Reducing the Rare Disease Diagnostic Odyssey: Identifying Potential Pulmonary Arterial Hypertension Patients Using the Problem List of Diagnosed Patients
Author(s)
Kyeryoung Lee, PhD, Augustine Annan, PhD, Isita Polamarasetti, MS, Hunki Paek, PhD, Liang-Chin Huang, PhD, Xiaoyan Wang, PhD.
IMO Health, Rosemont, IL, USA.
IMO Health, Rosemont, IL, USA.
OBJECTIVES: The prolonged “diagnostic odyssey” in rare diseases often takes 2-5 years or longer, imposing heavy burdens to patients and systems. We investigated whether profiles of conditions documented before diagnosis in confirmed Pulmonary Arterial Hypertension (PAH) patients could inform models to identify undiagnosed individuals earlier in their care pathway.
METHODS: We analyzed 1,170 PAH and 1,501 non-PAH patients from the IMO Health Terminology Search-log datasets. Clinical conditions recorded prior to PAH diagnosis were examined, and logistic regression assessed individual and co-occurring condition associations with subsequent PAH.
RESULTS: Of 41,584 unique clinical conditions recorded, 26 were identified as occurring before PAH diagnosis among all PAH patients. Ten of these showed the strongest individual associations, including palpitations, respiratory failure, hyperlipidemia, and heart failure (all p<0.0001), as well as anemia (p=0.001), dyspnea (p=0.008), and obstructive sleep apnea (p=0.033). In co-occurrence analyses, the triad of “palpitations, hyperlipidemia, and heart failure” was significantly more frequent in PAH patients (OR 2.14; 95% CI 1.08-4.27, p=0.030). Combinations involving hypertension also showed higher odds: “respiratory failure and hypertension” (OR 4.30; 95% CI 2.01-9.14, p<0.0001), or “anemia and hypertension” (OR 1.79; 95% CI 1.28-2.51, p=0.001), “hypertension, hyperlipidemia and gastroesophageal reflux disease (GERD)” (OR 2.79, 95% CI 1.11-7.00, p=0.028), “hypertension, dyspnea and heart failure” (OR 2.28, 95% CI 1.02-5.09, p=0.045), “hypertension, obstructive sleep apnea and GERD” (OR 1.97, 95% CI 1.03-3,76, p=0.040). Interestingly, when hypertension was absent, the co-occurrence of “hyperlipidemia and GERD” was significantly less frequent among PAH patients (OR 0.34, 95% CI 0.15-0.73, p=0.006). The predictive model achieved a weighted F1-score of 0.62 and an overall accuracy of 0.65.
CONCLUSIONS: We explored the feasibility of using pre-PAH diagnostic condition lists to predict undiagnosed PAH patients. Distinct pre-diagnostic and co-occurring condition patterns in PAH patients can inform predictive modeling, potentially enabling earlier identification and reducing the diagnostic journey in rare disease.
METHODS: We analyzed 1,170 PAH and 1,501 non-PAH patients from the IMO Health Terminology Search-log datasets. Clinical conditions recorded prior to PAH diagnosis were examined, and logistic regression assessed individual and co-occurring condition associations with subsequent PAH.
RESULTS: Of 41,584 unique clinical conditions recorded, 26 were identified as occurring before PAH diagnosis among all PAH patients. Ten of these showed the strongest individual associations, including palpitations, respiratory failure, hyperlipidemia, and heart failure (all p<0.0001), as well as anemia (p=0.001), dyspnea (p=0.008), and obstructive sleep apnea (p=0.033). In co-occurrence analyses, the triad of “palpitations, hyperlipidemia, and heart failure” was significantly more frequent in PAH patients (OR 2.14; 95% CI 1.08-4.27, p=0.030). Combinations involving hypertension also showed higher odds: “respiratory failure and hypertension” (OR 4.30; 95% CI 2.01-9.14, p<0.0001), or “anemia and hypertension” (OR 1.79; 95% CI 1.28-2.51, p=0.001), “hypertension, hyperlipidemia and gastroesophageal reflux disease (GERD)” (OR 2.79, 95% CI 1.11-7.00, p=0.028), “hypertension, dyspnea and heart failure” (OR 2.28, 95% CI 1.02-5.09, p=0.045), “hypertension, obstructive sleep apnea and GERD” (OR 1.97, 95% CI 1.03-3,76, p=0.040). Interestingly, when hypertension was absent, the co-occurrence of “hyperlipidemia and GERD” was significantly less frequent among PAH patients (OR 0.34, 95% CI 0.15-0.73, p=0.006). The predictive model achieved a weighted F1-score of 0.62 and an overall accuracy of 0.65.
CONCLUSIONS: We explored the feasibility of using pre-PAH diagnostic condition lists to predict undiagnosed PAH patients. Distinct pre-diagnostic and co-occurring condition patterns in PAH patients can inform predictive modeling, potentially enabling earlier identification and reducing the diagnostic journey in rare disease.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
MSR179
Topic
Methodological & Statistical Research, Real World Data & Information Systems
Topic Subcategory
Artificial Intelligence, Machine Learning, Predictive Analytics
Disease
No Additional Disease & Conditions/Specialized Treatment Areas, Rare & Orphan Diseases