Using language to identify a bladder pain component in women with Dysmenorrhoea-Related Pelvic Pain: A cross-sectional study.

From the 750 participants with DRPP, 468 women reported bladder pain (DRPPB+). This group was sub‐categorised into 281 participants with bladder pain present, and 187 participants reporting bladder pain only with a full bladder. Two hundred and eighty two women reported no bladder pain (DRPPB−). Participants were sub‐categorised by the presence of reported endometriosis: DRPPB+ 196 (41.9%) self‐reported endometriosis 5 (1.1%) self‐reported excluded endometriosis 267 (57.1%) unknown endometriosis status DRPPB− 107 (37.9%) self‐reported endometriosis 2 (0.7%) self‐reported excluded endometriosis 173 (61.3%) unknown endometriosis status Most women with bladder pain reported pain involving other pelvic organs: bowel, vulval, dyspareunia and ‘other’ pelvic pain. Within the DRPPB+ cohort, 52 (21.37%) women reported the presence of all other types of pain represented in the survey (listed above; Table  1 ). Only 12 (4.26%) women with DRPPB− reported the presence of all types of pain. Presence of other locations of pain The survey collected 316 000 words of responses to descriptions of participants' pain. When describing dysmenorrhoea, bloating ( P  = 0.014) and pressure ( P  = 0.031) were more strongly associated with DRPPB+ compared to DRPPB−. DRPPB− subjects were more likely to use the word excruciating ( P  < 0.001) to describe their dysmenorrhoea compared to DRPPB+ (Table  2 ). Pain descriptors used to describe dysmenorrhoea from free‐form text responses in positive dysmenorrhoea‐related pelvic pain with a bladder pain component (total) and without a bladder component The bold values are statistically significant. An analysis of variance test showed bloating ( P  = 0.030) had a significant association with participants reporting bladder pain all the time, compared with bladder pain when holding on and DRPPB−. Dull ( P  = 0.0077) was more significantly associated with both bladder pain when holding on and DRPPB−, compared to those with bladder pain all the time. Excruciating ( P  = 0.0021) was more strongly associated with DRPPB− than the other two cohorts. Some women with unknown endometriosis status described suspicions that endometriosis may be present and currently undiagnosed, with reasons including financial constraints in seeking and securing medical diagnosis, or feeling dismissed by their PHCP with ‘constant battles to have my pain recognised and validated by people in the medical field’ [Participant 19]. Radiating ( P  = 0.028) was the only descriptor with a statistically significant association to bladder pain and a reported diagnosis of endometriosis, compared to those without bladder pain and endometriosis. For women with bladder pain, the frequency of pain descriptors was obtained using AntConc concordance software. Descriptors of bladder pain were related to pressure, distending, and stretching. Most commonly, fullness ( n  = 80, 17.09%), sharp ( n  = 79, 16.88%), pressure ( n  = 62, 13.24%), emptying ( n  = 54, 11.54%), aching ( n  = 44, 9.40%), stabbing ( n  = 43, 9.12%) and burning ( n  = 42, 8.97%) were used. The descriptors from the pre‐determined list that were statistically significantly associated with the presence of bladder pain were pounding ( P  < 0.001), tingling ( P  < 0.001), unbearable ( P  = 0.001), burning ( P  = 0.02), stabbing ( P  = 0.010) and cramping ( P  = 0.021), using the Pearson χ 2 test (Table  3 ). Dull and aching were not significantly associated with the presence of bladder pain symptoms. Pain descriptors selected from a predetermined list in positive dysmenorrhoea‐related pelvic pain with a bladder pain component (total) and without a bladder component Unbearable ( P  = 0.026), burning ( P  = 0.044) and tingling ( P  = 0.003) showed a statistically significant association between DRPPB+ and a reported endometriosis diagnosis, compared to DRPPB− and endometriosis, using the Pearson χ 2 test (Table  4 ). While not statistically significant, the descriptors stabbing , cramping , and aching were words chosen frequently by both cohorts, indicating need for qualitative interpretation. Pain descriptors selected from a predetermined list in those with bladder pain and reported endometriosis

Our research has taken a symptom‐based approach best suited to the needs of PHCP seeking to better understand the language used by women with DRPP. This study identifies that women with a bladder pain component to their DRPP describe dysmenorrhoea differently from women without a bladder pain component. The descriptors women use for their self‐perceived bladder pain were also analysed for comparison, and these can be identified by directly asking patients about their bladder pain. The descriptors used for dysmenorrhoea could alert PHCP to a potential co‐existing bladder component, and a possible underlying bladder condition. This is particularly important in a time‐limited primary health care context. Chronic overlapping pain conditions have multiple aetiologies and diverse and complex clinical manifestations. 16 DRPP fits this definition. While several discrete aetiologies are associated with DRPP, there is also a clear overlap between pelvic visceral pain conditions. In this study, participants showed complex pelvic pain with multiple symptoms. A higher number and variety of pain descriptors used may reflect a higher complexity of pelvic pain within an individual, with more types of pelvic pain perceived in different pelvic organs/locations. 17 Hypervigilance, rumination and catastrophising are all constructs associated with central sensitisation 18 and may reveal themselves in language, but this was not explored here. Dysmenorrhoea is the most common presenting symptom accompanying endometriosis, 19 and recurrent dysmenorrhoea may be a precursor to CPP. A study of 100 women with dysmenorrhoea only found that 16.4% transitioned to CPP within a year, 20 and 50% had transitioned to CPP within 12 years. 20 In the current study, the presence or absence of reported endometriosis did not affect the language women used, regardless of their bladder pain status. The one exception was radiating , which was used significantly more frequently by women with reported endometriosis and DRPPB+ ( P  = 0.028). Laparoscopic surgery to remove endometriosis lesions is well established as part of pelvic pain management. However, commonly women report more than one type of pelvic pain, and the extent of endometriosis lesions correlates only weakly with the severity of pain. 21 Despite full surgical excision of endometriosis lesions, pain is not always relieved. 21 This increases risks and economic burdens of surgery for women least likely to benefit. 22 The enhanced understanding of language used by women with a perceived bladder pain component may better determine the origin of pain symptoms. Women may achieve more timely investigation, improved identification and management of pelvic pain symptoms with medical and surgical interventions targeted to those who would benefit most. 23 The use of language analysis to aid medical diagnosis is a cost‐efficient way to improve equity of access to care within low‐resource settings where the information provided by laparoscopic assessment may not be available. 7 Language can also enhance information‐gathering and empathy‐building. 24 The written answers allowed participants to have time in their own environment to reflect anonymously, and there were no word limits on their responses. Since the study was advertised in pelvic pain social media forums, it captured responses from a range of women with DRPP. There are currently no available studies that have analysed the language of pelvic pain in women. The 316 000 words of responses collected by this survey can be used to design future focus group studies. This self‐selected descriptive data aid further understanding of pelvic pain language with potential clinical implementation. Given the open‐ended questions and spontaneity of word choices, the frequency of pelvic pain descriptors chosen was sometimes low. However, the large cohort size made it possible to study small differences between word choices that might not have otherwise been evident, and enabled a greater variety of words to be captured, particularly in an exploratory study. The written word may differ from spoken dialogue, which is the dominant channel of patient consultations. Pain descriptors can overlap semantically, eg twisting and squeezing . This study treated each descriptor as an independent entity, rather than grouping them into semantic families of words. Subsequent research may identify semantic groups of terms which are collectively associated with specific pain conditions. This survey was advertised on social media platforms that may bias women where endometriosis is their main diagnosed pathology. The data contained some respondents where endometriosis was explicitly excluded. In other cases, respondents may simply not have recorded their endometriosis status, although the emphasis with which respondents wrote about their endometriosis suggested its prominence in their pain experience. Verified clinical diagnoses would require comparison to consultation notes and operation reports. This is an initial guide for PHCP seeking accurate descriptions of patients' pelvic pain to assist in their differential diagnoses, investigations, and management. Where further developed, this could potentiate a wider patient‐centred language tool for women with pelvic pain, 9 facilitating empowered communication between women and their PHCP 25 within a time‐restricted setting. This language tool could inform a pre‐consultation questionnaire to assist PHCPs carry out a more focused, efficient 26 and effective clinical consultation, enabling better pain diagnosis and management.

Chronic pelvic pain (CPP) affects up to 26% of the world's female population, 1 yet for over 60% of affected women no definitive diagnosis of the cause of pain is made. 2 CPP, defined as non‐cyclical lower abdominal or pelvic pain present on most days for >3–6 months, 3 has compounding impacts on well‐being, personal relationships, educational opportunities, work attendance and professional development. 4 , 5 Dysmenorrhoea‐Related Pelvic Pain (DRPP) is a term used within this research study to describe CPP where dysmenorrhoea is one of the symptoms present. The term DRPP is used here to distinguish the target group from women with dysmenorrhoea as a stand‐alone symptom, and from women with broader CPP where dysmenorrhoea may not be present. While DRPP is often associated with the presence of endometriosis, an analysis of 14 pain‐related symptoms in 168 Australian women with DRPP referred to a gynaecologist found a similar symptom profile in women both with and without a lifetime diagnosis of endometriosis. 6 The only statistically significant difference between these groups was the frequency of bladder symptoms, which were present in 20% of women with laparoscopically confirmed endometriosis, compared to 50% of women with laparoscopically excluded endometriosis ( P  = 0.005). 6 This suggests that a proportion of women presenting with DRPP may have an undiagnosed underlying bladder pain component contributing to their overall pain experience that is not identified during routine laparoscopic diagnosis. These women may be undergoing repeat surgery where pain persists, for a pelvic pain condition that may have responded to non‐surgical methods targeting symptoms related to the bladder. 7 The diagnostic potential of language in clinical medicine has been a growing area of research since Melzack (1975), especially in the absence of other proof of the existence, intensity, nature, and location of reported pain. Sullivan & Derbyshire 8 state ‘[…] pain is fundamentally subjective with self‐report providing the most complete and reliable access to another's pain’. Specific patterns of language to aid clinical decision‐making are well developed within some clinical contexts. A peri‐umbilical intensifying colicky pain migrating to the right abdomen that becomes sharp and constant suggests appendicitis. 9 However, a broadly based, widely recognised specific pattern of language use for DRPP, including where a bladder component may contribute to pain symptoms, does not currently exist. 10 Identifying specific patterns of language to improve communication between patients and primary health care practitioners (PHCP), often working in time‐restricted and low‐resource settings has the potential to assist diagnosis. This study reports on the language used by a cohort of women with DRPP who differ according to the presence or absence of bladder symptoms, and aims to identify bladder involvement within the perception of pain.

This was a mixed methods cross‐sectional study using an online survey ( www.surveymonkey.com ) to collect structured and free text data. Australian and New Zealand women with DRPP aged 18–49 years completed a voluntary online survey titled ‘The Language of Pelvic Pain in Women’. To maximise recruitment of women with DRPP, the survey was advertised through social media platforms including Pelvic Pain Foundation of Australia, Endometriosis New Zealand, Endometriosis Australia, the Canberra Endometriosis Network, QENDO, EndoActive, and Body in Mind. These posts explained the research intent and nature of the survey. Women with pelvic pain self‐enrolled via a hyper‐link, connecting them to study details, risks and benefits of involvement, data storage plan, investigator contact details, and required electronically signed informed consent before access to the survey. Ethics approval was obtained through the Human Research Ethics Committee (University of Adelaide: 2017‐100). The research was conducted in accordance with the Declaration of Helsinki. Participants had access to anxiety/depression and sexual/ domestic abuse support services. The survey was designed to take 30 min to complete, and invited women to describe their pain in both structured free text and guided formats. The survey captured patient demographics, menstrual cycle history (age at menarche, age when dysmenorrhoea began), the average number of days per month with pain, and the functional deficits of pain, based on interference with daily tasks. A free‐text section followed asking women to describe potential pain triggers. To identify basic patterns of language use, participants were individually invited to describe their different subsets of pain (period, bladder, bowel, vulval, dyspareunia, and ‘other’ pelvic pain). Participants rated the intensity of each subgroup of pain on a validated numerical rating scale 11 from zero to ten (ten being the worst pain imaginable). Free‐form text boxes requested that participants describe each subgroup of pain and its effect on their life, with no word limit. This open forum is the start of identifying basic patterns of language use. Participants were asked to select from a list of eight pre‐determined relevant descriptors that best described their DRPP in general. As the language of pelvic pain is limited in the literature, descriptors were selected in consultation with a clinically experienced gynaecologist and pain specialist, and in combination with the McGill Pain Questionnaire, 12 a tool for rating pain developed by Melzack and Torgerson in 1971. The list of descriptors followed the free‐form responses, so the value of descriptors provided earlier was not diminished. During the six‐month period of availability in 2018, 1034 participants responded to the survey. Of these participants, 750 were included in this study, based on requirements of being an adult of reproductive age (18–49 years), having a communicative proficiency of English, having DRPP, eg dysmenorrhoea as a component of their pelvic pain, and having completed the relevant parts of the survey (listed below). Participants were excluded given their age (≥50 years), the absence of DRPP, missing responses in the dysmenorrhoea or bladder pain free‐text questions, or a history of pelvic mesh surgery since the pathophysiology of this type of pain is entirely separate. 13 Dysmenorrhoea was defined to participants as pain during periods. All participants had pelvic pain, and women with dysmenorrhoea (DRPP) were analysed further. Participants with DRPP were divided into two groups: with (DRPPB+) and without bladder pain (DRPPB−). Bladder pain was defined by self‐report based on responses to: Do you have bladder pain? Yes/Only when holding on/No Are you able to tell us what your bladder pain feels like? Do you have bladder pain? Yes/Only when holding on/No Are you able to tell us what your bladder pain feels like? This study is based in a PHCP setting where investigative efforts and specialist consultations may be incomplete, and often prior to invasive or surgical procedures. Hence, endometriosis was defined as an unprompted self‐report of this condition. Participants were de‐identified and allocated a numerical code. The transfer table linking participants' personal data to the codes was stored in a password‐protected Microsoft Excel file accessible only to the researchers. The frequency of discrete pain descriptors from the free‐text boxes were initially analysed qualitatively using AntConc concordance software. 14 Each descriptor was allocated a binary code (0 = absence, 1 = presence). StataCorp Stata Statistical Software 15 was used to compare the frequency of these descriptors. The free‐form and predetermined descriptors between DRPPB+ and DRPPB− were analysed using Pearson χ 2 tests. Logistic regression tests were used to compare variance across more than two groups (bladder pain present, bladder pain when holding on and no bladder pain). The effect of self‐report of endometriosis on descriptor use was analysed using Pearson χ 2 tests. A P‐ value of <0.05 was used to determine statistical significance.

Table S1. Pain descriptors used to describe dysmenorrhoea from free‐form text responses in those with bladder pain, bladder pain when holding on and no bladder pain. Table S2. Pain descriptors used to describe dysmenorrhoea from free‐form text responses in women with positive dysmenorrhoea‐related pelvic pain with a bladder pain component (DRPPB+) and reported endometriosis. Table S3. Pain descriptors selected from a predetermined list in those with bladder pain, with bladder pain only when holding on and without bladder pain. Figure S1. Flow chart of the division of participants.