Two thirds of patients may not need routine 12-month specialist review in an epilepsy clinic: A cross-sectional study of clinic appointments

Objectives: Timely access to specialist outpatient clinics can be difficult to achieve as outpatient services are often oversubscribed leading to unacceptable wait times. New patients, or those with emergent issues may wait for appointments whilst existing patients are booked in for routine reviews “ just in case ” there is a problem, using considerable clinic resources. We investigated routine 12-month review appointments to assess whether these appointments changed patient management. Methods: The medical records of 100 randomly selected adult patients attending annual review appointments over 12 months at a publicly-funded specialist outpatient epilepsy clinic in Melbourne, Australia were audited. Demographic and clinical data as well as information about the content of each appointment were analysed to determine whether the appointment resulted in changes to epilepsy management (eg medication change), administrative actions (eg drivers license approval) or the provision of information or education. Logistic regression was performed to assess what clinical factors were associated with changes in patient care arising from the 12-month review appointment. Results: Almost half (47%) of appointments resulted in no change to patient care and 37% had only administrative outcomes, such as the completion of a regulatory driving report. Only 16% of appointments resulted in a change in medical management. The only factor that independently predicted a change in medical management was the occurrence of a seizure in the previous year. The only factor independently associated with not having any change in medical management or administrative action was having an unknown seizure type. Conclusions/ significance: Only a small number of patients experience a change in medical management when attending a 12-month epilepsy clinic appointment, with a need for management change associated with the presence of ongoing seizure. Outpatient services should limit the use of routine annual follow up to those patients most likely to need intervention or support, creating “ just in time ” capacity for timely access to review as issues arise.


Introduction
Timely access to specialist epilepsy care for diagnosis, commencement of medication, and evaluation for epilepsy surgery is associated with better patient outcomes for patients with epilepsy [1].Public hospital outpatient services, frequently used to deliver epilepsy care, are commonly over-subscribed with long waiting lists [2,3].This issue is further exacerbated in epilepsy clinics given epilepsy is a chronic disease and diagnosis may often be made at a young age and persist throughout life.The incidence of a first seizure is as high as 10 % throughout the lifetime, with at least a third of these patients going on to develop epilepsy [4].Insurance claim data indicates there may be significant delays to commencing treatment for patients with epilepsy [5], providing further support for the hypothesis that services are unlikely to be meeting the demand.Given the high volume of new referrals to epilepsy services, novel strategies are required to manage clinical demand and facilitate safe discharge from clinic.Hence, measures to optimise use of existing resources are required to enable appropriate access, both to new patients and patients with ongoing needs requiring specialist input.
Capacity constraints across the neurological workforce have been identified in a number of developed nations, including Australia [6][7][8].Simpson-Yap et al. [7], predicted an ongoing increase in the disparity between neurology specialist supply and demand but predicted that an increase in workforce alone is insufficient to meet this demand.It is estimated that there were 620 adult neurologists in Australia in 2000 [7], however there are no reliable estimates of those recognised as epilepsy specialists.In addition, although the cost burden of epilepsy in Australia is significant [9], there is no reliable information about how to optimally manage this demand.
It has been shown that after commencing a first medication for a new diagnosis of epilepsy, up to 50 % of patients will achieve seizure freedom on first-line, anti-epileptic therapy.However, approximately 35 % of patients will continue to have seizures despite successive trials of medical therapy [10,11].Patients who respond early to medication and remain stable may not require routine specialist appointments, and ongoing needs (for example, regular regulatory approval for driving) may be addressed by primary care clinicians [12].
Despite these known differences in the disease trajectory for patients with epilepsy, it is common for outpatient clinic models to routinely provide patients with annual reviews with a specialist neurologist as standard practice.These "just in case" appointments enable a patient to consult with a specialist in the event this should be required.However, the need for specialist input is unpredictable, and a fixed, regular appointment with a neurologist does not assist the patient at the time a new issue arises and needs to be addressed more urgently.An alternative is to reduce the use of prescheduled, routine follow up appointments, but this approach carries risks of missing important opportunities to discuss epilepsy related health issues and may leave the patient disconnected from the clinic [13].It is possible that these consultations identify emerging issues that have important implications for care, or play a vital role in supporting patients to manage their disease.These concerns, and potential risks, need to be weighed up against potential benefits in redesigning models of care with the intention of maximising the reach of limited resources.
We aimed to investigate the value of "just in case" appointments with a neurologist, scheduled to maintain regular contact with the patient within the clinic, by auditing routine, annual appointments in an outpatient specialist epilepsy clinic.We describe the patients who attended an annual review and determine the clinical utility of these consultations by analysing the actions taken by specialists during these appointments.

Study design and Setting
We conducted a cross-sectional retrospective audit using electronic medical records for all patients, including those with a non-epilepsy diagnosis (eg.PNES), attending annual specialist epilepsy review with a neurologist, expert in epilepsy management, at a publicly-funded outpatient epilepsy clinic in Melbourne, Australia.Eligible patients were randomly selected from a cohort of all patients attending clinic for a follow up appointment during 2019 and 2020, who had not had any contact with the clinic in the previous 11 months.Patients were excluded from analysis if they failed to attend their appointment for any reason or if they had attended a review appointment less than 12 months prior to the appointment.For patients who attended two eligible appointments during the study period, only one appointment was analysed.Institutional approval for the audit was provided (QA 22-008).The study is reported consistent with STROBE (Strengthening of reporting of observational studies in epidemiology) checklist for crosssectional studies [14].

Participants
A report of all patients who attended the epilepsy clinic for review appointments with a neurologist during the trial period was obtained from the health service clinic database and filtered to include only those for whom a minimum of 11 months had elapsed since their previous appointment.These patients were then sorted into random order using the random number generation command within Excel, and files audited from the top of the list until the required sample size was reached.

Variables and data source
Data were collected from the hospital's electronic medical records.Three categories of data were collected: demographic information (age, sex); clinical information relating to the patient's medical history and clinic attendance (time since first contact with clinic, number of seizures in past year, epilepsy classification according to International League Against Epilepsy criteria [15], medications, number of anti-epileptic medications); and the outcome of the appointment.Data were collected by one researcher and checked by a second researcher independently.
Outcomes of appointment were classified into one of the following four hierarchical categories: Change in medical management − any change to medical management including change in anti-seizure medication type or dose, referral for procedure, identification of a new clinical issue (such as pregnancy or identification of a new co-morbid condition), or other actions to address side effects or comorbidity.
Administrative Actioncommunication with regulatory bodies (for example, related to driving or employment) or provision of supporting documentation relating to the patient's medical condition, without any change to epilepsy management.
Information only -Documentation in the medical history that recorded an exchange of information, such as education about selfmanagement or community support services but no medical or administrative action.
No change -No change to epilepsy management, administrative actions or record of information exchange documented in the medical history associated with the appointment.
Where there was more than one outcome across multiple categories (such as a medication change and driver's license review), it was classified into the category highest on the list (change in medical management).

Data analysis
Sample Size Calculation: To assess change in medical management, using a conservative expectation of 50 % of appointments resulting in a documented change, 97 patients would be sufficient to estimate the proportion of patients receiving a change to medical management as a result of the consultation within 10 % accuracy with 95 % confidence [16].A sample size of 100 was also expected to be sufficient for the proposed regression analysis, using the rule of thumb that there should be a minimum of 10 cases for each variable to be included [17].
Analysis: Patient demographic and outcome data were analysed descriptively.We then performed bivariate analysis of factors to explore associations with changes in medical management alone, or with either a medical change or administrative action.Factors that approached statistical significance (p < 0.25) in the univariate analysis were then entered into a multi-variate analysis using logistic regression.

Participant characteristics
A total of 100 annual specialist epilepsy review appointments with a neurologist were randomly selected from a possible 177 appointments.The average time between the appointment and the previous appointment was 351 days (SD 17 days) Baseline characteristics of the sample are described in Table 1, which shows a predominance of males (60 %).The average age of the patients overall was 47.2 years (age range 19 -77 years).Males were an older cohort (males: 49.6, females: 43.5 years).The majority of patients (68 %) had been attending the clinic for over two years.Most patients (88 %) were seizure-free in the preceding year, with 4 % having had one seizure, 3 % having had two and 5 % experiencing more than two seizures, and 75 % were taking a singular medication.Of the 11 patients experiencing seizures in the last 12 months, 5 had epilepsy with generalised seizures, 5 with epilepsy with focal seizures and 1 had an unclassified epilepsy syndrome.Half of patients analysed had focal epilepsy, 27 % generalised, and 23 % had either unclassified epilepsy or an unconfirmed diagnosis including possible seizure mimic.

Appointment outcomes
During the appointment there was a change to medical management in 16 patients, including changes to medications or steps taken to address a new clinical issue.There were a further 37 appointments in which the neurologist took some form of administrative action, such as notification to a licensing authority regarding work or driving safety (Table 2).
Of the remaining 47 appointments audited, the outcomes of 30 were classified as "information only".Records described information exchange such as discussion of side effects or comorbidities but no change in plan or intervention by the neurologist.There were 17 appointments classified as "no change", with no reported discussion, administrative or medical action taken.

Predictors of intervention at 12 months
To explore factors associated with greater levels of intervention at a 12-month review appointment, a series of bivariate analyses were used to explore associations between baseline variables and two binary outcomes: (1) change to medical management (yes or no) and (2) occurrence of either a medical or administrative change (yes or no), indicating some form of active intervention by a neurologist resulting from the appointment (Table 3).

Outcome one: Change to medical management
Binary logistic regression was performed to determine whether the experience of seizures in the last year, age or type of epilepsy predict a change in medical management at 12-month review (Table 4).
The occurrence of a seizure in the previous year was the only independent predictor of a change to management (p = 0.009), with patients who had experienced a seizure having 7.16 times the odds (95 % CI 1.68 to 20.69) of having a change to management compared to a person who had not had a seizure in the previous year.
The model explained up to 19.1 % (pseudo R 2 ) of the variation and correctly classified 85 % of cases.

Outcome 2: Occurrence of either a medical or administrative change
A second binary logistic regression model was performed to determine whether the experience of seizures in the last year, age or type of epilepsy predict any active intervention by the neurologist at 12-month review, including either a change in medical management or an administrative outcome (Table 5).
When including both administrative actions and medical changes by the neurologist, the only factor that independently predicted a tangible outcome from the appointment was the type of seizure: the 23 patients with unknown or unclassified seizure type had 0.14 times the odds (95 %CI 0.04 to 0.49) of having any kind of intervention at 12 months compared to those with focal seizures.This model explained up to 16.4 % (pseudo R 2 ) of the variance and correctly predicted 64 % of cases.

Discussion
This audit of 12-month review appointments in an epilepsy clinic indicates that only 16 % of patients experienced a change in medical management and 37 % received an administrative action, typically related to driving, as a result of their appointment.This finding suggests that, for many patients, routine 12-month review appointments at an epilepsy clinic did not lead to significant change in management of their condition and were likely provided for administrative reasons or "justin-case" a new issue arose.The only independent predictor of a change in medical management was the occurrence of a seizure in the previous year.When considering administrative interventions as well as changes in medical management, patients with unknown seizure type were less likely to have some kind of change or action from the appointment compared with those who had confirmed focal or generalized epilepsy.
"Just in time" is a concept arising from inventory management, referring to the concurrent arrival and use of resources, essentially eliminating wasted space and time in a manufacturing process [18].In the epilepsy clinic where this study was conducted, the resource of specialist neurologist care is made available to patients through routine appointments "just in case" any problems have arisen in the past year [19].However, the practice of neurologists booking annual reviews contributes to the system being overloaded, blocking access to patients who would benefit from a "just in time" appointment as well as contributing to long waiting lists for new patients needing to access the  service.
Our study suggests that the utility of these annual "just in case" appointments may be limited, with resources better utilised to assess patients with new or emergent issues.The widespread use of waitlists as a feature of public epilepsy clinics suggests a resource-constrained environment where supply and demand are not well-balanced and rationing of review appointments may be required [2].Reducing low-value review appointments has been shown to be beneficial in increasing efficiency in other outpatient clinic settings [20].This study adds to the evidence that divesting from automatic provision of "just in case" reviews is potentially a safe way to manage increasing demand, particularly if accompanied by new processes enabling "just in time" access as the need arises [13].
Our results show that a select number of patients did experience a change of management at their 12-month review appointment.The key predictor of a change in management was ongoing seizures, indicating, unsurprisingly, that those patients with unstable, or medication refractory epilepsy are more likely to require active management by an epilepsy specialist.Prioritising patients at greatest need for medical intervention enhances access to specialist care for those who are likely to derive greatest benefit [1,13].By limiting routinely scheduled review appointments to patients who meet the definition for drug resistant epilepsy [21], clinics are likely to meet the needs of those who have the greatest requirement for expert neurologist input.
In contrast, those patients who only require completion of administrative actions related to driving and work, are much less likely to require specialist expertise.Primary care providers often have a welldeveloped relationship with their patients and holistic understanding of their needs.They are well-placed to have discussions about these issues, provided that they are adequately supported by regulatory and indemnity systems and equipped with the required information to make recommendations.By prioritising clinic access for medical over regulatory need, demand for clinic review could be substantially reduced, freeing up access for new patients with new medical issues.However, this approach requires clinics to have the capacity and flexibility to review patients when new issues arise.Patients with pharmaco-responsive epilepsy remain at risk of seizure recurrence even after a long duration of seizure freedom [22].In addition, patients with pharmaco-responsive epilepsy may elect to cease medical therapy after a prolonged period of seizure freedom which requires expert advice and guidance [23].Furthermore, major life changes, such as pregnancy in women with epilepsy, is a time when expert navigation is required to weigh up the risks and benefits of treatment options [24].Freeing up 12-month review appointments may improve access for patients at highest need, but must be coupled with new models of care which improve timely assessment, such as "open access" clinics [13], to avoid adverse impacts on long term patient care and clinic viability [25].
The observation that patients with "unclassified" epilepsy were less likely to have a change in management is likely due to the patients in this cohort who had a non-epilepsy diagnosis.For the purpose of this study, patients with an unclassified seizure disorder and those with a nonepilepsy diagnosis, predominantly PNES, were grouped together for analysis due to low overall numbers in these categories.Given there is minimal difference in time to treatment failure in patients with generalised or focal epilepsy versus unclassified epilepsy [11,26] it is not expected that the unclassified seizure group influenced this result.Those Comparisons between groups completed using t-tests for continuous variables and chi-squared tests for categorical variables.* Differences considered statistically significant at p < 0.05.# Factors included in binary logistic regression models due to p < 0.25.

Table 4
Logistic regression analysis of factors predicting a change in medical management at 12-month review.patients with PNES, however, are likely to attend clinic for ongoing support and guidance but would not require adjustments to antiepileptic therapy, which is not required for management, nor necessarily require communication for regulatory approval regarding driving [27].We note there are overall low numbers of patients with PNES in the cohort.Although not addressed in this study, it is likely that this is due to a general policy to not provide long term follow-up for this patient group once a confirmed diagnosis has been made and appropriate supports have been accessed.In addition, after an initial PNES diagnosis a number of patients will have relatively benign course after diagnosis and initial management [28].Unpublished data from our first seizure service show comparable rates of referral for a new diagnosis of PNES.An important consideration is that this audit may not have captured less tangible benefits that may be experienced by patients who attend a 12-month review.For example, many patients may take a high degree of comfort from reassurance provided in these consultations or receive information that is beyond the scope of what is typically provided by their primary care physician.However, it is also important to consider the relative return on investment within the broader context of community need for a scarce and specialised resource.There is clear evidence to demonstrate that delays in access to epilepsy care are associated with worse patient outcomes [1]; resource constraints make it necessary to weigh up the potential benefit arising from a timely consultation versus the reassurance provided in a 12-month review.Future studies could investigate the effect of introducing a decision algorithm, including seizure status, epilepsy type, and other relevant patient factors, to help prioritise how appointments are made.There may also be scope to explore multi-disciplinary models of care utilising other types of health professionals to provide supportive care, so that specialist neurology resources can be targeted most effectively, or enhance the use of self-management strategies.[28,29].
This study has some other limitations that also need to be considered.Our observations may not be generalizable to all services, such as higher acuity services which focus on a more complex or medication refractory cohort, so guidance must be targeted to the characteristics of the patient cohorts.If a "just in time" model is implemented, facilities must be provided to enable ready access and responsive care for those patients such as described in John etal [13].
In conclusion, outpatient clinics require clear parameters about models of care that they provide, including the role of review appointments.There are some services that can only be provided by specialist neurologists, including diagnosis of epilepsy, initial management and stabilisation of the disease, and overseeing the management of chronic and intractable epilepsy.For other issues such as monitoring of health over time, identification of new issues requiring further investigation, regulatory intervention, supportive care and education, others in the healthcare team can share the load.An emphasis on discharge to community-based supports with clear and easy re-entry process would assist in moving clinics towards a "just in time" model where high value care is preserved for those who require it most.

Table 1
Baseline characteristics of participants.

Table 5
Logistic regression analysis of factors predicting a change in either medical management or an administrative intervention by the neurologist at 12-month review.
* Compared to focal seizure type.