Saiki, Lori S.; Cloyes, Kristin G.
Urinary incontinence affects from 26.5% to 51% of midlife women (Morrill et al., 2007; Nygaard et al., 2008). Disclosure rates of incontinence concerns are relatively poor, with estimates ranging from only 10% to 60% of affected women (Dugan et al., 2001; Kinchen et al., 2003; Morrill et al., 2007). Themes identified as factors contributing to a lack of disclosure of incontinence symptoms and the negative effects of incontinence symptoms on social, interpersonal, and personal quality of life include embarrassment (Kinchen et al., 2003; Melville, Wagner, Fan, Katon, & Newton, 2008), fear of what others may think (Li, Low, & Lee, 2007; Nicolson, Kopp, Chapple, & Kelleher, 2008), and feeling old and less functional as a woman and as an adult (Garcia, Crocker, & Wyman, 2005; Hägglund & Ahlström, 2007; Kralik, Koch, & Telford, 2001; Peake & Manderson, 2003). Two predominant themes identified in studies of the lived experience of incontinence are privately controlling the symptoms (Bradway, 2005; Wang, Chen, Jou, & Tsao, 2011) and, subsequently, hiding the condition from others (Hayder & Schnepp, 2010; Zeznock, Gilje, & Bradway, 2009). These recurring themes suggest that women may assess the social risks that disclosure brings before seeking care for urinary incontinence.
Self-disclosure involves risk assessment and uncertainty about the response from the recipient of the disclosed information (Saiki & Lobo, 2011). Healthcare providers may facilitate the movement from perceiving uncertainty as a danger to framing uncertainty as an opportunity (Mishel, 1990). Self-presentation theory suggests that people rely on communicated impressions of self to ascertain safety in interactions (Goffman, 1959). Identifying women’s communication strategies used to probe for safety in disclosing incontinence is critical to providing care.
Women’s public blogging offers a natural language data source to explore women’s communication patterns surrounding urinary incontinence. Blog text closely resembles conversational speech patterns (Neustein, 2007). Language surrounding health concerns is significant data and offers critical insights to inform interventions. Frequently occurring patterns in word usage contain information about what is in the forefront of people’s thoughts, feelings, or beliefs when they use language to communicate (Neuendorf, 2002). Because language is a social activity used to communicate thoughts, feelings, beliefs, and intentions, the words people use as they present with health concerns mean something. The purpose of this study was to explore empirical patterns of word use in women’s blog communications about incontinence for insights into social risk and self-presentation by women living with incontinence. How do women with incontinence present themselves in public communication? What insights can be gained regarding specific communication patterns present in blog text that healthcare professionals could potentially use to facilitate disclosure of incontinence?
This study is a descriptive, empirical content analysis of women’s blog text about urinary incontinence. Content analysis is a systematic, procedural set of methods in which empirical analyses of extant textual data (words in a text) inform interpretation of the intentions or characteristics of the author or the audience as well as the messages or meanings of text (Krippendorff, 2004; Neuendorf, 2002). Content analysis has the potential to utilize rich qualitative data to empirically support or challenge current healthcare practices (Cloyes, Berry, Reblin, Clayton, & Ellington, 2012). Content analysis of publically archived blog text may provide a unique perspective on women’s self-presentation of life with incontinence. Patterns in self-presentation may suggest cues that could be used to facilitate disclosure of incontinence.
The blogs used to create the textual data for this study were archived posts published in the public domain. The blogs were created with the intention of being available for public reading and response. Blogs were de-identified and aggregated as one text document for analysis. No quotes or identifiable phrases or passages were used to support the analysis. No information was available from the blog posts to identify any demographic traits other than gender and adulthood. No human beings were enrolled as human subjects, linked with identifiers to the blog posts, contacted for further information, or otherwise involved in any aspect of the research. For these reasons, the University of Utah Institutional Review Board and the University of New Mexico Human Research Protection Office waived human subjects’ review for this content analysis of publically archived blog text about incontinence.
A purposive, convenience sampling strategy found 23 blog texts through Google and Wordpress search engines using the search phrase “blogs about incontinence.” Blogs included for analysis were written in English by 19 women about their own incontinence (inclusive dates: 2006–2011). Self-report of incontinence is a valid and reliable means of identifying cases (Sandvik, Seim, Vanvik, & Huskaar, 2000). Blogs excluded for analysis were written by health professionals and the incontinence products industry or if the incontinence was a result of dementia, spinal cord injury, or prostate cancer. Informational content imported from other sources (interview with health professional, review of anatomy/physiology of the pelvic floor) was also excluded. The compiled blog text was analyzed by General Inquirer, Yoshikoder, usingenglish.com, NVivo9, and WordStat software (N = 16,629 words).
Content Analytic Strategies
The aggregated blog text was analyzed for both manifest content (words present in the text) and latent content (inferred meaning of words used in the text) with word frequency, valence, and key-word-in-context (KWIC) analysis. Valence refers to the positive or negative connotation of a word. KWIC analysis examined word context by close reading of 15-word concordances (frames) around the key word of interest. Patterns in KWIC word co-occurrence were tested for significance by Z scores (p < .05), an appropriate measure of relative proportion. Latent content was first explored using established construct dictionaries developed based on extensive research and application in the disciplines of linguistics and psychology, particularly social psychology (Krippendorff & Bock, 2009). The dictionaries used in this study were the Linguistic Inquiry and Word Count (Pennebaker & King, 1999) and the Lasswell dictionary (General Inquirer, 2012). A custom latent construct dictionary was also developed. Construct definitions were derived from previous conceptual development of self-disclosure (Saiki & Lobo, 2011) and informed by Mishel’s (1990) reconceptualized uncertainty in illness theory and Goffman’s (1959) self-presentation theory. Word lists used to identify the construct in the blog text were developed by an iterative process that considered in vivo word frequencies, close reading of concordances, and valence (Saldaña, 2009). Three rounds of coding resulted in the final custom latent construct dictionary. Definitions and word lists for the latent constructs—social risk assessment, audience, divulgence, response, and self-concept—are presented in Table 1.
Cluster analysis was performed using WordStat. In WordStat, clusters are derived from an index of similarity (word co-occurrence or proximity in the text): Jaccard’s coefficient = a / (a + b + c), where a represents cases where both items occur and b and c represent cases where one item is found but not the other (Provalis Research, 2011, p. 100). The parameter for co-occurrence was preset at a 15-word window around the key word. This window is fairly narrow to search for words used in close proximity and, therefore, interpreted to be close in meaning association, either intended or unintended (Krippendorff, 2004) by the blogger.
The word tag cloud of the 100 most frequently used words in the blog text confirmed that a major theme was “my bladder incontinence,” and bloggers frequently used “I” and “you” (Figure 1; NVivo9). A “hard word” tag cloud (>2 syllables, n = 478; usingenglish.com) indicated frequent use of the words “incontinence,” “seriously,” “embarrassing,” “generally,” “horrible,” “everyone,” “exercises,” “pregnancy,” and “menopause.”
The most frequently used verbs in the blog text were variants of “urinate” (n = 72), forms of the cognitive verb “know” (n = 38), and the affective verb “feel” (n = 35). Other frequently occurring verbs included forms of “tell” (n = 25) and “say” (n = 21). From the Linguistic Inquiry and Word Count dictionary, pronoun use was the largest category (12.96%) in the blog text, with “self” (which includes others with I such as “we,” “our,” and “us”) used more often than “I” (8.42% vs. 8.09%).
From the Lasswell dictionary, time/space (6.95%) and enlightenment (3.82%) were the most frequent constructs. Enlightenment is defined as an understanding of interpersonal and cultural social interactions, and “enlightenment gained through shared thought and information” was the largest subcategory present in the blog text. Physical well-being (2.84%) was the third most frequently represented overall category.
Word frequencies of the custom latent constructs were also analyzed. Social risk assessment accounted for 1.1% of the blog text, compared with references to incontinence (a major focus of the blog text at 1.4% of the text). In contrast, only 0.3% of the text was about pelvic floor muscle therapy (PFMT), and 0.2% of the text referred to seeking care from healthcare providers. The construct divulgence accounted for 0.6% of the text. Positive self-concept words (1.1% of the blog text) were often in the context (verified by KWIC analysis) of “health” and being physically “fit” (38%). Positive self-concept words were oriented toward improvement, personal agency, and assertion of normalcy. Negative self-concept words (1.3% of the blog text) were strongly negatively charged feelings that arose from incontinence (e.g., “horrible,” “devastating,” “mortified,” “crying,” and “unfeminine”).
Comparative frequencies of key words in 15-word frames were calculated. Response (see Table 1) from the audience that contextually reflected facilitating social connection (e.g., “help,” “kindly”) was more frequently found with close (65%) connections than with community (20%) or healthcare providers (15%). Blocking social connection (e.g., “annoyed,” “silence”), however, was frequently found with the community category (78%). The difference in the proportion of close contacts demonstrating facilitating social connections versus close contacts demonstrating blocking social connections was significant (Z = −2.68, p = .004). Community contacts were significantly more likely to be represented in the blocking social connections category than the facilitating social connections (Z = −2.97, p = .002). Differences between the frequency of facilitative (e.g., “helped,” “listened”) and blocking reactions from healthcare providers were not significant (Z = −0.28, p = .39). Examples of blocking language used in reference to providers were “up to me,” “on my own,” or a description of raising the topic year after year to no avail. Of the 22 references to PFMT found in the blog text, the valence of two (9.1%) were positive (e.g., “improved,” “taught me how”), 14 (63.6%) were negative (e.g., “don’t work,” “unnatural”), and six (27.3%) were neither positive nor negative (e.g., “need to do,” “supposed to do”).
Cluster analysis of words from the latent constructs divulgence, audience, and social risk assessment (criterion for clustering set at greater than three co-occurrences) resulted in three major clusters (Figure 2); cry and horrible formed an isolated cluster that did not include audience or divulgence. The second major cluster, with subclusters inclusive of “husband” and “friend” (close audience), included embarrassed/embarrassing and then “joke-awful-taboo” and “told/problem(s)-mention/coping.” The third major cluster, with “blog” and “people” (community audience), included “talk,” “share,” “story,” and “hope,” with an associated subcluster of “depressing” and “issues.”
The general impression gleaned from the manifest word frequency analyses is that the blog text was “about my seriously embarrassing generally horrible bladder incontinence beginning with pregnancy or menopause” and is indicative of an overall negative impact of incontinence, supporting findings of previous studies (e.g., Kinchen et al., 2003; Melville et al., 2008). This content analysis adds the following insights. A concentration of pronouns that merge self with others in inclusive context, inclusive cognitive language, and a strong representation of social references to others combined to indicate that the primary content thread found in women’s blog text about incontinence was a focus on securing social connection and a sense of place in the human community. This interpretation is further supported by the frequent use of words from Lasswell’s enlightenment category, indicating a strong thread of the importance of social connectedness in the blog text. The high frequency of words representing Lasswell’s physical well-being was an unanticipated finding. The custom latent construct positive self-concept contained many words that described the blogger as otherwise healthy, fit, capable, intelligent, and living a productive life. This construct provided an interesting counterpoint to the negative aspects of living with incontinence. Some women might present themselves as healthy, active, and physically fit rather than reveal the degree of symptom impact on their quality of life.
Cluster analysis results indicated that women with incontinence may reach out to others by initially testing the social waters for risks to self-esteem or safety in disclosure. Partners and friends (and even providers) may be joked with or told about the embarrassing problems related to incontinence but are not privy to the extent of emotional distress that surrounds the personal experience of incontinence. The cluster analysis indicates that sharing stories about incontinence in the public forum of blogs was a hopeful act. This finding suggests a new strategy to improve women’s experiences when seeking care for incontinence concerns that would have been missed with word frequency analysis alone. Results from this cluster analysis suggest that any mention of incontinence in the clinical setting should be considered a request for information or support.
This preliminary study was limited by the nature of the data collected and the number of archived blogs available for analysis. No demographic information was collected. Therefore, social or cultural factors that may influence perceptions of the impact of incontinence on quality of life (Welch, Botelho, & Tennstedt, 2011) were not analyzed. The researchers could not probe for clarification in language or word use. The researchers were not able to discern motivations to write the blogs or reasons why women would or would not seek care for incontinence.
Empirical trends in patterns of word choice in women’s naturally occurring, public blog communications about incontinence offer new insights in how women may assess the social risks to disclosing incontinence. On the basis of study findings, women with incontinence may present as being healthy and fit rather than risk a possible negative reaction by the healthcare provider. Further research is needed to explore whether providers might facilitate disclosure of urinary incontinence by first acknowledging the woman’s strengths, thereby creating a sense of safety and acceptance. Probing for safety in disclosure may take the socially neutral form of making a casual reference to incontinence or adopting a joking manner. This study suggests that the level of symptom distress reported during a wellness examination or on a symptom bother questionnaire may not represent the level of hidden distress that women with incontinence actually feel. This study also suggests that there is a need for more effective training and ongoing support for PFMT as an early intervention strategy. Cluster analysis indicates that communicating about incontinence is a hopeful act, although it may be perceived as a social risk. Any attempts by women to initiate a conversation regarding urinary incontinence should be encouraged in a sensitive manner.
Bradway C. (2005);. Women’s narratives of long-term urinary incontinence. Urologic Nursing. , 25:, 337–344.
Cloyes K. G. Berry P. H. Reblin M. Clayton M. Ellington L. (2012);. Exploring communication patterns among hospice nurses and family caregivers: A content analysis of in-home speech interactions. Journal of Hospice & Palliative Nursing. , 14:, 426–437.
Dugan E. Roberts C. P. Cohen S. J. Preisser J. S., Davis C. C., Bland D. R., Albertson E. (2001);. Why older community-dwelling adults do not discuss urinary incontinence with their primary care physicians. Journal of the American Geriatrics Society. , 49:, 462–465.
Garcia J. A. Crocker J. Wyman J. F. (2005);. Breaking the cycle of stigmatization: Managing the stigma of incontinence in social interactions. Journal of Wound, Ostomy and Continence Nursing. , 32:, 38–52.
Goffman E. (1959); . The presentation of self in everyday life. New York, NY: Anchor Books.
Hägglund D. Ahlström G. (2007);. The meaning of women’s experience of living with long-term urinary incontinence is powerlessness. Journal of Clinical Nursing. , 16:, 1946–1954.
Hayder D. Schnepp W. (2010);. Experiencing and managing urinary incontinence: A qualitative study. Western Journal of Nursing Research. , 32:, 480–496.
Kinchen K. S. Burgio K. Diokno A. C. Fultz N. H., Bump R. Obenchain R. (2003);. Factors associated with women’s decisions to seek treatment for urinary incontinence. Journal of Women’s Health. , 12:, 687–698.
Kralik D. Koch T. Telford K. (2001);. Constructions of sexuality for midlife women living with chronic illness. Journal of Advanced Nursing. , 35:, 180–187.
Krippendorff K. (2004); . Content analysis: An introduction to its methodology. ( 2nd ed.). Thousand Oaks, CA: Sage.
Bock M. A.(Eds.). (2009); . The content analysis reader. Los Angeles, CA: Sage.
Li F. L. W. Low L. P. L. Lee D. T. F. (2007);. Coping with stress incontinence: Women’s experiences of everyday inconveniences. International Journal of Urological Nursing. , 1:, 112–119.
Melville J. L. Wagner L. E. Fan M.-Y. Katon W. J. Newton K. M. (2008);. Women’s perceptions about the etiology of urinary incontinence. Journal of Women’s Health. , 17:, 1093–1098.
Mishel M. H. (1990);. Reconceptualization of the uncertainty in illness theory. Image: The Journal of Nursing Scholarship. , 22:, 256–262.
Morrill M. Lukacz E. S. Lawrence J. M. Nager C. W. Contreras R. Luber K. M. (2007);. Seeking healthcare for pelvic floor disorders: A population-based study. American Journal of Obstetrics & Gynecology. , 197:, 86.e1–86.e6.
Neuendorf K. A. (2002); . The content analysis guidebook. Thousand Oaks, CA: Sage.
Neustein A. (2007);. Sequence package analysis: A new method for intelligent mining of patient dialog, blogs and help-line calls. Journal of Computers. , 2:, 45–51.
Nicolson P. Kopp Z. Chapple C. R. Kelleher C. (2008);. It’s just the worry about not being able to control it! A qualitative study of living with overactive bladder. British Journal of Health Psychology. , 13:, 343–359.
Nygaard I. Barber M. D. Burgio K. L. Kenton K. Meikle S. Schaffer J. Brody D. J. (2008);. Prevalence of symptomatic pelvic floor disorders in US women. Journal of the American Medical Association. , 300:, 1311–1316.
Peake S. Manderson L. (2003);. The constraints of a normal life: The management of urinary incontinence by middle aged women. Women & Health. , 37:, 37–51.
Pennebaker J. W. King L. A. (1999);. Linguistic styles: Language use as an individual difference. Journal of Personality and Social Psychology. , 77:, 1296–1312.
Provalis Research. (2011); . WordStat user. ’s manual. Montreal, QC: Provalis Research.
Saiki L. S. Lobo M. L. (2011);. Disclosure: A concept analysis. Journal of Advanced Nursing. , 67:, 2713–2722.
Saldaña J. (2009); . The coding manual for qualitative researchers. (1st ed.). Los Angeles, CA: Sage.
Sandvik H. Seim A. Vanvik A. Huskaar S. (2000);. A severity index for epidemiological surveys of female urinary incontinence: Comparison with 48-hour pad-weighing tests. Neurourology and Urodynamics. , 19:, 137–145.
Wang Y.-H. Chen S.-H. Jou H.-J. Tsao L.-I. (2011);. Doing the best to control: The experiences of Taiwanese women with lower urinary tract symptoms. Nursing Research. , 60:, 66–72.
Welch L. C. Botelho E. M. Tennstedt S. L. (2011);. Race and ethnic differences in health beliefs about lower urinary tract symptoms. Nursing Research. , 60:, 165–172.
Zeznock D. E. Gilje F. L. Bradway C. (2009);. Living with urinary incontinence: Experiences of women from “The Last Frontier.”. Urologic Nursing. , 29:, 157–163,