Efficacy of a Language-Concordant Health Coaching Intervention for Latinx with Diabetes

Abstract

Objective

To describe the effect of a language-concordant health coaching intervention for Spanish-speaking patients with limited English proficiency (LEP) and uncontrolled Type 2 Diabetes (T2D) on glycemic control, anxiety, depression, and diabetes self-efficacy.

Methods:

64 patients with T2D were randomly assigned to a control or intervention group. Outcomes were assessed by blood work and surveys pre and post intervention.

Results:

The mean sample age was 47.8 years ±11.3 and 81% were female. HbA1c was not significantly different between groups at baseline. The intervention group's HbA1c was significantly lower at times 2 and 3 than in the control arm (p<.01 and p<.001). There were significant reductions in the intervention group's mean HbA1c levels from baseline 10.37 to midpoint 9.20, p < .001; and from baseline 10.42 to study end 8.14, p < .001. Depression and anxiety scores significantly decreased (p<.05 and p<.00), and diabetes self-efficacy significantly increased (p<.001).

Conclusion:

Health coaching led to statistically significant and clinically meaningful decreases in HbA1c, depression, and anxiety scores among LEP Latinx adults with uncontrolled T2D.

Practice Implications:

Heath coaching can be conducted in primary care clinics by nurses or advanced practice nurses. The short-term intervention tested here could be adapted to the clinical setting.

1. Introduction

Over 25 million residents in the United States (US) self-identify as speaking English less than “very well” [1], which is commonly defined as Limited English Proficiency (LEP) [2]. When health care providers cannot communicate in the first or primary language spoken by LEP individuals, effective and equitable care can be compromised, and life-threatening miscommunications can occur [3, 4]. One effective option to close linguistic and cultural gaps in the provision of health services is the use of language-concordant (LC) health care providers – that is, providers who speak the same language as their LEP patients [4, 5]. The use of LC health care providers has been found to predict positive provider and patient attitudes towards each other [6], more counseling on diet and exercise [6–11], increased likelihood of discussing mental health issues [12, 13],, and greater patient adherence to provider recommendations [6]. The need for Spanish-speaking LC health care providers is particularly critical as Spanish is the most common language spoken by LEP individuals in the US [1] and the US Hispanic population is projected to grow substantially in the coming decades [14].

Compared to non-Hispanic white Americans, Hispanic Americans are disproportionately affected by type 2 diabetes (T2D) [15] and experience higher rates of diabetes-related complications [16, 17]. Hispanics with T2D also experience high rates of psychosocial comorbidities such as anxiety [18] and depression [19] which are associated with out-of-range glycemic control, micro and macrovascular disease, and poor quality of life [19, 20]. These adverse outcomes hold true for many patients who are already in the healthcare system and have a regular primary care provider [21]. There is substantive evidence that having access to care alone is not related to better diabetes outcomes [22, 23]. Early stages of T2D can be well controlled by changes to lifestyle factors such as diet, exercise, alcohol use, smoking, and stress management [24–27]. Support for lifestyle changes must be offered in addition to standard treatments such as insulin and/oral medications. Changing one's lifestyle can be a constant, difficult struggle [28], and may be especially so for LEP patients that do not have a LC health care provider for effective collaboration and intensive health education as recommended [16, 29, 30].

The purpose of our study was to describe the effect of a language-concordant (LC) health coaching intervention for Spanish-speaking patients with LEP and uncontrolled T2D using a randomized control, 2-group pre and post-test design. The primary hypothesis was that those patients receiving health coaching would show decreased HBA1c (primary outcome), decreased anxiety and depression, and increased self-efficacy (secondary outcomes) compared to those in the comparison group.

2. Methods

2.1. Trial Design

After approval from the institutional review boards of the University of Arizona and the Ohio State University and review of the ethics committees at the clinical sites, the study was implemented as a randomized controlled trial. The protocol was largely maintained as proposed except for the inclusion criteria for the health coaches. Initially, the health coaches were slated to be graduate nursing students enrolled in the nurse practitioner (MS) or Doctor of Nursing practice (DNP) programs [31, 32] in nursing schools. However, at the time of the study we did not have enough Spanish-speaking nursing students who met the language fluency criteria. Therefore, we expanded the criteria to include senior nursing students (BSN) and nurse practitioners in practice who met the criteria for health coaches.

2.2. Sample Size Calculations and Randomization

The original power analysis, conducted in Pass 13 [33] was based, in part, on the pool available who could be trained to conduct the coaching. We anticipated there to be 15 student coaches per year resulting in a final student sample of 45 coaches. Given the anticipated 2.5 patients per student coach, we anticipated an intervention sample of patients of 37.5 per year for a total of 112 intervention participants and 100 comparison group participants. Given the smaller than expected pool of eligible health coaches, we conducted new power analyses to address our primary aim: to assess the efficacy of the health coaching model at the patient level. The minimum sample size of 41 patients had greater than 80% power to detect a small-medium size effect (dz=.04) in paired t-tests testing 4 outcomes: Glycemic control (HbA1c), Anxiety (GAD-7), Depression (PHQ-9), and Diabetes Self-Efficacy. The statistician generated the randomization sequence using an online random number generator [34]. Participants were randomly assigned to intervention or control group using block randomization (block size =6). Prepared study packets were assigned a study identification number and group number (intervention 1 or control 2) for each site. Study promotoras utilized the next available study packet for the enrollee, thus automatically assigning an ID and group to the participant.

2.3. Participants and Procedures

2.3.1. Eligibility and Recruitment

Patients were identified from the electronic medical records of Federally Qualified Health Centers (FQHC) that served large Spanish-speaking populations located in Columbus, Ohio and Tucson, Arizona. Patient eligibility criteria were being a patient at the recruitment site for at least six months; diagnosed with T2D with current HbA1c >7% (54 mmol/mol); stated preference for communication in Spanish; and age over 21 years (study of adults with T2D). A list of eligible patients was generated at six-month intervals at each clinic. A study promotora (a lay community member who received specialized training to recruit) contacted the patients from the list and explained the study. Interested patients met the promotora at the clinic for more information and signed an informed consent form if recruited. Recruitment occurred between 2015 and 2017. As patients consented to be in the study, they were randomly assigned (using a prepared randomization list) to the intervention or comparison group.

2.3.2. Intervention and Control Group Procedures

Both groups were given standard diabetes care (consultations with a primary care provider, a dietitian, a pharmacist, and a diabetes educator). Standard care was as defined by the clinic. While materials provided were in Spanish, the conversations with providers may or may not have been in Spanish (depending on the availability of Spanish-speaking providers). All participants also underwent a brief physical examination (blood pressure, height, weight, waist circumference), completed a baseline survey (anxiety, depression, diabetes self-efficacy and demographics), and received diabetes management materials in Spanish. All participants also had blood tests (lipids and HbA1c) completed if they did not have an HbA1c in the past 3 months or lipids in the past 6 months in the electronic medical record). Those in the control group were then told their next visit would be in six months. Those in the intervention group received the nurse-led health coaching intervention.

The health coaching intervention was designed to assist patients with diabetes self-management through biweekly health coaching phone calls and 3 in-person visits (baseline, mid-point, and closing). The intervention was delivered over a period of 6 months (one coaching call every two weeks) by a health coach specially trained in methods of effective communication in Spanish using sociolinguistic strategies with chronically ill patients. The curriculum is described in detail elsewhere [35]. Health coaches were trained over the course of 8 weeks using a hybrid in-person and online format. The health coaching curriculum was organized as a 3-credit hour course with a pass/fail grade. Assigned instructors were experienced in coaching techniques and second-language instruction. All students were required to be 'high novice' or higher on Spanish fluency in speaking, reading, and writing via standardized computer testing [36]. Briefly, topics covered diabetes management (treatment and lifestyle); health coaching principles; goal setting; techniques of motivational interviewing; communication based on moving patients along the behavior change continuum; and addressing mental health concerns. Patients and coaches collaborated on discrete communicative tasks such as goal setting and diabetes self-management education topics. Expected learner outcomes and weekly content of the health coaching course are presented in Tables 1 and 2. Self-management for the purposes of our intervention was described as patients implementing behaviors to control modifiable lifestyle factors such as diet, physical activity, medication adherence, and stress management. To monitor intervention fidelity, all coaching sessions were recorded. A Spanish-speaking clinical instructor listened to the recordings and gave constructive feedback to the coach prior to the next coaching session. The study protocol including the intervention was pilot tested successfully and is reported elsewhere [37].

Table 1.

Expected Student Learning Outcomes for Spanish Language Health Coaching Course.

Expected Learning Outcomes	Course Content Modules
1. Apply fundamental elements of clinical conversations.	• Establishing Rapport and Relationship-Building in the Chronic Care Encounter.
2. Apply fundamental elements of clinical conversations.	Establishing Rapport and Relationship Building in the Chronic Care Encounter.
3. Apply fundamental elements of clinical conversations.	• Culturally Appropriate Listening Strategies in the Chronic Care. • Cultural Perceptions of Health Professional Roles within Telephone-based Health Coaching. • Cultural Perceptions of Family Solidarity in the Telephone-based Health Coaching Encounter.
4. Apply fundamental elements of clinical conversations.	• Culturally Appropriate Listening Strategies in the Chronic Care. • Culturally Appropriate Management of Goals, Expectations, Successes and Failures in Telephone-based Health Coaching.
5. Engage in clinical conversations in Spanish demonstrating patient-centered care and communication with instructor assistance, when needed.	• Motivational Interviewing and the Cultural Values of Latino Patients. • Telephone-based Health Coaching and Culturally Appropriate Verbal Expressions of Empathy. • Culturally Appropriate Management of Goals, Expectations, Successes and Failures in Telephone-based Health Coaching.
6. Proficiently conduct a motivational interview with a standardized patient in Spanish.	• Motivational Interviewing and the Cultural Values of Latino Patients

Table 2.

Topical Outline for the Health Coaching Course.

Topics by Week	Topical Outline
Week 1
Foundations of Patient-Centered Interactions	• Communication processes 101 • What does patient-centered communication do for us? • Patient-centered communication (PCC)
Week 2
Listening Strategies	• Self-assessment • About listening and the importance of listening • Becoming a better listener • Active listening
Escuchar y explicar: Los conocimientos básicos de la salud en la consejería de diabetes	Estrategias de comunicación para • Aconsejar sobre los conocimientos básicos de la salud de pacientes con diabetes • Comunicar conceptos claves en el buen manejo de la enfermedad • Verificar el nivel de comprehensión del paciente
[English translation]  Listen and Explain: The Basics of Health in Diabetes Counseling	Communication strategies to • Advise on health basics for patients with diabetes • Convey key concepts for good management of the disease • Check the comprehension level of the patient
Week 3
Motivational Interviewing	• Situating the Transtheoretical Model of Change inside MI • Enacting MI
Quiero … pero no puedo. Auto-eficacia y disposición al cambio en la consejería	Estrategias comunicativas para • Evaluar la auto-eficacia del paciente con diabetes • Promover la auto-eficacia Aconsejar sobre disposición para cambia
[English translation]  I want to... but I can’t. Self-efficacy and willingness to change in counseling	Communication strategies to • Assess the self-efficacy of the patient with diabetes • Promote self-efficacy • Counsel on the willingness to change
Week 4
Mediated Motivational Interviewing	• Mini review of MI • Mediated communication • Telemedicine • Linking mediated communication & MI • Importance of nonverbal communication in the mediated MI
Week 5
Communication Accommodation Theory	• What is Communication Accommodation Theory (CAT)? • Convergence v. divergence • How understanding CAT can help your patients and you
Lenguaje y comunicación en la interacción médico-paciente	• Los primeros tres minutos  o La interrupción  o La interrogación  o El uso de continuadores discursivos • La variatión dialectal  o La variabilidad léxica en el español de América  o Manejando la variabilidad léxica
[English translation]  Language and communication in the doctor-patient interaction	• The first three minutes  o The interruption  o The interrogation  o Using transitional phrases • Dialect variation  o Lexical variation in American Spanish  o Managing lexical variability
Week 6
El eufemismo y el tabú lingüístico	• Describir el tabú lingüístico • Identificar las manifestaciones del tabú lingüístico  o Eufemismo  o Disfemismo  o Abuso verbal (insultos) • Identificar el proceso lingüístico de eufemización • Repasar algunos eufemismo de uso común en el español de América Latina y los Estados Unidos
[English translation]  Euphemism and taboo language	• Describe taboo language • Identify the types of the taboo language  o Euphemism  o Dysphemism  o Verbal abuse (insults) • Identify the linguistic process of euphemizing • Review some euphemisms commonly used in Spanish in Latin America and the United States.

An additional two weeks were included in the course for readings and the final exam

2.4. Measurement

All measures were assessed at baseline, mid and post intervention for the intervention group or at baseline and study end for the control group.

2.4.1. Primary Outcome:

HbA1c, the glycated hemoglobin test, 3-month average level of blood sugar. HbA1c was assessed at baseline and post-intervention for both study groups.

2.4.2. Secondary Outcomes

The Patient Health Questionnaire (PHQ9) measures depression severity with Likert-type items ranging from 0 (not at all) to 4 (everyday), with higher scores indicative of greater depression [38]. A score between 10–14 is moderate depression.

The General Anxiety Disorders (GAD-7) Scale is a 7-item, 4-point Likert-type scale ranging from (0) Not at all to (3) Every day, with higher scores indicating greater functional impairment related to the patienťs experience of anxiety. A score of greater than or equal to 10 indicates moderate anxiety [39].

The Diabetes Self-Efficacy Scale measures self-perceptions held by people with diabetes about their personal competence, power, and resourcefulness for successfully managing their diabetes. The 8 10-point Likert-type items were used in this study, with lower scores indicative of higher self-efficacy [40]. The secondary outcomes were only measured with the intervention group.

2.4.3. Demographic Measures

Sociodemographic variables collected were biological sex, age, race education, relationship status, employment, income, number of children and, of those, how many lived with them.

3. Results

Figure 1 describes the flow of participants through the trial.

Figure 1:

CONSORT Flow Diagram

3.1. Preliminary Analyses

The three mental health measures were strongly positively correlated: PHQ9 and GAD-7, r=.773, p<.01; PHQ and Diabetes Self-Efficacy, r=-.648, p<.01; and GAD-7 and Diabetes Self-Efficacy, r= -.379, p<.05. These correlations indicate that while co-occurring, none were so highly correlated that they could be said to be measuring the same underlying construct [41].

3.2. Sample Demographic and Health Characteristics

The mean age was 47 years old (SD=11.3), 81% were female, and 79% were white Hispanics. A third had earned a High School or Technical School Diploma, 42% were unemployed, and 61.2% earned $20,000 or less annually (See Table 3). All participants were Spanish speakers with LEP.

Table 3.

Patient Demographic Characteristics at Baseline Stratified by Group.

	Control Group n=39		Intervention Group n=25		Chi-Square Test
	N	%	N	%
Biological Sex
Female	21	84	31	80
Male	4	16	8	20	0.203
Race
Caucasian/White	21	84	28	72
African American/Black			2	5
Asian			2	5
Multiracial	4	16	7	4	0.667
Relationship Status
Single	6	24	7	18
Married/Co-habiting	15	60	25	64
Divorced/Widowed	4	16	4	10	0.867
Education
8th grade or less	6	24	15	38
Some high school	4	16	5	13
High School Graduate	7	28	9	23
Technical school Graduate	7	28
Bachelor’s Degree (4 years)			1	2.5	0.289
Employment Status
No	9	36	17	44
Yes (Part Time)	11	44	14	36
Yes (Full Time)	5	20	8	20	0.812
Income
$20,000 or less	18	72	20	51
$20,001 - $40,000	2	0.08	9	23
$40,001 - $60,000			1	2.5
$100,001 and above			1	2.5	0.486
	X	SD	X	SD	t-test
Age	45.67	11.15	50.03	11.42	1.29
Children	2.6	1.42	2.97	1.9	0.729
Children at Home	1.94	1.39	1.52	1.19	1.05

There were no statistically significant differences.

There were no significant differences in any health characteristics (BP, Heart rate, weight, BMI, cholesterol, triglycerides, HDL and LDL) beween the intervention and control groups.The participants were likely to be obese with average BMI of 33, which is congruent with patients with poor controlled T2D (based on HbA1c at baseline). Almost two thirds of all participants reported being on at least one diabetes-related medication and all participants took, on average, 4.3 (SD=3.72) medications daily. More than half reported at least one other diagnosed chronic health condition, such as hypertension and, on average, participants had 3.6 (SD=2.38) comorbities (See Table 4).

Table 4.

Patient Clinical Characteristics at Baseline Stratified by Group.

	Control Group		Intervention Group
	M	SD	M	SD	T Test
Number of Chronic Conditions (1–9)	3.60	2.44	3.59	2.32	0.0168
	N	%	N	%	Chi-Square Test
Chronic Conditions
Hypertension	10	36.0	9	25.6
Hyperlipidemia	6	32.0	6	15.4
Vitamin D Deficiency	6	12.0	4	16.0
Depression	8	16.0	6	15.4
Diabetes Neuropathy	3	8.0	2	8.0	1.46, p>.05
Medications
Metformin	24	61.5	16	64.0
Atorvastatin	21	53.8	16	64.0
Lisinopril	23	58.9	12	48.0
Glyburide	17	43.5	12	48.0	.658, p>.05

3.3. Outcomes of Health Coaching

All analyses were conducted according to the intent-to-treat principle. Given the small sample size, we lacked sufficient power to fit large models comparing the HbA1c trajectories between the groups over time. The intervention group participants experienced statistically significant decreases in the primary outcome, HbA1c levels, over time – from 10.37% at baseline to 8.14% at time 3 (F_(1,31)=104.6, p<.0001). As illustrated in Figure 2 and Table 5, there were statistically significant differences in HbA1c between the intervention group (Mean=9.20, SD=.78, n=39) and the control group (Mean=9.77, SD=.70, n=25; t= 3.04, p<.001, d_Cohen=.769). Similarly, there were statistically significant differences in HbA1c between the intervention group and the control group at time 3 (Mean=8.14, SD=.31, n=33 and Mean=9.52, SD=.65, n=25, t=9.90, p<.001, d_Cohen=2.71)

Figure 2:

HbA1c Levels over Time Stratified by Intervention Group in the RCT

Table 5.

HbA1c levels between and within Intervention and Control Groups.

Independent t-tests
	T1			T2			T3
	Mean	SD	t-test	Mean	SD	t-test	Mean	SD	t-test
Control	10.02	1.46		9.77	0.7		9.52	0.65
Intervention	10.37	0.293	1.18	9.20	0.78	3.04**	8.14	0.31	9.90***
Paired t-tests within Group
	T1		T2		T3
	Mean	SD	Mean	SD	Mean	SD	T1 to T2	T2 to T3	T1 to T3
Control	10.02	1.46	9.77	0.70	9.52	0.65	2.13*	ns	ns
Intervention	10.37	0.293	9.20	0.78	8.14	0.31	3.98***	5.45***	10.69***

ns=not significant

^*p<.05

^**p<.01

^***p<.001

T1=baseline; T2=study midpoint; T3=study end

The paired t-test analyses of the intervention group showed significant reductions in mean HbA1c levels from baseline (Mean=10.37, SD=.29) to midpoint (Mean=9.20, SD=.65, n=37, t=3.98, p < .001, d_Cohen=.58); from midpoint (Mean=9.20, SD=.78) to end (Mean=8.32, SD=.31, n=33, t=5.45, p < .001, d_Cohen =.89); and from baseline (Mean=10.37, SD=.29) to end (Mean=8.32, SD=.31, n=25, t=10.69, p < .001, d_Cohen=2.06). There were no statistically significant differences by biological sex.

As detailed in Table 6, there was a statistically significant decrease in two secondary outcomes: in depression symptoms (PHQ9), from a mean of 6.62 (SD=1.74) at baseline to a mean of 4.17 (SD=1.58) at time 3 (n= 21, t=4.95, p<.001, n=33, d_Cohen=.77); and in anxiety symptoms (GAD-7), from a mean of 6.38 (SD=5.14) at baseline to 4.71 (SD=3.96) at time 3 (n=21, t=2.46, p<.05, n=33,d_Cohen=.348). There was a statistically significant increase in diabetes self-efficacy from 6.31 (SD=2.14) to 7.59 (SD=1.80, n=21, t=3.78, p<.001, d_Cohen=.533). There were no statistically significant differences by biological sex.

Table 6.

Changes in Secondary Outcomes Across Time in the Intervention Group.

	Time 1		Time 2		Time 3		Paired t-tests Change in Mean Score
	Mean	SD	Mean	SD	Mean	SD	T1 to T2	T2 to T3	T1 to T3
Mental Health
PHQ9		1. 7		1. 2		1. 5
Intervention	6.62	4	3.92	3	4.17	8	2.67*	ns	4.95***
GAD7		5. 1		4.0		3. 9
Intervention	6.38	4	4.07	3	4.71	6	ns	ns	2.46*
Diabetes Self-efficacy		2. 1		1.5		1. 8
Intervention	6.31	4	7.33	7	8.23	0	ns	ns	3.78***

ns=not significant

^*p<.05

^**p<.01

^***p<.001

T1=baseline; T2=study midpoint; T3=study end

4. Discussion and Conclusions

4.1. Discussion

The nurse-led health coaching intervention for diabetes self-management for LEP Spanish-speaking patients with T2D was successful in decreasing HbA1c and depression, and in increasing diabetes self-efficacy. The intervention group that received health coaching decreased, on average, almost a full point in HbA1c over the course of the intervention, with no corresponding decreases in control group participants. It is important to note that despite the significant decrease in HbA1c over the course of the study – at time 3 – the intervention group's average of HbA1c, 8.23, still indicated uncontrolled diabetes (despite these patients having access to medical care). All participants were on at least one medication and two-thirds were on at least one diabetes-related medication. In future studies, controlling type of medications and quantity would allow for additional testing of the effect of coaching. HbA1c control is critical to the management of T2D in adults and is impacted by multiple factors of medication, diet and physical activity [42], some of which may have needed more intensive intervention.

Mental health issues also increase the diabetes burden [38]. As such, being able to decrease both HbA1c and depression shows great promise for a health coaching intervention. As indicated in the literature, an integrated language curriculum seemed to increase our success in terms of improved patient outcomes when compared to other Spanish-based diabetes self-management programs [43–45].

Anxiety and depression are often closely related [46, 47] with decreases in one leading to better management of the other. Our study found significant decreases in mean depression scores and anxiety in the intervention group. It is possible that, as HbA1c decreased, participants have felt healthier or happier with their self-management skills, leading to a downtrend in depressive symptoms and anxiety. Future interventions could increase focus on specific mental health counseling.

We anticipated that the intervention's focus on coaching toward goal setting and action steps to manage T2D would increase participant self-efficacy. That assumption held true for the intervention group. Self-efficacy has been shown to be positively influenced by other interventions [48, 49]; however, the key is increased confidence or self-efficacy leads to change in clinical outcomes such as HbA1c. In our study, the primary outcomes of HbA1c, as noted, did change positively. The control group, as expected, did not significantly decrease in HbA1c, depression, or anxiety or increase in self-efficacy. There were no changes in usual care offered to these patients during the study at either clinic. However, it must be noted that usual care for diabetes management may have included written materials in Spanish but the other aspects such as discussion with a provider may have been mediated by an interpreter (depending on the availability of a Spanish-speaking provider). Given that the two groups were comparable at baseline on demographics and HbA1c control, we can attribute these changes in the intervention group primarily to the health coaching. It is heartening that even with the small sample size we saw clinically meaningful significant changes in HbA1c and depression. In retrospect, measuring actual medication adherence could have enhanced the findings reported.

Having opened the health coaching training to nurse practitioners in the clinical setting, we established that such training need not occur only in the graduate school environment. We found it was feasible to educate practicing nurse practitioners was conducted as continuing professional development with CEU credits provided by an accredited college of nursing. This study had several limitations. Due to challenges in recruiting heath coaches, the study was extended over a much longer time period than originally intended. The graduate course on health coaching was first offered as an elective in the nurse practitioner/Doctor of Nursing Practice program at two colleges of nursing. At the first college, there was not a large enough pool of nursing students who met the criteria in terms of Spanish language fluency. At the second college, an additional issue was the changes in curriculum led to electives (such as the health coaching course) being taken off the program plan. In the latter case, if students were interested in becoming health coaches, they had to sign up for an extra course, which cost them money and time (and in some cases affected their financial aid packages). As such, we expanded recruitment to nurse practitioners in practice. The upside of the expanded recruitment was, as noted previously, the ability to train coaches outside a graduate program. The inability to recruit sufficient numbers of NP students who were fluent in Spanish shows the importance of increasing diversity in the nursing profession. Current estimates show nurses from minority backgrounds represent 19.2% of the registered nurse (RN) workforce, comprising 80.8% White/Caucasian; 6.2% African American; 7.5% Asian; 5.3% Hispanic; 0.4% American Indian/Alaskan Native; 0.5 Native Hawaiian/Pacific Islander; 1.7% two or more races; and 2.9% other nurses [50]. Given projections that minority populations could become the majority by 2045 [14], increasing the diversity of the nursing profession with respect to race/ethnicity as well as cultural sensitivity and language ability has been declared a top priority by professional nursing organizations. We posit that access to second-language based health education strategies such as the coaching course described in the current study may be a strategy to increase the ability of nurses and advanced practice nurses to care for an increasing diverse patient population.

Nurse practitioners and nurse practitioner students have busy schedules, often working full time or part time with full-time coursework. This led to limits on the nurses' ability to work with the participants' schedules. For example, one coach was only able to talk for limited hours after standard business hours. Ideally, coaches should be extremely flexible in their availability to participants. Participants themselves had tight schedules, with competing life demands that led to challenges keeping up with the coaching calls. In future iterations of the intervention, we suggest implementing coaching as part of a clinic's diabetes management initiative, allowing providers time as part of their clinic schedule, and thereby reducing the time burden on both providers and patients. We also propose adding incentives, such as tangible rewards from health insurance (perhaps a discount on premiums or medications). While this may be an ambitious suggestion, the continuing problem of diabetes control among those most at risk demands innovative and equitable solutions.

4.2. Conclusion

There are few studies of language concordant coaching for self-management and HbA1c control for LEP persons with diabetes. Our findings show that a nurse-led, Spanish-language program can effectively help patients manage HbA1c, anxiety and depression. Nurse-promotoras/es may enhance recruitment and retention of patients to the program, and the use of phone-based coaching makes it feasible for busy people.

4.3. Practice Implications

There are many factors in the successful self-management of type 2 diabetes, including nutrition, weight, and physical activity. We primarily addressed HbA1c control by focusing strategies on diet, movement and medication adherence and showed that HbA1c control can, at the very least, be impacted positively by a Spanish-language, nurse-led coaching program. It may not be practical for many practices to utilize advanced practice nurses' time in this capacity, but registered nurses could clearly lead such coaching. Nurse practitioners could be certified to train registered nurses and lay promotoras/es in the system. Overall, we present Spanish language health coaching as a feasible and effective strategy to assist patients with limited English proficiency with type 2 diabetes self-management.

Highlights.

• Nurse-led health coaching can help patients manage type 2 diabetes (T2D).

• An intervention was tested for Spanish speakers with limited English proficiency.

• The intervention led to meaningful decreases in average blood sugar levels.

• Depression and anxiety scores were also positively impacted by the intervention.

• Language concordant coaching for managing T2D is a feasible and effective strategy.