Medical cost and frequent mental distress among the non-elderly US adult population

Abstract

Background

Frequent mental distress (FMD) is an important measure of perceived poor mental health. With the rising cost of health care, it is not uncommon for working adults to delay seeking care. The objective of this study was to determine the relationship between avoidance of medical care due to cost and FMD among the non-elderly US population.

Methods

We analyzed data from 282 044 non-elderly US population from a 2008 Behavioral Risk Factor Surveillance System survey. Multivariable logistic regression models were used to assess the association between avoidance of medical care due to cost and FMD adjusted for covariates.

Results

The overall prevalence of FMD in the non-elderly population was 11.1%; whereas it was 24.2% for those reporting avoiding medical care due to cost. Approximately 18% of the population had no health insurance coverage and the prevalence of FMD was significantly greater in this group. The odds of FMD were >2-fold elevated for respondents who were unable to see a doctor because of cost (adjusted odds ratio: 2.40, 99% confidence interval: 2.19, 2.63).

Conclusions

These findings highlight the need for affordable medical care for reducing mental distress and improving population health.

frequent mental distress, BRFSS, medical cost

Introduction

One of two primary overarching goals of the Healthy People 2010 initiative, a public health strategic plan intended to improve the health of the population, is to increase years and health-related quality of life (HRQOL),¹ and understanding factors that influence HRQOL is important in developing strategies to meet this goal.

A general decline in HRQOL was observed in the US population from 1993 to 2001, although significant variations existed among demographic and socioeconomic subgroups.²

Frequent mental distress (FMD) is an important measure of perceived poor mental health and has been used by the Centers for Disease Control and Prevention (CDC), as part of healthy days measures, to identify unmet mental health needs, track mental health over time and develop population health promotion programs.^3,4 It has been shown to be useful in assessing population health in terms of evaluating outcomes of primary care.^5–7

Mental health influences physical, social and emotional wellbeing of individuals.⁴ Mental illness, such as depression and anxiety, are diagnosable and treatable; however, poor mental health generally remains sub-clinical and difficult to diagnose. Studies have linked poor mental health to poor physical health, adverse cardiovascular outcomes and impaired psychosocial functioning.^8–10 This could have tremendous social and economic cost to the society.

In the USA, ∼18% of the non-elderly population lacks health-care coverage,¹¹ a growing segment of the insured has inadequate coverage¹² and costs are increasing for those covered by employer-sponsored plans.¹³ Additionally, an estimated 17% of adults in the USA avoid health-care costs by postponing care, not filling prescriptions, not seeing specialists and not following a doctor's advice.¹⁴

We used data from the CDC's 2008 Behavioral Risk Factor Surveillance System (BRFSS) to examine the association between the avoidance of medical care because of cost and FMD among the non-institutionalized non-elderly adult population of the USA.

Methods

Study population

The BRFSS is an annual state-based telephone survey of adults aged 18 years and older conducted by state health departments with methodological help provided by CDC. The BRFSS questions are developed jointly by CDC and the states. The survey questionnaire consists of (i) a core component, which include questions asked by all states; (ii) optional modules, which each state can choose to include and (iii) state-specific questions, which each individual state develops and administers to their state population.¹⁵ Each state conducts monthly telephone interviews according to the protocol provided by CDC. Each randomly selected telephone number is dialed up to 15 times on 4 different occasions on weekdays, weekends and weeknights. States submit their data to CDC and they make it available to public on their website (http://www.cdc.gov/brfss) after appropriate cleanup and validation check. To ensure quality control, 5% of telephone interviews were randomly selected for verification callbacks. The questions used in this analysis primarily came from the core component questionnaire.¹⁶ The 2008 BRFSS sample comprised 414 509 adults 18 years and older living in households in all US states, the District of Columbia, Puerto Rico, Guam and the US Virgin Islands. For this study, however, we initially selected 285 384 participants who were <65 years old and residing in 50 US states and the District of Columbia. The survey was conducted by computer assisted telephone interviews using a disproportionate stratified sample design with one telephone number representing a sample record. Telephone numbers were assigned to strata based on the density (high density or medium density) of household telephone numbers in a block of 100 numbers. Since BRFSS use a complex survey design, the following formula was used to calculate the final weight.

 

where ‘FINALWT is the final weight assigned to each study respondent; STRWT is the inverse of the sampling fraction of each stratum; 1/NPH is the inverse of the number of residential telephone numbers in the respondent's household; NAD is the number of adults in the respondent's household and POSTSTRAT is the number of people in an age-by-sex or age-by-race/ethnicity-by-sex category in the population of a region or a state divided by the sum of the preceding weights for the respondents in the same age-by-sex or age-by-race/ethnicity-by-sex category. It adjusts for non-coverage and non-response and forces the sum of the weighted frequencies to equal population estimates for the region or state’.¹⁵ The response rate for the 2008 BRFSS survey was 53.3%.

Dependent variables

The dependent variable of the study was FMD, which was defined based on the question: ‘Now thinking about your mental health, which includes stress, depression and problems with emotions, for how many days during the past 30 days was your mental health not good?’^3,4 The responses were coded as 1 if respondents reported 14 or more of the previous 30 days of mental health as not good, 0 otherwise. The 14-day cutoff is used by clinicians as a marker of clinical depression and anxiety disorders.³ There were 3340 missing values in the FMD variable and these observations were dropped from the analysis. The final study sample comprised 282 044 observations. BRFSS HRQOL measures have demonstrated acceptable construct validity, criterion validity and internal consistency.^17–19

Independent variables

The primary independent variable of the study is avoidance of medical care because of cost. Responses from the BRFSS core question ‘was there a time in the past 12 months when you needed to see a doctor but could not because of cost?’ was used as the data source for the primary independent variable. Since this is a yes/no question, responses were dichotomized.

Covariates

A number of important variables that have been linked to FMD were included in the study as potential confounders. Among these were health coverage; demographic characteristics including age groups (18–24, 25–34, 35–44, 45–54 and 55–64 years), sex (male and female), race/ethnicity (White, non-Hispanic, Black, non-Hispanic, Hispanic and Other race/ethnicity), marital status (grouped into married/unmarried couple, divorce/widowed/separated and never married), employment status (employed, unable to work, unemployed, homemaker, students and retired), income categories (<15, 15–<24.9, 25–34.9, 35–49.9 and 50K and above) and education (grouped into college or technical, no high school, high school and some college or technical education); behavioral and lifestyle characteristics such as, physical activity, smoking, alcohol use and body mass index (BMI) and chronic disease status including asthma, diabetes and cardiovascular disease (CVD). Health coverage was defined based on the question ‘Do you have any kind of health care coverage, including health insurance, prepaid plans such as HMOs, or government plans such as Medicare?’ Physical activity was defined based on responses to the question ‘during the past month, other than your regular job, did you participate in any physical activities or exercises such as running, calisthenics, golf, gardening, or walking for exercise?’ Alcohol use was defined based on responses to two questions: ‘during the past 30 days, have you had at least one drink of any alcoholic beverage such as beer, wine, a malt beverage or liquor? and ‘ …during the past 30 days, on the days when you drank, about how many drinks did you drink on the average?’

Data analysis

BRFSS use complex sample survey design. Therefore, all data analysis was performed using the STATA statistical software package version 11.0 (STATA Corp, College Station, TX.) taking into account survey characteristics (primary sampling unit, strata and sampling weight) of the data. The study sample was described using weight descriptive measures. The relationships between the primary independent variable and each of the dependent variables were tested using univariate logistic regression analysis followed by multiple logistic regression analysis controlling for potential confounding effects of the covariates. Because of a large sample, P < 0.01 was used to test for statistical significance and 99% confidence intervals were constructed in the logistic regression model.

Results

The total number of respondents that made up the sample was 282 044. Table 1 describes the characteristics of the study population and prevalence estimates of FMD for each independent variable. The prevalence of FMD in this population of adults under 65 years old was 11.1% [99% confidence interval (CI): 10.8, 11.4]. Almost 16% were unable to visit doctor because of cost and the prevalence of FMD in this group was 24.2%, more than twice the overall prevalence of FMD. The prevalence of FMD was significantly higher among those without health insurance (14.8%, 99% CI: 13.9–15.8). The majority of the respondents reported participating in physical activities and the prevalence of FMD was significantly lower in this group when compared with those who did not reported engaging in physical activities (P < 0.001). The estimated prevalence of FMD was 20% or greater for study participants who reported to be divorced/widowed/separated, unable to work or unemployed, reported earning <$15 000/year, were current smokers and had current asthma and CVD (Table 1).

Table 1

Descriptive characteristics of the study population and prevalence estimates of frequent mental distress

	Study population		Prevalence of FMD
	Weighted (%)	na	Weighted (%) (99% CI)
All	100.0	282 044	11.1 (10.8, 11.4)
Unable to see a doctor because of cost
No	84.0	241 389	8.6 (8.3, 8.9)
Yes	15.8	40 174	24.2 (23.1, 25.3)
Health insurance coverage
No	17.8	39 461	14.8 (13.9, 15.8)
Yes	81.8	241 932	10.2 (9.9, 10.5)
Physical activity
No	23.7	70 129	17.8 (17.0, 18.6)
Yes	76.2	211 694	8.9 (8.6, 9.3)
Sex
Male	50.0	109 160	9.0 (8.6, 9.4)
Female	50.0	172 884	13.1 (12.7, 13.5)
Age (years)
18–24	14.8	13 366	10.8 (9.7, 11.9)
25–34	21.9	38 270	10.8 (10.1, 11.6)
35–44	22.8	60 069	10.9 (10.3, 11.5)
45–54	22.7	81 928	11.9 (11.4, 12.5)
55–64	17.1	84 978	10.7 (10.2, 11.3)
Race/ethnicity
White, non-Hispanic	66.1	218 721	10.8 (10.5, 11.1)
Black, non-Hispanic	10.3	23 701	12.8 (11.8, 13.9)
Hispanic	15.6	20 015	11.2 (10.1, 12.3)
Other race/ethnicity	7.3	17 116	10.4 (9.4, 11.6)
Marital status
Married/unmarried couple	65.2	178 580	9.0 (8.7, 9.4)
Divorced/widowed/separated	12.6	59 957	19.5 (18.7, 20.4)
Never married	21.9	42 300	12.1 (11.3, 13.0)
Education
College or technical school graduate	34.5	104 302	6.6 (6.3, 7.0)
Did not graduate high school	10.5	21 261	17.8 (16.5, 19.1)
High school graduate	27.6	77 751	12.8 (12.2, 13.5)
Some college or technical school	27	78 008	12.2 (11.7, 12.9)
Employment status
Employed	69.3	196 889	8.3 (8.0, 8.6)
Unable to work	5.4	20 796	39.6 (37.8, 41.3)
Unemployed	6.8	15 073	19.9 (18.4, 21.6)
Homemaker	7.9	21 431	10.5 (9.6, 11.6)
Student	5.7	6257	9.9 (8.4, 11.7)
Retired	4.3	20 567	8.0 (7.2, 8.9)
Income
$50 000 or more	47.2	131 526	6.8 (6.5, 7.2)
<$15 000	8.0	22 696	24.7 (23.2, 26.3)
$15 000–$24 999	12.4	34 157	17.4 (16.4, 18.5)
$25 000–$34 999	9.1	26 586	12.7 (11.7, 13.8)
$35 000–$49 999	12.6	39 293	10.5 (9.7, 11.3)
Refused/missing	10.7	27 786	11.5 (10.4, 12.6)
BMI
Normal (BMI 18.51–24.99)	33.5	90 215	9.8 (9.3, 10.3)
Underweight (BMI ≤ 18.5)	1.7	4002	13.6 (11.3, 16.4)
Overweight (BMI 25.00–29.99)	34.1	95 757	9.8 (9.3, 10.4)
Obese (BMI ≥ 30)	26.2	78 772	14.3 (13.6, 14.9)
Smoking status
Never	58.5	153 672	8.1 (7.7, 8.5)
Current smoker	20.4	57 263	19.8 (18.9, 20.6)
Former smoker	20.8	70 238	10.8 (10.3, 11.4)
Drinking
None	44.6	127 902	12.5 (12.0, 12.9)
1–2 drinks	35.3	108 635	8.6 (8.2, 9.1)
3–4 drinks	10.6	25 769	10.5 (9.6, 11.5)
5 or more	5.6	11 371	15.3 (13.7, 17.1)

^aUnweighted n.

The odds of FMD were more than two times greater for study participants who were not able to visit their doctor due to cost [Model 4 adjusted odds ratio (OR) = 2.40, 99% CI: 2.19, 2.63]. The unadjusted odds of FMD were 35% lower for study participants with health coverage; however, after adjusted for potential confounders the odds ratio increased to 1.25 (99% CI: 1.13, 1.40) (Model 4; Table 2). There was no interaction between the cost of medical care and health insurance coverage (P = 0.412).

Table 2

Unadjusted and adjusted odds ratios of association between independent variables and FMD

	Model 1	Model 2	Model 3	Model 4
	OR (99% CI)	OR (99% CI)	OR (99% CI)	OR (99% CI)
Unable to see a doctor because of cost	3.40 (3.17, 3.65)	3.06 (1.82, 3.33)	2.59 (2.37, 2.82)	2.40 (2.19, 2.63)
Health insurance coverage	0.65 (0.60, 0.71)	1.23 (1.11, 1.35)	1.25 (1.13, 1.38)	1.25 (1.13, 1.40)

Note: health insurance coverage was included in Models 2–4 as a default confounding variable. Model 1: unadjusted univariable odds ratio. Model 2: model adjusted for lifestyle factors including physical activity, BMI, smoking and drinking behavior. Model 3: model adjusted for socio-demographic characteristics including age, sex, race/ethnicity, education, employment and household income. Model 4: model adjusted for all variables in Models 2 and 3 and the presence of comorbidities including asthma, diabetes and coronary heart disease.

Discussion

Main findings of this study

Access to health care is an important public health indicator outlined in Healthy People 2010 and improving access to care by reducing barriers is a significant public health goal.¹ The results of the study indicate that non-elderly individuals that reported avoiding medical care because of cost were more likely to report FMD, despite controlling for health insurance coverage, lifestyle factors, socio-demographic characteristics and presence of co-morbidities.

The odds of FMD were >2-fold greater among the non-elderly population who were unable to seek medical care due to cost (adjusted OR = 2.40, 99% CI: 2.19–2.63) and 25% greater among those with health coverage (adjusted OR: 1.25, 99% CI: 1.13–1.40). Zahran et al. ⁵ analyzed BRFSS data aggregated from 2003 through 2005 to identify risky health behaviors and demographic characteristics associated with HRQOL among young adults aged 18–24 years. More than 11% reported that they had not been able to see a doctor when needed at least once in the previous year because they could not afford care. Those affected by financial barrier to care reported more physically and mentally unhealthy days than those who were not affected by the barrier. Weinick et al.²⁰ reported that 6.8% of US adults postponed needed care because of cost; the proportion increased to 17% when other cost-related barriers such as not filling prescriptions, not seeing a specialist when necessary or using an alternative form of care were included. Similarly, Jarret et al.²¹ indicated that men have reported not receiving primary care and not seeing specialists when referred because of cost, and cost has been linked to failure to receive follow-up treatment for chronic conditions.²² In addition, trends indicate that the likelihood of non-elderly adults with chronic conditions having unmet medical needs because of cost increased from 1997 to 2006.²³ It is pertinent to note that the two age groups between 35 and 54 years of age in this study represented the largest groups of uninsured population and these groups are more prone to chronic health conditions which drive health-care costs when medical treatment is sought at a later stage in disease course.

Financial access to medical care is often synonymous with having health-care insurance.²⁴ However, health insurance does not necessarily provide all necessary financial resources needed to pay for care and those covered are frequently required to pay copayments or deductibles. Results from the RAND Health Insurance Experiment indicate that demand for medical services decreased as out of pocket costs increased.²⁵ Similarly, Wong et al.²⁶ found that when categorized by level of cost sharing, those in the high cost share group were less likely to seek medical care for minor symptoms (OR = 0.39, P < 0.001) and severe symptoms (OR = 0.22, P < 0.01) when compared with those with no cost share. These observations suggest that health-care costs affect care-seeking decisions regardless of insurance status.

Some limitations of this study should be noted. BRFSS is a telephone-based survey and generally excludes those who do not own a landline; who mostly belongs to low socioeconomic groups. As more people abandon landlines in favor of cell phone-only usage, selection bias may occur threatening the validity of survey results. However, beginning in 2010, CDC has started to include cell phone-only households in the BRFSS surveys. Institutionalized population, which may have higher prevalence of mental distress, is also excluded from the BRFSS survey. The degree to which individual respondents over- or understated their responses may have resulted in misclassification bias. However, any such misclassification is likely to be non-differential as there is no reason to believe that individuals with FMD are more likely than those without FMD to overstate or understate if they avoided medical care due to cost. Selection bias could be present if non-respondents were more likely to have FMD. In the absence of data on non-respondents, it was not possible to determine the extent of the bias. Finally, because of cross sectional in nature, we cannot make any conclusion regarding the direction of causality between FMD and avoidance of medical care due to costs.

In conclusion, this study found that avoidance of health care due to cost is associated with FMD in the non-elderly US population. This effect is independent of the health insurance coverage. Future research, using better study designs, is needed to evaluate the role health-care cost plays in seeking medical care and the impact of avoided medical care on FMD.

Conflict of interest

None declared.