Urinary incontinence (UI) is a frequent problem in midlife and older women.1 Approximately 45% of midlife women between the ages of 45 and 55 years report UI occurring at least a few times per month, and approximately 15% report daily UI.2 Urinary incontinence can affect women's lives negatively, leading to decreased quality of life3–6 and significant morbidity such as functional decline7 and increased risk of falls and fractures in the elderly.8
Many effective treatments are available to women for any level of UI, including behavioral (eg, limiting fluid intake, changing voiding habits), pelvic floor muscle exercises, pessaries, medications, and surgery.9 In cross-sectional studies, less than half of women with UI tell health care providers about their problem.10–13 These studies have identified factors associated with UI treatment-seeking such as male gender,11 greater social support,14 and higher use of other health care services.14
Gaps in knowledge about disparities in UI treatment-seeking still exist. It is not known if changes in UI characteristics and physical, social, and psychological factors over time may affect whether women seek UI treatment. Even less clear is whether UI duration, changes in UI frequency, and other factors associated with seeking treatment might differ between racial or ethnic, socioeconomic, educational backgrounds, or a combination of these. The present analyses explored potential disparities in UI treatment-seeking in a racial or ethnically diverse cohort of community-dwelling midlife women followed longitudinally over 9 years.
MATERIALS AND METHODS
This is a secondary analysis of data from the Study of Women's Health Across the Nation from 1996 to 2007. The Study of Women's Health Across the Nation is a multicenter, multiracial or ethnic prospective cohort study investigating longitudinally the biological and psychosocial characteristics and common symptoms of midlife women across the menopausal transition. Briefly, from a sample of women recruited by random digit dialing and list-based sampling, each site of seven sites recruited approximately 450 women for the Study of Women's Health Across the Nation, consisting of approximately 50% white (non-Hispanic) women and 50% women from one designated minority racial or ethnic group. In this study, we included all women from six of the clinical sites (Pittsburgh, Oakland, Los Angeles, Detroit, Chicago, Boston) representing white (non-Hispanic), black, Chinese, and Japanese women. We could not include the white (non-Hispanic) and Hispanic women from the New Jersey site because questionnaires were not administered there during the years of this investigation. Study of Women's Health Across the Nation enrollment inclusion criteria included: age 42–52 years; self-identification as black; Japanese, Chinese, Hispanic, or white; and premenopausal or early perimenopausal status. Enrollment exclusion criteria included: inability to speak English, Japanese, or Cantonese; hysterectomy, bilateral oophorectomy, or both; on any hormonal medications; or pregnant or lactating. The study protocol was approved by all institutional review boards, and participants provided written, informed consent.
Based on data from Study of Women's Health Across the Nation questionnaires that were self-administered (6th-grade reading level) at an approximately annual in-person visit, we created variables to describe UI characteristics as follows. We defined responses to the following question: “In the last month, about how many times have you leaked urine, even a small amount?” as daily for “almost daily or daily,” weekly for “several days per week,” monthly for “less than 1 day per week,” and no clinically significant UI as “less than once per month” or none. We defined UI type by responses to the question “under what circumstances does leakage occur?” We defined women who reported leakage with “coughing, laughing, sneezing, jumping up and down, with physical activity” as having stress UI, whereas women who reported leakage with “when you have the urge to void and can't reach the toilet fast enough” as having urge UI. We defined mixed UI when women responded positively to both circumstances of leakage. Finally, women who reported UI at any time point and then reported an increase in frequency from one annual visit to the next were considered to have worsening UI (ie, from no regular UI to monthly or more, from monthly to weekly or more, or from weekly to daily). We considered women who reported a decrease in frequency of UI from one annual visit to the next as having improving UI. In the same in-person visit self-administered questionnaire, participants reported on treatment-seeking behaviors for UI by responding yes or no to the following question that the Study of Women's Health Across the Nation included at baseline and in visits 7, 8, and 9: “Have you ever discussed your leakage with a doctor, nurse, or other health care professional?” after they had responded “yes” to a stem question: “Have you ever leaked urine, even a small amount?”
Covariates included self-reported race or ethnicity, income, and education level; level of difficulty in paying for basics; depressive15 and anxiety symptoms16; social support or network17; and experience of discrimination18; medical history; and health care utilization. We calculated body mass index (weight (kg)/[height (m)]2) based on measurements obtained by a calibrated balance beam scale and by a stadiometer. The Study of Women's Health Across the Nation defined menopause status using Stages of Reproductive Aging Workshop criteria.19 We created summary variables for the patterns of UI and these other variables like in prior publications.2,20,21
Women in our main analytic sample were those who remained in the study, reported UI at any visit, and responded to the treatment-seeking questions in visits 7, 8, or 9 or were presumed to have not sought treatment (Fig. 1). We examined bivariate associations with seeking treatment using χ2 tests for categorical variables and t-tests for continuous variables. We then used multiple logistic regression analysis, exploring forward, backward, and stepwise methods to examine the associations between demographic, psychosocial, physical, health care, and UI characteristic variables and UI treatment-seeking across the first 9 years of cohort observation. We chose our independent variables a priori based on the literature2,6,14 and from our bivariate analysis when P was ≤.3. We selected our models based on best fit, defined as the lowest Akaike Information Criterion.22
We examined interactions in two ways. First, we developed stratified models using our primary independent variables: racial or ethnic category, socioeconomic status, and education. Second, we explored interaction between UI frequency and UI duration and each of these variables within our main model by entering these combinations as interaction terms.
Because women were questioned about UI treatment-seeking behaviors at baseline and then not until visit 7, we could not determine when during that 8-year interval women actually sought treatment and thus which factors preceded treatment-seeking behavior. To address this, we secondarily examined multivariable models in a subsample of women who had not reported seeking treatment before visit 7, following them through visit 9.
A subset of Study of Women's Health Across the Nation–incontinent women had missing data because, although they reported UI one or more times before visit 7, they answered the stem question “Have you ever leaked urine, even a small amount” at visit 7, 8, 9, or all of these negatively (n=364) and thus did not respond to our treatment-seeking questions. To evaluate the potential bias this nonresponse could have on our results, we compared two sampling strategies. First, we counted the 364 women as missing and did not include them in what we defined as the limited analytic sample. Second, we assumed that these 364 incontinent women who responded negatively to the stem question would not likely have sought treatment. For these women, we imputed “no UI treatment-seeking” and included them in an expanded analytic sample with the primary results presented here. We compared the multivariable models for the limited and expanded analytic samples to examine the effect of our assumption.
Of the 3,302 women enrolled in the Study of Women's Health Across the Nation longitudinal cohort, 2,234 (68%) reported monthly or more UI at any visit between baseline and year 9. Of these women who reported UI, a total of 1,520 (68%) reported seeking UI treatment, most between baseline and year 9 (Fig. 1). Compared with the 2,251 women who remained in the Study of Women's Health Across the Nation through year 9, women who dropped out of the study or had missing data were more likely to be white, have lower annual income and education at baseline, have higher body mass indexes, and have either no medical insurance or public insurance. They were also more likely to report daily UI of urge or mixed type with a higher level of bothersomeness (data not shown).
The characteristics of incontinent women who did and did not seek UI treatment are shown in Table 1. Longitudinal characteristics of UI associated with seeking treatment in bivariate analyses were a longer duration of UI and more years reporting daily UI (Table 2).
In multivariable analyses, women had higher odds of seeking treatment when they reported experiencing symptoms for more than 7 years, when UI worsened over time, and when UI was at least daily and bothersome just before seeking treatment. Importantly, women who saw physicians regularly or had more visits for preventive women's health care were also more likely to discuss their UI with a health care provider. When stratifying our models by race or ethnicity (Table 3), by level of difficulty paying for basics, and by educations level (data not shown), we found few unique factors associated with seeking treatment. Likewise, in our primary multivariable models, interactions between UI frequency (P=.77) and duration (P=.94) and race or ethnicity as well as UI frequency (P=.60) and duration (P=.31) and difficulty paying for basics were not statistically significant. However, our power for detecting interactions between these variables was low (0.40 with α=0.05).
For our secondary analysis of incontinent women who had not sought treatment for their UI before year 7, overall, approximately 12% reported seeking UI treatment between visits 7 and 9 (approximately 6% per year). Like with our primary analysis, we found that frequency of UI in the year before seeking treatment and longer duration of UI had the strongest association with UI treatment-seeking. Among this smaller subset of women, we found no differences in the rates of seeking treatment by racial or ethnic group (data not shown).
The 364 incontinent women who did not answer the treatment-seeking questions because they answered “no” to the stem question “Have you ever leaked urine, even a small amount?” and for whom we imputed “no treatment-seeking” in our expanded analytic sample differed from those for whom we had complete data. They were more likely to be black or Asian, not born in the United States, and have a lower education level and annual household income. More than 50% of these 364 women had reported UI at least twice during the previous years; however, they were less likely to report daily UI, less likely to report mixed UI symptoms, and had a lower level of UI bothersomeness. When we compared our multivariable models of the expanded analytic sample with those in the limited analytic sample, we found few differences. Black women were statistically more likely than white women to seek treatment in our limited analytic sample (odds ratio [OR] 1.59, 95% confidence interval [CI] 1.18–2.16) than in our expanded sample (adjusted OR 1.25, 95% CI 0.77–2.04). Frequent UI in the year before seeking treatment was strongly associated with seeking treatment in our expanded analytic sample, but not in our limited analytic sample. Otherwise, point estimates for all other variables were similar.
In this 9-year, longitudinal study of midlife women, we found no clear or consistent evidence of racial, educational, or socioeconomic disparities in seeking treatment for UI. Rather, longitudinal characteristics of UI had the strongest association with UI treatment-seeking behavior. In cross-sectional studies, women with more frequent and bothersome UI symptoms were more likely to seek care.14,23,24 Over time, we found that worsening symptoms, longer duration of symptoms, and persistent symptoms of daily UI were associated with women seeking UI treatment.
We found that women with more health care contact were more likely to report seeking treatment for their UI.14 Women who seek care for other health concerns may be more likely to seek care for UI or more health care visits may provide more opportunity to discuss UI. Higher health care use may also represent a closer doctor–patient relationship. Of particular importance to gynecologists, the results of our study suggest that regular women's preventive health visits have the benefit of increasing the opportunity for incontinent women to discuss their UI.
Our study results do not support previous findings that black women or women with lower socioeconomic circumstances are either less likely to seek care or seek care only at a higher level of bother or UI frequency than white women or women of higher socioeconomic resources.25 Some of the intriguing findings in our bivariate analyses such as experience of discrimination, higher importance of spirituality, anxiety symptoms, and greater social support, which may be related to race or ethnicity, socioeconomic status, or education level, did not persist after adjustment in multivariable models but warrant further investigation.
The Study of Women's Health Across the Nation has collected a rich variety of data that allowed a multilayered exploration of changes in UI characteristics and UI treatment-seeking in a diverse community-based sample of women over 9 years. However, our study had some important limitations. Although our UI questions were very similar to those in validated questionnaires,26,27 such instruments were not available at the initiation of the Study of Women's Health Across the Nation. Women with UI were more likely to drop out of our study. Race or ethnicity, income, and education play a role in longitudinal cohort retention; white women with lower income and education were more likely to drop out of the Study of Women's Health Across the Nation. These factors may also have affected questionnaire responses because more black, Asian, and immigrant women with lower incomes and education did not respond to our treatment-seeking questions as a result of inconsistent UI reporting. Hispanic women could not be included in this analysis. The women who participated in the Study of Women's Health Across the Nation over 9 years represent those who actively engage in regular study visits and thus may be more likely to engage the medical system. Only 5% of our cohort reported not having health insurance, half the national average for this age group and timeframe,28 and thus our cohort had ready access to care. For all these reasons, our study participants were probably more likely to seek care than the general population. Some of our analyses were limited by small numbers. For some of our analyses, we did not know exactly when in the intervening period between baseline and visit 7 a woman sought UI treatment; patterns of UI (such as worsening or no change in UI) may have crossed over the point of UI treatment-seeking rather than clearly preceded it. However, our analysis of longitudinal UI characteristics in women who sought care for the first time after visit 7 yielded similar results, suggesting minimal effect of this uncertainty.
The results of our investigation are important for research, public health outreach, and the clinical care of women. Our findings demonstrate some important problems in UI epidemiologic research. We showed higher rates of inconsistent UI reporting among black, Asian, and immigrant women on our questionnaires, suggesting differences in either meaning given to their UI problem or problems in interpretation of the questionnaire. Consideration should be given to developing instruments for assessing and validating change in UI over time in diverse populations. For public health educators, messages that encourage women to seek care for worsening or a longer duration of UI may be important for encouraging women to seek care. Clinicians should be more vigilant questioning infrequent health care system users about this common condition that may be affecting quality of life. While women may not report their UI problem at its onset, clinicians can use preventive women's health visits to educate patients about returning for worsening of UI symptoms, thus establishing a relationship that encourages seeking treatment for UI at future health care visits.
1. Hannestad YS, Rortveit G, Sandvik H, Hunskaar S; Norwegian EPINCONT study, Epidemiology of Incontinence in the County of Nord-Trøndelag. A community-based epidemiological survey of female urinary incontinence: the Norwegian EPINCONT study. Epidemiology of Incontinence in the County of Nord-Trondelag. J Clin Epidemiol 2000;53:1150–7.
2. Waetjen LE, Liao S, Johnson WO, Sampselle CM, Sternfield B, Harlow SD, et al.. Factors associated with prevalent and incident urinary incontinence in a cohort of midlife women: a longitudinal analysis of data: study of women's health across the nation. Am J Epidemiol 2007;165:309–18.
3. Botlero R, Bell RJ, Urquhart DM, Davis SR. Urinary incontinence is associated with lower psychological general well-being in community-dwelling women. Menopause 2010;17:332–7.
4. Margareta N, Ann L, Othon L. The impact of female urinary incontinence and urgency on quality of life and partner relationship. Neurourol Urodyn 2010;28:976–81.
5. Salvatore S, Serati M, Laterza R, Uccella S, Torella M, Bolis PF. The impact of urinary stress incontinence in young and middle-age women practising recreational sports activity: an epidemiological study. Br J Sports Med 2009;43:1115–8.
6. Minassian VA, Yan X, Lichtenfeld MJ, Sun H, Stewart WF. Predictors of care seeking in women with urinary incontinence. Neurourol Urodyn 2012;31:470–4.
7. Omli R, Hunskaar S, Mykletun A, Romild U, Kuhry E. Urinary incontinence and risk of functional decline in older women: data from the Norwegian HUNT-study. BMC Geriatr 2013;13:47.
8. Brown JS, Vittinghoff E, Wyman JF, Stone KL, Nevitt MC, Ensrud KE, et al.. Urinary incontinence: does it increase risk for falls and fractures? Study of Osteoporotic Fractures Research Group. J Am Geriatr Soc 2000;48:721–5.
9. Abrams P, Andersson KE, Birder L, Brubaker L, Cardozo L, Chapple C, et al.. Fourth International Consultation on Incontinence Recommendations of the International Scientific Committee: Evaluation and treatment of urinary incontinence, pelvic organ prolapse, and fecal incontinence. Neurourol Urodyn 2010;29:213–40.
10. Shaw C. A review of the psychosocial predictors of help-seeking behaviour and impact on quality of life in people with urinary incontinence. J Clin Nurs 2001;10:15–24.
11. Roberts RO, Jacobsen SJ, Rhodes T, Reilly WT, Girman CJ, Talley NJ, et al.. Urinary incontinence in a community-based cohort: prevalence and healthcare-seeking. J Am Geriatr Soc 1998;46:467–72.
12. Hagglund D, Walker-Engström ML, Larsson G, Leppert J. Quality of life and seeking help in women with urinary incontinence. Acta Obstet Gynecol Scand 2001;80:1051–5.
13. Seim A, Sandvik H, Hermstad R, Hunskaar S. Female urinary incontinence—consultation behaviour and patient experiences: an epidemiological survey in a Norwegian community. Fam Pract 1995;12:18–21.
14. Burgio K, Ives DG, Locher JL, Arena VC, Kuller LH. Treatment seeking for urinary incontinence in older adults. J Am Geriatr Soc 1994;42:208–12.
15. Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas 1977;1:385–401.
16. Gold EB, Colvin A, Avis N, Bromberger J, Greendale GA, Powell L, et al.. Longitudinal analysis of the association between vasomotor symptoms and race/ethnicity across the menopausal transition: study of women's health across the nation. Am J Public Health 2006;96:1226–35.
17. Sherbourne CD, Stewart AL. The MOS social support survey. Soc Sci Med 1991;32:705–14.
18. Williams DR, Spencer DIS, Jackson JS. Race, social identity, and physical health: interdisciplinary explorations. Oxford (United Kingdom): Oxford University Press; 1999.
19. Soules MR, Sherman S, Parrott E, Rebar R, Santoro N, Utian W, et al.. Executive summary: Stages of Reproductive Aging Workshop (STRAW). Fertil Steril 2001;76:874–8.
20. Waetjen LE, Feng WY, Ye J, Johnson WO, Greendale GA, Sampselle CM, et al.. Factors associated with worsening and improving urinary incontinence across the menopausal transition. Obstet Gynecol 2008;111:667–77.
21. Waetjen LE, Ye J, Feng WY, Johnson WO, Greendale GA, Sampselle CM, et al.. Association between menopausal transition stages and developing urinary incontinence. Obstet Gynecol 2009;114:989–98.
22. Akaike H. Information measures and model selection. Int Stat Inst 1983;22:277–91.
23. Fultz NH, Herzog AR. Self-reported social and emotional impact of urinary incontinence. J Am Geriatr Soc 2001;49:892–9.
24. Sampselle CM, Harlow SD, Skurnick J, Brubaker L, Bondarenko I. Urinary incontinence predictors and life impact in ethnically diverse perimenopausal women. Obstet Gynecol 2002;100:1230–8.
25. Berger MB, Patel DA, Miller JM, Delancey JO, Fenner DE. Racial differences in self-reported healthcare seeking and treatment for urinary incontinence in community-dwelling women from the EPI study. Neurourol Urodyn 2011;30:1442–7.
26. Brown JS, Bradley CS, Subak LL, Richter HE, Kraus SR, Brubaker L, et al.. The sensitivity and specificity of a simple test to distinguish between urge and stress urinary incontinence. Ann Intern Med 2006;144:715–23.
27. Lukacz ES, Lawrence JM, Buckwalter JG, Burchette RJ, Nager CW, Luber KM. Epidemiology of prolapse and incontinence questionnaire: validation of a new epidemiologic survey. Int Urogynecol J Pelvic Floor Dysfunct 2005;16:272–84.
28. National Center for Health Statistics. Health, United States 2010. Hyattsville (MD): National Center for Health Statistics; 2011.