We conducted a web-based survey on Qualtrics [23] to understand BLV people’s current practices and lived experiences of tracking and using their PHD. We ensured that our survey was accessible to screen readers (eg, JAWS, NVDA, and VoiceOver), the main assistive technology tools for BLV people that narrate digital content as synthesized speech.

We started the survey with questions about basic background information, such as the vision history and status, list of devices they owned, self-reported health conditions, and the level of interest in their own health. Before asking questions about PHD, we defined the term “Personal Health Data” as health-related data that people can track about themselves outside the clinic [7] and provided examples, such as exercise (eg, step count and miles run), sleep, diet, heart rate, weight, and blood pressure data. As we wanted to ask different sets of questions based on people’s experience with PHD tracking, we asked a question to categorize respondents into 1 of the 3 groups: current tracker (CT), past tracker (PT), and no experience (NE).

Our main survey asked about four aspects of PHD tracking experience: (1) types of data, (2) tracking methods and tools, (3) reasons for tracking (or not), and (4) barriers to tracking. For the data types and tracking methods and tools, we asked respondents to select items from a comprehensive list and let them also specify any missing items using the “Others” option. We used open-ended questions to elicit the reasons for tracking (or not) and barriers to tracking. Textbox 1 presents the key questions for each aspect; we have slightly rephrased some of the questions for different groups based on the respondents’ answers to prior questions. A full list of questions can be found in Multimedia Appendix 1.

We recruited people who were interested in PHD and met the following inclusion criteria: people who were (1) blind or LV (ie, those with mild to severe distance vision impairment); (2) aged at least 18 years; and (3) able to participate in the survey via a computer, mobile device, or phone. We provided an option to take the survey over the phone, in which case a researcher read out the questions and recorded responses. We recruited respondents via the National Federation of the Blind’s mailing list, researchers’ personal network, and web-based BLV communities on Facebook. The survey was available on the web during April and May 2022. As it was created in English, the survey was accessible to those who spoke English.

We initially had a total of 203 (198 via the web and 5 via phone) survey responses. Of the 184 completed responses, we excluded 2 (1.1%) that were submitted from the same IP address. On further inspection of the quality of the responses, we noticed a suspicious data entry pattern. A total of 26 responses repeated the same answer (“No,” “n,” or “I don’t really want to say”) for several open-ended optional questions, with most having the same set of answers across several response IDs. As it is highly unlikely that different individuals submit the same answers for multiple open-ended questions that are optional, we excluded these from our data set, resulting in a total of 156 responses.

We used open coding to analyze qualitative data from open-ended survey responses. Through several rounds of iterative refinements, we generated 71 codes that captured representative ideas in the data. We then used affinity diagramming to categorize similar codes and drew higher-level themes over several refinements. For quantitative data, such as demographics, PHD type, and data capturing methods, we characterized the data using descriptive statistics.

Our survey study was reviewed and approved by the University of Maryland, College Park’s Institutional Review Board under #1854542. We maintained the confidentiality of our respondents by using a security-accredited survey provider, deidentifying survey results during our analysis phase and in our reporting, and solely using password-protected servers and workstations for data storage.

To motivate participation, we used raffles, offering a US $50 Amazon gift card to 2 respondents and a US $20 Amazon gift card to 10 respondents. Before individuals took our survey, they were informed of our study’s procedures, potential risks and benefits, and their individual rights as participants and were given the opportunity to opt-out.

Table 1 summarizes the demographic characteristics of 156 respondents. Nearly an equal number of respondents who self-identified as totally blind (TB; 79/156, 50.6%) or LV (77/156, 49.4%) participated in our survey. Approximately two-thirds of them were CT (102/156, 65.4%), 19.2% (30/156) were PT, and 15.4% (24/156) had NE tracking PHD. More than half of the respondents were visually impaired since birth (83/156, 53.2%), and approximately one-third became visually impaired after birth and had been visually impaired for at least 10 years (51/156, 32.7%). Most respondents (151/156, 96.8%) rated their health condition as fair or better. Respondents varied in age, with 22.4% (35/156) aged 35 to 44 years and 21.2% (33/156) aged 55 to 64 years; 68.6% (107/156) were female, 29.5% (46/156) were male, and 1.9% (3/156) were nonbinary. More than two-thirds of the respondents were White (107/156, 68.6%), 10.9% (17/156) were Asian, and 5.8% (9/156) were Black or African American. In terms of education, 59.6% (93/156) of the respondents had a bachelor’s degree or higher.

Exercise (115/132, 87.1%) and weight (101/132, 76.5%) were the 2 most popular items the CT and PT respondents reported tracking (Figure 1). Other highly ranked items were sleep (63/132, 47.7%), heart rate (61/132, 46.2%), food (60/132, 45.5%), water (58/132, 43.9%), and blood pressure (55/132, 41.7%). On average, CT respondents tracked 5.8 (SD 2.9; range 1-14) items, and PT respondents tracked 4.7 (SD 2.6; range 1-11) items. In responding to the question “What other PHD do you want to track?,” the CT and PT respondents commonly mentioned stress (36/132, 27.3%), cholesterol (31/132, 23.5%), blood pressure (30/132, 22.7%), and mood (29/132, 22%). We noted that 27.3% (36/132) of respondents selected “None (I don’t have anything else I want to track),” which is tied with “stress” for the most selected. Refer to Figure 2 for further details.

For 24 NE respondents, we asked what type of PHD they wanted to track. The top 5 items were weight (17/24, 71%), stress (16/24, 67%), exercise (15/24, 63%), food intake (15/24, 63%), and water intake (15/24, 63%; Figure 3). NE respondents wanted to track 7.3 (SD 4.2) items on average, demonstrating high interest in tracking PHD.

Figure 4 shows the tracking methods for the CT and PT respondents. A majority of participants used smartphone apps (90/132, 68.2%), with the most frequently mentioned apps being Apple Health (n=45), Fitbit (n=15), and MyFitnessPal (n=15). However, few respondents also mentioned Weight Watchers (n=4); Google Fit (n=3); and Strava (n=3), an outdoor cycling app. Approximately half of the CT and PT respondents reported using wearable devices (68/132, 51.5%); Apple Watch was the most popular device (n=41), followed by Fitbit (n=22). Among the CT who used some sort of technology for tracking (n=87), 59 (68%; 34 very satisfied and 25 somewhat satisfied) respondents reported that they were satisfied with the current technology (Figure 5).

Our survey with BLV individuals shows that there is not only substantial interest but also important needs for tracking PHD by the respondents (Textbox 2). Unless otherwise noted, the total sample size for this section was 156 respondents.

A total of 33 respondents perceived tracking as an important tool for preventative health. Of this group, 23 described tracking as a means of maintaining their current state, whether sustaining a healthy status or preventing deterioration into a health problem. Awareness of the medical issues in their family’s history and aging were also listed as reasons why tracking was needed.

A means to improve oneself was another desired benefit for trackers. For some, improving one’s health was noted as a step to external goals, such as enabling new activities like aerial arts. General health improvement was a tracking reason for 24 respondents, with weight loss being the most mentioned. Some shared additional benefits of tracking, which helped them toward better health, such as being coached with “information to figure out how [to] effectively manage my health.” Tracking systems also provided welcomed encouragement that helped respondents remain on their health journey. For example, 1 respondent whose data led to a sustained effort said the following (as a respondent’s anonymized ID, we used the last 4 letters from the identifier the survey system automatically creates. We used “CT” to denote participants who are “current trackers” and “TB” to denote those who are “totally blind”):

A total of 50 respondents wished to use tracking as a means to learn about themselves, whether simply because of curiosity, or to view what progress was made in their activity. Of this group, the majority (34/50, 68%) found tracking to be an important diagnostic tool for their health, keeping track of various measurements and being made aware of any outlying incidents. Several respondents also valued the averages provided in the tracking systems for measurement comparison.

A group of 22 respondents expanded their analysis and wished to use tracking data for outcome analysis to examine how different actions affected their health. This ranged from inspecting the correlations of activity toward health to the effectiveness of their medications and treatments. One respondent shared how tracking was used to validate the proper use of medical equipment:

A total of 13 respondents mentioned the additional use of tracking systems as a historical record keeper as a reason for tracking. Of the 13 respondents, 2 saw this historical data as a means of sharing information with their doctors.

Different aspects of using tracking data with others were mentioned, such as social activities, including step count challenges. A total of 3 respondents tracked PHD to benefit others, striving to remain healthy to continue caring for family or clients in their lives, or merely to lessen the need for others’ aid.

Some tracking systems aid BLV trackers in adapting to their disability. A majority of respondents (67/132, 50.8%) relied on a screen reader (either on a smartphone or computer) to access the tracked data (Table 2). They augmented inaccessible devices with their smartphones, which provided an accessible screen reader for transferred data, and others relied on the added haptic feedback to alert them to various reminders and alerts. When buttons were mislabeled and screen readers were not helpful, BLV respondents learned by trial and error to memorize the functionalities of the mislabeled buttons.

Tracking systems also provided other types of organizational help, such as one user who appreciated how they were able to “funnel as much as possible into my Apple Health app as the primary hub,” and another who liked the ability to favorite specific health statistics to add to their feed.

Self-tracking has been extensively studied in the health informatics community [24-29], but the experience of BLV people was rarely represented. We contributed to an understanding of BLV people’s personal health tracking motivation, practices, needs, and challenges. Our survey showed that BLV respondents have strong desires and needs to track their PHD, and many of them are already tracking their data despite many hurdles. Popular tracking items—exercise, weight, sleep, and food—and the reasons for tracking reported in this study were similar to those of sighted people [3]. However, BLV people experienced diverse accessibility challenges in tracking and accessing data, failing to effectively benefit from self-tracking technologies.

Challenges persist throughout the different phases of self-tracking, starting from identifying accessible tools to difficulty in tracking and reviewing data. The main barriers our respondents experienced included suboptimal tracking experiences and insufficient benefits against the extended burden for BLV people. Some of these accessibility challenges were unique to BLV users, making already burdensome self-tracking even more challenging. For example, a common challenge faced by sighted users was to identify a personal tracking app that would help them meet their goals. To address this problem, sighted users would try out multiple apps before they settle on a tool of their choice—a process called “trialing” [30]. Such an approach is unlikely to be viable for BLV people because it requires considerable efforts to check whether a tool is accessible before they can assess the tool’s efficacy.

A recurring theme we identified in the data was whether a technology is accessible is not all black and white, but it exists on a spectrum (refer to the themes under “Knowing It Could Be Better” in Textbox 3). Even if BLV users spot an accessible technology, it is unclear to what extent the technology is accessible: tracking might be possible, but there may be an information delay in accessing the tracked data; data may be accessible only from an app but not from a wearable device, or the feedback (eg, a chart) may not be accessible at all. Contrary to our expectations and to the reported barriers from the open-ended survey responses, most current BLV trackers (59/87, 68%) were satisfied with the personal health technologies that they are using. Of the 34 highly satisfied respondents, 10 (30%) answered the questions specifically for their current setup, that is, they did not consider systems with barriers that had already been pruned from their setup. However, for others, perhaps the contrasting result indicates that despite many accessibility challenges BLV users experience, some people could have successfully used work-arounds to meet their tracking goals or might have been satisfied with subpar experience because they had low expectations to begin with. Further research is warranted to examine the gap between the high level of satisfaction and the accessibility challenges reported and to what extent the currently supported self-tracking features are being used and what key accessibility challenges have high impact on user satisfaction had they been addressed.

Similarly, the relationship between barriers to tracking and people’s current tracking status may prove to be another avenue of research worth investigating further. Although a large majority of NE respondents (22/24, 92%) expressed interest in collecting PHD, additional studies are needed to find more details on the barriers that limit these individuals from beginning to track and possible solutions to overcome them. Although our survey found similar barriers reported by PT and CT members, future work is needed to fully investigate why some persevere when confronted with barriers and others discontinue. It is conceivable that a decreased level of importance assigned to one’s health or circumstantial disadvantages that intensify some barriers are the reasons that contribute to discontinuation, but this inquiry would benefit from studies that allow for more depth beyond this initial survey.

Our participants appeared to be technologically savvy, as there were 94.2% (147/156) of smartphone users compared with the US average of 85%, and 96.8% (151/156) laptop or desktop device owners compared with the US average of 77% [31]. Moreover, 44.2% (69/156) of our survey respondents reported that they currently own wearable devices, which is much higher than the US average of 21% [2]. Although not generalizable to all BLV people, this shows our respondents’ high interest in personal health tracking. The challenges we reported regarding PHD tracking still hold because the lived experience of interacting with personal health technologies can come from those who had or have been using the technologies. Furthermore, fewer tech-savvy groups will likely experience the identified accessibility challenges if they were to engage in self-tracking.

However, our findings also indicate that a large variability exists among BLV individuals. Further studies are warranted to understand how general BLV individuals engage in personal health tracking while not conflating the 2 groups—TB and LV people—as their vision capabilities (eg, whether one has residual vision vs not) and respective solutions (eg, GUI based and screen reader based) could be different [32]. Furthermore, we need to devise ways to help BLV individuals with less technical proficiency onboard with self-tracking, such as identifying appropriate tools that meet their individual tracking goals.

We collected data through a web-based and phone survey to recruit more respondents more easily. To ensure the data quality, we carefully inspected the quality of responses and excluded 28 suspicious responses—2 with duplicate IP addresses and 26 with a suspicious pattern. However, as we did not observe or interact with the respondents, the inherent concern about the data quality of web-based surveys remains.

On the basis of our survey of 156 BLV users’ PHD tracking experiences, we reported quantitative and qualitative findings that contribute to an in-depth understanding of their motivations, tracking practices, challenges, and work-arounds. These findings shed light on design opportunities and areas to focus on making PHD technologies more accessible to BLV individuals. We strongly encourage technology companies and health researchers to invest their efforts in making PHD technologies accessible to BLV people, as a step toward ensuring equitable access to health data for all.