Abstract
Background:
Parkinson’s disease (PD) has been hypothesized to be associated with certain personality traits, including conscientiousness and punctuality. However, research aimed at quantifying these traits is largely derived from questionnaire-based personality inventories rather than real-world observations.
Objective:
To explore the presence of a parkinsonian personality profile by assessing the no-show rate of patients with PD versus other neurological disorders.
Methods:
We extracted data from our electronic health record for all neurology appointments over a 78-month interval. Additionally, we obtained primary care appointment data for the same patients over the same timeframe. For each appointment we collected appointment date/time, check-in time, provider, age, sex, insurance type, days between appointment date and scheduling, diagnosis code, and no-show status.
Results:
19,433 unique patients (400 with PD) accounting for a total of 252,347 outpatient appointments were included in our analysis. The overall no-show rate for PD patients was 3% versus 7.4% for patients with other neurologic disorders (OND). No show rates for PD patients were lower than those with OND for both neurology appointments (2.7% versus 13.6%) and for primary care visits (3.1% versus 5.9%).
Conclusions:
Patients with PD have lower no-show rates than patients with OND. Additionally, the no-show rate for patients with PD did not differ between their neurology and primary care appointments, confirming that patient’s personality rather than provider traits account for this difference, and supporting the presence of a parkinsonian personality.
Keywords
INTRODUCTION
Over the last one hundred years various authors have speculated that Parkinson’s disease (PD) is associated with certain personality traits including punctuality, conscientiousness and reduced novelty seeking [1–3]. Initial data to support this theory was limited to case reports. More recent support has come from standardized, quantitative, personality inventories. It is, however, uncertain how questionnaire-based personality profiles correlate with the real-world behaviors of PD patients. Current evidence has demonstrated that cigarette smokers are less likely to develop PD, and a possible explanation for this finding is that specific personality traits such as cautiousness and lack of novelty seeking mitigate against tobacco use [4]. Other explanations have also been proposed [5]. We also know that a minority of PD patients on dopaminergic drugs develop problematic reward-seeking and impulsive behaviors, such as pathological gambling [6, 7]. Dysregulation of the mesocorticolimbic dopamine system is thought to underlie impulse control disorders associated with PD therapies, and conversely, reduced novelty seeking associated with PD may result from a dopaminergic deficiency state [1, 8].
We hypothesized that if PD was in fact associated with a personality profile consisting of high conscientiousness and punctuality, then PD patients would miss fewer clinic appointments and have fewer late arrivals than patients with other neurological disorders. We therefore conducted the present study to assess whether PD was an independent predictor of no-show and non-punctual arrival status.
METHODS
This study was approved under exempt review by the Carilion Clinic Institutional Review Board. Carilion Clinic is a not-for-profit academic medical center associated with the Virginia Tech Carilion School of Medicine. The institution offers residency training in neurology and serves patients affected by the full spectrum of neurological disorders.
De-identified data were extracted from the electronic record (Epic) for all adult (age 18+ years) neurology patient appointments at our Institution (Carilion Clinic) over a 78-month interval. For each appointment, we collected appointment date/time, check-in (arrival) time, provider, age, sex, insurance type, days between appointment date and scheduling (wait time from booking to appointment), diagnosis code, and no-show status. As part of an integrated health system, we were also able to obtain primary care (PCP) appointment data for the same patients over the same timeframe. Neurological diagnoses were grouped into PD and the other neurologic disorders (OND) for statistical analysis. The range of other OND encompassed the full breath of outpatient neurology, including headache syndromes, demyelinating disease, epilepsy, neuromuscular disorders, stroke, cognitive behavioral neurology and other movement disorders.
Analyses of the data were performed at both the visit-level and patient-level. Since many patients had multiple visits, logistic regression for correlated data was done using visit level data to assess the effect of PD on no-show (yes/no), adjusted for age, gender, type of payor, provider, appointment month, appointment day of the week, and time between appointment and scheduling. At the patient level, Poisson regression was done to evaluate the effect of PD on the incidence of no-shows (number of no-shows out of total number of scheduled appointments), adjusted for age and gender. SAS 9.4 (SAS Institute, Cary NC) and R 3.3.2 (R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org) were used for statistical analyses.
RESULTS
We analyzed a total of 19,433 patients (400 with PD) and a total of 252,347 appointments (6,181 for patients with PD). Analysis included 205,098 primary care appointments (4,202 for patients with PD) and 47,249 neurology appointments (1,979 for patients with PD). PD patients accounted for 2.1% of the neurology clinic’s population and accounted for 20.6% of movement disorder clinic visits. Characteristics of the study population are provided in Table 1.
Characteristics of the study population
Medicaid is a federal and state funded program that covers some medical costs for families/individuals in the United States who have limited financial resources. Medicare is a federally funded program that covers some medical costs for patients aged 65 years and older, patients receiving government disability, and those with specific medical conditions including chronic hemodialysis. Commercial insurance is funded by patients and/or their employers. For this cohort, the payor “Missing” means that payor information was not captured by the EHR. Wait time for appointments was defined as the interval between the date of scheduling and date of appointment.
PD patients arrived on average 12±22 minutes prior to their appointment time versus 11±24 minutes for OND patients (p < 0.001). PD patients had an overall no-show rate of 3% versus 7.4% for patients with OND (p < 0.001). Only 25% of our 400 PD patients had ever missed an appointment compared to 43% of the OND population (p < 0.001). PD patients had lower no-show rates than those with OND both in our neurology clinic (2.7% versus 13.6%, p < 0.001), and with primary care (3.1% versus 5.9%, p < 0.001). No-show rates for patients with PD did not differ between primary care and neurology appointments (3.1% vs. 2.7%, p = 0.460).
Poisson and linear regression modeling confirmed PD independently conferred a relative reduction in risk of missing clinic appointments. No-show risk was also lower for women, patients with commercial insurance, older age, and a shorter wait time for appointments (Tables 2 and 3).
Poisson regression modeling of no-show rates for patient level data
Incidence Rate Ratios with 95% confidence intervals are shown. For every decade of advancing age, the Incidence Rate Ratio of a no-show decreases by 0.74, a 26% relative reduction in risk of a no-show. Male gender conferred an 8% relative increase in risk of a no-show. PD diagnosis conferred a 35% relative reduction in risk of a no-show.
Logistic regression modeling of no-show rates for visit level data
Odds ratios and 95% confidence intervals are shown. The odds of missing a clinic appointment were lower for patients who had been diagnosed with PD. The odds of missing a clinic appointment were highest for patients without commercial insurance and for those with long wait times. Wait time for appointments was defined as the interval between the date of scheduling and date of appointment.
DISCUSSION
In this single-center, cross-sectional study we found that PD patients had significantly lower clinic no-show rates compared to non-PD patients after adjusting for age, gender, payor, provider, and wait time from booking to appointment. PD patients also arrived for appointments earlier than those with other neurological disorders, though the difference in arrival time was modest and not likely to be noted by clinicians.
Strengths of our study include the relatively large sample size and the fact that it provides a novel means of assessing and quantifying behaviors that previously had been assessed largely through questionnaires.
Some methodological limitations should be noted. This was a study of de-identified data, and for that reason International Statistical Classification of Disease (ICD) codes from our neurology clinic were used to identify PD rather than formal diagnostic criteria, as might be used in the setting of a prospective clinical trial [9, 10]. While all PD patients were diagnosed by a neurologist, and most by a single, fellowship-trained movement disorder specialist, it is likely that diagnostic accuracy is lower with our methodology, and it is not known how diagnostic errors might impact study results. While diagnosis codes were entered into the electronic record by neurologists, all other data used in our analysis were entered by non-clinical staff as part of routine patient care. It should be noted that patients who missed every appointment in our clinic could not be included in our analysis because no diagnostic code would ever have been entered by a neurologist.
Additionally, data was acquired retrospectively, and likely there are confounders for which we cannot account. The persistent rather than paroxysmal nature of PD symptoms, the availability of effective PD therapies, and the frequent need of medication adjustments all may foster a lower no-show rate in our neurology clinic independent of a specific PD-related personality profile. The fact that the no-show rates for PD patients did not differ between their neurology and primary care appointments, however, could indicate that patient traits rather than provider/clinic factors drive the difference. We also anticipate the presence of cognitive impairment and availability of caregiver support would impact odds of missing an appointment, and this data could not be included in our analysis.
If PD is associated with a specific personality profile, it is uncertain to what extent these traits are mediated by dopaminergic deficiency. Prior studies in PD have correlated results of personality questionnaires with presynaptic striatal dopamine dysfunction and found mixed results [1, 11]. Our study included no treatment data, and we therefore are unable to determine whether PD therapies influence clinic no-show rates. Likewise, because we cannot reliably date PD onset, we cannot determine whether the lower rate of missed appointments is a premorbid feature of PD. Other work has suggested that personality changes in PD pre-date clinically overt motor symptoms, and a recent study has shown a similar personality profile in patients with REM behavior disorder, a known risk factor for PD [1, 12].
Over the last several years, electronic medical records have gained widespread use and contain a massive amount of clinical data. While data mining studies are not likely to yield definitive knowledge about PD, the growing volume of electronic medical record data present new opportunities to test hypotheses in a “real-world” setting. Acknowledging the limitations inherent in our design, this study supports the presence of a PD personality profile consisting of conscientiousness and punctuality, as suggested by prior studies using standardized personality testing.
CONFLICT OF INTEREST
The authors have no conflict of interest to report.