Introduction
Brazil faces alarming traffic-related mortality rates, which are approximately ten times higher than those of the safest countries with up to 110 people killed in fatal crashes per day (Carvalho & Guedes, 2023). In addition to the road safety actions related to the built environment, engineering and human factors, it is essential to also consider underlying individual issues that may be impacting crash events, specifically health. A growing public health issue in Brazil is chronic kidney disease (CKD), with prevalence in the general population estimated to grow by 5 percent annually (Thomé et al., 2019). According to the census conducted by the Brazilian Society of Nephrology (SBN) in 2021, an estimated 144,779 patients regularly undergo dialysis therapy (Nerbass et al., 2022). The most common form of renal replacement therapy is haemodialysis (HD), which is performed in specific centres. This modality requires the patient to be present at the HD centres at least three times a four hour treatment per week, and treatment is recognised as negatively impacting their quality of life and life expectancy (Al-Mansouri et al., 2021).
Mobility, fundamental for independence and quality of life (Shumway-Cook et al., 2002), is defined as a person’s ability to move, whether from a chair to a bed, to the yard, street, or any other place they wish, using their own body, equipment, or means of transportation (Mayo & Mate, 2022). Efficient mobility is integrated into the World Health Organization’s (WHO) concept of public health and includes cognitive, psychosocial, physical, environmental, and financial aspects, significantly influenced by culture, gender, and personal history (Mayo & Mate, 2022; Webber et al., 2010; WHO, 2001).
A key predictor of quality of life is directly related to the degree of mobility to the sessions and can be complemented by the levels of accessibility to transportation and the individual’s own needs (Santos et al., 2017). For patients on HD, barriers to mobility are time spent in weekly treatment, increased dependence on family members, anxiety, and fatigue after sessions (Megari, 2013). In addition, CKD patients can have comorbidities (e.g., neurocognitive disorders, episodes of hypoglycaemia, sleep apnoea, etc.), as well as dialysis-related complications (e.g., nausea, hypotension, excessive fatigue), that can pose risks to themselves and other road users, when the patient is moving in traffic environments, particularly as the driver of a vehicle (Kepecs et al., 2018; Lacson Jr & Brunelli, 2011; Vats & Duffy, 2010).
Dialysis sessions shorter than four hours are associated with increased morbidity and mortality (Flythe et al., 2013). Thus, transportation plays an important role in the effectiveness of treatment. Patients must be able to travel, whether by driving, public transport, or other means, in a way that supports adequate clinical stability and allows them to complete full treatment sessions even when travel times are long (Blagg, 2005; Go et al., 2004; Ritt et al., 2007).
End-stage CKD has widespread systemic effects and is associated with higher mortality, more hospitalisations, and increased cardiovascular complications (Tylicki et al., 2022). It also requires strict treatment adherence and regular travel to the HD centre, regardless of the patient’s mobility. This highlights the importance of studying how patients access transportation to HD centres, especially in Brazil, where diverse geographical barriers may complicate travel.
Rather than blaming patients with CKD for their involvement in crashes (Carvalho & Guedes, 2023), this study evaluated patients’ choices in relation to travelling to the HD centre. It aims to identify CKD patients who continue driving despite having a condition and treatment that may impair their driving abilities, both before and shortly after a haemodialysis session (Kepecs et al., 2018; Lacson Jr & Brunelli, 2011; Vats & Duffy, 2010). Given the scarcity of studies in Brazil examining how haemodialysis patients commute and the extent to which they continue driving despite having impaired clinical conditions, this research addresses an important gap with potential implications for both clinical care and traffic safety. We hypothesise that a considerable number of HD patients maintain driving habits despite the clinical and cognitive impairments related to CKD and its comorbidities, which may pose risks not only to themselves but also to public road safety.
Methods
Study Overview
This was a quantitative, descriptive, cross-sectional study conducted with adult patients undergoing haemodialysis (HD) in São Paulo, Brazil. The study investigated transportation habits, driving history, and factors potentially related to traffic safety in this population. Data were generated through structured interviews between April 2021 and March 2022 and analysed using descriptive and inferential statistical methods.
Population and Setting
According to official data from the Brazilian Ministry of Health (DATASUS), approximately 8,800 patients were undergoing regular haemodialysis in the city during the study period (April 2021 to March 2022) (DATASUS, 2025). Access to dialysis centres was restricted in response to the COVID-19 pandemic. Eight HD centres were available for recruitment.
The interviews were conducted in accredited haemodialysis centres situated in different regions of the city, allowing access to a diverse patient population in terms of sociodemographic and clinical characteristics. Participants were recruited during routine haemodialysis sessions at both public and private facilities, with clinically stable patients approached while waiting for or undergoing treatment. All interviews were conducted in person by a trained physician, using a standardised questionnaire, based on the patient’s availability and comfort.
Eligibility Criteria and Selection
Eligibility criteria included adult patients (≥18 years old) undergoing regular haemodialysis therapy in accredited centres of São Paulo, Brazil, who were clinically stable and agreed to participate by signing the informed consent form. There were no restrictions regarding sex, age, ethnicity, or socioeconomic status. Exclusion criteria were: patients with acute clinical complications on the day of the interview (such as intercurrent medical conditions that could interfere with the questionnaire); history of stroke with residual neurological deficits, particularly language impairment; or, cognitive dysfunction that precluded reliable responses. Participants were consecutively recruited during their scheduled HD sessions, either in the morning or afternoon shifts, across all weekdays. Given the cross-sectional design of the study, no follow-up procedures were planned after the interview.
No formal sample size calculation was performed given the descriptive and exploratory design of the study. Nevertheless, the sample was considered adequate to allow reliable estimates of proportions with acceptable precision, as reflected in the use of 95 percent confidence intervals for the main outcomes.
Variables and Data Sources
The primary outcome was self-reported driving status after initiation of HD. Exposures and predictors included demographic variables (age, sex, education, employment status, marital status), clinical characteristics (aetiology of CKD, comorbidities, and dialysis-related complications), and transportation-related factors. Potential confounders considered were mobility limitations, access to transportation, and frequency of HD sessions.
All data were generated through a structured interview conducted by a trained physician using a standardised questionnaire. Demographic, transportation, and traffic-related data were obtained by patient self-report, while clinical information (including aetiology of CKD and comorbidities) was confirmed by medical records when available. Events such as traffic violations, collisions, vehicle damage, and intradialytic complications were reported for the six months preceding the interview.
Age was categorised into four groups (<20, 20-39, 40-59, and ≥60 years). Educational level (none, incomplete elementary, elementary, middle school, college) was classified according to the Brazilian educational system.
Comorbidities were assessed through a predefined checklist covering cardiovascular (e.g., hypertension, myocardial infarction, arrhythmia), endocrine (e.g., type I and type II diabetes, hypercholesterolemia), and sleep-related disorders (e.g., insomnia, sleep apnoea, restless legs syndrome). Intradialytic complications (e.g., hypotension, arrhythmia, nausea, tremors, insomnia, restless legs syndrome) were recorded as present or absent. Driver’s licence status and type were classified according to the official Brazilian system (categories A-E and combinations).
Bias Control
Several measures were implemented to minimise potential sources of bias. To reduce selection bias, patients were consecutively recruited from multiple HD centres, encompassing both public and private institutions, and approached during routine dialysis sessions to ensure representativeness. To limit information and recall bias, face-to-face interviews were conducted by trained physicians using a standardised questionnaire, with predefined categories. Whenever possible, clinical information (aetiology of CKD, comorbidities, dialysis schedule) was cross-checked with medical records. To mitigate response bias, participants were assured of confidentiality and anonymity, encouraging honest reporting of driving history and traffic incidents. Despite these efforts, we acknowledge that some degree of self-reporting bias may still be present, particularly regarding sensitive issues such as traffic violations.
Quantitative Variables
Quantitative variables such as age, duration of HD treatment, and weekly hours of dialysis were initially analysed as continuous data. Normality of distribution was assessed using the Kolmogorov-Smirnov test. When normally distributed, results were presented as mean and standard deviation; otherwise, median and interquartile range were reported. For comparative analyses, some variables were grouped into categories (e.g., age by decades, HD duration in months or years) to facilitate interpretation and comparability across subgroups.
Statistical Analysis
Descriptive statistics were used to summarise demographic, clinical, and transportation-related variables. Continuous variables were assessed for normality using the Kolmogorov-Smirnov test; normally distributed data were expressed as mean and standard deviation, and non-normally distributed data as median and interquartile range. Categorical variables were presented as absolute and relative frequencies.
To examine associations, the Student’s t-test was used for continuous variables with normal distribution, and the chi-square test or Fisher’s exact test for categorical variables, as appropriate. Proportions were compared using 95 percent confidence intervals, and non-overlapping intervals were considered statistically different. Potential confounders, such as age, sex, and presence of comorbidities, were considered in stratified analyses to assess their influence on the outcomes.
No formal subgroup or interaction analyses were pre-specified. Missing data were minimal, did not exceed 5 percent for any variable, and were handled by listwise deletion, without imputation. Given the cross-sectional design, there was no loss to follow-up. No sensitivity analyses were conducted. Given the descriptive and cross-sectional design, no multivariable adjustment was performed, instead potential confounders such as age, sex, and comorbidities were considered in stratified descriptive analyses. All statistical analyses were performed using JASP Team (2025) (Version 0.13.1), and p-values <0.05 were considered statistically significant.
Results
From the HD centres that were accessible during the COVID-19 restrictions in place from April 2021 to March 2022, 472 participants were invited to participate at 4 HD centres. The final sample comprised 439 participants (93% of the total number of people approached) who met eligibility criteria, agreed to participate, and completed the structured interview. Patients were excluded if they presented acute intercurrent clinical conditions on the day of the interview or residual neurological deficits from prior stroke that impaired communication.
Before undergoing renal replacement therapy through haemodialysis, 243 (55.4%) of the 439 participants in this study were driving. After starting haemodialysis, 157 (64.6%) of them continued to drive. Among those 157 participants who kept driving, 117 (74.5%) drove even on dialysis days, and 40 (25.5%) did not drive on dialysis days.
Table 1 presents the main socio-demographic characteristics of the study participants. More participants were male (56.0%), and the largest proportion was 60 year olds (42.8%) (Mean = 55 years; Median = 57 years; Mode = 45 years). Half the participants (53.1%) had up to 4 years of formal education. The majority of the participants (n=373, 85%) were not working and receiving government financial assistance due to CKD and its complications.
Table 2 presents details on participants’ main diseases by age. Each CKD patient on HD in this study had an average of 4 to 5 pre-existing conditions. Cardiovascular diseases were predominant (28.3%), with hypertension being the most prevalent in this subgroup (64.3%), followed by sleep disorders (16.4%), where insomnia accounted for 46.8 percent in this subgroup, followed by sleep apnoea/snoring (27.8%) and Restless Legs Syndrome/Willis-Ekbom Disease (21.3%). Endocrine diseases were the third most frequent (14.7%), with diabetes mellitus being predominant in this subgroup (51.5%).
Hypertension was reported as the aetiology of CKD in 41.3 percent of the patients, followed by diabetes (22.8%), while other causes appeared in proportions below 10 percent, and an indeterminate cause in 13.5 percent of the cases. As for the duration of treatment for haemodialysis patients, 26.0 percent mentioned being in treatment for 1 to 3 years, with the highest percentage among those aged 40 to 59 years (Table 3). Additionally, 19.1 percent reported being in therapy for over 10 years.
One patient (aged 19 years) did not report any complications, leaving 438 patients who reported a total of 478 complications (1.1 complications per person). The most frequent intradialytic complications were insomnia (33.5%) and hypotension (29.9%), followed by restless legs syndrome (15.3%), nausea/vomiting (9.6%), palpitations/arrhythmia (6.5%), and tremors (4.6%).
Table 4 presents details on driving behaviour. Before undergoing renal replacement therapy through HD, 243 (55.3%) of the 439 participants stated they were driving. After starting HD, 157 (64.6%) of participants continued to drive, and among these, 117 (74.5%) drove even on dialysis days, while 40 (25.5%) did not. Among the 157 who kept driving, most were male (86.0%), and 16 (10.2%) reported driving without a valid driver’s licence. Of those who drove on dialysis days, 31.6 percent were aged 60 years or older, 56.4 percent had completed high school or higher, and 20.5 percent were regularly employed.
Only 14 of the drivers reported being involved in traffic crashes in the past year, while 24 reported minor damages, such as bumps during parking manoeuvres or collisions with fixed objects. The most commonly used daily modes of transportation were cars (48.1%) and buses (34.6%). On dialysis days, 32.4 percent of patients used cars and 29.8 percent used buses, followed by 15.0 percent using taxi/Uber services and 11.9 percent institutional vans.
Discussion
This study describes the transportation habits and driving status of patients with end-stage CKD undergoing haemodialysis in São Paulo, Brazil. A CKD patient undergoing renal replacement therapy with HD needs to travel to the treatment clinic three times a week, which necessitates a restructuring of their routines. With a focus on the home-to-clinic-to-home journey, and we obtained a general profile of these patients with special attention to those driving before and after starting HD. Our main finding was that a significant proportion (two-thirds) of HD patients continued to drive. Among these drivers, approximately three quarters reported driving even on dialysis days, which exposes them to potential complications related to dialysis. Furthermore, almost all these drivers reported experiencing one or more sleep disorders, which may further impair driving ability and increase the risk of traffic crashes.
Before starting HD, half the participants (55.3%) were driving, which corresponds to the proportion of licensed drivers in the city of São Paulo (DETRAN, 2021). Driving while undergoing renal replacement therapy may also be common in other countries (Varela et al., 2015). An Australian study reported that 76.8 percent of patients were driving, of whom 76.3 percent were at risk of driving impairment post dialysis (e.g., dizziness, numbness, falling asleep while driving (Graver et al., 2021).
Similar to what is observed internationally, the proportion of women with end-stage CKD in our sample was lower than that of men, and the proportion of women who continue to drive is markedly lower than that of men. Although this study did not investigate why some patients stop driving after starting HD, it is possible to infer, as in other studies, that this may be related to CKD and HD. The impact on their abilities or clinical occurrences during or after sessions may make patients reconsider their ability to drive safely (Kepecs et al., 2018; Vats & Duffy, 2010). The significant difference observed between the proportions of men and women who continue driving after starting HD may be associated with differences in perception and caution in facing problems and men’s underutilisation of available help in the healthcare system (Addis & Mahalik, 2003).
The selected centres predominantly serve public patients, with extensive coverage by the Brazilian Unified Health System (Sistema Único de Saúde - SUS) (Nerbass et al., 2022; Thomé et al., 2019).
The clinical, demographic, and epidemiological characteristics of the participants in this study are comparable to those described in other studies (Godinho et al., 2006; Gupta et al., 2021; Nerbass et al., 2022). End-stage CKD was more prevalent in male patients, and the most common aetiologies were High Blood Pressure (HBP) and Diabetes Mellitus (DM) (Gupta et al., 2021; Nerbass et al., 2022). The incidence of end-stage CKD increases with age, and people over 75 years old have more than double the incidence compared to those aged 45 to 64 years (Gupta et al., 2021), making it an increasingly important variable in light of the aging global population (WHO, 2020).
Longevity also exposes individuals to other diseases beyond end-stage CKD, creating a set of variables (visual impairment, motor problems of neurological or orthopaedic origin, cognitive ability) that increase the risk of crashes while driving or may limit the person’s ability to perform the complex task of operating a motor vehicle (Lacson Jr & Brunelli, 2011; Vats & Duffy, 2010). Continuing to drive may be an even greater necessity for the Brazilian population, particularly due to the multiple diversities present in Brazil, such as long distances, lack of public transportation, and financial difficulties (Godinho et al., 2006; Graver et al., 2021). These factors demand greater autonomy from the patient to undergo treatment and may explain why nearly two-thirds of them continue to drive after starting HD.
With an estimated population of 46.6 million people, the State of São Paulo has 25.9 million licensed drivers (DETRAN, 2021; IBGE, 2010), which is more than half of the population (55.5%). Similarly, half of the participants in this study hold a National Driver’s License (CNH – Carteira Nacional de Habilitação), indicating that the current sample reflects the general population of the State of São Paulo regarding driving behaviour.
During HD, it is common for patients to experience clinical complications, especially in the first few months of treatment (Levy et al., 2009), as seen in nearly a quarter of the participants in this study, who had not yet completed 12 months of treatment. Some complications may have a potential impact on driving ability, considering their intensity, duration, or expected recovery time. Hypotension, nausea, vomiting, dizziness, weakness, leg numbness, fatigue, and palpitations are complications frequently experienced by patients undergoing treatment (Graver et al., 2021; Vats & Duffy, 2010). Our data on clinical complications were frequent and similar to those in other studies, supporting the interpretation that these patients may be at increased risk of crashes, although this should be viewed with caution given the descriptive design and self-reported nature of the data.
The criteria adopted regarding driving safety are quite strict in some studies, where a single clinical complication during or after dialysis is enough to classify the patient as a high-risk driver (Graver et al., 2021). Knowing that the initiation of renal replacement therapy presents more clinical complications for the patient, it is important to pay closer attention in the first few months of treatment to those patients who choose to be their own driver on the home-clinic-home trips.
Diseases in general, such as those prevalent in the population of this study, affect the ability to drive by compromising cognitive functions, and most of the diseases etiologically related to end-stage CKD are also associated with cognitive decline. This makes these patients particularly prone to impairments while driving (Kepecs et al., 2018; Shen et al., 2017). Adding to this unfavourable scenario are the clinical complications and sleep disorders, with complaints of obstructive sleep apnoea and snoring present in nearly one-third of the participants in this study. Therefore, these diseases and their corresponding functional impairments could have been present even before the end-stage CKD phase, contributing to a worsening in health status, cognition, and mortality (Arvanitakis et al., 2004; Drew et al., 2015; Graver et al., 2021; Knopman et al., 2001; Posner et al., 2002; Solomon et al., 2009).
Sleep disorders had a similar prevalence across all age groups, with insomnia, sleep-disordered breathing (snoring and obstructive sleep apnoea), and restless legs syndrome being particularly prominent. Sleep disorders frequently lead to drowsiness and impaired attention, which, when combined with other clinical problems, whether acute (during the HD session) or chronic (uraemia, subacute vascular lesions, progressive cognitive impairment, complications of the disease that led to CKD, among others), can dangerously impair the ability to perform such a complex task as driving (Shen et al., 2017). Therefore, addressing and providing medical guidance at the dialysis clinic is essential, aiming to recognise and offer recommendations to these patients, thereby contributing to preventing crashes in this population.
In the United States of America (US), the American Medical Association and the National Highway Traffic Safety Administration of the US Department of Transportation recommend that physicians take an active role in patient safety regarding driving. They recognise the main reasons that expose patients and others to the risk of automotive crashes. Additionally, they have prepared an educational manual for physicians, the ‘Physician’s Guide to Assessing and Counselling Older Drivers,’ which provides useful and objective information for managing these cases. Some tools, such as the ‘Am I a Safe Driver?’ questionnaire included in the mentioned Guide, can be useful for detecting problems, as patients’ awareness of their deficiencies may be improved (Graver et al., 2021). The presence of drowsiness while driving or fainting, for any reason, suggests a high risk for driving (Joseph, 2013; Vats & Duffy, 2010), especially in a population already facing increased mortality due to cognitive dysfunction (Drew et al., 2015).
There are still few studies involving populations undergoing renal replacement therapy and driving, and a broader research effort is needed to better control the impact of sleep disorders on driving, with emerging technologies such as kinematic driving sensors playing a potentially important role in the monitoring of drivers with chronic conditions (Mukherjee et al., 2024; Sharwood et al., 2011). Still, all studies report the negative impact of this treatment on the ability and sense of safety while driving (Lacson Jr & Brunelli, 2011; Vats & Duffy, 2010). The crash rate is high (Graver et al., 2021), emphasising the importance of safe transportation and the careful attention of the physician, so that the end-stage CKD patient is viewed holistically in their fragility and limitations, thereby ensuring treatment efficiency (Go et al., 2004; Graver et al., 2021; Lacson Jr & Brunelli, 2011; Vats & Duffy, 2010).
Our data highlight a reality that must be managed rationally, as HD patients, like any citizen, need to travel to carry out their activities, but they must do so under conditions that meet the demands of this practice, considering the risks posed to others and themselves. The majority of patients who were driving before HD continue to drive after starting treatment, including on dialysis days. Among the patients who no longer drive, approximately 4 out of 5 reported stopping due to HD. The main comorbidities associated with the HD patients in this study were hypertension, diabetes mellitus, and sleep disorders, notably insomnia, and Restless Legs Syndrome/Willis-Ekbom Disease (RLS/WED). Physicians involved in treating HD patients must be vigilant about acute peri dialysis events that may impair safe driving, as well as any signs of compromised driving ability in these patients. Regulatory authorities and traffic medicine specialists need to be aware of the realities faced by this population and work towards achieving a harmonious solution.
Study strengths and limitations
Among the advantages of our study, we include the evaluation of a relatively large number of patients treated within the public health system, allowing for greater external validity of our data. Additionally, the sample has a proportion of patients who were driving before HD that is similar to that of the population of the State of São Paulo, suggesting that we obtained a representative sample.
Taken together, our findings indicate that a considerable proportion of HD patients remain active drivers despite a high prevalence of comorbidities and treatment-related complications that may impair driving safety. However, these results should be interpreted with caution. The descriptive and cross-sectional design, the reliance on self-reported data, and the possibility of selection bias during the COVID-19 pandemic limit the strength of causal inferences. While our results are consistent with previous studies from other countries, further analytic and longitudinal research are needed to confirm these associations and to better understand the impact of clinical complications and sleep disorders on driving safety in this population.
The limitations of this study include the difficulty of validating crash data, which patients may underreport, and the fact that data collection occurred in the context of the COVID-19 pandemic, when access to dialysis centres was restricted. This situation may have introduced selection bias, possibly favouring the inclusion of more clinically stable patients. In addition, information was obtained by self-report, which is subject to recall and social desirability biases, especially for sensitive issues such as traffic violations or driving without a valid licence, potentially leading to underestimation of these events. The sample was not randomly selected, which may limit the generalisability of the findings, although its sociodemographic profile is consistent with the dialysis population of São Paulo. Another limitation is that only unadjusted estimates were reported; no multivariable models were applied, restricting control of potential confounders and leaving room for residual confounding. Finally, complications were recorded as reported by the patients, and multiple symptoms could be listed, which may have slightly overestimated their frequency. Taken together, these limitations suggest that while the direction of bias may underestimate the frequency of adverse driving events, the main conclusion remains consistent, that a significant proportion of HD patients continue to drive despite clinical conditions that may compromise road safety outcomes for themselves and other road users.
Conclusion
In summary, this study shows that a considerable proportion of patients with end-stage CKD continue to drive after initiating haemodialysis, including on dialysis days, despite frequent comorbidities and treatment-related complications that may compromise driving safety. These findings highlight the need for clinicians to routinely address transportation and driving habits as part of the clinical management of HD patients. Educational strategies for patients and families, as well as guidance from regulatory and traffic medicine authorities, are essential to promote safe mobility and reduce the risk of crashes. Public health policies should consider the reality of these patients, who must travel regularly to dialysis centres, ensuring that transportation is both accessible and safe.
Acknowledgements
The authors acknowledge Maria Helena Prado de Mello Jorge for her assistance.
AI tools
AI tools (ChatGPT-5.2) were used to improve English expression and clarity.
Author contributions
Conception: GFP Execution: AAC, AAC Drafted the article: AAC, AAC Analyses and interpretation of the reported study: GFP, AAC, KC Revised it critically for intellectual contents: GFP, AAC, KC
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Ethics approval
The study protocols were approved by the Research Ethics Committee of the Federal University of São Paulo (Approval: 4.359.193/2020) on 25 October 2020.
Data availability statement
All materials, data, and study protocols associated with this manuscript are available from the corresponding author upon reasonable request. The data supporting the findings of this study were collected and stored using the REDCap platform (Research Electronic Data Capture) hosted at Universidade Federal de São Paulo (UNIFESP) and can be accessed upon request at: https://redcap.unifesp.br.
Conflicts of interest
The authors declare that there are no conflicts of interest.