Abstract
Positive and negative affect are both associated with health outcomes. Using validated measures, we examined associations between affect, self-reported measures of health, and objective measures of systemic inflammation in a cross-sectional sample of outpatient subjects recruited from an urban county hospital. Participants (n=1055) recruited from the Grady Trauma Project in Atlanta, GA underwent standardized interviews including self-report measures of psychiatric symptoms and physical health. A subset (n=246) consented to an assay of serum C-reactive protein (CRP). Regression models including positive affect as the predictor variable with covariates of age, gender, income, trauma load, depression and PTSD symptoms, were significantly associated with physical health domain scales of the Short Form-36 Health Survey (SF-36) of general health (R2=0.212; p<0.001) and physical functioning (R2=0.154; p=0.013). No association was observed using negative affect as the predictor variable. While greater serum CRP concentrations were associated with less positive affect (r=−0.137; p=0.038), this relationship did not remain significant (p=0.250) when controlling for demographic variables, body mass index, trauma load, and psychiatric symptoms. Future studies using larger samples or samples with more variance for CRP and positive and negative affect may be helpful in investigating the relationship between CRP and positive and negative affect. Our results support the hypothesis that positive affect contributes beneficially to physical health. Development of strategies to enhance positive affect in at-risk populations may be a meaningful way to improve their health.
Keywords: PTSD, c-reactive protein, immune response, positive affect, negative affect, health
Introduction
The literature examining negative affective states and their resulting outcomes has historically greatly outweighed comparable studies of positive affective states (Diener et al., 1999). Through large community samples, much evidence has accrued in the medical literature describing the association of psychiatric disorders such as depression, that are characterized by negative affective state, with large global disability burdens including coronary heart disease (Hemingway and Marmot, 1999) and diabetes (Golden et al., 2004). Conversely, a growing body of data examining the impact of positive affect on human biology has demonstrated a modest, but beneficial, influence of positive affect on biomarkers of autonomic, neuroendocrine, and immune physiology (Dockray and Steptoe, 2010), measures of general physical health (Pressman and Cohen, 2005; Rasmussen et al., 2009) and health or disease progression in clinical populations of patients with diabetes (Robertson et al., 2012) or cardiovascular disease (Huffman et al., 2017).
The relationships between affect and health have been consistently observed in a variety of international studies across socioeconomic and racial/ethnic cohorts. An overall higher life satisfaction score in a cohort of healthy Finnish adults from the Finnish Twin Study was negatively related to mortality (Koivumaa-Honkanen et al., 2000). Lack of positive affect was associated with increased mortality in a British community sample recruited as part of the Health and Lifestyle Study (Huppert and Whittington, 1995) as well as a North Carolina cohort of urban and rural-dwelling, African-American and white elders (Blazer and Hybels, 2004). In the same North Carolina cohort, it was also found that while increasing scores on a depressive scale were associated with increased risk of stroke, increasing scores on a positive affect scale were even more strongly associated with decreased risk of stroke (Ostir et al., 2001). In a Mexican-American elder cohort, lack of positive affect, rather than increased negative affect, was predictive for development of physical disability (Ostir et al., 2000).
A recent meta-analysis of prospective observational cohort studies examining the association between positive well-being and mortality found positive affect and positive-trait-like disposition to be associated with reduced mortality in both healthy and ill populations (Chida and Steptoe, 2008). More significantly, this association was found to be independent of the effect of negative affect, consistent with previous studies identifying an independent effect of positive affect on health (Diener et al., 1985).
In the present study, we examined the relationship between positive and negative affect and self-reported subjective health using the Short Form Health Survey (SF-36) in a predominantly African-American, heavily trauma-exposed and impoverished population recruited from the primary care waiting rooms of a large county hospital. We hypothesized that increasing positive affect would be associated with better health outcomes, and conversely, that increasing negative affect would be associated with worsened health outcomes. To test this hypothesis, we developed a statistical model accounting for the effects of demographic factors, depressive symptoms, and post-traumatic stress symptoms, as our study cohort has high rates of current and lifetime prevalence of major depressive disorder and post-traumatic stress disorder (PTSD) (Gillespie et al., 2009). In addition to our subjective measurement of health and its association with positive and negative affect, we also examined the association of positive and negative affect with an objective measure of health, serum C-reactive protein (CRP) level, a biomarker of systemic inflammation linked to increased medical and psychiatric risk (Miller et al., 2009; Penninx et al., 2003) in a sub-sample of study participants.
Methods
Study Design and Participants
Participants in the current study were recruited as a part of the Grady Trauma Project, a 5-year National Institutes of Health-funded study of risk factors and resilience in PTSD (Bradley et al., 2011; Gillespie et al., 2009). Data were collected between 2009 and 2014. Participants were recruited from general primary care and obstetric/gynecologic clinic waiting rooms during normal business hours at Grady Memorial Hospital, a large, publically-funded county hospital serving a low-income, primarily African-American population in Atlanta, Georgia.
Inclusion criteria included English-language fluency and ability to give informed consent. Verbal and written consent was obtained for all participants. Trained interviewers approached participants while they were waiting for medical appointments. Approximately 60% of participants approached at this phase agreed to participate. Study participants completed a series of self-report questions verbally, in cooperation with study interviewers, over a period of 45 to 75 minutes (dependent on participants’ trauma history and current symptom reports). All procedures in this study were approved by the institutional review boards of Emory University School of Medicine and Grady Memorial Hospital, Atlanta, Georgia.
Our sample (N=1055) includes individuals who completed self-report affect and health outcome measures. However, as participants were allowed to decline questions they did not wish to answer, or did not complete all measures due to the time constraints inherent to interviewing in a clinical waiting room, the number of participants for each individual analyses varies. A subgroup of participants (n=246) who completed this initial interview was invited to participate in a secondary phase of the study that included structured interviews, history & physical examination, laboratory measurements, and physiologic studies that comprised additional study areas of the Grady Trauma Project. This included venipuncture to assess serum CRP levels during the secondary phase of the study that occurred on average 1–2 weeks post-initial assessment. Body Mass Index (BMI) was calculated based on measurements obtained at the time of physician history and physical exam (mean = 32.05, SD = 7.87, range 18.38 – 70.41).
Psychological Measures
A full description of the battery of self-report measures obtained during the interview has previously been described (Gillespie et al., 2009). See Table 1 for detailed descriptive statistics on psychological measures. Demographic information was assessed using a locally-developed demographics form (Gillespie et al., 2009). Trauma exposure was measured using the Traumatic Events Inventory (TEI; Gillespie et al., 2009), a 14-item screening measure of the total number of different types of trauma an individual has been exposed to in their lifetime (e.g., domestic violence, serious car accident, child abuse). Symptoms of depression were assessed with the Beck Depression Inventory (BDI), a 21-item self-report inventory measuring frequency and severity of depression symptoms (Beck et al., 1988). Depression symptom scores range from 0–63, with 0–13 considered minimal, 14–19 mild, 20–28 moderate, and 29–63 severe. Current post-traumatic stress symptoms were assessed using the modified PTSD Symptom Scale (PSS). The PSS is a 17-item self-report measure which assesses the presence/absence of the 17 symptoms contained in the DSM-IV diagnostic criteria for PTSD during the two weeks preceding assessment, with scores ranging from 0–51. A total score higher than 13 is indicative of probable PTSD (Foa et al., 1993).
Table 1.
Demographic characteristics of total sample and serum CRP subsample in frequency (n) and percentage (%), and mean ± SD for the total sample and serum CRP subsample.
| Demographics. | Total Sample (n=1055) | Serum CRP Sample (n=246) | ||
|---|---|---|---|---|
| Sex | N | % | N | % |
| Female | 699 | 66.3 | 105 | 42.7 |
| Male | 355 | 33.7 | 141 | 57.3 |
| Race | ||||
| African-American | 972 | 92.1 | 226 | 91.9 |
| Hispanic | 6 | 0.6 | 2 | 0.8 |
| Asian | 2 | 0.2 | 1 | 0.4 |
| Caucasian | 45 | 4.3 | 11 | 4.5 |
| Mixed | 15 | 1.4 | 4 | 1.6 |
| Other | 12 | 1.1 | 2 | 0.8 |
| Employment | ||||
| Unemployed | 812 | 77.5 | 197 | 80.1 |
| Employed | 236 | 22.5 | 49 | 19.9 |
| Education | ||||
| <12th grade | 257 | 24.4 | 62 | 25.2 |
| 12th grade or HS | 368 | 34.9 | 90 | 36.6 |
| Graduate | ||||
| GED | 57 | 5.4 | 19 | 7.7 |
| Some College or Tech | 230 | 21.8 | 48 | 19.5 |
| School | ||||
| Tech School Grad | 44 | 4.2 | 11 | 4.5 |
| College Graduate | 77 | 7.3 | 13 | 5.3 |
| Graduate School | 15 | 1.4 | 2 | 0.8 |
| Income | ||||
| $0 – 249 | 309 | 29.3 | 79 | 32.1 |
| $250 – 499 | 96 | 9.1 | 32 | 13.0 |
| $500 – 999 | 293 | 27.8 | 68 | 27.6 |
| $1000 – 1999 | 221 | 20.9 | 46 | 18.7 |
| $2000 or more | 110 | 10.4 | 14 | 5.7 |
| Mean ± SD | Mean (SD) | Range | Mean (SD) | Range |
| Age | 42.5 (12.3) | 18–65 | 41.1 (12.2) | 18–65 |
| PANAS positive | 37.2 (9.4) | 10–50 | 39.6 (8.7) | 10–50 |
| PANAS negative | 21.8 (9.1) | 10–50 | 22.0 (9.7) | 10–49 |
| Overall trauma exposure (TEI) | 5.1 (3.9) | 0–14 | 5.5 (3.3) | 0–14 |
| Depressive Symptom Severity (BDI) | 15.7 (12.5) | 0–58 | 15.7 (13.0) | 0–57 |
| PTSD Symptom Severity (PSS) | 13.7 (12.4) | 0–50 | 15.5 (12.7) | 0–50 |
| Body Mass Index (BMI) | - | - | 32.1 (7.87) | 18–70 |
| Serum CRP | - | - | 5.1 (4.8) | 0.03–18.8 |
General Sample: Does not sum to 1055; Sex: 1 subject did not report; Race: 3 subjects did not report; Employment: 7 subjects did not report; Education: 7 subjects did not report; Income: 26 subjects did not report; CRP Sample: Does not sum to 246; Sex: 0 subject did not report; Race: 16 subjects did not report; Employment: 0 subjects did not report; Education: 1 subjects did not report; Income: 7 subjects did not report
Positive and negative affect was measured using the Positive and Negative Affect Schedule (PANAS), a 20-item self-report measure of positive and negative affect (Watson, D. et al., 1988). The PANAS, developed as a 20-item scale in 1988 (Watson, D. et al., 1988), measures two dominant dimensions of emotional experience as positive and negative affect that account for most of the variance in self-rated affect. Across multiple studies, the PANAS has shown good reliability and validity (Crawford and Henry, 2004). Cronbach’s alpha coefficients for the present study were 0.90 for positive affect subscale and 0.88 for negative affect subscale, demonstrating good internal validity. The two PANAS subscales were used to assess positive affect and negative affect in our sample.
Self-report of physical health status was measured using the Short Form Health Survey (SF-36), a 36-item self-report of health status (Ware Jr and Sherbourne, 1992). This survey consists of 8 scales of eight health domains, four of which are part of the summary measure of physical health, and the other four of mental health. The four physical health scales are categorized as physical functioning (10 items), role limitations due to physical health (role-physical; 4 items), bodily pain (2 items), and general health (5 items). Each item can be standardized to a 0 to 100 range, and then all items of each scale can be summed and averaged to obtain a scaled score, which is norm-based to a mean of 50, and standard deviation of 10. Construct validity, internal consistency, and reliability of the SF-36 and each of its individual domain scales has been demonstrated in numerous large general populations with divergent socio-demographic features and rates of disease states along with efficacy in the detection of differences in health between groups (Sullivan et al., 1995; Ware Jr, 2000).
Measurement of Serum C-reactive Protein (CRP)
Blood samples were obtained during the morning between the hours of 9:00 AM and 10:00 AM by venipuncture and stored at -80° C until time of assay. Serum CRP levels were measured using an immunoturbidometric assay from Sekisui Diagnostics (www.sekisuidiagnostics.com) on the Beckman AU480 chemistry analyzer, with an inter-assay coefficient of variation (CV) of 5.2% and an intra-assay CV of 3.1%. CRP levels averaged 5.14 ± 4.77 mg/L and ranged from 0.03 to 18.84 mg/L. CRP was collected on average within two weeks of psychological assessment.
Statistical Analysis
First, bivariate correlations were used to assess the association between positive and negative affect and measures of health including subjective ratings of general health, bodily pain, physical functioning, and physical limitations, as well as BMI and serum CRP. A series of linear regression models were then used to examine the associations of positive and negative affect with physical health outcomes. Potential confounders that have been previously associated with positive and negative affect, including age, sex, income, trauma exposure, and current depressive and post-traumatic symptoms, were controlled for in linear regression analyses.
Separate models examining the association between positive and negative affect independent of sociodemographic variables (i.e., age, gender, and income) with each physical health outcome were first run. Then, four models including both positive and negative affect were run covarying for sociodemographic variables to determine the unique effects of positive and negative affect on outcome variables. Finally, stepwise models were run to examine associations of variables of interest and covariates with the four physical health domain scales of the SF-36 and serum CRP; step 1) age, sex, income, and trauma load, step 2) current PTSD and depression symptoms, and step 3) positive or negative affect (run separately). Positive and negative affect were entered as the third and final step to examine the incremental validity in our predictive linear regression model. BMI was included as a covariate in step 1 for our stepwise regression models predicting CRP concentrations. The data were analyzed using SPSS (v.21).
Results
Sample characteristics
Study participants predominantly self-identified as African-American in both the primary sample (92.1%) and in the serum CRP subsample (91.9%; Table 1). Rates of participation by other racial groups were similar between the primary sample and subsample. Female participants (66.3%) outnumbered male participants in the primary sample and male participants (57.3%) outnumbered female participants in the subsample. With respect to the primary sample, most participants were unemployed (77.5%) with household income less than $2000 per month and education level of high school diploma or less (69.7%). Socioeconomic characteristics for the CRP subsample were similar. Study participants experienced high levels of trauma with 95.1% of the sample reporting exposure to at least one type of traumatic event (see Table 1 for descriptive details). Psychiatric symptom counts were also high (mean BDI score = 15.7 ± 12.4 and mean PSS score = 13.7 ± 12.4 in the primary sample; similar rates were found in the subsample.
Bivariate Correlations of Positive and Negative Affect with Physical Health Outcomes
Bivariate correlations were used to calculate Pearson correlation coefficients between positive and negative affect subscales of the PANAS and the SF-36 physical health domain scales. The correlation between positive and negative affect in the overall sample was 0.340 (p<0.001). All four SF-36 physical health domain scales: general health (n=1031, r=0.288), physical functioning (n=821, r=0.231), bodily pain (n=1041, r=0.229), role-physical (n=832, n=0.234) were positively correlated with positive affect (p<0.001, Table 2). Conversely, the same four scales: general health (n=1033, r=−0.240), physical functioning (n=823, r=−0.146), bodily pain (n=1043, r=−0.235), and role-physical (n=834, r=−0.194) were inversely correlated with negative affect (p<0.001, Table 2).
Table 2.
Main effects of SF-36 scales and Serum CRP on positive and negative affect.
| Positive Affect | Negative Affect | |||||
|---|---|---|---|---|---|---|
| n | r | p-value | N | r | p-value | |
| General Health | 1031 | 0.288 | <0.001 | 1033 | −0.240 | <0.001 |
| Physical Functioning | 821 | 0.231 | <0.001 | 823 | −0.146 | <0.001 |
| Bodily Pain1 | 1041 | 0.229 | <0.001 | 1043 | −0.235 | <0.001 |
| Role-Physical | 832 | 0.234 | <0.001 | 834 | −0.194 | <0.001 |
| Serum CRP | 245 | −0.137 | 0.038 | 246 | 0.038 | 0.551 |
| TEI Total Score | 942 | −0.150 | <0.001 | 942 | 0.256 | <0.001 |
| PSS Total Score | 1022 | −0.288 | <0.001 | 1024 | 0.527 | <0.001 |
| BDI Total Score | 1037 | −0.512 | <0.001 | 1038 | 0.636 | <0.001 |
Lower self-reported pain is reported as higher scores on the respective SF-36 measure
Bivariate Correlations of Positive and Negative Affect with Trauma load, PTSD, and Depressive Symptoms
Trauma load, (n=942, r=−0.150, p<0.001), PTSD (n=1022, r=−0.288, p<0.001), and depression (n=1037, r=−0.512, p<0.001) symptoms were negatively correlated with positive affect (Table 2). Trauma load, (n=942, r=0.256, p<0.001), PTSD (n=1024, r=0.527, p<0.001), and depression (n=1038, r=0.636, p<0.001) symptoms were positively correlated with negative affect (Table 2).
Bivariate Correlations of Positive and Negative Affect with serum CRP
Serum CRP was inversely correlated with positive affect (n=245, r=−0.137, p=0.038). However, serum CRP was not correlated with negative affect (n=246, r=0.038, p=0.551).
Linear Regression Models of Affect and Physical Health Outcomes
Initial linear regression models controlling for sociodemographic variables only showed that positive affect was significantly associated with all four physical health outcome variables (all p’s<.001; see Table 3). Similarly, negative affect was significantly associated with all four physical health outcome variables independent of sociodemographic variables (all p’s<.001; see Table 3). To examine the unique effects of positive and negative affect, additional models were run including both positive and negative affect while controlling for sociodemographic variables; as shown in Table 3, results were similar with both positive and negative affect showing significant associations across all four physical health outcomes (all p’s<.05).
Table 3.
Linear regression analyses of positive and negative affect with measures of physical health controlling for sociodemographic variables
| b | SE b | β (beta) | F | R2 | |
|---|---|---|---|---|---|
| Positive Affect | |||||
|
| |||||
| General Health | 68.014*** | 0.158 | |||
| Age | −0.546 | 0.059 | −0.272 | ||
| Sex | 1.072 | 1.516 | 0.021 | ||
| Income | 1.386 | 0.524 | 0.078** | ||
| PANAS Positive | 0.634 | 0.077 | 0.243*** | ||
|
| |||||
| Bodily Pain1 | 28.472*** | 0.112 | |||
| Age | −0.622 | 0.075 | −0.252 | ||
| Sex | −1.426 | 1.932 | −0.022 | ||
| Income | 1.072 | 0.667 | 0.048 | ||
| PANAS Positive | 0.606 | 0.098 | 0.188*** | ||
|
| |||||
| Physical Functioning | 28.250*** | 0.142 | |||
| Age | −0.700 | 0.080 | −0.293*** | ||
| Sex | −3.441 | 2.013 | −0.058 | ||
| Income | 2.067 | 0.727 | 0.095** | ||
| PANAS Positive | 0.552 | 0.104 | 0.178*** | ||
|
| |||||
| Role – Physical | 29.787 | 0.122 | |||
| Age | −0.616 | 0.088 | −0.237*** | ||
| Sex | 0.976 | 2.204 | 0.015 | ||
| Income | 2.413 | 0.796 | 0.102** | ||
| PANAS Positive | 0.655 | 0.113 | 0.195*** | ||
|
| |||||
| Negative Affect | |||||
|
| |||||
| General Health | 67.803*** | 0.161 | |||
| Age | −0.605 | 0.059 | −0.301*** | ||
| Sex | 1.110 | 1.516 | 0.022 | ||
| Income | 1.554 | 0.522 | 0.087** | ||
| PANAS Negative | −0.667 | 0.077 | −0.250*** | ||
|
| |||||
| Bodily Pain1 | 65.067*** | 0.131 | |||
| Age | −0.688 | 0.074 | −0.278*** | ||
| Sex | −1.200 | 1.909 | −0.019 | ||
| Income | 1.107 | 0.655 | 0.050 | ||
| PANAS Negative | −0.787 | 0.098 | −0.239*** | ||
|
| |||||
| Physical Functioning | 14.926*** | 0.128 | |||
| Age | −0.748 | 0.081 | −0.313*** | ||
| Sex | −3.940 | 2.025 | −0.066 | ||
| Income | 2.200 | 0.733 | 0.101** | ||
| PANAS Negative | −0.405 | 0.105 | −0.130*** | ||
|
| |||||
| Role – Physical | 32.917*** | 0.117 | |||
| Age | −0.682 | 0.088 | −0.262*** | ||
| Sex | 0.495 | 2.198 | 0.008 | ||
| Income | 2.430 | 0.796 | 0.102** | ||
| PANAS Negative | −0.650 | 0.113 | −0.192*** | ||
|
| |||||
| Both | |||||
|
| |||||
| General Health | 52.791*** | 0.190 | |||
| Age | −0.569 | 0.059 | −0.283*** | ||
| Sex | 1.513 | 1.492 | 0.030 | ||
| Income | 1.182 | 0.517 | 0.666* | ||
| PANAS Positive | 0.477 | 0.080 | 0.183*** | ||
| PANAS Negative | −0.481 | 0.081 | −0.181*** | ||
|
| |||||
| Bodily Pain1 | 40.286*** | 0.146 | |||
| Age | −0.657 | 0.073 | −0.266*** | ||
| Sex | −0.884 | 1.898 | −0.014 | ||
| Income | 0.813 | 0.655 | 0.037 | ||
| PANAS Positive | 0.389 | 0.103 | 0.120*** | ||
| PANAS Negative | −0.655 | 0.103 | −0.198*** | ||
|
| |||||
| Physical Functioning | 16.780*** | 0.148 | |||
| Age | −0.709 | 0.080 | −0.297*** | ||
| Sex | −3.282 | 2.009 | −0.055 | ||
| Income | 1.923 | 0.728 | 0.088** | ||
| PANAS Positive | 0.470 | 0.110 | 0.152*** | ||
| PANAS Negative | −0.250 | 0.110 | −0.080* | ||
|
| |||||
| Role – Physical | 25.361*** | 0.140 | |||
| Age | −0.640 | 0.087 | −0.246*** | ||
| Sex | 1.227 | 2.183 | 0.019 | ||
| Income | 2.122 | 0.792 | 0.089** | ||
| PANAS Positive | 0.493 | 0.119 | 0.146*** | ||
| PANAS Negative | −0.487 | 0.119 | −0.144*** | ||
P<.05
P<.01
P<.001
Lower self-reported pain is reported as higher scores on the respective SF-36 measure
Stepwise Models with Positive Affect and Physical Health Outcomes
Results for the full stepwise models including positive affect with all covariates are shown in Table 4. An overall model accounting for age, gender, income, trauma load, PTSD and depression symptoms, and positive affect was statistically significant and accounted for 21.2% of the variance in general health outcomes (p<0.001). In Step 3, positive affect had a significant positive association with general health outcomes above and beyond demographic and trauma variables and psychiatric symptoms (p=0.001; Table 4).
Table 4.
Stepwise linear regression analyses of predictor variables for positive affect with measures of physical health
| b | SE b | β (beta) | F-change | R2 | R2 change | |
|---|---|---|---|---|---|---|
| General Health | ||||||
|
| ||||||
| Step 1: | 29.023*** | 0.119 | 0.119 | |||
| Age | −0.561 | 0.066 | −0.277*** | |||
| Sex | 0.647 | 1.683 | 0.013 | |||
| Income | 1.913 | 0.567 | 0.109** | |||
| Trauma Load | −0.972 | 0.240 | −0.131*** | |||
| Step 2: | 44.856*** | 0.202 | 0.083 | |||
| Age | −0.575 | 0.063 | −0.283*** | |||
| Sex | 2.937 | 1.623 | 0.058 | |||
| Income | 1.048 | 0.548 | 0.060 | |||
| Trauma Load | −0.397 | 0.258 | −0.054 | |||
| PSS Score | 0.083 | 0.085 | 0.042 | |||
| BDI Score | −0.658 | 0.081 | −0.330*** | |||
| Step 3: | 11.305** | 0.212 | 0.010 | |||
| Age | −0.549 | 0.063 | −0.271*** | |||
| Sex | 2.991 | 1.613 | 0.059 | |||
| Income | 0.921 | 0.546 | 0.052 | |||
| Trauma Load | −0.378 | 0.257 | −0.051 | |||
| PSS Score | 0.067 | 0.085 | 0.034 | |||
| BDI Score | −0.537 | 0.088 | −0.269*** | |||
| PANAS Positive | 0.317 | 0.094 | 0.118** | |||
|
| ||||||
| Bodily Pain | ||||||
|
| ||||||
| Step 1: | 23.635*** | 0.098 | 0.098 | |||
| Age | −0.677 | 0.082 | −0.272*** | |||
| Sex | −1.431 | 2.100 | −0.023 | |||
| Income | 1.381 | 0.704 | 0.064 | |||
| Trauma Load | −1.163 | 0.298 | −0.127*** | |||
| Step 2: | 43.499*** | 0.180 | 0.082 | |||
| Age | −0.713 | 0.078 | −0.287*** | |||
| Sex | 1.183 | 2.025 | 0.019 | |||
| Income | 0.435 | 0.681 | 0.020 | |||
| Trauma Load | 0.028 | 0.322 | 0.003 | |||
| PSS Score | −0.352 | 0.107 | −0.144** | |||
| BDI Score | −0.507 | 0.101 | −0.206*** | |||
| Step 3: | 0.504 | 0.181 | 0.001 | |||
| Age | −0.706 | 0.079 | −0.284*** | |||
| Sex | 1.196 | 2.026 | 0.019 | |||
| Income | 0.402 | 0.682 | 0.019 | |||
| Trauma Load | 0.034 | 0.322 | 0.004 | |||
| PSS Score | −0.356 | 0.107 | −0.146** | |||
| BDI Score | −0.475 | 0.111 | −0.193*** | |||
| PANAS Positive | 0.084 | 0.118 | 0.025 | |||
|
| ||||||
| Physical Functioning | ||||||
|
| ||||||
| Step 1: | 21.737*** | 0.116 | 0.116 | |||
| Age | −0.757 | 0.090 | −0.312*** | |||
| Sex | −3.834 | 2.230 | −0.065 | |||
| Income | 2.658 | 0.791 | 0.123** | |||
| Trauma Load | −0.325 | 0.319 | −0.038 | |||
| Step 2: | 12.002*** | 0.147 | 0.031 | |||
| Age | −0.764 | 0.089 | −0.315*** | |||
| Sex | −2.069 | 2.223 | −0.035 | |||
| Income | 2.058 | 0.788 | 0.095** | |||
| Trauma Load | 0.192 | 0.357 | 0.022 | |||
| PSS Score | −0.042 | 0.122 | −0.018 | |||
| BDI Score | −0.415 | 0.114 | −0.177*** | |||
| Step 3: | 6.161* | 0.154 | 0.008 | |||
| Age | −0.735 | 0.090 | −0.303*** | |||
| Sex | −1.940 | 2.215 | −0.033 | |||
| Income | 1.956 | 0.786 | 0.091* | |||
| Trauma Load | 0.217 | 0.356 | 0.025 | |||
| PSS Score | −0.052 | 0.121 | −0.022 | |||
| BDI Score | −0.297 | 0.123 | −0.126* | |||
| PANAS Positive | 0.324 | 0.130 | 0.102* | |||
|
| ||||||
| Role – Physical | ||||||
|
| ||||||
| Step 1: | 17.112*** | 0.092 | 0.092 | |||
| Age | −0.690 | 0.100 | −0.259*** | |||
| Sex | −0.127 | 2.461 | −0.002 | |||
| Income | 2.616 | 0.871 | 0.111** | |||
| Trauma Load | −0.735 | 0.352 | −0.078* | |||
| Step 2: | 24.409*** | 0.154 | 0.062 | |||
| Age | −0.718 | 0.096 | −0.270*** | |||
| Sex | 2.427 | 2.411 | 0.037 | |||
| Income | 1.685 | 0.853 | 0.071* | |||
| Trauma Load | 0.310 | 0.386 | 0.033 | |||
| PSS Score | −0.318 | 0.131 | −0.125* | |||
| BDI Score | −0.455 | 0.123 | −0.177*** | 3.630 | 0.158 | 0.005 |
| Step 3: | ||||||
| Age | −0.693 | 0.097 | −0.261*** | |||
| Sex | 2.534 | 2.407 | 0.039 | |||
| Income | 1.595 | 0.853 | 0.068 | |||
| Trauma Load | 0.328 | 0.385 | 0.035 | |||
| PSS Score | −0.324 | 0.131 | −0.127* | |||
| BDI Score | −0.357 | 0.133 | −0.139** | |||
| PANAS Positive | 0.270 | 0.142 | 0.078 | |||
P<.05
P<.01
P<.001
Lower self-reported pain is reported as higher scores on the respective SF-36 measure
For bodily pain, the overall model including age, gender, income, trauma load, PTSD and depression symptoms, and positive affect was not statistically significant (p=0.478; Table 4). Adding positive affect in the final step further did not significantly explain additional variance in bodily pain above and beyond other variables included in the model (Table 4).
An overall model accounting for age, gender, income, trauma load, PTSD and depression symptoms, and positive affect was statistically significant and accounted for 15.4% of the variance in physical functioning (p<0.05; Table 4). In the final step, positive affect had a significant positive association with physical functioning above and beyond demographic and trauma variables and psychiatric symptoms (p=0.013; Table 4).
For physical role limitations, the overall model including age, gender, income, trauma load, PTSD and depression symptoms, and positive affect was not statistically significant (p=0.057; Table 4). Adding positive affect in the final step further did not significantly explain additional variance in physical role limitations above and beyond other variables included in the model, although the association approached significance (p=0.057; Table 4).
Stepwise Models with Negative Affect and Physical Health Outcomes
Results for the full stepwise models including negative affect with all covariates are shown in Table 5. An overall model predicting general health outcomes and accounting for age, gender, income, trauma load, PTSD and depression symptoms, and negative affect was not statistically significant (p=0.084). Adding negative affect in the final step further did not significantly explain additional variance in overall health above and beyond other variables included in the model (Table 5). For bodily pain, the overall model including age, gender, income, trauma load, PTSD and depression symptoms, and negative affect was not statistically significant (p=0.524). Adding negative affect in the final step further did not significantly explain additional variance in bodily pain above and beyond other variables included in the model (Table 5). Similarly for physical functioning, the overall model including age, gender, income, trauma load, PTSD and depression symptoms, and negative affect was not statistically significant (p=0.698). Adding negative affect in the final step further did not significantly explain additional variance in physical functioning above and beyond other variables included in the model (Table 5). Lastly, for physical role limitations, the overall model including age, gender, income, trauma load, PTSD and depression symptoms, and negative affect was also not statistically significant (p=0.848). Adding negative affect in the final step did not significantly explain additional variance in physical role limitations above and beyond other variables included in the model (Table 5).
Table 5.
Stepwise linear regression analyses of predictor variables for negative affect with measures of physical health
| b | SE b | β (beta) | F-change | R2 | R2 change | |
|---|---|---|---|---|---|---|
| General Health | ||||||
|
| ||||||
| Step 1: | 29.023*** | 0.119 | 0.119 | |||
| Age | −0.561 | 0.066 | −0.277*** | |||
| Sex | 0.647 | 1.683 | 0.013 | |||
| Income | 1.913 | 0.567 | 0.109** | |||
| Trauma Load | −0.972 | 0.240 | −0.131*** | |||
| Step 2: | 44.856*** | 0.202 | 0.083 | |||
| Age | −0.575 | 0.063 | −0.283*** | |||
| Sex | 2.937 | 1.623 | 0.058 | |||
| Income | 1.048 | 0.548 | 0.060 | |||
| Trauma Load | −0.397 | 0.258 | −0.054 | |||
| PSS Score | 0.083 | 0.085 | 0.042 | |||
| BDI Score | −0.658 | 0.081 | −0.330*** | |||
| Step 3: | 2.996 | 0.205 | 0.003 | |||
| Age | −0.578 | 0.063 | −0.285*** | |||
| Sex | 2.948 | 1.621 | 0.058 | |||
| Income | 1.072 | 0.547 | 0.061 | |||
| Trauma Load | −0.385 | 0.258 | −0.052 | |||
| PSS Score | 0.108 | 0.086 | 0.055 | |||
| BDI Score | −0.587 | 0.090 | −0.294*** | |||
| PANAS Negative | −0.184 | 0.106 | −0.069 | |||
|
| ||||||
| Bodily Pain | ||||||
|
| ||||||
| Step 1: | 23.635*** | 0.098 | 0.098 | |||
| Age | −0.677 | 0.082 | −0.272*** | |||
| Sex | −1.431 | 2.100 | −0.023 | |||
| Income | 1.381 | 0.704 | 0.064 | |||
| Trauma Load | −1.163 | 0.298 | −0.127*** | |||
| Step 2: | 43.499*** | 0.180 | 0.082 | |||
| Age | −0.713 | 0.078 | −0.287*** | |||
| Sex | 1.183 | 2.025 | 0.019 | |||
| Income | 0.435 | 0.681 | 0.020 | |||
| Trauma Load | 0.028 | 0.322 | 0.003 | |||
| PSS Score | −0.352 | 0.107 | −0.144** | |||
| BDI Score | −0.507 | 0.101 | −0.206*** | |||
| Step 3: | 0.406 | 0.181 | 0.001 | |||
| Age | −0.715 | 0.078 | −0.287*** | |||
| Sex | 1.184 | 2.026 | 0.019 | |||
| Income | 0.449 | 0.681 | 0.021 | |||
| Trauma Load | 0.033 | 0.322 | 0.004 | |||
| PSS Score | −0.340 | 0.108 | −0.140** | |||
| BDI Score | −0.475 | 0.113 | −0.193*** | |||
| PANAS Negative | −0.085 | 0.133 | −0.026 | |||
|
| ||||||
| Physical Functioning | ||||||
|
| ||||||
| Step 1: | 21.737*** | 0.116 | 0.116 | |||
| Age | −0.757 | 0.090 | −0.312*** | |||
| Sex | −3.834 | 2.230 | −0.065 | |||
| Income | 2.658 | 0.791 | 0.123** | |||
| Trauma Load | −0.325 | 0.319 | −0.038 | |||
| Step 2: | 12.002*** | 0.147 | 0.031 | |||
| Age | −0.764 | 0.089 | −0.315*** | |||
| Sex | −2.069 | 2.233 | −0.035 | |||
| Income | 2.058 | 0.788 | 0.095** | |||
| Trauma Load | 0.192 | 0.357 | 0.022 | |||
| PSS Score | −0.042 | 0.122 | −0.018 | |||
| BDI Score | −0.415 | 0.114 | −0.177*** | |||
| Step 3: | 0.150 | 0.147 | 0.000 | |||
| Age | −0.764 | 0.089 | −0.315*** | |||
| Sex | −2.078 | 2.225 | −0.035 | |||
| Income | 2.064 | 0.789 | 0.096** | |||
| Trauma Load | 0.197 | 0.358 | 0.023 | |||
| PSS Score | −0.035 | 0.123 | −0.015 | |||
| BDI Score | −0.393 | 0.128 | −0.167** | |||
| PANAS Negative | −0.057 | 0.147 | −0.018 | |||
|
| ||||||
| Role – Physical | ||||||
|
| ||||||
| Step 1: | 17.112*** | 0.092 | 0.092 | |||
| Age | −0.690 | 0.100 | −0.259*** | |||
| Sex | −0.127 | 2.461 | −0.002 | |||
| Income | 2.616 | 0.871 | 0.111** | |||
| Trauma Load | −0.735 | 0.352 | −0.078* | |||
| Step 2: | 24.409*** | 0.154 | 0.062 | |||
| Age | −0.718 | 0.096 | −0.270*** | |||
| Sex | 2.427 | 2.411 | 0.037 | |||
| Income | 1.685 | 0.853 | 0.071* | |||
| Trauma Load | 0.310 | 0.386 | 0.033 | |||
| PSS Score | −0.318 | 0.131 | −0.125* | |||
| BDI Score | −0.455 | 0.123 | −0.177*** | |||
| Step 3: | 0.037 | 0.154 | 0.000 | |||
| Age | −0.717 | 0..097 | −0.270*** | |||
| Sex | 2.434 | 2.413 | 0.037 | |||
| Income | 1.682 | 0.854 | 0.071 | |||
| Trauma Load | 0.307 | 0.386 | 0.032 | |||
| PSS Score | −0.321 | 0.132 | −0.126* | |||
| BDI Score | −0.467 | 0.139 | −0.182** | |||
| PANAS Negative | 0.031 | 0.160 | 0.009 | |||
P<.05
P<.01
P<.001
Lower self-reported pain is reported as higher scores on the respective SF-36 measure
Positive Affect and Serum CRP
A separate stepwise linear regression with serum CRP as the dependent variable, accounting for age, sex, BMI, income, trauma load, current PTSD and depression symptoms, and positive affect as the predictor was run. The final step of this model including positive affect (F=1.333, df=172, p=0.250) was not significant (Table 6). Only BMI was a significant predictor of CRP concentrations (t=4.736, p<0.001; Table 6).
Table 6.
Stepwise linear regression analyses of predictor variables of positive affect with serum CRP concentrations.
| b | SE b | β (beta) | F-change | R2 | R2 change | |
|---|---|---|---|---|---|---|
| Serum CRP | ||||||
|
| ||||||
| Step 1: | 6.958*** | 0.172 | 0.172 | |||
| Age | −0.032 | 0.044 | −0.053 | |||
| Sex | 1.562 | 1.060 | 0.109 | |||
| Income | −0.783 | 0.380 | −0.145* | |||
| Trauma Load | 0.096 | 0.161 | 0.043 | |||
| BMI | 0.302 | 0.063 | 0.352*** | |||
| Step 2: | 0.078 | 0.173 | 0.001 | |||
| Age | −0.033 | 0.044 | −0.055 | |||
| Sex | 1.547 | 1.074 | 0.108 | |||
| Income | −0.784 | 0.384 | −0.146 | |||
| Trauma Load | 0.105 | 0.179 | 0.048 | |||
| BMI | 0.299 | 0.064 | 0.348*** | |||
| PSS Score | −0.021 | 0.061 | −0.036 | |||
| BDI Score | 0.022 | 0.058 | 0.038 | |||
| Step 3: | 1.333 | 0.180 | 0.007 | |||
| Age | −0.037 | 0.044 | −0.061 | |||
| Sex | 1.436 | 1.078 | 0.101 | |||
| Income | −0.783 | 0.384 | −0.145 | |||
| Trauma Load | 0.090 | 0.180 | 0.041 | |||
| BMI | 0.301 | 0.063 | 0.349** | |||
| PSS Score | −0.015 | 0.061 | −0.026 | |||
| BDI Score | −0.007 | 0.063 | −0.013 | |||
| PANAS Positive | −0.088 | 0.076 | −0.095 | |||
Discussion
We observed significant relationships between positive and negative affect and all four domains of physical health functioning, and CRP concentrations as a measure of general inflammatory state, with effect sizes ranging from 0.14–0.29. Our results show that even after conservatively controlling for the effects of demographic variables, trauma exposure, and current PTSD and depression symptoms, a significant association remains between positive, but not negative, affect and self-reported general health ratings and physical functioning as measured by self-report via the SF-36. To the best of our knowledge, our study is the first reported examination of the relationship between positive and negative affect and health outcomes in a predominantly African-American, heavily trauma exposed, and low socio-economic status population. Our findings are consistent with previous studies in which increased trait positive affect was associated with fewer symptoms of illness and better self-reported health across a variety of clinical populations in ambulatory as well as inpatient settings (Pressman and Cohen, 2005). Other reports have also identified an association between trait positive affect and improved health outcomes in healthy populations along with evidence suggesting that the effects of positive emotion are independent of, and possibly stronger, than the effects of negative emotion on health outcomes (Andreasson et al., 2013; Pressman and Cohen, 2005).
Negative affect was also associated with worse reported health, increased bodily pain, worse physical functioning, and greater physical role limitations. However, these associations did not remain significant when we controlled for the effects of demographic factors, trauma load, and depression and PTSD symptoms. The loss of association between negative affect and these sub-scales of the SF-36 when controlling for psychiatric symptoms may be a consequence of overlap between negative affect and symptoms of anxiety, post-traumatic stress and especially depression (Watson, David et al., 1988). These results also suggest that the effects of negative affect on these health outcomes seem to be fully accounted for by the presence of psychiatric symptoms. As such, this is an important limitation of our statistical model and by extension, our study findings. Previous reports have suggested that chronic pain negatively influences affect/mood rather than negative affect predisposing one to pain (Gaskin et al., 1992). However, causation cannot be determined from our main effects model results given the cross-sectional nature of our analysis.
In a subset of participants from our primary sample, we observed a significant inverse correlation between serum CRP levels and positive affect, but did not observe significant correlation between serum CRP and negative affect. However, the association between positive affect and serum CRP did not remain significant when examined in our stepwise regression model controlling for demographic variables, trauma load, and depressive and post-traumatic stress symptoms. Our failure to find a correlation between negative affect and CRP is particularly surprising considering the extent of previous reports describing an association between depression and PTSD with CRP (Fernandes et al., 2016; Michopoulos et al., 2015; Valkanova et al., 2013), and could be due to inadequate statistical power, considering the smaller size of our subsample.
The significance of the association between increased age and both decreased positive and negative affect in our model may reflect findings of natural changes in measured affect in longitudinal population samples over time (Mroczek and Kolarz, 1998). The significant association between increased depressive symptoms and decreased positive affect, and between increased depressive symptoms and increased negative affect may reflect prior findings that depressive symptoms may result from combination of both low positive affect and high negative affect (Watson, David et al., 1988). Even considering the high trauma burden observed in our sample, the effect of current depressive symptoms appeared to be more contributory in our model to both positive and negative affect as compared to current PTSD symptoms. The high trauma burden in our cohort itself can be theorized to lead to adverse health outcomes through negative physiological effects such as observed increase in inflammatory markers. However, as described before, affect was found to have a separate effect from trauma exposure and PTSD symptoms in our statistical models.
The central strengths of our study are the relatively large sample size and the demographic focus on lower socioeconomic status, heavily trauma-exposed African-American primary care patients. Studying this particular sample may limit generalizability to other populations, yet the homogeneity of the sample may increase validity of our results to similar populations of lower socioeconomic status, highly traumatized African-Americans, that tend to be understudied populations in biomedical research. As recruitment occurred in primary care settings, this population may reflect a more general population than a specific disease or mental health population.
Limitations of our study include somewhat superficial measurement of pain through the SF-36, rather than with a more comprehensive pain index such as the McGill Pain Questionnaire (MPQ) (Katz and Melzack, 2011). In addition, our measures of health outcomes were based on subjective self-report without objective correlation. Related to this concern, previously reported studies of patient symptom reporting and affect indicate that when objective measures of disease are held constant, individuals with high trait positive affect report fewer severe symptoms from disease (Cohen et al., 2003). As such, the subjective reporting style of the SF-36 and the positive affect scale may confound the current results. Consequently, we may progressively underestimate the true burden of disease in individuals with increased positive affect. An additional study limitation is related to our assessment of affect with the PANAS. The PANAS places a significant demand on the ability of research participants to discriminate closely related emotions and the capacity of their vocabulary to correctly recognize terminology describing distinct, yet closely-related, emotions. Variability in the recognition and discrimination between closely-related emotions and the interpretation of emotion terminology by study participants may be a random effect, systematic effect, or some combination of both. In the case of a random effect, such variability would be expected to limit statistical power to detect, as opposed to confound, our measurement of positive and negative affect on health measures. Systematic effects on response to the PANAS may also be present and related possibly to effects of trauma on educational achievement and vocabulary. Another study limitation is the cross-sectional manner in which study data were acquired, limiting our capacity to consider causal relationships between study variables. Additionally, because we had not collected data on smoking status of participants we could not include that as a covariate in CRP analyses. Finally, serum CRP was not obtained at the same time the self-report psychological measures (depression, PTSD, and affect) and the length of time between assessment sessions varied by participant. As a result, serum CRP levels may be more reflective of the psychological and medical status of subjects at the time of collection as opposed to the time of their initial study assessment and it is impossible to know how much affect in particular varied between those sessions. However, we would anticipate that these differences represent random, as opposed to systematic, error that may have contributed to the negative findings in our subsample. Our clinical interview could be argued to have induced a temporary change in state affect, in either a positive or negative direction, and would be prohibitively difficult to objectively quantify.
In summary, we found an independent association between positive affect and self-reported physical health in a large, homogeneous, highly trauma-exposed and impoverished population cohort. Studies have examined the interaction and possible moderating effect of positive affect on stress, pain, and negative affect, and this method could serve as a model for future examination of our study data (Zautra et al., 2005). Growing public interest in positive psychology and related formal and informal psychotherapeutic interventions may reflect a greater empiric realization of the contribution of positive emotions and individual traits to general well-being. Including more physiologic markers of disease severity including autonomic, cardiovascular, immunologic, and endocrine parameters in future analysis could provide further insights and possible confirmation of the effect of positive affect on physical health outcomes. Following the stress-buffering hypothesis, the contribution of individual resilience to physical health and physiological markers should be further examined. Positive psychology and its uses in not only ill, but healthy and at-risk populations, holds significant promise as a powerful tool for public health in physical and mental wellness. As our understanding of what is considered pathology – negative health outcomes and PTSD – has grown, we believe that the focus of research has and will shift to concepts such as positive health, resilience, and posttraumatic growth.
Acknowledgments
The current study was supported by MH096764, MH071537 (KJR), MH082256 (CFG), HD071982 (BB), MH102890 (APL), HD085850 (VM),the Emory and Grady Memorial Hospital General Clinical Research Center, NIH National Centers for Research Resources (M01RR00039), NARSAD (CFG), The Burroughs Welcome Fund (KJR), the Howard Hughes Medical Institute (KJR), the Atlanta Clinical Translational Science Institute, the NIH National Centers for Research Resources (M01 RR00039), and the Emory University General Clinical Research Center at Grady Hospital. APW is supported by the VA CDA IK2CX000601. The contents of this manuscript do not reflect the views of the Department of Veterans Affairs or the United States Government. This study would not have been possible without the research expertise and technical assistance of all the staff, volunteers, and participants of the Grady Trauma Project.
Footnotes
Disclosures: All authors have no conflicts of interest.
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 citable 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
- Andreasson AN, Szulkin R, Unden AL, von Essen J, Nilsson LG, Lekander M. Inflammation and positive affect are associated with subjective health in women of the general population. J Health Psychology. 2013;18(3):311–320. doi: 10.1177/1359105311435428. [DOI] [PubMed] [Google Scholar]
- Beck AT, Steer RA, Carbin MG. Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review. 1988;8(1):77–100. [Google Scholar]
- Blazer DG, Hybels CF. What Symptoms of Depression Predict Mortality in Community-Dwelling Elders? Journal of the American Geriatrics Society. 2004;52(12):2052–2056. doi: 10.1111/j.1532-5415.2004.52564.x. [DOI] [PubMed] [Google Scholar]
- Bradley B, DeFife JA, Guarnaccia C, Phifer J, Fani N, Ressler KJ, Westen D. Emotion dysregulation and negative affect: association with psychiatric symptoms. The Journal of Clinical Psychiatry. 2011;72(5):685–691. doi: 10.4088/JCP.10m06409blu. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chida Y, Steptoe A. Positive psychological well-being and mortality: a quantitative review of prospective observational studies. Psychosomatic Medicine. 2008;70(7):741–756. doi: 10.1097/PSY.0b013e31818105ba. [DOI] [PubMed] [Google Scholar]
- Cohen S, Doyle WJ, Turner RB, Alper CM, Skoner DP. Emotional style and susceptibility to the common cold. Psychosomatic Medicine. 2003;65(4):652–657. doi: 10.1097/01.psy.0000077508.57784.da. [DOI] [PubMed] [Google Scholar]
- Crawford JR, Henry JD. The positive and negative affect schedule (PANAS): construct validity, measurement properties and normative data in a large non-clinical sample. The British Journal of Clinical Psychology / the British Psychological Society. 2004;43(Pt 3):245–265. doi: 10.1348/0144665031752934. [DOI] [PubMed] [Google Scholar]
- Diener E, Larsen RJ, Levine S, Emmons RA. Intensity and frequency: dimensions underlying positive and negative affect. Journal of Personality and Social Psychology. 1985;48(5):1253. doi: 10.1037//0022-3514.48.5.1253. [DOI] [PubMed] [Google Scholar]
- Diener E, Suh EM, Lucas RE, Smith HL. Subjective well-being: Three decades of progress. Psychological Bulletin. 1999;125(2):276. [Google Scholar]
- Dockray S, Steptoe A. Positive affect and psychobiological processes. Neuroscience & Biobehavioral Reviews. 2010;35(1):69–75. doi: 10.1016/j.neubiorev.2010.01.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fernandes BS, Steiner J, Molendijk ML, Dodd S, Nardin P, Goncalves CA, Jacka F, Kohler CA, Karmakar C, Carvalho AF, Berk M. C-reactive protein concentrations across the mood spectrum in bipolar disorder: a systematic review and meta-analysis. Lancet Psychiatry. 2016;3(12):1147–1156. doi: 10.1016/S2215-0366(16)30370-4. [DOI] [PubMed] [Google Scholar]
- Foa EB, Riggs DS, Dancu CV, Rothbaum BO. Reliability and validity of a brief instrument for assessing post-traumatic stress disorder. Journal of Traumatic Stress. 1993;(6):459–473. [Google Scholar]
- Gaskin M, Greene AF, Robinson ME, Geisser M. Negative affect and the experience of chronic pain. Journal of Psychosomatic Research. 1992;36(8):707–713. doi: 10.1016/0022-3999(92)90128-o. [DOI] [PubMed] [Google Scholar]
- Gillespie CF, Bradley B, Mercer K, Smith AK, Conneely K, Gapen M, Weiss T, Schwartz AC, Cubells JF, Ressler KJ. Trauma exposure and stress-related disorders in inner city primary care patients. General Hospital Psychiatry. 2009;31(6):505–514. doi: 10.1016/j.genhosppsych.2009.05.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Golden SH, Williams JE, Ford DE, Yeh HC, Sanford CP, Nieto FJ, Brancati FL. Depressive symptoms and the risk of type 2 diabetes the atherosclerosis risk in communities study. Diabetes Care. 2004;27(2):429–435. doi: 10.2337/diacare.27.2.429. [DOI] [PubMed] [Google Scholar]
- Hemingway H, Marmot M. Psychosocial factors in the aetiology and prognosis of coronary heart disease: systematic review of prospective cohort studies. BMJ. 1999;318(7196):1460–1467. doi: 10.1136/bmj.318.7196.1460. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Huffman JC, Legler SR, Boehm JK. Positive psychological well-being and health in patients with heart disease: a brief review. Future Cardiology. 2017 doi: 10.2217/fca-2017-0016. [DOI] [PubMed] [Google Scholar]
- Huppert FA, Whittington JE. Symptoms of psychological distress predict 7-year mortality. Psychological Medicine. 1995;25(05):1073–1086. doi: 10.1017/s0033291700037569. [DOI] [PubMed] [Google Scholar]
- Katz J, Melzack R. The McGill Pain Questionnaire: Development, psychometric properties, and usefulness of the long-form, short-form, and short-form-2. In: Turk DC, Melzack R, editors. Handbook of pain assessment. Guilford Press; New York: 2011. pp. 45–66. [Google Scholar]
- Koivumaa-Honkanen H, Honkanen R, Viinamäki H, Heikkilä K, Kaprio J, Koskenvuo M. Self-reported life satisfaction and 20-year mortality in healthy Finnish adults. American Journal of Epidemiology. 2000;152(10):983–991. doi: 10.1093/aje/152.10.983. [DOI] [PubMed] [Google Scholar]
- Michopoulos V, Rothbaum AO, Jovanovic T, Almli LM, Bradley B, Rothbaum BO, Gillespie CF, Ressler KJ. Association of CRP genetic variation and CRP level with elevated PTSD symptoms and physiological responses in a civilian population with high levels of trauma. The American Journal of Psychiatry. 2015;172(4):353–362. doi: 10.1176/appi.ajp.2014.14020263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Miller AH, Maletic V, Raison CL. Inflammation and its discontents: the role of cytokines in the pathophysiology of major depression. Biological Psychiatry. 2009;65(9):732–741. doi: 10.1016/j.biopsych.2008.11.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mroczek DK, Kolarz CM. The effect of age on positive and negative affect: a developmental perspective on happiness. Journal of Personality and Social Psychology. 1998;75(5):1333. doi: 10.1037//0022-3514.75.5.1333. [DOI] [PubMed] [Google Scholar]
- Ostir GV, Markides KS, Black SA, Goodwin JS. Emotional well-being predicts subsequent functional independence and survival. Journal of the American Geriatrics Society. 2000 doi: 10.1111/j.1532-5415.2000.tb04991.x. [DOI] [PubMed] [Google Scholar]
- Ostir GV, Markides KS, Peek MK, Goodwin JS. The association between emotional well-being and the incidence of stroke in older adults. Psychosomatic Medicine. 2001;63(2):210–215. doi: 10.1097/00006842-200103000-00003. [DOI] [PubMed] [Google Scholar]
- Penninx BW, Kritchevsky SB, Yaffe K, Newman AB, Simonsick EM, Rubin S, Ferrucci L, Harris T, Pahor M. Inflammatory markers and depressed mood in older persons: results from the Health, Aging and Body Composition study. Biological Psychiatry. 2003;54(5):566–572. doi: 10.1016/s0006-3223(02)01811-5. [DOI] [PubMed] [Google Scholar]
- Pressman SD, Cohen S. Does positive affect influence health? Psychological bulletin. 2005;131(6):925. doi: 10.1037/0033-2909.131.6.925. [DOI] [PubMed] [Google Scholar]
- Rasmussen HN, Scheier MF, Greenhouse JB. Optimism and physical health: a meta-analytic review. Annals of Behavioral Medicine. 2009;37(3):239–256. doi: 10.1007/s12160-009-9111-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Robertson SM, Stanley MA, Cully JA, Naik AD. Positive emotional health and diabetes care: concepts, measurement, and clinical implications. Psychosomatics. 2012;53(1):1–12. doi: 10.1016/j.psym.2011.09.008. [DOI] [PubMed] [Google Scholar]
- Sullivan M, Karlsson J, Ware JE., Jr The Swedish SF-36 Health Survey—I. Evaluation of data quality, scaling assumptions, reliability and construct validity across general populations in Sweden. Social science & Medicine. 1995;41(10):1349–1358. doi: 10.1016/0277-9536(95)00125-q. [DOI] [PubMed] [Google Scholar]
- Valkanova V, Ebmeier KP, Allan CL. CRP, IL-6 and depression: a systematic review and meta-analysis of longitudinal studies. Journal of Affective Disorders. 2013;150(3):736–744. doi: 10.1016/j.jad.2013.06.004. [DOI] [PubMed] [Google Scholar]
- Ware JE., Jr SF-36 health survey update. Spine. 2000;25(24):3130–3139. doi: 10.1097/00007632-200012150-00008. [DOI] [PubMed] [Google Scholar]
- Ware JE, Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36): I. Conceptual framework and item selection. Medical Care. 1992:473–483. [PubMed] [Google Scholar]
- Watson D, Clark LA, Carey G. Positive and negative affectivity and their relation to anxiety and depressive disorders. Journal of Abnormal Psychology. 1988;97(3):346. doi: 10.1037//0021-843x.97.3.346. [DOI] [PubMed] [Google Scholar]
- Watson D, Clark LA, Tellegen A. Development and validation of brief measures of positive and negative affect: the PANAS scales. Journal of Personality and Social Psychology. 1988;54(6):1063–1070. doi: 10.1037//0022-3514.54.6.1063. [DOI] [PubMed] [Google Scholar]
- Zautra AJ, Johnson LM, Davis MC. Positive affect as a source of resilience for women in chronic pain. Journal of Consulting and Clinical Psychology. 2005;73(2):212. doi: 10.1037/0022-006X.73.2.212. [DOI] [PMC free article] [PubMed] [Google Scholar]