Lifestyle Profile Scale for Gravida with Urinary Incontinence: scale development and psychometric testing in Chinese Mainland Authors: Minna Mao, Ping Xu, Ping-ping Guo, Xiaojuan Wang, Hongyan Wang, Suwen Feng Published: 2024 ABSTRACT Abstract AbstractObjectivesA significant number of women may experience urinary incontinence (UI) during pregnancy. While lifestyle modifications are recommended as important measures for preventing and treating UI, little is known about the lifestyle status of gravida with UI. This study aimed to develop the Lifestyle Profile Scale for Gravida with Urinary Incontinence (LPG-UI) and to evaluate its psychometric properties.DesignIt was an instrument design study carried out in two stages. In stage 1, a literature review was conducted to generate the initial item pool. Subsequently, a Delphi survey was conducted to obtain expert consensus and reduce the items. In stage 2, further item reduction was performed through item analysis. Then, psychometric properties of the instrument were evaluated, including structural validity, internal consistency, split-half reliability, test–retest reliability and criterion validity.SettingThis study was conducted in obstetric clinics of a tertiary maternity hospital in Zhejiang Province, mainland China.ParticipantsA sample of 447 pregnant women with UI were enrolled in stage 2.ResultsA 24-item, 7-dimension lifestyle profile scale was ultimately obtained. The content validity was satisfactory. The split-half reliability and total-scale Cronbach’s alpha coefficient were 0.890 and 0.819, respectively. The criterion validity was 0.598. The seven-factor model, which accounted for 63.667% of the total item variance, demonstrated a good fit with the data (χ2/df =1.672, root mean square error of approximation=0.053, goodness of fit=0.884).ConclusionsThe LPG-UI has sufficient validity and reliability for assessing the lifestyle status of pregnant women with UI.Keywords: Pregnant Women, Urinary incontinences, Health Education, Nursing Care, Health, Urogynaecology STRENGTHS AND LIMITATIONS OF THIS STUDY A novel instrument for assessing lifestyle status in gravida with urinary incontinence (UI) was developed and validated. A novel instrument for assessing lifestyle status in gravida with urinary incontinence (UI) was developed and validated. The Lifestyle Profile Scale for Gravida with UI is practical and reliable for clinicians to assess the lifestyle behaviours of pregnant women with UI and, accordingly, to provide individualised lifestyle advice. The Lifestyle Profile Scale for Gravida with UI is practical and reliable for clinicians to assess the lifestyle behaviours of pregnant women with UI and, accordingly, to provide individualised lifestyle advice. The participants in this study were recruited from one tertiary maternity hospital in one country. The participants in this study were recruited from one tertiary maternity hospital in one country. Introduction Urinary incontinence (UI), which places an enormous burden on the healthcare system and society,1 has become a worldwide public health problem.2 Pregnancy and parturition are considered important risk factors for the development of UI.1 The prevalence of UI at various stages of pregnancy ranges from 9% to 75%, and the incidence increases with increasing gestational age, averaging approximately 41%.3 As one of the most important clinical manifestations of pelvic floor dysfunctions, UI adversely affects the quality of life of pregnant women, primarily their physical, emotional and social relationships.4 5 And a significant percentage of women with persistent postpartum UI may have symptoms from their first pregnancy, and the percentage can reach 61.1%.6 According to clinical treatment guidelines, non-surgical treatment is the first-line approach for patients with UI, with lifestyle intervention being the recommended important step in the first-line treatment of UI.7,9 Effective lifestyle measures include maintaining a balanced diet, modifying fluid intake, reducing caffeine intake, engaging in regular physical activity, addressing constipation and obesity, and quitting smoking.7,9 Research shows that a healthy lifestyle can alleviate UI,10 and further intensive lifestyle intervention can reduce the prevalence of UI.11 Pregnant women have multiple opportunities for lifestyle-related health education and interventions due to their ongoing contact with healthcare professionals throughout the 9-month pregnancy period and their motivation to safeguard the health of the fetus.12 If lifestyle recommendations are provided by medical staff, the incidence of UI during pregnancy and after delivery is likely to decrease.13 Previous studies have developed lifestyle assessment tools to evaluate people’s health behaviours and provide a reference and effective intervention basis for lifestyle management.14 These tools predominately cover physical activity, diet, stress management, social relationships, and health responsibility domains.14 However, lifestyle assessment tools for UI are lacking. The existing lifestyle scales, such as the Health Promoting Lifestyle Profile (HPLP), are not specific to UI and are not suitable for assessing the lifestyle behaviours of pregnant women or people at high risk of UI. For example, the National Institute for Health and Clinical Excellence recommended adjusting fluid intake in women with UI9; however, this adjustment is not represented in the nutritional dimension of the HPLP.15 Furthermore, existing lifestyle scales do not incorporate assessment measures related to urination and defecation behaviours. Therefore, it is essential to develop a lifestyle scale specifically for pregnant women with UI. The use of a comprehensive model that incorporates physical, psychological and social aspects is one of the most effective ways to lifestyle modification and promote health.16 Pender’s health-promoting model (HPM) is one of the classical models that explain why people behave the way they do and how to change their behaviours.17 In this study, we developed the Lifestyle Profile Scale for Gravida with UI (LPG-UI) based on HPM and evaluated its psychometric properties. It would be helpful for medical workers to determine the lifestyle status of pregnant women with UI and to provide specific lifestyle health education. Methods This study had two stages: (1) scale development and (2) psychometric testing (online supplemental material figure 1). The Chinese version of the Health Promotion Lifestyle Profile II scale (HPLP-II-R) was used for the validity criterion. Scale development Item generation The initial items for the scale were generated by a literature review. We searched for articles published in 12 electronic databases—the PubMed, Web of Science, CINAHL, Embase, PsycINFO, Cochrane Library, Scopus, China National Knowledge Infrastructure, WanFang, VIP, Yiigle and CBM databases—from inception until 27 February 2022. The following Medical Subject Headings terms and keywords were used: ‘urinary incontinence’, ‘pelvic floor disorders’, ‘pelvic floor disease’, ‘pelvic floor dysfunction’, ‘life style’, ‘lifestyle’, ‘health education’ and ‘self-management’. The inclusion criteria consisted of articles that were (1) pertinent to the research topic (deals with the lifestyle content of women with UI) and (2) published in either English or Chinese. Two researchers screened titles and abstracts of search results to identify studies that potentially meet the inclusion criteria. Articles that were not available in full text, or studies published in letters, case reports, comments or focused on UI associated with other conditions (such as UI with male or military women, and UI that happened in patients who had a stroke or diabetes) were excluded. The literature search flow chart is shown in online supplemental material figure 2. Delphi survey and content validity examination A two-round Delphi survey was conducted to preliminarily evaluate and assess the content validity of the items in this study. An objective sampling method was used to select experts with experience and training in urogynaecology from across the country. An expert correspondence questionnaire was used. The experts rated the relevance and importance of the items on a scale ranging from 1 to 4 and 1 to 5 (1=very irrelevant, 4=highly relevant; 1=very unimportant, 5=highly important). Experts were also encouraged to provide feedback on wording, item allocation and scale construction and suggest any additions, modifications or deletions. The content validity index (CVI) was determined by calculating the relevance score. The number of experts with relevance scores ≥3 was divided by the total number of experts to determine the item-level content validity index (I-CVI), and the mean CVI of all items was considered the scale-level content validity index (S-CVI). An acceptable I-CVI was greater than 0.78, and an acceptable S-CVI was ≥0.9.18 We removed items with a rating of less than 0.78. Pilot test Pilot testing was carried out with a convenience sample drawn from a tertiary women’s hospital in August 2022. The inclusion criteria for pregnant women were as follows: (1) symptoms of urine leakage have occurred at least once in the last month, (2) were aged >20 years and (3) had basic reading and comprehension skills. The exclusion criteria were women with urine leakage before pregnancy; and women with pregnancy complications for which lifestyle modifications may be suggested, such as gestational diabetes mellitus, hypertension, cervical dysfunction and high risk of preterm delivery. The questionnaire was anonymous and contained only the pilot version of the LPG-UI. The researchers accompanied the participants during the questionnaire completion process, promptly addressed any questions raised by the participants and recorded the time taken to complete the questionnaire. The data from the pilot test were not included in the final psychometric analysis. Psychometric tests Sample and data collection According to sample size calculation principles, the sample size should be 5–10 times the number of items.19 We estimated the sample size based on the number of pilot LPG-UI items, ranging from 185 to 370. Considering a 10% loss to follow-up rate, a total of 204–407 participants were required for this study.20 The data were collected during two distinct periods. The first period was September 2022−November 2022. The data collected during this period were classified as dataset A and used for item analysis and exploratory factor analysis (EFA). Of the included participants, 34 were chosen at random for a retest 2 weeks later. The second period was January 2023−March 2023. The data collected during this period were classified as dataset B and used for performing confirmatory factor analysis (CFA) to verify the structure of the scale. The inclusion and exclusion criteria, sampling methods and sampling sites were consistent with those of the pilot study. Two researchers explained the study’s purpose, obtained informed consent when recruited participants, then distributed and collected the paper questionnaire. The researchers reviewed the collected questionnaires for missing responses, promptly requesting any necessary information from the participants. Those participants, who participated in the retest validity measurement 2 weeks later, were distributed and collected questionnaires by email or social software for the second time. Participation in the survey was voluntary and anonymous, and participants were able to withdraw at any time. Patient and public involvement None. Measures General questionnaire: a self-developed questionnaire was used to collect general demographic information, which included age, educational level, parity, trimester and so on. The HPLP-II-R: the HPLP-II-R was translated and revised by Cao based on the HPLP-II for evaluating health promotion lifestyle-related behaviours in mainland China.15 It has 40 items and six dimensions, including health responsibility (11 items), physical activity (eight items), nutrition (six items), spiritual growth (five items), interpersonal relations (five items) and stress management (five items) dimensions. For each dimension, Cronbach’s α coefficient ranged from 0.63 to 0.81, and the split-half (odd-even) reliability ranged from 0.64 to 0.78.15 The items were assessed using a 4-point Likert scale ranging from 1 (never) to 4 (always). The HPLP-II-R showed good reliability in this study (Cronbach’s α=0.93). LPG-UI (pilot version): the pilot version of the LPG-UI was obtained after the Delphi survey and pilot study and included 37 items, with item scores ranging from 1 (completely inconsistent) to 5 (completely consistent). Data analysis The collected data were entered into Excel 2019 by two individuals. IBM SPSS V.25.0 and AMOS V.26.0 were used for the data analysis. Missing values were inferred using means. Questionnaires with more than three missing values or clear-cut answer patterns were promptly removed during data entry.21 22 Descriptive statistics, such as frequency (%), mean and SD, were used to analyse participants’ demographic characteristics. Validity and reliability were assessed using specific methods, which are outlined below. Item analysis Item analysis encompassed analyses based on the following: (1) items with a critical ratio (CR) <3.0 and a p value >0.05 were considered for deletion23; (2) an item was considered for retention when its score demonstrated a significant correlation with the total scale score or when the item-total correlation value fell within the range of 0.30 to 0.8023; (3) an item was considered for deletion when Cronbach’s α showed substantial improvement when it was removed (with an increase of 0.5 or more in the overall scale’s α)23; (4) an item was considered for deletion with the coefficient of variation (CV) less than 0.15 in the discrete trend method; and (5) an item was considered for deletion if its factor loading was less than 0.4 or its factor contained less than 3 items or it loadings in more than one factor.24 We deleted items step by step while performing EFA. Items that satisfied the factor analysis method and met three or more of the five test method criteria were retained. Construct validity Construct validity was assessed using EFA and CFA. The sample’s suitability for factor analysis was examined using the Kaiser-Meyer-Olkin (KMO) test and Bartlett’s test of sphericity. Principal component analysis and the varimax orthogonal rotation method were used to conduct EFA on dataset A. The number of factors was not constrained. Factors with eigenvalues greater than 1 and items with factor loadings greater than 0.4 were retained. The item removal process was conducted iteratively during multiple runs of the EFA. The items were reviewed, categorised and compared with the initial attributes and elements identified through a literature review and interviews. Through thorough discussions and consensus among researchers, unqualified items were removed. We subsequently conducted CFA with dataset B to compare and assess the model fitness via the maximum likelihood method. Items with a factor loading <0.32 were considered for deletion.25 The suitability of the model was evaluated by multiple indicators: the χ2 test and df ratio (χ2/df <3), root mean square error of approximation (RMSEA<0.08), goodness-of-fit index (GFI>0.90), comparative fit index (CFI>0.90), Tucker-Lewis index (>0.90), normed fit index (NFI>0.90), incremental fit index (IFI>0.90) and relative fit index (>0.90).26 Criterion validity Pearson’s correlation coefficient was calculated using the total LPG-UI and HPLP-II-R scores. A positive correlation between the LPG-UI and HPLP-II-R scores was hypothesised. Internal consistency reliability The Cronbach’s α coefficient was computed to assess the extent to which all scale items consistently measured the underlying construct. In general, a Cronbach’s alpha coefficient greater than 0.7 is acceptable.27 Test−retest reliability Test−retest reliability was assessed by calculating the intraclass correlation coefficient (ICC). The time interval between the two rounds was approximately 2 weeks. The ICCs were categorised as follows: poor (below 0.40), fair (0.4–0.59), good (0.6–0.74) and excellent (0.75–1.00).28 Split-half reliability Split-half reliability was obtained by dividing items into two parts according to odd and even numbers and then calculating the Cronbach’s α coefficient. Results Scale development Item generation Through a comparative analysis of the findings obtained from the literature review, the initial 77 items and 10 dimensions of the specific lifestyle factors that may affect the development of UI during pregnancy were drafted (online supplemental material table 1). The generation of dimensions was carried out under the guidance of HPM. According to HPM, lifestyle can be influenced by personal characteristics and experiences, behavioral-specific perceptions and affects.17 The specificities of the pregnant population and the special behaviour habits of UI patients were taken into account. In combination with pregnancy and UI-related guidelines, we set dimensions that included fluid management, defecation management, urination management, hygiene management and daily life management. At the same time, considering that UI patients are currently not optimistic about their medical-seeking behaviour, and the reality that urine leakage can cause shame and increase social pressure, the contents of health responsibility and stress management were added, and the contents on interpersonal communication for pregnant women were refined. The initial items of the LPG-UI were measured on a 5-point Likert scale (1=completely inconsistent, 2=basically inconsistent, 3=partially consistent, 4=basically consistent, 5=completely consistent). Delphi survey and content validity examination In total, 15 experts participated from May 2022 to August 2022. Consensus among participants was achieved after two rounds of the survey. (The demographic characteristics of the Delphi participants were shown in online supplemental material table 2. In the initial round, we meticulously deliberated and refined the questionnaire’s dimensions and items, ensuring alignment with the experts’ suggestions. Necessary modifications regarding the expression of the items were made. A total of 41 items were eliminated, and eight new items were incorporated. In the second round, an additional seven items were removed, leading to the scale comprising 37 items and six dimensions. We adjusted the dimensions as follows after the Delphi method and discussion of the research group: (1) ‘food intake management’ and ‘fluid intake management’ were merged into ‘diet management’; (2) ‘urination management’ and ‘defecation management’ were merged into ‘excretion management’; (3) retained items in the ‘daily life management’ dimension were classified into the ‘health responsibility’ dimension; and (4) ‘hygiene management’ was removed from the scale due to the deletion of four items with a high CV or low CVI, and the remained one item was transferred to the ‘Health Responsibility’ dimension based on experts’ advice and group discussions. The I-CVI for all the retained items was consistently 0.80 or higher, while the S-CVI reached an acceptable level of 0.90. Pilot study findings: participant feedback Of the 34 questionnaires distributed, 30 valid questionnaires were collected. Participants indicated that the items were easy to understand, and the questionnaire completion time ranged from 4 to 8 min, with an average of 6 min. Psychometric tests Characteristics of the participants In total, 485 pregnant women were surveyed, and 447 questionnaires were used in the statistical analysis. In the first period, 225 pregnant women were surveyed, resulting in the collection of 205 valid questionnaires (91.11%). These data served as dataset A. In the second period, 260 participants were surveyed, and 242 (93.08%) participants completed the questionnaires. These data were classified as dataset B. The participants’ ages ranged from 21 to 46 years, with an average age of 30.77 years. Most participants had a college education or higher (n=417, 87.42%). The gestational age ranged from 8 to 41 weeks. The mean gestation duration was 31.95 weeks, and most participants were primiparas (n=313, 70.02%) (table 1). Table 1Participants’ demographic information (n=447)CharacteristicsGroupDataset A (n=205)Dataset B (n=242)Age (years)Mean (SD)30.53 (3.70)31.01 (4.18)Range21–4222–46Highest level of educationN (%)N (%)Junior high school and below3 (1.5)4 (1.6)High school13 (6.3)10 (4.1)College/university159 (77.6)178 (73.6)Postgraduate30 (14.6)50 (20.7)Gestational age (weeks)Mean (SD)32.28 (6.26)31.62 (6.35)Range8–4111–40ParityN (%)N (%)Primipara147 (71.7)166 (68.6)Nulliparous58 (28.3)76 (31.4)Single pregnancyN (%)N (%)Yes192 (93.7)236 (97.5)No13 (6.3)6 (2.5) Participants’ demographic information (n=447) Construct validity Item analysis Item analysis and EFA were conducted concurrently, with the item deletion process being executed incrementally through four iterations of EFA and the application of the other four item analysis methods. Based on the statistical results, all the retained items met the criteria. In the four rounds of factor analysis, items 14, 24 and 29 had factor loadings below 0.4, items 4, 5, 21, 22 and 30 were located in a factor containing less than 3 items, respectively, and item 36 was loaded in more than one factor; these items were deleted. Four items (3, 13, 32 and 35) that did not meet three or more of the five testing criteria were also excluded. Specifically, apart from item 19, which had a CR of 2.96, the ratio for the remaining items fell within the range of 3.300–9.996. Furthermore, except for items 19 and 20, which exhibited coefficient values of 0.217 and 0.232, respectively, the item-total correlation values for the remaining items ranged from 0.322 to 0.599. The factor loading of all the items ranged from 0.536 to 0.799. No individual item was observed to have a substantial impact on Cronbach’s α when removed (table 2). Table 2Item statistics results of LPG-UI (37 items, n=205)ItemCritical rationCorrelation coefficientDiscrete trendCVFactor analysisCronbach’s alpha if item deletedResult14.046*0.346*0.4270.7550.822Retained24.994*0.388*0.4910.7220.821Retained32.875*0.286*0.5560.6200.823Deleted42.753*0.243*0.3420.7550.825Deleted54.440*0.368*0.2330.4630.821Deleted65.359*0.409*0.2280.7610.819Retained76.299*0.463*0.2170.7020.818Retained84.921*0.393*0.2280.6330.820Retained98.684*0.540*0.1810.7560.816Retained109.996*0.573*0.1790.7970.816Retained118.544*0.599*0.2100.7300.814Retained127.333*0.533*0.2320.7990.816Retained134.260*0.380*0.1430.6620.821Deleted142.986*0.206*0.285<0.40.827Deleted157.449*0.497*0.1930.7190.817Retained166.088*0.382*0.1700.7150.820Retained175.776*0.384*0.2130.7010.820Retained186.053*0.420*0.3110.5360.819Retained192.926*0.217*0.2790.7770.826Retained203.383*0.232*0.3380.7980.827Retained211.8780.145*0.1210.6680.825Deleted223.746*0.347*0.2370.4250.821Deleted233.300*0.322*0.2500.6220.822Retained245.383*0.433*0.190<0.40.819Deleted254.113*0.402*0.4570.8090.820Retained266.223*0.496*0.4280.8770.817Retained276.282*0.493*0.4020.8150.817Retained286.808*0.508*0.3960.6720.816Retained295.459*0.445*0.236<0.40.818Deleted303.630*0.247*0.3180.7880.825Deleted313.976*0.387*0.2770.7220.820Retained320.9500.0900.4520.7480.831Deleted334.955*0.423*0.2620.5950.819Retained345.505*0.387*0.2520.7790.820Retained351.9540.1300.2330.6550.828Deleted366.341*0.418*0.361<0.40.820Deleted377.306*0.486*0.2490.5440.817Retained*P < 0.01.CV, Coefficient of VariationLPG-UILifestyle Profile Scale for Gravida with UI Item statistics results of LPG-UI (37 items, n=205) P < 0.01. CV, Coefficient of VariationLPG-UILifestyle Profile Scale for Gravida with UI Exploratory factor analysis The suitability of dataset A for EFA was confirmed by KMO and Bartlett’s tests. Bartlett’s test was significant (χ2=2479.459, p<0.001) and KMO value was 0.742, indicating that the samples were eligible for factor analysis. The first EFA yielded 11 common factors with eigenvalues >1, which explained 62.97% of the overall variance. We scrutinised the items in the 11 factors, excluding one that cross-loaded on two factors (item 36 showed cross-loading on both factors 4 and 11), two items exhibited low factor loadings (items 14 and 29) and three self-factorial items (items 4, 5 and 21). Combined with the results of the item analysis, four items (items 3, 13, 32 and 35) that failed to meet at least three of the screening indicators except factor analysis were excluded. After running the EFA several times, item 24 was removed (loading value <0.4), and items 22 and 30 were tied together into an independent factor that only comprised these two items. After discussion by the researchers, these two items were deleted. After these modifications, the EFA with 24 items unveiled seven factors, each with eigenvalues surpassing 1.0, explaining 63.667% of the total variance. We pinpointed the defining traits of the seven factors with the highest factor loading values and assigned a factor name that encompassed all items within each factor. The first factor was labelled as ‘health responsibility’ because it refers to attitude and seeking help from medical workers to prevent and treat UI. The second factor was titled ‘food intake management’ and concerns fruit, vegetable, protein and crude fibre intake. Factor 3 was named ‘stress management’, indicating gravidas’ ways of coping with stress and attitudes when they face UI. Factor 4, ‘urination management’, assesses gravidas’ urination habits and urine colour. Factor 5, ‘fluid intake management’, assesses gravidas’ water intake and habits. Factor 6 was named ‘defecation management’ because it reflects the bowel evacuation habits and can assess the presence of constipation. Factor 7 was given the name ‘activity management’ because it reflects the activity status of gravidas, which may affect pelvic floor muscle elasticity and UI (table 3). The full version of the LPG-UI can be found in the online supplemental material table 3. Table 3Factorial load matrix of exploratory factor analysis (n=205)CodeItemsFactor loadingF1F2F3F4F5F6F7 26Obtain health information related to urinary incontinence actively0.877 25Consult or report to medical staff in time if urine leakage occurs0.821 27Perform measures to avoid urine leakage actively0.804 28Treat diseases related to urinary incontinence actively0.689 12Eat coarse fibre in moderation 0.814 10Eat fresh vegetables every day 0.791 9Eat fresh fruit every day 0.745 11Eat protein-rich foods every day 0.731 34Trust yourself to deal with your urine leakage 0.768 31Discover the bad mood caused by urine leakage, and adjust timely 0.732 33Take measures to reduce the impact of urine leakage on work and life 0.612 37My family and friends can provide support if I need help 0.599 16I won't hold urine on purpose 0.786 17My urine is pale yellow in colour 0.704 15Perform preventive urination before sleeping and outing 0.703 6My daily water intake between 1500 and 2000 mL 0.802 7Drinking small amounts of water throughout the day rather than large amounts in short bursts 0.759 8Drink water during the day and reduce it at night 0.573 20Defecation is time-consuming and laborious 0.832 19Defecate less than three times a week 0.797 18Have a habit of regular defecate 0.556 1Develop a proper exercise plan and follow it 0.790 2Participate in moderate-intensity physical activity three or more times per week and accumulate more than 150 min 0.747 23Active weight control to avoid excessive growth 0.610Eigenvalue5.2801.8561.6471.4931.3901.063Variance explained (%)21.9997.7356.8386.2215.7934.428Cumulative percentage %21.99940.38747.22553.44659.23963.667F1, Hhealth responsibility; F2, Ffood intake management; F3, Sstress management; F4, Uurination management; F5, Ffluid intake management; F6, Ddefecation management; F7, Aactivity management. Factorial load matrix of exploratory factor analysis (n=205) F1, Hhealth responsibility; F2, Ffood intake management; F3, Sstress management; F4, Uurination management; F5, Ffluid intake management; F6, Ddefecation management; F7, Aactivity management. Confirmatory factor analysis Based on the findings of the fourth EFA, the 24-item, 7-factor scale was subjected to confirmatory component analysis by AMOS software using Dataset B. The model was found to fit since the χ2/df value was equal to 1.672 (χ2=386.34, df=231, p<0.001). All the items except items 8 and 23 presented factor loadings exceeding 0.32. The seven-factor model fell within the acceptable range and provided a good fit to the data (RMSEA=0.053, RMR=0.064, GFI=0.884, CFI=0.902, IFI=0.905 and NFI=0.793). The detailed fit indices are provided in online supplemental material table 4. Criterion-related validity The total LPG-UI score exhibited moderate correlations with the total HPLP-II R score and its six dimensions’ scores (r=0.335 to 0.598, p<0.01) (table 4). Table 4Correlations of the LPG-UI and HPLP-II R subscalesHPLP-II RDimension 1Dimension 2Dimension 3Dimension 4Dimension 5Dimension 6LPG-UI0.598*0.393*0.453*0.400*0.408*0.435*0.335**P < 0.01.HPLP-II RThe Chinese version of the Health Promoting Lifestyle Profile II ScaleLPG-UILifestyle Profile Scale for Gravida with Urinary Incontinence Correlations of the LPG-UI and HPLP-II R subscales P < 0.01. HPLP-II RThe Chinese version of the Health Promoting Lifestyle Profile II ScaleLPG-UILifestyle Profile Scale for Gravida with Urinary Incontinence Reliability testing Internal consistency reliability The Cronbach’s α for the total LPG-UI score was 0.819, and for each subfactor, it ranged from 0.612 to 0.848 (table 5), affirming the consistent measurement of the constructed concept across all factors. Table 5Reliability analysis of the LPG-UISubscale/total scaleNumbers of itemsInternal consistencyTest–retest reliabilityCronbach’s α coefficientICCF140.8420.877*F240.8480.862*F340.6770.646†F430.6751.0*F530.6790.578†F630.6120.771*F730.6220.925*Total scale240.8190.849**P <0.01.†P <0.05.ICC, Intraclass Correlation CoefficientLPG-UILifestyle Profile Scale for Gravida with Urinary Incontinence Reliability analysis of the LPG-UI P <0.01. P <0.05. ICC, Intraclass Correlation CoefficientLPG-UILifestyle Profile Scale for Gravida with Urinary Incontinence Split-half reliability As some dimensions consisted of only three items, we opted not to compute the split-half reliability for each dimension and instead focused on calculating the split-half reliability for the overall scale. The split-half reliability of the overall LPG-UI was 0.890. Test–retest reliability The ICC for the overall score remained at 0.849 (p<0.01), confirming the LPG-UI’s stability. The ICC values for factors 1 through 7 ranged from 0.578 to 1.0 (table 5), demonstrating strong test–retest reliability. Discussion The LPG-UI is a novel lifestyle assessment tool designed for pregnant women experiencing UI. The results indicated that the LPG-UI exhibited favourable internal consistency (Cronbach’s α: 0.819), good stability (ICC: 0.849) and sufficient or acceptable validity (including content validity, structural validity and criterion-related validity). The seven-factor model explained 63.667% of the total variance, and the parameters of the model-fitting degree reached ideal values in CFA. In summary, the LPG-UI is a valid and reliable assessment tool for evaluating the lifestyle status of pregnant women with UI. In the field of UI, studies of health-related quality of life (HRQOL) are prevalent. HRQOL is widely used to understand how diseases influence the lives of individuals, but it differs from the concept of a health-promoting lifestyle.29 A health-promoting lifestyle is defined as any activity or behaviour aimed at achieving higher levels of health, well-being and happiness.30 Therefore, scales for these two concepts measure different aspects. Specifically, we focused more on the behavioural health-promoting lifestyle status rather than the degree of the negative impact of UI on quality of life. While existing lifestyle scales could be used to assess pregnant women with UI, they do not capture the specific lifestyles of this group, such as changes in water intake, physical activity, urination frequency and stigma. The LPG-UI is more targeted and can be used to comprehensively assess specific lifestyles through seven dimensions. In terms of dimensions, we generated 10 dimensions guided by HPM and literature review. Factor analysis revealed that the LPG-UI was a multidimensional scale consisting of seven factors, which contradicted the initial item generation and the findings of the Delphi survey. The existing contradictions revolved around whether to merge ‘food intake management’ and ‘fluid intake management’ into ‘diet management’ and ‘urination management’ and ‘defecation management’ into ‘excretion management’. The fluid intake habits of UI patients are unique. Approximately 17–40% of women who experience UI report limiting their fluid intake as a means of managing their urinary symptoms.31 Items regarding liquids were more correlated with each other than items about other foods. Thus, it was plausible that ‘fluid intake management’ could be its own dimension. As for ‘urination management, healthy urination behaviour was essential for preventing and alleviating lower urinary tract symptoms.32 However, urination management was not strongly correlated with defecation behaviour, and factor analysis provided further evidence of this. We therefore maintained separate dimensions for urination and defecation management. Wang developed the Web-based Toileting Behavior (TB-WEB) scale, which was specifically designed to assess urination behaviour.33 Compared with the TB-WEB, the LPG-UI assesses not only urination behaviour but also other behaviours that may affect UI, which is more comprehensive. We did not adhere to the original dimensions exactly, and the final dimensions were conceptually equivalent to one of the original 10 dimensions or a combination of two dimensions. Similar developments have occurred with other reported scales.34 The results of the reliability tests showed that the LPG-UI is a reliable instrument. The internal consistency and split-half reliability of the LPG-UI was >0.8, indicating that the LPG-UI had good homogeneity and internal consistency. For each dimension, three dimensions comprising three items had coefficients greater than 0.6 but less than 0.7. Cronbach’s α is sensitive to the number of items in a scale, so it is common to observe lower α values in factors with a small number of items.35 Most studies take a reliability greater than 0.7 as a qualified indicator, but a coefficient greater than 0.6 can represent a moderate degree of internal consistency.36 The ICC of the LPG-UI (0.646–1.0) was assessed and revealed good scale stability. The HPLP-II R was used to evaluate the criterion-related validity of the LPG-UI. Both scales are based on HPM theory and focus on the assessment of lifestyle; thus, the HPLP-II seemed to be an appropriate scale for assessing the criterion validity of the LPG-UI. The correlation of the LPG-UI with the HPLP-II was significant (Pearson Correlation: 0.598), which means that the LPG-UI has good criterion-related validity. This scale holds significance for effective personal lifestyle management. First, this scale contains most of the content of the individual’s lifestyle behaviour, which emphasises trying to find their hidden, unrealised unhealthy behaviours. UI may have a negative effect on the emotional status of pregnant women and restrict their social functions,37 and the status quo of their medical-seeking behaviour is discouraging.38 The LPG-UI also focuses on the individual’s stress management and health responsibilities which are required special attention. For the reasons described above, the researchers believe that the LPG-UI can be used to assess a gravida’s current status of healthy lifestyle, both overall and for each of the seven factors, to identify where improvement is needed, and to develop what to teach or train based on the resulting profile. Furthermore, it can be used to evaluate the effectiveness of intervention programmes run by various providers for pregnant women with UI. This study has several limitations. First, since this study was conducted at a single centre and the sample size was not large, the representation of the sample may be limited. A follow-up study should further increase the sample size and source to increase the external validity. Second, although a targeted literature review and group discussion were carried out in the construction stage of the item pool, we did not conduct semistructured interviews with patients, which may have made the expression of the items insufficiently grounded. Conclusion The LPG-UI is a five-point scale consisting of 24 items and 7 dimensions (health responsibility, food intake management, stress management, urination management, fluid intake management, defecation management and activity management). The total LPG-UI score ranges from 24 to 120, and a higher score signifies a healthier lifestyle. We anticipate that the LPG-UI will serve as an assessment tool for evaluating the lifestyles of pregnant women with UI, gauging the effectiveness of UI-related health education and providing a foundation for the development of educational programmes. supplementary material Footnotes Funding: This study was funded by the Scientific Research Foundation of the National Health Commission (WKJ-ZJ-2211) and Medical Science and Technology Project of Zhejiang Province (grant number: 2022KY188 and 2024KY1146). Prepublication history and additional supplemental material for this paper are available online. To view these files, please visit the journal online (https://doi.org/10.1136/bmjopen-2024-085493). Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Not applicable. Ethics approval: This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of Women’s Hospital School of Medicine Zhejiang University in Zhejiang Province, China (code of ethics approved: no. IRB-20210342-R). Participants gave informed consent to participate in the study before taking part. Data availability free text: Not applicable. Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research. Contributor Information Minna Mao, Email: maomn@zju.edu.cn. Ping Xu, Email: 13147655763@163.com. Ping-ping Guo, Email: 694532381@qq.com. Xiaojuan Wang, Email: 11818206@zju.edu.cn. Hongyan Wang, Email: whysammy@zju.edu.cn. Suwen Feng, Email: fengsw@zju.edu.cn. Data availability statement No data are available. References Supplementary Materials Data Availability Statement No data are available.