Here, we report a retrospective analysis of RWD to improve the understanding of PSP progression and management across healthcare settings and the disease journey. This study was unique in its tailored approach of partnering with PSP patient communities and individual advocates for participant recruitment. PSP-specific recruitment approaches, including hosting a live book reading and coauthoring blog posts with advocates for PSP, are detailed in the Methods section. Every step of the research process, from recruitment to account creation and research consent, was completed remotely through various modalities, including the AllStripes website, social media, internet advertising, email, and/or telephone. This approach reduced some of the logistical barriers to participating in research, as individuals with PSP could participate without traveling to research sites. This method, coupled with a direct partnership with PSP communities and advocates, enabled the recruitment of a larger cohort of PSP patients from a broader geographical area than many traditional site-based studies [1, 8, 9]. The short time for recruitment is another advantage of this approach compared to traditional site-based recruitment, which can take several years [1] and relies on patients attending appointments at participating research sites. Recruitment for in-person research studies is a particular challenge in studies of PSP due to the older age and potential mobility problems and difficulty traveling for PSP patients. Transportation is commonly cited as a barrier to participating in research for older individuals [22]. By consenting online to participate in medical records–based research, travel was removed as a barrier to participating in this study.

Medical records from multiple healthcare centers across the disease journey were collected, and data points were abstracted directly from these records, which allowed for broader data collection than many previous studies of PSP progression and management that relied on record collection from a single or limited number of care centers [4, 18,19,20,21]. Patients with rare diseases frequently have long and complex care journeys requiring medical appointments in multiple care settings, including community care providers, specialist centers, academic centers, and centers of excellence. Individuals may therefore have medical records located in a range of settings, and RWD studies focusing only on the medical records from larger medical centers omit a potentially significant portion of the individual’s care journey. The current study collected medical records from a mean of 2.9 healthcare centers per individual, with 23.6% of participants having documents in four or more healthcare centers, allowing for more comprehensive data capture of the participant’s medical history, management, and disease progression than would have been possible from a defined number of care centers.

Our findings are generally consistent with what has been previously reported in the literature. For this cohort, the median time from symptom onset to diagnosis was 3.7 years, which is consistent with other studies showing significant delays between symptom onset and PSP diagnosis of 3.5 to 4.9 years [4, 15, 19]. In agreement with previous reports of disease onset to symptom development [18], the earliest symptoms experienced by our cohort were falls and speech problems, where first fall was the most common symptom to occur at disease onset. The most common symptoms experienced prediagnosis and at any time during the disease course were very similar to other studies of PSP symptoms [4, 18, 19, 23], with falls, gait problems, and mobility problems as the most common symptoms prediagnosis and overall. Similar to other studies, speech problems and dysarthria were experienced by 79.2% of participants. However, 47.2% of all participants experienced these symptoms prior to diagnosis, in contrast with other studies, which report a smaller proportion of patients experiencing speech problems prior to diagnosis [19, 23]. This difference may arise from differences in milestone definition and/or date of symptomatic onset; for example, O’Sullivan et al. (2008) defined these milestones as “unintelligible speech” and “severe dysphagia” [19], which would likely have a later onset than an initial speech problem or dysphagia.

In addition to a high burden of symptoms associated with PSP, participants in our cohort also had documentation of many comorbidities, with hypertension and depression occurring in 51.4 and 44.4% of the cohort, respectively. While hypertension is a common comorbidity in the general population [24], a recent study demonstrated a modest yet significant association between hypertension and PSP and showed that the presence of hypertension at least 10 years prior to the onset of PSP symptoms was approximately 1.5 times more common in PSP cases than in matched controls [25]. Depression is also well recognized to be associated with PSP [26, 27]. Depression is more common in PSP patients than in the general population of adults 65 and over [28], affecting approximately 60% of PSP patients, as shown in a recent systematic review [27]. Hopelessness and withdrawal, specific subscales of depression, have also been shown to be higher in individuals with PSP than in those with other neurodegenerative diseases [29], and depression has been identified as an independent determinant of reduced quality of life in individuals with PSP [30]. The effects of a high comorbidity burden on increasing HRU are well documented for the general population [31, 32]; however, the specific impacts of multimorbidity on the HRU of patients with PSP have not yet been sufficiently investigated.

Very little research into the HRU of individuals with PSP has been published to date. A recent study analyzed claims data to quantify healthcare utilization among individuals with PSP and demonstrated significantly increased yearly healthcare spending for individuals with PSP compared to controls [33]. Our study adds to these findings by characterizing changes in HRU after individuals receive a diagnosis, including an increase in the utilization of assistive devices and supportive care that would be challenging to assess in claims or EHR databases. The majority of participants (86.1%) in our cohort were prescribed levodopa or carbidopa at any time. Given that up to half of individuals with PSP are initially misdiagnosed with Parkinson’s disease [15] due to the shared early clinical features of the two conditions, levodopa prescription before PSP diagnosis could be due to a misdiagnosis of Parkinson’s disease or part of a levodopa test to rule out Parkinson’s disease [34]. Approximately 30% of participants in this study still received levodopa after being diagnosed with PSP, despite one of the characteristics of PSP being poor levodopa responsiveness [35, 36]. It has been documented that some patients still derive modest benefit from levodopa or other dopamine agonists, particularly those with the PSP Parkinson phenotype [36,37,38]. Thus, some patients may continue on the drug postdiagnosis due to the potential for even a slight improvement, especially in the absence of a disease-modifying therapy. The PSP phenotype was not well documented in the records for most participants in the current study, suggesting that better education is required to make healthcare providers more aware of PSP phenotypes; however, it is likely that the majority of participants had the PSP-RS phenotype based on their reported symptoms and the fact that PSP-RS is the most common PSP phenotype [39].

PSP is well documented to result in functional disability. In a large cross-sectional study using physician-reported questionnaire data on over 800 patients with PSP across multiple countries, 60% of patients were using a nonwheelchair walking aid, and 23% were using a wheelchair at the time of analysis (median 31 months from symptom onset) [40]. In the current cohort, 90.3% of participants had documentation of using at least one assistive device during the course of their disease, with almost half requiring a wheelchair and nearly half using an assistive device prior to the month of PSP diagnosis. This degree of functional disability has a significant impact on the disease progression and prognosis for an individual with PSP. A previous study investigated the relationship between five clinical disability milestones, including the inability to walk unassisted, and reported a significant association between earlier occurrence of disability and shorter survival time, even after adjusting for sex and age at onset [20].

Due to the many motor, cognitive, neuropsychiatric, neurocognitive, and behavioral symptoms of PSP [41], it is unsurprising that many individuals with PSP require assistance from a caregiver. Most participants in this cohort required some form of supportive care (86.1%), most commonly assistance from an unpaid caregiver while living at home (66.7%). A recent study reported a similar proportion of PSP patients with a caregiver (73%), but only 24% of the cohort used a professional caregiver [39]. The majority of PSP patient caregivers in other studies were family members [41, 42], which is also consistent with our findings. This reality underscores the economic and psychosocial burden experienced by the families of those living with PSP.

While most participants in our cohort underwent at least one surgery or procedure (84.7%), two of the most common surgeries and procedures, cataract surgery and colonoscopy, are likely to be more age-related than PSP-specific [43, 44]. Alternatively, it is possible that cataract surgery may have resulted from a misdiagnosis of cataracts in PSP patients who had blurred or double vision related to their PSP, not cataracts. The second most common procedure documented was a swallow study, which is considered the gold standard diagnostic tool for the evaluation of dysphagia and is consistent with the high proportion of participants with dysphagia in this cohort. Dysphagia is a common symptom of PSP and increases the risk of aspiration, pneumonia, and death [45, 46]. As such, regular swallow studies are recommended to assess the risk of dysphagia and evaluate the patient’s need for diet modifications or feeding tube requirements [47].

RWD from medical records is characterized by several inherent limitations that may have impacted the results of this study. While every effort was made to collect the complete medical record of every participant in the study, some documents may have been lost, destroyed, or were otherwise irretrievable from participants’ healthcare centers. Data quality depended primarily on the accurate reporting of information on clinical symptoms by neurologists and movement disorder specialists, which may be prone to clinician error, differences in documentation processes, or incomplete reporting. The postdiagnosis follow-up time was not equal and varied across all participants from 0 months to over 5 years, which may have impacted our findings since individuals were likely at different stages of their disease progression during the study period. In addition, this study may be characterized by potential selection bias since recruitment of participants relied heavily on partnership with PSP patient communities and individual advocates. As such, recruitment was biased toward English-speaking individuals who were connected with the wider PSP patient community with access to a computer and the internet. Finally, only individuals who had received a diagnosis of PSP and had not participated in an interventional trial for PSP were eligible for inclusion in the study. Therefore, undiagnosed and misdiagnosed individuals and those who previously participated in an interventional trial were not captured in this analysis.