20 61 68 0.60 0.52 ^ 0.68 0.77 0.67^ 0.85
Pad-test
Jorgensen et al. [1987] 16 18 1 14 0.94 0.73 ^ 0.99 0.44 0.28 ^ 0.61
One additional paper did not provide su⁄cient data to be included in the analysisd
48-hr pad-test for USI in women
Versi et al. [1996] 57 12 5 31 0.92 0.82 ^ 0.97 0.72 0.57^ 0.83
Urinary diary for DO in womene
Contreras Ortiz et al. [1993] 23 33 3 158 0.88 0.71^ 0.96 0.83 0.77^ 0.87
Ultrasound
observed leakage for USI in womenf
Dietz et al. [1999] 10 3 6 18 0.63 0.35 ^ 0.85 0.86 0.64 ^ 0.97
Dietz and Wilson [1998] 66 9 13 29 0.84 0.74 ^ 0.91 0.76 0.60 ^ 0.89
Dietz and Clarke [2001] 33 2 2 15 0.94 0.81^ 0.99 0.88 0.64 ^ 0.99
Quinn et al. [1989] 87 6 3 28 0.97 0.91^ 0.99 0.82 0.66 ^ 0.93
Pooled (RE) 0.89 0.84 ^ 0.93 0.82 0.73 ^ 0.89
LRþ 4.94 [95% CI: 3.88 ^ 6.01]

Neurourology and Urodynamics DOI 10.1002/nau

TABLE III. (Continued)
True False False True
Reference positives positives negatives negatives Sensitivity 95% CI Speci¢city 95% CI
g
Ultrasound
bladder neck descent for USI in women
Chen et al. [1997] 27 15 10 50 0.73 0.56 ^ 0.86 0.77 0.65 ^ 0.87
Bergman et al. [1988a] 38 2 6 45 0.86 0.73 ^ 0.95 0.96 0.86 ^1.00
Bergman et al. [1988b] 30 3 2 26 0.94 0.79 ^ 0.99 0.90 0.73 ^ 0.98
Pooled (RE) 0.84 0.76 ^ 0.90 0.86 0.79 ^ 0.91
LRþ 6.00 [95% CI: 4.72 ^ 7.28]
X-ray
observed leakage for USI in women
Pelsang and Bonney [1996] 37 29 24 69 0.61 0.47^ 0.73 0.70 0.60 ^ 0.79
Scotti et al. [1990] 53 18 35 68 0.60 0.49 ^ 0.71 0.79 0.69 ^ 0.87
Pooled (RE) 0.60 0.52 ^ 0.68 0.74 0.68 ^ 0.81
LRþ 2.31 [95% CI: 1.62 ^3.00]
X-ray
bladder neck descent for USI in women
Grischke et al. [1991] 20 20 14 30 0.59 0.41^ 0.75 0.60 0.45 ^ 0.74
Bergman et al. [1988] 32 15 0 12 1.00 0.89 ^1.00 0.44 0.26 ^ 0.65
Pooled (RE) 0.79 0.67^ 0.88 0.55 0.43 0.66
LRþ 1.76 [95% CI: 0.90 ^ 2.61]
Full bladder clinical stress test for USI in women (supine)h
Hsu et al. [1999] 29 1 2 9 0.94 0.79 ^ 0.99 0.90 0.56 ^1.00
Kadar [1988] 14 5 4 14 0.78 0.52 ^ 0.94 0.74 0.49 ^ 0.91
Scotti and Myers [1993] 68 10 13 54 0.84 0.74 ^ 0.91 0.84 0.73 ^ 0.92
Pooled (RE) 0.85 0.78 ^ 0.91 0.83 0.74 ^ 0.90
LRþ 5.00 [95% CI: 3.79 ^ 6.21]
Single-channel urodynamics (SCU)i
Standing SCU for DO in women
Sand et al. [1991] 43 15 8 34 0.84 0.71^ 0.93 0.69 0.55 ^ 0.82
Sand et al. [1987] 15 8 46 134 0.25 0.15 ^ 0.37 0.94 0.89 ^ 0.98
Sutherst and Brown [1984] 35 7 0 58 1.00 0.90 ^1.00 0.89 0.79 ^ 0.96
Pooled (RE) 0.63 0.55 ^ 0.71 0.88 0.84 ^ 0.92
LRþ 12.00 [95% CI: 10.58 ^13.42]
Supine SCU for DO in elderly women
Fonda et al. [1993] 17 7 4 15 0.81 0.58 ^ 0.95 0.68 0.45 ^ 0.86
Ouslander et al. [1988] 61 10 23 43 0.73 0.62 ^ 0.82 0.81 0.68 ^ 0.91
Pooled (RE) 0.74 0.65 ^ 0.82 0.77 0.66 ^ 0.86
LRþ 3.57 [95% CI: 2.41^ 4.73]
Supine SCU for DO in elderly men
Fonda et al. [1993] 20 0 1 6 0.95 0.76 ^1.00 1.00 0.54 ^1.00
Ouslander et al. [1988] 21 1 4 5 0.84 0.64 ^ 0.96 0.833 0.36 ^1.00
Pooled (RE) 0.89 0.76 ^ 0.96 0.92 0.62 ^1.00
LRþ 18.20 [95% CI: 12.62 ^ 23.78]
Supine SCU for USI in women
Resnick et al. [1996] 19 10 3 64 0.86 0.67^ 0.95 0.86 0.77^ 0.92
Ambulatory Urodynamics for USI in women
Davis et al. [1998] 7 38 2 3 0.78 0.45 ^ 0.94 0.07 0.03 ^ 0.19
Sitting urethral pressure pro¢le (UPP) for USI in women
Swift and Ostergard [1995] 32 1 33 42 0.49 0.37^ 0.62 0.98 0.88 ^1.00
Richardson [1986] 30 43 5 59 0.86 0.70 ^ 0.95 0.58 0.48 ^ 0.68
Pooled (RE) 0.62 0.52 ^ 0.72 0.70 0.61^ 0.77
LRþ 9.38 [95% CI: 6.91^11.84]
Supine UPP for USI in womenj
Versi [1990] 54 21 16 81 0.77 0.66 ^ 0.85 0.79 0.71^ 0.86
a
An additional six papers did not provide su⁄cient data to be included in the analysis.
b
An additional three papers did not provide su⁄cient data to be included in the analysis.
c
One additional paper did not provide su⁄cient data to be included in the analysis.
d
ICS 1-hr pad-test for any leakage in women.
e
One additional paper did not provide su⁄cient data to be included in the analysis.
f
Two additional papers did not provide su⁄cient data to be included in the analysis.
g
Two additional papers did not provide su⁄cient data to be included in the analysis.
h
One additional paper did not provide su⁄cient data to be included in the analysis.
i
One additional paper did not provide su⁄cient data to be included in the analysis.
j
Two additional papers did not provide su⁄cient data to be included in the analysis.

Neurourology and Urodynamics DOI 10.1002/nau

680 Martin et al.
found this to be 0.60 (0.53 ^ 0.69) sensitive and 0.75 (0.69 ^ 0.82)
speci¢c.Two studies used the descentof thebladder neck during
stress and found this to be 0.98 (0.93 ^1.00) sensitive and 0.56
(0.46 ^ 0.68) speci¢c for the diagnosis of USI inwomen.
Stress test. Sixstudieswereidenti¢edthatcomparedtheuse
ofastresstest with multi-channelurodynamicsforthediagnosis
ofUSIandthequalityofreportingofthestudiesinthisgroupwas
high with all six papers presenting full contingency tables.
However there were di¡erences in the way the stress tests were
performed.Two studies used the supine stress test, one with the
bladder ¢lled with 200 ml saline, the other with an empty
bladder. Two papers used a standing stress test, both with a
full bladder (>200 ml). One paper performed the stress test in
both the supine and standing position with a full bladder. One
paper performed a cough-stress test with the patient sitting in
the erect position, however, the diagnosis was also dependant
on the result of single-channel urodynamics. As a result of these
di¡erences each type of stress test was analyzed separately.The
only stress test to have su⁄cient data to allow calculation of
measures of diagnostic accuracy was the full bladder supine
clinical stress test. The data from three studies was pooled to
provide a sensitivity of 0.85 (0.78 ^ 0.91) and speci¢city of 0.83
(0.74 ^ 0.90) for the diagnosis of USI in women compared with
multi-channelurodynamics.
Single-channel urodynamics. Eightstudies wereidenti¢ed
that compared the use of single-channel urodynamics with
multi-channel urodynamics. All studies were conducted in a
secondary care setting. Three of these studies investigated
elderly populations (greater than 70,60,and 65 years).
Data from two papers were combined to provide a pooled
sensitivity of 0.74 (0.65 ^ 0.82) and speci¢city of 0.77 (0.66 ^
0.86) for the diagnosis of DO in elderly women using supine
single-channel cystometry. Data from three papers were
combined to provide a pooled sensitivity of 0.63 (0.55 ^ 0.71)
and speci¢city of 0.88 (0.84 ^ 0.92) for the diagnosis of DO in
women using standing single-channel cystometry.
Urethral pressure pro¢ling. Two studies investigated the
use of UPP for the diagnosis of USI in women. The data were
combined to provide apooled sensitivityof 0.62 (0.52 ^ 0.72) and
speci¢city of 0.70 (0.61^ 0.77) for the diagnosis of USI in women
by urethralpressure pro¢ling inthe sitting position.
Ambulatory urodynamics. Ambulatory urodynamic
monitoring is the monitoring of leakage, £ow recordings, and
pressure in the bladder and abdomen with or without pressure
in the urethra in an ambulatory setting [Harvey et al., 2001]. Six
studies were identi¢ed that compared the use of ambulatory
urodynamics with standard multi-channel urodynamics.
Neurourology and Urodynamics DOI 10.1002/nau
However, due to the variability of the samples used in this
group, some studies included only female patients, some only
male and others studied a mixed population, it was not possible
to combine the data from these papers. The sensitivities and
speci¢cities demonstrated by these studies are heterogeneous.
It is not possible therefore, to draw any conclusions about the
e⁄cacyofambulatory urodynamics.
DISCUSSION
This is the ¢rst systematic review of methods of diagnosing
urinary incontinence. The search strategy used to identify
relevant papers was based on the Cochrane and NHS CRD
strategies used for identifying studies of diagnostic perfor-
mance, which is well validated. It is important, for consistency
and accuracy, for systematic reviews of diagnostic methods to
use these strategies.
The literature dealing with the diagnosis of urinary incon-
tinence is highly fragmented and there is a lack of literature
dealing with the diagnosis of urinary incontinence or bladder
outlet obstruction in men.The quality of the reporting-studies
included was poor in terms of both the reporting the study
methodology as well as the study results. This restricted the
number of studies suitable for inclusion in meta-analyses:
38% of the 121 relevant papers did not present data in a form
that could allow its inclusion in a meta-analysis and because of
this clinical interpretation was di⁄cult. Few studies reported
how they were funded: the amount and source of funding that
a study has, can a¡ect the likelihood of it being published and
this is one potential source of publication bias.
Two potential limitations of the review are that included
papers were restricted to those published in English. Due to
time constraints it was not possible to include papers in other
languages, time restrictions also prevented hand searching of
key journals. These two factors may have resulted in some
relevant papers being excluded from the review.
Primary Care Tests
The pooled sensitivity and speci¢city values of clinical his-
tory for diagnosis of women suggest that a clinical history is
highly sensitive but less speci¢c in diagnosing USI. These ¢nd-
ings suggest that a large proportion of women with USI can be
correctly diagnosed in primary care and that initiating low
risk, low-cost behavioral treatment at this stage may be appro-
priate.
It is not possible to recommend the use of a particular vali-
dated scale for diagnosing urinary incontinence due to insu⁄-
cient and fragmented evidence; the eight papers identi¢ed in
the review studied seven di¡erent scales. Future work in this
area should concentrate on validating existing scales, in parti-
cular those prepared by the WHO-sponsored International
Consultation on Incontinence and recommended by the ICS,
rather than attempting to develop new instruments.

Little evidence was identi¢ed on the performance of the
diagnostic care tests commonly used in primary care such as
urinary diaries, pad-tests, and validated scales and much of the
existing evidence on these tests could not be synthesized due
to the large number of di¡erent tests that were investigated.
For example, eight studies compared a pad-test with multi-
channel urodynamics, however four di¡erent types of pad-
tests were used and of the eight studies only two reported data
in a form that could be included in a statistical analysis.
There is a need for the development and standardization of
scales, pad-tests, and diaries for use in the diagnosis and mea-
surement of severity of urinary incontinence. It should also be
pointed out that, even if not having huge value for diagnosis,
simple primary care investigations have a valuable function as
a measure of severity and impact of urinary incontinence.
Secondary Care Tests
A number of papers compared the clinical stress test with
multi-channel urodynamics for the diagnosis of USI resulting
in a high sensitivity and speci¢city. These studies performed
the clinical stress test with an arti¢cially ¢lled bladder, which
increases the invasiveness of the test, however if the test could
be performed with a naturally full bladder, with no signi¢cant
detriment to diagnostic accuracy, then this could be a useful
non-invasive diagnostic test that could be utilized in both pri-
mary and secondary care. Research into such a test could be of
clinical interest.
A large amount of literature was identi¢ed that dealt with
imaging the lower urinary tract for the diagnosis of USI
through ultrasound and X-ray methods. Ultrasound was
found to be the most e¡ective method of imaging leakage from
and movement of the bladder. This suggests that ultrasound
imaging may be a useful screening test, especially considering
the lower costs and risks associated with it, however, ultra-
sound should not be considered as a replacement for the
gold-standard of full multi-channel urodynamic testing.
The studies included in the review do not give details on
which clinical personnel performed the ultrasound investiga-
tions, therefore, it is di⁄cult to conclude whether ultrasound
could be performed as accurately in primary care by non-spe-
cialist sta¡. Further research into the accuracy and utility of
ultrasound for the diagnosis of urinary incontinence in both
secondary and primary care is required. A test that requires
signi¢cant and specialized training would unlikely to be suita-
ble for use as a screening tool.
It is not clear why imaging by ultrasound should be more
e¡ective than by X-ray. A possible reason for this could be that
women ¢nd it di⁄cult to reproduce clinical symptoms, such
as leakage due to coughing or straining, during X-ray testing
conditions. Further investigation is required into the diagno-
sis of urinary incontinence by X-ray imaging as just four
papers in the review dealt with this.
There is an issue with ambulatory urodynamics that it is
thought by some experts to be more sensitive than standard
Neurourology and Urodynamics DOI 10.1002/nau
Systematic Review of Diagnostic Assessments for UI 681
multi-channel urodynamics and should be the true gold stan-
dard for the diagnosis of urinary incontinence. However, the
view of the ICS is that ambulatory urodynamics is overly sen-
sitive but not very speci¢c in detecting urinary leakage.
Ambulatory urodynamics has not been standardized by the
ICS and therefore cannot be recommended for routine clinical
practice. The International Consultation on Incontinence
(ICI) group on urodynamics in 2002 concluded that further
study of the place and advantages of ambulatory monitoring
was necessary [Homma et al., 2002].
Within the review, the number of studies undertaken in pri-
mary care was far fewer than those undertaken in secondary
care settings. This has important implications for interpreta-
tion of the ¢ndings. The studies undertaken in secondary care
are mainly undertaken on referred patients attending as out-
patients.They are very di¡erent from undi¡erentiated patients
presenting in primary care. It is likely that referred patients
have already undergone some form of diagnostic process and
therefore, using various diagnostic assessment tools with this
population may produce greater levels of sensitivity and
speci¢city than in a mainly un-referred, undi¡erentiated
population.
It is critical to make a distinction between tests and assess-
ment methods that can be undertaken in primary care and
those that can only be undertaken in secondary care. The
majority of diagnostic and assessment processes can be under-
taken in primary care and comprise clinical history taking, the
use of scales, physical examination, and simple tests such as
diaries and pad-tests. These tests are simple, are low in cost
and carry low risks. Such tests may o¡er useful information
on severity which when combined with history may provide
su⁄cient information to commence primary care interven-
tions (which are low cost and low risk).
Research Recommendations
There is clearly a need for large scale, high quality, primary
studies evaluating the systematic use of a number of diagnostic
methods in a primary care setting in order that the results of
this systematic review can be veri¢ed or not. Such studies
should not only include an assessment of clinical e¡ectiveness,
in this case diagnostic accuracy, but also include an assessment
of costs and quality of life/patient acceptance/satisfaction to
inform future health policy decisions. Only a small number
of studies investigated the diagnosis of UI using an algorithm
or battery of tests. Such a ‘‘common sense’’ approach which
mirrors clinical practice warrants further investigation.
Care should be taken when reporting studies of diagnostic
accuracy.The recommendations of the standards for reporting
diagnostic accuracy (STARD) initiative should be followed to
ensure the accuracy and completeness of reporting design and
results [Bossuyt et al., 2003]. This will result in more studies
ful¢lling the inclusion criteria for future evidence-based
investigations. The updating of this review should be consid-
ered within 4 ^ 6 years.
682 Martin et al.
CONCLUSION
A large proportion of women with USI can be diagnosed in
primary care from clinical history alone. Initiating low risk,
low-cost treatments on the basis of this diagnosis may be
appropriate, however, higher risk interventions such as sur-
gery should be based upon the gold-standard of full multi-
channel urodynamic testing. In secondary care, ultrasound
imaging may o¡er a valuable alternative to urodynamics inves-
tigation. The clinical stress test is e¡ective in the diagnosis of
USI. More research is needed on the diagnostic and screening
tests used in primary care and on diagnosis of men.
ACKNOWLEDGMENTS
We acknowledge the contributions of all the members of the
investigative team: David Turner, Christine Shaw, Francine
Cheater, and Christopher Chapple.
REFERENCES
Bergman A, McKenzie C, Ballard CA, et al. 1988. Role of cystourethrography
in the preoperative evaluation of stress urinary incontinence in women.
J Reprod Med 33:372 ^ 6.
Abrams P, Cardozo L, Fall M, et al. 2003. The standardisation of terminology
in lower urinary tract function: Report from the standardisation sub-com-
mittee of the International Continence Society. Urology 61:37^ 49.
Bergman A, Ballard CA, Platt LD. 1988a. Ultrasonic evaluation of urethrove-
sical junction in women with stress urinary incontinence. J Clin Ultra-
sound 16:295 ^300.
Bergman A, McKenzie CJ, Richmond J, et al. 1988b. Transrectal ultrasound
versus cystography in the evaluation of anatomical stress urinary inconti-
nence. BJU 62:228 ^34.
Bossuyt P, Reitsma J, Bruns D, et al. 2003. Towards complete and accurate
reporting of studies of diagnostic accuracy: The STARD initiative. BMJ
326:41^ 4.
Cantor TJ, Bates CP. 1980. A comparative study of symptoms and objective
urodynamic ¢ndings in 214 incontinent women. BJOG 87:889 ^ 92.
Chen GD, Su TH, Lin LY. 1997. Applicability of perineal sonography in ana-
tomical evaluation of bladder neck in women with and without genuine
stress incontinence. J Clin Ultrasound 25:189 ^ 94.
Contreras Ortiz O, Lombardo RJ, Pellicari A. 1993. Non-invasive diagnosis
of bladder instability using the Bladder Instability Discriminant Index
(BIDI). Zentralbl Gynakol 115:446 ^ 9.
Cundi¡ GW, Harris RL, Coates KW, et al. 1997. Clinical predictors of urinary
incontinence in women. Am J Obstet Gynecol 177:262 ^ 7.
Davis G, McClure G, Sherman R, et al. 1998. Ambulatory urodynamics of
female soldiers. Mil Med 163:808 ^12.
De Muylder X, Claes H, Neven P, et al. 1992. Usefulness of urodynamic inves-
tigations in female incontinence. Eur J Obstet Gynecol Reprod Biol
44:205 ^ 8.
Dietz HP, Clarke B. 2001. Translabial color Doppler urodynamics. Int Uro-
gynecol J Pelvic Floor Dysfunct 12:304 ^ 7.
Dietz HP, Wilson PD. 1998. Anatomical assessment of the bladder outlet and
proximal urethra using ultrasound and videocystourethrography. Int Uro-
gynecol J Pelvic Floor Dysfunct 9:365 ^ 9.
Dietz HP, McKnoulty L, Clarke B. 1999. Translabial color Doppler for ima-
ging in urogynecology: A preliminary report. Ultrasound Obstet Gynecol
14:144 ^ 7.
Diokno AC, Wells TJ, Brink CA. 1987. Urinary incontinence in elderly
women: Urodynamic evaluation. J Am Geriatr Soc 35:940 ^ 6.
Neurourology and Urodynamics DOI 10.1002/nau
Diokno AC, Normolle DP, Brown MB, et al. 1990. Urodynamic tests for
female geriatric urinary incontinence. Urology 36:431^ 9.
FitzGerald MP, Brubaker L. 2002. Urinary incontinence symptom scores
and urodynamic diagnoses. Neurourol Urody 21:30 ^5.
Fonda D, Brimage PJ, D’Astoli M. 1993. Simple screening for urinary incon-
tinence in the elderly: Comparison of simple and multichannel cystome-
try. Urology 42:536 ^ 40.
Grischke EM, Anton H, Stolz W, et al. 1991. Urodynamic assessment and
lateral urethrocystography. A comparison of two diagnostic procedures
for female urinary incontinence. Acta Obstet Gynecol Scand 70:225 ^ 9.
Harvey MA, Kristjansson B, Gri⁄th D, et al. 2001. The Incontinence Impact
Questionnaire and the Urogenital Distress Inventory: A revisit of their
validity in women without a urodynamic diagnosis. Am J Obstet Gynecol
185:25 ^31.
HommaY, Batista J, Bauer D, et al. 2002. In: Abrams P, Cardozo L, Khoury S,
Wein A, editors. Incontinence: 2nd international consultation on inconti-
nence. 2nd edition. Plymouth, UK: Plymbridge Distributors. Ltd. 317^ 72.
Hsu TH, Rackley RR, Appell RA. 1999. The supine stress test: A simple
method to detect intrinsic urethral sphincter dysfunction. J Urol
162:460 ^3.
Ishiko O, Hirai K, Sumi T, et al. 2002. The urinary incontinence score in the
diagnosis of female urinary incontinence. Int J Gynaecol Obstet 68:131^ 7.
Jorgensen L, Lose G, Andersen JT. 1987. One-hour pad-weighing test for
objective assessment of female urinary incontinence. Obstet Gynecol
69:39 ^ 42.
Kadar N. 1988. The value of bladder ¢lling in the clinical detection of urine
loss and selection of patients for urodynamic testing. BJOG 95:698 ^
704.
Klovning A, Hunskaar S, Eriksen BC. 1996. Validity of a scored urological
history in detecting detrusor instability in female urinary incontinence.
Acta Obstet Gynecol Scand 75:941^5.
Korda A, Krieger M, Hunter P, et al. 1987. The value of clinical symptoms in
the diagnosis of urinary incontinence in the female. Aust NZ J Obstet
Gynaecol 27:149 ^51.
Kujansuu E, Kauppila A. 1982. Scored urological history and urethrocysto-
metry in the di¡erential diagnosis of female urinary incontinence. Ann
Chir Gynaeco 71:197^ 202.
Lagro-Janssen AL, Debruyne FM, vanWeel C. 1991.Value of the patient’s case
history in diagnosing urinary incontinence in general practice. Br J Urol
67:569 ^ 72.
Lemack GE, Zimmern PE. 1999. Predictability of urodynamic ¢ndings
based on the Urogenital Distress Inventory-6 questionnaire. Urology
54:461^ 6.
Martin JL, Williams KS, Abrams KR, et al. 2006. Systematic review and eva-
luation of methods of assessing urinary incontinence. Health Technol
Assess 10.
McGrother CM, Donaldson MMK, Wagg A, et al. 2003. Continence. In: Ste-
vens ARJMJ, editor. Health care needs assessment: The epidemiologically
based needs assessment reviews. Abingdon, UK: Radcli¡e Medical Press
Ltd.
NHS Centre for Reviews and Dissemination. 2001. Undertaking systematic
reviews of research on e¡ectiveness. CRD guidelines for those carrying
out commissioning reviews. CRD Report Number 4. York: NHS Centre
for Reviews and Dissemination 1^50.
Niecestro RM, Wheeler JS, Nanninga J, et al. 1992. Use of stresscath for diag-
nosing stress incontinence. Urology 39:266 ^ 9.
Ouslander J, Staskin D, Raz S, et al. 1987. Clinical versus urodynamic diag-
nosis in an incontinent geriatric female population. J Urol 137:68 ^ 71.
Ouslander J, Leach G, Abelson S, et al. 1988. Simple versus multichannel
cystometry in the evaluation of bladder function in an incontinent geria-
tric population. J Urol 140:1482 ^ 6.
Pelsang RE, Bonney WW. 1996. Voiding cystourethrography in female stress
incontinence. AJR Am J Roentgenol 166:561^5.
Quinn MJ, Fanrsworth BA, Pollard WJ, et al. 1989. Vaginal ultrasound in the
diagnosis of stress incontinence: A prospective comparison to urody-
namic investigations. Neurourol Urodyn 8:302 ^3.

Ramsay IN, Hilton P, Rice N. 1993. The symptomatic characterization of
patients with detrusor instability and those with genuine stress inconti-
nence. Int Urogynecol J 4:23 ^ 6.
Resnick NM, Brandeis GH, Baumann MM, et al. 1996. Misdiagnosis of urin-
ary incontinence in nursing home women: Prevalence and a proposed
solution. Neurourol Urodyn 15:599 ^ 618.
Richardson DA. 1986. Value of the cough pressure pro¢le in the evaluation of
patients with stress incontinence. Am J Obstet Gynecol 155:808 ^11.
Sand PK, Hill RC, Ostergard DR. 1987. Supine urethroscopic and standing
cystometry as screening methods for the detection of detrusor instability.
Obstet Gynecol 70:57^ 60.
Sand PK, Hill RC, Ostergard DR. 1988. Incontinence history as a predictor of
detrusor stability. Obstet Gynecol 71:257^ 60.
Sand PK, Brubaker LT, Novak T. 1991. Simple standing incremental cystometry
as a screening method for detrusor instability. Obstet Gynecol 77:453 ^7.
Sandvik H, Hunskaar S, Vanvik A, et al. 1995. Diagnostic classi¢cation of
female urinary incontinence: An epidemiological survey corrected for
validity. J Clin Epidemiol 48:339 ^ 43.
Scotti RJ, Myers DL. 1993. A comparison of the cough stress test and single-
channel cystometry with multichannel urodynamic evaluation in genuine
stress incontinence. Obstet Gynecol 81:430 ^3.
Neurourology and Urodynamics DOI 10.1002/nau
Systematic Review of Diagnostic Assessments for UI 683
Scotti RJ, Ostergard DR, Guillaume AA, et al. 1990. Predictive value of ure-
throscopy as compared to urodynamics in the diagnosis of genuine stress
incontinence. J Reprod Med 35:772 ^ 6.
Sunshine TJ, Glowacki GA. 1989. Clinical correlation of urodynamic testing
in patients with urinary incontinence. J Gynecol Surg 5:93 ^ 8.
Sutherst JR, Brown MC. 1984. Comparison of single and multichannel cysto-
metry in diagnosing bladder instability. Br Med J Clin Res Ed. 288:1720 ^
2.
Swift SE, Ostergard DR. 1995. Evaluation of current urodynamic testing
methods in the diagnosis of genuine stress incontinence. Obstet Gynecol
86:85 ^ 91.
Turner DA, Shaw C, McGrother CW, et al. 2004. MRC incontinence team.
The cost of clinically signi¢cant urinary storage symptoms for commu-
nity dwelling adults in the UK. BJU Int 93:1246 ^52.
Versi E. 1990. Discriminant analysis of urethral pressure pro¢lometry data
for the diagnosis of genuine stress incontinence. BJOG 97:251^ 9.
Versi E, Orrego G, Hardy E, et al. 1996. Evaluation of the home pad test in the
investigation of female urinary incontinence. BJOG 103:162 ^ 7.
Whiting P, Rutjes AWS, Dinnes J, et al. 2004. Development and validation of
methods for assessing the quality of diagnostic accuracy studies. Health
Technol Assess 8.