Clinical and economic burden among older adults with acromegaly in the United States

Medicare beneficiaries with acromegaly use significantly more healthcare services and have higher healthcare costs than beneficiaries without acromegaly. Skilled nursing facility care, HHA visits, and non-ED outpatient visits are around two-times more common among beneficiaries with acromegaly than among those without acromegaly. Total all-cause healthcare costs are more than double among Medicare beneficiaries with acromegaly compared with beneficiaries without acromegaly. Outpatient pharmacy costs are the largest contributor to the cost differences between those with and without acromegaly.

Comorbidity burden was higher among Medicare beneficiaries with acromegaly than what has been published in commercially insured acromegaly populations [20]. This result is not surprising as adults older than 65 years old typically have higher rates of comorbidities compared with adults less than 65 years old [30]. Additionally, we observed that older patients with acromegaly have higher rates of comorbidities relative to a matched reference cohort of older patients without acromegaly. Broder et al. found that acromegaly-related complications (i.e., colon neoplasms, musculoskeletal, cardiovascular disease, sleep apnea and hypopituitarism) and cardiovascular risk factors (i.e., diabetes, hypertension, hypertriglyceridemia) were associated with statistically significant increases in overall healthcare costs and odds of hospitalization and emergency department visits [31]. Addressing symptoms and disease progression through more effective treatment options could be an area of potential cost-savings.

Overall, the rates of HCRU and costs found in the current work were similar to published real-world studies. Published HCRU ranged from 4.0–34.6% of patients with inpatient hospitalization and 0.6–28.3% with ED visits [19–21]; however, these studies used commercial data and did not examine burden among Medicare beneficiaries. The only published article which analyzed all-cause and acromegaly-related utilization and costs reported that in a commercially insured population, 18.8% of patients had acromegaly-related inpatient hospitalization and 1.8% had acromegaly-related ED visits [20], comparable to what we found in the current study (9.8 and 3.5%, respectively).

There is some potential for statistically significant differences between patients with acromegaly and the disease-free references to be false positives due to our reporting of multiple hypothesis tests. However, most of the outcomes tested were found to be significant at p < 0.001; therefore, it is unlikely that the significance of those findings would change following the implementation of a multiple comparisons correction method. Outcomes for which lower p-values are reported should be interpreted with additional caution as a p-value adjustment for those outcomes is more likely to impact the study’s findings.

The proportion of patients with pituitary surgery in the current study (3.3%) was lower than in previous claims-based analyses (5.3–31.8%) [20,31,32]. Rates of surgery in claims-based analyses, particularly those conducted in the US setting, are lower than reported in medical registries (92.0–92.7%) or patient surveys (71–81%) because claims data rarely cover the lifetime medical history of a patient [18,33,34]. The current study analyzed Medicare claims data and reported the proportion of prevalent acromegaly cases among Medicare beneficiaries who had surgery during a 1-year observation period. Among previous claims-based analyses, differences in the observed proportion of patients who received pituitary surgery may be attributable to differences in the studies’ designs. First, previous studies included patients under the age of 65 and patients who are 65 years old or older only make up about 10% of those cohorts [20,31,32]. Older patients are likely to have different rates of surgery compared with younger patients due to patient preferences for surgery and more contraindications for surgery among older patients relative to younger patients [35,36]. Second, in the current study, a randomly selected claim with an acromegaly diagnosis was used as the index date over a 5-year identification period; therefore, the current study’s cohort is composed of patients in various states of the acromegaly disease course. One strength of the current study’s design is that it observed the pituitary surgery rate for Medicare beneficiaries diagnosed with acromegaly during 1 year across the full spectrum of the disease course. Conversely, in the studies authored by Broder et al. and Burton et al. the first observed acromegaly diagnosis was used as the index date and their cohorts are biased toward patients who are both younger (due to their patient cohorts which contains patients from 1 year old to 65+ years old) and earlier in the acromegaly disease course (due to their study protocols which selected the first observed diagnosis as the index date) [31,32]. Therefore, Broder et al. and Burton et al. were more likely to observe surgeries under their study designs relative to the current study design [31,32].

Medicare beneficiaries without surgery may also have had milder disease, which is not uncommon in older patients with acromegaly [12,17,37]. Lower rates of surgery in our cohort relative to younger cohorts are in line with conservative disease management practices related to the documented higher risk of mortality and poor post-surgery outcomes in an older population [38,39]. Ceccato et al. compared acromegaly surgical rates, using medical records from an endocrinology unit in Italy, among patients for whom the age of onset was less than 65 years old and those 65 years and older and found more than double the proportion of patients with surgery in the younger age group (86 vs 31%) [17]. As this published source using data other than claims also shows lower rates of surgery among older adults, it is unlikely that the current study’s lower rates were due to missing data. Additionally, because all of the patients in this study can have their healthcare reimbursed by Medicare, it is unlikely that beneficiaries would have paid for surgery out-of-pocket and any surgery reimbursed by Medicare was available in the data used in the current study. New medical therapies may be especially valuable for maintaining disease control among older patients with acromegaly for whom surgical intervention is less common.

Similar to the current study, a high rate of nontreatment was also seen in published sources, ranging from 55.2% and 49.1% of patients without treatment [20,32]. Beneficiaries who were not on medications during the observation period may have acromegaly that is in remission, had a surgery that successfully reduced IGF-1 and GH levels, or discontinued therapy due to cost, healthcare barriers or patient dissatisfaction with current treatment options [40]. For example, Medicare beneficiaries with mild disease may have opted out of current available treatment due to preferred route of administration (e.g., preferred an oral medication over an injectable medication) or avoidance of adverse events associated with available treatments [12,41]. Studies show that 58–81% of patients with acromegaly have disease recurrence following medication discontinuation [42,43]. Furthermore, other studies have linked treatment dissatisfaction with low treatment adherence [44,45]. Future studies focusing on treatment satisfaction among older patients with acromegaly and treatment patterns and the proportion of older patients who achieve biochemical control and symptom management with or without treatment could provide further insight into the high rates of nontreatment.

The current study’s total costs (all-cause and acromegaly-related) were comparable to those published in previous studies using older commercial data, which ranged from $24,900 to $48,341 [19,20,31,32,42,43]. Placzek et al. reported acromegaly-related total costs (medical + pharmacy) of $14,470 (inflation-adjusted to 2023 USD), slightly less than the $19,889 found in the current study. The main driver of the difference between that cost and the current study appears to be related to therapy costs, as Placzek et al. reported prescription (pharmacy) costs of $8158 (inflation-adjusted to 2023 USD) compared with the $15,487 we found; additionally, our therapy costs also included surgery and radiation [20].

This retrospective cohort study uses claims data to estimates of healthcare utilization and costs associated with acromegaly in US Medicare beneficiaries 65 years and older. Patient identification was based on diagnosis and procedure codes, rather than being clinically validated, so misdiagnosis of acromegaly was possible. Results from common acromegaly diagnostic testing, such as IGF-1 serum or pituitary MRI, are not available in claims data. Using a linked data source, such as linked electronic health records (EHR) and claims data, may offer additional and valuable clinical detail [44–50]. However, linked EHR and claims data can be limited in their usefulness due to variations in data quality and completeness based on differences in data structure and collection methods, terminology, and definitions across providers, sites and systems [44,45,47,48,51]. Additionally, while lab data can be found in linked EHR and claims data, it can still be limited in its availability and usefulness due to differences in the lab tests and codes used by different local laboratories and a lack of consistency in how and when results are recorded and interpreted [45,49,51]. Sample size, which was a key consideration in the current study as acromegaly is a rare disease, is another potential limiting factor in the usefulness of linked EHR and claims data [44,45,49]. As described earlier, Medicare RIFs are an ideal, large data source for capturing older adults with rare diseases such as acromegaly due to wide coverage in the US, and the methodology for patient identification used in the current study has been utilized in other published work [31,32]. While a 1-year observation/follow-up period is a commonly used time frame in real-world studies [17,19,20,43], restricting the analysis to a 1-year time frame limits the study’s ability to inform whether beneficiaries were ever on any medications. However, 1-year evaluations of HCRU and costs provide valuable insights for patients, physicians, and payers and this information is relevant for budgeting, staffing and resource allocation for each stakeholder. As this study relied on insurance claims for services provided, undiagnosed patients or patients not seeking acromegaly-related care would not have been captured. This could have resulted in underestimated utilization and costs. Additionally, we only measured direct healthcare costs and did not examine indirect healthcare costs that may add to the burden among older adults with acromegaly, such as loss of productivity, caregiver burden and reduced quality of life. Lastly, results may not be generalizable to different populations with acromegaly, such as those lacking healthcare coverage and patients under the age of 65 years. Results may also not be generalizable to beneficiaries with incident cases of acromegaly as the cohort included patients at varying stages of disease progression.

Medicare beneficiaries with acromegaly have substantial healthcare resource use and costs compared with controls without acromegaly. The higher rates of utilization, costs and comorbidities among beneficiaries with acromegaly compared with their counterparts without acromegaly indicates a need for additional effective therapeutic options for patients with acromegaly. Effective disease management, through biochemical control and symptom alleviation, may reduce the utilization and cost burden associated with acromegaly. Future studies examining the impact of treatment choices and patterns on utilization and costs could provide more insight into this unmet need.

TP Quock, SK Rattana, IE Paulson: Acquisition of data; design of the work; interpretation of data; revising the work for important intellectual content; final approval. E Chang, AK Das, A Speller, MH Tarbox, MS Broder: Design of the work; analysis of data; interpretation of data; drafting and revising the work for important intellectual content; final approval.

Funding for this study and publication-related fees was received from Crinetics Pharmaceuticals.

TP Quock, SK Rattana, IE Paulson are employees of and hold stock in Crinetics Pharmaceuticals, Inc. at the time of this study. At the time this research was conducted, E Chang, AK Das, A Speller, MH Tarbox and MS Broder were employees of Partnership for Health Analytic Research, LLC (now ADVI Health), which was paid by Crinetics to perform the research described in this manuscript. PHAR also discloses financial relationships with the following commercial entities outside of the submitted work: Abbvie, Akcea, Amgen, Astellas, AstraZeneca, Delfi Diagnostics, Dompe, Exact Sciences Corporation, Genentech, Gilead, GRAIL, Ionis, Janssen, Nobelpharma, Novartis, Pfizer, Recordati, Regeneron and Sanofi US Services. The authors have no other competing interests or relevant affiliations with any organization or entity with the subject matter or materials discussed in the manuscript apart from those disclosed.

No funded writing assistance was utilized in the production of this manuscript.

This study received approval for full waiver of Health Insurance Portability and Accountability Act (HIPAA) authorization from the Western Institutional Review Board.

This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 Unported License. To view a copy of this license, visit https://creativecommons.org/licenses/by-nc-nd/4.0/