Lifetime Treatment Contact and Delay in Treatment Seeking After First Onset of a Mental Disorder

The literature has found consistently that most people who report having a mental disorder in the past 12 months also report not receiving help for their illness. In the United States, for example, Wang and others (1) have estimated that only 32.9% of people with a mood disorder and 21.7% of those with an anxiety disorder in the past year had used mental health services during that period. In the Netherlands, similar rates—37.4% and 22.2%, respectively—were found (2). However, less frequently cited findings indicated that a majority of people with a common mental disorder had used health services for their mental health problems at some point during their lifetime (3,4). This finding held true for mood disorders but was less pronounced for phobias and substance use disorders (5).

These findings prompted several authors to focus on delay in initial treatment rather than on failure to make contact for treatment. A new generation of general population surveys on mental health that use the Composite International Diagnostic Interview (CIDI) 3.0, which includes more questions than earlier versions about treatment for specific disorders, has made it possible to study this issue in greater depth. A landmark paper by Wang and colleagues (6), who analyzed data from 15 countries in the World Mental Health (WMH) Survey Initiative, reported the proportion of respondents with lifetime disorders in developed countries who made prompt treatment contact in the year of onset. These percentages ranged from 28.8% to 52.1% for mood disorders, 11.2% to 36.4% for anxiety disorders, and 2.0% to 18.6% for substance use disorders. For those who made contact, median delays varied from one to six years for mood disorders, three to 28 years for anxiety disorders, and six to 28 years for substance use disorders. Thus, despite the fact that a majority of persons with a mood disorder eventually make treatment contact during their life, it may take several years after onset to do so.

The WMH surveys have also been examined for predictors of failure to make initial treatment contact. Men, older cohorts, people who are younger at disorder onset, and residents of developing countries had lower rates of initial treatment contact (6). In the United States, poorly educated and never-married people also had lower odds of treatment contact (1). Data from the National Comorbidity Survey Replication (NCS-R) suggest that in the United States, predictors of failure to ever make a treatment contact were very similar to predictors of duration of delay. However, no results were presented (1), so predictors associated with longer delays in treatment seeking after disorder onset remain unknown.

A related issue that has not been studied in the WMH surveys is the extent to which people perceive having received helpful treatment and if so, whether it occurred at a timely moment. This is an important topic because positive perceptions of the care received form people's attitudes toward mental health care and, in the long run, influence their decisions about seeking treatment if problems recur (7).

Data from the Netherlands Mental Health Survey and Incidence Study−2 (NEMESIS-2) were used to study these issues in more detail. Following the example of the WMH Survey Initiative, the NEMESIS-2 made use of the CIDI 3.0, a diagnostic interview that made it possible to assess treatment seeking and delays in treatment seeking, including receipt of helpful treatment, after first onset of a mental disorder.

Methods

Sample

For the NEMESIS−2, a multistage, stratified random sampling procedure was applied. First, a random sample of municipalities was drawn. Second, a random sample of addresses of private households was drawn from postal registers in these municipalities. Third, an interview was conducted with the member of the household who was aged 18 to 64 and sufficiently fluent in the Dutch language and whose birthday most closely preceded the first contact with the household. Addresses of institutions were excluded. Persons temporarily living in institutions, however, could be interviewed later, during the fieldwork, if they had returned home.

Fieldwork

The Minister of Health, Welfare and Sport sent a letter to selected households (N=11,349) to introduce and recommend the study. An accompanying brochure explained its goals in more detail; it also referred to a Web site for respondents. Shortly after receiving this letter, households were contacted by telephone or visited in person. At least ten visits or phone calls at different times and on different weekdays were made. This resulted in a nationally representative sample (weighted response rate of 65.1%; N=6,646), although younger respondents were somewhat underrepresented (8).

The interviews, averaging 95 minutes, were conducted by laypersons using a laptop computer, and almost all were held at the respondent’s home. The fieldwork was performed from November 2007 to July 2009.

The study was approved by a medical ethics committee. After having been informed about the study aims, respondents provided written informed consent. A more detailed description of the design and fieldwork was previously published (8).

Diagnostic instrument

DSM-IV diagnoses were made by using the CIDI 3.0, a fully structured lay-administered diagnostic interview (9). In the Netherlands, the CIDI 3.0 was first used for the European Study on the Epidemiology of Mental Disorders (ESEMeD), which is part of the WMH Survey Initiative. The CIDI 3.0 version used for the NEMESIS−2 was an improvement on the one used for the Dutch ESEMeD.

The disorders considered by this study included mood disorders (major depression, dysthymia, and bipolar disorder), anxiety disorders (panic disorder, agoraphobia without panic disorder, social phobia, specific phobia, and generalized anxiety disorder), substance use disorders (alcohol or drug abuse and dependence), and impulse-control disorders (attention-deficit hyperactivity disorder [ADHD] and oppositional defiant disorder). Impulse-control disorders were assessed only for respondents aged 18–44 because of concerns about recall bias among older respondents (10). Lifetime prevalence and age of onset were assessed separately for each disorder.

Clinical calibration studies in various countries (11) found that the CIDI 3.0 assesses mood, anxiety, and substance use disorders with generally good validity compared with blinded clinical reappraisal interviews. A clinical reappraisal interview carried out in a subsample of the NCS-R also found a valid assessment of ADHD by the CIDI 3.0.

Lifetime treatment contact

At the end of each CIDI diagnostic section, respondents were asked whether they ever talked to a general medical doctor or other professional about the disorder under investigation and if so, how old they were when they first did so. The term “other professional” was broadly meant to include psychologists, counselors, spiritual advisors, herbalists, acupuncturists, and any other healing professional.

Helpful treatment

The respondents who had ever talked to a general medical doctor or other professional about the disorder under investigation were also asked if they considered any treatment that they had received for the disorder to have been helpful and if so, how old they were when they first received helpful treatment.

Predictors

Age at disorder onset was coded in four categories corresponding to important stages of life (0–12, 13–19, 20–29, and 30–64) (12), similar to age categories used in other studies (1). Age at disorder onset, age at interview, sex, and education level were used as predictors of lifetime treatment contact and delay in seeking treatment after disorder onset. These variables were also used as predictors of receipt of lifetime helpful treatment and delays in seeking helpful treatment after initial treatment contact.

Statistical analysis

The data were weighted to correct for differences in the response rates among several population groups and for differences in the probability of selection of respondents within households (8), so that it would be possible to generalize the results to the national population. Robust standard errors were calculated by using the first-order Taylor-series linearization method, as implemented in Stata 11, in order to obtain correct 95% confidence intervals and p values (13).

First, summary statistics were used to describe lifetime treatment contacts and delays in seeking treatment after disorder onset. Second, Cox regression analyses were used to examine predictors of treatment contact separately for persons with any mood, any anxiety, or any substance use disorder as well as for respondents with any mental disorder. In these analyses the effect of two predictors varied over time regarding lifetime treatment contact (interaction effect). As a consequence, these effects are presented for two different periods (the first two years after disorder onset and more than two years after disorder onset). As an example, older age groups had significantly lower odds of treatment contact than respondents aged 18–24, and this effect became especially apparent more than two years after onset among the persons with any anxiety disorder or any mental disorder. Predictors of ever making a treatment contact among persons with any impulse-control disorder could not be studied because of the small number of people with these disorders.

Third, generalized linear models (GLMs) with a gamma distribution and a log link function were used to study predictors of delay in initial treatment contact after disorder onset separately for persons with any mood, anxiety, or substance use disorder or any disorder who eventually made a treatment contact for the disorder. Because many respondents reported seeking treatment in the same year of disorder onset, these GLMs fitted the rather skewed data better than an ordinary least-squares regression model or other GLM. The results of the GLMs are reported as the mean number of additional years from disorder onset to first treatment contact for a particular group—for example, women—compared with a reference group—in this case, men.

Fourth, summary statistics were used to describe treatment seeking and delays in treatment seeking for helpful treatment received after first treatment contact. Fifth, Cox regression analysis was used to examine predictors of helpful treatment received, and a GLM with a gamma distribution and a log link function was used to study predictors of delay in helpful treatment received; both analyses were conducted among the subsample of respondents with any mental disorder, respectively, who made a treatment contact and who received helpful treatment for their disorder.

Results

The proportion of respondents with a lifetime treatment contact ranged from 16.7% for persons with a substance use disorder to 81.8% for persons with a mood disorder (Table 1). The proportion of respondents with a lifetime treatment contact was 55.3% for persons with an impulse-control disorder and 60.9% for persons with an anxiety disorder. Rates of lifetime treatment contact by individual disorder were highest for persons with dysthymia (91.4%) and lowest for persons with alcohol abuse (6.5%).

The percentage of respondents who made prompt initial treatment contact (during or before the year of onset) ranged from 16.5% for persons with an impulse-control disorder to 68.1% for persons with a mood disorder. This proportion varied substantially within the main groups of disorders, from 12.1% for ADHD to 68.9% for major depression. Respondents with mood disorders, panic disorder, generalized anxiety disorder, and drug abuse delayed treatment by a median of zero years, compared with a median delay of 12 and 19 years, respectively, among persons with social and specific phobia (Table 1).

The most consistent predictors of lifetime treatment contact among people with a mental disorder were age and age at onset (Table 2). Age was significantly related to lifetime treatment contact for all main groups of mental disorders (mood, anxiety, substance use, and any disorder). Older age groups had significantly lower odds of treatment contact than respondents aged 18 to 24, and among persons with an anxiety disorder or any mental disorder, this effect became especially apparent more than two years after disorder onset.

Respondents with onset before age 30 were also significantly less likely to make treatment contact than respondents with later onset for all four disorders. This effect became especially apparent in the first two years after the onset of the disorders, except substance use disorder.

The predictors gender and education showed less consistent findings. Women had significantly higher chances than men of making treatment contact for any mental disorder, but this pattern was not seen for mood, anxiety, and substance use disorders. Professionals with higher-level vocational education and university graduates were significantly less likely than persons with a primary or basic vocational education to make treatment contact for a substance use disorder but not for a mood or anxiety disorder.

Age and age at onset were the most consistent predictors of delay in initial treatment contact (Table 3). Older age groups had significantly longer treatment delays for all groups of mental disorders. For persons with a mood disorder, additional treatment delay (mean additional years of delay) varied between .70 for those between 25 and 34 and 5.98 for those between 55 and 64. Additional years of delay among persons with an anxiety or a substance use disorder were even larger. Respondents with earlier ages at onset for all groups of mental disorders also had significantly longer delays. Additional years of treatment delay for persons with a mood disorder varied between 1.92 for those with disorder onset between 20 and 29 and 17.00 for those with the earliest onset (before age 13). The delays were longer among persons with an anxiety disorder and peaked at 20.48 years among those with the earliest disorder onset.

The findings for gender were less consistent. Women had a shorter treatment delay than men for all four disorders, but the difference was significant only among men and women with any mental disorder. Education level was not significantly related to delay (Table 3).

The proportion of respondents who received helpful treatment ranged from 33.5% for those with a substance use disorder to 69.5% for those with a mood disorder (Table 4). Among individual disorders, the percentage of persons who received lifetime helpful treatment was highest for major depression (70.1%) and lowest for oppositional defiant disorder (16.2%).

The proportion of respondents who received helpful treatment in the year of initial treatment contact ranged from 40.3% for persons with an impulse-control disorder to 64.7% and 66.4%, respectively, for persons with a mood disorder and a substance use disorder. Median years of delay between initial treatment contact and receipt of helpful treatment among respondents who got helpful treatment were very low (median=0 for persons with a mood, anxiety, or substance use disorder, and median=1 for persons with an impulse-control disorder) (Table 4).

Women were slightly more likely than men to receive helpful treatment (Table 5). Age and education were not significantly related to receipt of lifetime helpful treatment. Respondents with earlier ages at onset had lower odds of getting helpful treatment for their disorder, but this difference was significant only for the category of earliest onset.

Despite the fact that women were more likely to report helpful treatment than men, they experienced somewhat longer delays than men before receiving it. Age was not strongly related to delays of helpful treatment. Only persons aged 45 to 54 experienced significantly longer delays than persons aged 18 to 24. Respondents with earlier ages at onset also experienced significantly longer delays. Additional treatment delay varied between 1.19 years for persons with disorder onset between ages 20 and 29 to 3.80 years for persons with the earliest onset (before age 13). Education was not significantly related to delay of helpful treatment.

Discussion

This study confirms earlier findings that a majority of people with a mood or anxiety disorder had used services for their mental health problems at some point during their lives. It also strengthens earlier statements (1) that suggested—without quantified results—that predictors of never making a treatment contact and of treatment delay are very similar. New findings are the relatively high rates of helpful treatment received among those with a mood or anxiety disorder and the relatively short delays in receiving helpful treatment for any mental disorder after initial treatment contact.

Limitations

Although the NEMESIS-2 sample was representative of the Dutch population on most parameters, people with an insufficient mastery of Dutch, those with no fixed address, and institutionalized people were underrepresented. Hence, our findings are not generalizable to these groups, a problem common to most population studies on service use.

Initial treatment contact was assessed broadly. Respondents could have talked to a medical doctor or other professional about the disorder under investigation. This means that this study was not strictly about treatment seeking for specialty care but about seeking treatment in general for specific mental disorders.

Recall problems may have compromised respondents’ estimations of having sought or delayed seeking treatment and of having received helpful treatment. Respondents who did not seek treatment for their mental disorder may have been more likely to forget their symptoms or to perceive them as normal variations of their mood than those who received treatment. We could not investigate this possibility, but in such a case it would probably have resulted in an overestimation of the proportion of respondents with a lifetime mental disorder who made initial treatment contact. People who are asked about the dates of events commonly demonstrate a memory bias called telescoping, in which recent events tend to be perceived as more remote and distant events are perceived as more recent. It is not known if telescoping occurred, but any such bias would probably have resulted in an underestimation of the duration of delays (1,6).

In the CIDI 3.0, no data are available on the nature, intensity, and duration of treatments obtained for specific mental disorders. Given this limitation and increasing evidence that not all treated respondents received minimally adequate treatment for their disorder, the present findings should be interpreted as overestimations of the proportion of respondents who received adequate treatment and underestimations of the delay in receiving adequate treatment. Although it is conceivable that those who did not seek treatment or postponed seeking treatment had mild disorders, it is also possible that persons with serious mental disorders were deprived of help for too long. Moreover, evidence is building that symptoms can worsen when mental disorders, even mild ones, are not treated in time. Therefore, on the basis of these findings, we can't make a judgment about the seriousness of the consequences of not seeking or delaying seeking treatment.

High treatment rates and short delays

Treatment rates were somewhat lower but delays were shorter in this study than in the NCS-R (1). No other developed country besides the United States reports lifetime treatment rates for specific mental disorders (not to be confused with treatment for mental health problems in general). Despite the fact that the treatment rates for mood and anxiety disorders were somewhat lower in the Netherlands than in the United States, these rates were still quite high. Results from the NCS-R indicate that between 88.1% and 94.2% of persons with mood disorders and 50.1% and 95.3% of respondents with anxiety disorders eventually made treatment contact for their particular disorder (1), compared with 75.3% and 91.4% and 40.6% and 90.1%, respectively, of persons in this study. In both studies, rates of treatment contact were lower for substance use disorders and impulse-control disorders.

Compared with data from 15 countries included in the WMH Survey Initiative (6), the findings showed that the Netherlands is among the top three countries with the shortest treatment delays for all three main groups of mental disorders (mood, anxiety, and substance use). The United States is not among these leading countries.

Variation among disorders in initial treatment contact

People with major depression, panic disorder, and generalized anxiety disorder showed high rates of initial treatment contact and short delays in treatment seeking. Reasons for this finding could be that these disorders often manifest themselves by serious, recurrent, or persistent symptoms that interfere with patients' daily activities or are seen as a threat to their health (1,5,14,15). Another explanation is that in recent decades many initiatives have been undertaken to increase the recognition and treatment awareness of depressive disorders among patients as well as general practitioners.

In line with earlier research (1,15) people with a specific or social phobia showed lower rates of initial treatment contact and longer delays in treatment seeking compared with persons with other anxiety disorders. This could be explained by the generally early onset of phobias and the fact that youngsters are dependent on their parents or other adults for initiating a first treatment contact, the fewer functional impairments associated with specific phobias (5), and a fear of providers or treatments that involve social interaction, such as group therapy and waiting rooms, which can be an expression of social phobias (1,15).

The low lifetime treatment rates and long treatment delays among those with alcohol dependence are consistent with earlier research (1,15). Addicted people often deny their problems and feel a strong resistance to seek treatment. A new finding is the low rate of helpful treatment reported by those with substance use disorders. This could be related to the chronicity of addiction and the paucity of effective options for treatment of addictions.

Despite the fact that respondents with childhood ADHD were more likely to make a treatment contact for their disorder than those with oppositional defiant disorder, they had somewhat longer treatment delays after disorder onset. This finding was consistent with earlier research (1,15). The difference in treatment rates could be the result of the increasing public awareness of ADHD and the perception, both on the part of patients and of society at large, that behavioral problems are less the domain of mental health care than of other sectors, such as social services or justice (1). The difference in duration of treatment delays might be explained by the fact that adults are more troubled by the behavior of children with oppositional defiant disorder than with ADHD.

Sex differences in treatment seeking

Consistent with previous research, women had significantly higher odds than men of making treatment contact for any mental disorder (1,6). A plausible explanation is that women more often recognize vague symptoms, such as stress, as mental health problems (16). Women also had a shorter treatment delay. A sex difference in prompt treatment seeking was seen less consistently in other studies (1,3,17). A new finding is that women were slightly more likely than men to report helpful treatment. This could be explained by the higher prevalence rates of mood and anxiety disorders among women and the greater availability of effective treatments for these disorders than for externalizing disorders, which are more common among men.

Cohort differences in treatment seeking

Consistent with previous research, older age groups had a significantly lower likelihood of treatment contact (1,6) and longer treatment delays (1,3,5,17). It is conceivable that older people are less positive about mental health care (7), are less exposed to public campaigns initiated to increase people's knowledge of mental health problems and of the availability of effective treatment opportunities, and took less advantage of the expansion of mental health services in recent decades than younger adults.

Onset of disorder and treatment seeking

Consistent with previous research, respondents with earlier ages at onset were significantly less likely to make treatment contact (1,6) and had significantly longer treatment delays (1,3,5,17). A frequently cited explanation is that youngsters are dependent on their parents or other adults for realizing a first treatment contact, and recognition of mental disorders is often low among these adults, unless symptoms are extreme (1). Another possibility is that children and adolescents with long-lasting mental health problems view their problems as normal variations of mood or may have learned to cope with them and, as a consequence, do not perceive a need to seek professional help for their mental health problems (1,18). Another important factor is the lower availability of child mental health services.

A new finding is that respondents with an earlier age at onset benefited less often from professional care for mental disorders and received it less promptly. It is conceivable that mental disorders are more difficult to detect among children than adults because of the challenges in unraveling the normal upheaval of childhood, such as transition to another school and first love or work experience, from a mental illness. So when mental health problems in childhood arise, it might take some time to find the proper therapy.

Educational differences in treatment seeking

In all analyses, only one educational difference was significant: people with higher-level vocational education and university graduates were less likely than persons with a primary or basic vocational education to make treatment contact for a substance use disorder. A possible explanation is that people with less education more often have somatic illnesses and as a consequence more promptly seek help when substance use problems arise. This finding was partly in line with findings from the NCS-R (1) indicating that people who are more highly educated are more likely to make treatment contact for major depression and panic disorder but are less likely to make initial treatment contact for substance use disorders. Future research should study to what extent educational differences exist in the nature, intensity and duration of treatments received for specific mental disorders, because there is some evidence that less highly educated people might benefit less from certain types of therapy (19).

Predictors of treatment seeking and treatment delay

Predictors of lifetime treatment contact and of treatment delay were similar. These results were in line with earlier suggestions that were made—without quantified results—on the basis of the NCS-R (1). The same predictors were found for receiving treatment perceived as helpful and for delay of helpful treatment, although the results were less significant, probably because of the relatively small number of respondents who got helpful treatment. This means that men, older cohorts, and respondents with an earlier age of onset were at a disadvantage at different times in the help-seeking process, from first treatment contact to receipt of helpful treatment.

Conclusions

Notwithstanding the high probability of lifetime treatment contact and the relatively short delay in seeking treatment after disorder onset in the Netherlands, at least four groups appeared to remain unnecessarily out of reach of professional care or to wait too long for treatment. They included men, persons aged 55 to 64, persons with a substance use disorder, and persons with disorder onset before age 13. These groups, as well as parents of children with a mental disorder, might have an interest in better information about how to recognize symptoms of mental disorders at an early stage and about the effective treatment options available.