IBS patients’ treatment expectancy and motivation impacts quality of the therapeutic alliance with provider: Results of the IBS Outcome Study

Abstract

Background:

In the absence of a satisfactory medical or dietary treatment, the quality of the therapeutic alliance between IBS patients and their provider is deemed critical to managing refractory IBS. Surprisingly, little research has been conducted on the nature of the therapeutic alliance, factors that influence it, or practical strategies to improve it. This study sought to identify actionable variables that impact therapeutic alliance in patients with refractory IBS.

Methods:

Subjects included a total of 436 Rome III-diagnosed IBS patients (80% F, M age 5 41 years) who completed a battery of clinical measures at the beginning of the acute treatment phase of an NIH behavioral trial. Pretreatment candidate predictor variables were organized into four categories: sociodemographic, extraintestinal, interpersonal, clinical (e.g., symptom severity, pain intensity), cognitive (e.g., treatment motivation, expectancy for improvement). Alliance was assessed by patient and clinician-rated measures of the Working Alliance Inventory (WAI) after first treatment session.

Results:

Patient reports of alliance were most strongly and consistently predicted by patient access to interpersonal support β = .16 (95% CI = .07/.25), motivation for symptom improvement, β = .12 (95% CI = .02/.21), and expectancy of symptom improvement, β = .35 (95% CI = .25/.44). Therapist ratings of alliance also were predicted by patient expectancy of symptom improvement, β = .16 (95% CI = .05/.26).

Conclusions:

When managing IBS, a focus on dynamic factors of treatment motivation, social support, and treatment expectancy may be useful in improving the quality of the therapeutic alliance between patient and provider.

Irritable bowel syndrome (IBS) is a common, painful, often disabling GI disorder whose cardinal symptoms (abdominal pain, diarrhea and/or constipation) affects up to 40 million Americans (1), making it the most common GI disorder seen by gastroenterologists and primary care physicians. Its adverse impact on work productivity and quality of life (2) is comparable to life-threatening conditions such as diabetes mellitus and hepatitis (3). The day-to-day burden of IBS is increased by limited efficacy of medical (4) or dietary treatment(5, 6). Inter-individual variation in treatment response to available medical therapies suggests that their therapeutic benefit depends on a variety of non-specific factors including the quality of the interaction between doctor and patient (i.e., therapeutic alliance).

The therapeutic alliance refers to the quality and nature of the patient(6)–provider relationship, their collaborative interaction, and bond that emerges during treatment(7). In the absence of curative agents, a strong patient-doctor relationship is regarded as “paramount” (8) for engaging patients, optimizing health outcomes, increasing patient satisfaction (9), and enhancing motivation for change particularly among treatment-resistant patients (10). That said, there is surprisingly little research on the alliance in IBS, or for that matter, other GI diseases. In one of the few published studies, the quality of physician-patient interactions based on chart review of physician notes was associated with reduced repeated office visits for IBS (11), a finding that is particularly important in a value-based health care environment that places a premium on efficiency and efficacy (4, 12).

One can argue that the quality of therapeutic alliance is so well accepted at a general level (13) that it has impeded our understanding of its critical components or factors that influence its formation. If predictors that influence alliance are identified, efforts can be made to strengthen the quality of the relationship between IBS patients and doctors and the effectiveness of treatments they prescribe. Thus, it is important to progress beyond questions of whether alliance is important to outcome but to questions of what are its critical components and how alliance develops (14).

We sought to identify actionable variables that may influence therapeutic alliance in patients with refractory IBS (i.e., persistent symptoms, impaired quality of life, and repeated consultations despite medical therapy) (4). Candidate variables with potential predictive value were drawn from the broader literature and organized into four conceptually distinct categories (See Figure 1): sociodemographic (age, gender, education level); extra-intestinal (number of medical, psychiatric comorbidities, somatization, emotional distress, neuroticism); interpersonal (e.g., social support, interpersonal problems); clinical (predominant bowel type, duration of illness, average pain intensity, number of prior procedures, IBS symptom severity), cognitive (perceived distress, fear of GI symptoms, pain catastrophizing, self-efficacy, anxiety sensitivity) including expectancy of and motivation for improvement. Drawing from the literature (15, 16), we predicted a hierarchy of factors that are related to alliance. The more distal factors such as socio-demographics and extra-intestinal medical comorbidities at the base of the triangle should have the weakest relationship to therapeutic alliance. Cognitive variables, particularly treatment motivation, located at the top of the triangle should be the most proximal influences on therapeutic alliance. We also were interested in knowing whether the same – or different -- predictors governed therapeutic alliance as measured by provider versus patient. Therefore, we assessed therapeutic alliance from both the patient and provider perspective using parallel versions of the same measure as part of a NIH funded behavioral trial. To minimize the extent to which alliance ratings were confounded by treatment response and maximize generalizability of study findings to drug and non-drug treatments, alliance data were collected at the beginning of the treatment delivery phase of an NIH-funded behavioral trial.

Figure 1.

Proposed Categories for Influenes on Therapuatic Alliance Organized from Most Distal to Most Proximate

Methods

Participants

Participants included 436 IBS patients recruited at two tertiary academic medical centers in Buffalo, NY and Chicago, IL as part of an NIH-funded clinical trial of two behavioral self-management treatments, the details of which can be found elsewhere (17–19). Participants were enrolled primarily through local media coverage, community advertising, and physician referral. To be eligible for the study, all participants must have met the Rome III diagnostic criteria(19)as determined by a board-certified gastroenterologist. Because this study was conducted as part of a clinical trial for moderate to severely affected patients with IBS participants also must have reported IBS symptoms of at least moderate intensity (i.e., symptom frequency of at least twice weekly for a minimum duration of 6 months and causing life interference). Subjects with presence of a comorbid organic GI illness that would adequately explain GI symptoms; mental retardation; current or past diagnosis of schizophrenia or other psychotic disorders; current diagnosis of depression with suicidal ideation; current diagnosis of psychoactive substance abuse were excluded. Institutional review board approval (UB, May 19, 2009; NU, December 19, 2008) and written, signed, informed consent was obtained before study initiation. The study was completed in full compliance with the Declaration of Helsinki.

Procedure

After a brief telephone interview to determine eligibility, participants were scheduled for a medical examination administered by a board-certified gastroenterologist to confirm the diagnosis of IBS based on Rome III criteria (20) and testing drawn from a battery of psychometrically validated measures detailed below.

Measures

Comorbidity.

Medical comorbidity was assessed using a comorbidity checklist that covers 112 medical conditions common to IBS patients (21) assessing diagnosed conditions present in the past 3 months. Number of medical comorbidities yields a total comorbidity score. Psychiatric comorbidity was assessed with MINI International Neuropsychiatric Interview (MINI). The MINI (22) is a fully structured diagnostic interview that generates psychiatric diagnoses according to the fourth edition of the DSM (23) assessing six classes of conditions: anxiety disorders, mood disorders, somatization disorders, anxiety-depression, and antisocial personality disorder. Scores were created by summing the number of disorders in which the patient met criteria. Somatization was measured using the Screening for Somatoform Symptoms-7(SOMS-7, 24). The SOMS-7 includes a total of 53 medically unexplained physical symptoms (e.g., dry mouth, heart palpitations). Respondents are asked to report the symptoms that have been present during the past 7 days. The total number of endorsed unexplained somatic symptoms yields a somatization symptom count. Higher scores indicate more frequent unexplained somatic complaints. Neuroticism. The biological disposition to experience emotions strongly was assessed using the Trait subscale of the State-Trait Personality Inventory (25). Its 20 items require subjects to indicate how they generally feel by rating the frequency of their feelings of anxiety on a four-point scale ranging from 1 (almost never) to 4 (almost always). Higher scores reflect higher predisposition to experience strong emotions across situations.

Social Variables.

Three conceptually distinct aspects of social functioning were assessed. The Kraus Emotional Support Scale (KESS, (26) is a four-item scale to assess the amount of social support that is available to the patient. Participants rate their social support on a six-point scale 1(very often) to 5 (did not want/need) and 6 (don’t’ know). The four items are averaged to create a negative interaction index. High scores indicate more frequent negative interactions. Inventory of Interpersonal Problems (IIP, (27) is a 32-item measure that assess a person’s most salient interpersonal difficulties, such as, social inhibition, non-assertiveness, and neediness. Participants rate their difficulties on a five-point scale ranging from 0 (nota at all) to 4 (extremely). Higher scores indicate higher levels of interpersonal problems. Interpersonal Support Evaluation List (ISEL-12, (28) is a 12-item measure that assesses the perceived availability of four types of social support (i.e., appraisal, belonging, self-esteem, tangible support). The scale ranges from 1 (definitely false) and 4 (definitely true) and higher scores indicate higher levels of perceived support.

Clinical Variables.

IBS symptom severity was measured using the 5- item Irritable Bowel Syndrome Severity Scale (IBS-SSS) (29). The IBS-SSS is a validated rating scale that requires participants to rate the overall severity of their GI symptoms on five, 100-point visual analog scales. Scales measure the severity of abdominal pain, frequency of abdominal pain, severity of abdominal distension, dissatisfaction with bowel habits, and interference with quality of life. Items are summed with total scores ranging from 0 to 500. Higher scores indicate more severe symptoms during the past week. Abdominal pain intensity over the previous 7 days was measured with an 11-point numerical rating scale (NRS), where 0=no pain and 10=worst possible pain (30).

Cognitive Variables.

Fear of GI symptoms was assessed using the 15-item Visceral Sensitivity Index (VSI, (31). Items were rated on a 6-point Likert scale (0 = strongly disagree to 5 = strongly agree) and specify a possible negative consequence to the experience of GI symptom-specific anxiety. Scores ranged from 0 (no fear of GI symptoms) to 75 (strong fear of GI symptoms). Higher scores indicate greater fear of GI symptoms. The Perceived Stress Scale(32) used was the four-item version. Participants indicate on a five-point scale ranging from 0 (never) to 4 (very often) the degree to which situations in their lives are appraised as stressful. Scores range from 0 to 16 and higher scores indicated higher levels of perceived stress. The IBS Self-Efficacy Scale is 25-items that measure patients’ confidence in their ability to control and manage IBS symptoms (33). Each item was scored on a 7-point scale, 1 (strongly disagree) to 7 (strongly agree). Items were summed to derive an overall total score of IBS self-efficacy ranging from 25 to 175. Higher scores indicate higher levels of self-efficacy. The 2-item version of the Pain Catastrophizing subscale of the abbreviated Coping Strategies Questionnaire (CSQ, 34) asks patients to rate the frequency with which they engage in thoughts that index catastrophizing during pain episodes. Respondents rate each item using a scale ranging from 0 (never do) to 7 (always do). Higher scores indicate higher levels of pain catastrophizing. Anxiety Sensitivity Inventory (ASI; (35) was used to measure the belief that anxiety and arousal symptoms have dangerous consequences. The 16 items of the Anxiety Sensitivity Inventory are rated on a six-point scale (0 = very little, 5 = very much) and yield a range of scores ranging from 0 to 80 with higher scores signifying greater anxiety sensitivity.

Motivation and Treatment Expectancy.

On the basis of prior behavioral research, assessing motivation incorporates two qualities: valance (i.e., how important is this outcome?) and strength (i.e., how likely is this outcome to occur?). To this end, patients were asked to rank order their primary treatment goals from a list of 12 choices (e.g., reduce the severity of IBS symptoms, get a medical diagnosis for GI symptoms, learn ways I can manage IBS symptoms on my own). They were then asked to rate how important their top ranked goal was (1 = not important, 2 = slightly important, 3 = moderately important, 4 very important, 5 = extremely important). Patients then rated their confidence in achieving their highest ranked goal. A third global motivation item asked patients to rate their motivation to learn ways to control or reduce their IBS symptoms (1 = not motivated and 10 = greatest motivation I ever had). Treatment success expectancy was assessed with an item from the IBS version (36) of the Credibility / Expectancy Questionnaire(37) asking how much improvement the patient felt would really occur as rated from 0% to 100%.¹

Dependent Variable

The Working Alliance Inventory – short form (WAI-Short Revised; (38, 39) is a 12-item self-report questionnaire of the therapeutic alliance that measures clinician and patient agreement on overall treatment goals, the steps to achieve these goals, and the emotional quality of their bond. Responses range from 1 (never) to 5 (always) and higher scores indicate a stronger working alliance. This questionnaire is comprised of three subscales assessing a focus on agreeing on the goals of therapy, the tasks of therapy, and the bond that develops between the therapist and the patient. Clinicians completed a parallel version. Both alliance measures were completed at the end of session 1 before the full scope of treatment strategies were implemented.

Data analysis

To test the influence of the different domains (socio-demographics, extra-intestinal medical comorbidities, interpersonal variables, clinical variables, cognitive variables, motivational/expectancy variables) on therapeutic alliance, we conducted a series of stepwise regression analyses. In each step, we entered the variables representing one of the domains (see Table 1 for a description of the measures within each domain). The variables that achieved statistical significance in each step were retained in the subsequent steps. The order in which we entered the blocks of variables was based on the model presented in Figure 1. The triangle in the figure indicates a hierarchy of factors that influence therapeutic alliance with the more distal factors such as socio-demographics and extra-intestinal medical comorbidities at the base of the triangle and cognitive variables, particularly expectancy and treatment motivation, at the top of the triangle as they should be the most proximal influences on therapeutic alliance. Blocks were entered into the regression analysis beginning with the more distal factors in the first blocks and the more proximal factors in the later blocks while controlling for those variables that showed statistical significance in the earlier blocks. Power to detect a small effect size of R² =.03 is adequate. Even if all thirty variables are maintained in the final equation, with an alpha set at .05, and a desired power of at least .80, the minimum sample size necessary to detect a significant effect of that size is 208 participants, less than half of our sample size of 436.

Table 1.

Variables and Measures

Domain	Construct	Measure	Mean (SD)/ %
Sociodemographic	Age	Years	41.39 (14.79)
	Gender	Self-identified (1=male, 2=female)	80.3% Female
	Race	Self-identified (0=white, 1=minority)	7.6% Minority
	Education	Highest level achieved	2.2% College degree
	Employment	Self-identified (0=unemployed,1= employed)	63.5% Employed
Extraintestinal	Psychiatric comorbidity	Count of psychical disorders endorsed	1.21 (1.62)
	Medical comorbidity	Count of medical disorders endorsed	4.64 (4.92)
	Somatization	SOMS-7	11.87 (6.35)
	Stress	How often stressed past 12 months	2.27 (0.91)
	Neuroticism	Trait Anxiety Inventory	20.05 (5.88)
	Self-rated health	1= excellent, 5= poor	2.84 (0.89)
Social	Emotional Support	Kraus Emotional Support Scale (KESS)	11.83 (2.92)
	Interpersonal Difficulties	Inventory of Interpersonal Problems	34.34 (18.10)
	Social Support	Interpersonal Support Evaluation List	40.04 (6.44)
Clinical	IBS Type	Dummy coded IBS-C & IBS-D variables	29.8% C, 43.1% D
	Duration of IBS Symptoms	Self-reported years	17.05 (14.39)
	Average Pain	Average abdominal pain past 7 Days (0–10)	5.01 (1.97)
	Symptom Severity	IBS-SSS Total Score	281.86 (72.08)
	Number of Diagnostic Tests	Self-reported number of past GI medical tests	4.31 (3.27)
	Previous IBS Treatment	Self-reported (0=no, 1= yes)	70.0% Yes
Cognitive	Catastrophizing	Pain Catastrophizing Scale (PCSQ)	2.58 (1.67)
	Perceived Stress	Perceived Stress Scale (PSS)	6.73 (3.21)
	Anxiety Sensitivity	Anxiety Sensitivity Scale (ASI)	23.28 (11.59)
	Visceral Sensitivity	Visceral Sensitivity (VSI)	44.59 (14.11)
	Self-Efficacy	IBS Self-efficacy IBS_SE	98.95 (25.01)
Motivation/Expectancy	Motivation to Achieve Tx Goal	(1 -10)	9.21 (1.25)
	Importance of Tx Goal	(1 -100)	71.97 (20.06)
	Likelihood achieving of Tx Goal	(1 -100)	61.66 (26.59)
	Expectancy of Tx Success	(0-100)	58.71 (21.55)

Results

Patient Total Working Alliance Score

The first column of Table 2 presents the results of the analysis on the overall therapeutic alliance score. In the first step of the analysis, the Working Alliance Inventory total score was regressed on the socio-demographic variables, F (5, 411) = 2.08, p =.06, R² = .03, RSME = 1.000. The only individual variable to reach statistical significance was gender which was then retained in all additional steps in the analysis. When extra-intestinal medical comorbidities were added in the next step, F (7, 416) = 1.42, p =.19, R² = .03, RSME = 0.984, no variables reached statistical significance. Social variables were entered in the next step, F (4, 422) = 6.40, p < .01, R² = .06, RSME = 0.968. Gender and interpersonal support were significantly related to alliance and were retained into the next step. While controlling for gender and interpersonal support, clinical factors related to IBS were entered in the next step, F(9, 414) = 4.39, p < .01, R² = .09, RSME = 0.959. Therapeutic alliance was significantly predicted by interpersonal support and symptom severity. Therapeutic alliance was next regressed on gender, interpersonal support, symptom severity and all cognitive factors, F(8, 413) = 5.86, p < .01, R² = .10, RSME = 0.950. In this step interpersonal support, symptom severity, and self-efficacy were significantly related to alliance. Finally, controlling for gender, interpersonal support, symptom severity, and self-efficacy, therapeutic alliance was regressed on motivational and expectancy factors, F(8, 406) = 14.23, p < .01, R² = .22, RSME = 0.858. Interpersonal support, symptom severity, motivation to reduce or control symptoms, and expectancy of treatment success were all significantly related to alliance.

Table 2.

Results for Stepwise Regression of WAI Total and Subscales on Variables Within the Five Domains

	Alliance	Goals	Tasks	Bonding
	β (95% CI)	β (95% CI)	β (95% CI)	β (95% CI)
Demographics
Age	−.08 (−.18/.01)	−.07 (−.17/.02)	−.07 (−.17/.02)	−.08 (−.18/.02)
Gender	.12 (.05/.54)*	.12 (.06/.55)*	.11 (.04/.53)*	.09 (−.02/.47)
Race	.03 (−.26/.46)	.03 (−.23/.49)	.04 (−.23/.49)	.00 (−.37/.35)
Education	−.05 (−.09/.03)	−.09 (−.12/.01)	−.06 (−.10/.02)	.03 (−.04/.08)
Employment	−.03 (−.13/.08)	−.03 (−.13/.07)	−.04 (−.14/.06)	.02 (−.08/.12)
Extraintestinal
Gender	.09 (−.02/.48)	.09 (−.01/.48)	.09 (−.03/.47)	--
Psychiatric comorbidity	.03 (−.10/.15)	.02 (−.11/.14)	.00 (−.12/.13)	.07 (−.06/.19)
Medical comorbidity	.04 (−.07/.15)	.10 (−.02/.21)	.05 (−.06/.16)	.01 (−.11/.12)
Somatization	.05 (−.06/.17)	.05 (−.06/.16)	.07 (−.05/.18)	.02 (−.09/.14)
Stress	−.10 (−.20/.01)	−.09 (−.20/.02)	−.09 (−.20/.02)	−.09 (−.20/.02)
Neuroticism	−.08 (−.22/.05)	−.07 (−.21/.06)	−.07 (−.20/.06)	−.10 (−.24/.03)
Self-rated health	−.01 (−.12/.09)	−.01 (−.12/.10)	−.01 (−.11/11)	−.05 (−.15/.06)
Social
Gender	.09 (.00/.47)	.10 (.02/.49)*	.09 (.01/.48)*	--
Emotional Support	.04 (−.07/.14)	.05 (−.06/.16)	.04 (−.07/.15)	.03 (−.08/.13)
Interpersonal Problems	.09 (−.01/.19)	.09 (−.02/.19)	.11 (.01/.21)*	.03 (−.07/.13)
Interpersonal Support	.21 (.10/.33)**	.16 (.04/.27)**	.19 (.08/.31)**	.21 (.10/.33)**
Clinical
Gender	.06 (−.09/.39)	.07 (−.06/.43)	.07 (−.06/.42)	--
Interpersonal Problems	--	--	.10 (.00/.21)	--
Interpersonal Support	.20 (.11/.30)**	.15 (.05/.24)**	.23 (.13/.34)**	.22 (.13/.32)**
IBS- C	.03 (−.18/.31)	.01 (−.22/.27)	.01 (−.24/.26)	.05 (−.14/.36)
IBS- D	.07 (−.08/.37)	.06 (−.11/.35)	.09 (−.05/.40)	.03 (−.17/.29)
Duration of IBS Symptoms	−.02 (−.11/.07)	−.03 (−.12/.07)	−.01 (−.10/.09)	−.01 (−.11/.08)
Average Pain	.10 (.00/.21)	.14 (.03/.24)*	.11 (.01/.22)*	.03 (−.07/.14)
Symptom Severity	.14 (.03/.24)*	.10 (−.01/.21)	.12 (.01/.23)*	.15 (.04/.26)**
Number of Diagnostic Tests	.00 (−.09/.10)	.01 (−.09/.10)	−.02 (−.12/.07)	−.01 (−.10/.09)
Previous IBS Treatment	.01 (−.18/.23)	.06 (−.07/.35)	.00 (−.22/.20)	−.01 (−.23/.19)
Cognitive
Gender	.08 (−.03/.44)	.08 (−.03/.44)	.08 (−.05/.43)	--
Interpersonal Problems	--	--	.11 (−.01/.22)	--
Interpersonal Support	.21(.11/.30)**	.16 (.06/.26)**	.21(.10/.31)**	.23 (.13/.33)**
Symptom Severity	.15 (.04/.25)**	--	.11 (.00/.22)	.15 (.04/.25)**
Average Pain	--	.17 (.07/.26)**	.12 (.01/.22)*	--
Catastrophizing	.03 (−.08/.15)	.01 (−.11/.12)	.03 (−.09/.14)	.01 (−.11/.12)
Perceived Stress	−.04 (−.14/.07)	−.04 (−.15/.07)	−.09 (−.20/.02)	−.02 (−.13/.09)
Anxiety Sensitivity	.08 (−.04/.19)	.06 (−.05/.18)	.07 (−.04/.19)	.09 (−.02/.21)
Visceral Sensitivity	.09 (−.03/.20)	.16 (.04/.27)**	.03 (−.09/.14)	.04 (−.08/.16)
IBS Self-Efficacy	.11 (.01/.21)*	.11 (.01/.21)*	.13 (.02/.22)*	.06 (−.04/.16)
Motivation
Gender	.03 (−.16/.29)	.04 (−.12/.32)	.03 (−.14/.30)	--
Interpersonal Problems	--	--	.10 (.00/.19)	--
Interpersonal Support	.16 (.07/.25)**	.12 (.03/.21)**	.17 (.08/.27)**	.20 (.11/.30)**
Symptom Severity	.13 (.03/.23)**	--	.12 (.02/.23)*	.12 (.02/.22)*
Average Pain	--	.07 (−.03/.16)	.01 (−.09/.11)	--
Visceral Sensitivity	--	.15 (.06/.24)**	--	--
IBS Self-Efficacy	.06 (−.03/.15)	.07 (−.02/.16)	.06 (−.03/.16)	--
Importance of Tx Goal	−.01 (−.11/.09)	−.02 (−.12/.07)	−.07 (−.17/.03)	.03 (−.08/.14)
Likelihood Achieving Tx Goal	−.03 (−.13/.07)	−.01 (−.11/.09)	.02 (−.01/.19)	−.06 (−.16/.04)
Motivation Achieve Tx Goal	.12 (.02/.21)*	.12 (.02/.21)*	.09 (−.01/.19)	.11 (.01/.21)*
Expectancy of Tx Success	.35 (.25/.44)**	.35 (.26/.45)*	.37 (.28/.47)**	.18 (.08/.28)**

Variables in italics significant at previous steps and carried over. Dashes indicate variable not entered in equation predicting that particular outcome variable. Bolded are consistently statistically significant across all subscales. CI = Confidence Interval.

^*p <.05

^**p < .01

Working Alliance Subscales

Parallel analyses to those described above were conducted on each of the Working Alliance Inventory subscales. Results of these analyses tended to mirror those for the overall score although there were some differences.

Goals. The second column of Table 2 presents the results of the analysis on the goals subscale of the WAI. In the first step, the goals subscale was regressed on the socio-demographic variables, F (5, 410) = 2.63, p =.02, R² = .03, RSME = 0.997. The only variable to reach statistical significance was gender which was retained through all subsequent steps. When extra-intestinal comorbidities were added in the next step, F (7, 405) = 1.93, p =.06, R² = .03, RSME = 0.981, no variables reached statistical significance. Controlling for gender, social variables were entered in the next step, F (4, 421) = 4.67, p < .01, R² = .04, RSME = 0.968. Gender and interpersonal support were significantly associated with the WAI goals subscale and were retained into the next step. Clinical factors related to IBS were entered next, F(9, 413) = 3.66, p < .01, R² = .07, RSME = 0.964. The WAI goals subscale was significantly related to interpersonal support and average abdominal pain, both of which were retained into the next step along with gender. The WAI goals subscale was next regressed on cognitive factors, F(8, 413) = 6.08, p < .01, R² = .11, RSME = 0.946, while controlling for gender, interpersonal support, and average abdominal pain. When cognitive factors were added to the equation, interpersonal support, average abdominal pain, fear of GI symptoms (visceral sensitivity), and self-efficacy were significantly related to alliance goals. When the goals subscale was regressed on motivational and expectancy factors while controlling for variables significant in previous steps, interpersonal support, fear of GI symptoms (VS), motivation to reduce or control symptoms, and expectancy of treatment success were significantly related to alliance goals, F(9, 405) = 14.32, p < .01, R² = .24, RSME = 0.846. Tasks. The third column of Table 2 presents the results of the analysis on the tasks subscale of WAI. In the first step, the tasks subscale was regressed on the sociodemographic variables, F (5, 411) = 2.21, p =.05, R² = .03, RSME = 0.996, and again the only individual variable to reach statistical significance was gender. Extra-intestinal medical comorbidities were added in the next step, F (7, 406) = 1.46, p = .18, R² = .03, RSME = 0.990. While controlling for gender, social variables were entered next, F (4, 422) = 5.70, p < .01, R² = .05, RSME = 0.976. Gender, interpersonal problems, and interpersonal support were significantly related to task alliance scores. Clinical factors related to IBS were entered next, F(10, 412) = 4.10, p < .01, R² = .09, RSME = 0.963. The WAI tasks subscale was significantly related to interpersonal support, average abdominal pain, and symptom severity all of which were retained into the next step along with interpersonal problems and gender. When cognitive factors were added to the equation, F(10, 409) = 4.89, p < .01, R² = .11, RSME = 0.951, interpersonal support, average abdominal pain intensity, and IBS self-efficacy, were significantly related to task alliance. Finally, when the WAI tasks subscale was regressed on the variables significant in past steps and motivational and expectancy factors, interpersonal support, symptom severity, and expectancy of treatment success were the significant predictors, F(10, 402) = 13.03, p < .01, R² = .25, RSME = 0.938. Emotional Bonding. The fourth column of Table 2 presents the results of the analysis on the tasks subscale of WAI. The WAI bonding subscale was regressed on the sociodemographic variables, F (5, 410) = 1.43, p =.21, R² = .02, RSME = 1.004, and no variables were significant predictors. When extraintestinal medical comorbidities were added in the next step, F (6, 406) = 0.83, p = .55, R² = .01, RSME = 0.999, gain no variables reached statistical significance. Social variables were entered next, F (3, 423) = 6.76, p < .01, R² = .05, RSME = 0.979, and interpersonal support significantly predicted bonding alliance. Controlling for interpersonal support, clinical factors related to IBS were entered next, F(8, 414) = 4.04, p < .01, R² = .07, RSME = 0.967. The WAI bonding subscale was significantly predicted by interpersonal support and symptom severity both of which were retained into the next step. When cognitive factors were added to the equation, F(7, 413) = 5.23, p < .01, R² = .08, RSME = 0.961, only interpersonal support and symptom severity were significantly associated with the alliance bonding subscale. Lastly, when the WAI bonding subscale was regressed on motivational and expectancy factors, F(6, 409) = 9.28, p < .01, R² = .12, RSME = 0.936, interpersonal support, IBS symptom severity, motivation, and treatment expectancy were significantly associated with bonding.

Therapist Working Alliance Ratings

Because patient and therapist ratings of overall therapeutic alliance were correlated, r (433) = .27, p <.001, we also examined whether patient variables were related to the therapist ratings of the total working alliance score. The same analyses that were presented above were conducted except that the outcome examined was the therapist ratings of therapeutic alliance. In the first step of the analysis, the therapist rated WAI total score was regressed on patient socio-demographic variables and no variables significantly predicted alliance, F (5, 412) = 1.11, p =.35, R² = .01, RSME = 1.001. When therapist ratings were regressed on extra-intestinal medical comorbidities again no variables were significant, F (6, 408) = 0.72, p =.63, R² = .01, RSME = 0.940. No patient social variables were significantly related to therapist ratings of alliance, F (3, 426) = 0.93, p = .43, R² = .01, RSME = 0.999, nor were patient clinical factors, F(7, 420) = 0.94, p =.48, R² = .02, RSME = 0.998. No patient cognitive variables, F (5, 421) = 1.25, p =.29, R² = .02, RSME = 0.990, predicted therapist ratings of alliance, however, while motivational variables were not related to therapist ratings, patient expectancy of treatment success was related to therapist ratings of alliance, β = .16 (95% CI = .05/.26), t (417) p =.003, F (4, 417) = 2.91, p =.02, R² = .03, RSME = 0.983.

Discussion

For a disorder like IBS whose pathophysiology and treatment recommendations is unclear, the one thing that most GEs can agree upon is the value of a quality therapeutic alliance and treatment motivation (40) (41). So it is striking that both what makes for a quality alliance has eluded formal investigation. We address this void by undertaking a study of the predictors of therapeutic alliance based on a hierarchical set of predictors ranging from the most distal sociodemographic factors, through clinical, interpersonal, and cognitive factors to the more proximal cognitive factors including treatment expectancy and motivation.

Among demographic variables, females showed higher scores in overall alliance as well as on their perception that the clinician shared their goals (WAI Goals subscale) and were in agreement about the steps taken to help improve their IBS (Tasks subscale). Extraintestinal conditions were unrelated to therapeutic alliance. As predicted, interpersonal support was related to greater overall alliance as well as to each of its components (i.e., emotional bond [Bond subscale], tasks, goals). Severity of IBS symptoms and pain intensity were related to different aspects of the therapeutic alliance with severity of IBS symptoms associated with the emotional bond (e.g., therapist’s warmth, empathy, involvement) with the clinician and pain intensity related to collaboration around goals and agreement about the steps needed to realize goals. Among cognitive variables, fear of GI symptoms (VSI) - one of most robust predictors of IBS patients’ experience (42) -- predicted goal-related alliance and self-efficacy was related to both goals and task-related alliance. Finally, patient qualities that had the strongest relationships with the quality of therapeutic alliance came from treatment expectancy and motivation variables. Motivation to achieve the patient’s most important treatment goal was correlated with level of agreement with goals and emotional bond between patient and provider. Treatment expectancy of success was strongly related to all dimensions of the therapeutic alliance.

As predicted interpersonal factors, particularly levels of social support, had strong and consistent impact on therapeutic alliance. Patients who reported greater levels of social support were more likely to describe alliance in a positive manner. These relationships are not surprising. After all, beyond its clinical decision making function, a clinic visit is fundamentally an interpersonal endeavor. The psychological process that underlies conditions of social support outside of a clinical setting are likely to be played out in a clinic setting. An IBS patient who, for example, reacts defensively when support needs fall short at home or work may respond similarly to a doctor who provides insufficient reassurance or guidance. A GE who provides validation and empathy may fulfill a patient’s broader need for social support and engenders trust in the process. Regardless, these data extend findings from psychosocial literature that highlights the importance of social support in forming a strong therapeutic bond during psychological treatments (e.g. (43) (44)). In research that has assessed its predictive influence on outcomes of psychosocial therapies, change in social support has been found to be a better predictor of symptom relief than change in therapeutic alliance (45). Whether this pattern of data extends to IBS and medical therapies is important as it would suggest that the formation of social networks outside of the clinical encounter is an important treatment goal, both as a way of strengthening therapeutic alliance and resilience to pathogenic stressors (46).

A major finding of this study was the relationship between treatment motivation and the quality of therapeutic alliance at the beginning of treatment. Motivation to control or reduce IBS symptoms was a consistent predictor of different aspects of the working alliance: the tasks around which treatment was structured, the collaborative interaction between patient and provider, and their emotional bond. Whether this pattern of data extends to providers of other orientations (e.g. GEs) and treatment modalities is unknown.

Given the importance of the therapeutic alliance in managing IBS (9) and the observed relationship to motivation, data raise the question (44)of whether the quality of relationship can be improved by increasing motivation and, if so, how? This question first depends on understanding that motivation is not a fixed trait (“she’s just not motivated”) like eye color or race. Rather motivation is a dynamic process that fluctuates over time and can vary from situation to situation. In other words, it is a state and therefore changeable over time. A person may describe herself as motivated at one time and ambivalent about treatment at other times. Treatment motivation fluctuates as a result of internal and external factors including social relationships, job status, finances, family and friends, a sense that s/he has been heard, the belief that the message will be helpful, level of importance attached to the prescribed behavior, and confidence that it can be changed.

The observed relation between motivation and alliance suggest that a specific therapeutic approach called motivational interviewing (MI) may help GEs manage refractory IBS patients. MI is based on the premise that patients are best able to adopt behavioral changes for managing the day-to-day demands of chronic illness like IBS when the motivation to change come from within themselves rather than being imposed on them by others. Originally developed as an intervention for addictive behaviors (47), MI has been effectively extended to support a variety of non-psychiatric health problems including weight loss, medication adherence, sexual health, diabetes, and Inflammatory Bowel Disease (IBD) (48). It has been argued that for patients with IBD, MI can be implemented when providers “want to improve relationships with patients and be perceived as more empathic and skilled at communication” (49). As such, it is clinically relevant to the GE whose most challenging patients often struggle to manage the day-to-day burden of chronic GI disease influenced by actionable behavioral factors (e.g., IBS, Crohn’s Disease, etc.) including medication compliance, dietary changes, increasing physical activity, and other behavior changes critical to optimal disease management.

MI is a collaborative, goal-oriented approach based on the science of health behavior change focusing on the skills patients and providers need to elicit change. MI is designed to produce rapid, internally motivated health behavior change by helping patients confront and resolve their ambivalence about changing, identify their reasons for changing, and providing structure for making changes. Key elements and assumptions of motivational interviewing include the following:

Establish Empathy.

Creating a treatment environment that is collaborative, empathic, and non-judgmental creates the foundation for behavior change. Shifting from a hierarchical to collaborative style allows patients to be open to considering alternatives. An important part of creating this environment is utilizing communication skills and providing empathy for the patient’s current struggles with their health. A full delineation is beyond the scope of this paper and the reader is encouraged to review the work of Drossman (50) who has detailed key communication skills (e.g., reflective listening, open-ended questions, empathic responding) also featured in a successful MI approach.

Normalize Resistance.

MI approaches resistance, or heightened ambivalence, as a natural part of therapeutic process that often unfolds unevenly. A patient can recognize the importance of a health behavior change and still feel unsure about making that change. An everyday IBS example includes a patient’s sincere desire to reduce symptoms but reluctance to eliminate favorite, symptom-triggering foods. Dismissing the patient as “just not motivated” does not eliminate the problem. By the same token, resistance that arises from the emotional unpleasantness (fear, disgust) to the prospect of undergoing a colonoscopy does not diminish its medical necessity; however, those feelings need to be addressed to engage the patient in so that s/he takes advantage of preventative care. In these cases, resistance can be a sign of the patient’s internal struggle between where they are now and where they (and the provider) would like them to be.

Addressing Resistance.

Physicians are typically trained to adopt a directive interactional style that emphasizes the value of persuasive communication through repeating information or highlighting the importance of making a lifestyle change (e.g., to reduce IBS symptoms you need to manage stress better, adhere to the FODMAP diet, and take medication consistently or to reduce your A1c you need to remove refined carbohydrates from your diet). This approach assumes that the sticking point in adopting health behavior is providing something the patient does not have (e.g., accurate information or knowledge, etc.). This approach has its place, but is not the solution to resistance which requires mobilizing change resources the patient already possesses but has not accessed. The GE who adopts an MI stance recognizes that resistance is a signal to respond differently and should not directly be opposed unless the clinician’s goal is to increase resistance. Rather, the MI clinician values the importance of understanding, exploration, and patience as a tool for helping patients reach the conclusion that behavior change is the best way to achieve their goals.

Develop Discrepancy.

It is important for the GE to understand that ambivalence is difficult to resolve because it is rooted in the patient’s internal tug of war between the costs and benefits about changing a behavior. Through MI techniques such as developing discrepancy, the GE can tip the balance in favor of behavior change that supports positive outcomes. The key to developing discrepancy is trusting and supporting the patient in doing his/her own discovery around the importance of adopting a target behavior rather than pointing out and advising him/her of what could seem important or meaningful from the clinician’s perspective. Health behavior change is motivated when the patient perceives discrepancy between present behavior and important personal goals or beliefs because it incentivizes the patient to take action to reduce the dissonance between a status quo behavior that is not working and an alternative one that supports a health-related target goal (51). A clinician can develop discrepancy by engaging the patient in a brief discussion about the reasons for and against changing in a non-judgmental way that helps to uncover organically the discrepancy between the patient’s current behavior and a target behavior. “So part of you wants to make behavior changes to improve your IBS symptoms, and another part of you feels like you are too busy and it will be one more thing that you can’t have or do because of IBS.”

Rolling with Resistance/Avoiding Argumentation.

When exploring ambivalence and developing discrepancies, it is best to avoid argumentation. Persuasion, threat, or confrontation are not effective methods as they all increase resistance by encouraging defensive responses in the very person (patient) who is most responsible for finding answers and solutions to her health problems. An alternative approach is to “roll with resistance” by expressing empathy about the difficulties of making changes and encouraging change talk. For example, “Perhaps XYZ is so important to you that you won’t give it up, no matter what the cost,” or “How will you know you do need to make a change? What will you notice happening in your life?” By rolling with resistance rather than attacking it, the provider is able to avoid a negative interaction or “wrestling” with the patient’s viewpoint, creating further discussion that feels more like a “dance” than a struggle.

Support Self-Efficacy.

Facilitating successful behavior change depends on patients having sufficient confidence (i.e., self-efficacy) to make a specific change. A person’s belief in their ability to manage a chronic condition or change a health habit – their self-efficacy -- determines whether a behavior is initiated, the amount of effort exerted in goal attainment, and the persistence experienced in the face of setbacks (52). Patients are unlikely to make health changes unless there is hope for success. Self-efficacy is one of the most important predictors of health outcomes whether it is smoking cessation (53, 54), diabetes management (55, 56), or treatment compliance (57) (58). The strength of self-efficacy beliefs have such a robust influence on health outcomes that it has been found to be a better predictor than actual health behavior changes that physicians prescribe (59).

A unique aspect of this study is that it links patients’ confidence in their ability to manage IBS (IBS self-efficacy) to the quality of the therapeutic alliance. Patients with stronger self-efficacy for managing IBS described their relationship with the provider more positively. Individuals exhibit very different levels of self-efficacy to changing a target health behavior. Some are confident, others pessimistic, while others fall somewhere in between. Still others find their self-efficacy fluctuates from one extreme to another. Because confidence can be undermined by past failures and disappointments, providers can foster self-efficacy by inquiring about successful behavior change from the patient’s past, and highlighting personal strengths that are transferable to achieving a future health goal. For example, if a patient successfully quit smoking in the past, highlighting their persistence and determination, and asking them to identify what worked in that situation may induce higher self-efficacy. Another strategy is to ask the patient to identify a person who has been able to accomplish a difficult behavior change to draw inspiration that drives behavioral persistence around a goal. Provider’s expressing confidence in the patient’s abilities to change also can have a positive impact on self-efficacy and the likelihood that s/he will adopt behavior change. For example, saying to a patient “I know that this change will be challenging, but I wouldn’t ask something of you I didn’t think you could accomplish” can facilitate a cognitive shift that allows the patient to begin to believe that change is possible and can even become a self-fulfilling prophecy. Patients give voice to the principle of self-efficacy when they say, “I didn’t think I could be successful at first but I knew my GE thought I could and that was what got me through until I began to believe in myself again” and foster the adoption of health behavior changes.

Like all providers, GEs understandably prefer to treat a consistently motivated patient, but MI recognizes that patients present for treatment oftentimes feeling ambivalent. Ambivalence and/or low motivation does not mean that the patient is uncooperative, lazy, comfortable in the sick role, mentally ill, or seeking secondary gain. Instead, it typically means that the patient is struggling to make sense of change under what for them is difficult (threatening) circumstances. Shifting to a more benign perspective provides a constructive “frame” that reduces conflict, enhances empathy, and makes mutually beneficial behavior change more likely. Utilizing the key elements and assumptions listed above, the GE is better positioned to direct the patient to adopt the desired behavior change whether it is scheduling a feared but necessary diagnostic procedure, dietary changes, medication compliance, increasing physical activity, or other lifestyle changes that are the cornerstone of disease management. By seeing even the most challenging patients in a compassionate and dignified light, the GE affords her/himself the clinical “legroom” to mobilize their patient change resources rather than assuming the burden of behavior change which is fundamentally the patient’s responsibility.

In addition to motivation, treatment expectancy emerged as a consistent predictor of the total alliance score and the three subscales for both clinicians and patients. Patents who reported stronger expectations of improvement were more likely to describe alliance in positive terms. GEs can enhance positive expectancy by discussing with patients their expectations about the treatment process early on in treatment. Patients should be encouraged to state their specific expectations, treatment goals, and concerns (60). Sharing concerns other patients have stated can help patients feel more comfortable expressing their own concerns, and encouraging the patient to identify a few specific and measurable goals for treatment fosters positive treatment expectancy(61). Additionally, a positive treatment expectancy is initiated by the GE through presentation of a strong rationale for prescriptive treatment (62). This strong rationale may be established by offering a non-technical understanding of the research findings of the prescriptive treatment(63). Finally, if treatment is complex and multifaceted from the patient’s perspective, it is helpful to explain to the patient in clear terms without medical jargon why certain treatment is being suggested and the logic behind this treatment plan(64).

There are several limitations inherent in this study that bear on interpretation of its findings. Expectancy was measured using a single item from an established treatment expectancy questionnaire; however, we understand that treatment expectation is multifaceted and includes credibility beliefs as well as prognostic expectations (61). Future research that assesses both aspects of treatment expectations is important for understanding more fully how treatment beliefs impact (and are impacted by) the working alliance between patient and provider. Because of the relative demographic homogeneity of our sample (mostly white, female, chronically ill, educated patients seeking nondrug treatment), our results may not generalize to a broader, more diverse population. To increase the generalizability of our findings, it will be important for follow-up studies to draw from populations other than those used in the present study. Subjects were recruited from a behavioral trial, which may mean our data reflect a selection bias in favor of individuals who volunteer for such research studies.

Additionally, we focused on alliance early in treatment versus alliance that has developed over time. There are several reasons why we focused on early alliance. First, research suggest that quality of the therapeutic alliance is formed early in treatment (65). Second, assessing variables before the full scope of treatment is implemented minimizes a potential confound between alliance and outcome (i.e. the quality of alliance is influenced by amount of symptom change). Third, by focusing on the alliance that develops at the end of the first session before treatment effects materialize, we are better able to generalize our findings to other treatment modalities (e.g., pharmacotherapy) beyond behavioral interventions although the pattern of alliance data for drug treatments is similar to non-drug treatments (66).

An additional source of bias that may arise from structured RCTs is the use of treatment protocols that specify the tasks, goals, and interactional style required to implement behavioral treatments as prescribed. In many respects, the features of a detailed treatment manual maps onto the core dimensions of a quality therapeutic alliance. Whether these features typify the care delivered in non-research settings where treatment encounters are likely less prescriptive is unknown. One might claim that the results apply only to patients who undergo behavioral treatment although the broader health research would argue against this belief. First, alliance, like placebo improvement, is a non-specific treatment variable (67) that underlies treatment independent of its technical components(68, 69). Second, the impact of alliance is seen in medical-oriented treatments for multiple disease states ranging from diabetes, cardiovascular disease, and asthma, (67–71). The breadth of effect across disease states and treatment modalities argues against the claim that the observed relationship between alliance and its predictors is specific to patients who undergo behavioral self-management carried out by a GI psychologist. Whether this pattern of results extends to providers of non-behavioral treatment for IBS is unknown although certainly motivational research developed in the addiction area has been meaningfully extended to other non-behavioral disease states (e.g., diabetes, cardiovascular disease, asthma) for the benefit of medical providers like clinical gastroenterologists who lack graduate training in psychology(72) but seek practical tools for managing motivationally challenged medical patients.. One lesson learned from the broader motivational literature is that the underlying dynamics of motivation cut across different disease states and providers and we look forward moving beyond anectodate and observation to systematic research on the motivational processes that underpin the patient-GE relationship. Additionally, we believe our findings and its conceptual framework are useful to gaining understanding of how the patient –provider relationship operates to facilitate therapeutic change. Further research is needed to determine whether the pattern of data found in this study extend to other treatment modalities (dietary, medical, exercise) and providers of other disciplines which were limited to psychologists in the present study. Until then out data underscore the importance of alliance and the factors that may improve it.

The study was not designed to establish causal relationships between significant predictors and working alliance. At best, our data can be construed as suggestive of a possible causal relationship between studied variables. Future research using longitudinal and experimental designs will better assess any causal effects of social support, motivation, and expectancy on alliance.

Conclusion

In spite of these limitations, this is study contributes to the literature by shedding light on the what makes for a quality working relationship (well defined tasks geared toward a shared goal by patient and provider with an emotional bond) and the factors that contribute to it from the patient perspective. Future research is needed to see whether these dimensions impact health outcome.

Study Highlights.

What is known?

• IBS is a common GI condition whose course is believed to be strongly influenced by the quality of the therapeutic relationship between patient and provider

• Surprisingly little IBS research has been conducted on the nature of the therapeutic alliance, factors that impact it, or strategies to improve it.

What is new here?

• Interpersonal support, severity of IBS symptoms, abdominal pain intensity, treatment expectancy, and treatment motivation impact the quality of therapeutic alliance between IBS patients and their provider.

• GEs can improve the quality of therapeutic alliance by adopting practical strategies that enhance patients’ motivation and treatment expectancies.

Abbreviations

IBS

Irritable Bowel Syndrome

GI

Gastrointestinal

NIH

National Institutes of Health

MINI

MINI International Neuropsychiatric Interview

SOMS-7

Screening for Somatoform Symptoms-7

KESS

The Kraus Emotional Support Scale

IIP

Inventory of Interpersonal Problems

ISEL-12

Interpersonal Support Evaluation List

IBS-SSS

Irritable Bowel Syndrome Symptom Severity Scale

NRS

Numerical Rating Scale

VSI

Visceral Sensitivity Index

CSQ

Coping Strategies Questionnaire

ASI

Anxiety Sensitivity Inventory

WAI

The Working Alliance Inventory

GE

Gastroenterologist

SNS

Sympathetic Nervous System

HPA

Hypothalamic Pituitary Adrenal Axis

MI

Motivational Interviewing

CI

Confidence Interval