Perceptions of Shared Decision Making and Decision Aids Among Rural Primary Care Clinicians

Abstract

Background

Shared Decision Making (SDM) and Decision Aids (DAs) increase patients’ involvement in healthcare decisions and enhance satisfaction with their choices. Studies of SDM and DAs have primarily occurred in academic centers and large health systems, but most primary care is delivered in smaller practices and over 20% of Americans live in rural areas where poverty, disease prevalence and limited access to care may increase the need for SDM and DAs.

Objective

To explore perceptions and practices of rural primary care clinicians regarding SDM and DAs.

Design

Cross sectional survey.

Setting and Participants

Primary care clinicians affiliated with the Oregon Rural Practice-based Research Network (ORPRN).

Results

Surveys were returned by 181 of 231 eligible participants (78%), 174 could be analyzed. Two-thirds of participants were physicians, 84% practiced family medicine, and 55% were male. Sixty five percent of respondents were unfamiliar with the term “SDM”, but following definition, 97% reported they found the approach useful for conditions with multiple treatment options. Over 90% of clinicians perceived helping patients make decisions regarding chronic pain and health behavior change as moderate/hard in difficulty. Although 69% of respondents preferred that patients play an equal role in making decisions, they estimate this happens only 35% of the time. Time was reported as the largest barrier to engaging in SDM (63%). Respondents were receptive to using DAs to facilitate SDM in printed (95%) or web-based formats (72%) and topic preference varied by clinician specialty and decision difficulty.

Conclusions

Rural clinicians recognized the value of SDM and were receptive to using DAs in multiple formats. Integration of DAs to facilitate SDM in routine patient care may require addressing practice operation and reimbursement.

Many medical decisions are in “gray” areas, where alternative management strategies exist and scientific evidence does not provide a clear answer about which choice is best. In these circumstances, reasonable people make different choices based on individual values and preferences.¹ Patients and clinicians routinely confront such healthcare choices, including: managing perimenopausal symptoms ², treatment of early prostate cancer, ^{3, 4} and surgical versus medical management of severe, painful knee osteoarthritis. ⁵

Shared Decision Making (SDM) is a process by which patient and clinician work together to reach an informed decision by considering individual patient preferences and values as well as medical evidence. This approach is considered to be a key strategy for delivering healthcare that is patient-centered as well as evidence based. A current challenge is finding ways to facilitate this SDM process in practice environments where clinician time and attention are highly constrained and where information for patients, though abundant, may be of uncertain quality.^{6, 7}

A sizable body of research has established the usefulness and impact of adopting the SDM approach and employing Decision Aids (DAs) to enhance medical decision making. ^8–10 This progress has led to increased interest in exploring SDM, including the use of DAs, at the state and national policy level. These policies aim to raise the standard of informed consent, hopefully improving outcomes and decreasing costs associated with elective medical procedures.^{1, 11}

Implementing DA tools into clinical practices may facilitate widespread dissemination of the SDM approach. ¹² Although three-fourths of primary care practices employ five or fewer physicians ¹³ and 21% of the US population lives in rural and frontier geographic areas ¹⁴ most studies of SDM and DAs have been conducted in academic centers or large health systems in metropolitan areas.⁹ Rural residents might benefit from SDM and DAs because of barriers due to distance and reduced access to care,¹⁵ poverty, lack of insurance coverage,^{16, 17} and disease as well as risk factor prevalence.^18–22

To address this gap in the literature we undertook this survey to explore the perceptions and practices of rural primary care clinicians regarding the use of SDM and DAs in their practices. We hypothesized that clinicians practicing in rural practices would be less aware of and receptive to SDM principles and DAs compared to groups studied previously.

Methods

Setting and Participants

We invited clinicians affiliated with the 49 member clinics of the Oregon Rural Practice-based Research Network (ORPRN) to complete the study. Practices are eligible to join ORPRN if they provide primary care in rural and frontier Oregon communities. ^{23, 24} ORPRN clinicians, two thirds of them physicians and one third physician assistants or nurse practitioners, provide primary care to over 235,000 patients. ORPRN practices range in size from one to 17 clinicians, mainly in family medicine, with a median personnel size of three clinicians supported by eight clinic staff. Sixty percent of ORPRN practices are physician-owned businesses. ORPRN practices reflect the geographic, demographic, and ownership diversity of primary care clinics in rural Oregon.²³

We initially identified 240 primary care clinicians affiliated with these 49 practices. After eliminating clinicians who had retired or were no longer at a member practice, 231 clinicians were eligible.

Survey Design and Administration

The survey consisted of 21 open-ended and fixed-response questions designed to assess five domains, including: 1. Familiarity with SDM; 2. Perceptions of SDM; 3. Assisting patients with medical decisions; 4. Perceptions of DAs; and 5. Demographic and practice characteristics (see Appendix A). Definitions for SDM and DAs were provided prior to question prompts for domains 2 and 4 respectively. This survey was funded by the Foundation for Informed Medical Decision Making (FIMDM) and was completed in preparation for a study of implementing FIMDM DAs in practice. Six questions were modeled after a national survey conducted in 2008 by FIMDM ⁷ and one question was modeled after a FIMDM questionnaire for use before and after viewing DAs. We developed the list of SDM and DA topics based on a current list of DAs available from FIMDM.^a Since ORPRN members include pediatricians, in addition to family medicine and internal medicine clinicians, the research team added four common conditions (i.e., ADD/ADHD, asthma, contraception, and immunizations) which might be useful in pediatric or adolescent care but which are not currently included in the repertoire of FIMDM DAs. Primary care clinician members of the study team reviewed the survey for face validity. The survey was pilot tested with family medicine clinicians from Oregon Health & Science University (OHSU) who practice in settings similar to those of ORPRN members.

We administered a cross-sectional survey using both a web-based version (hosted on Survey Monkey [Ref: www.surveymonkey.com]) and a printed mailed version. Weekly emails describing the survey and providing a link to the online instrument were sent to eligible clinicians during September 2009. The ORPRN Network Director sent personal follow-up invitations to those not responding in October 2009. Field-based ORPRN research assistants contacted office managers at the 49 practices during September and October 2009 and mailed paper surveys when requested. Data collection closed December 7, 2009.

Two response incentives were used. Clinicians completing the survey could elect to receive a $20.00 gift card to their choice of a bookstore, coffee shop or grocery store. Additionally, clinics achieving a response rate of 70% or greater by eligible clinicians received a clinic-wide pizza party sponsored by ORPRN. The OHSU Institutional Review Board approved this study (IRB #5479).

Results

Participants

Participant demographics, which are similar to general ORPRN clinician membership, are detailed in Table 1. Of the eligible clinicians, 181 of 231 (78%) returned the survey and 174 responses were complete and could be analyzed.^b Forty of the 49 clinics (82%) participating in the survey achieved a response rate of 70% or greater by eligible clinicians.

Table 1.

Respondent training, specialty and demographics (N = 174)

	Survey Respondents n (%)a	General ORPRN Membershipb,c,d
Respondent Training
Physician (MD/DO)	116 (67)	94 (63)
Physician Assistant	32 (18)	31 (21)
Nurse Practitioner	26 (15)	25 (17)
Respondent Specialty
Family Medicine	147 (84)	122 (81)
Pediatrics	16 (9)	19 (13)
Internal Medicine	11 (6)	4 (3)
Clinician Age (years)
≤ 35	36 (21)	30 (22)
36–45	46 (27)	36 (26)
46–55	43 (25)	33 (24)
≥ 55	46 (27)	37 (27)
Male Gender	94 (55)	76 (51)

Note: Due to rounding percentage totals may not add to 100.

^aThree respondents did not provide an age or gender on SDM/DA survey

^bNumbers from 2009–2011 ORPRN Member Survey

^cThree clinicians were missing specialty and two listed psychiatry

^d14 clinicians were missing age

Familiarity and Use of Shared Decision Making

When participants were asked about their familiarity with the term “shared decision making” on a four item scale almost two-thirds of respondents “didn’t know much about it” (48%) or had “never heard of it” (17%). Although the majority of respondents were not familiar with the term, after reading a definition nearly all respondents (97%) reported that SDM was “useful” (54%) or “extremely useful” (43%) in conditions with multiple treatment options. As summarized in Figure 1, the majority of respondents (range: 60% – 85%) reported that SDM was very useful for five categories of medical decisions - changing behaviors, taking new medications, cancer screening, surgery and managing chronic conditions. Despite the high importance placed on SDM, in only two of these five decisions (presented in Figure 1) did over half of respondents say that they “always” used SDM (65% for changing lifestyle behaviors and 56% for surgery).

Figure 1. Importance of utilizing a shared decision making process for various types of medical decisions.

Note: N = 173 for “Cancer screening tests.” N = 174 for other medical decisions listed. “Not” includes responses for “Not too Important” and “Not at all Important”.

Perceptions of Decision Responsibility

As presented in Figure 2, 69% of respondents thought that the patient and clinician should play equal roles managing preference sensitive health conditions. No respondents indicated that clinicians alone should make these decisions. However, when asked about who actually makes these decisions in practice, respondents indicated that they believed clinicians often play the larger role with 42% reporting that decisions are made by clinicians alone. One respondent illustrated this point, commenting: “Whenever possible and appropriate I follow this philosophy [of SDM]. However, I find some patients still want their provider to do all the decision making (‘you're the doctor’).” Another wrote, “[SDM] Makes practice more time consuming and tedious, but [it is] extremely important (if only for sake of compliance).” There were no significant associations between clinician age and gender with reports of who should or actually makes medical decisions.

Figure 2. Clinician views (%) about who should make, and who actually does make, clinical decisions.

Note: N = 173 should decide, N = 172 actually decides. Percentages may not add to 100 because of rounding.

Barriers and Facilitators for Use of SDM

In response to five potential barriers to the incorporation of SDM into practice, two thirds of the respondents “agree” or “strongly agree” that there is insufficient time for detailed discussion with patients and that patients have difficulty understanding all that they need to know to make decisions. Fewer than one-fifth of respondents reported that they would prefer that patients rely on the clinician’s own recommendations and one in ten agreed that lack of patient interest in a SDM process or lack of a trusted source of information for patients were barriers.

Helping Patients with Medical Decisions

As detailed in Table 2, respondents most frequently identified chronic pain and health behavior change as areas where helping patients make decisions was “hard.” In contrast, 41% – 52% of the respondents reported that contraception, immunizations, asthma, and osteoarthritis were cases where it was “easy” to help patients make decisions. Many decisions, such as depression, diabetes, low back pain, heart failure, menopause and coronary artery disease, fell into the “moderately” difficult category for the majority of respondents.

Table 2.

Perceived Difficulty of Helping Patients Make Medical Decision, Arranged in Descending Order by “Hard”, n (%), N = 174.

Condition/Topic	Difficulty Rating
	Hard	Moderate	Easy	N/A	Missing
Chronic pain	108 (63)	49 (29)	4 (2)	11 (6)	2
Health behavior change	88 (51)	69 (40)	14 (8)	1 (1)	2
Dementia	82 (47)	56 (32)	10 (6)	25 (15)	1
Weight loss surgery	49 (28)	78 (45)	21 (12)	25 (15)	1
Low back pain	47 (27)	103 (60)	18 (10)	5 (3)	1
ADD/ADHD	44 (25)	86 (50)	27 (16)	16 (9)	1
Prostate cancer screening	29 (17)	79 (46)	47 (27)	17 (10)	2
Prostate Cancer	28 (16)	74 (43)	38 (22)	32 (19)	2
Depression	24 (14)	117 (68)	30 (17)	1 (1)	2
Diabetes	24 (14)	104 (60)	43 (25)	2 (1)	1
Breast cancer	22 (13)	72 (42)	51 (30)	27 (16)	2
Advance directives	21 (12)	82 (47)	57 (32)	13 (8)	1
Benign prostatic hyperplasia	14 (8)	82 (48)	57 (33)	18 (11)	3
Menopause	13 (8)	93 (54)	46 (27)	19 (11)	3
Heart failure	13 (8)	98 (57)	41 (24)	21 (12)	1
Colon cancer screening	11 (6)	80 (47)	66 (38)	15 (9)	2
Immunizations	11 (6)	77 (45)	85 (49)	0 (0)	1
Coronary artery disease	10 (6)	94 (54)	55 (32)	14 (8)	1
Contraception/Birth control	6 (4)	64 (37)	90 (52)	13 (8)	1
Uterine bleeding/fibroids	6 (4)	89 (52)	52 (30)	25 (15)	2
Osteoarthritis	2 (1)	83 (49)	69 (41)	15 (9)	5
Asthma	2 (1)	89 (52)	80 (47)	1 (1)	2

Perception and Use of Decision Aids

Nearly all respondents in our sample (91%) reported that they were currently using DAs. Use of brochures and handouts from outside entities or created internally by clinicians and staff were reported most frequently (85% and 45% respectively) while 33% utilized web-based DAs and 2% used DVD/VHS formats. When respondents were asked about the preferred format for DAs to use before, during, or after the clinician visit over 90% of clinicians indicated they would “likely” or “very likely” use printed brochures or handouts, whether internally developed or from an external source (94% and 95% respectively), 72% would use web based DAs, and 38% reported they were likely to use DVD/VHS formats. As indicated in Figure 3, approximately 90% of respondents would be “much” or “somewhat” more interested in DAs if there were reimbursement for clinician or support staff time (91% and 90% respectively).

Figure 3. Impact of Reimbursement and Access on Clinician Interest in Using Decision Aids in Clinical Practice to Facilitate SDM (%).

Note: N = 171 Access to service and reimbursed for support staff; N = 170 reimbursed for decision aids and reimbursed for clinician time.

Using the same list that respondents had rated for difficulty helping patients make decisions, we asked them to indicate the top five topics for which they thought a DA would be useful for their patients. As summarized in Table 3, across all respondents the top five ranked conditions for which a DA was preferred were chronic pain, health behavior change, advance directives, diabetes and low back pain. Only 2% of clinicians (3 out of 174) indicated that they did not find a condition listed for which DAs would be helpful. Although the sample size is small, pediatrician preferences appeared to differ from other clinicians by rating ADD/ADHD (100%, n = 16), asthma (81%, n = 13) and immunizations (81%, n = 13) as the topics for which a DA would be most helpful.

Table 3.

Clinician Preference for “Top 5 DA” Topics by All Respondents and by Primary Care Specialty, n (%).

Condition/Topic	All respondents (N = 174)	Primary Care Specialty
		Family Medicine (N = 147)	Internal Medicine (N = 11)	Pediatrics (N = 16)
Chronic pain	111 (64)	103 (70)	7 (64)	1 (6)
Health behavior change	107 (62)	90 (61)	7 (64)	10 (63)
Advance directives	78 (45)	72 (49)	5 (46)	1 (6)
Diabetes	78 (45)	68 (46)	5 (46)	5 (31)
Low back pain	69 (40)	64 (44)	3 (27)	2 (13)
ADD/ADHD	67 (39)	49 (33)	2 (18)	16 (100)
Depression	64 (37)	51 (35)	3 (27)	10 (63)
Immunizations	62 (36)	47 (32)	2 (18)	13 (81)
Prostate cancer screening	62 (36)	58 (40)	4 (36)	0 (0)
Dementia	60 (35)	53 (36)	7 (64)	0 (0)
Colon cancer screening	56 (32)	54 (37)	2 (18)	0 (0)
Asthma	53 (31)	38 (26)	2 (18)	13 (81)
Menopause	53 (31)	51 (35)	2 (18)	0 (0)
Contraception/Birth control	51 (29)	43 (29)	1 (9)	7 (44)
Weight loss surgery	49 (28)	44 (30)	4 (36)	1 (6)
Coronary artery disease	41 (24)	39 (27)	2 (18)	0 (0)
Heart failure	38 (22)	36 (25)	2 (18)	0 (0)
Breast cancer	36 (21)	33 (22)	3 (27)	0 (0)
Benign prostatic hyperplasia	31 (18)	29 (20)	2 (18)	0 (0)
Prostate cancer	29 (17)	25 (17)	4 (36)	0 (0)
Osteoarthritis	27 (16)	26 (18)	1 (9)	0 (0)
Uterine bleeding/fibroids	24 (14)	23 (16)	1 (9)	0 (0)
None of these	3 (2)	2 (1)	1 (9)	0 (0)

Discussion

This survey found that while most rural primary care clinicians were unfamiliar with the term “Shared Decision Making,” when presented with a definition the majority reported this approach is “useful” or “extremely useful” when patients face more than one management option. One respondent commented on this discrepancy between familiarity with the terminology and actually understanding and using the approach, stating: “I didn't realize this [SDM] had a name. Involving patients in their healthcare decisions is simply how I was taught to practice.” Clinicians saw SDM as being very important across a range of medical decisions and preferred that patient and clinician make these decisions equally. However, they perceived that in actual practice, clinicians play the greater role.

Respondents were receptive to using DAs to facilitate informed patient choice and they reported interest in having access to DAs in various formats. Although an earlier survey on SDM and DAs by FIMDM had 402 participants, only 12% were from rural areas, ⁷ making our study the largest survey sample of rural U.S. primary clinicians on this topic. Our rural respondents reported higher current use of DAs than clinicians in the national FIMDM survey (91% versus 43%).⁷ Although we provided a definition of DAs prior to this question to ensure respondents had the same construct in mind, this 48% difference may relate to variations in perceptions of what constitutes a “decision aid.” Respondents may have considered educational materials as DAs even if they did not meet all parts of a more formal definition applied among academicians.^{9, 25} Clinicians may value these types of materials as part of a discussion to elicit preferences and engage patients in decision making.

Our sample of rural clinicians was also more likely to say that they would prefer paper based DAs compared to the FIMDM study sample which had equal preference for DVD, online, or print based.⁷ However, when responses on the FIMDM survey were stratified, the authors found that clinicians working in practices that were rural, smaller or located on the US west coast were more likely to prefer print and in-office DAs to online aids.⁷ We speculate that our respondents may prefer printed DAs because of familiarity with this format, perceived simplicity of work-flow integration, concerns regarding administration costs for other DA formats, limited access to internet technologies in the clinic, or assumptions about patient access to technology. The rural-urban “digital divide” has been well described, ^{26, 27} and rural residents are less likely to have broadband internet access and to engage in health-related internet use. ²⁸ Additionally, because patients are likely to view DVDs or online DAs outside of the clinician office, and clinicians have traditionally shared information during an office visit, changes in mindset and work practices will be required to allow patients to review evidence-based options. In combination, these factors may lead to patterns of practice more geared toward the use of paper-based DAs in settings similar to ORPRN practices.

As reported in previous studies, our respondents identified patient characteristics as prominent barriers to engaging in a SDM process. ^{7, 29, 30} Rural residents are less likely to be insured and have a lower average educational attainment than their urban counterparts. Research demonstrates that when discussing prenatal screening, less educated women were less likely to engage in SDM.³¹ In addition, rural residents must often travel considerable distances for specialty care. This may increase the need for information about management options as well as influence decisions. For example, women with stage I or II breast cancer who lived at greater distances from a radiation oncology facility were more likely to undergo a mastectomy rather than more limited surgery plus radiation therapy which would require additional travel burden and the associated costs.³²

Our results are consistent with other research regarding adoption of digital tools in clinical practice. Electronic health record adoption is slowed in practices where the benefits confer on others while the costs are incurred by the practice itself. ³³ Financial incentives are not currently aligned to support SDM and the thoughtful use of DAs for preference sensitive conditions. ³⁴ We expect that until the benefits of SDM to an individual practice or clinician at least meet the cost of adoption, uptake will be sluggish. Physician incentives under the provisions of the 2010 Affordable Care Act to increase engagement in SDM through Accountable Care Organizations may help to alleviate this misalignment.^35–37 While most clinicians in our survey and in the FIMDM survey⁷ reported that potential facilitators for DA use included reimbursement for clinician and staff time spent during a clinical encounter discussing information and for providing the DA to patients, even measures such as this do not ensure full adoption. We found that about 10–20% of clinicians said that reimbursement for time or the DA itself would not make a difference in their interest in using DAs. For some clinicians other factors may mitigate against adoption even if financial incentives become more favorable. Alternatively, some may believe that the use of SDM and DAs are a standard of providing high quality care or make practice easier and that they should not receive special reimbursement for these services.

We expected that clinicians would be more likely to choose DAs for conditions they regarded as difficult, but this was not always the case. Some selected DAs for topics that they had described as “easy” while DAs were not chosen for some topics reported to be “hard.” For example, more than 60% of respondents selected DAs for chronic pain and health behavior change, and these were considered “hard” conditions. However, the next three most commonly preferred DA topics were advance directives (49%), diabetes (46%), and low back pain (44%), which respondents had indicated were of “moderate” difficulty in regard to helping patients make decisions (47%, 60%, and 60% respectively). Although immunizations and asthma were considered “easy” conditions across all respondents, over 80% of pediatric clinicians indicated that DAs for these topics would be helpful.

We speculate that other reasons besides decision difficulty may explain clinicians’ expectations of DA usefulness. DAs may provide a means to convey complex technical material with which a clinician is less familiar (as with surgical procedures), and to present balanced information where treatment options are largely equivalent or supporting evidence is uncertain (as with hormone replacement therapy for immediately postmenopausal women or treatment of early prostate cancer for men). They might also be useful to ‘break the ice’ for topics that often have significant emotional content (such as dementia or chronic pain) or to save time where the amount of information to be conveyed is substantial or frequently encountered in practice (as in immunizations, asthma, or contraception). In sum, the desirability of a DA may relate to the underlying confidence and competence of the clinician dealing with the topic, the strength of evidence about the intervention, the emotional aspects of the decision, and the expected value of a DA in terms of saving clinician time or effort. For these rural clinicians the ease or difficulty of specialty referral and geographic access may also influence the desirability of various DA topics.

This is the first study of which we are aware that has examined the knowledge and preferences of pediatric clinicians for SDM and DAs. Although the number of pediatric clinicians in our study was small, they indicated a strong interest in topics related to child health issues, such as ADHD treatment and immunizations. These tend to be conditions where clinicians must spend considerable time during office visits negotiating care options and which may generate patient or parent concerns.^{38, 39} Pediatric and adult care providers shared interest in DAs for topics such as asthma, diabetes and health behavior change. Future work should engage pediatric and family medicine clinicians to develop DAs for child health-specific topics, as few DAs for children and their families currently exist.

Our study is one of only a few to explore the views of rural clinicians, including non-physician health professionals, on SDM and DAs.³⁰ However, generalizability may be limited by the fairly small sample size and focus on a single state. Although the data suggest interesting questions about potential differences in perceptions of and receptivity toward SDM and DAs based on gender, specialty, training, availability of services, and familiarity with the SDM concept, our study was not designed to address these issues. Additionally, it is possible we did not include potential DA topics that would have been rated more highly by participants. We attempted to mitigate this by utilizing a list of DA topics available from FIMDM as well as adding four topics we thought would be applicable to pediatric clinicians.

This survey measured clinician attitudes and self-reported behaviors but did not measure actual use of SDM or DAs in practice, nor the content or quality of DAs used by ORPRN clinicians, where there might be considerable variation. Our survey also did not evaluate local contextual factors which might influence the use of SDM or DAs in actual practice such as availability of specialty care, travel distance to services, insurance status or other factors that may be especially important in rural areas. These factors are important considerations for future survey research as well as in implementation studies.

Implications

Our results indicate that while most rural primary care clinicians believe that medical decisions should be shared equally with patients, they perceive that the clinician more often makes actual decisions. Our sample of rural clinicians recognized the value of SDM and was receptive to using DAs in multiple formats. Respondents also reported interest in having DAs on various topics, including those that were both “hard” and “easy” to help patients make medical decisions. SDM and use of supportive resources such as DAs have great potential to improve patient care, communication and satisfaction, but only if they become a more integral part of routine care.

Our findings provide a foundation for interventions to increase the use of SDM and DAs in rural practices. Although there is a trend for increasing size of primary care practices and decreased physician ownership,⁴⁰ 88% of visits to US office-based physicians are to practices with nine or fewer physicians.⁴¹ There are few studies regarding interventions that are effective for implementing SDM and the use of DAs in small to medium sized, non-academic clinical practices ⁶ although the majority of primary care clinicians practice in these settings.¹³ We believe further work is needed to inform “Best Practices” for integrating DAs and SDM in non-academic primary care settings.