Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Apr 1.
Published in final edited form as: J Psychiatr Res. 2020 Jan 29;123:54–61. doi: 10.1016/j.jpsychires.2020.01.009

The importance of physical and mental health in explaining health-related academic role impairment among college students

Chelsey R Wilks a,b, Randy P Auerbach c, Jordi Alonso d,e,f, Corina Benjet g, Ronny Bruffaerts h, Pim Cuijpers i, David D Ebert j, Jennifer G Green k, Claude A Mellins c, Philippe Mortier l,m,n, Ekaterina Sadikova b, Nancy A Sampson b, Ronald C Kessler b,*
PMCID: PMC7047531  NIHMSID: NIHMS1558000  PMID: 32036074

Abstract

Research consistently documents high rates of mental health problems among college students and strong associations of these problems with academic role impairment. Less is known, though, about prevalence and effects of physical health problems in relation to mental health problems. The current report investigates this by examining associations of summary physical and mental health scores from the widely-used Short-Form 12 (SF-12) Health Survey with self-reported academic role functioning in a self-report survey of 3,855 first-year students from five universities in the northeastern United States (US; mean age 18.5; 53.0% female). The mean SF-12 physical component summary (PCS) score (55.1) was half a standard deviation above the benchmark US adult population mean. The mean SF-12 mental component summary (MCS) score (38.2) was more than a full standard deviation below the US adult population mean. Two-thirds of students (67.1%) reported at least mild and 10.5% severe health-related academic role impairment on a modified version of the Sheehan Disability Scale. Both PCS and MCS scores were significantly and inversely related to these impairment scores, but with nonlinearities and interactions and much stronger associations involving MCS than PCS. Simulation suggests that an intervention that improved the mental health of all students with scores below the MCS median to be at the median would result in a 61.3% reduction in the proportion of students who experienced severe health-related academic role impairment. Although low-cost scalable interventions exist to address student mental health problems, pragmatic trials are needed to evaluate the effectiveness of these interventions in reducing academic role impairment.

Keywords: College students, health-related academic role impairment, mental health, physical health

1. Introduction

Epidemiological research consistently finds high rates of mental health problems among college students in the United States (US; Cho et al., 2015; Hunt and Eisenberg, 2010; Kendler et al., 2015) and across the world (Auerbach et al., 2016; Auerbach et al., 2019; Auerbach et al., 2018) as well as significant associations of these problems with decrements in academic role performance (Alonso et al., 2018; Alonso et al., 2019; Bruffaerts et al., 2018), and discontinuing college (Arria et al., 2013; Eisenberg et al., 2009). Specific mental health problems such as Attention Deficit Hyperactivity Disorder (ADHD), depression, and sleep disturbances are typically found to be the most important mental disorders in these studies (Merkt and Gawrilow, 2016; Gormley et al., 2019; Hysenbegasi et al., 2005; Gaultney, 2010).

Research has also shown that physical health problems are associated with decrements in academic role performance among college students (Dryer et al., 2016; Ruthig et al., 2011; El Ansari and Stock, 2010). This research is much less extensive and fine-grained than the research on mental health problems, presumably based on the fact that the vast majority of college students are in good physical health. As a result of this fact, a single yes-no measure is often used of either any disability or any chronic condition physical health problem in studying associations between physical health problems and academic performance, whereas more complex multivariate models assessing the joint effects of diverse conditions on academic performance are often used in studies of mental disorders.

Normative data suggest that the prevalence of mental and physical health problems might be more comparable among college students than suggested by the different approaches used to examine their associations with academic performance. Specifically, inspection of the Physical Component Summary (PCS) and Mental Component Summary (MCS) subscales in the Medical Expenditures Panel Survey (MEPs; Cohen et al., 2009), an annual nationally representative sample of the US Civilian non-institutionalized population, shows that mean physical health component scores are only slightly higher than mean mental health component scores among the youngest respondents (ages 20-29) (Hanmer and Kaplan, 2016). Furthermore, the majority of primary care visits on college campuses are for physical health problems rather than mental health problems (Turner and Keller, 2015), although college students tend to underuse psychological services even when they endorse symptoms of a psychiatric disorder (Bruffaerts et al., 2019). Based on these findings, it is plausible to think that physical health problems might be more important relative to mental health problems in affecting the academic performance of college students than implied by the literature. We are unaware, though, of any research that has investigated this issue systematically by comparing either relative prevalence or relative importance of these two sets of health problems in accounting for decrements in the academic performance of college students.

We carried out such an investigation by examining the associations of SF-12 PCS and MCS scores with reports about health-related academic role impairments in a self-report survey of first-year college students from five universities in the northeast US. The students were surveyed as part of the first phase of the WHO World Mental Health Surveys International College Student (WMH-ICS) Initiative (Cuijpers et al., 2019). Prior cross-national WMH-ICS reports documented high lifetime and 12-month prevalence of mental disorders in the US as well as other participating countries (Auerbach, et al., 2018) along with academic role impairment associated strongly with these disorders (Alonso, et al., 2019). Building on this earlier work, we examine the more highly aggregated MCS score rather than measures of specific mental disorders in order to make an even-handed comparison with the single summary measure of physical disorder available in the survey. Both these measures are used to predict student reports of health-related academic role impairment.

2. Methods

2.1. Sample

All incoming first year students in the five participating colleges and universities (n=20,583) were invited to participate in a web-based self-report health survey between October 2017 and March 2019. All but one school invited first year students to participate in October with the other school inviting first year students in March. A total of 3,855 students completed the survey (18.7% response rate). Participants were excluded from analysis if they were under age 18 (n = 7), older than age 22 (n = 68), were currently or previously married (n = 27), and either had a child, were pregnant, or had a pregnant partner (n = 9) based on the rarity of these characteristics in the sample. The final analysis sample included 3,761 respondents. All participants provided written informed consent prior to participation. All study procedures were approved by the human subjects boards of all involved organizations. The investigation was carried out in accordance with the latest version of the Declaration of Helsinki.

2.2. Measures

2.2.1. Physical and mental health:

Physical and mental health in the four weeks before the survey was assessed with the SF-12, a widely-used 12-question self-report scale designed to assess perceived health (Ware et al., 1996). Separate physical health component and mental health component summary scores were constructed from SF-12 responses. The PCS and MCS both have a theoretical range of 0-100, with higher scores indicating better health, and have been normed to have a mean of 50 and a standard deviation of 10 in the total US population (Gandek et al., 1998; Ware et al., 1996; Ware et al., 1995). The SF-12 is an abbreviated measure of the SF-36 and the SF-12 achieved strong R-squares with the original SF-36 on both the PCS (0.91) and MCS (0.94) (Ware et al., 1996). The SF-36, in turn, has been shown to have good construct validity in that the two broad dimensions of physical and mental health found in much previous research was replicated in dimensional analyses of the SF-36 and these dimensional scales were shown to have similar patterns of association as clinician measures of physical and mental health with a wide range of correlates (McHorney et al., 1993). In addition, PCS scores are correlated strongly with objective disorder severity measures among patients with a wide range of physical disorders, whereas MCS scores are correlated strongly with objective disorder severity measures among patients with a wide range of mental disorders (Coons et al., 2000).

2.2.2. Health-related academic role impairment:

Health-related role impairments in the 30 days before the interview was assessed with a revised version of the Sheehan Disability Scale (SDS; Sheehan et al., 1996), a short self-report visual analogue scale of impairments in functioning across multiple role domains. The revised SDS asked respondents to rate on a 0-10 scale the extent to which problems with their health impaired their functioning in each of a series of life domains (e.g., quality of school work, social life, close personal relationships), using a rating system in which a score of 0 was labeled no impairment, scores in the range 1-3 were labeled mild impairment, 4-6 moderate impairment, 7-9 severe impairment, and 10 very severe impairment. We modified the original SDS wording, which combined work and school, to ask separately about each. We focused here on responses to the academic role impairment question, collapsing responses into nested categories of any (1-10 versus 0) and severe (7-10 versus 0-6) in the total sample and subsample estimates of more than mild among those with any health-related academic role impairment (4-10 versus 1-3) and severe among those with more than mild impairment (7-10 versus 4-6). No data as yet exist on the validity of these reports compared to objective academic performance measures, such as grade point average, but comparable studies of objective work performance measures among employed people show that the SDS is one of the most valid self-report scales of work performance (Mateen et al., 2017). Ongoing WMH-ICS methodological studies are collecting comparable data for college students, but results are not yet available.

2.2.3. Control variables:

All models included dummy control variables for schools along with controls for the following socio-demographic variables: age (continuous); gender (male, female, and self-reported “other”); race/ethnicity (Hispanic, Non-Hispanic Black, Non-Hispanic White, other); nativity (nested categories of the student being foreign-born, at least one parent being foreign-born, at least one grandparent being foreign-born, and all grandparents being native-born); and highest parental education (high school or less, some college, college graduate, some post-baccalaureate education, doctorate or other professional degree).

2.3. Analysis methods

The survey data from each college were post-stratified to match the distribution of the entire first-year class on the cross-classification of sex and race/ethnicity in order to adjust for discrepancies between the sample and the population on these variables. These were the only post-stratification variables available across all schools. Item-level missing data were then multiply imputed (MI) using the fully conditional specification method with 20 imputations per respondent (van Buuren, 2007). MI logistic regression analysis was used to estimate the associations of PCS and MCS scores with the four dichotomous measures described above of health-related academic role impairment, controlling for the socio-demographic variables described above. Eight nested logistic models were estimated for each outcome: Separate linear and nonlinear models for PCS predicting the outcome (M1-M2); comparable models for MCS (M3-M4); models for the linear additive (M5) and interactive (M6) associations of PCS and MCS with the outcome; and models for the nonlinear additive (M7) and interactive (M8) associations of PCS and MCS with the outcome. Nonlinearities were modeled as incremental regression splines for the lowest and highest quartiles of the PCS and MCS distributions in addition to linear terms. Differences in comparative model fit were evaluated with MI-adjusted likelihood ratio χ2 tests. The logits and logits +/− two standard errors of best-fitting models were exponentiated and presented as odds-ratios (ORs) with 95% confidence intervals (95% CIs). Statistical significance was evaluated consistently using .05-level two-sided MI-adjusted tests.

To aid in the interpretation of the interactive spline models, we calculated population attributable risk proportions (PARPs; Greenland and Drescher, 1993) for best-fitting models to estimate the effects of hypothetical interventions to improve respondent scores on either the PCS or MCS while holding the other score constant. Such estimates assume provisionally that PCS and MCS scores are causal risk factors for academic impairment and that the effects of hypothetical interventions to increase these scores are captured by the logistic regression coefficients. These simulations were carried out for six hypothetical interventions that: improved the MCS scores of students in the bottom quartile of the distribution to equal either the observed 25th percentile score or the median; improved the MCS scores of students in the bottom half of the distribution to equal the median; and improved the PCS scores in the same three ways. Each scenario improved only one of the two SF-12 scores while holding the other score constant. Population attributed risk proportions (PARPs) in academic role impairment due to these hypothetical interventions were calculated by dividing the difference between observed and predicted proportions by observed proportions. The jackknife repeated replication (JRR) simulation method (Rust and Rao, 1996) was used to estimate standard errors of PARPs using colleges as strata and random subsamples of respondents within colleges as sampling error calculation units. All analyses were carried out in SAS Version 9.4 (SAS Institute Inc., 2014).

3. Results

3.1. Sample description

Item-level missing data ranged between 0.1% (age) and 1.7% (race/ethnicity) across the variables considered here. The 3,761 students in the analysis sample had a mean age of 18.5 and were 53.0% female, 13.9% Hispanic, 9.3% Non-Hispanic Black, 43.0% Non-Hispanic White, and 33.7% defined themselves as Non-Hispanics of “other” races that we did not ask them to specify. (Table 1) More than one-fourth of students (26.3%) were not born in the US, whereas 26.9% were first generation, 16.0% second generation, and the remaining 30.7% third or later generation. No data were collected to disaggregate students not born in the U.S. to distinguish immigrants from international students. Most students (83%) came from families in which at least one parent was a college graduate and a majority had at least one parent with either a masters (31.3%) or doctoral/professional (27.5%) degree. Additional demographic characteristics of the sample are summarized in Appendix Table 1.

Table 1.

Student characteristics by sex and in total

Total (n=3,761) Male (n=1,739) Female (n=1,995) Other (n=27) F-test
Est (SE) Est (SE) Est (SE) Est (SE) Fnum df, dem df
I. Socio-demographics
 Age (Mean) 18.5 (0.0) 18.6 (0.0) 18.4 (0.0) 18.3 (0.1) 2, 13871420.4 = 40.4***
 Race (%)
  Hispanic 13.9 (0.6) 13.4 (1.0) 14.3 (0.8) 12.6 (5.6) 2, 869390 = 0.3
  Non-Hispanic Black 9.3 (0.6) 8.8 (0.9) 9.8 (0.7) 12.2 (6.8) 2, 11178.4 = 0.7
  Non-Hispanic Other 33.7 (0.9) 33.7 (1.6) 33.9 (1.1) 28.3 (9.0) 2, 14271.8 = 0.2
  Non-Hispanic White 43.0 (0.9) 44.1 (1.6) 42.0 (1.1) 46.8 (9.2) 2, 19340.5 = 0.9
 Nativity (%)
  Respondent not born in US 26.3 (0.8) 26.6 (1.4) 26.2 (1.0) 10.7 (6.1) 2, 27870421.8 = 1.6
  Respondent born in US but at least 1 parent not 27.0 (0.8) 25.2 (1.4) 28.7 (1.0) 8.2 (4.9) 2, 42789669.7 = 5.0**
  Respondent and both parents born in US, but at least 1 grandparent not 16.0 (0.6) 15.7 (1.0) 16.2 (0.8) 27.0 (7.8) 2, 11491047.6 =1.2
  Respondent and both and all 4 grandparents born in US 30.7 (0.9) 32.4 (1.5) 28.9 (1.0) 54.1 (9.2) 2, 52901755.7 = 5.8**
 Parental education (%)
  High school or less 8.3 (0.5) 8.0 (0.7) 8.6 (0.6) 3.8 (3.1) 2, 6583957.3 = 0.5
  Some college 8.7 (0.5) 8.1 (0.8) 9.1 (0.6) 17.4 (7.4) 2, 14109191.1 = 1.7
  Bachelor’s degree 24.2 (0.8) 24.2 (1.4) 24.1 (1.0) 28.8 (8.5) 2, 23819798.2 = 0.2
  Master’s degree 31.4 (0.9) 31.8 (1.5) 30.9 (1.0) 35.8 (8.9) 2, 26708709.1 = 0.3
  Doctorate/Professional degree 27.5 (0.8) 27.8 (1.3) 27.3 (1.0) 14.2 (6.3) 2, 137430111.3 = 1.2
II. Perceived health and health-related academic role impairment
  PCS (Mean) 55.1 (0.1) 55.3 (0.2) 54.9 (0.1) 50.7 (1.8) 2, 2252502.3 = 9.0***
  MCS (Mean) 38.2 (0.2) 40.4 (0.3) 36.3 (0.2) 30.4 (2.1) 2, 5649358.3 = 72.7***
  Modified Sheehan Scale (%)
  None 32.9 (0.9) 40.8 (1.6) 26.3 (1.0) 11.1 (5.7) 2, 2547120.8 = 46.1***
  Mild 36.0 (0.9) 33.6 (1.4) 38.2 (1.1) 24.1 (8.2) 2, 8933099.6 = 5.1**
  Moderate 20.6 (0.7) 16.9 (1.1) 23.6 (0.9) 40.3 (9.3) 2, 3759109.1 = 15.4***
  Severe 10.5 (0.5) 8.7 (0.8) 11.9 (0.7) 24.6 (7.6) 2, 35559164.1 = 7.7***
  More than mild/Any 46.4 (1.1) 43.2 (1.8) 48.1 (1.3) 72.9 (9.1) 2, 12290775.6 = 6.0**
  Severe/More than mild 33.7 (1.5) 33.9 (2.7) 33.6 (1.8) 37.9 (10.9) 2, 13001475.9 = 0.1
Open in a new tab
*

p < .05;

**

p < .01;

***

p < .001

Abbreviations: MCS, Mental Component Summary; PCS, Physical Component Summary; SE, standard error; US, United States

The PCS mean in the sample is 55.1 (SE = 0.1), which is about half a standard deviation better than the mean of 50 in the overall US adult population. The PCS standard deviation is 6.1 compared to 10 in the general population. The MCS mean, 38.2 (SE = 0.2), is significantly lower than the PCS mean and is over a full standard deviation worse than the mean in the overall US adult population. The MCS standard deviation is 12.9 compared to 10 in the general population. There is a small, albeit statistically significant, negative correlation between MCS and PCS scores (r = −0.24, p < .001).

Roughly two-thirds of the sample (67.1%) reported health-related academic role impairment, including 36.0% mild, 20.6% moderate, and 10.5% severe. Among individuals with any health-related academic role impairment, 46.4% reported that the impairment was more than mild and 33.7% of those with more than mild academic role impairment reported that it was severe.

3.2. Associations of PCS and MCS scores with health-related academic role impairment

3.2.1. Comparative model fit:

Inspection of comparative model fit shows that M8 (the model with all nonlinearities and the interaction between PCS and MCS) is the best model predicting severe health-related academic role impairment and more than mild academic role impairment among students with any academic role impairment, whereas M7 (the model with all terms other than the interaction) is the best model predicting the other outcomes. (Table 2) The interaction is not significant in predicting any of the outcomes when we assume linear marginal effects (i.e., M6 vs. M5; χ21 = 0.2-2.8, p = .89-.09), but emerges as significant in predicting severe impairment and more than mild impairment among students with any impairment when we allow for nonlinear marginal effects (i.e., M8 vs. M7; χ21 = 5.1-18.5, p = .020-<.001).

Table 2.

Comparisons of fit across nested models (MI-adjusted likelihood ratio χ2 tests)a

Parameters df Any impairment More than mild impairment/Any Severe impairment/More than mild Severe impairment
I. Models (χ2LR)
 Model 1 PCSb 1 4552.0 3345.4 1457.7 2402.6
 Model 2 PCS, PCSs1,c PCSs3d 3 4273.2 3304.9 1453.9 2352.8
 Model 3 MCSe 1 3388.9 3064.9 1388.0 1996.9
 Model 4 MCS, MCSs1,f MCSs3g 3 3347.1 3039.4 1358.2 1989.6
 Model 5 PCS, MCS 2 3228.4 2872.9 1355.1 1864.1
 Model 6 PCS, MCS, PCS*MCS 3 3226.2 2871.7 1352.3 1864.1
 Model 7 PCS, PCSs1, PCSs3, MCS, MCSs1, MCSs3 6 3180.5 2849.1 1320.3 1852.9
 Model 8 PCS, PCSs1, PCSs3, MCS, MCSs1, MCSs3, PCS*MCS 7 3177.1 2843.9 1301.9 1851.2
II. Model comparisons (χ2LR difference)
  Model 5 vs 1 MCS 1 1323.6*** 472.6*** 102.6*** 538.5***
  Model 5 vs 3 PCS 1 160.6*** 192.1*** 33.0*** 132.8***
  Model 2 vs 1 PCSs1, PCSs3 2 278.8*** 40.5*** 3.7 49.7***
  Model 4 vs 3 MCSs1, MCSs3 2 41.8*** 25.5*** 29.8*** 7.3*
  Model 7 vs 5 PCSs1, PCSs3, MCSs1, MCSs3 4 47.8*** 23.8*** 34.7*** 11.2**
  Model 7 vs 4 PCS, PCSs1, PCSs3 3 166.6*** 190.4*** 37.9*** 136.7***
  Model 7 vs 2 MCS, MCSs1, MCSs3 3 1092.6*** 455.9*** 133.6*** 499.9***
  Model 6 vs 5 PCS*MCS 1 2.2 1.2 2.8 0.02
  Model 8 vs 7 PCS*MCS 1 3.4 5.1* 18.5*** 1.8
Open in a new tab
*

p < .05;

**

p < .01;

***

p < .001

Abbreviations: df, degrees of freedom

a

All models adjusted for age, gender, school, nativity, parent education and race/ethnicity

b

PCS = SF-12 Physical Component Summary

c

PCSs1 = SF-12 Physical Component Summary spline (<25th percentile)

d

PCSs3 = SF-12 Physical Component Summary spline (>75th percentile)

e

MCS = SF-12 Mental Component Summary

f

MCSs1 = SF-12 Mental Component Summary spline (<25th percentile)

g

MCSs3 = SF-12 Mental Component Summary spline (>75th percentile)

3.2.2. Model coefficients:

Examination of model coefficients shows that the ORs of PCS and MCS are consistently less than 1.0, indicating that improvements in both physical and mental health are associated with reductions in academic role impairment. (Table 3) This broad pattern is consistent with the gross associations between quartiles of the PCS and MCS distributions with the outcomes. (Appendix Table 2) With the exception of the model predicting any impairment, the ORs of PCS with impairment increase monotonically with increasing mental health. Similarly, with the exception of the model predicting any impairment, the ORs of MCS with impairment increase monotonically with increasing physical health. But the significant PCS x MCS interactions are consistently greater than 1.0, indicating that the generally negative associations between each type of health and impairment weaken with decreases in the other type of health.

Table 3.

Best model associations of Physical and Mental Health Component Summaries predicting academic role impairment

Any impairment More than mild impairment/Any Severe impairment/More than mild Severe impairment
OR (95%CI) OR (95%CI) OR (95%CI) OR (95%CI)
I. SF-12 PCS splines
 0–25th percentile 0.9*** (0.8-0.9) 0.8*** (0.7-0.9) 0.8*** (0.7-0.9) 0.9*** (0.9-1.0)
 25th–75th percentile 0.9* (0.9-1.0) 0.8*** (0.8-0.9) 0.8*** (0.8-0.9) 0.9*** (0.8-0.9)
 >75th percentile 1.0 (0.8-1.1) 0.9* (0.8-1.0) 0.9* (0.8-1.0) 0.9 (0.8-1.0)
  χ23 166.6*** 27.2*** 37.7*** 136.7***
II. SF-12 MCS splines
   0–25th percentile 1.0* (0.9-1.0) 0.8*** (0.6-0.9) 0.7*** (0.6-0.8) 0.8*** (0.8-0.9)
   25th–75th percentile 0.8*** (0.8-0.8) 0.8*** (0.7-0.9) 0.7*** (0.6-0.8) 0.9*** (0.8-0.9)
   >75th percentile 0.9*** (0.9-1.0) 0.8* (0.8-1.0) 0.8* (0.7-0.9) 1.0 (0.9-1.0)
    χ23 1,092.6*** 30.7*** 60.2*** 499.9***
   PCS × MCS interaction -- 1.3* (1.0-1.6) 1.6*** (1.3-2.1) --
Open in a new tab
*

p < .05;

**

p < .01;

***

p < .001

Abbreviations: CI, confidence interval; MCS, Mental Component Summary; OR odds ratio; PCS, Physical Component Summary

3.2.3. Population attributable risk proportions:

The simulations estimate that each of the six hypothetical interventions would result in a significant reduction in each of the three components of academic role impairment (i.e., any impairment, more than mild impairment among students with any, severe impairment among students with more than mild impairment) as well in the overall proportion of students with severe academic role impairment. (Table 4) Comparisons of the three pairs of interventions to improve MCS and PCS in similar ways across these four outcomes show that the estimated effects would be significantly different in 11 out of 12 cases, in 10 of which the intervention to improve mental health would lead to a significantly greater decrease in academic role impairment than the comparable intervention to improve physical health.

Table 4.

Population Attributable Risk Proportions for best-fitting models simulating the effects of intervening on either SF-12 Physical Component Summary scores or Mental Component Summary scores in reducing academic role impairmenta

Any impairmentb More than mild impairment/Anyc Severe impairment/More than mildc Severe impairmentb
PARP (SE) T (SE)d PARP (SE) T (SE)d PARP (SE) T (SE)d PARP (SE) T (SE)d
I. SF-12 PCS
 If below Q1, shift to Q1 3.9 (0.5) 5.6 (1.2) 4.4 (1.5) 12.2 (2.2)
 If below Q1, shift to Median 5.4 (0.6) 11.5 (1.1) 8.2 (1.5) 21.6 (3.1)
 If below Median, shift to Median 6.1 (0.9) 14.3 (1.7) 10.4 (1.9) 24.6 (3.3)
II. SF-12 MCS
  If below Q1, shift to Q1 0.7 (0.3) −3.2 (0.2)*** 9.0 (1.1) 3.4 (1.3)* 11.0 (3.1) 6.6 (5.9) 38.7 (3.2) 26.5 (7.5)*
  If below Q1, shift to Median 7.9 (0.8) 2.5 (0.5)*** 17.5 (1.4) 6.0 (1.6)* 22.1 (2.1) 13.9 (3.3)*** 53.5 (4.3) 31.9 (14.1)*
  If below Median, shift to Median 12.1 (1.1) 6.0 (1.0)*** 21.7 (1.7) 7.4 (2.9)* 27.3 (2.4) 16.9 (4.7)* 61.3 (5.1) 36.7 (18.4)*
Open in a new tab
*

p < .05;

**

p < .01;

***

p < .001

Abbreviations: MCS, Mental Component Summary; PARP, Population Attributable Risk Proportion; PCS, Physical Component Summary; SE, standard error

a

All models adjusted for continuous age, male gender, school, nativity, parent education and race/ethnicity

b

Model for PARP includes adjusting variables and SF-12 MCS splines (at the 25th, 50th and 75th percentiles) and SF-12 PCS splines (at the 25th, 50th and 75th percentiles)

c

Model for PARP includes adjusting variables and SF-12 MCS splines (at the 25th, 50th and 75th percentiles), SF-12 PCS splines (at the 25th, 50th and 75th percentiles), and interaction between SF-12 MCS and PCS

d

T-tests compare the impact of congruous shifts in SF-12 MCS vs SF-12 PCS scores.

The largest estimated effect in the simulations is for the intervention to improve the mental health of students with scores below the MCS median to be at the median. Such an intervention would be expected to result in a 61.3% reduction in the proportion of students who experience severe health-related academic role impairment. The estimated effects of a comparable intervention to improve student physical health is a 24.6% reduction in severe health-related academic role impairment. Decomposition of the MCS effect suggests that nearly two-thirds of the total effect on severe role impairment would be due to increasing the mental health of students with scores below the 25th percentile to the 25th percentile (63%; 38.7/61.3), whereas the remainder would be due to increasing the mental health of these same students to the median (24%; [53.5-38.7]/61.3) and increasing the mental health of students with scores between the 25th and 50th percentiles to the median (13%; [61.3-53.5]/61.3). Although these improvements in mental health would be associated with significant reductions in health-related academic role impairments across the range of impairment levels, the largest proportional reductions in all cases would be in severe role impairment among students with more than mild role impairment.

A decomposition of the total comparable PCS effect suggests that about half of the effect on severe role impairment would be due to increasing the physical health of students below the 25th percentile to the 25th percentile (50%; 12.2/24.5), whereas the remainder would be due to increasing the physical health of these same students to the median (38%; [21.6-12.2]/24.5) and increasing the physical health of students with scores between the 25th and 50th percentiles to the median (12%; [24.5-21.6]/24.5). Unlike the situation with a mental health intervention, where the largest proportional reductions in severe academic role impairment would be due to reducing severe impairment among students with more than mild impairment, the major effect of a physical health intervention would be in reducing more than mild impairment among students with any impairment.

4. Discussion

To the best of our knowledge, this is the first study to examine the relative importance of physical and mental health in accounting for health-related academic role impairment among college students. Results indicate that college students have significantly better physical health and significantly worse mental health than the overall US adult population, that most students have at least mild health-related academic role impairment, that physical and mental health are both significantly and inversely associated with this impairment, and that these associations are for the most part a good deal stronger for mental than physical health. The latter result suggests that successful interventions to increase student mental health would lead to considerably greater improvements in academic role performance than would successful interventions to increase student physical health.

The finding that mean PCS scores are higher than in the adult general population is not surprising given the young age of the sample. The finding that mean MCS scores are lower than in the adult general population was also expected given evidence of high Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) disorder prevalence in other recent surveys of college students (Cho et al., 2015; Hunt and Eisenberg, 2010; Kendler et al., 2015). However, we were nonetheless surprised to find that the mean MCS score was more than a full standard deviation below the mean in the adult general population. However, these results are quite different from those in the MEP survey reviewed in the introduction (Cohen et al., 2009; Hanmer and Kaplan, 2016). It is noteworthy that the MEPS reported results by age and sex but not in a way that distinguished college students from other young adults, making it impossible for us to know if the results found in the current study are idiosyncratic to this sample or generalize to all college students in the US. We are exploring these possibilities in a series of surveys in a larger sample of institutions that include non-respondent follow-up assessments.

We were surprised to see that PCS and MCS scores were negatively (albeit modestly) correlated given other data suggesting positive correlations between the two types of problems in the general population (Scott et al., 2007). This might reflect both the fact that physical health problems in this age range are typically not severe enough to influence mental health and that the effects of mental disorders on chronic physical disorders take more time to emerge (Scott et al., 2016), although both of these possibilities imply the existence of specifications that need to be investigated in epidemiological samples with broader age ranges to see if they are confirmed.

Our analysis of nonlinear associations and interactions suggests that identifying and successfully treating students with the worst mental health would have the most impact on academic performance. However, it would presumably be more difficult to treat these students than to treat students with milder emotional problems. Pragmatic trials are needed to investigate this issue. Given that there are 22 million college students in the US and our results suggest that many of them experience mental health problems sufficiently severe to be associated with impaired academic performance, scalable interventions will have to be centrally involved in addressing this enormous problem of unmet need for treatment (Harrer et al., 2019). Existing research suggests that guided online interventions can be as effective as face-to-face psychotherapy in treating mild-moderate common mental disorders (Carlbring et al., 2018), but it is not known if this is equally true among college students. A challenge in answering this question is that college students have a number of psychological barriers to seeking treatment for emotional problems such as embarrassment, cost, and inconvenience of treatment that will have to be overcome before broad-based interventions can be implemented (Ebert et al., 2019a). Interventions delivered through computational software offer a scalable way of addressing some of these concerns because they can be delivered privately via computer or smart phone in the student’s home or dorm room at a relatively low cost and at times that are convenient for the student. Interventions offered to everyone emphasizing mind health and optimal performance rather than need for treatment or offered as a course in psychological skills training could potentially avoid the inherent stigma associated with seeking help for psychological problems (Cuijpers et al., 2009). In addition to these possibilities, we are exploring a number of other innovative ways of increasing treatment uptake and retention (Ebert et al., 2019b) and expanding group psychoeducational intervention program to be delivered in a variety of settings and mediums. The next phase of WMH-ICS will involve carrying out a series of pragmatic trials to determine whether these interventions are successful in improving college student mental health (Cuijpers et al., 2019).

The results should be interpreted within the context of several limitations. First, the response rate was low and the sample might be biased in the direction of students with psychological problems having a higher probability of participation than other students. Future college health surveys need to develop methods to improve response rates. Second, the sample consisted only of first year students from five northeastern universities, limiting generalizability of findings. Third, both academic role impairment and health functioning were assessed using self-report questionnaires rather than objective assessments (e.g., physical examinations, administrative reports of grade point average), although both the SF-12 and SDS have been shown to have high concordance with clinician ratings (Salyers et al., 2000; Sheehan et al., 1996). Relatedly, the SF-12 measures are summary scores that provide no insights into the specific physical and mental disorders that should be the focus of clinical attention in order to reduce academic role impairments. Fourth, substance use disorders are not included in these measures even though substance problems are known to be common among college students and to influence academic role performance (Auerbach et al., 2018; Bruffaerts et al., 2018). Future research needs to determine which specific mental disorders account for the strong associations documented here between SF-12 MCS scores and academic functioning as well as to include information about alcohol and drug use disorders. Fifth, all but one of the surveys were carried out in October. This limits the external validity of results because the mental health of first-year college students is known to decrease over the course of the school year (Pritchard et al., 2007; Sax et al., 2004). Future WMH-ICS surveys are being implemented in random replicates across the entire academic year to address this problem.

4.3. Conclusions

Within the context of these limitations, the results suggest that mental health problems account for a much higher proportion of academic role impairment than physical health problems among university students. As universities begin to grapple with the growing recognition of the important role mental health plays in student success, more attention will need to be placed on broad-based interventions designed to improve the mental health of the student body. Given the magnitude of the problem, scalable solutions are needed. More research is needed to evaluate the effects of innovative scalable mental health interventions and develop methods to triage college students in need of treatment into the least expensive interventions.

Supplementary Material

1

Acknowledgments

Conflicts of Interest

Dr. Ebert reports to have received consultancy fees/served on the scientific advisory board for Sanofi, Novartis, Minddistrict, Lantern, Schoen Kliniken, and two German health insurance companies (BARMER, Techniker Krankenkasse). He is also a stakeholder in the Institute for health training online (GET.ON), which aims to implement scientific findings related to digital health interventions into routine care. In the past 3 years, Dr. Kessler received support for his epidemiological studies from Sanofi Aventis; was a consultant for Datastat, Inc., Johnson & Johnson Wellness and Prevention, Sage Pharmaceuticals, Shire, Takeda; and served on an advisory board for the Johnson & Johnson Services Inc. Lake Nona Life Project. The remaining authors have no conflicts to declare.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Alonso J, Mortier P, Auerbach RP, Bruffaerts R, Vilagut G, Cuijpers P, Demyttenaere K, Ebert DD, Ennis E, Gutiérrez-García RA, Green JG, Hasking P, Lochner C, Nock MK, Pinder-Amaker S, Sampson NA, Zaslavsky AM, Kessler RC; WHO WMH-ICS Collaborators, 2018. Severe role impairment associated with mental disorders: results of the WHO World Mental Health Surveys International College Student Project. Depress Anxiety, 35, 802–814. 10.1002/da.22778 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Alonso J, Vilagut G, Mortier P, Auerbach RP, Bruffaerts R, Cuijpers P, Demyttenaere K, Ebert DD, Ennis E, Gutiérrez-García RA, Green JG, Hasking P, Lee S, Bantjes J, Nock MK, Pinder-Amaker S, Sampson NA, Zaslavsky AM, Kessler RC; WHO WMH-ICS Collaborators, 2019. The role impairment associated with mental disorder risk profiles in the WHO World Mental Health International College Student Initiative. Int. J. Methods Psychiatr. Res, 28, e1750. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Arria AM, Caldeira KM, Vincent KB, Winick ER, Baron RA, O’Grady KE, 2013. Discontinuous college enrollment: Associations with substance use and mental health. Psychiatr. Serv, 64, 165–172. 10.1176/appi.ps.201200106 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Auerbach RP, Alonso J, Axinn WG, Cuijpers P, Ebert DD, Green JG, Hwang I, Kessler RC, Liu H, Mortier P, Nock MK, Pinder-Amaker S, Sampson NA, Aguilar-Gaxiola S, Al-Hamzawi A, Andrade LH, Benjet C, Caldas-de-Almeida JM, Demyttenaere K, Florescu S, de Girolamo G, Gureje O, Haro JM, Karam EG, Kiejna A, Kovess-Masfety V, Lee S, McGrath JJ, O’Neill S, Pennell BE, Scott K, Ten Have M, Torres Y, Zaslavsky AM, Zarkov Z, Bruffaerts R, 2016. Mental disorders among college students in the World Health Organization World Mental Health Surveys. Psychol. Med, 46, 2955–2970. 10.1017/s0033291716001665 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Auerbach RP, Mortier P, Bruffaerts R, Alonso J, Benjet C, Cuijpers P, Demyttenaere K, Ebert DD, Green JG, Hasking P, Lee S, Lochner C, McLafferty M, Nock MK, Petukhova MV, Pinder-Amaker S, Rosellini AJ, Sampson NA, Vilagut G, Zaslavsky AM, Kessler RC; WHO WMH-ICS Collaborators, 2019. Mental disorder comorbidity and suicidal thoughts and behaviors in the World Health Organization World Mental Health Surveys International College Student Initiative. Int. J. Methods Psychiatr. Res, 28, e1752 10.1002/mpr.1752 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Auerbach RP, Mortier P, Bruffaerts R, Alonso J, Benjet C, Cujipers P, Demyttenaere K, Ebert DD, Green JG, Hasking P, Murray E, Nock MK, Pinder-Amaker S, Sampson NA, Stein DJ, Vilagut G, Zaslavsky AM, Kessler RC; WHO WMH-ICS Collaborators, 2018. The WHO World Mental Health Surveys International College Student Project: Prevalence and distribution of mental disorders. J. Abnorm. Psychol, 127, 623–638. 10.1037/abn0000362 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bruffaerts R, Mortier P, Auerbach RP, Alonso J, Hermosillo De la Torre AE, Cuijpers P, Demyttenaere K, Ebert DD, Green JG, Hasking P and Stein DJ, 2019. Lifetime and 12-month treatment for mental disorders and suicidal thoughts and behaviors among first year college students. Int. J. Methods Psychiatr. Res, 28, e1764 10.1002/mpr.1764 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bruffaerts R, Mortier P, Kiekens G, Auerbach RP, Cuijpers P, Demyttenaere K, Green JG, Nock MK, Kessler RC, 2018. Mental health problems in college freshmen: Prevalence and academic functioning. J. Affect. Disord, 225, 97–103. 10.1016/j.jad.2017.07.044 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Carlbring P, Andersson G, Cuijpers P, Riper H, Hedman-Lagerlof E, 2018. Internet-based vs. face-to-face cognitive behavior therapy for psychiatric and somatic disorders: An updated systematic review and meta-analysis. Cogn. Behav. Ther, 47, 1–18. 10.1080/16506073.2017.1401115 [DOI] [PubMed] [Google Scholar]
  10. Cho SB, Llaneza DC, Adkins AE, Cooke M, Kendler KS, Clark SL, Dick DM, 2015. Patterns of substance use across the first year of college and associated risk factors. Front Psychiatry, 6, 152 10.3389/fpsyt.2015.00152 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Cohen JW, Cohen SB, Banthin JS, 2009. The medical expenditure panel survey: A national information resource to support healthcare cost research and inform policy and practice. Med. Care, 47(7 Suppl 1), S44–50. 10.1097/MLR.0b013e3181a23e3a [DOI] [PubMed] [Google Scholar]
  12. Coons SJ, Rao S, Keininger DL, Hays RD, 2000. A comparative review of generic quality-of-life instruments. Pharmacoeconomics, 17, 13–35. 10.2165/00019053-200017010-00002 [DOI] [PubMed] [Google Scholar]
  13. Cuijpers P, Auerbach RP, Benjet C, Bruffaerts R, Ebert D, Karyotaki E, Kessler RC, 2019. The World Health Organization World Mental Health International College Student Initiative: An overview. Int. J. Methods Psychiatr. Res, 28, e1761 10.1002/mpr.1761 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Cuijpers P, Munoz RF, Clarke GN, Lewinsohn PM, 2009. Psychoeducational treatment and prevention of depression: The “coping with depression” course thirty years later. Clin. Psychol. Rev, 29, 449–458. 10.1016/j.cpr.2009.04.005 [DOI] [PubMed] [Google Scholar]
  15. Dryer R, Henning MA, Tyson GA, Shaw R, 2016. Academic achievement performance of university students with disability: Exploring the influence of non-academic factors. International Journal of Disability, Development and Education, 63, 419–430. 10.1080/1034912X.2015.1130217 [DOI] [Google Scholar]
  16. Ebert DD, Mortier P, Kaehlke F, Bruffaerts R, Baumeister H, Auerbach RP, Alonso J, Vilagut G, Martinez KU, Lochner C, Cuijpers P, Kuechler AM, Green J, Hasking P, Lapsley C, Sampson NA, Kessler RC; WHO World Mental Health-International College Student Initiative collaborators, 2019a. Barriers of mental health treatment utilization among first-year college students: First cross-national results from the WHO World Mental Health International College Student Initiative. Int. J. Methods Psychiatr. Res, 28, e1782 10.1002/mpr.1782 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Ebert DD, Franke M, Kählke F, Küchler AM, Bruffaerts R, Mortier P, Karyotaki E, Alonso J, Cuijpers P, Berking M, Auerbach RP, Kessler RC, Baumeister H; WHO World Mental Health - International College Student collaborators, 2019b. Increasing intentions to use mental health services among university students. Results of a pilot randomized controlled trial within the World Health Organization’s World Mental Health International College Student Initiative. Int. J. Methods Psychiatr. Res, 28, e1754 10.1002/mpr.1754 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. El Ansari W, Stock C, 2010. Is the health and wellbeing of university students associated with their academic performance? Cross sectional findings from the United Kingdom. Int. J. Environ. Res. Public Health, 72, 509–527. 10.3390/ijerph7020509 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Eisenberg D, Golberstein E, Hunt JB, 2009. Mental health and academic success in college. The B.E. Journal of Economic Analysis & Policy, 9. doi: 10.2202/1935-1682.2191 [DOI] [Google Scholar]
  20. Gandek B, Ware JE, Aaronson NK, Apolone G, Bjorner JB, Brazier JE, Bullinger M, Kaasa S, Leplege A, Prieto L, Sullivan M, 1998. Cross-validation of item selection and scoring for the SF-12 health survey in nine countries: Results from the IQOLA project. J. Clin. Epidemiol, 51, 1171–1178. [DOI] [PubMed] [Google Scholar]
  21. Gaultney JF, 2010. The prevalence of sleep disorders in college students: Impact on academic performance. J. Am. Coll. Health, 59, 91–97. 10.1080/07448481.2010.483708 [DOI] [PubMed] [Google Scholar]
  22. Gormley MJ, DuPaul GJ, Weyandt LL, Anastopoulos AD, 2019. First-year GPA and academic service use among college students with and without ADHD. J. Atten. Disord, 23, 1766–1779. 10.1177/1087054715623046 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Greenland S, Drescher K, 1993. Maximum likelihood estimation of the attributable fraction from logistic models. Biometrics, 49, 865–872. [PubMed] [Google Scholar]
  24. Hanmer J, Kaplan RM, 2016. Update to the report of nationally representative values for the noninstitutionalized US adult population for five health-related quality-of-life scores. Value Health, 19, 1059–1062. 10.1016/j.jval.2016.05.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Harrer M, Adam SH, Baumeister H, Cuijpers P, Karyotaki E, Auerbach RP, Kessler RC, Bruffaerts R, Berking M, Ebert DD, 2019. Internet interventions for mental health in university students: A systematic review and meta-analysis. Int. J. Methods Psychiatr. Res, 28, e1759 10.1002/mpr.1759 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Hunt J, Eisenberg D, 2010. Mental health problems and help-seeking behavior among college students. J. Adolesc. Health, 46, 3–10. 10.1016/j.jadohealth.2009.08.008 [DOI] [PubMed] [Google Scholar]
  27. Hysenbegasi A, Hass SL, Rowland CR, 2005. The impact of depression on the academic productivity of university students. J. Ment. Health Policy Econ, 8, 145. [PubMed] [Google Scholar]
  28. Kendler KS, Myers J, Dick D, 2015. The stability and predictors of peer group deviance in university students. Soc. Psychiatry Psychiatr. Epidemiol, 50, 1463–1470. 10.1007/s001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Mateen BA, Doogan C, Hayward K, Hourihan S, Hurford J, Playford ED, 2017. Systematic review of health-related work outcome measures and quality criteria-based evaluations of their psychometric properties. Arch. Phys. Med. Rehabil, 98, 534–560. 10.1016/j.apmr.2016.06.01327-015-1031-4 [DOI] [PubMed] [Google Scholar]
  30. McHorney CA, Ware JE Jr., Raczek AE, 1993. The MOS 36-Item Short-Form Health Survey (SF-36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs. Med. Care, 31, 247–263. 10.1097/00005650-199303000-00006 [DOI] [PubMed] [Google Scholar]
  31. Merkt J, Gawrilow C, 2016. Health, dietary habits, and achievement motivation in college students with self-reported ADHD diagnosis. J. Atten. Disord, 20, 727–740. 10.1177/1087054714523127 [DOI] [PubMed] [Google Scholar]
  32. Pritchard ME, Wilson GS, Yamnitz B, 2007. What predicts adjustment among college students? A longitudinal panel study. J. Am. Coll. Health, 56, 15–22. 10.3200/JACH.56.1.15-22 [DOI] [PubMed] [Google Scholar]
  33. Rust KF, Rao JN, 1996. Variance estimation for complex surveys using replication techniques. Stat. Methods Med. Res, 5, 283–310. [DOI] [PubMed] [Google Scholar]
  34. Ruthig JC, Marrone S, Hladkyj S, Robinson-Epp N, 2011. Changes in college student health: Implications for academic performance. J. Coll. Stud. Dev, 52, 307–320. 10.1353/csd.2011.0038 [DOI] [Google Scholar]
  35. Salyers MP, Bosworth HB, Swanson JW, Lamb-Pagone J, Osher FC 2000. Reliability and validity of the SF-12 health survey among people with severe mental illness. Med. Care, 1141–1150. doi: 10.1097/00005650-200011000-00008 [DOI] [PubMed] [Google Scholar]
  36. SAS Institute Inc., 2014. SAS 9.4 Output delivery system: User’s guide. Cary, N.C.: SAS Institute. [Google Scholar]
  37. Sax LJ, Bryant AN, Gilmartin SK, 2004. A longitudinal investigation of emotional health among male and female first-year college students. Journal of the First-Year Experience & Students in Transition, 16, 39–65. [Google Scholar]
  38. Scott KM, Bruffaerts R, Tsang A, Ormel J, Alonso J, Angermeyer MC, Benjet C, Bromet E, de Girolamo G, de Graaf R, Gasquet I, Gureje O, Haro JM, He Y, Kessler RC, Levinson D, Mneimneh ZN, Oakley Browne MA, Posada-Villa J, Stein DJ, Takeshima T, Von Korff M, 2007. Depression-anxiety relationships with chronic physical conditions: Results from the World Mental Health Surveys. J. Affect. Disord, 103, 113–120. 10.1016/j.jad.2007.01.015 [DOI] [PubMed] [Google Scholar]
  39. Scott KM, Lim C, Al-Hamzawi A, Alonso J, Bruffaerts R, Caldas-de-Almeida JM, Florescu S, de Girolamo G, Hu C, de Jonge P, Kawakami N, Medina-Mora ME, Moskalewicz J, Navarro-Mateu F, O’Neill S, Piazza M, Posada-Villa J, Torres Y, Kessler RC, 2016. Association of mental disorders with subsequent chronic physical conditions: World Mental Health Surveys from 17 countries. JAMA Psychiatry. 73, 150–158. 10.1001/jamapsychiatry.2015.2688. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Sheehan DV, Harnett-Sheehan K, Raj BA, 1996. The measurement of disability. Int. Clin. Psychopharmacol, 11 Suppl 3, 89–95. [DOI] [PubMed] [Google Scholar]
  41. Turner JC, Keller A, 2015. College health surveillance network: Epidemiology and health care utilization of college students at US 4-year universities. J. Am. Coll. Health, 63, 530–538. 10.1080/07448481.2015.1055567 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. van Buuren S, 2007. Multiple imputation of discrete and continuous data by fully conditional specification. Stat. Methods Med. Res, 16, 219–242. 10.1177/0962280206074463 [DOI] [PubMed] [Google Scholar]
  43. Ware JE, Kosinski M, Keller SD, 1995. SF-12: How to score the SF-12 physical and mental health summary scales, second ed The Health Institute, New England Medical Center, Boston, Massachusetts [Google Scholar]
  44. Ware J Jr., Kosinski M, Keller SD, 1996. A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med. Care, 220–233. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1558000-supplement-1.docx (31.1KB, docx)

RESOURCES