Physicians’ Influence on Primary Care Patients’ Reluctance to Use Mental Health Treatment

According to data from the World Health Organization, a quarter of primary care visits are for mental health issues (1). Research in Israel and elsewhere indicates that about half of patients with a mental health issue seek care not from a mental health professional (2,3) but from primary care physicians (4), mainly because of perceived stigma (5). Because positive attitudes toward mental health treatment lead to actual service use (6,7), we wished to determine whether primary care physicians might positively affect their patients’ willingness to use mental health services.

The objectives of this study were, first, to examine differences between primary care patients who differ in degree of positivity in their attitudes (positive, ambivalent, or negative) toward mental health professionals and second, among patients with ambivalent or negative attitudes, to examine characteristics of those who change or do not change their attitudes after primary care physicians recommend specialized mental health treatment.

Methods

The study population consisted of primary care patients who, at the time of the study (2002–2003), were between ages 25 and 75 and visited one of eight selected clinics in Israel’s largest health maintenance organization (HMO) during the year before the interview. For this study, which is part of a larger study about primary care utilization (8), a decision was made to exclude younger adults (<25) because of relatively few medical problems or, if in active military service, use of army medical services. Older adults (>75) also were excluded due to generally higher utilization rates for somatic problems. The clinics represent a cross-section of the Israeli population. The analysis is based on face-to-face interviews with 902 Jewish patients. Only patients who signed informed consent forms (77% of those eligible) were included. The study protocol and instruments were reviewed and approved by the institutional review board of the HMO as well as by the medical directors of the eight selected clinics. [Additional details are available online in the data supplement to this report.]

We administered two mental health assessment instruments, for which reliability and validity have been established. A shortened version of the Brief Symptom Inventory provided a global measure of distress, which is known as the Global Severity Index (GSI). The Composite International Diagnostic Interview–Short Form (CIDI-SF) provided diagnoses of depression and anxiety. [Additional details are available in the online data supplement.]

Patients self-rated their mental health over the previous 12 months using a five-category scale on which responses range from excellent to poor. Patients who rated their mental health as poor were considered subjectively psychologically distressed. Patients who reported working less than usual or being less productive due to psychological distress were characterized as dysfunctional.

Patients’ attitudes toward mental health treatment were assessed with a vignette concerning an individual with emotional difficulties (for example, unhappy, depressed, loss of appetite, or sleeping problems). Respondents were asked two questions: if a friend was in such a situation, would the respondent advise treatment from a mental health professional (such as a psychologist or social worker)? Second, respondents were asked to place themselves in the same situation and consider whether they would seek mental health treatment for themselves.

Attitudes toward treatment were categorized as positive, uncertain, or negative. Because significant agreement was found between answers to the two questions (κ=.50, p<.001), they were combined into one variable with three categories: positive (the answer to both questions was positive or one answer was positive and the other uncertain), uncertain (the answer to both questions was uncertain or one answer was positive and the other negative), or negative (the answer to both questions was negative or one answer was negative and the other uncertain).

Patients initially uncertain or negative were asked about their willingness to follow their primary care physician’s recommendation to get treatment from a mental health professional. The responses were yes, uncertain, or no. To examine the influence of the primary care physician’s recommendation, follow-up responses were compared with the initial responses, and a two-category variable was created: positive influence (became positive after receiving the primary care physician’s recommendation) or no influence (did not become positive after receiving the primary care physician’s recommendation).

Sociodemographic questions relevant to mental health treatment seeking, as documented in various publications (6,9,10) concerned gender, place of birth, age, years of education, marital status, religious observance, and sufficiency of family income to cover the costs of daily living.

Initially, demographic variables were assessed for significant bivariate associations with two dependent variables, the readiness to get treatment from a mental health professional and the influence of the primary care physician on the patient’s readiness to get treatment from a mental health professional (cross-tabulation).

Pairwise tests of the equality of row proportions were performed. The results were based on two-sided z tests with a significance level of .05. The p values of the tests were adjusted with the Bonferroni method.

For each dependent variable, the independent variables found to be significantly associated in the previous analysis were entered together into a logistic regression in order to check their association with the dependent variables after analyses controlled for the effect of all other independent variables entered into the model. The dependent variable—the attitude toward getting treatment from a mental health professional—had three categories. Thus a multinomial logistic regression was performed when each regression was done twice, each time with a different reference category. Data analyses were performed with SPSS/PC, version 19.0. Two-sided tests of significance were used in all analyses.

Results

Sociodemographic and clinical characteristics of the study population are provided in Table 1. Overall, about half (54%) expressed readiness to get treatment from a mental health professional when needed, a third (33%) were uncertain, and 13% were unwilling to do so.

Bivariate analyses revealed that the readiness of primary care patients for mental health treatment was significantly associated with gender, place of birth, age, education, religious observance, subjective psychological distress, and a diagnosis of anxiety (Table 1). These variables were included in multinomial logistic regressions with the exception of diagnosis of anxiety because of its high correlation with subjective psychological distress (χ²=79.76, df=1, p<.001). [A table in the online data supplement provides results of the multinomial logistic regressions.]

Among patients who were subjectively psychologically distressed according to self-evaluations, 88% (N=61 of 69) also showed distress according to the GSI or received a CIDI-SF diagnosis of depression or anxiety. In contrast, among those who were distressed according to the GSI or had a diagnosis of depression or anxiety, only 21% (N=61 of 292) acknowledged that they were psychologically distressed.

Multinomial logistic regressions indicated that the probability of an uncertain or a negative attitude toward specialized mental health treatment was similar to a negative one, except for gender; women had a higher probability of a positive or uncertain attitude compared with men, who had more negative attitudes. The probability of having an uncertain or negative attitude rather than a positive one was significantly higher among patients born in the former Soviet Union, adults ≥65 years, younger adults aged 25–34, patients without high school education, and those whose religious observance was ultra-Orthodox. Another predictor of uncertain or negative attitudes was not feeling subjectively psychologically distressed.

Among patients who initially had an uncertain or negative attitude, 28% (116 of 418) were influenced positively by their physician. More specifically, among those who expressed uncertainty, a third (32%, or 96 of 297) changed their attitude to positive after receiving the physician’s recommendation. In contrast among participants with negative attitudes, a similar change occurred for less than one-fifth (17%, 20 of 121).

The influence of primary care physicians on patients’ attitudes toward mental health treatment was significantly correlated in bivariate analyses with place of birth, diagnosis of depression, the four-category variable of psychiatric diagnosis, and initial attitude toward treatment. All of these variables except diagnosis of depression remained significant in the logistic regression. [Additional details are available in a table in the online data supplement.]

The logistic regression indicated that the probability of changing attitudes toward specialized mental health treatment from negative or uncertain to positive after the physician’s recommendation was significantly higher among patients born in the former Soviet Union or Western countries and among those with both anxiety and depression. As expected, those who were initially uncertain about mental health care tended to be significantly more affected by their physician’s recommendation than those with negative attitudes.

Discussion

Studies in the Western world have shown that a large proportion of people with serious emotional problems are willing to consult a mental health professional (7,9). This study is consistent with these studies in that only a relatively small percentage (14%) of patients had a negative attitude toward using mental health services. With regard to Israel, the National Health Survey—a general population survey conducted in the same period as this study—revealed that close to half of respondents with mood-related or anxiety diagnoses had a positive attitude toward mental health treatment (2). In this study, almost 60% of participants with psychopathology had positive attitudes, as expressed by a willingness to get treatment. This difference might be explained by different populations. That is, primary care patients may have more severe mental problems than those in the general population and may be more receptive to mental health treatment.

A related finding revealed that those who subjectively perceived psychological distress were more receptive to treatment than others, even more than those with clinical diagnoses who did not perceive that they were distressed. This finding is consistent with studies showing that a major influence of help-seeking behavior is perceived need (6,11).

This study also revealed that women, respondents aged 35–64 years, and those with a higher education had more positive attitudes about treatment, a finding consistent with other studies (7,9). With regard to country of origin, immigrants from the former Soviet Union were especially hesitant to get treatment, perhaps reflecting cultural and political stigmas (12). In addition, other research showed a tendency for Soviet Jewish immigrants (13) to express mental health problems through somatic complaints and to seek help from primary care physicians.

To the best of our knowledge, the relationship between religious observance and readiness to get mental health treatment has not been examined. This study revealed that religiously observant patients were less likely than others to have positive attitudes toward treatment. This might reflect a stigma among more religious Jews (10), who prefer to seek mental health assistance from religious authorities (such as rabbis or Kabbalists).

Many patients prefer to use primary care physicians for mental health care (4). In this study, even among patients with uncertain or negative attitudes toward specialty mental health treatment, 28% were influenced positively by their physician. This is consistent with an Israeli survey that found an increase in the willingness to access mental health treatment when their physicians recommended it (14). Among patients not inclined to get mental health treatment, the physician’s recommendation had greater impact on those with more severe clinical diagnoses. In regard to sociodemographic characteristics, the probability of being more influenced by the physician’s recommendation was higher among patients from the former Soviet Union, who are known to respect authority figures (15).

One limitation of this survey is the absence of data concerning the actual use of mental health services, especially by those positively influenced by their physician’s recommendations. It would have been preferable to also have had an analysis of Arabs; however, the small number of Arab patients precluded this possibility. Finally, the absence of objective work measurements related to distress also affected this analysis.

Conclusions

A positive attitude toward treatment is a prerequisite for outpatient mental health treatment seeking. Therefore, it is important to help patients change negative or uncertain attitudes toward such treatment, especially if patients have psychiatric diagnoses. A major finding of this study is the positive impact that primary care physicians’ recommendations had on reluctant patients. Therefore, encouraging physicians to discuss mental health issues would likely promote more positive attitudes and thereby increase patients’ willingness to use mental health services when needed.