Abstract
Berglund M, Gonzalez-Izquierdo A, Denaxas S, B. Cord Lethebe, Sajobi TT, Engbers JDT, Wiebe S, Josephson CB. Epilepsia. 2024, doi:10.1111/epi.18105. Online ahead of print. Objective: The incidence of late-onset epilepsy (LOE) is rising, and these patients may use an excess of healthcare resources. This study aimed to measure pre-/post-diagnostic healthcare use (HCU) for patients with LOE compared to controls. Methods: This was an observational open cohort study covering years 1998–2019 using UK population-based linked primary care (Clinical Practice Research Datalink [CPRD]) and hospital (HES) electronic health records. The participants included patients with incident LOE enrolled in CPRD and 1:10 age-, sex-, and general practice-matched controls. The exposure was incident LOE (diagnosed at age ≥65) using a 5-year washout. The main outcome was all HCU (primary care [PC], accident and emergency [A&E], admitted patient and outpatient care) using inverse proportional weighting to PC use and HCU by setting. An interrupted time-series analysis was used to examine pre-/post-diagnostic HCU between patients with LOE and controls over 4 years on either side of the diagnosis/matching date. An adjusted mixed-effects negative binomial regression was used for post-diagnosis HCU interactions. Results: Of 2,569,874 people ≥65 years of age, 1048 (4%) developed incident LOE. Mean weighted total HCU increased by 32 visits per patient-year (95% confidence interval [95% CI]: 13–50, p = .003) until LOE diagnosis, and then dropped by a mean of 60 visits per patient-year (95% CI: −81 to −40). There was an acute rise and fall over the 1–2 years immediately after pre-/post-diagnosis. Incident HCU remained higher for LOE compared to controls post-diagnosis (adjusted incidence rate ratio: 1.72; 95% CI: 1.65–1.70; p < .001), including A&E, outpatient, and admitted care. Significance: HCU demonstrates an acute on chronic rise over the 4 years before diagnosis of LOE. To what extent the partial reversal of the acute pre-diagnosis rise, and the mediators of the accelerated increase compared to controls are attributed to epilepsy, comorbid and bidirectional disease states, or a combination of both warrants further exploration.
Commentary
Patients with epilepsy face many dangers above and beyond the general population. The burden of seizures can be compounded by other physical and psychiatric comorbidities. This may translate into elevated rates of healthcare utilization for their myriads of seizure- and non-seizure conditions. Studying trends in healthcare utilization among people with epilepsy, compared to people without epilepsy, may yield valuable insight into successes and shortcomings in healthcare. Trends in healthcare utilization may only be magnified in late-onset epilepsy (LOE), which poses many special challenges related to increasing polypharmacy, multimorbidity and frailty, and etiologies such as stroke, dementia, and neoplasia.
Berglund et al. 1 recently described the trajectory of healthcare utilization before versus after a diagnosis of LOE. They conducted a retrospective cohort study containing outpatient, inpatient, and emergency room-linked electronic health records. From their ∼1 million starting population ≥65 years old with linkable data from across the United Kingdom, they identified about 1000 people who developed LOE. The investigators employed a weighted composite measure of healthcare utilization whereby each primary care encounter (including all appointments, telephone calls, messages, lab requisitions, etc.) counted as 1, outpatient visits (other than primary care) counted as 10, admission episodes counted as 43, and emergency room visits (not leading to admission) counted as 68. Their primary outcome consisted of a weighted sum of all such events.
Patients showed an interesting trend surrounding the date of the first epilepsy diagnosis code. The mean weighted number of events per 6 months was approximately 45 when follow-up began 4 years before diagnosis. Over the next couple of years, this event rate accelerated upwards, spiking at about 130 around the time of diagnosis. This declined back down but leveled out to a “new normal” around 80 over the next several years (higher than their pre-epilepsy baseline). This trend was recapitulated when considering the “unweighted” number of healthcare events, essentially any given type of healthcare event (primary care, outpatient, admission, or emergency), and different timeframes (3-, 6-, or 12-month event windows).
The investigators also compared these data to 10:1 (thus about 10,000) age-, sex-, and general practitioner (thus roughly sociodemographic region)-matched controls without epilepsy. Controls showed a completely different trajectory—just slow, linearly increasing event rates over the years. Also interestingly, cases and controls began at approximately the same event rate 4 years prior to epilepsy diagnosis (or matching date), or epilepsy cases slightly higher than controls. Curves clearly started to diverge thereafter, even starting years before epilepsy diagnosis during a pre-diagnosis smoldering stage.
The data illustrate an important aspect of the patient's journey, not previously described so clearly until now. This includes an approximately 3-fold increase in healthcare event rates surrounding the time of diagnosis, but even years prior to that with a rise before the more sudden uptick. This is in addition to the persistently elevated event rate thereafter compared to the patient's pre-diagnosis baseline. However, a higher healthcare event rate 8 years later (4 years pre- and 8 years post-diagnosis) is not surprising for patients that are already ≥65 years old, also seen in the control trajectory. Still, seeing these trends laid out supports the already-known heightened burden of chronic disease faced by our patients before, during, and after diagnosis.
The sharp decline in event rates after diagnosis is also at least as interesting as the sharp rise beforehand. Although not directly testable from these data, the authors theorize that one explanation could be that undiagnosed seizures drove increasing utilization before diagnosis, whereas after diagnosis the initiation of antiseizure medication could have mitigated seizure-related events thus driving utilization back down. This is possible, as delayed epilepsy diagnosis is a known pervasive problem, especially plausible in older adults for whom diagnosis may be more challenging. 2
Still, interpreting these data is not so straightforward. One issue is that while utilization is readily measurable, it is not necessarily inherently good or bad. Elevated utilization could seem bad, suggesting preventable emergency visits, overuse, or unmet medical needs. Unfortunately, though, medical necessity is not so easily measured, thus we can only assume that actual utilization serves as a proxy for need or quality. 3 Greater utilization could be a favorable reflection of a system with excellent healthcare access for those who need it, much like the absence of healthcare (not measurable from claims) could signal underutilization (the lack of seeking appropriate services) due to either patient or system factors. In other fields, some have attempted to address these research problems by restricting study outcomes to “preventable” 4 “non-ambulatory-sensitive” or “non-discretionary” readmissions (in which emergent care would almost certainly be necessary), thus it is also a bit difficult to compare epilepsy versus non-epilepsy groups without such a measure of a person's inherent healthcare-seeking attitudes or access, even if adjusting for a wide range of demographics and comorbidities.
Moreover, given this was a 65-year-old cohort studied over 8 years, death or hospice enrollment may reduce the “sicker” group's apparent healthcare utilization.
However, the most important limitation seems to be that these presented data do not distinguish between event diagnoses or contain prescription data to test the sequence of events. Did patients have a sharp rise in seizure-related visits (even if measured by nonspecific diagnoses before an epilepsy code), followed by initiation of antiseizure medications, followed by persistent antiseizure medication use, followed by the observed decline in seizure-related events? Or did the rise and fall have nothing to do with antiseizure medications or seizures? The sequence of events could inform the degree to which causation versus reverse causation or detection bias could be at play and would be critical to understanding before drawing clearer conclusions from data.
We all appreciate both the acute and chronic nature of epilepsy. These data remind us that the time surrounding an epilepsy diagnosis can be particularly perilous, consisting of a spike in healthcare contact, as the patient seeks the right set of healthcare professionals for their seizure care and possible etiologies or comorbidities. While these data do not define the forces driving these findings, they serve as a call to further explore the full causal pathway. Also, these data highlight that despite the downturn in event rates, patients did not come away from an epilepsy diagnosis (or even controls, over time) “good as new.” This could reflect the accumulation of medical conditions necessitating greater healthcare contact. For example, their adjusted model understandably found that brain tumors conferred the greatest increase in event rates compared to any other studied variables. Future research is needed to dissect the composition of post-diagnosis visits to understand the degree of preventability.
