Using Real World Data (RWD) to Identify Care Needs in an Adolescent and Young Adult (AYA) Lymphoma Population: A Match Cohort Study

Adolescent and young adult (AYA) patients with malignancies are a unique population within oncology with differences in survivorship outcomes as compared to adult and pediatric counterparts. The National Comprehensive Cancer Network (NCCN) identified eight critical domains in AYA care: mental health, cognitive function, physical exercise, fatigue, immunizations/infection prevention, pain, sexual function and sleep disorders. However, it can be difficult to establish a clinical benchmark of these domains at an institutional level without time-consuming chart review. With the advent of real-world data generated from the electronic medical record (EMR), feasible large scale data assessment can be used to identify care needs and direct future prospective studies and interventions.

Here we demonstrate the feasibility of a match cohort study using EMR comparing metrics in the above NCCN domains in AYA lymphoma patients within the University of Rochester Medical Center. We identified 172 AYA lymphoma patients diagnosed between 2013 and 2019 and compared them to 516 controls over that time period using 3:1 control matching, both at time of diagnosis and one year post-diagnosis. Controls were identified within the EMR with exclusion criteria of no cancer diagnoses and fewer than an average of ten medical visits per year over the study period to control for other chronic disorders.

Demographic characteristics for both groups were matched for age at diagnosis date, sex, and race. The resultant populations had an average age of 29 (SD 6), male predominance (58.72%), and Caucasian predominance (82.56%). The lymphoma group had a slightly higher rate of federally funded insurance as compared to controls (25% vs. 20%).

In mental health domains, lymphoma group had a higher rate of depression at baseline (19.77% vs. 1.16%; p <0.0001) but anxiety was at comparable rates (8.72% vs. 10.85% p <0.46) with similar rates of antidepressant medication prescriptions (27.33% vs. 25.97% p < 0.75). At one year post-diagnosis, anxiety and depression rates were near identical but rates of antidepressant prescriptions had increased in the lymphoma group (30.81% vs. 22.65% p < 0.03.)

In physical health domains, lymphoma group had similar rates of BMI in the overweight and obese range as compared to controls and less than five percent of the group had a sleep medicine, nutrition, or cognitive assessment at baseline or one year post-diagnosis. Vaccination rates, as measured by influenza and pneumovax adherence, were low both at diagnosis and at one year post- diagnosis in the lymphoma group (18.6% and 21.5%, respectively).

In the pain domain, the lymphoma group had similarly low rates of alcohol abuse, opiate abuse and chronic pain diagnoses as compared to controls (between 1 and 4% without statistically significant differences). Interestingly, pain scores recorded at visits on the 0 to 10 scale were noticeably lower for the lymphoma group both at diagnosis and one year post-diagnosis (4.65% vs 11.24% p=0.009; 5.23% vs 11.43% p=0.02) however prescription of any analgesic medication was higher in the lymphoma group (76.16% vs. 40.98% p<0.0001; 63.95% vs. 30.62 p<0.0001).

Lastly, in the reproductive and sexual health domain, the lymphoma group had high rates of consultation with reproductive endocrinology at diagnosis as compared to controls (31.40% vs. 1.74% p<0.0001) suggestive of early referral between initial suspicion of cancer and tissue diagnosis. The overall use of the services however were still at relatively low rates. Additionally, long-acting removable contraception rates, while higher in the lymphoma population (11.27% vs. 4.23 p<0.02) were still relatively low. Finally, rates of erectile dysfunction diagnoses were similar in lymphoma cases as compared to controls (0.99 vs. 0.66 p< 0.7128) without significant differences in tadalafil/sildenafil prescriptions at one year post-diagnosis (2.97% vs. 1.32% p<0.15)

These data support the feasibility of using the electronic medical record to track institutional level ancillary markers of care for prospective study and population-specific interventions. Future studies will be directed at collaborations to address these care gaps and a prospective survey of AYA patients to identify patient-centered interventions improving care for this critical population.