Factors Associated With Nonadherence to Pelvic Floor Physical Therapy Referral for the Treatment of Pelvic Pain in Women

Abstract

Background

Patients with pelvic pain due to pelvic floor myofascial pain syndrome are often referred for pelvic floor physical therapy, the primary treatment option. However, many patients do not adhere to the treatment.

Objective

The purpose of this study was to examine the adherence rate and outcomes of patients referred for physical therapy for pelvic floor myofascial pain syndrome and identify risk factors associated with nonadherence.

Design

This was a retrospective cohort study.

Methods

ICD-9 codes were used to identify a cohort of patients with pelvic floor myofascial pain syndrome during a 2-year time period within a single provider's clinical practice. Medical records were abstracted to obtain information on referral to physical therapy, associated comorbidities and demographics, and clinical outcomes. “Primary outcomes” was defined as attendance of at least 1 visit. Secondary outcomes included attendance of at least 6 physical therapist visits and overall improvement in pain. Statistical analysis was performed using chi-square, Fisher exact, and independent t tests. Nonparametric comparisons were performed using Wilcoxon signed rank test. Multivariate analysis was completed to adjust for confounders.

Results

Of the 205 patients, 140 (68%) attended at least 1 session with physical therapy. At least 6 visits were attended by 68 (33%) patients. Factors associated with poor adherence included parity and a preexisting psychiatric diagnosis. The odds of attending at least 1 visit were 0.75 (95% confidence interval = 0.62–0.90) and 0.44 (95% confidence interval = 0.21–0.90), respectively. Patients who attended ≥ 6 visits were more likely to have private insurance (78%) and travel shorter distances to a therapist (mean = 16 miles vs 22). Patients with an improvement in pain (compared with those who were unchanged) attended an average of 3 extra physical therapist visits (mean = 6.9 vs 3.1).

Limitations

Limitations include reliance on medical records for data integrity; a patient population derived from a single clinic, reducing the generalizability of the results; the age of the data (2010–2012); and the likely interrelatedness of many of the variables. It is possible that maternal parity and psychiatric diagnoses are partial surrogates for social, logistic, or economic constraints and patient confidence.

Conclusions

Initial adherence to pelvic floor physical therapy was less likely for multiparous women and women with a history of psychiatric diagnosis. Persistent adherence was more likely with private insurance or if the physical therapist location was closer. Pain improvement correlated with increased number of physical therapist sessions.

Pelvic pain, defined as noncyclic pain that is present for at least 3 to 6 months, is one of the most common patient reports encountered by gynecologists and affects nearly 15% of women.¹ Pelvic pain can affect a woman's quality of life, her ability to work, and her ability to care for herself and her family.² The etiology can be complex, and the differential can be broad.³ Systems commonly involved include gynecologic, gastrointestinal, urologic, psychologic, neurologic, and musculoskeletal.⁴ One recent study estimated that 22% of patients with chronic pelvic pain have an associated musculoskeletal cause.⁵

Pelvic floor myofascial pain and spasm are part of the broader category of pelvic floor disorders, of which pelvic floor physical therapy (PFPT) is considered the first-line therapy and is widely accepted.⁶ Other pelvic floor disorder conditions include pelvic organ prolapse, overactive bladder, and incontinence. The efficacy of PFPT for the treatment of pelvic floor disorders has been well established; however, patient knowledge of this treatment option is often poor.^7–10 Additionally, adherence to PFPT has been reported as very low despite the fact that this is a low-risk treatment option with proven success.¹¹

Several studies have tried to evaluate PFPT adherence and identify barriers to adherence.^7,12 Alewijnse et al looked at predictors of long-term PFPT adherence in women with urinary incontinence and found that women were more likely to adhere if they had worse disease (more incontinence), had little-to-no sex education prior, and attended at least the initial PFPT evaluation and short-term physical therapy.¹³ Similarly, a study by Chen et al evaluated adherence in Taiwanese women with incontinence.¹¹ They also found women with more severe incontinence were more adherent, as were those who had higher self-efficacy with the home-based part of the therapy. A consensus statement on the issue of patient adherence to PFPT was published in 2015 and sponsored by the International Continence Society.⁷ The authors first note a lack of high-quality studies in this area and state further research is needed. However, they were able to conclude that patient-related factors were the most significant barrier to PFPT adherence factors such as positive perception of efficacy, perception of social pressure, and severity of disease. The most commonly studied indications for PFPT adherence were female incontinence and post-prostatectomy problems in males.

The clinical interaction with female pelvic floor myofascial pain syndrome patients may be different and is not well-studied. Significant clinical experience has revealed that most patients require extensive education on this diagnosis and the process of PFPT to maximize adherence with treatment. Despite focused counseling to all patients in this regard, a subset of patients chooses not to follow-up with PFPT recommendations despite apparent initial agreement with the plan.¹⁴ A recent study by Shannon et al randomized patients to an extra educational session prior to PFPT referral and still found greater than 50% nonadherence in both arms.¹² This study included patients with all pelvic floor disorders.

Patients with pelvic pain from pelvic floor disorders are an understudied group within this set of disorders. There may be additional unknown barriers to adherence in this group compared with other pelvic floor disorders. It is clear from prior studies that simply increasing education is not enough to improve adherence. As noted in the consensus statement, researchers need to identify all potential barriers to PFPT adherence to address individual patient needs and improve disease resolution.

The primary objective of this study was to determine adherence rates with PFPT in patients with pelvic pain from pelvic floor myofascial pain syndrome and to identify risk factors for nonadherence. In addition, our secondary outcome was to assess changes in pain scores and the association with PFPT attendance.

Methods

This study was a retrospective cohort of patients seen at an academic medical center in a single-provider's practice between July 1, 2010 and October 15, 2012. Inclusion criteria were age ≥ 18 years old, subjective report of chronic pelvic pain (defined as the presence of pain in the pelvic region for at least 3 months duration and occurring with or without menses), physical exam demonstrating pelvic floor myofascial pain syndrome (defined as reproduction of patient's pain when digital pressure applied to either levator ani, obturator internus, and/or piriformis muscles), initial agreement to pursue PFPT, and referral to PFPT. Women were excluded if they did not have insurance coverage (or private means of pay) for physical therapy or had attended PFPT in the preceding 6 months.

Patients were initially identified by ICD9 billing codes (729.1, 625.9, 625.0, 728.85, and 625.1) to be reviewed for possible study inclusion. Utilizing an electronic medical record database, a department billing data manager cross-matched the specific provider's billing data with the desired codes. Each individual chart was then reviewed by hand to assess inclusion/exclusion criteria. Of the 389 identified patients, 205 met these criteria.

Prior to referral to PFPT, each patient received the same educational information regarding the role of PFPT for pelvic floor myofascial pain syndrome and what to anticipate during the treatment. This was done by the same provider (A.Y.). The primary outcome was defined as initial adherence with at least 1 visit to PFPT. Two secondary outcomes were also defined: (1) full adherence to at least 6 PFPT visits, and (2) improvement in pelvic pain after referral for PFPT. For our study, we used a cutoff of 6 physical therapy visits to define full adherence. Most PFPT studies will report 10 visits as a complete therapy plan; however, we chose to decrease our requirement to 6 given that this is a real-world study, to allow for differences in insurance coverage, and to assure an adequate number of patients in order to determine which patient factors contributed to poor adherence.^8,15 Mean number of PFPT visits reported in similar studies for similar indications is 6 (where attendance at one-half of prescribed visits defined adherence).¹⁶

Data were collected by chart review and included demographic data (age, race, parity, insurance status), medical history (menopause status, complaint of dysuria or dyspareunia, psychiatric diagnosis), and PFPT information (distance travelled to PFPT using zip codes, and physical therapy adherence rates based on physical therapist progress reports). Referral to a particular physical therapist was based on a patient's insurance status and the proximity to their home or work address. Prior to referral, each therapist was vetted by the physician (A.Y.) to assure the type of therapy provided. Because our practice is an academic center with a large catchment area, several PFPT practice locations are included in the data. Attendance at physical therapy was determined by contacting the physical therapist to whom each patient was referred.

The included variables are consistent with those utilized in analytic models in previously published studies regarding PFPT attendance in urogynecology patients.¹⁶ Given that this is the first PFPT adherence study in a strictly pelvic pain population, other variables included in the analytic models were based on studies of adherence to medical therapies in a pelvic pain population and notable comorbidities affecting treatment outcomes.^17,18

Sample size was estimated based on previous research in this area. In previous studies regarding PFPT adherence, 50% attendance is common (when considering attendance at half of prescribed visits as adherent).¹⁶ Using 50% as a control, we estimated that chronic pain patients may have poorer attendance. Using a power of 80% (alpha = 0.05, 2-tailed) to detect a 20% difference in adherence rates (our primary outcome), the sample size was 186 patients.

Statistical analysis was performed using chi-square, Fisher exact, and independent t tests where appropriate. Nonparametric comparisons were performed using Wilcoxon signed rank test. Multivariate analysis was completed to adjust for confounders. One-way ANOVA was used to look at the difference in means among categories with more than 2 groups. Variables with P  < .05 were considered statistically significant. All statistical analyses were conducted using STATA v 12 (Stata, College Station, TX, USA).

This review was performed with the approval of the Vanderbilt University Medical Center Research Review Board.

Role of the Funder

This study was supported by CTSA award (ref. no. UL1TR000445) from the National Center for Advancing Translational Sciences. Its contents are solely the responsibility of the authors and do not necessarily represent official views of the National Center for Advancing Translational Sciences or the National Institutes of Health. The funders had no role in the design, conduct, or reporting of this study.

Results

A total of 205 patients met inclusion criteria for the study. Patient demographics and frequencies are summarized in Table 1. Our cohort attended a mean of 4 physical therapist visits (range = 0–57). For comparison, the table is divided into 3 groups: patients who went to no physical therapist visits (31%), those who went to 1–5 visits (35%), and those who went to 6 or more visits (33%). The distribution of age is relatively uniform across the 3 groups, with only 10% of patients over age 55 years. Patients who attended ≥ 6 visits were slightly but significantly (P = .06) older than the other 2 groups. Race, parity, and menopause status were evenly distributed across the groups.

Table 1.

Participant Demographics^a

Characteristic	Group With No PFPT (n = 65) Mean/Median (SD) or %	Group With 1–5 PFPT Visits (n = 72) Mean/Median (SD) or %	P b	Group With ≥ 6 PFPT Visits (n = 68) Mean/Median (SD) or %	P c
Age, mean (SD)	36.4 (11.1)	35.8 (12.6)	.50	39.6 (13.0)	.06
Race
White	53 (82)	59 (82)	.75	57 (84)	.26
Black	7 (11)	6 (8)		10 (15)
Other	5 (7)	7 (10)		1 (1)
Parity
0	7 (11)	25 (35)	.11	26 (38)	.29
1 or 2	28 (43)	23 (32)		21 (31)
3+	30 (46)	24 (33)		21 (31)
Premenopausal	40 (62)	54 (75)	.12	42 (62)	.29
Health insurance
Private	42 (65)	47 (65)	.99	53 (78)	.05
Medicare	5 (8)	6 (8)		8 (12)
Medicaid	15 (23)	16 (22)		5 (7)
Self-pay	3 (5)	3 (4)		2 (3)
Distance traveled, mi,d mean (SD)	22 (39)	22 (38)	.84	16 (41)	.02
Dysuria	6 (9)	8 (11)	.70	7 (10)	1.0
Dyspareunia	50 (77)	58 (81)	.35	46 (68)	.18
Dysmenorrhea	23 (35)	26 (36)	.93	15 (22)	.05
Pain with bowel movement	10 (15)	8 (11)	.46	8 (12)	.78
Psychiatric history	34 (52)	21 (29)	.006	29 (43)	.73
BMI, mean (SD)	25 (20)	26 (6)	.24	28 (14)	.008
Smoking	21 (32)	16 (22)	.18	9 (13)	.03
History of physical abuse	2 (3)	2 (3)	.87	6 (9)	.08
History of sexual abuse	5 (8)	3 (4)	.33	7 (10)	.29

^aValues are presented as numbers (percentages) of participants unless otherwise indicated. BMI = body mass index; PFPT = pelvic floor physical therapy.

^bThe reference category for the group with 1–5 PFPT visits was the group with no PFPT.

^cThe reference category for the group with ≥ 6 PFPT visits was the group with <6 visits.

^d1 mi = 1.6 km.

Patients who went to 6 or more visits had a higher percentage of private insurance (P = .05). Similarly, those who traveled shorter distances to their visits also went to more visits. The patients who attended less than 6 visits traveled a mean of 22 miles, whereas those with 6 or more visits traveled a mean of 16 miles (P = .02). Although 75% of patients lived within 40 miles of their physical therapist office, at least one-half of the cohort had to drive a minimum of 20 miles.

Within the entire cohort, the most commonly reported pain symptoms were dyspareunia (75%) and dysmenorrhea (31%). However, there were more complaints of dysmenorrhea among patients with fewer than 6 visits (35%) than those fully adherent to physical therapy (22%) (P = .05). Patients with a history of psychiatric illness or smoking status were less likely to present for any PFPT visits (P = .006). A history of physical or sexual abuse was not significantly related to the number of physical therapist visits attended. Patients who were smokers were less likely to be fully adherent (P = .03).

The results from a multivariate analysis of risk factors for physical therapist attendance are displayed in Table 2. Two variables, parity (which was defined in the model as having at least 1 child; odds ratio = 0.75; 95% confidence interval = 0.62–0.90) and history of psychiatric illness (odds ratio = 0.44; 95% confidence interval = 0.21–0.90), were statistically significant in their association with physical therapy attendance. Patients with a greater number of children and those with a psychiatric history were less likely to attend physical therapy at all. The association of parity and physical therapist attendance is further demonstrated in the Figure, which shows that the greater number of children, the fewer visits attended.

Table 2.

Risk Factors for Nonattendance to Pelvic Floor Physical Therapy (No Visits vs Any Visits)^a

Variable	Adjusted Odds Ratio (95% CI)
Age	1.01 (0.99–1.04)
Parity	0.75 (0.62–0.90)
Insurance status	1.83 (0.84–4.00)
Distance traveled	1.00 (0.99–1.01)
Psychiatric illness	0.44 (0.21–0.90)
History of physical abuse	2.16 (0.33–14.22)
History of sexual abuse	0.55 (0.14–2.15)

^aBold type indicates variables that are statistically significant. CI = confidence interval.

Figure.

Association of parity and physical therapy attendance.

Follow-up pain status was obtained on 125/205 (61%) patients. Mean pain score (on a Visual Analog Scale, 1–10 scale) decreased from 3.6 (SD = 3.3) to 2.5 (SD = 3.2) for the entire cohort with follow-up. Of the patients who followed-up, 66% reported improvement in pain. Table 3 shows the mean number of physical therapist visits among those who reported improvement, no change, and worsening pain. Patients who reported an improvement in pain had a mean of 7 physical therapy visits, whereas those who reported no improvement had a mean of 3 visits. This is statistically significant (P = .0003). More physical therapist visits were attended by those who worsened, although this group had only 6 patients, and is likely confounded by a worse pretherapy condition.

Table 3.

Association of Number of Physical Therapy Visits With Change in Pain Status

Pain Status (No. of Participants)	Mean No. of Physical Therapy Visits (SD)	P
Improved (n = 83)	6.9 (5.4)	.0003a
Unchanged (n = 36)	3.1 (4.5)
Worsened (n = 6)	12.7 (22.3)

^aComparison of group that had improved pain status with group that had unchanged pain status.

Discussion

Overall, 140 (68%) of our referred patients did attend at least 1 PFPT visit, whereas only 33% completed at least 6 visits. This is consistent with prior literature on this subject from other academic referral centers. Farahani et al reported on a urinary incontinence population (N = 254) referred for PFPT, of whom only 46% completed therapy.¹⁹ Similarly, another group studied the effect of prereferral consultation with a therapist prior to PFPT, and found the same adherence rates in the intervention and control groups: 67% attended 1 visit whereas only 43% completed all visits.²⁰ Finally, a recent study by Shannon et al of patients with all indications for PFPT referral found 66% of patients attended 1 visit whereas only 29% completed the PFPT course.¹⁶ The only significant factor associated with nonadherence in this study was race. However, most patients were referred for incontinence-related diagnoses. The authors did not stratify attendance rates by particular diagnosis.

The results from our study suggest that nonadherence with PFPT recommendations has 2 separate parts. In particular, the question of which patient factors may be correlated with attending any PFPT visit is different from the question of which factors influence how many visits a given patient attended. Our results suggest that the characteristics of patients who followed recommendations and attended at least 6 PFPT sessions were different from those who attended only 1 to 5 sessions and the group who did not attend any sessions. The patients who did attend at least 6 PFPT sessions were more likely to have private health insurance, travel less distance to attend a PFPT session, not complain of dysmenorrhea, have a lower BMI, and be nonsmokers. Some of these findings were not unexpected. In our area of practice, there were numerous pelvic floor-specialized physical therapists who accepted only private insurance. Although there were physical therapists who accepted public insurance payments or self-pay, anecdotally, patients did report an association with nonadherence and lack of insurance.

The inverse association between a complaint of dysmenorrhea and attendance to PFPT may be due to the fact that there are many effective treatments for dysmenorrhea that may significantly lessen any musculoskeletal component. As a result, once starting medical therapy, patients may have seen significant improvement without attending as many or even any PFPT sessions. Lower (more normal) BMI and nonsmoking status may be signs of a patient's increased attention to their health and hence related to increased adherence. Additionally, it also may be related to increased baseline physical activity and hence willingness to undergo a physical therapy solution. The significance of the parity correlation with lack of attendance was initially surprising. Our clinical experience suggests that women of greater parity tend to have greater challenges scheduling medical visits and appointments for themselves due to childcare or work constraints. In particular, for more detail, more specific social and economic factors will be worth investigating in a future study regarding PFPT adherence. Because this is the first study to our knowledge to look specifically at PFPT adherence for pelvic pain only, there are no prior studies to compare with in regards to associated risk factors. Most studies in this area involve incontinence and prolapse. In those patients, disease severity, better social support, and shorter distance to therapy have been shown to be associated with higher rates of attendance.^21,22 Social support (as it possibly relates to childcare) and distance to therapy were also risk factors for nonadherence in our population.

Among patients with urinary incontinence, 2 studies found a positive association with adherence and the severity of disease.^11,13 We did not find an association in our analysis between pelvic pain severity or type of pain (dysmenorrhea, dyspareunia, and dysuria) and increased PFPT adherence. The study by Alewijnse found that patients who received short-term PFPT were more likely to still adhere to long-term therapy.¹³ We did find, similarly, that most patients who presented to the initial PFPT evaluation and early treatment were more likely to attend at least 6 visits.

When comparing the group of patients who only attended 1 to 5 sessions with those who did not attend any, the only significant characteristic of the patients who attended at least 1 to 5 PFPT sessions was that these patients were less likely to have reported a psychiatric diagnosis. This finding supports a practice of directed counseling to this patient group, emphasizing that at least 1 session can have positive health benefits. It is plausible that this patient population may have additional challenges in trusting and adhering to medical provider recommendations. It is worth noting that no specific psychiatric diagnosis was singled out in our study.

As an additional point, we looked at the outcome of patient-reported pain improvement regardless of PFPT attendance status at their follow-up visit. We found that 66% of women noted improvement. In contrast, 29% did not note improvement. An analysis of the mean number of physical therapist visits between these 2 groups showed that those who noted improvement on average attended 3 more sessions than those who did not note improvement. Furthermore, the group who saw improvement, on average, attended more than 6 sessions. This value was statistically significant. This suggests that in this population of referred patients, increased visit attendance is correlated with improved pain symptoms.

The main strength of this study is that it is the only study we are aware of that sought to determine adherence to PFPT for treatment of pelvic pain, with particular focus on patient characteristics and risk factors. Previous studies have focused mainly on urinary or bowel complaints in this population.²³ Additional strengths of this study include the number of factors examined to assess both a medical and social association with PFPT adherence, the length of follow-up for most of our patient population, and the diversity of our patient population. We included women of all socioeconomic backgrounds and insurance status.

This study has several limitations. First is its retrospective nature, which contributes to reliance on medical records for data integrity. Second, because this patient population was derived from a single provider's clinic, this limits the generalizability of the results, and our findings may not apply to other pelvic pain populations. Additionally, the age of the data (2010–2012) might limit the applicability of the results to current times. This time frame was chosen due to changes in the primary provider's practice and referral patterns after 2012 and was an attempt to limit heterogeneity in the study population and bias in the results. Finally, another limitation is that many of our variables are likely interrelated. It is possible that maternal parity and psychiatric diagnoses are likely partial surrogates for social, logistic, or economic constraints and patient confidence affecting physical therapy adherence.

Until more in-depth, possibly prospective studies are published in this area, these data suggest that gynecology providers may be able to identify patients at risk for nonadherence. Given the proven results associated with PFPT adherence in the treatment of pain associated with pelvic floor muscle spasm, physicians should target at-risk patients and provide close follow-up and education, thereby potentially decreasing nonadherence rates. Additionally, our results may help guide physical therapists to identify patients who may be at risk for not completing the full prescribed course and to intervene early to prevent dropout.

Author Contributions

Concept/idea/research design: F. Aguirre, J. Heft, A. Yunker

Writing: F. Aguirre, A. Yunker

Data collection: F. Aguirre, J. Heft

Data analysis: F. Aguirre, J. Heft, A. Yunker

Project management: F. Aguirre, A. Yunker

Funding

This study was supported by CTSA award No. UL1TR000445 from the US Department of Health and Human Services, National Institutes of Health, and National Center for Advancing Translational Sciences. Its contents are solely the responsibility of the authors and do not necessarily represent official views of the National Center for Advancing Translational Sciences or the National Institutes of Health.

Disclosures

The authors completed the ICJME Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest.