Social Psychiatry and Psychiatric Epidemiology

, Volume 46, Issue 5, pp 403–411

Development and validation of a 6-day standard for the identification of frequent mental distress

Authors

    • Department of Veterans AffairsCenter of Excellence for Suicide Prevention, VAMC Canandaigua
    • Department of PsychiatryUniversity of Rochester
  • Hua He
    • Department of PsychiatryUniversity of Rochester
  • Cynthia A. Claassen
    • Department of Veterans AffairsCenter of Excellence for Suicide Prevention, VAMC Canandaigua
    • Department of PsychiatryUniversity of Rochester
  • Kerry Knox
    • Department of Veterans AffairsCenter of Excellence for Suicide Prevention, VAMC Canandaigua
    • Department of PsychiatryUniversity of Rochester
  • Xin Tu
    • Department of PsychiatryUniversity of Rochester
Original Paper

DOI: 10.1007/s00127-010-0204-4

Cite this article as:
Bossarte, R.M., He, H., Claassen, C.A. et al. Soc Psychiatry Psychiatr Epidemiol (2011) 46: 403. doi:10.1007/s00127-010-0204-4

Abstract

Purpose

The goals of the current study were to assess the concurrent validity of a single-item measure of general mental distress with established, multi-item mental health measures used in population-level surveillance and to establish the optimal cutpoint for determining psychological distress (previously identified as frequent mental distress) using recently available data from the Behavioral Risk Factor Surveillance System survey.

Methods

Data for this study were obtained from the core questionnaire and two optional modules available as part of the 2006 and 2007 Behavioral Risk Factor Surveillance System (BRFSS) surveys. Frequent mental distress (FMD) was identified by the number of days of self-reported poor mental health during the last 30 days. Comparisons of the number of days with poor mental health and positive scores for measures of depression and serious mental illness were calculated to identify the most efficient cutpoint for establishing FMD.

Results

Comparisons of results obtained from ROC analyses using the PHQ-8 and K6 reported 0.867 (95% CI 0.861–0.872) and 0.840 (95% CI 0.836–0.845) of the area under the curve, respectively, suggesting good accuracy. Using the Youden index, 6 days of poor mental health in the past 30 days, rather than the existing 14-day standard, was identified as the point at which the sum of the sensitivity and specificity was greatest.

Conclusion

Results from this study suggest that a 6-day standard (FMD-6) can be used as a valid and reliable indicator of generalized mental distress with strong associations to both diagnosable depressive symptomology and serious mental illness.

Keywords

SurveillanceMental healthPsychological distressMethodology

Introduction

An estimated 25 million U.S. adults suffer from some form of mental illness [1]. Estimates of the economic burden in lost earnings associated with mental illness exceed 190 billion dollars annually [2] and increase to more than 300 billion when extended to incarcerated and homeless populations [3]. The tremendous social and economic burden associated with mental illness reaches beyond lost wages and institutional costs. According to estimates calculated from the Global Burden of Disease Study [4], mental illness is a leading contributor to global morbidity, second only to cardiovascular disease. Despite the overwhelming social and economic burden associated with mental illness, data on the prevalence or correlates of mental disorders among the U.S. adult population are rare.

Several measures of general and clinically relevant psychiatric symptomatology have been validated for use in population-level studies. However, while these empirically validated, multi-item measures of psychological distress are available and used in certain population-level studies, the length and expense of multiple-item measures makes them prohibitive for many surveillance and research activities. For example, the Centers for Disease Control and Prevention’s (CDC) Behavioral Risk Factor Surveillance System (BRFSS), which is the largest public health surveillance system in the U.S., limits its use of multiple-item indicators to topic-specific, optional survey modules that are only administered to certain subsamples of survey participants [5].

In the absence of empirically validated multiple-item measures of mental health symptomology, population-wide estimates of the prevalence and correlates of psychological problems in the BRFSS have been based on abbreviated measures, such as the single-item quality of life-derived frequent mental distress (FMD). Conceptualized as a report of 14 or more days of poor mental health in the 30 days preceding survey administration [69], FMD is one of four health-related quality of life (HRQoL) measures routinely available in BRFSS data, and is also used in some other population-based surveys. While not specific to mental illness, these measures of HRQoL have been shown to have excellent test–retest reliability [10], have been validated in other populations [11], and are strongly correlated to a number of behavioral and health-related outcomes [1214].

Despite their utility as indicators of general psychological distress, none of the single item HRQoL-derived measures has been validated against existing multi-item measures. Further, the decision to use a 14-day cutoff to designate frequent mental distress was originally based on the number of days required to diagnose clinically significant depressive symptomology, and has not been informed by comparisons with other available measures of mental health. If such a criterion were derived empirically, the resulting threshold could be used to provide needed information on the prevalence and correlates of poor mental health among all survey participants, facilitate identification of emerging risk populations, and provide needed information for the development of prevention programs among indicated populations. More generally, comparisons between multi-item measures of psychopathology and FMD would provide empirical evidence of the relationship between number of days of self-reported poor mental health and established indicators of specific diagnoses, including depression. The goals of the current study were to assess the concurrent validity of this single-item measure of general mental distress with established, multi-item mental health measures used in population-level surveillance and to establish the optimal FMD cutpoint for determining psychological distress using BRFSS data.

Multi-item population-level mental health measures

The nine-item Patient Health Questionnaire (PHQ-9), is a measure of depressive symptomology used in population-level studies, and can be scored to identify mild–severe depression with treatment recommendations corresponding to the level of symptom severity [15]. Scores on the PHQ-9 can range from 0 to 27, with a cutpoint of 10 suggestive of moderate to severe depression [15]. A previous assessment of the epidemiologic properties of the PHQ-9 reported a sensitivity of 88% for major depression with identical specificity (88%) using a cutpoint of 10 or greater [15]. Studies using a cutpoint of 10 or greater have estimated the prevalence of depressive symptoms to range between 2.5 and 38%, depending on the characteristics of the study population [16]. The PHQ-9 has been widely adopted in clinical [16] and epidemiologic studies [1720], and has been translated into multiple languages, including Spanish, German, Dutch, Italian, and Thai [2124]. An alternate 8-item (PHQ-8) version of the Patient Health Questionnaire, excluding a single-item question of thoughts of death or self-harm, has also been developed and has been shown to have almost identical sensitivity, specificity, and positive predictive value using a recommended cutpoint of ≥10 [15]. In 2006, the PHQ-8 was administered to a random-digit dial sample of non-institutionalized U.S. adults and an estimated 8.6% of participants were identified as having a score of 10 or greater [25].

Another mental health measure designed for use in the general population is the 6-item K6 measure of non-specific serious mental illness (SMI). Also developed for use for population-level surveys, the K6 is a self-report measure of clinically relevant symptom clusters consistent with diagnoses of serious mental illness (e.g., schizophrenia, bipolar I, other psychotic illness) as described in the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) [26]. Comparisons of the K6 and more detailed 10-item (K10) measures with blinded clinical assessments using the Structured Clinical Interview (SCID) resulted in the recommendation of a cutoff score of greater or equal to 13 as the most efficient screen for symptoms consistent with the presence of serious mental illness (SMI). Using this method, the K6 was estimated to have a sensitivity of 0.36 and specificity of 0.96 [27]. When administered to a telephone sample in the Boston area, scores on the K6 showed an estimated 12-month SMI prevalence of 7.1% among adults of age 18 years and older [27]. Additional estimates of the 12-month prevalence of symptoms consistent with SMI have ranged between 6.1 and 17.1%, depending on sample characteristics and precipitating events [28, 29].

Method

Data for this study were obtained from the core questionnaire and two optional modules available as part of the 2006 and 2007 Behavioral Risk Factor Surveillance System (BRFSS) surveys. Coordinated by the CDC, BRFSS utilizes a nationally representative sample of non-institutionalized adults to collect data annually from all U.S. states, Washington D.C., Guam, the U.S. Virgin Islands, and Puerto Rico. In 2006 and 2007, the median response rates were 51.4% (range 35.1–66.0%) and 50.6% (range 26.9–65.4%), respectively [30, 31]. Data on depression and serious psychological distress were obtained from two topic-specific optional modules incorporating the above-described, previously validated measures of mental health (PHQ-8; K6). PHQ-8 data on symptoms of depression were obtained from the 2006 Anxiety and Depression module administered as part of the BRFSS in 39 states. K6 data on symptoms consistent with a diagnosis of a serious mental illness were obtained from measures available in the 2007 Mental Illness and Stigma module administered as part of the BRFSS in 26 states. A complete list of participating states for each year and detailed information on BRFSS methodology are available from the CDC website (http://www.cdc.gov/brfss). The PHQ-8 included questions about the frequency of symptoms of depression during the past 2 weeks prior to survey administration with response options ranging from 0 to 14 days. Specifically, respondents are asked to identify the number of days during the past 2 weeks that they felt “little pleasure in doing things”, “down depressed or hopeless”, “tired or had little energy”, “bad about yourself or that you were a failure or had let yourself or your family down”, had trouble falling asleep or staying asleep or sleeping too much”, “had a poor appetite or ate too much”, “had trouble concentrating on things, such as reading the newspaper or watching TV”, or “moved or spoken so slowly that other people could have noticed. Or, the opposite, being so fidgety or restless that you were moving around a lot more than usual”. Consistent with previous use [25], the number of reported days was recoded into the following categories: 0–1 day = 0; 2–6 days = 1; 7–11 days = 2; 12 days or more = 3. The PHQ-8 was scored by taking the sum for all responses, and a score of 10 or greater was considered “positive”. The K6 included questions about the frequency of serious psychological distress in the past 30 days. Response categories included none, a little, some, most or all of the time. Respondents were asked to identify how often during the past 30 days they felt nervous, hopeless, restless or fidgety, “so depressed nothing could cheer you up”, “that everything was an effort”, or worthless. The K6 was scored by converting responses into the following categories: none = 0, a little = 1, some = 2, most of the time = 3, all of the time = 4. A score of 13 or more was considered to be “positive”.

Depressive symptomology was identified in this study using a PHQ-8 cutoff score of greater than or equal to 10. Serious mental illness was identified with a K6 cutoff score equal to or greater than 13. Frequent mental distress (FMD) was identified by the number of days of self-reported poor mental health reported in response to the question, “Now thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good?” Comparisons of the number of days with poor mental health, depression (PHQ-8), and serious psychological distress (K6) were calculated to identify the optimal cutpoint for establishing mental distress in the last 30 days.

To establish the convergent validity of FMD for assessing mental health status, comparisons of the number of days with poor mental health and other empirically validated measures of psychological distress were calculated using receiver operating characteristic (ROC) curve analyses, calculations of sensitivity, specificity, positive predictive value, negative predictive value, and logistic regression. ROC curve analyses can be used to measure the overall discrimination ability of a continuous outcome indicator to correctly classify those with and without the condition. Within our context, ROC curves were constructed by plotting the sensitivity on the Y-axis and (1 − specificity) on the X-axis, while the sensitivity and specificity were calculated for each value of the number of days with poor mental health in the past month. The most important summary index of the ROC curve is the area under the ROC curve (AUC). An ROC curve with AUC > 0.5 means the test is informative in the sense that it is better than classifying subjects in a completely random fashion. In general, the closer the curve to the upper-left corner [point (0, 1)], the bigger the AUC and the more accurate the test. For each possible threshold based on the sample, we computed estimates of the corresponding sensitivity and specificity, as well as the positive predictive value and negative predictive value [32]. In this study, separate ROC curve analyses were performed to determine how well FMD discriminated depressive symptomology (PHQ-8 ≥ 10) and serious mental illness (K6 ≥ 13).

Validity for the number of days with poor mental health and depressive symptomology (PHQ-8 ≥ 10) or serious mental illness (K6 ≥ 13) was assessed by calculating estimates of sensitivity and specificity for each unique value for the number of days with poor mental health during the last 30 days (range 0–30). Sensitivity is an estimate of the ability of a test or screen to correctly identify those who have a specific condition (i.e., poor mental health within our context) [33]. Specificity, in contrast, represents the ability of a test or measure to correctly identify those who are negative when the condition is not present [33]. The optimal cutpoint for identifying frequent mental illness was identified by maximizing the Youden index, the sum of sensitivity, and specificity [34]. The positive and negative predictive values for the number of days in the past 30 with poor mental health were also calculated. The positive predictive value (PPV) is an estimate of the proportion of subjects with the condition among participants who are identified as being positive [33]. Similarly, the negative predictive value (NPV) is an estimate of the proportion of those without the condition among participants who are identified as having a negative value on a test or outcome [33]. Unlike sensitivity and specificity, PPV and NPV are a function of the underlying prevalence of the condition in the population. Consistent with estimates obtained from previous population-based studies, PPV and NPV were calculated using prevalence estimates of 8.6% [25] for the PHQ-8 measure of depressive symptomatology and 7.1% [27] for the K6 measure of serious mental illness.

To identify similarities or differences in the associations between each measure of poor mental health and key demographic and economic characteristics, four multivariate logistic regressions were conducted separately for each measure of poor mental health, controlling for individual characteristics, including age (18–29, 30–39, 40–49, 50–59, 60–69, and 70+ years), gender, race/ethnicity (non-Hispanic white, non-Hispanic black, non-Hispanic multiracial, non-Hispanic other, and Hispanic), education (high school diploma or less vs. some college or more), marital status (married/cohabitating vs. single/divorced/widowed/separated), and physical health status (good or better vs. fair or poor). Odds ratios were used to assess the relationships between measures of mental health and the covariates considered.

All analyses conducted as part of this study used information from the de-identified dataset available from the CDC website analyzed with SAS (ver. 9.2) and SPSS (ver. 16.0) whenever appropriate. Internal IRB approval for the use of BRFSS data was obtained from the Department of Veterans Affairs.

Results

In 2006, a total of 170,814 (48%) BRFSS participants provided data for both the PHQ-8 measure of depressive symptomology and number of days with poor mental health. In 2007, a total of 141,560 (33%) BRFSS participants provided data for both the K6 measure of serious mental illness and number of days with poor mental health (FMD). Among those participating in the 2006 Anxiety and Depression module, 8.95% (95% CI 8.62–9.27) had a PHQ-8 score equal to or greater than 10, indicating the presence of moderate–severe depressive symptomology. In 2007, 3.96% (95% CI 3.73–4.20) of those participating in the mental illness and stigma module had a K6 score greater than or equal to 13, suggesting the presence of symptoms consistent with a serious mental illness [26, 27].

Comparisons of the AUCs using the PHQ-8 (Fig. 1) with a cutpoint of greater than or equal to 10 as the gold standard test and the K6 (Fig. 2) with a cutpoint greater than or equal to 13 as the gold standard test reported 0.867 (95% CI 0.861–0.872) and 0.840 (95% CI 0.836–0.845) of the area under curve, respectively, suggest good accuracy [35]. While the number of days with poor mental health (FMD) had similar accuracy when compared to both the PHQ-8 and K6 measures of depressive symptomology and serious mental illness, the AUC was significantly higher for number of days with poor mental health when compared to the PHQ-8 measure of depressive symptomology.
https://static-content.springer.com/image/art%3A10.1007%2Fs00127-010-0204-4/MediaObjects/127_2010_204_Fig1_HTML.gif
Fig. 1

ROC curve analysis for poor mental health and PHQ-8

https://static-content.springer.com/image/art%3A10.1007%2Fs00127-010-0204-4/MediaObjects/127_2010_204_Fig2_HTML.gif
Fig. 2

ROC curve analysis for poor mental health and K6

Table 1 presents results from calculations of sensitivity, specificity, PPV, and NPV separately for the number of days of poor mental health in the past 30 days and depressive symptomology (PHQ-8) and serious mental illness (K6). The optimal cutpoint for a dichotomous measure of frequent mental distress was determined by identifying the point at which the sum of sensitivity and specificity was greatest. Positive and negative predictive values were calculated using the prevalence estimates for the PHQ-8 and K6 obtained from results of previous studies [25, 26]. As shown in Table 1, 4 and 5 days of poor mental health were the points at which the sum of the sensitivity and specificity was greatest when a positive score for the PHQ-8 was used as the gold standard. When a positive value for the K6 was used as the gold standard, 6 days of poor mental health in the past 30 days was identified as the point at which the sum of the sensitivity and specificity was greatest. Importantly, the summary measure used to identify the optimal cutpoint for identifying mental distress was significantly lower than the current 14 day standard. Based on these calculations, a cutpoint of 6 or more days of poor mental health in the past 30 days was identified as the optimal indicator of mental distress. Overall, 15.54 and 15.22% of BRFSS participants reported 6 or more days of poor mental health in the past 30 days prior to survey completion in 2006 and 2007, respectively.
Table 1

Sensitivity and specificity of the number of days with poor mental health

No. of days

Comparison with PHQ-8a

Comparison with K6b

Sensitivity

Specificity

Sum

PPV

NPV

Sensitivity

Specificity

Sum

PPV

NPV

0

1.00

0.00

1.00

0.09

1.00

0.00

1.00

0.07

0.99

1

0.82

0.73

1.54

0.23

0.98

0.86

0.71

1.57

0.19

0.99

2

0.80

0.76

1.57

0.25

0.97

0.86

0.74

1.60

0.21

0.98

3

0.78

0.82

1.60

0.30

0.97

0.84

0.80

1.63

0.24

0.98

4

0.75

0.85

1.61

0.34

0.97

0.82

0.83

1.65

0.27

0.98

5

0.74

0.87

1.61

0.36

0.97

0.81

0.84

1.65

0.28

0.98

6

0.69

0.90

1.59

0.42

0.97

0.78

0.88

1.66

0.33

0.98

7

0.68

0.91

1.59

0.42

0.97

0.77

0.88

1.65

0.34

0.98

8

0.65

0.92

1.57

0.45

0.96

0.75

0.90

1.64

0.36

0.98

9

0.64

0.92

1.56

0.45

0.96

0.74

0.90

1.64

0.36

0.98

10

0.64

0.92

1.56

0.45

0.96

0.74

0.90

1.64

0.36

0.97

11

0.58

0.94

1.52

0.51

0.96

0.70

0.92

1.62

0.41

0.97

12

0.57

0.94

1.52

0.51

0.96

0.69

0.92

1.62

0.41

0.97

13

0.57

0.95

1.51

0.51

0.96

0.69

0.92

1.61

0.41

0.97

14

0.57

0.95

1.51

0.51

0.96

0.69

0.92

1.61

0.41

0.97

15

0.54

0.95

1.49

0.52

0.95

0.67

0.93

1.60

0.42

0.97

16

0.44

0.96

1.41

0.55

0.95

0.58

0.95

1.53

0.47

0.97

17

0.44

0.97

1.41

0.56

0.95

0.58

0.95

1.53

0.47

0.97

18

0.44

0.97

1.41

0.56

0.95

0.58

0.95

1.53

0.47

0.97

19

0.44

0.97

1.40

0.56

0.95

0.57

0.95

1.52

0.47

0.97

20

0.44

0.97

1.40

0.56

0.95

0.57

0.95

1.52

0.47

0.97

21

0.37

0.97

1.34

0.57

0.94

0.50

0.96

1.46

0.49

0.96

22

0.37

0.97

1.34

0.57

0.94

0.49

0.96

1.45

0.49

0.96

23

0.36

0.97

1.34

0.57

0.94

0.49

0.96

1.45

0.49

0.96

24

0.36

0.97

1.34

0.57

0.94

0.49

0.96

1.45

0.49

0.96

25

0.36

0.97

1.34

0.57

0.94

0.49

0.96

1.45

0.49

0.96

26

0.34

0.98

1.31

0.57

0.94

0.46

0.96

1.42

0.49

0.96

27

0.34

0.98

1.31

0.57

0.94

0.45

0.96

1.42

0.49

0.96

28

0.33

0.98

1.31

0.57

0.94

0.45

0.96

1.42

0.50

0.96

29

0.33

0.98

1.30

0.58

0.94

0.44

0.97

1.41

0.50

0.96

30

0.32

0.98

1.30

0.57

0.93

0.44

0.97

1.40

0.50

0.96

aThe prevalence used for computation of PPV and NPV was 9.1%

bThe prevalence used for computation of PPV and NPV was 7.2%

Positive and negative predictive values were calculated for each unique value for the number of days of poor mental health in the past 30 days using prevalence estimates obtained from previous population-based surveys. As shown in Table 1, scores for the PPV were moderate when using a cutpoint of 6 days or more. Specifically, the PPV was 0.42 when compared to a positive score on the PHQ-8 and 0.33 when compared to a positive score on the K6. However, the NPV scores were markedly higher with estimates of 0.97 and 0.98 when compared to the PHQ-8 and K6.

Results from multivariate logistic regression models for 6 and 14-day indicators of mental distress, depressive symptomology, serious mental illness, and key demographic variables are shown in Table 2. Overall, results from these analyses suggest that all four indicators of poor mental health are similarly associated with all demographic measures and indicators of physical health status included in this study. Overall, poor mental health was higher among younger age groups when compared with participants over the age of 70 years. Males were significantly less likely than females to score above the identified cutoff score for all four measures. Those with at least some college, married or cohabitating, or who reported good or better health status were also significantly less likely than the comparison groups to have a score above the cutoff point for any of the four measures. There was some disparity in the racial and ethnic associations. There were no statistically significant associations between serious mental illness (K6 ≥ 13) and those who were non-Hispanic black, non-Hispanic other, or Hispanic. When the associations between key demographic measures and physical health status were compared, the only statistically significant differences between the 6 and 14-day standards of mental distress were for those between the ages of 18 and 29 years those who reported good or better physical health.
Table 2

Multivariate associations between indicators of poor mental health and individual characteristics

 

Mental distress (6+ days)

Mental distress (14+ days)

Depressive symptomology (PHQ-8)

Serious mental illness (K6)

OR

95% CI

OR

95% CI

OR

95% CI

OR

95% CI

18–29 years

4.42

4.25–4.60

3.99

3.80–4.18

5.14

4.75–5.57

4.23

3.74–4.77

30–39 years

4.23

4.08–4.38

4.15

3.98–4.33

4.92

4.56–5.30

4.12

3.68–4.60

40–49 years

3.81

3.61–3.93

3.83

3.69–3.99

4.65

4.33–4.98

4.26

3.86–4.71

50–59 years

3.25

3.15–3.56

3.38

3.26–3.51

3.93

3.67–4.20

3.82

3.48–4.20

60–69 years

1.88

1.81–1.94

1.93

1.86–2.01

1.91

1.77–2.05

1.87

1.69–2.08

Male

0.64

0.63–0.65

0.68

0.66–0.69

0.62

0.60–0.64

0.81

0.77–0.86

Non-Hispanic black

0.73

0.71–0.76

0.71

0.68–0.74

0.82

0.77–0.87

1.04

0.96–1.14

Non-Hispanic multiracial

0.84

0.82–0.87

0.81

0.78–0.84

0.87

0.82–0.93

0.88

0.81–0.96

Non-Hispanic other

1.35

1.27–1.44

1.39

1.30–1.49

1.52

1.38–1.68

1.04

0.89–1.21

Hispanic

0.94

0.90–0.99

0.94

0.89–0.99

0.97

0.88–1.06

0.89

0.78–1.01

College education

0.83

0.81–0.84

0.78

0.77–0.80

0.66

0.64–0.69

0.53

0.50–0.56

Married/cohabitating

0.63

0.62–0.64

0.60

0.59–0.62

0.57

0.55–0.59

0.50

0.47–0.53

Good health

0.21

0.20–0.21

0.18

0.17–0.18

0.13

0.13–0.14

0.13

0.12–0.18

Reference group is 70+ years of age, female, non-Hispanic white, high school education or lower, single, fair or poor health

Odds ratios for depressive symptomology and serious mental illness were calculated for both the 6 and 14-day cutpoint for number of days with poor mental health. In 2006, the odds ratio of having depressive symptomology above the threshold value of 10 was 20.60 (95% CI 19.82–21.51) for those with 6 days or more days of poor mental health and the odds ratio was 22.54 (95% CI 21.68–23.44) for those with 14 days or more days of poor mental health in the last 30 days. In 2007, the odds ratio of having serious mental illness was 25.24 (95% CI 23.69–26.89) for those with 6 or more days of poor mental health and 26.80 (95% CI 25.28–28.42) for those with 14 or more days of poor mental health.

Discussion

The two objectives of this study were first to provide empirical evidence of the associations between a quality of life derived single-item measure of mental distress and existing multi-item measures of poor mental health and second to assess the validity of the existing 14 day standard for the identification of mental distress. Findings from this study suggest that a single-item measure derived from a question designed to measure mental health-related quality of life can be used as a valid indicator of mental distress. Further, comparisons of the epidemiologic characteristics of the number of days of poor mental health and previously validated measures of depressive symptomology and serious mental illness suggest that the existing 14-day standard used to categorize frequent mental distress may not be the optimal cut point for the identification of clinically significant symptoms of mental disorders.

The quality of life derived measure used in this study is currently available as a core item in the Behavioral Risk Factor Surveillance System survey and in some other large population surveys. The validation of a single item measure of mental health that is congruent with multi-item measures of depression and serious mental illness is important to understanding the epidemiology of symptoms of psychological distress at the population level. An empirically validated measure of mental distress that is regularly available in existing data sources has the potential to contribute to improved surveillance of poor mental health among priority populations, identification of emerging risk groups and characteristics, and the development of etiologic research. While the use of a single-item measure such as the proposed 6-day standard for identification of frequent mental distress (FMD-6) may not be appropriate for clinical screening, the cost associated with the development of population surveys and declining survey participation rates support the use of single-item indicators at this level of surveillance.

The prevalence of mental distress, as identified by the FMD-6, was approximately 15% for both years of data included in this study. While this number is notably higher than the estimated prevalence for either depressive symptomology or serious mental illness alone, the calculated specificity and PPV for this measure suggest that a percentage of those who have a positive score for the FMD-6 may not meet established criteria for either of these two more explicit measures of mental health while still experiencing frequent mental distress. No attempt was made in these analyses to identify and test for other types of diagnosable psychological problems that might be associated with frequent mental distress, and there are obviously mental health issues other than depression or serious mental illness. Prior research conducted in the US population has established an overall, past-month, estimate of the self-reported prevalence of mental disorders to range between 12.3 and 17.3%, depending on age group, suggesting that the estimated FMD-6 BRFSS rate of 15% is well within range [36].

Statistically, it is also important to note the comparatively modest gains in both sensitivity and NPV associated with the use of a 14-day standard. Given the increases in sensitivity and NPV associated with the use of a 6-day standard, the relative burden of a positive score for those who may not meet more stringent criteria should be balanced by the improvement in negative scores for those who may otherwise be identified as having symptoms consistent with clinically significant depression or serious mental illness.

There are several limitations to this study that should be considered. First, all data are based on self-reported symptoms during a specific time frame that have not been corroborated through clinical interviews. It is possible that assessment either in person or with alternate instruments might yield different results. While the BRFSS is designed to yield nationally representative estimates, all states did not participate in either the 2006 Anxiety and Depression or 2007 Mental Illness and Stigma optional modules. It is possible that the results reported in this study are not representative of the prevalence or correlates of poor mental health among the larger population. Data for the PHQ-8 and K6 were collected during different years. Thus, relationships between FMD and either depressive symptomology or serious mental illness were calculated independently using data from the appropriate BRFSS survey year, and all three measures were not available for any one participant. This methodological limitation precluded simultaneous consideration of the associations between FMD, depressive symptomology, and serious mental illness. Finally, the validity of FMD as an indicator of either depressive symptomology or serious mental illness is partially dependent on the epidemiological properties of these instruments. For example, a positive score for the FMD-6 has a PPV of 0.42 for a positive score for moderate–severe depressive symptomology as measured by the PHQ-8. Thus, the ability of FMD to accurately identify clinically relevant depression is mitigated by the validity and PPV of the PHQ-8 as a measure of depression.

In summary, results from this study suggest that a 6-day standard (FMD-6) can be used as a valid and reliable indicator of generalized mental distress with strong associations to both diagnosable depressive symptomology and serious mental illness. Previous assessments of ultra-short or brief measures of depression [37] and other mental disorders [3840] have concluded that these measures perform as well as longer, more detailed assessments and have endorsed the use of such measures for the purpose of identifying populations with conditions such as depression and anxiety [37, 40]. Brief quality of life assessments have also shown strong associations with anxiety and depression when compared to self-report symptom severity [41] or scores obtained using multi-item measures [42]. The availability of a single-item measure of symptoms consistent with depression or SMI in the BRFSS survey and other existing data systems enhances the ability of researchers and public health professionals to identify and assess the burden of poor mental health among large populations without the cost or methodological compromise associated with the use of condition-specific multi-item measures. However, future studies of the validity and epidemiologic qualities of this measure using other data sources are needed. Additional studies of the relationships between FMD and other measures of poor mental health mental are therefore encouraged.

Copyright information

© Springer-Verlag 2010