Introduction

Many people can fully recover from severe mental illness (SMI). There are times, however, when it becomes necessary to ask for help. For example, help with medication may be needed, or help in recognizing and addressing signs of relapse. Or help may be needed with housing or employment, or in dealing with difficult emotions. Across these situations, however, some people may be more willing to ask for help than others. Such willingness may be viewed as a crosscutting tendency, one that spans requests for different types of assistance from different people. What makes one person ask for help more generally and another decide to go it alone? Relative to people who are reluctant to ask for help, maybe people who are more willing to ask for support are better off because needs are being met. For example, maybe they are happier, or maybe they are more likely to stay out of psychiatric hospitals.

However predisposition to request aid is unstudied among people with SMI. We therefore addressed three research questions: (1) What are characteristics of people who are willing to ask for help? (2) Are people who are less willing to ask for help also more likely to be admitted to psychiatric hospitals, perhaps because they are not getting the assistance they need to maintain stability? (3) Are people who are more willing to ask for help also more likely to be happy, perhaps because they are not struggling on their own?

The literature on help-seeking behavior is informative. However much of it addresses something somewhat different—the pursuit of professional mental healthcare instead of our focus on a more generic willingness to request assistance (e.g., responses to survey items such as “I ask for help when I need it”). Studies on help-seeking distinguish between formal help-seeking from professionals (e.g., mental health service providers) and informal help-seeking from family and friends. Most investigations (about 66%) address the former (professionals) and the remainder addresses both (professionals as well as family and friends: Rickwood and Thomas 2012). We address both (together in single “willingness” construct). Most studies on help-seeking (see Rickwood and Thomas 2012 for a review) do not focus on people with SMI, and those that do rarely focus on specific types of mental illness (e.g., bipolar disorder). We also focus on SMI more generally.

The literature on help-seeking among people with SMI makes four pertinent points. First, people with greater illness severities are more likely than healthier counterparts to seek assistance (Bebbington et al. 2000), and perceived need for aid is the best predictor of attempts to obtain it (Eisenberg et al. 2011). Second, a desire to handle problems independently is the most common explanation offered for not seeking assistance (Mojtabai et al. 2011). “I can handle it myself” is a common refrain. Third, stigma keeps people from asking for help (Clement et al. 2015; Savage et al. 2016; Schomerus and Angermeyer 2008), for requests for aid can give away the presence of a mental illness that is judged harshly by society. Fourth, people who normalize their problems are less liable to seek help. That is, help-seeking is avoided among people who believe that symptoms of SMI are not serious or are commonplace (Biddle et al. 2007; Savage et al. 2016).

We offer several hypotheses. In accordance with prior research (Rickwood and Thomas 2012), we hypothesized that older individuals would be more likely than younger counterparts to ask for help. Younger individuals may have less experience and less insight into problems. If there is no problem, then why ask for help? Also based on prior research (Sosulski and Woodward 2013), we hypothesized that women would be more likely than men to request assistance. Men might be socialized more often to fix their own problems rather than show weakness by asking for help. Many men are taught to appear strong by hiding intrapersonal struggle. We also hypothesized that people with higher levels of internalized stigma would be less willing to ask for help, for other studies have found an association (see a review by Clement et al. 2015). If people internalize stigma (e.g., “persons with SMI are strange”), then they may be unlikely to request assistance because they want to hide psychopathology. Based on findings by Sosulski and Woodward (2013), we also hypothesized that people with substance use disorders would be less likely than other people to ask for help. Maybe the frequent isolation of substance abuse precludes requests for aid, or maybe help is instead pursued in a drink or drug (and not from other people).

Finally, we hypothesized that people with more social support are more likely to ask for help. There are more people to ask, and there is greater selection of individuals to choose from. In addition, members in a stronger support network can more often help people to identify need for assistance (e.g., “I think you need help”) especially when the need is burdensome to network members (e.g., “this is taking a toll on me:” Villatoro et al. 2014). However the opposite may instead be true. People with weaker social networks might ask for help more often (Yeung et al. 2013) precisely because the help is not readily available within the network. In other words, help must be requested elsewhere (i.e., outside the network) because there is a lack of in-network people who could freely give support.

We used Andersen’s (1995) model of health service use as a theoretical framework, except that we altered the outcome. Instead of predicting actual use of help (service use), we substituted a generic willingness to ask for help, and then added two follow-up variables (happiness and psychiatric hospitalization following willingness to request assistance). Shown in Fig. 1 are: (1) predisposing characteristics (age, sex, race, education); (2) need based characteristics (presence of substance use disorder, depression, serious psychological distress, or health-related difficulties); and (3) enabling characteristics (income, neighborhood desirability, social support, and lack of internalized stigma) that promote: (a) willingness to request assistance; and (b) subsequent happiness and psychiatric hospitalization.

Fig. 1
figure 1

Conceptual framework

Full size image

Our study was exploratory in nature, for: (1) we focused on a topic that has received little consideration; (2) our sample size was limited; and (3) our research design was cross-sectional rather than longitudinal, which constrained our ability to firmly establish causation. Our main purpose was to call attention to the need for further investigation, and our end-of-article discussion focuses largely on ideas for future research. Our sample was unique, for study participants were members of Fountain House, a mental health service agency in New York City that offers a broad range of activities on a daily basis. It is a rich social environment, where evening and weekend gatherings are available, and where staff is even encouraged to socialize with members outside of traditional 9–5 h. Thus most peers and staff support each other on a continual basis. What are characteristics of people who are willing to ask for help in such a socially rich environment, and is such willingness associated with certain outcomes (happiness; psychiatric hospitalization)?

Methods

All study participants had SMI and they spoke English. We trained a doctoral student in social work to administer our structured interview to Fountain House members. Participants were not paid, and were recruited to be in the study if they were available on days of data collection. In other words, we came to Fountain House and recruited members to be in the study on the very same day. This investigation was part of a larger study on social network formation (the creation of interpersonal connections). Written informed consent and IRB approval was obtained.

Measures

Dependent Variables

We had three dependent variables. Our primary dependent variable was willingness to ask for help. It was assessed using a 3-item subscale of the Recovery Assessment Scale (Giffort et al. 1995): (1) “I am willing to ask for help;” (2) “I ask for help when I need it;” and (3) “I know when to ask for help.” Items were measured on a 5-point Likert-type scale (1 = strongly disagree to 5 = strongly agree). Scores on the three items were summed and divided into quartiles. Least willing study participants scored in the bottom quartile, somewhat willing participants scored in the 2nd quartile, and most willing participants scored in the top 2 quartiles. Sample size limitations precluded us from examining the top 2 quartiles separately. The Recovery Assessment Scale was developed by service consumers in order to capture the way people experience recuperation from mental illness. It has very good internal consistency (Cronbach’s alpha of 0.76–0.97, depending on the study), test–retest reliability, and interrater reliability (Salzer and Brusilovskiy 2014).

Our remaining two dependent variables were: (1) psychiatric hospitalization, where study participants were asked if they had been admitted in the past 12 months for an overnight stay for problems with emotions, nerves, mental health, or use of alcohol or drugs; and (2) happiness, where study participants were asked to take all things together and report how happy they are currently (very happy, pretty happy, not too happy, not at all happy).

Independent Variables

Presence of a substance use disorder was assessed using the CAGE (acronym based on its 4 items: Ewing 1984). The yes/no items on the CAGE include: (1) “Have you ever felt that you needed to cut down on your drinking or drug use?” (2) “Have people annoyed you by criticizing your drinking or drug use?” (3) “Have you felt bad or guilty about your drinking or drug use?” and (4) “Have you ever had a drink or used drugs first thing in the morning to steady your nerves or to get rid of a hangover (eye-opener)?” Participants who endorsed two or more items were counted as having a substance use disorder. The CAGE is adequately correlated (0.48–0.70) with other substance abuse screening instruments (Dhalla and Kopec 2007).

Presence of depression was assessed using the 10-item Center for Epidemiological Studies Short Depression Scale (CES-D-10). Items include: (1) “I was bothered by things that don’t usually bother me;” (2) “I had trouble keeping my mind on what I was doing;” and (3) “I felt depressed.” On a 4-point Likert-type scale (0 = rarely or none of the time to 3 = all of the time), study participants reported on these feelings in the past week. The cutoff score for depression is 10 or greater. The CES-D-10 has strong psychometric properties (see Bjorgvinsson et al. 2013).

Serious psychological distress was measured by the Kessler-6 Scale (Kessler et al. 2003). With 6 items on a 5-point Likert-type scale (0 = none of the time to 4 = all of the time), study participants were asked how often in the past 30 days they felt a certain way (e.g., nervous, hopeless, restless or fidgety). Possible scores range from 0 to 24, with scores of 14 or higher indicating serious psychological distress. The K6 has strong psychometric properties (see Cornelius et al. 2013).

Health-related difficulties were assessed by asking study participants to rate the degree to which health problems interfered with life and activities with close friends and family members during the last 30 days (5-point Likert-type scale from 0 = not at all to 4 = extremely).

Neighborhood desirability was assessed using the Social Position items on the Collaborative Psychiatric Epidemiology Survey (CPES: Inter-university Consortium for Political and Social Research 2016). Items include: (1) “people in my neighborhood can be trusted;” (2) “I feel safe being out alone in my neighborhood during the night;” and (3) “people sell or use drugs in my neighborhood.” The 7 Items were measured on a 4-point Likert-type scale (1 = very true to 4 = not true at all). Items were summed and divided into quartiles. People scoring in the top quartile were considered to be residents of the least desirable neighborhoods. Information on psychometric properties of scales for neighborhood desirability and health-related difficulties was unavailable.

Social support was assessed using the Abbreviated (5-item) Medical Outcomes Study Social Support Survey (MOS-SSS: Sherbourne and Stewart 1991). On a 5-point Likert-type scale (1 = none of the time to 5 = all of the time), study participants were asked to report how often certain forms of support were available if needed (e.g., someone to get together with for relaxation; someone to love and make you feel wanted). Items were summed and divided into quartiles. People in the top quartile were considered to have the most support. The MOS-SSS has strong psychometric properties (see Moser et al. 2012).

Internalized stigma was assessed using the Internalized Stigma of Mental Illness Inventory-10 item version (ISMI-10: Ritsher et al. 2003). The ISMI-10 measures the extent to which people with mental illness apply negative stereotypes about psychopathology to themselves. Items on the ISMI-10 include: (1) “I don’t socialize as much as I used to because my mental illness might make me look or behave ‘weird;’” (2) “I can have a good, fulfilling life despite my mental illness;” and (3) “mentally ill people tend to be violent.” All items of the ISMI-10 were measured on a 4-point Likert-type scale (1 = strongly disagree to 4 = strongly agree). Mean scores exceeding 2.50 reflected a high degree of internalized stigma. The ISMI-10 has a high level of internal consistency (Cronbach’s alpha of 0.75: Boyd et al. 2014).

Analysis

We used logistic regression to examine willingness to ask for help as it relates to all of the independent variables in the first column of Table 1, and then used logistic regression again to examine: (1) past-year psychiatric hospitalization as it relates to willingness to ask for help; and (2) happiness as it relates to willingness to ask for help.

Table 1 Prediction of willingness to ask for help among adults (N = 150) with severe mental illness
Full size table

Results

Among the 150 persons with severe mental illness, four factors were associated with willingness to ask for help (Table 1). First, people between the ages of 50 and 59 were 4.34 times as likely (CI 1.32–14.31, p < .05) to ask for help as people between the ages of 18 and 39, Hispanics were 8.51 times as likely (CI 2.05–35.36, p < .01) as Whites to ask for help. Third, relative to individuals with the lowest incomes (bottom quartile), those with the highest incomes (top quartile) were 7.23 times as likely (CI 1.76–29.97, p < .01) to ask for help. Fourth, relative to individuals with the lowest levels of social support (bottom quartile of the MOS-SSS), those with the highest levels were 3.82 times as likely (3rd quartile: CI 1.17–12.46, p < .05) and 12.36 times as likely (top quartile: CI 3.01–50.85, p < .001) to ask for help.

Relative to individuals who were less willing to ask for help (n = 78, or 52%), people who were willing to ask for help (n = 72, or 48%) were 2.07 times as likely (CI 1.01–4.26, p < .05) to report being very happy or pretty happy (as opposed to not too happy or not at all happy). Seventy-eight percent (56/72) of people who were willing to ask for help were very happy or pretty happy, while 63% (49/78) of people who were less willing to ask for help were very happy or pretty happy. Findings for psychiatric hospitalization in the last year were not significant. In other words, people who were less willing to ask for help were no more likely to be hospitalized than counterparts who were willing to request aid.

Discussion

Relative to Caucasians, we found that Hispanics were much more likely to be willing to ask for help. More specifically, although more research is needed, perhaps Hispanics are more willing to ask family members for assistance. In Latino families, Villatoro et al. (2014) discuss familismo, which includes a tendency to provide support and advice during difficult times. In contrast, Caucasians are raised more often to be independent, to more singlehandedly go forth in the world, and may be less willing to ask for help.

We also found that people with the highest incomes were much more likely to be willing to ask for assistance than people with the lowest incomes. Although the high incomes in this group weren’t especially high (all except one were under $48,000), perhaps people with more money are more willing to ask for help because they can afford to purchase needed supports (i.e., they can buy the help). Or perhaps people with more money have greater confidence in their ability to obtain assistance because they have developed confidence in their ability to generate income. After all, if they are able to successfully earn more money or at least collect it (e.g., through SSI), then they may feel like they can successfully obtain other types of support when they ask for it. Greater self-efficacy may be key. However it remains unclear why people with lower incomes are not more willing to ask for help, for poverty often demands a greater need for assistance. More research is needed, and employment data (unavailable in our study) must be collected.

We found that people with higher levels of social support were much more likely to be willing to ask for help than people with the lowest level of social support. This substantiates our hypothesis that willingness to ask for help is promoted by having more support available, or by having greater selection of supports for diverse needs. Our alternative hypothesis was not supported. Weaker social support does not appear to promote willingness to ask for assistance, even though there might at times be greater need to ask for help elsewhere when such help is not offered in one’s own limited support system. More research is needed in order to investigate the different circumstances under which weaker versus stronger support systems promote willingness to request aid.

More research is also needed in order to explain our age-related finding. People of ages 50–59 were more likely to be willing to ask for help than younger counterparts (18–39). Although we hypothesized that older individuals would be more willing to ask for assistance (perhaps because they have more insight that they need it), it remains unclear why people in their forties or over the age of 59 did not differ from younger counterparts. (Findings for people in their forties approached statistical significance, and would likely have been significant with a larger sample size.) Further investigation is needed, and studies must also separate people in their twenties from people in their thirties, for sample size limitations precluded us from making this distinction.

Three of our hypotheses were not substantiated. First, contrary to prior research on help-seeking behavior (Sosulski and Woodward 2013), women were no more likely than men to be willing to ask for help. Even though women may be taught more often that vulnerability is acceptable, and even though men may be socialized more often to go it alone, we were unable to find gender differences. Second, contrary to prior research by Sosulski and Woodward (2013), willingness to ask for help did not differ among people with versus without substance use disorders. Addicts may isolate more, and many may seek help from substances instead of from people, but apparently they are not less willing to ask for assistance. Third, even though people with internalized stigma may wish to keep the mental illness hidden from others, they were not less likely than other people to be willing to ask for help. This contradicted prior research on help-seeking behavior (Clement et al. 2015; Savage et al. 2016; Schomerus and Angermeyer 2008).

Although willingness to ask for help was not associated with psychiatric hospitalization, it did relate to happiness. People with SMI might be happier if needs are met after asking for assistance. More research is needed on other outcomes stemming from willingness to request aid. For example, researchers could study: (1) intervention adherence, for other people can provide transportation to needed supports when help is requested, for example, or provide help (when requested) in taking medications as prescribed; (2) satisfaction with care, for intervention approval may increase when appeals are made for assistance; (3) rapport with service providers, for relationships with practitioners could benefit from requests for help, especially if such requests are granted; (4) homelessness, if housing-related resources are lost less frequently because pleas are made for support; and (5) suicidal ideation, if help is requested before ideas relating to self-harm are formulated.

In addition to studying these outcomes, we propose nine other ideas for future research. First, researchers must study illness severity, for people with greater symptomatology may be more likely than healthier individuals to request assistance (Bebbington et al. 2000). Second, researchers should examine levels of perceived independence (e.g., “I don’t need any help”), for a desire to handle problems alone is the most common explanation offered for not seeking assistance (Eisenberg et al. 2012; Mojtabai et al. 2011). Third, future studies could examine past experiences with assistance. If people have been helped in the past, then they may be likely to request help again (Savage et al. 2016). Similarly, expectations should be examined. Does one expect to get help when asked? Will the help be valuable?

Fourth, researchers must examine insight. One needs to know that help is needed before requests for assistance are made (Yeung et al. 2013), and fifth, future studies should examine tendency to normalize symptoms of SMI (e.g., “everyone goes through this”), for such normalization can prevent willingness to request help (Biddle et al. 2007; Savage et al. 2016). If people believe that problems will dissipate without assistance, or if they believe that symptoms are not serious, then there is little motivation to ask for support (Eisenberg et al. 2012). Sixth, researchers should examine stages in the help-seeking process. For example, people may be less willing to ask for help when a problem is just beginning. However willingness could grow as the problem continues or exacerbates (Villatoro et al. 2014).

Seventh, researchers could examine proximity of people within a social network (Yeung et al. 2013). People might be more willing to ask for help from nearby versus far-away individuals, or perhaps if nobody is close geographically then there is a greater need to ask for help. Similarly, moving from one location to another (mobility) could also affect willingness to ask for help (Yeung et al. 2013). A stranger in a new area is likely to be in need of assistance. Eighth, acculturation should also be examined. Recent immigrants might be more willing to ask for assistance, or perhaps they are less willing if they are more isolated. Ninth, research is needed on interventions that increase willingness to ask for help, for we found that such willingness relates to happiness. Interventions that increase social support could also be enhanced, for we found that social support was highly associated with inclination to invite aid.

Our study has limitations. First, because of our cross-sectional design, we were unable to definitively establish causal direction in relation to: (1) presence of substance use disorder, depression, serious psychological distress, or health-related difficulties, for people might develop these problems following an unwillingness to ask for help (rather than vice versa); (2) social support, for people might have less support after they are unwilling to request assistance (rather than vice versa); and (3) happiness, for unhappier people might be less willing to seek aid (instead of willingness leading to happiness). However we followed Andersen’s (1995) theoretical framework, and (in all of the cross-sectional studies that use the model) use of help could always precede need-based or enabling characteristics. Only predisposing characteristics (e.g., age, sex, race) must come first. Nevertheless, longitudinal research is needed in order to firmly establish causality. In addition, sample size limitations prevented us from completing some analyses. For example we were unable to examine certain age differences, or differences between the top two quartiles of willingness to ask for help. Moreover, we were unable to examine important variables in our study of factors that might be associated with willingness to ask for help (or possible correlates of this willingness such as happiness or prevention of psychiatric hospitalization). For example, we did not assess: (1) shortage of services; (2) lack of access to available care; (3) services used outside of Fountain House; (4) stigma relating to asking for or receiving help, which differs from internalized stigma; (5) success in the past of receiving help, which could be associated with future willingness to request aid; (6) diagnoses, although everyone had SMI (schizophrenia, bipolar disorder, or major depressive disorder); (7) needs relating to side effects of medication, for example, or other needs that were not assessed in the study (i.e., apart from substance use disorder, depression, serious psychological distress, or health related difficulties); or (8) experiences relating more generally to poverty or marginalization in society. Finally, our findings may not generalize to other agencies or locations where people with SMI could be more isolated. Fountain House members are surrounded by people who can provide assistance when asked. For example, evening and weekend activities are available, and staff is even encouraged to socialize with members outside of normal working hours. In other words, relative to some other people with SMI, Fountain House members might be more willing to ask for help than because there is greater availability of assistance. More research is needed in other agencies and in other locations.

Conclusions

Hispanics were more likely than Caucasians to be willing to ask for help, and people with the highest incomes (relative to the lowest) were also more likely. In addition, individuals with more social support were more likely than people with the least social support to be willing to request assistance, and people between the ages of 50–59 were more likely than younger counterparts (ages 18–39). Finally, people who were willing to ask for help were more likely than less willing individuals to report being very happy or pretty happy. More research is needed in a wide variety of areas in order to better understand predisposition to seek aid, and interventions are needed that promote it.