Introduction

Driving is an important instrumental activity of daily living (ADL), which allows older adults to access healthcare resources, shop for daily necessities, and socialize and engage in recreational activities. However, studies show that the burden of chronic diseases may lead to driving cessation, which may in turn lead to a steeper decline in health and symptoms of depression among older adults [1,2,3,4]. While there is paucity of literature studying gender and ethnic disparities in driving behavior and related health outcomes [5], the limited data hints at a widening of the gap with age, with women and nonwhite men being more likely to stop driving than white men [6]. Also, limited data demonstrates that women belonging to ethnic minorities make up a large proportion of never-driver older adults and are underutilizers of healthcare resources [7, 8].

Our study is a comparative analysis of self-reported driving status and comorbid health conditions of older African Americans receiving care at a safety net hospital in Atlanta, to study utilization of healthcare resources and chronic disease burden. We hypothesized that utilization of healthcare resources would be lower in nondriving older adults due to transportation constraints, and therefore, chronic disease burden would be higher in this group compared with drivers of similar age at Grady Hospital.

Materials and Methods

We conducted a chart review of patients aged 65 and older who sought care at the Emma Darnell Geriatric Center at Grady Hospital during the period 2/1/2016–2/1/2017. The study was approved by the Institutional Review Board at the Emory University and the Research Oversight Committee at the Grady Health System. Data obtained from chart review included age, gender, race/ethnicity, healthcare insurance information, functional/ADL datasheet (which included data on self-reported ADLs, hearing and vision impairment, and driving status), number of visits to the Emma Darnell Geriatric Clinic during the 12-month period, and associated visit diagnoses. The original Katz [9] and Lawton [10] ADL scales were modified based on available ADL data to calculate scores. Exclusion criteria included patients under age 65 at the start of the study period (2/1/2016) or those with missing data on driving status.

Analysis

Raw data were obtained on 868 patients. After exclusion of patients who were under 65 at the start of the study period or had missing data about driving status, 690 patient charts were included in the final analysis. For patients with multiple visits, data was analyzed based on the most recent clinic visit. Self-reported driving status was analyzed and compared for all variables. Driving status was graded as “driving” if the individual reported driving anytime in the past year, even if not driving to the appointment. Driving status was graded as “nondriving” if the individual reported not actively driving in the past year. Prevalence of disease conditions in drivers and nondrivers was compared using Pearson’s chi-squared tests. Multiple logistic regression analysis was used to determine diseases in drivers and nondrivers across age, gender, and other demographic characteristics. Analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC). To determine the difference in cumulative chronic diseases between drivers and nondrivers, a “score” was computed from the available data on whether an individual had a specific chronic condition or not. The seven variables, which included leading seven visit diagnoses, were hypertension, diabetes, cognitive disorders, chronic kidney disease (CKD), chronic obstructive pulmonary disease (COPD), cerebrovascular disease (CVD), and cancer. The maximum score was seven while the minimum was zero. These visit diagnoses were compiled from all clinic visits during the study period.

Variables

Functional status scores were calculated using a modified version of the Katz index of independence in ADL [9] and the Lawton-Brody Instrumental ADL scale [10]. As our dataset did not have information on “mode of transportation” and “continence,” a modified scoring excluding these categories was used for the analysis. Additionally, due to inadequate information on each of the categories, a score of 1 was assigned for independency and 0 for partial or full dependency in each ADL category. The mean values were obtained by dividing the total score by the number of categories, 7 for Lawton scale and 5 for Katz scale.

The utilization of healthcare was measured by the mean number of Geriatric Clinic visits per year among patients with single and multiple visits. The T test was conducted to assess for statistical significance. Health insurance information was categorized with driving status in cross tabulation tables. Any face-to-face clinic encounters with either physician, nurse practitioner, or physician-assistant were included in the number of clinic visits. The number of clinic visits per year was analyzed by the type of healthcare insurance using basic descriptive statistics.

Results

Data from the medical records of 690 patients who had at least one visit to Emma Darnell Geriatric Clinic, were aged 65 or older on 2/1/2016, and had information about driving frequency included in their medical records was reviewed. Demographics are listed in Table 1. One hundred and fifty-seven (23%) of a total 690 patients included in our analysis were drivers, and the remaining 533 (77%) were nondrivers. Approximately 95% of individuals in both the driving and nondriving groups identified themselves as African Americans. Seventy-six percent were women. Twenty percent of the women in our cohort reported being drivers compared with 32% of the men. Fifty percent of our patients were enrolled in Medicare Parts A and B, while 40% were enrolled in a Medicare advantage plan. The remaining 10% had only Medicare B or Medicaid.

Table 1 Demographic, functional, and sensory status
Full size table

Utilization of Heathcare Resources

As seen in Table 1, the utilization of healthcare resources, as measured by mean number of Geriatric Clinic visits in a year, was similar among drivers (1.2 mean visits per year) and nondrivers (1.3 mean visits per year). While a total of 268 patient visits to Geriatric Clinic in the year were by participants in a Medicare advantage plan, only one participant had 4 visits in the year, whereas 205 out of 268 Medicare advantage plan participants (76%) had only one clinic visit a year.

Sensory Impairment

The prevalence of self-reported visual impairment was slightly higher among nondrivers compared with drivers (34% vs 29%) (see Table 1). Also, 17 people reported being blind, all in the nondriving group. Corrective lenses were used by 54% of the drivers and 35% of the nondrivers.

The prevalence of self-reported hearing impairment was 14% among drivers and 16% among nondrivers (Table 1). Only two patients reported being deaf, both in the nondriver group.

Functional Impairment

The use of assistive devices was more prevalent among nondrivers compared with drivers. A total 26% of the nondrivers vs 8% of the drivers used a walker; 11% of the nondrivers vs 1.2% of the drivers used a wheelchair. The ADL scores indicate lower levels of functional dependency in the driving group compared with the nondriving group, with mean modified Katz score being 4.9 among the driving group vs 4.2 among the nondriving group and mean modified Lawton score being 6.6 among the driving group vs 4.4 among the nondriving group (see Table 1).

Comorbid Conditions

The leading medical diagnoses associated with clinic visits included hypertension, diabetes, cognitive disorders, CKD, COPD, CVD, and cancer (Table 2). The nondrivers were found to have greater odds of cognitive impairment compared with drivers in our group (OR 2.69, 95% CI 1.532–4.736). However, there was no statistically significant difference in prevalence of any other diagnoses between the driving and nondriving groups.

Table 2 Chronic diseases and driving status
Full size table

We analyzed the distribution of cumulative chronic diseases across the driving statuses. While individuals with 4 or more chronic conditions were less likely to drive, the results were not statistically significant. This effect could likely be due to a greater number of nondrivers in our sample.

Discussion

In our cohort of primarily African American older adults receiving care at the outpatient Geriatric Clinic of a safety net hospital, two-thirds of which were women, barely one-fifth of patients reported driving a motor vehicle. While the drivers were functionally more independent and less likely to have cognitive impairment than the nondrivers, the number of Geriatric Primary Care Clinic visits in a year was low in both groups when compared with the reported utilization of primary care services of 4–8 visits per person per year in older adults [11, 12]. Additionally, the number of clinic visits was also low for patients enrolled in Medicare advantage plans, which generally cover transportation to appointments. A previously published study of utilization of healthcare resources in rural neighborhood in North Carolina’s Appalachian counties identified possession of driver’s license, having family/friend available for driving, and access to public transportation (esp. in older women) as predictors of higher number of chronic care and regular care visits in a year [13]. While our patient population is urban, a majority reside in the underserved areas of the Metro Atlanta, where there may be limited access to public transportation. Additionally, considering prevalence of functional, sensory, and cognitive impairment in a large percentage of our patient population, they may not be able to utilize public transportation independently and may need a caregiver to accompany them to medical appointments.

Our study had a few limitations. (1) The data on comorbid conditions were obtained by compiling visit diagnoses, which may not be reflective of the full comorbid profile or problem list of the patients. (2) The driving status question did not specifically assess for driving to appointments. While older adults may drive to nearby grocery stores or gas stations, they may be less willing to drive to traffic-congested zones like hospital-based clinics in downtown neighborhoods and therefore may still be dependent on alternative modes of transportation or a driver when visiting these locations. (3) Our data did not include information on past driving status to differentiate between never-drivers and former drivers. (4) We defined utilization of healthcare by number of visits to the Geriatric Primary Care Clinic at Grady Hospital. Our data did not include information about urgent care or emergency room visits.

There is need for further research to study the impact of transportation-related barriers on utilization of healthcare among older adults living in underserved urban communities, including availability of a caregiver/companion and reliable and affordable transportation. Additionally, while alternative modes of transportation including ridesharing e-hail services are gaining increasing popularity in urban centers, knowledge about and use of these services is generally low in older Americans, esp. those belonging to low-income groups [14]. This opens up a venue for researchers to collaborate with and pilot ridesharing options which may be available to and affordable for older adults living in underserved urban communities.

Conclusion

Approximately 77% of older adults getting care at a safety net hospital in Atlanta reported not driving a motor vehicle. The drivers in our sample were functionally more independent and were less likely to have cognitive impairment compared with the nondriving group. However, utilization of healthcare resources, as measured by the number of Geriatric Clinic visits per year, was low in both groups and was not affected by enrollment in a Medicare advantage plan.