Content uploaded by Abigail Newlands
Author content
All content in this area was uploaded by Abigail Newlands on Feb 22, 2024
Content may be subject to copyright.
Received: 29 August 2023
|
Accepted: 5 February 2024
DOI: 10.1002/nau.25426
CLINICAL ARTICLE
Evaluating the quality of life impact of recurrent urinary
tract infection: Validation and refinement of the
Recurrent UTI Impact Questionnaire (RUTIIQ)
Abigail F. Newlands BSc, MSc
1
|Melissa Kramer BBus, BA
2
|
Lindsey Roberts BSc, MSc, PhD, CPsychol
3
|Kayleigh Maxwell BSc, MSc
4
|
Jessica L. Price
2
|Katherine A. Finlay PhD, CPsychol, AFBPsS
1
1
School of Psychology and Clinical
Language Sciences, University of
Reading, Reading, UK
2
Live UTI Free Ltd, Sandyford, Dublin,
Ireland
3
School of Psychology, University of
Buckingham, Buckingham, UK
4
Department of Psychology, Faculty of
Natural Sciences, University of Stirling,
Stirling, UK
Correspondence
Katherine A. Finlay, PhD, CPsychol,
AFBPsS, School of Psychology and
Clinical Language Sciences, University of
Reading, Reading, RG6 7BE, UK.
Email: katherine.finlay@reading.ac.uk
Abstract
Background and Aims: Recurrent urinary tract infection (rUTI) has
significant negative consequences for a wide variety of quality of life (QoL)
domains. Without adequate validation and assessment of the unique insights
of people living with rUTI, clinical results cannot be fully understood. The
Recurrent UTI Impact Questionnaire (RUTIIQ), a novel patient‐reported
outcome measure of rUTI psychosocial impact, has been robustly developed
with extensive patient and clinician input to facilitate enhanced rUTI
management and research. This study aimed to confirm the structural validity
of the RUTIIQ, assessing its strength and bifactor model fit.
Methods: A sample of 389 adults experiencing rUTI (96.9% female, aged
18–87 years) completed an online cross‐sectional survey comprising a
demographic questionnaire and the RUTIIQ. A bifactor graded response
model was fitted to the data, optimizing the questionnaire structure based on
item fit, discrimination capability, local dependence, and differential item
functioning.
Results: The final RUTIIQ demonstrated excellent bifactor model fit
(RMSEA = 0.054, CFI = 0.99, SRMSR = 0.052), and mean‐square fit indices
indicated that all included items were productive for measurement
(MNSQ = 0.52–1.41). The final questionnaire comprised an 18‐item general
“rUTI QoL impact”factor, and five subfactor domains measuring “personal
wellbeing”(three items), “social wellbeing”(four items), “work and activity
interference”(four items), “patient satisfaction”(four items), and “sexual
Neurourol Urodyn. 2024;1–13. wileyonlinelibrary.com/journal/nau
|
1
This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial‐NoDerivs License, which permits use and distribution in any
medium, provided the original work is properly cited, the use is non‐commercial and no modifications or adaptations are made.
© 2024 The Authors. Neurourology and Urodynamics published by Wiley Periodicals LLC.
Abbreviations: CFI, comparative fit index; DIF, differential item functioning; EFA, exploratory factor analysis; GRM, graded response model; IRT,
item response theory; MHRM, Metropolis‐Hastings Robbins‐Monro (estimation method); MNSQ, mean square; QoL, quality of life; RMSEA, root
mean square error of approximation; rUTI, recurrent urinary tract infection; RUTIIQ, Recurrent Urinary Tract Infection Impact Questionnaire;
RUTISS, Recurrent Urinary Tract Infection Symptom Scale; SRMSR, standardized root mean square residual.
wellbeing”(three items). Together, the general factor and five subfactors
explained 81.6% of the common model variance. All factor loadings were
greater than 0.30 and communalities greater than 0.60, indicating good model
fit and structural validity.
Conclusions: The 18‐item RUTIIQ is a robust, patient‐tested questionnaire
with excellent psychometric properties, which capably assesses the patient
experience of rUTI‐related impact to QoL and healthcare satisfaction.
Facilitating standardized patient monitoring and improved shared decision‐
making, the RUTIIQ delivers the unique opportunity to improve patient‐
centered care.
KEYWORDS
bifactor model, chronic pain, item response theory, patient experience, patient‐reported
outcomes, psychosocial outcomes, women's health
1|INTRODUCTION
Recurrent urinary tract infection (rUTI) is defined as
experiencing two or more UTIs in six months or three or
more in a year,
1
and affects more than 100 million people
worldwide annually.
2
Given considerable negative
impact to a broad range of quality of life (QoL)
domains,
3–6
and significant socioeconomic implica-
tions,
4,7–9
there is an urgent need to validate the unique
rUTI patient experience and incorporate QoL assessment
into clinical management and research.
5,10–12
Without
the inclusion of rUTI‐specific patient‐reported outcome
measures (PROMs), evaluation of clinical testing out-
comes and symptoms are limited in their real‐world
application. Such tools are required to improve patient
monitoring, shared decision‐making, and rUTI
management.
13,14
The Recurrent UTI Impact Questionnaire (RUTIIQ)
is a new PROM evaluating rUTI‐related impact to QoL.
15
The RUTIIQ was developed in accordance with gold‐
standard PROM development recommendations by the
COnsensus‐based Standards for the selection of health
Measurement INstruments (COSMIN) initiative,
16,17
with extensive, international expert clinician and patient
input (see Figure 1for methodology; Stages I–IV have
been published in Newlands et al., 2023).
15
A five‐factor
structure was identified by exploratory factor analysis
(EFA) of pilot data, comprising: “personal wellbeing,”
“social wellbeing,”“work and activity interference,”
“patient satisfaction,”and “sexual wellbeing.”
15
The
RUTIIQ demonstrates excellent psychometric properties,
including strong test–retest reliability (intraclass correla-
tion coefficient, ICC = 0.66–0.91; computed based on a
single‐rating, absolute‐agreement, two‐way mixed effects
model
18
), internal consistency (Cronbach's α=0.81–0.96),
content validity (item content validity indices, I‐CVI =
0.75–1.00), and concurrent validity with related QoL
measures (Spearman's ρ=0.69–0.76).
15
As the required next step in the PROM development
process, in preparation for ongoing work to determine
the questionnaire's clinical responsiveness to interven-
tion, the current study aimed to build on preliminary
testing of the RUTIIQ by confirming its structural
validity and identifying areas for refinement.
16
A bifactor
structure was hypothesized, expecting a general “rUTI
QoL impact”factor that explains the common variance
between all items, and five specific, uncorrelated
subfactors aligning with the factors identified through
EFA, each explaining the unique influence of a specific
construct beyond the general factor.
19,20
Bifactor model-
ling enables the evaluation of general scores based on all
items represented by a general trait, as well as individual
domain scores for items represented by specific
traits.
19–21
2|MATERIALS AND METHODS
2.1 |Study design and participants
Adults meeting the diagnostic criteria for rUTI,
1
with
advanced or proficient fluency in English based on the
Common European Framework of Reference for Lan-
guages,
22
participated in an online cross‐sectional survey
(N= 389, 96.9% female biological sex; see Table 1for
characteristics). Definition of rUTI was presented to
participants using the Recurrent UTI Symptom Scale
(RUTISS),
23,24
with participants required to report at
2
|
NEWLANDS ET AL.
15206777, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/nau.25426 by Test, Wiley Online Library on [22/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
least two symptomatic episodes of UTI in the past six
months or at least three in the past year.
1
Inclusion of
participants based on self‐report of symptoms reflected
recent microbiological findings that standard urine
culture misses up to 58% of true infections and published
recommendations regarding the need to prioritize patient
symptom reporting.
25–27
This approach enabled access to
the larger, more heterogeneous sample required for item
response theory (IRT) analysis. Exclusion criteria com-
prised a current diagnosis of interstitial cystitis. Aiming
to validate the confirmatory factor structure and make
refinements to minimize respondent burden, RUTIIQ
data were collected for psychometric analyses including
IRT (see Section 2.3).
19,28,29
At least 250 participants were required to facilitate
multidimensional IRT analysis with a graded response
model (GRM) for ordered polytomous data
30
; sampling
adequacy was exceeded (see Figure S1 for sampling flow
diagram). Participants were mostly recruited via news-
letters and social media posts from a key UTI stakeholder
group, Live UTI Free (84.8%, n= 300), participants and
clinicians sharing the study information on social media
or by word of mouth (9.51%, n= 37), and online UTI‐
focused support groups (5.66%, n= 22).
2.2 |Procedure
After reviewing the study information and ethical
considerations, participants gave electronic consent and
completed a screening questionnaire to confirm elig-
ibility to participate (see Figure S1). Eligible participants
completed the preliminary RUTIIQ, a 30‐item self‐report
questionnaire assessing rUTI‐related impact to QoL with
five domains: “personal wellbeing,”“social wellbeing,”
“work and activity interference,”“patient satisfaction”
with UTI‐related medical care, and “sexual wellbeing”
(optional domain, preceded by a pre‐qualifying question:
“Do you feel your UTI(s) has/have impacted your sex life
in the past 2 weeks?”).
15
Participants used an 11‐point
Likert scale ranging from 0 (“strongly disagree”)to10
(“strongly agree”) to rate their level of agreement with
FIGURE 1 Methodology employed to develop and validate the Recurrent Urinary Tract Infection Impact Questionnaire (RUTIIQ). The
current study reports the methodology and findings from Stage V. Results from Stages I–IV are published in Newlands et al., 2023.
15
NEWLANDS ET AL.
|
3
15206777, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/nau.25426 by Test, Wiley Online Library on [22/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
statements about rUTI impact. After completing the
RUTIIQ, a debrief form signposted participants to
support resources.
2.3 |Data handling and statistical
analysis
The final sample comprised 389 participants (Figure S1).
Eighty‐nine participants did not complete the optional
“sexual wellbeing”questions, responding “no”or “prefer
not to say”to the pre‐qualifying question. This left a total
of 300 included datapoints for multidimensional IRT
analysis. Responses to the “patient satisfaction”questions
were reverse scored, thus higher scores indicated higher
levels of rUTI QoL impact for all domains. As outlined in
this section, a preliminary model was identified, making
refinements based on item and model fit, before
reconducting IRT analysis with a final model. Statistical
terminology definitions are available in Table S1.
TABLE 1 Participant demographic characteristics.
Characteristic n%
Biological sex
Female 377 96.9
Male 12 3.08
Gender
Female 374 96.1
Male 12 3.08
Nonbinary 2 0.51
Prefer not to say 1 0.26
Country of residence
United Kingdom 153 39.3
United States 147 37.8
Canada 26 6.68
Australia 8 2.06
Ireland 5 1.29
Greece 4 1.03
India 4 1.03
Spain 4 1.03
Other
a
38 9.77
Ethnicity
Asian (including Asian American,
Asian British)
10 2.57
Black (including African, African
American, Caribbean, Black British)
3 0.77
Hispanic or Latino American 5 1.29
Mixed ethnicity or multiple ethnic groups 4 1.03
Native Hawaiian or other Pacific Islander 2 0.51
White (including Caucasian,
White British, White European)
340 87.4
Other ethnicity 4 1.03
Prefer not to say 21 5.40
Fluency in English
Native or bilingual 337 86.6
Advanced or proficient 52 13.4
Relationship status
Married or in a civil partnership 199 51.2
In a relationship (unmarried) 117 30.1
Single 44 11.3
Divorced 14 3.60
Widowed 6 1.54
Separated 5 1.29
Other 2 0.51
TABLE 1 (Continued)
Characteristic n%
Prefer not to say 2 0.51
Highest level of education
Some high school/secondary school 7 1.80
High school/secondary school 65 16.7
Bachelor's degree or equivalent 168 43.2
Master's degree or equivalent 104 26.7
Doctoral level training or equivalent 16 4.11
Other professional qualification(s) 22 5.66
Prefer not to say 7 1.80
Annual household income (GBP)
No current income 12 3.08
£1–£9999 15 3.86
£10 000–£24 999 37 9.51
£25 000–£49 999 102 26.2
£50 000–£74 999 48 12.4
£75 000–£99 999 40 10.3
£100 000 or more 59 15.2
Prefer not to say 76 19.5
Note:N= 389.
a
Other countries where n≤3 comprise the following 29 countries listed
alphabetically: Angola, Argentina, Austria, The Bahamas, Belgium, Croatia,
Czech Republic, Denmark, Finland, France, Germany, Iceland, Israel, Italy,
Jersey, Malawi, Mexico, Netherlands, New Zealand, Nigeria, Norway,
Romania, Serbia, Slovakia, South Africa, Sweden, Thailand, Turkey, and
Ukraine.
4
|
NEWLANDS ET AL.
15206777, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/nau.25426 by Test, Wiley Online Library on [22/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.3.1 |Preliminary model identification
IRT analysis was conducted in R using the mirt
package,
31
fitting a confirmatory bifactor model to assess
the plausibility of assessing and scoring a general,
overarching factor as well as individual domains (sub-
factors).
19,21
The confirmatory model specified one
general factor onto which all 30 items were expected to
load (“rUTI QoL impact”), and five orthogonal subfactors
aligning with the factors identified during EFA
15
:
“personal wellbeing”(items A1–A4), “social wellbeing”
(items B1–B5), “work and activity interference”(items
C1–C7), “sexual wellbeing”(items D1–D4), and “patient
satisfaction”(items E1–E10). Recommendations for IRT
modelling with polytomous scales (i.e. scales with more
than two response options) by Toland et al. and by Reeve
and Fayers were followed, identifying a suitable
approach based on the RUTIIQ's ordered, polytomous
Likert‐type scale which successively increases from 0
(“strongly disagree”)to10(“strongly agree”).
19,30,32
A
bifactor GRM for ordered polytomous scales was thus
fitted, which is also specifically recommended for PROM
evaluation.
30,32,33
The Metropolis‐Hastings Robbins‐
Monro (MHRM) estimation method, the mathematical
algorithm recommended to estimate multidimensional
IRT models with more than three expected factors, was
employed to estimate item and model parameters.
19
2.3.2 |Model assumption checks
The intercept parameters, which govern the choice of the
next response category over the previous one (e.g.,
responding 10 vs. 9), were examined to assess partici-
pants' use of the 11‐point Likert scale.
19
It was expected
that the parameters would successively decrease as the
response categories (therefore the latent trait of QoL
impact) increased.
19
IRT analysis assumes that there is local independence
of items after controlling for a latent construct. In other
words, after controlling for the factor influencing a
respondent's choice of response, there should be no
statistically significant association or correlation between
items.
34–36
Item pairs that do not meet this assumption
are said to exhibit local item dependence, or LID. Yen's
Q
3
statistics were computed for each item pair within the
bifactor model, with values above 0.50 indicating
LID.
34–36
It was expected that the Yen's Q
3
statistics
should be less than 0.50 for all item pairs except those in
which both items measure the same subfactor or “testlet”
trait (e.g., A1–A2, B1–B2, etc.).
19,37
Finally, it is assumed that an item should be
interpreted in the same way across different subgroups,
known as item invariance.
38
Likelihood ratio χ
2
analysis
was conducted to check for item invariance, or the
absence of differential item functioning (DIF).
38
This
examined the extent to which each item performs
differently within the model based on biological sex
(female vs. male), age (older vs. younger than the
median, 42 years old), household income (> £25 000 vs.
< £25 000), level of education (university degree or above
vs. school or lower), and current antibiotic use (yes vs.
no).
38
Model parameters were freely estimated across
categorical groups and pvalues were adjusted using the
Bonferroni correction.
39
A statistically significant group
difference (χ
2
,p
adj
< 0.05) indicated the presence of
DIF.
38
2.3.3 |Model fit and performance
Standardized item factor loadings were expected to be
greater than 0.30 and communalities greater than 0.60,
indicating good fit and structural validity.
40
Mean‐square
(MNSQ) outfit statistics were examined to evaluate item
fit, with values between 0.50 and 2.00 indicating
acceptability for measurement.
41
Item slope (discrimina-
tion) parameters (α), which can be interpreted similarly
to factor loadings in classical factor analysis,
42
suggested
which items performed best within the model in terms of
differentiating between respondents' level of QoL
impact.
19,43
Minimum α= 0.65 was expected to suggest
at least “moderate”discrimination capability, with high-
er values indicating better performance.
43
Overall model fit was evaluated by computing the C
2
statistic of goodness of fit for ordinal data, with a non‐
statistically significant result suggesting good model fit.
44
This test is sensitive to sample size, thus making model
fit inferences based on the following indices is usual: root
mean square error of approximation (RMSEA
C2
;“good
fit”≤0.06), Comparative Fit Index (CFI; “good fit”≥0.95),
and standardized root mean square residual (SRMSR;
“good fit”≤0.06).
44,45
2.3.4 |Model refinement
The RUTIIQ was thus refined and finalized applying the
following strategy.
19,34–41,43–45
First, an item was proposed
for deletion if it: (i) demonstrated statistically significant
DIF (p
adj
< 0.05), (ii) showed poor item fit (MNSQ < 0.50 or
>2.00), (iii) indicated low discrimination capability
(α< 0.65), (iv) demonstrated poor factor loading (<0.30),
or (v) contributed insufficiently to the common model
variance (communality, h
2
< 0.60). While some level of LID
was expected due to related items within subfactors or
NEWLANDS ET AL.
|
5
15206777, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/nau.25426 by Test, Wiley Online Library on [22/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
testlets (e.g., items A1–A4),
19,37
in cases of LID (Q
3
> 0.50)
between item across different subfactors (e.g., items
A1–B1), one item from the pair was considered for deletion
based on which demonstrated stronger performance and fit.
Finally, the RMSEA, CFI, and SRMSR model fit indices
were assessed.
IRT analysis was re‐conducted iteratively after mak-
ing each proposed deletion until a confirmed model was
reached. The final, refined version of the RUTIIQ was
created according to this (see Table 2for included items;
the full RUTIIQ and scoring instructions are available
from the corresponding author).
2.3.5 |Reliability
The internal consistency of the final RUTIIQ was
evaluated by computing Cronbach's alpha (α) coefficients
for the general “rUTI QoL impact”factor and for
each subfactor, with α> 0.70 indicating acceptable
reliability.
46
2.3.6 |Readability
The minimum literacy level for comprehension of the
final RUTIIQ was estimated with the Automated
Reliability Index (ARI), a readability assessment tool
suitable for nonnarrative text such as PROMs.
47
3|RESULTS
3.1 |Participants
Most participants reported female biological sex (96.9%,
n= 377; Table 1), and were aged between 18 and 87 years
old (M= 45.4, SD = 17.1). Participants resided in 37
countries, mainly the United Kingdom (39.3%, n= 153)
and United States (37.8%, n= 147). Approximately three‐
quarters (74.0%, n= 288) reported a bachelor's degree or
higher, and approximately a third (37.8%, n= 147)
reported an annual household income above £50 000.
Participants reported an average of 3.62 UTIs in the past
6 months (SD = 2.90), and 7.06 in the past year
(SD = 5.91).
3.2 |Preliminary bifactor model
The preliminary bifactor model, composed of one general
factor and five orthogonal subfactors, converged success-
fully with MHRM estimation.
3.2.1 |Preliminary model assumption checks
All model assumption checks were passed: (i) successively
decreasing model intercept parameters confirmed consist-
ent use of the 11‐point Likert scale (see Table S2); (ii) no
TABLE 2 List of final 18 items included in the Recurrent
Urinary Tract Infection Impact Questionnaire (RUTIIQ).
Updated item
number Item
Because of my UTI(s)…
A1 I have experienced feelings of anxiety.
A2 I have experienced feelings of low mood or
depression.
A3 I have felt hopeless about the future.
B1 I have avoided socializing more than I used to.
B2 I have felt embarrassed in social situations.
B3 I have felt that I am no longer close to anyone.
B4 I have felt anxious in social situations.
C1 I regularly missed full or partial days of work,
home responsibilities or studying.
C2 The kind or amount of work I could do was
limited.
C3 It was more difficult than usual to concentrate
on my work.
C4 It was more difficult than usual to handle my
workload.
Thinking about my UTI‐related medical care…
D1 I have felt confident about being able to get the
medical care I need.
D2 I have felt like my medical concerns are taken
seriously.
D3 I have felt I could access UTI testing and
treatment quickly enough.
D4 I have had easy access to the medical
specialists I need.
E1*I have avoided sexual activity to minimize risk
of developing or worsening UTI symptoms.
E2*I have felt unable to enjoy sexual activity due
to my UTI(s).
E3*I have been concerned about the impact of my
UTI(s) on my sex life and/or sexual
relationship(s).
Note: The full RUTIIQ and scoring instructions are available from the
corresponding author.
*Questions about sexual wellbeing (items E1–E3) are optional and only to be
answered by respondents who report that their UTI(s) has/have impacted
their sex life in the past 2 weeks. The “Thinking about my UTI‐related
medical care”prompt only applies to items D1‐D4, and not to items E1‐E3.
6
|
NEWLANDS ET AL.
15206777, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/nau.25426 by Test, Wiley Online Library on [22/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LID was identified between items within different
subfactors, meeting the assumption of local independence
of items (Q
3
< 0.50), (iii) no DIF was identified based on
age, biological sex, household income, education level, or
current antibiotic use, meeting the assumption of item
invariance (χ
2
,p
adj
> 0.05; see Table S3).
3.2.2 |Preliminary model fit and
performance
The common variance was strongly represented, with
model and item fit suggesting necessary areas for refine-
ment. The general “rUTI QoL impact”factor and the five
subfactors collectively accounted for 79.5% of the common
variance, with the general factor accounting for 35.6% and
subfactors accounting for between 2.00% and 19.1% each
(Table S2). All items loaded onto the general factor and one
subfactor with standardized factor loadings above 0.30
(range = 0.32–0.84, Table S2), indicating strong fit and
structural validity, except item A4 assessing sleep disrup-
tion (general factor loading = 0.69, subfactor loading =
0.11). All items contributed sufficiently to the common
model variance, demonstrating communalities above 0.60
(Table S2), except item A4 (h
2
= 0.49) and item C3 assessing
self‐pressure to work despite illness (h
2
= 0.57).
All item MNSQ fit statistics were between 0.50 and 2.00
for good fit, except item E4 assessing confidence in treatment
decisions (MNSQ = 2.54; see Table S2). All except one item
(A4) indicated at least “moderate”discrimination capability,
demonstrated by slope parameters α> 0.64, with 25 items
(83.3%) indicating at least “high”discrimination (α> 1.35).
43
The CFI suggested good model fit (0.97). As the C
2
goodness of fit test (C
2
(375, N= 300) = 1124.26, p< 0.05),
RMSEA (0.082, 95% CI [.076, 0.087]) and SRMSR (0.064)
suggested inadequate model fit, results indicated it was
necessary to refine the model.
3.2.3 |Preliminary model refinement
The refinement strategy outlined in Section 2.3 was
applied, removing poor fitting items one at a time, and
re‐running the analysis after each proposed deletion to
assess the impact on model and item fit. Overall, 12 items
were removed: A4, B1, C2, C3, C6, D4, E1, E4, E6, E8, E9,
and E10 (see Table S4).
3.3 |Final bifactor model
A final 18‐item confirmatory bifactor model was identi-
fied (see Figure 2), comprising one general “rUTI QoL
impact”factor and five orthogonal subfactors. The final
18‐item RUTIIQ was created in line with this model with
updated item numbering (see Table 2for final included
items; the full RUTIIQ and scoring instructions are
available from the corresponding author). This briefer
measure demonstrates strong psychometric properties
while seeking to minimize respondent burden.
As detailed in the scoring instructions that accom-
pany the RUTIIQ (available from the corresponding
author), an overall RUTIIQ impact score and five
individual domain scores for specific QoL traits may be
computed. Individual domain scores are computed by
summing the item scores relevant to each domain,
48
with
a maximum possible score range of 0–30 for the
“personal wellbeing”and “sexual wellbeing”domains,
and 0–40 for the “social wellbeing,”“work and activity
interference,”and “patient satisfaction”domains. There
may be no individual score for the “sexual wellbeing”
domain since it is optional. The overall RUTIIQ impact
score, a transformed score with maximum range 0–100
may be computed to simplify interpretation and compar-
ison, especially in cases where respondents have opted
not to answer the optional “sexual wellbeing”items.
48
To
calculate a RUTIIQ impact score, administrators may
sum each individual domain score, divide this sum by the
number of completed items, and then multiply this by 10.
The observed scores in this sample (N= 389) demon-
strated heterogeneity in rUTI‐specific QoL impact,
ranging the full breadth of possible scores (Table 3).
3.3.1 |Final model assumption checks
All model assumption checks were passed: (i) evidence of
successively decreasing intercept parameters and no
disordered thresholds (Table 4); (ii) no LID was
identified between items within different subfactors,
meeting the assumption of local independence of items
(Q
3
< 0.50); (iii) the assumption of item invariance was
satisfied, evidenced by finding no DIF based on age,
biological sex, household income, education, or current
antibiotic use, meeting the assumption of item invariance
(χ
2
,p
adj
> 0.05; Table S5).
3.3.2 |Final model fit and performance
The common variance was strongly represented, and
both model and item fit were excellent. The general
‘rUTI QoL impact’factor and five subfactors collectively
accounted for 81.6% of the common variance (Table 4).
The general factor accounted for 41.2% of the common
variance, and the subfactors accounted for between 3.8%
NEWLANDS ET AL.
|
7
15206777, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/nau.25426 by Test, Wiley Online Library on [22/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
and 13.2% each. All standardized factor loadings and
communalities were greater than 0.30 and 0.60, respec-
tively, indicating excellent fit and structural validity
(Table 4). All item MNSQ fit statistics indicated
productivity for measurement, falling between 0.52 and
1.41 (M= 0.81, SD = 0.24; see Table 4). All items
indicated at least ‘moderate’discrimination capability
(α> 0.64), with 15 (91.7%) demonstrating at least ‘high’
capability and performance (α> 1.35; Table 4).
43
The model RMSEA (0.054, 95% CI [.042, 0.064]),
SRMSR (0.052) and CFI (0.99) demonstrated excellent
model fit.
44,45
While the C
2
goodness of fit test produced a
statistically significant result (C
2
(117, N= 300) = 217.87,
p< 0.05), model fit inferences were based on the RMSEA,
SRMSR and CFI due to the test's sensitivity to sample
size.
44,45
3.3.3 |Reliability
Internal consistency was excellent for the 18‐item general
rUTI QoL impact' factor (α= 0.92).
46
Reliability was
similarly strong for the five subfactors, with findings
ranging from α= 0.80–0.93 (Table S6).
FIGURE 2 This diagram demonstrates the bifactor structure represented by the 18 items included in the Recurrent Urinary Tract
Infection Impact Questionnaire (RUTIIQ). All items load onto the general factor colored in blue on the left: “rUTI quality of life impact.”
Each item also loads onto a subfactor colored in yellow on the right, each assessing a specific recurrent urinary tract infection (rUTI) quality
of life trait. Standardized factor loadings (>0.30), communalities (>0.60), and model fit indices (RMSEA = 0.054, CFI = 0.99,
SRMSR = 0.052) indicate excellent fit and structural validity (see Table 4).
8
|
NEWLANDS ET AL.
15206777, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/nau.25426 by Test, Wiley Online Library on [22/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3.3.4 |Readability
The ARI for the final RUTIIQ is 7.0, demonstrating that
this PROM is appropriate for people with a reading age of
12 years old or above (approximately US 7th grade, UK
Key Stage 3/year 8).
47
4|DISCUSSION
The RUTIIQ is a psychometrically strong measure of
patient‐reported QoL outcomes and rUTI healthcare
experience. A rigorous PROM development methodology
following best‐practice recommendations by COSMIN
maximized international patient and clinician input
throughout,
16,17
with exploratory analysis indicating
excellent internal consistency, test–retest reliability,
content validity, construct validity, and structural valid-
ity.
15
The high‐quality statistical approaches utilized to
refine the RUTIIQ demonstrated the strength of its factor
structure.
The final, optimized 18‐item RUTIIQ demonstrates a
well‐fitted bifactor structure that minimizes respondent
burden. A general factor evaluates “rUTI QoL impact,”
with five subfactors measuring “personal wellbeing,”
“social wellbeing,”“work and activity interference,”
“patient satisfaction,”and “sexual wellbeing.”Simple
scoring and administration instructions are provided
with the questionnaire, available from the corresponding
author. The RUTIIQ can be utilized within a number of
clinical and research contexts, including providing rapid
quantitative insights into key QoL domains impacted by
rUTI, assessing longitudinal change in patient QoL
outcomes in response to new and existing interventions,
exploring patient‐specific responses to antibiotic treat-
ment approaches, and identifying underlying domains
that may benefit from targeted medical and/or psycho-
social management and intervention. It is recommended
that in both clinical and research settings, the RUTIIQ is
administered alongside a validated rUTI‐specific patient‐
reported outcome measure of rUTI symptom presenta-
tion to capture the full breadth of the rUTI patient
perspective, such as the 15‐item Recurrent UTI Symptom
Scale (RUTISS).
23,24
The bifactor model and item fit statistics were
excellent,
44,45
highlighting the strength of the RUTIIQ
and its structural validity. The final 18 items each loaded
highly onto a specific subfactor in addition to the general
factor, demonstrating that they can also be assessed as
separate, individual domains.
44,45
Internal consistency
and reliability of the general factor and subfactors were
high, meeting gold‐standard recommendations.
17
Further research would address certain limitations. It
is recognized that most participants were Caucasian and
reported a high level of education and household income,
thus further research is necessary to establish cross‐
validation of the RUTIIQ and develop translations for
non‐English speaking populations or lower socio-
economic status respondents. It is acknowledged that
some RUTIIQ respondents may opt not to complete the
sexual wellbeing items; this is in accordance with the
APA Ethics Code and UK Government Social Research
published recommendations on asking personal ques-
tions,
49,50
and the remaining subscales maintain validity.
Although rUTI is significantly more prevalent in
females,
8
additional testing of the RUTIIQ would be
TABLE 3 Observed Recurrent Urinary Tract Infection Impact Questionnaire (RUTIIQ) scores.
Score Items*NMSD Observed range
Overall RUTIIQ impact score A1–E3 389 61.4 23.3 0–100
Individual domain scores
Personal wellbeing A1–A3 389 20.1 8.94 0–30
Social wellbeing B1–B4 389 17.8 13.0 0–40
Work and activity interference C1–C4 389 21.2 13.7 0–40
Patient satisfaction
a
D1–D4 389 28.3 10.6 0–40
Sexual wellbeing
b
E1–E3 300 26.0 5.63 0–30
Note: Higher scores indicate greater level of rUTI‐related impact to quality of life. All observed scores ranges align with the maximum possible ranges for each
score type.
Abbreviations: M, mean; SD, standard deviation.
a
Reverse scored.
b
The ”sexual wellbeing”domain is optional, therefore not all participants responded to these questions.
*Item numbers as per the refined, final version of the RUTIIQ (see Table 2; the full RUTIIQ and scoring instructions are available from the corresponding
author).
NEWLANDS ET AL.
|
9
15206777, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/nau.25426 by Test, Wiley Online Library on [22/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
TABLE 4 Bifactor graded response model item parameter estimates, fit statistics, and factor structure of the 18‐item Recurrent Urinary Tract Infection Impact Questionnaire (RUTIIQ).
Item*
Slope Intercept Standardized factor loading
h
2
Item MNSQ
outfitα
G
α
S1
α
S2
α
S3
α
S4
α
S5
c
1
c
2
c
3
c
4
c
5
c
6
c
7
c
8
c
9
c
10
GS1 S2 S3 S4 S5
A1 2.54 1.22 7.39 5.64 4.96 4.62 3.97 3.26 2.12 1.08 –0.32 –1.23 0.77 0.37 0.73 0.78
A2 5.51 3.84 13.0 10.5 8.70 7.56 6.34 4.87 3.69 1.67 –0.87 –3.05 0.80 0.56 0.94 0.56
A3 3.37 2.15 6.90 5.26 4.41 3.84 3.20 2.47 1.50 0.20 –1.04 –2.27 0.78 0.50 0.85 0.75
B1 4.01 1.87 5.80 4.48 3.52 2.93 2.50 1.57 0.68 –0.33 –1.94 –3.03 0.85 0.39 0.87 0.69
B2 2.34 2.44 3.52 1.88 1.20 0.57 0.24 –0.38 –1.02 –1.74 –2.75 –3.63 0.62 0.64 0.80 0.75
B3 1.79 1.61 1.44 0.50 0.01 –0.39 –0.89 –1.43 –1.91 –2.36 –3.09 –3.37 0.61 0.55 0.67 0.73
B4 4.45 3.32 5.70 3.85 2.56 1.95 1.23 0.36 –0.65 –2.00 –3.42 –4.54 0.77 0.57 0.92 0.60
C1 2.45 1.90 3.43 2.32 1.74 1.17 0.85 0.11 –0.50 –1.14 –1.79 –2.33 0.69 0.54 0.77 0.98
C2 3.69 3.15 5.94 4.64 3.72 2.93 2.18 1.36 0.72 –0.22 –1.54 –2.57 0.72 0.61 0.89 0.72
C3 3.72 2.95 6.92 5.85 4.64 3.74 2.76 1.98 1.29 0.50 –1.00 –2.05 0.74 0.59 0.89 0.71
C4 3.49 2.80 5.94 4.64 3.72 2.93 2.18 1.36 0.72 –0.22 –1.54 –2.57 0.73 0.58 0.87 1.03
D1 2.42 4.64 11.3 9.18 7.76 6.71 5.46 4.50 3.24 2.04 0.23 –2.14 0.44 0.84 0.90 0.72
D2 1.37 2.87 6.39 5.36 3.97 3.23 2.76 1.80 1.17 0.76 –0.59 –1.73 0.38 0.80 0.78 1.03
D3 1.30 2.03 5.14 4.36 3.67 3.15 2.64 2.07 1.47 0.82 0.05 –1.01 0.44 0.69 0.67 1.19
D4 1.26 2.39 6.17 5.49 4.24 3.27 3.04 2.12 1.59 0.86 –0.19 –1.21 0.40 0.75 0.72 0.88
E1 2.28 4.11 10.5 9.67 8.71 8.11 7.13 6.35 6.05 4.77 3.02 1.49 0.46 0.82 0.88 0.52
E2 2.55 4.27 9.84 9.34 8.53 7.65 6.91 5.75 5.03 4.06 2.66 0.85 0.49 0.81 0.90 0.55
E3 1.72 1.73 7.37 6.22 5.68 4.80 4.29 3.90 3.18 2.33 1.12 –0.13 0.58 0.58 0.67 1.41
ECV 0.41 0.04 0.07 0.08 0.13 0.09
Note:α= slope (or discrimination) parameters; higher slopes indicate greater discrimination. “moderate”discrimination capability: α= 0.65–1.34; “high”discrimination capability: α= 1.35–1.69; “very high”
discrimination capability: α≥1.70.
41
c
1
–c
10
= intercept parameters; these should successively decrease in value between c
1
and c
10
to demonstrate consistent use of the 11‐point scale.
19
Item MNSQ fit statistics
between 0.50 and 2.00 are interpreted as acceptable for measurement, with statistics closer to 1.0 indicating best fit to the model with the least distortion.
40
Abbreviations: ECV, explained common variance; G, general factor (rUTI quality of life impact); h
2
, communality; MNSQ, mean square (item fit statistics); S1, sub‐factor 1 (personal wellbeing); S2, sub‐factor 2 (social
wellbeing); S3, sub‐factor 3 (work and activity interference); S4, sub‐factor 4 (patient satisfaction); S5, sub‐factor 5 (sexual wellbeing).
*Item numbers as per the refined, final version of the RUTIIQ (see Table 2; the full RUTIIQ and scoring instructions are available from the corresponding author). Items D1–D4 (patient satisfaction) have been reverse
scored.
10
|
NEWLANDS ET AL.
15206777, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/nau.25426 by Test, Wiley Online Library on [22/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
beneficial to better understand its psychometric propert-
ies in males. Early indications suggest that the RUTIIQ is
accessible and appropriate for use by males living with
rUTI, however it should be noted the small proportions
of males recruited for this study mean that broader
extrapolation of the RUTIIQ to male‐specific rUTI
presentation is currently limited, reflective of a paucity
of male‐specific literature in UTI. While the sample size
was adequate, a larger sample may have further
improved the model parameter estimates, thus future
research could seek to validate the model within a larger
patient sample and in a clinical context.
30
The next stage
in PROM development, assessing the responsiveness and
clinical interpretability of the RUTIIQ, is ongoing.
16
5|CONCLUSION
The RUTIIQ is a psychometrically valid 18‐item ques-
tionnaire assessing patient‐reported personal wellbeing,
social wellbeing, work and activity interference, patient
satisfaction, and sexual wellbeing. Its simple scoring
facilitates standardized patient monitoring and quantifi-
cation of QoL impact. This brief patient‐reported
outcome measure offers a unique opportunity to
critically assess and prioritize the rUTI patient perspec-
tive, supplementing clinical management by improving
shared decision‐making and highlighting psychosocial
challenges requiring intervention.
AUTHOR CONTRIBUTIONS
All authors contributed to the study conceptualization
and methodological design. Abigail F. Newlands under-
took the study investigation, data collection, and project
administration. Abigail F. Newlands and Katherine A.
Finlay conducted formal data analysis and interpretation.
Melissa Kramer and Jessica L. Price contributed to the
study resources and participant recruitment. Abigail F.
Newlands prepared the original draft manuscript, and all
authors reviewed and approved the final manuscript.
ACKNOWLEDGMENTS
We would like to thank Live UTI Free for supporting this
study and recruitment.
CONFLICT OF INTEREST STATEMENT
Melissa Kramer is CEO of Live UTI Free Ltd.; however,
no financial incentives have been received.
DATA AVAILABILITY STATEMENT
The datasets used and analyzed during the current study
are available from the corresponding author on reason-
able request.
ETHICS STATEMENT
Ethical approval was granted by the School of Psychology
and Clinical Language Sciences Research Ethics Committee,
University of Reading (project reference no.: 2022‐115‐KF).
All participants were provided with study information sheets
and debrief forms, and electronically signed their consent to
take part before participation.
ORCID
Abigail F. Newlands http://orcid.org/0000-0002-
4718-0075
Melissa Kramer http://orcid.org/0000-0002-9242-5203
Lindsey Roberts http://orcid.org/0000-0001-5277-2377
Kayleigh Maxwell http://orcid.org/0000-0002-
8747-7201
Jessica L. Price http://orcid.org/0000-0002-0487-0826
Katherine A. Finlay http://orcid.org/0000-0002-
8997-2652
REFERENCES
1. Bonkat G, Bartoletti R, Bruyére F, et al. EAU Guidelines on
Urological Infections 2020. European Association of Urology
Guidelines 2020 Edition. European Association of Urology
Guidelines Office; 2020.
2. Zeng Z, Zhan J, Zhang K, Chen H, Cheng S. Global, regional,
and national burden of urinary tract infections from 1990 to
2019: an analysis of the global burden of disease study 2019.
World J Urol. 2022;40(3):755‐763. doi:10.1007/s00345-021-
03913-0
3. National Health Service. Urinary tract infections (UTIs). 2022.
Accessed January 4, 2023. https://www.nhs.uk/conditions/
urinary-tract-infections-utis/
4. Wagenlehner F, Wullt B, Ballarini S, Zingg D, Naber KG.
Social and economic burden of recurrent urinary tract
infections and quality of life: a patient web‐based study
(GESPRIT). Exp Rev Pharmacoeconom Outcome Res.
2018;18(1):107‐117. doi:10.1080/14737167.2017.1359543
5. Naber KG, Tirán‐Saucedo J, Wagenlehner FME. Psychosocial
burden of recurrent uncomplicated urinary tract infections.
GMS Infect Dis. 2022;10:10. doi:10.3205/id000078
6. Boeri L, Capogrosso P, Ventimiglia E, et al. Six out of ten
women with recurrent urinary tract infections complain of
distressful sexual dysfunction—a case‐control study. Sci Rep.
2017;7(1):44380. doi:10.1038/srep44380
7. Ciani O, Grassi D, Tarricone R. An economic perspective on
urinary tract infection: the “costs of resignation”.Clin Drug
Invest. 2013;33(4):255‐261. doi:10.1007/s40261-013-0069-x
8. Foxman B. Urinary tract infection syndromes. Infect Dis Clin
North Am. Mar 2014;28(1):1‐13. doi:10.1016/j.idc.2013.09.003
9. François M, Hanslik T, Dervaux B, et al. The economic burden
of urinary tract infections in women visiting general practices
in France: a cross‐sectional survey. BMC Health Serv Res.
2016;16(1):365. doi:10.1186/s12913-016-1620-2
10. Moskowitz D, Lee U. Patient distress in women with recurrent
urinary tract infections: how can physicians better meet
patients needs? Curr Urol Rep. 2018;19(12):97. doi:10.1007/
s11934-018-0849-1
NEWLANDS ET AL.
|
11
15206777, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/nau.25426 by Test, Wiley Online Library on [22/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
11. Grigoryan L, Mulgirigama A, Powell M, Schmiemann G. The
emotional impact of urinary tract infections in women: a
qualitative analysis. BMC Womens Health. 2022;22(1):182.
doi:10.1186/s12905-022-01757-3
12. Maxwell K, Roberts L, Kramer M, Price J, Newlands A,
Finlay KA. Psychosocial burden and healthcare disillusion-
ment in recurrent UTI: a large‐scale international survey of
patient perspectives. Front Urol. 2023;3:1264299. doi:10.3389/
fruro.2023.1264299
13. Sosland R, Stewart JN. Management of recurrent urinary tract
infections in women: how providers can improve the patient
experience. Urology. 2021;151:8‐12. doi:10.1016/j.urology.2020.
06.059
14. Meadows KA. Patient‐reported outcome measures: an over-
view. Br J Commun Nurs. 2011;16(3):146‐151. doi:10.12968/
bjcn.2011.16.3.146
15. Newlands AF, Roberts L, Maxwell K, Kramer M, Price JL,
Finlay KA. Development and psychometric validation of a
patient‐reported outcome measure of recurrent urinary tract
infection impact: the recurrent UTI impact questionnaire.
Qual Life Res. 2023;32:1745‐1758. doi:10.1007/s11136-023-
03348-7
16. Mokkink LB, De Vet HCW, Prinsen CAC, et al. COSMIN risk
of bias checklist for systematic reviews of patient‐reported
outcome measures. Qual Life Res. 2018;27(5):1171‐1179.
doi:10.1007/s11136-017-1765-4
17. Prinsen CAC, Mokkink LB, Bouter LM, et al. COSMIN
guideline for systematic reviews of patient‐reported outcome
measures. Qual Life Res. 2018;27(5):1147‐1157. doi:10.1007/
s11136-018-1798-3
18. Qin S, Nelson L, McLeod L, Eremenco S, Coons SJ. Assessing
test–retest reliability of patient‐reported outcome measures
using intraclass correlation coefficients: recommendations for
selecting and documenting the analytical formula. Qual Life
Res. 2019;28(4):1029‐1033. doi:10.1007/s11136-018-2076-0
19. Toland MD, Sulis I, Giambona F, Porcu M, Campbell JM.
Introduction to bifactor polytomous item response theory analysis.
JSchPsychol. 2017;60:41‐63. doi:10.1016/j.jsp.2016.11.001
20. Chen FF, Hayes A, Carver CS, Laurenceau JP, Zhang Z.
Modeling general and specific variance in multifaceted
constructs: a comparison of the bifactor model to other
approaches. J Pers. 2012;80(1):219‐251. doi:10.1111/j.1467-
6494.2011.00739.x
21. Dunn KJ, McCray G. The place of the bifactor model in
confirmatory factor analysis investigations into construct
dimensionality in language testing. Front Psychol.
2020;11:1357. doi:10.3389/fpsyg.2020.01357
22. Council of Europe. Common European Framework of Refer-
ence for Languages: Learning, Teaching, Assessment. Cam-
bridge University Press; 2001.
23. Newlands AF, Roberts L, Maxwell K, Kramer M, Price JL,
Finlay KA. The Recurrent Urinary Tract Infection Symptom
Scale: development and validation of a patient‐reported
outcome measure. BJUI Compass. 2023;4(3):285‐297. doi:10.
1002/bco2.222
24. Newlands AF, Kramer M, Roberts L, Maxwell K, Price JL,
Finlay KA. Confirmatory structural validation and refinement
of the recurrent urinary tract infection symptom scale. BJUI
Compass. 2024;5(2):240‐252. doi:10.1002/bco2.297
25. Brubaker L, Chai TC, Horsley H, Khasriya R, Moreland RB,
Wolfe AJ. Tarnished gold—the “standard”urine culture:
reassessing the characteristics of a criterion standard for
detecting urinary microbes. Front Urol. 2023;3:1206046.
doi:10.3389/fruro.2023.1206046
26. Pickard R, Chadwick T, Oluboyede Y, et al. Continuous low‐
dose antibiotic prophylaxis to prevent urinary tract infection
in adults who perform clean intermittent self‐catheterisation:
the AnTIC RCT. Health Technol Assess (Rockv). 2018;22(24):
1‐102. doi:10.3310/hta22240
27. Harding C, Mossop H, Homer T, et al. Alternative to
prophylactic antibiotics for the treatment of recurrent urinary
tract infections in women: multicentre, open label, rando-
mised, non‐inferiority trial. BMJ. 2022;376:e068229. doi:10.
1136/bmj-2021-0068229
28. Immekus JC, Snyder KE, Ralston PA. Multidimensional item
response theory for factor structure assessment in educational
psychology research. Front Educ. 2019;4:45. doi:10.3389/feduc.
2019.00045
29. Edelen MO, Reeve BB. Applying item response theory (IRT)
modeling to questionnaire development, evaluation, and
refinement. Qual Life Res. 2007;16(1):5‐18. doi:10.1007/
s11136-007-9198-0
30. Reeve BB, Fayers PM. Applying item response theory
modelling for evaluating questionnaire item and scale
properties. In: Fayers PM, Hays RD, eds. Assessing Quality of
Life in Clinical Trials: Methods and Practice. 2nd ed. Oxford
University Press; 2005.
31. Chalmers RP. mirt: a multidimensional item response theory
package for the R environment. J Stat Softw. 2012;48(6):1‐29.
doi:10.18637/jss.v048.i06
32. Samejima F. The general graded response model. In: Nering M,
Ostini R, eds. Handbook of Polytomous Item Response Theory
Models.Routledge;2010:77‐107.
33. Thissen D, Reeve BB, Bjorner JB, Chang C‐H. Methodological
issues for building item banks and computerized adaptive
scales. Qual Life Res. 2007;16(1):109‐119. doi:10.1007/s11136-
007-9169-5
34. Christensen KB, Makransky G, Horton M. Critical values for
Yen's Q3: identification of local dependence in the rasch
model using residual correlations. Appl Psychol Meas.
2016;41(3):178‐194. doi:10.1177/0146621616677520
35. Davidson M, Keating JL, Eyres S. A low back‐specific version
of the SF‐36 Physical Functioning scale. Spine. 2004;29(5):
586‐594. doi:10.1097/01.brs.0000103346.38557.73
36. ten Klooster PM, TAAL E, van de Laar MAFJ. Rasch analysis
of the Dutch health assessment questionnaire disability
index and the health assessment questionnaire II in patients
with rheumatoid arthritis. Arthritis Care Res. 2008;59(12):
1721‐1728. doi:10.1002/art.24065
37. DeMars CE. Application of the bi‐factor multidimensional
item response theory model to testlet‐based tests. J Educ Meas.
2006;43(2):145‐168. doi:10.1111/j.1745-3984.2006.00010.x
38. Schneider L, Strobl C, Zeileis A, Debelak R. An R toolbox for
score‐based measurement invariance tests in IRT models.
Behav Res Methods. 2022;54(5):2101‐2113. doi:10.3758/s13428-
021-01689-0
39. Bland JM, Altman DG. Statistics notes: multiple significance
tests: the Bonferroni method. BMJ. 1995;310(6973):170.
12
|
NEWLANDS ET AL.
15206777, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/nau.25426 by Test, Wiley Online Library on [22/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
40. Tabachnick BG, Fidell LS. Using Multivariate Statistics. 6th ed.
Pearson; 2012.
41. Linacre JM. What do infit and outfit, mean‐square and
standardized mean? Rasch Meas Trans. 2002;16(2):871‐882.
42. Hsieh C‐A, Von Eye AA, Maier KS. Using a multivariate
multilevel polytomous item response theory model to study
parallel processes of change: the dynamic association between
adolescents' social isolation and engagement with delinquent
peers in The National Youth Survey. Multivar Behav Res.
2010;45(3):508‐552. doi:10.1080/00273171.2010.483387
43. Baker FB, Kim S. The Basics of Item Response Theory Using R.
Springer International Publishing; 2017.
44. Cai L, Monroe SL A new statistic for evaluating item response
theory models for ordinal data. CRESST Report 839. National
Center for Research on Evaluation, Standards, and Student
Testing; 2014.
45. Maydeu‐Olivares A, Joe H. Assessing approximate fit in
categorical data analysis. Multivar Behav Res. 2014;49(4):
305‐328. doi:10.1080/00273171.2014.911075
46. Cronbach LJ. Coefficient alpha and the internal structure of tests.
Psychometrika. 1951;16(3):297‐334. doi:10.1007/BF02310555
47. Kincaid JP, Delionbach LJ. Validation of the automated
readability index: a follow‐up. Hum Fact. 1973;15(1):17‐20.
doi:10.1177/001872087301500103
48. Widaman KF, Revelle W. Thinking thrice about sum scores,
and then some more about measurement and analysis. Behav
Res Methods. 2023;55(2):788‐806. doi:10.3758/s13428-022-
01849-w
49. American Psychological Association. Ethical Principles of
Psychologists and Code of Conduct. http://www.apa.org/
ethics/code/index.html
50. Government Social Research. Ethics in Social Research: The
views of Research Participants. 2007: P6124.
SUPPORTING INFORMATION
Additional supporting information can be found online
in the Supporting Information section at the end of this
article.
How to cite this article: Newlands AF, Kramer
M, Roberts L, Maxwell K, Price JL, Finlay KA.
Evaluating the quality of life impact of recurrent
urinary tract infection: Validation and refinement
of the Recurrent UTI Impact Questionnaire
(RUTIIQ). Neurourol Urodyn. 2024.
doi:10.1002/nau.25426
NEWLANDS ET AL.
|
13
15206777, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/nau.25426 by Test, Wiley Online Library on [22/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License