Data were analyzed from the 2006–2010 Behavioral Risk Factor Surveillance System (BRFSS) and the BRFSS Adult Asthma Call-back Survey (ACBS).^{11, 12} BRFSS is an ongoing, state-based, random-digit-dialed telephone survey of non-institutionalized U.S. adults aged ≥18 years.¹¹ Before 2011, BRFSS interviews were conducted by landline telephone.¹¹ Detailed information about BRFSS’ multistage sampling design and sample weighting is available elsewhere.¹¹ ACBS is a follow-up telephone survey administered to BRFSS participants who indicated that they ever had asthma (i.e., they responded ‘yes’ to the question, “have you ever been told by a doctor, nurse, or other health professional that you had asthma?”).¹³ ACBS is conducted approximately two weeks after BRFSS.¹³ Jurisdiction-specific Institutional Review Board (IRB) requirements apply to each participating state, territory, and the District of Columbia; informed consent was obtained.¹³ Analyses of BRFSS and ACBS data are exempt from review by the IRB at the Centers for Disease Control and Prevention.

To obtain a sufficiently large sample to produce stable estimates, Adult BRFSS and ACBS data were combined from 2006–2010.¹³ Data were weighted to represent adults aged ≥18 years with asthma from 40 states, the District of Columbia, and Puerto Rico (i.e., U.S. areas from which Adult BRFSS and ACBS data were available during 2006–2010).¹² Additional information on participating states, weight calculations, and response rates is located online and in prior publications.^{12, 14, 15}

This study was focused on adults aged ≥65 years with active asthma (Figure 1). As in previous analyses of BRFSS and ACBS data,^16–18 respondents were classified as having active asthma if they reported in ACBS that at least one of the following occurred during the past 12 months: talked to a doctor or other health professional about his or her asthma, took asthma medication, or experienced any symptoms of asthma.

For background information, the study population of adults aged ≥65 years with active asthma was compared to adults aged 18–64 years with active asthma (Online Repository Table E1). The same definition for active asthma was applied to both groups.

Respondent characteristics obtained from BRFSS data were age (at time of interview), sex, race/ethnicity, U.S. census region, annual household income, and smoking history. Smoking history was assessed using responses to the questionnaire item, “Have you smoked at least 100 cigarettes in your entire life?”

Clinical comorbidity information regarding coronary artery disease (CAD), diabetes, and obesity were obtained from BRFSS data. Respondents were classified with CAD if they answered “yes” to ≥1 of the following two questionnaire items: “Has a doctor, nurse, or other health professional ever told you that you had a heart attack, also called a myocardial infarction?” or “Has a doctor, nurse, or other health professional ever told you that you had angina or coronary heart disease?” History of diabetes was determined using the questionnaire item “Has a doctor, nurse, or other health professional ever told you that you had diabetes?” Gestational diabetes, prediabetes, and borderline diabetes were not included in the definition of diabetes for this analysis. Individuals with a body mass index (BMI) ≥30 kg/m² calculated from their reported height and weight were categorized as obese.

Characteristics obtained from ACBS responses included time since asthma diagnosis and asthma medication use in the past three months. Time since asthma diagnosis was determined from the questionnaire item, “How long ago was it [when you were first told by a doctor or other health professional that you had asthma]?” Asthma medication use in the past three months was classified into four categories: none, short-acting β agonist (SABA) only, ≥1 asthma controller medication (i.e., inhaled corticosteroids, long-acting β agonists, leukotriene modifiers, methylxanthines, cromolyn, nedocromil, or immunomodulators), and oral corticosteroid only (without SABA or asthma controller medication[s]).

Asthma control was assessed using previously published methods that were adapted from the 2007 National Asthma Education and Prevention Program Guidelines for the Diagnosis and Management of Asthma (EPR-3) and account for the ACBS design (Table E2).^{14, 19, 20} Briefly, three impairment measures were assessed: daytime symptoms, night-time symptoms, and SABA use (other than for preventing exercise-induced bronchospasm) during the past three months. Asthma control was defined using the highest impairment level identified for each adult across these three measures.

Asthma severity was measured according to EPR-3 guidelines and previously published methods, which classified asthma severity using information on asthma control and controller medication use.^{15, 19, 20} Briefly, respondents whose asthma was well controlled without asthma controller medication use in the past three months were categorized as having intermittent asthma. Persistent asthma included all respondents who reported using asthma controller medication (regardless of asthma control status) and respondents not reporting asthma controller medication use but meeting criteria for not well controlled or very poorly controlled asthma. Additional information on defining asthma control and severity using ACBS data is available in prior publications.^{14, 15, 20}

Information on asthma-related hospitalizations was obtained from the questionnaire item, “During the past 12 months, that is since [1 year ago today], have you had to stay overnight in a hospital because of your asthma? Do not include an overnight stay in the emergency room.” Information on asthma-related ED/UCV was obtained from the questionnaire item, “During the past 12 months, have you had to visit an emergency room or urgent care center because of your asthma?”

Regarding clinical comorbidities, ACBS data were used to identify respondents with a history of chronic obstructive pulmonary disease (COPD) or depression. Respondents were classified with COPD if they answered “yes” to ≥1 of the following three questionnaire items: “Have you ever been told by a doctor or health professional that you have chronic obstructive pulmonary disease, also known as COPD?”, “Have you ever been told by a doctor or other health professional that you have emphysema?”, or “Have you ever been told by a doctor or other health professional that you have chronic bronchitis?” History of depression was determined using the questionnaire item “Have you ever been told by a doctor or other health professional that you were depressed?”

Regarding potentially modifiable factors, cost as a barrier to health care was investigated using questionnaire responses concerning inability to buy medication, see a primary care physician, or see a specialist for asthma in the past 12 months. Home environment variables were related to secondhand smoke, pets allowed in respondents’ bedrooms, reports of seeing or smelling mold in the home in the past 30 days, and report of seeing cockroaches, mice, or rats inside the home in the past 30 days.

As in previous analyses, uninformative responses (i.e., don’t know, missing, refused) were considered negative responses (i.e., not having the relevant outcome).^{16, 18, 20, 21}

Sampling weights were used to calculate population estimates representative of 40 states, the District of Columbia, and Puerto Rico (Table E2). These sampling weights accounted for BRFSS and ACBS nonresponse and unequal sampling probabilities.¹⁷ To properly account for the complex sample survey design, the entire dataset (N = 74,209 ever diagnosed with asthma) was retained in the final analyses. Given this investigation’s focus on adults aged ≥65 years with active asthma, only results generated from the stratum of the population of 14,076 older adults with active asthma (Figure 1) are presented in the main tables.

Complex sample survey procedures were used to compare continuous and categorical variables (PROC SURVEYMEANS and PROC SURVEYFREQ in SAS, respectively).^{22, 23} Multivariate logistic regression was performed to calculate odds ratios and 95% confidence intervals (CIs) for associations between asthma-related hospitalizations or ED/UCV and all variables of interest; results were adjusted for sex, race/ethnicity, annual household income, smoking history, and asthma severity. Statistical significance was set at a P value <0.05. Results based on <50 unweighted respondents (i.e., percentages based upon a denominator <50) or for which the relative standard error (standard error/prevalence estimate; a measure of an estimate’s reliability) was >0.3 were suppressed.¹² Identical procedures were used in secondary analyses examining potential sex differences in variables relevant to asthma control, as well as sensitivity analyses (stratification by COPD status; modified active asthma definition using a 3-month time frame for asthma medication use instead of the conventional 12-month time frame). These analyses were conducted in SAS version 9.3 (SAS Institute Inc., Cary, NC).

Population attributable fractions (PAFs) were calculated using the formula PAF = (p[RR-1])/(p[RR-1]+1), where RR is the relative risk and p is the proportion of the study population with the risk factor; RR was estimated by the adjusted OR. Similarly, PAF 95% CIs were estimated using adjusted OR 95% CIs.

The percentage of older adults with ≥1 asthma-related hospitalization varied significantly by geographic region (P=.03; Table I). Compared to older adults with no asthma-related hospitalizations in the past year, those with ≥1 asthma-related hospitalization were more likely to be Black, female, and report annual household incomes <$25,000. Additional demographic information is presented in Figure E1. Persistent asthma, very poorly controlled asthma, and use asthma controller medication was more common among older adults with ≥1 asthma-related hospitalization. The two groups did not differ significantly in smoking history (P=.06) or age (weighted average ± standard error=73.1±0.5 years for ≥1 asthma-related hospitalization vs. 73.4±0.1 years for no asthma-related hospitalizations, P=.5).

In multivariate analyses (Table II), asthma-related hospitalizations among older adults were associated with COPD, CAD, depression, and diabetes, but not obesity. Compared to older adults without asthma-related hospitalizations, adjusted odds were higher among older adults with asthma-related hospitalization(s) for being unable to see a primary care doctor for asthma, see a specialist for asthma, or buy medication for asthma because of cost. Also, signs of mold or cockroaches in the home were associated with asthma-related hospitalizations. Estimates of PAFs (Table E3) indicated the variable with the largest overall PAF was COPD (0.63; 95% CI=0.50–0.73); of modifiable factors examined, the largest PAF was associated with financial barriers to buying asthma medication (0.14; 95% CI=0.08–0.21).

Factors associated with asthma-related ED/UCV were similar to those for hospitalizations (Table III). COPD, CAD, depression, and diabetes were more common among older adults with ≥1 asthma-related ED/UCV, as were the inability to see a primary care doctor for asthma, see a specialist for asthma, or buy medication for asthma because of cost. Compared to older adults without asthma-related ED/UCV, adjusted odds were higher among older adults with asthma-related ED/UCV for signs of mold in the home. In contrast to asthma-related hospitalizations, asthma-related ED/UCV were not significantly associated with cockroaches in the home.

A sensitivity analysis of older adults with asthma but not COPD (n = 7,725) indicated financial barriers to asthma-related health care were significantly associated with asthma-related hospitalizations, but clinical comorbidities and home environment conditions were not (Table E4). In this population, obesity was the only variable examined that was found to be associated with asthma-related ED/UCV. Among older adults with both asthma and COPD (n = 6,351), CAD and financial barriers to asthma-related health care were significantly associated with asthma-related hospitalizations; besides these variables, depression was also associated with asthma-related ED/UCV in this population. Results from a sensitivity analysis using a modified definition of active asthma yielded results similar to the main analyses (Table E5), except diabetes and home environment conditions were not significantly associated with asthma-related hospitalizations or ED/UCV.

Compared to older men with active asthma, older women with active asthma were more likely to be Black and report annual household incomes <$25,000 (Table E6). The frequency of well controlled asthma was similar among men and women (40.2% [95% CI=37.2–43.2%] vs. 39.0% [95% CI=37.2–40.9%], respectively). Not well controlled asthma was more common among women (28.2% [95% CI=26.5–29.8%], compared to 23.5% [95% CI=21.0–26.0%] among men). In contrast, very poorly controlled asthma was more common among men (36.3% [95% CI=33.3–39.3%], compared to 32.8% [95% CI=31.1–34.5%] among women). Asthma medication use, asthma severity, and (weighted) average age did not differ significantly by sex. Men were more likely than women to report a lifetime smoking history of ≥100 cigarettes.

Compared to male older adults, adjusted odds were higher among female older adults for COPD, depression, and obesity (Table E7). In contrast, adjusted odds were lower among female older adults for CAD. Female sex among older adults was associated with being unable to see a primary care doctor for asthma or buy medication for asthma because of cost. No sex differences in home environment variables were observed.