Shared Decision Making About Starting anti-TNFs: A Pediatric Perspective

Abstract

Shared decision making (SDM) is central to patient-centered medicine and has the potential to improve outcomes for pediatric patients with inflammatory bowel diseases. We surveyed specialists about their use of SDM in the decision to start a tumor necrosis factor-α inhibitor in pediatric patients. Results were compared between those who reported using SDM and those who did not. Of 209 respondents, 157 (75%) reported using SDM. Physician/practice characteristics were similar between users and non-users. There were no statistically significant differences between groups in the components deemed important to the decision-making process nor the number of barriers or facilitators to SDM. Exploratory analyses suggested that physicians using SDM were more accepting of adolescent involvement in the decision-making process. Our results question the effectiveness of using reported barriers and facilitators to guide interventions to improve use of SDM, and suggest further work is needed to understand the adolescent role in decision-making.

Introduction

Shared decision making (SDM) is a process in which patients, caregivers, and clinicians make medical decisions together, supported by the best available evidence and aligned with patient values, preferences, and treatment goals.^1–4 Its use is associated with decreased decisional conflict, and increased patient and caregiver satisfaction, disease-related knowledge, and medication adherence.^5,6 Prior work demonstrates that SDM is not often being employed in the clinical setting, and that SDM occurs significantly less than patients and physicians report.^7–10 Patient, family, physician, and system variables may affect the incorporation of SDM into clinical practice.^11–13

SDM is particularly important in decisions with challenging trade-offs, such as starting a tumor necrosis factor-α inhibitor (anti-TNF) in pediatric patients. Anti-TNFs improve clinical outcomes and quality of life in many patients with chronic inflammatory conditions, but also have the potential for serious adverse effects, making this decision a particularly difficult one for families.^14–16 Decisions like this, where consequences may be significant, and patient or family preference may affect the decision are ideal for SDM.¹⁷

We assessed potential differences between physicians who self-reported using SDM in the anti-TNF decision making process (SDM physicians) and those who did not (non-SDM physicians). We hypothesized that SDM physicians would place more importance on decision-making components that are part of well-accepted SDM models^1–4 than non-SDM physicians. We further hypothesized that the SDM group would report fewer barriers and more facilitators to the SDM process. Having a better understanding of the perspectives of physicians who report using different decision-making styles will allow us to target and test interventions to increase SDM in pediatric chronic disease care.

Methods

Study Population:

A national sample of gastroenterologists and rheumatologists who care for pediatric patients with inflammatory bowel diseases (IBD) or juvenile idiopathic arthritis (JIA) and had prescribed an anti-TNF in the past year were surveyed about their use of SDM, key elements of the decision-making process, and barriers and facilitators to SDM. A detailed methods description can be found in the original publication.¹⁸

Survey Contents:

The survey assessed use of SDM through a single item which provided examples of a range of decision-making styles (see Table, Supplemental Digital Content 1, for treatment decision approach examples) and requested participants choose the example that best matched their style. Participants were then asked to rate the importance of different components of the decision-making process with regards to starting anti-TNFs. Finally, they were asked how difficult or helpful particular barriers and facilitators were to the SDM process with the following question: “To what extent do you perceive the following as difficulties/helpful during the anti-TNF treatment decision-making process?” (See Table, Supplemental Digital Content 2, for detailed individual survey items.) The survey separately addressed participants’ opinions in regards to SDM with parents (defined as parents or primary caregivers) and with adolescents (over age 11 years). The importance of decision-making components were rated on a 5-point Likert scale (from “not” to “extremely”) and barriers and facilitators were rated as difficult or helpful, respectively, on a 4-point scale (from “not at all” to “a great deal”). All responses were dichotomized for analysis (i.e. important or not, difficult or not, helpful or not).

Data Analysis:

Categorical variables were expressed as counts and percentages. Continuous variables were expressed as means and standard deviations (normal distribution) or medians with interquartile ranges (non-normal distribution). Chi-square and Fisher’s exact tests were used to compare physician characteristics across groups.

Physicians were categorized based on their response to the question, “When making decisions about treatment with anti-TNFs, my approach is usually most like…” Those who identified with a vignette consistent with SDM (Example 4) were identified as SDM physicians. Those who identified with any other decision-making style (Examples 1–3) were considered non-SDM physicians. See Table, Supplemental Digital Content 1, for treatment decision approach examples.

To summarize the importance assigned to components of the SDM process, we identified five components with parents and four with adolescent patients that were consistent with published models of SDM.^1–4 For parents these included (1) the doctor gives information to the parent; (2) the parent gives information about what is important to them to the doctor; (3) the doctor discusses pros and cons of treatment with parent; (4) the doctor gives a treatment recommendation to the parent; and (5) the doctor and family agree on treatment. For adolescents these included items one through four but with the word “patient” replacing the word “parent.” For each participant, we counted how many of the components were identified as important (“very” or “extremely” important) on the five-point Likert scale.

To limit the number of statistical tests, we summarized barriers and facilitators to SDM by adding up the total number of items identified as difficult or helpful to SDM. In other words, for each participant, we summed how many of the 19 and 14 barriers to SDM with parents and adolescents, respectively, were rated as “somewhat” or “a great deal” difficult, and how many of the seven and six facilitators to SDM with parents and adolescents, respectively, were identified as “somewhat” or “a great deal” helpful. We tested for differences in these totals between SDM and non-SDM groups using Mann-Whitney tests. We then conducted further descriptive analyses to compare whether SDM and non-SDM physicians diverged in the importance assigned to components of the decision process with parents compared to adolescents. No statistical tests were applied to this post-hoc analysis.

Regulatory Approval:

The original study was approved by the Cincinnati Children’s Hospital Medical Center Institutional Review Board.

Results

Of 209 respondents (66% response rate), 157 (75%) reported SDM as their typical approach to decisions about starting anti-TNFs. SDM and non-SDM physicians only differed by age, with SDM physicians being slightly younger (50 vs. 54 years old) (Table 1).

Table 1.

Participant characteristics by decision-making group.

Characteristic	SDM physicians (n=157)	Non-SDM physicians (n=52)	p value*
Age in years, mean (SD)	50.3 (11.2)	54 (11.5)	0.048
Sex, n (%)
Male	86 (55.5)	36 (69.2)
Female	69 (44.5)	16 (30.8)	0.08
Specialty, n (%)
Gastroenterology	79 (50.3)	26 (50)	0.97
Rheumatology	78 (49.7)	26 (50)
Fellowship completed, n (%)
Yes	147 (96.7)	51 (98.1)	1
No	5 (3.3)	1 (1.9)
Practice setting, n (%)
Private	6 (3.9)	1 (1.9)	0.53
Affiliated with academic center or children’s hospital	144 (95.4)	50 (96.2)
Other	1 (0.7)	1 (1.9)
Part time status, n (%)
Yes	14 (9.1)	6 (11.5)	0.6
No	140 (90.9)	46 (88.5)
Percent of primary patients with JIA/IBD, n (%)
<25	62 (40.3)	20 (38.5)	0.63
25–50	54 (35.1)	17 (32.7)
51–75	30 (19.5)	14 (26.9)
>75	8 (5.2)	1 (1.9)

SDM = shared decision making.

^*p <0.05 considered statistically significant.

Of the 5 items consistent with SDM with parents, most physicians in both groups rated them all as important. SDM physicians identified more of the 4 items consistent with SDM with adolescent patients as important compared to non-SDM physicians, though the difference was not statistically significant. SDM physicians found fewer barriers to SDM with parents and adolescents to be difficult compared to non-SDM physicians. Both groups reported the same number of facilitators to SDM with parents and adolescents as helpful. These differences were not statistically significant (Table 2).

Table 2.

Counts of important components to SDM (based on theoretical models of SDM), barriers, and facilitators to SDM by decision-making group.

	Number of components	SDM group median (IQR)	Non-SDM group median (IQR)	p value*
Important decision components, parent	5	5 (4,5)	5 (4,5)	0.83
Important decision components, adolescents	4	4 (3,4)	3 (2,4)	0.09
Number of barriers, parents	19	7 (4,11)	9 (4,12)	0.18
Number of barriers, adolescent	14	5 (2,9)	6.5 (3,9)	0.16
Number of facilitators, parents	7	6 (5,7)	6 (5,6)	0.25
Number of facilitators, adolescent	6	5 (4,6)	5 (4,6)	0.4

SDM = shared decision making. IQR = interquartile range.

^*p <0.05 considered statistically significant.

Finally, we compared whether SDM and non-SDM physicians assigned different degrees of importance to aspects of the shared decision making process with parents versus adolescents. Eighty-three percent of SDM physicians thought it was important for both parents and patients to give information to the doctor about what was important to them compared to 58% of non-SDM physicians. Similarly, SDM physicians were more likely to say it was important for the doctor to discuss pros and cons with the parent and the patient (90.8%) compared to the non-SDM group (76.9%). There was no difference between the SDM and non-SDM groups regarding the importance of the doctor sharing information with parents/patients; providing a treatment recommendation to the parents/patient; and insisting parents/patients accept treatment recommendations. No statistical tests were completed on these exploratory data.

Discussion

This is the first study in pediatrics to explore the differences between providers who self-report using SDM and those who report using other decision-making approaches. By better understanding these two provider groups, we hoped to identify targets for interventions to increase the uptake of SDM, potentially improving both decision- and healthcare-related outcomes. While self-identified SDM physicians reported fewer barriers to SDM compared to the non-SDM group, no statistically significant differences were found. Exploratory analyses suggested that SDM physicians seemed more accepting than others of adolescent involvement in the SDM process.

Our findings raise the question, why do physicians look so similar in terms of personal characteristics; opinions regarding factors important to decision-making; and perceived barriers and facilitators; yet report different decision-making styles when it comes to starting anti-TNFs? It is possible that physician self-report does not accurately reflect actual behavior, but rather social desirability bias.¹⁹ Prior work has shown that despite high levels of SDM reported by both physicians and patients, rates of SDM when assessed by audio and/or video-recording are much lower.^7–9 Thus, our participants may have over-reported their use of SDM, making the SDM and non-SDM groups appear more similar and biasing our results in a conservative direction.

Another possibility is that there are differences between the SDM and non-SDM groups that we did not measure. For example, a study of gastroenterologists caring for adult patients with IBD reported a lack of confidence in the data supporting SDM and its efficacy as a significant barrier.¹³ Other studies have reported increased physician motivation to use SDM and a belief that SDM would improve patient outcomes as facilitators.¹²

Finally, it is known that physician attitudes and behaviors regarding SDM vary depending on patient characteristics and clinical scenario.¹² This study partially eliminates this variability by focusing on a specific decision. However, in order to minimize participant burden, the survey does not deeply probe parent and adolescent socioeconomic, race, or behavioral health challenges that may affect the decision making style chosen within a given clinical encounter.

In our exploratory analyses we dove deeper into physician attitudes toward SDM with parents versus with adolescents. There was a trend toward physicians who used SDM reporting that adolescent involvement in decision making was more important than non-SDM physicians. If these findings were found to be significant, they could suggest the need for increased provider awareness regarding the importance of involving adolescents in their medical decisions as well as training on how to appropriately involve adolescents in SDM.

A limitation of this study is that we do not have any information on non-respondents. Additionally, our sample likely over-represents academic clinicians, but we are unaware of any systematic differences in attitudes toward SDM or decision-making behaviors between academic and community clinicians. Future work might survey members of broader pediatric gastroenterology and rheumatology societies.

Conclusions

SDM has the potential to improve a variety of decision- and health-related outcomes.^5,6 At its core, it addresses essential components of ethical medical care including the right of patients and families to understand their options and make an informed choice.²⁰ In this study, we aimed to better understand differences between those who report using SDM and those who do not in the often challenging decision to initiate an anti-TNF in pediatric patients with IBD and JIA, but found few significant differences between groups.

Our results suggest that interventions primarily focused on barriers and facilitators to SDM may not lead to practice change. Future work should be directed toward repeating the study with methods that allow direct observation of the use of SDM, correlating physician-reported practices and actual behaviors, exploring physician attitudes toward the efficacy of SDM, and querying physician openness to adolescent involvement in decision-making.

Supplementary Material

Supplemental Material 1

Supplemental Digita Content, Table 1, Treatment decision approach examples

Supplemental Material 2

Supplemental Digital Content, Table 2, Detailed survey items

What is known?

• Decisions with challenging trade-offs, like the decision to start an anti-TNF, are common in pediatric gastroenterology and rheumatology.

• Shared decision making (SDM) improves a variety of decision- and health-related outcomes.

• There is a paucity of data regarding the use of SDM in pediatric chronic disease care.

What is new?

• Pediatric physicians who reported using SDM versus those who did not are not significantly different in terms of personal characteristics, opinions on what is important in the decision making process, or identified numbers of barriers or facilitators to SDM.

• Using physician-reported barriers and facilitators to SDM may not be helpful in the design of interventions to increase its implementation in pediatric chronic disease care.