Physician Nutrition Advice and Referrals to Registered Dietitians

Abstract

Purpose. This study aims to examine the frequency and content of healthcare providers’ nutrition recommendations and referrals Registered Dietitian Nutritionists (RDN). Methods. Physicians, physician assistants, nurses, and other providers (> 18 years of age) currently practicing primarily in the United States received an email survey that assessed dietary recommendations for diabetes, hypertension, dyslipidemia, overweight/obesity, and general wellbeing, frequency and comfort level of providing nutrition advice, and RDN referrals. Chi-square and Student’s t-tests were used for analysis. Results. 154 physicians (61%), registered nurses/nurse practitioners/physician assistants (19.5%), and other providers (19.5%) were included. Those with nutrition education gave nutrition advice more than those without for some, but not all, conditions (P = .01). The Mediterranean diet was most frequently recommended, except for hypertension. The DASH diet was recommended to 47.7% of patients with hypertension. More providers gave dietary advice than referred to RDNs. Dietary advice was associated with RDN referrals for diabetes (P = .01) and wellbeing (P = .05). Providers with an RDN in their practice provided advice for diabetes more than those without (P = .01). Conclusion. Healthcare providers gave nutrition recommendations consistent with evidence-based guidance. RDN referrals occur in conjunction with dietary recommendations, not as replacement.

Past nutrition education, comfort, and having an RDN in their practice all influence whether healthcare providers give advice and refer to RDNs.

Purpose

Healthcare providers are increasingly being called upon to deliver nutrition advice to their patients,^1-5 and many clinical guidelines^6-10 specifically outline nutrition therapy for the treatment of non-communicable chronic diseases such as cardiovascular disease, hypertension, obesity, and diabetes. Medical doctors, particularly primary care physicians (PCPs), have traditionally tended to be a patient’s first point of contact, making them suitable gatekeepers for a patient’s treatment plan, which can include nutrition counseling. But more recently, an interdisciplinary approach has been recommended, with patients receiving nutrition counseling from more members of their care team than solely the PCP.^1,3,11 Indeed, a wide variety of healthcare providers understand that proper nutrition plays an essential role in disease prevention^5,12,13 and that poor diet quality, increased BMI, and low health-related quality of life have specifically been shown to be associated with increased mortality in adults. ¹⁴

While providers recognize the importance of nutrition in disease progression,^15-17 nutrition education has been historically underrepresented in health professional curricula, ¹⁸ with noted absences at medical schools and residency programs, ¹⁵ nursing schools, ⁵ and physician assistant programs.^2,13 Due to a lack of targeted nutrition education in these educational models and ongoing integration of care between specialties, one behavioral counseling strategy being widely recommended in clinical guidelines^3,6,7,19 is to refer patients to registered dietitian nutritionists (RDN), who are specifically trained to deliver both medical and individual nutrition therapy in acute and chronic care.^8,20

Perhaps due to inadequate education, healthcare providers’ attitudes and confidence in their ability to give nutrition advice are consistently reported to be low,^{12,13,17,21,22} and there are limited data examining the actual nutrition advice currently being given to patients. Of all providers, there is limited data examining physician nutrition practices,^12,17,23-25 and there is even less data examining the practices and advice of healthcare providers in other professions.^13,26,27 Moreover, most studies have examined healthcare providers outside of the United States.^24,28 Of those surveyed, physicians tend to believe nutrition is important, though counseling patients remains a challenge and medical doctors have varying opinions on giving nutrition advice to patients; with some finding it useful ²³ and others determining it as futile and time consuming. ²⁸ Given the current burden of non-communicable disease, it is important to determine whether the nutrition advice patients are receiving from physicians and other healthcare providers is evidence-based and consistent with best practice in the nutrition field.^6,7,9,29

Lastly, in addition to understanding what advice healthcare providers are giving patients, it is important to understand the referral practices to RDNs. This is particularly necessary due to increasing evidence that including an RDN as part of the integrated team is beneficial to high-risk patients,^30,31 and this data is currently lacking. Specifically, it is unclear how often physicians and other healthcare providers refer patients to an RDN in response to common non-communicable disease conditions (i.e., diabetes, hypertension, dyslipidemia, and overweight/obesity) as well as general preventive care and wellness.

This study adds to the current body of research by exploring the frequency and content of nutrition advice made by physicians and other healthcare providers to patients. It also examined how often physicians and other healthcare providers referred patients to RDNs.

Methods

Design and Sample

In this cross-sectional study, an anonymous closed online survey was conducted via Qualtrics (Core XM™), with data automatically captured by the survey program. Unique site visitor data was indicated via IP address to reduce multiple survey attempts and additional survey attempts were not allowed after initial submission. The survey was sent exclusively to members of the Listserv for the Institute of Lifestyle Medicine (ILM). The ILM is a non-profit professional education, research, and advocacy organization with a mission to reduce lifestyle-related deaths and disease through clinician-directed interventions with patients. ³² Subscribers to the Listserv include medical doctors, registered nurses, nurse practitioners, and other healthcare providers primarily based in the United States. Participants were excluded if they were under the age of 18 or not currently practicing in a clinical profession.

Survey

The authors developed the survey to target the following domains: nutrition education, comfort and perceived barriers of providing nutrition counseling, the specific dietary advice given to patients, and referral practices to RDNs. Questions pertaining to attitudes and reasoning for providing nutrition advice were adapted from Pojednic et al. ³³ The survey contained 18 questions on a single page, and was designed to take participants 7–10 minutes to complete. Subjects were able to review and change answers before submission. Demographic questions included age, profession, department/specialty, and number of years in practice. Nutrition education was assessed by whether respondents had received formal nutrition education (yes/no) and type of nutrition education (undergraduate, medical school, continuing education, and post-graduate). Comfort of providing nutrition advice to patients was assessed on a 5-point scale (1 = not comfortable at all to 5 = very comfortable). Perceived barriers to providing nutrition advice were assessed using a check all that apply response, that included time, compensation, patient compliance, patient nutrition knowledge, patient benefit, provider counseling skills, and provider nutrition knowledge. The specific nutrition advice given by providers was assessed for the following conditions: diabetes, hypertension, dyslipidemia, overweight/obesity, and general wellbeing. For each condition, providers were asked how often they prescribed a vegetarian/vegan diet, Mediterranean diet, Dietary Approaches to Stop Hypertension (DASH)/American Heart Association guidelines, low carbohydrate/ketogenic diet, and Dietary Guidelines for Americans. The diet choices were included based on current recommendations by leading health organizations as well as current dietary trends.^{6,7,9,10,29,34-36} Participants were permitted to choose multiple dietary patterns for each condition, as well as “none” or “other.” For each condition, participants were also asked how often they referred to RDNs and/or provide their own advice, using a 5-point scale (1 = never to 5 = always).

Prior to distribution to the study sample, the survey was tested for technical functionality and face validity on a group of ten physicians not on the ILM Listserv. The final survey was emailed to all individuals on the ILM list in March 2020, with 2 additional follow up reminders over the next 2 weeks (3 total email contacts). A reminder was also included in the April ILM Listerv email newsletter with a link to the survey. Participants who completed the survey were invited to enter a drawing to win a $50 gift card. This study was determined to be exempt by the Simmons University Institutional Review Board.

Analysis

Bivariate associations and descriptive statistics were conducted using IBM SPSS Statistics (Version 25.0. Armonk, NY: IBM Corp.) Due to sample size limitations, all practitioners were examined together and descriptive statistics were used to report the type of diet recommended for each condition. For questions with an “other” choice, “other” text responses were coded separately or into one of the original dietary pattern choices if deemed appropriate by at least 2 members of the research team. For example, 4 participants answered “low sodium” or “low sodium whole foods” for diets recommended for hypertension and were included in a new category containing both DASH and low-sodium diets, where responses such as “balanced diet” or “whatever my patient is interested in learning about” were kept in “other.” Categorical data was collapsed into bivariate variables, and chi-square tests were used to determine associations between responses, with nutrition education (yes/no), comfort of providing advice (not comfortable vs comfortable), and how often advice was given to patients (never/rarely/about half the time vs always/usually) as variables of interest. A P-value of < .05 was considered significant.

Results

The survey was sent to a total of 6769 individuals on the ILM Listserv in March and April of 2020. The open rates for the emails containing a link to the survey were 29% and 25%, respectively. A total of 235 participants responded to the online survey, yielding an initial response rate of 3.5%. After excluding 81 incomplete responses that did not answer questions pertaining to our primary outcomes (specific diet recommendations, frequency of advice given, and referral rates to RDNs), 154 responses were included in the analytic sample for a final response rate of 2.3% (Table 1).

Table 1.

Characteristics of the Study Sample from a Survey of Healthcare Professionals on the ILM ^a Listserv (n = 154).

Characteristic	Mean (SD)
Age (y)	52.1 (11.6)
Years in practice (y)	19.2 (11.2)
Characteristic	n (%)
Gender
Female	108 (70.1)
Male	46 (29.9)
Credential
Doctor of medicine or doctor of osteopathic medicine	94 (61.0)
Registered nurse, nurse practitioner or physician assistant	30 (19.5)
Other health professions b	30 (19.5)
Have received formal nutrition education in the past	1. (58.4)
Comfort with providing nutrition advice (n = 152)
Very comfortable	62 (40.8)
Not comfortable/Slightly comfortable/Comfortable	90 (59.2)

^aInstitute of Lifestyle Medicine (ILM).

^b‘Other health professions’ includes health coaches, physical therapists, one LDN, psychologists, exercise physiologists, one nutritionist, one SNAP-ED educator, one doctor of education, one bachelor of medicine, and various master’s degrees.

Our sample was primarily composed of Medical Doctors (MD) and Doctors of Osteopathy (DO) (61%). The average years in practice for all healthcare providers in our sample was 19.2 (SD = 11.2) years; 58.4% reported some formal nutrition education which included undergraduate, graduate, or medical school curriculum, online course, or other nutrition related certifications. All respondent demographic information can be found in Table 1.

A majority of respondents “always or usually” provided nutrition advice for general wellbeing (80.5%) or for patients diagnosed with overweight or obesity (89.5%) (Table 2). Notably, 90% of healthcare providers with nutrition education “always or usually” provided dietary advice for general wellbeing compared to 67.2% of those without nutrition education (P < .01). Nutrition education was not associated with frequency of dietary advice for any of the other conditions. Additionally, there was no association between years in practice and frequency of providing advice.

Table 2.

Frequency with Which Healthcare Professionals on the ILM ^a Listserv Report Referring to Registered Dietitians (RD) and Providing Dietary Advice for Specific Disease/Health Outcomes.

Condition (n = 154)	Always	Usually	About Half the Time	Seldom	Never
Diabetes
Refer to RD	18.2	26.3	23	18.4	13.8
Provide advice	62.3	24	5.2	3.9	4.5
Hypertension
Refer to RD	6.6	14.5	15.1	41.4	22.4
Provide advice	54.5	27.9	7.8	5.8	3.9
High cholesterol
Refer to RD	7.9	18.5	19.2	34.4	19.9
Provide advice	54.9	30.1	4.6	5.2	5.2
Overweight/Obesity
Refer to RD	15.1	27	19.1	23	15.8
Provide advice	63.4	26.1	5.9	1.3	3.3
General well-being
Refer to RD	6.6	8.6	13.8	41.4	29.6
Provide advice	56.5	24	11.7	3.2	4.5

Frequency of response (%).

^aInstitute of Lifestyle Medicine (ILM).

Overall, 40.3% of respondents stated that they were very comfortable providing nutrition recommendations. Those who reported receiving nutrition education were more likely to be “very comfortable” (47.7%) giving nutrition recommendations compared to 31.3% of those who did not receive nutrition education (P = .04). Furthermore, comfort level was a strong predictor of healthcare providers offering nutrition advice to patients. Providers who were “very comfortable” giving nutrition advice were more likely to report they actually provided advice for all conditions compared to those less comfortable; this included diabetes (95.2% vs 80%, respectively; P = .008), hypertension (93.5% vs 74.4%, respectively; P = .002), dyslipidemia (93.5% vs 78.7%, respectively; P = .012), overweight/obesity (98.4% vs 83.1%, respectively; P = .003), and general wellbeing (95.2% vs 70%, respectively; P < .01). Overall, MD/DO participants reported being “very comfortable” providing nutrition advice more often than all other respondents (47.3% vs 30.5%, respectively; P = .04).

When asked to select the specific diet they recommended, the Mediterranean diet was named by the highest percentage of all providers for each condition, regardless of education or comfort level (49.7% for diabetes, 53.2% for dyslipidemia, 54.5% for overweight/obesity, and 60.5% for general wellbeing). The exception was the DASH diet which was more commonly recommended for hypertension (47.7%) (Table 3). “Whole Food Plant Based” was the most frequent description for “other,” even with the inclusion of vegetarian/vegan as a dietary pattern option. Overall, respondents favored diets that limit or eliminate animal products and emphasize fruits and vegetables compared to more recently popularized diets like the ketogenic diet and/or (intermittent) fasting. ³⁶

Table 3.

Frequency with Which Healthcare Professionals on the ILM ^a listserv Recommend Specific Dietary Patterns for Each Disease/Health Outcome.

Condition (n = 154)	Vegetarian/Vegan	Whole Food Plant Based	Mediterranean	DASH/AHA	Dietary Guidelines for Americans	High Fat/Low Carb/Keto	Fasting	Other	None
Diabetes	37.9	9.8	49.7	22.2	15	10.5	1.3	12.5	9.8
Hypertension	37.3	11	45.1	47.7	11.8	3.9	0.6	8.6	7.8
High cholesterol	39.6	10.4	53.2	22.1	13.6	5.8	1.3	11	7.8
Overweight/Obesity	40.9	11	54.5	18.2	13.6	16.2	4.5	17.5	1.9
General wellbeing/Prevention	42.8	9.7	60.5	13.8	15.1	5.9	1.3	11.8	3.9

Dietary pattern (%).

Rows add to greater than 100% due to respondents choosing more than one dietary pattern.

^aInstitute of Lifestyle Medicine (ILM).

All healthcare providers referred patients to RDNs less frequently than they provided their own dietary recommendations. Between 80–89.5% of providers (depending on the condition) reported always or usually giving their own advice, while only 16.2–44.5% reported always or usually referring to an RDN. There was no association found between nutrition education or years in practice and RDN referrals. However, there was an association between a providers’ comfort level of providing dietary advice and the frequency with which they referred to an RDN, specifically for obesity. Of healthcare providers who reported being very comfortable providing nutrition advice, 32.3% referred to RDNs for obesity compared to 17% of those who were not comfortable (P = .042). There was no association between comfort and RDN referrals for any other condition.

Overall, diabetes was the health concern with the greatest percentage of respondents stating that they always/usually referred to RDNs, and there was a significant association between providing dietary advice and RDN referrals for diabetes (Table 2). Specifically, 50% of those who always or usually provided advice for diabetes always or usually referred to an RDN compared to 10% of those who never or usually provided advice (P = .001). A similar association was noted for dietary advice for wellbeing. Of those who always or usually provided nutrition advice for general wellbeing, 17.9% always or usually referred to an RDN compared to 3.4% of those who never or seldom provided advice (P = .05).

Lastly, having an RDN in the place of practice appeared to influence the frequency with which healthcare providers gave dietary advice. A greater percentage of healthcare providers with an RDN working in their practice reported always/usually provided nutrition advice for patients with diabetes when compared to those that do not have an RDN (93.5% vs 79.2%, respectively, P = .01).

Discussion

This is the first study, to the authors’ knowledge, that examines the specific dietary patterns that physicians and other healthcare providers primarily based in the United States offer to their patients for various chronic health conditions. The current study demonstrates that nutrition education is a predictor of providing advice for some, but not all, conditions. The Mediterranean diet was the most frequently recommended diet for all conditions except hypertension, for which DASH was most frequently recommended. Although providers gave advice more often than they referred to RDNs for all conditions, provision of dietary advice was positively associated with RDN referrals for diabetes and wellbeing. Lastly, providers with an RDN in practice reported always/usually providing advice for diabetes more frequently than those without an RDN in practice.

Overall, the most frequently recommended diets, Mediterranean and vegetarian diets, align with those promoted by authorities such as the American Diabetes Association (ADA) and the American Heart Association (AHA) for their respective conditions.^9,35 The ADA currently recommends several dietary patterns including vegan/vegetarian and the Mediterranean diet, which our findings indicate are the most frequently recommended for diabetes. For hypertension, providers are currently meeting the AHA guidelines by recommending the DASH/low-sodium diets most often. Although there is no prior research examining specific dietary recommendations by providers in the United States, our findings align with prior research in which Australian patients reported the rates that their doctors recommended low-sodium diets for hypertension. ³⁷

Of the conditions/disease states queried, overweight/obesity has the highest number of different diets suggested. This may be because there is no single agreed upon dietary pattern for treating obesity. ²⁹ Approaches to treat obesity vary, which may result in providers trying many different dietary patterns with their patients. ²⁹ Prior research has found that frequency of nutrition advice from providers increase with higher patient BMI, and given the lack of a specific diet recommendation for weight loss, this may provide insight on why the types of advice provided for overweight/obese individuals were more varied.^25,38,39

The results also suggest that providers are referring patients to RDNs in addition to giving their own advice, not as a replacement. Although it was hypothesized that those who provided advice themselves would not feel the need to refer, this was not the case. In line with the results of this research, a previous study found that medical students found collaboration with RDNs helpful when providing nutrition care. ⁴⁰ One possible explanation is that providers who care about nutrition enough to give recommendations may see added value in referring to RDNs. It is also possible that providers who know and often refer to RDNs place a higher importance on nutrition and therefore address it more frequently with their patients. Providers with a relationship to an RDN may have also been influenced by the RDN on the importance of diet therapy.

The distinction between those that refer to an RDN and those that do not is important to note and is a key outcome for understanding and addressing gaps in comprehensive patient care. Since those who do not provide nutrition advice are less likely to refer, many patients may not be given any nutrition guidance at all because they are neither counseled on nutrition by the provider nor referred to the nutrition expert (the RDN). While nutrition education has been historically underrepresented in medical education, the risk of diabetes, heart disease, and some forms of cancer are affected by unhealthful diet and, unfortunately, most doctors lack the nutrition knowledge to provide dietary therapy. ⁴¹ Incorporating formal nutrition education as part of the medical school curriculum has the potential to increase providers’ comfort and frequency of providing dietary advice, in addition to increasing effective partnerships with and referrals to dietitians, which could ultimately improve patient outcomes.

There are strengths and limitations to the current study. The primary strength is that the data address a major gap in current research. To our knowledge, this is the only study that currently provides insight into the specific nutrition recommendation practices of physicians and other healthcare providers primarily in the United States. A limitation of the study is the low response rate. This survey was conducted at the beginning of the COVID-19 pandemic (March and April 2020), which may have contributed to a reduction in survey completion. Another potential limitation is the survey population itself. The ILM’s primary mission ³² is “to reduce lifestyle related death and disease,” indicating that individuals on the Institute’s Listserv may be biased to already see the value of providing nutrition advice or referring to an RDN. Therefore, our results may be more representative of providers who already have a positive attitude toward nutrition therapy. Nevertheless, even in our amenable cohort, only 58.4% reported receiving nutrition education, and still only 40.3% felt very comfortable giving nutrition advice. As the majority of physicians and nurses do not receive nutrition education, ¹⁵ these results have implications in the larger field of healthcare professionals and the dietary advice they provide to patients.

Despite a small sample, this study provides important preliminary results about the specific dietary advice healthcare providers give for various health conditions and their referral practices to RDNs. In order to further examine the practices and recommendations of these providers, as well as to determine what factors influence these decisions, more research is needed that is more broadly reflective of the scope of healthcare providers. It may also be beneficial to examine what medications providers prescribe in tandem with diet, which this study did not do.

In conclusion, past nutrition education, comfort, and having an RDN in their practice all influence whether healthcare providers give advice and refer to RDNs. Our findings suggest that many healthcare providers are providing recommendations that align with those from leading health organizations. The frequency with which these recommendations are given is primarily associated with provider comfort and education, and dependent on patient condition. Finally, referrals to RDNs may be occurring in conjunction with provider dietary recommendations, not as a replacement. These findings suggest that in order to increase the strength and frequency of nutrition advice patients receive, it is imperative that providers be given adequate nutrition training and the resources to refer to RDNs as needed.

CME/CE Article Quiz.

American College of Lifestyle Medicine (ACLM) members can earn FREE CME/CE credit by reading this approved CME/CE article and successfully completing the online CME/CE activity. Non-members can earn CME/CE for $40 per article. Visit lifestylemedicine.org to join the ACLM.

Instructions.

1. AJLM CME/CE Articles and Quizzes are offered online only through the American College of Lifestyle Medicine and are accessible at lifestylemedicine.org/store. ACLM Members can enroll in the activity, complete the quiz, and earn this CME/CE for free. Non-members will be charged $40 per article.

2. A Passing score of 80% or higher is required in order to be awarded the CME/CE credit.