Acculturation Differences in Communicating Information about Child Mental Health Between Latino Parents and Primary Care Providers

Abstract

Background

Significant Latino-white disparities in youth mental health care access and quality exist yet little is known about Latino parents’ communication with providers about youth mental health and the role of acculturation in influencing this communication.

Methods

We estimated regression models to assess the association between time in the U.S. and the number of psychosocial issues discussed with the medical assistant (MA) and doctor, adjusting for child and parent mental health and sociodemographics. Other proxies of acculturation were also investigated including measures of Spanish and English language proficiency and nativity.

Results

Parent’s length of time in the U.S. was positively associated with their communication of: their child’s psychosocial problems with their child’s MA, stress in their own life with their child’s MA, and their child’s school problems with their child’s doctor. These differences were especially apparent for parents living in the U.S. for greater than ten years. Parent-child language discordance, parent and child nativity were also significantly associated with communication of psychosocial problems.

Discussion

Greater provider and MA awareness of variation in resistance to communicating psychosocial issues could improve communication, and improve the prevention, diagnosis and treatment of youth mental illness.

Background

Latino youth have higher lifetime rates of mood disorder than non-Latino white youth (1) and Latinos arriving in the U.S. as children have greater lifetime risk of psychiatric illness compared to Latinos spending their childhood elsewhere (2). These findings suggest that exposure to U.S. culture during childhood decreases the likelihood that Latino youth adopt the protective values and norms (e.g., familism) common to their first generation Latino parents (3). Improved awareness, prevention, and treatment of mental illness is needed among Latino youth, as is a greater alliance between health care professionals and Latino immigrant parents in the treatment of these youth.

However, Latino youth access mental health care services at approximately half the rate of white youth and these disparities have not improved over time (4). Latino children report fewer lifetime counseling visits than whites (5) as well as less antidepressant use (6) and stimulant use (7, 8). Improving communication about mental health between parents and medical professionals in primary care settings may be a viable pathway to reducing these disparities and lowering the burden of mental illness among Latino youth. Latino children have the highest rates of well-child visits among any racial/ethnic group (9) and the high frequency of recommended pediatrician visits (10) provide numerous opportunities for Latino parents to interact with primary care providers.

Latino Parents’ Willingness to Discuss Mental Illness with Primary Care Providers

More work is needed to understand the trust and expectations that Latino parents have in discussing mental health topics with medical professionals in primary care settings. Recent emphasis on patient-centered care has drawn attention to improving relationships between providers and patients (11) and to increasing engagement and retention in mental health care (12). Patient-centered efforts are likely to be successful among Latino families because of their already higher rates of acceptance and preference for mental health care counseling compared to whites (13), Latinos’ greater preference for individual and group mental health treatment compared to whites (14), and their higher likelihood of having mental illness treated in primary care settings (13, 15).

Primary care is a frequent “gateway” to mental health services for children and their families because primary care providers see a large number of children, often over a long period of time. Unfortunately, some studies suggest that only half of pediatric primary care visits include discussion of the child’s behaviors and mood, and even fewer visits include discussion of family or parent stress (16). Although both providers (17) and parents (18) may not feel that primary care is an appropriate place to discuss such concerns, research shows that parents are more satisfied with primary care visits that include discussion of psychosocial health concerns (19), that primary care providers can provide appropriate treatment and referrals for common pediatric mental health problems (20, 21), and that Latino parents were more willing than non-Latino parents to discuss their child’s mental health with the PCP (18).

Medical assistants (MAs) and other paraprofessionals increasingly play a role in team-based primary care, and in multicultural communities may be able to serve as “cultural bridges” between patients and providers (22, 23). Patients develop therapeutic relationships with MAs (24, 25), which may increase the effectiveness of specific interventions (26) and encourage patients to disclose mental health concerns and engage in care (26). However, we identified only 1 study that addressed attitudes towards differing mental health providers among Latinos, finding that Mexican-Americans much preferred to discuss issues with trained counselors rather than less well-trained paraprofessionals (27).

Acculturation and Latino Parents’ Communication with Medical Professionals

Acculturation, defined as a “culture learning process experienced by individuals who are exposed to a new culture or ethnic group” (28, p.102), is a dynamic process of adaptation involving culture learning and maintenance influenced by individual, group, and environmental factors (29). More than a unidirectional move by the individual from one culture to another, acculturation is a multidimensional and bidirectional construct involving different kinds and degrees of change (or no change) in a wide range of attitudes and behaviors (30).

Complicating the understanding of the role of acculturation is a disconnect between the development of multi-faceted and multidimensional theoretical constructs and the common practice of using proxy measures of acculturation (31, 32). Attempts to operationalize the multidimensional, bidirectional model have led to the development of well-validated scales of acculturation (33–35) that place greater emphasis on a broad range of culture-related attitudes and behaviors such as food and music preferences, religiosity, individualism vs. familism, friendships within/outside ethnic group, and sense of ethnic identity (36, 37). Analysis of measures used in multidimensional models have also increased understanding of which proxy variables provide accurate information in contexts where full acculturation scales are unable to be used. For example, length of residence in host country may be a useful proxy of acculturation in samples of only immigrant Latinos (38, 39), and English language proficiency has been shown to explain much of the variance in mental health among Latinos (40, 41), and is moderately correlated with acculturation scales (32). While these proxy measures do not capture the full range of cultural and socioeconomic changes that occur with greater exposure to U.S. customs and culture (40), a better understanding of how these proxies influence health care is needed as future studies will continue to use these measures given the impracticality of administering full acculturation scales in many intervention studies and community surveys (39).

Acculturation is an important predictor of mental illness (2, 3, 42–44) and we hypothesize that acculturation is also an important predictor of understanding Latino parents’ willingness to discuss mental health problems with primary care providers. The one study identified in this area was a qualitative study that identified that less acculturated Latino parents were not as inclined to interpret problem symptoms indicative of ADHD compared with more acculturated parents (45).

The objective of this paper is to analyze the relationship between factors relevant to acculturation and Latino parents’ communication of mental health issues in a team-based pediatric primary care setting, in order to contribute to the understanding of the underlying mechanisms of mental health care disparities and to examine a potential mechanism of reducing the greater lifetime risk of psychiatric illness faced by Latino youth living in the U.S. We test the hypothesis that increased time in the U.S. and other proxies of acculturation are positively associated with the likelihood of discussing youth mental health issues and the number of psychosocial constructs discussed with paraprofessionals and doctors. Secondarily, in an exploratory analysis, we compare the association of different proxies of acculturation status (parent time in the U.S., measures of Spanish and English language proficiency and nativity for youth and parents) with parent communication of youth mental health issues.

Data

Participants and Data Collection

We conducted secondary analysis of data from a study of mental health-related communication skills training for paraprofessional MAs (46). The training consisted of 3 one-hour meetings with MAs during which MAs participated in group discussions, watched video examples of interactions between MAs and patients, and practiced their skills through role play. The content of the training focused on learning skills to greet patients, orient them to the visit and illicit their health and mental health concerns, probe for more information when necessary, respond to patient concerns with empathy and support, and establish an agenda for the visit with the doctor (46).

The study took place at a federally qualified health center serving primarily Latino adults and youth in Washington, D.C. The previous study recruited 1 pre-training and 2 post-training cohorts of parents of youth ages 18 months to 16 years seeking general pediatric care. All MAs at the centers are native Spanish-speakers and Spanish-English bilingual. None of the pediatricians are native Spanish speakers, but all speak Spanish routinely with patients. Parents were predominantly female (88%), born in El Salvador (61%) and other Latin American countries (33.5%), and had low educational attainment (41% less than high school) (46). Parents and youth living in the U.S. for 6–10 years and greater than 10 years were significantly older than parents and youth living in the U.S. for 0–5 years (Table 1).

Table 1.

Rates of Communication with Medical Assistant (MA) and Doctor, Sociodemographic, and Health Status Variables by Time in U.S.

		# Years in U.S.
		0—5	6—10		>10
Any psychosocial problems discussed with the MA (n = 359) (%)		38.2	29.3		37.6
Count of psychosocial problems discussed with the MA		0.8	0.7		1.1
Did you talk to a MA about:
your child's behaviors (%)		33.8	24.0		31.2
your child's feelings (%)		11.8	11.3		22. 0
how your child is doing in school? (%)		10.3	14.7		20.6
how your child gets along with other children? (%)		11.8	15.3		16.3
any stresses and strains you are feeling yourself? (%)		1.5	5.3		9.9	**
any problems in your family? (%)		5.9	3.3		6.4
Any psychosocial problems discussed with the doctor (n = 358) (%)		68.1	74.7		74.8
Count of psychosocial problems discussed with the doctor		2.1	2.4		2.6
Did you talk to the doctor about:
your child's behaviors (%)		62.3	64.7		66.2
your child's feelings (%)		43.5	49.3		49.6
how your child is doing in school? (%)		29.0	45.3	**	54.7	**
how your child gets along with other children? (%)		33.3	42.7		46.0
any stresses and strains you are feeling yourself? (%)		21.7	20.7		25.2
any problems in your family? (%)		15.9	14.7		16.6
Parent's Age	17—24	34.6	13.0	**	9.2	**
	25—34	51.3	66.1	**	34.6	**
	35+	14.1	21.0		56.2	****
Child's Age	1—5	80.8	72.8		47.1	***
	6—11	14.1	22.2		35.3	****
	12—16	5.1	4.9		17.7	***
Round of interview (%)	1	41.0	32.1		32.0
	2	30.8	34.6		37.3
	3	28.2	33.3		30.7
Parent reported mental health score on GHMI and SDQ	Borderline or high (SDQ >= 12 or GHMI/BHMI >= 3)	37.2	42.6		28.8
PHQ-9 for parent > 5 (%)	Probable depression (PHQ > 5)	37.3	32.5		31.3
Child Health Fair or Poor (%)	Fair or Poor	20.5	17.9		20.9
Difficulty in Paying Bills (%)	Somewhat or Very Difficult	87.2	88.3		86.3
Do you have the money that you need (%)	Not enough	68.0	59.9		64.7

^**p<.05

^***p<.01

^****p<.0017 (Bonferonni-adjusted critical p-value)

Parents were recruited as they arrived for either scheduled or walk-in visits. If they agreed to be in the study and provided written consent, they completed a questionnaire (in English or Spanish) requesting demographic and other information. Following the visit, they were asked to provide ratings of their interactions with MAs and pediatricians. Recruiting procedures and additional details about data collection are described elsewhere (46). The study was approved by the Johns Hopkins School of Public Health Institutional Review Board.

Dependent variables

The parent was asked whether or not she/he had talked with an MA and a doctor about 6 psychosocial issues: (1) child’s behaviors, (2) child’s feelings, (3) how the child is doing in school, (4) how the child gets along with other children, (5) any stresses or strains that the parent is feeling, and (6) any problems in the family. We summarize these responses into 2 main dependent variables: Number of psychosocial issues discussed with the MA (0– 6) (α=.828), and number of psychosocial issues discussed with the doctor (0– 6) (α=.835). These questions were adapted from questions used in evaluations of pediatric consultations and other primary care trainings (47, 48).

Measures of Acculturation

We compared respondents on number of years that the parent respondent had lived in the U.S., both as a continuous measure and as a categorical measure to account for possible non-linearity in the response (<= 5, 6–10, > 10). This measure was chosen because of its common use in the immigrant mental health literature (e.g., 49, 50–52). Other factors related to acculturation were parent-child language discordance (mother speaks Spanish with friends but child speaks English at home), an indicator of whether the mother speaks both Spanish and English with family or friends, and nativity measures (U.S.-born parent, U.S.-born child, and discordance between the parent and child nativity [parent born outside of the U.S. and child born in the U.S.]). Exploratory factor analysis was used to assess the possibility of combining some or all acculturation-related variables into a reduced index. However, eigenvalues were all below 0.8 in these analyses (correlations were all below 0.3), suggesting that these scales contribute little to the explanation of variances in these variables. In another attempt to combine acculturation variables, a naïve method was used to create numerous additive combinations of these variables. However, none of these created scales were found to be reliable with all scales’ Cronbach’s alpha less than 0.6.

Other Independent Variables

To measure whether or not a child had mental health difficulties, two age-appropriate instruments were combined. Parents of children age 3–16 completed the Strengths and Difficulties Questionnaire (SDQ) (53), a measure of mental health difficulties during the past 6 months widely used among Spanish language populations and found to be highly specific (94.6%) and somewhat sensitive (63.3%) to the identification of individuals with a psychiatric diagnosis (54, 55) (α=.84). Parents of children age 18–36 months completed the gender-specific Mental Health Inventory (MHI) to report the child’s emotions and behaviors during the past 2 months. The MHI, created specifically for youth less than 3 years old, is extracted from the Child Behavior Checklist (CBCL) and has a mean test-retest reliability of .87 (56) (α=.48 for females and α=.49 for males). Based on U.S. norms, a child was considered to have mental health difficulties (moderate to high symptoms of mental illness) with SDQ ≥ 12 or MHI≥3 (57, 58). Children’s overall health was measured through parent report of whether the child’s health was excellent, very good, good, fair, or poor (recoded as fair/poor or not fair/poor). Parent mental health was measured using the Patient Health Questionniare-9 (PHQ-9) (α=0.85) (59) which has been validated in Spanish (60, 61).

Income barriers were measured using a yes/no indicator of whether the parent had difficulty paying bills or enough money for needed expenses. Parent’s age (categorized into 17–24, 25–34, 35+), child’s age (1–5, 6–11, 12–16), and round of interview (Round 1—pre-intervention, 2—post-intervention, or 3—follow-up) were also entered into regression models. The latter variable is used to adjust for the influence of the intervention on communication patterns.

Missing data

Of the 393 families reruited, 32 parents did not respond to the 6 “communication with MA” items and 33 parents did not respond to the 6 “communication with doctor” items. Only 2 additional respondents were missing on number of years in the U.S. These individuals were excluded from the analysis for a total of 359 individuals in the communication with MA analyses and 358 individuals in the communication with doctor analyses. Of the remaining sample, 20 individuals did not complete the PHQ-9. We imputed PHQ-9 scores for this group, applying multiple imputation methods using the mi procedure in Stata 12 (62). This technique creates 5 complete datasets, imputing missing values using a chained equations approach, analyzes each dataset, and uses standard rules to combine the estimates and adjust standard errors for uncertainty due to imputation (63, 64).

Analysis

To describe the sample, we first examined whether time in the U.S. was associated with communication with MAs and doctors and independent covariates of interest. Using chi-square tests, we separately compared the group in the U.S. for 6–10 years and >10 years to those in the U.S. 0–5 years on each of the dependent variables and covariates. Multiple comparisons lower the critical significance threshold to a Bonferonni-corrected p<.0017 (65), though significance at the p<.05 and p<.01 levels are highlighted for descriptive purposes.

Multivariate Regression Analysis

To identify the association between continuous time in the U.S. (number of years) and categorical time in the U.S. (0–5, 6–10, >10 years in the U.S.) and two dependent variables of interest (count of psychosocial issues discussed with the MA / with doctor), we estimated Poisson regression models, adjusting for children and parent mental health status, parent and child’s age, interview round, and income barriers. To explore the topics of communication that might be particularly malleable over time in the U.S., we assessed the association between time in the U.S. and discussion of the 6 psychosocial issues with the MA (6 regression models) and doctor (6 regression models), adjusting for parent and child health status and demographics.

Exploratory Comparison of Acculturation Measures

Poisson regression models of the number of psychosocial issues discussed with MAs and doctors were estimated as described above but replacing time in U.S. with each acculturation proxy measure. These coefficients are not directly comparable because they were estimated in separate models, but a qualitative assessment of differences in significance and magnitude provide preliminary evidence of differences across acculturation proxy measures.

Results

Unadjusted Differences by Time in the U.S. on Communication Measures

Table 1 presents results from an unadjusted comparison of our dependent variables, demographics, and health status measures between parents in the U.S. for 0–5 years, 6–10 years, and greater than 10 years. Approximately 1 in 3 parents in the sample discussed any psychosocial problem with an MA. Looking at psychosocial issues independently, parents in the U.S. >10 years were more likely to report discussing their own stresses and strains with an MA than parents in the U.S. for 0–5 years. More than 2 in 3 parents in the sample discussed psychosocial problems with the doctor, nearly twice the percentage of parents that discussed psychosocial problems with the MA. Parents in the U.S. for 6–10 years and greater than 10 years were more likely to discuss how their child was doing in school with the doctor than parents with 0–5 years in the U.S.

Multivariate Regression Results

When adjusted for parent and child age, health status, income, and study intervention, number of years in the U.S. was significantly associated with the number of psychosocial issues discussed with MAs (Table 2). Results were similar in direction but not significant for number of psychosocial issues discussed with doctors (p=0.094). Parents in the U.S. for more than 10 years communicated with MAs on more psychosocial issues (a mean of approximately 0.4 more) than parents in the U.S. for 0–5 years. The difference between parents in the U.S. for more than 10 years compared to parents in the U.S. for 0–5 years in count of psychosocial issues discussed with doctors was similar in direction but not statistically significant (p=0.063).

Table 2.

Poisson Regression Results Regressing Outcome Variables on # of Years in U.S. Conditional on other Parent and Child Characteristics

		Count of Psychosocial Problems Discussed with MA (n=359)			Count of psychosocial problems discussed with doctor (n=358)			Count of Psychosocial Problems Discussed with MA (n=359)			Count of psychosocial problems discussed with doctor (n=358)
		Coeff		SE	Coeff		SE	Coeff		SE	Coeff		SE
# of Years in U.S. (continuous variable)		0.02	**	0.01	0.01	^	0.01
# of Years in U.S. (referent 0–5)	6 —10							0.03		0.17	0.13		0.10
	>10							0.36	**	0.17	0.20	^	0.11
Parent's Age (referent 17–24)	25—34	−0.08		0.17	0.04		0.10	−0.09		0.18	0.01		0.11
	35+	−0.28		0.19	−0.02		0.12	−0.31		0.20	−0.04		0.12
Child Age (referent 1–5)	6—11	0.25	^	0.13	0.11		0.08	0.24	^	0.13	0.11		0.08
	12—16	0.44	**	0.19	0.23	^	0.12	0.42	**	0.19	0.23	^	0.12
Round of Interview (referent 1st)	2	−0.13		0.14	−0.08		0.08	−0.09		0.13	−0.08		0.08
	3	−0.30	**	0.15	0.07		0.09	−0.29	^	0.15	−0.08		0.09
Parent Reported Mental Health Score on GHMI and SDQ		−0.04		0.13	0.02		0.08	−0.04		0.13	0.01		0.08
PHQ for Parent		0.34	**	0.13	0.04		0.08	0.33	**	0.13	0.04		0.08
Child Health Fair or Poor		0.30	**	0.14	−0.01		0.09	0.28	**	0.14	−0.02		0.09
Somewhat or Very Difficult to Pay Bills		−0.50	**	0.16	−0.03		0.11	−0.51	**	0.16	−0.04		0.11
Not Enough Money for What Is Needed		−0.003		0.14	−0.11		0.08	0.00		0.14	−0.10		0.08
Constant		−0.02		0.22	0.82	**	0.14	0.11		0.22	0.82	**	0.14

^{^}p<.10

^**p<.05

In exploratory analyses of different types of communication, parent time in the U.S. was positively associated with parent communication with the MA about stress in his or her own life and parent communication with the doctor about issues at school (results available upon request).

Comparison of Acculturation Proxies

Having a parent born in the U.S. was a significant positive predictor of number of psychosocial issues communicated with an MA (Table 3). In contrast, a significant negative predictor of the number of psychosocial problems discussed with an MA was having language discordance between the parent and child, with the parent speaking predominately Spanish at home and the child speaking predominately English at home. Having a child born in the U.S. was also a positive predictor of the number of psychosocial problems communicated with the doctor.

Table 3.

Comparison of Coefficients^† Between Time in U.S. and Alternative Measures of Acculturation

	Count of psychosocial problems discussed with MA (n=359)			Count of psychosocial problems discussed with doctor (n=358)
	Coeff		SE	Coeff		SE
# Years in the U.S. (6 — 10) (n=141)	0.03		0.17	0.13		0.10
(>10) (n=135)	0.36	**	0.17	0.20	^	0.11
Language
Parent Speaks Spanish with Family but Child Speaks English at Home(n=50)	−0.61	**	0.20	0.008		0.10
Parent Speaks English and Spanish with Family and Friends (n=58)	−0.03		0.16	−0.11		0.10
Nativity
Parent Is Born in the U.S. (n=14)	0.71	**	0.26	0.13		0.19
Child Is Born in the U.S. (n=312)	0.18		0.21	0.37	**	0.14
Parent Not Born in the U.S. and Child Born in the U.S. (n=298)	−0.13		0.16	0.21	^	0.11

^**p<.05,

^{^}p<.10

^†Acculturation measures were entered individually into models of the 2 dependent variables

Discussion

The findings from this analysis provide new evidence that more recently immigrated Latino parents are not as inclined to discuss the mental health of their children and the psychosocial status of their family with primary care MAs and doctors compared with Latino parents that have been in the U.S. for a greater number of years. Specifically, compared to parents living in the U.S. for 0–5 years, Latino parents living in the U.S. for more than 10 years discussed more of their child’s psychosocial problems with their child’s MA, were more likely to discuss stress in the parent’s own life with their child’s MA, and were more likely to discuss their child’s problems at school with their child’s doctor. These results are relevant to the need to improve the identification of mental illness in primary care settings among minority populations (42, 66, 67). MAs and doctors might consider these relationships when identifying the health needs of immigrant Latino youth and be attuned to different communication behaviors regarding youth and family psychosocial problems.

Differences between proxies of acculturation and communication with providers identified in this study complement prior studies identifying that less acculturated parents were not as inclined to interpret problem symptoms indicative of ADHD compared with more acculturated parents (45) and that less acculturated youth with attention deficit-hyperactivity disorder (ADHD) were less likely to use stimulant medications than more highly acculturated youth (68). A potential reason for the reticence of less acculturated parents to disclose factors related to their child’s mental health and family’s psychosocial issues relates to concerns over privacy. With greater acculturation in the U.S., it is often the case that immigrants move out of enclaves with a high concentration of immigrants from their same country (69, 70), and therefore may be less concerned that information revealed in the clinic will spread to neighbors and family. To reduce Latino-white disparities in youth mental health care (4) and lower the elevated rates of mood disorder among Latino youth (1), future theoretical work is needed to better understand how Latinos’ attitudes towards mental health care and communication with medical professionals changes with greater exposure to U.S. norms. Further research is also needed to examine the strength of the association between parent/provider communication, diagnosis, and initiation of mental health treatment.

It is important to note that Latino parents were approximately twice as likely to talk to doctors about psychosocial problems than with MAs, although these reservations become less pronounced with more time in the U.S. Prior studies have found that Latinos have generally positive attitudes toward discussing their child’s mental health with pediatric primary care providers. This should be encouraged, and mental health should be added to the checklist of concerns to be reviewed during the primary care visit. On the other hand, there has been a demonstrated reluctance here and in one prior study to talk to MAs (46). This may be due to their lower status than doctors or feared loss of confidentiality due to shared membership in a small community (46). While MAs may be able to elicit some types of information from Latino parents and respond with empathy and support in a manner that encourages the disclosure of mental health concerns, these parents may also wish to reserve the discussion of other topics with professionals whom they perceive to be more qualified. MAs should approach mental health concerns with sensitivity and not push parents into discussions that may ultimately make them feel uncomfortable.

The intersection of willingness to communicate about medical and mental health issues and language, nativity, time in the U.S., acculturative stress, and ethnic identity is complex. We conducted a preliminary exploration of which acculturation proxies were relevant to patient-provider communication, heeding the advice of Guarnaccia et al. (71) and Abraido-Lanza et al. (36) to more fully and systematically assess these relationships for specific health care issues. The range of magnitude and significance of coefficients identified in this study’s results show the diverse ways in which acculturation can be measured and how these different proxy measures provide different windows into the acculturative process (71). We identified that nativity is a significant predictor of likelihood to communicate with providers about psychosocial issues. Being a U.S.-born parent is positively associated with the count of psychosocial issues discussed with an MA, whereas having a U.S.-born child is positively associated with the count of psychosocial problems discussed with a doctor. As in prior studies (72, 73), English language proficiency was an especially important marker of acculturation as it relates to communication with providers. In our study, discordant parent-child language fluency (parent speaks Spanish at home but child speaks English at home) was a significant negative predictor of communication with MAs about psychosocial issues, suggesting that linguistic isolation may be especially acute in these discordant dyads and contribute to feelings of helplessness or fear of divulging information.

A significant limitation of this study is that the proxies of acculturation used are less informative than more complex conceptual models of acculturation that account for bi- or multidirectional changes in attitudes (36). However, while measuring acculturation using proxy variables may overlook some of the nuances of acculturation captured by more complicated models, it allows researchers to judge the general relationship between exposure to U.S. culture and changes in health behaviors. In areas where little research has been conducted on the impact of acculturation, the use of proxy variables is generally accepted (15), and can help to indicate avenues for future research using more complex models. A second limitation is the potential bias of combining the SDQ and MHI as a singular measure of mental health difficulties across age groups. Sensitivity analyses that remove the mental health difficulties measure or only use the SDQ (on a subsample of 286 youth 3–16 years of age) produce regression estimates for acculturation proxies that are similar in direction and magnitude as the original analysis.

In summary, our study of Latino parent/child dyads in a FQHC in Washington DC identified that greater number of years in the U.S. was positively associated with parents’ likelihood to communicate with MAs about psychosocial issues related to their child’s mental health. These results were similar in direction but not significance in analyses of communication with doctors. These positive associations between time in the U.S. and communication were consistent across other proxy measures of acculturation, suggesting the evolving nature of trust with providers over the course of acculturation. Providers and MAs ought to be keenly aware of this variation in resistance to communicating psychosocial issues because of the importance of communication and trust in the diagnosis and treatment of youth mental illness.