Abstract
Background
Individuals with psychotic symptoms experience substantial morbidity and have shortened life expectancies; early treatment may mitigate the worst effects. Understanding care preceding a first psychotic disorder diagnosis is critical to inform early detection and intervention.
Study Design
In this observational cohort study using comprehensive information from the Massachusetts All-Payer Claims Database, we identified the first psychotic disorder diagnosis in 2016, excluding those with historical psychotic disorder diagnoses in the prior 48 months among those continuous enrollment data. We reviewed visits, medications, and hospitalizations 2012–2016. We used logistic regression to examine characteristics associated with pre-diagnosis antipsychotic use.
Study Results
There were 2505 individuals aged 15–35 years (146 per 100 000 similarly aged individuals in the database) with a new psychotic disorder diagnosis in 2016. Most (97%) had at least one outpatient visit in the preceding 48 months; 89% had a prior mental health diagnosis unrelated to psychosis (eg, anxiety [60%], depression [60%]). Many received psychotropic medications (77%), including antipsychotic medications (46%), and 68% had a visit for injury or trauma during the preceding 48 months. Characteristics associated with filling an antipsychotic medication before the psychotic disorder diagnosis included male sex and Medicaid insurance at psychosis diagnosis.
Conclusions
In this insured population of Massachusetts residents with a new psychotic disorder diagnosis, nearly all had some healthcare utilization, visits for injury or trauma were common, and nearly half filled an antipsychotic medication in the preceding 48 months. These patterns of care could represent either pre-disease signals, delays, or both in receiving a formal diagnosis.
Keywords: first episode psychosis, healthcare utilization, psychotropic medications
Introduction
Patients with psychotic disorders experience substantial morbidity and have shortened life expectancies compared to those without psychosis.1 Early intervention after a first episode of psychosis can improve disease trajectory,2,3 and there have been several policy efforts to increase access to treatment after a first episode of psychosis, such as increasing funding for specialty clinics.4 Lack of recognition of psychotic symptoms leading to increased duration of untreated psychosis is associated with poor illness prognosis and reduced effectiveness of treatment interventions.5–9
Despite recognition of the seriousness of psychosis and promising interventions, the vast majority of our knowledge about the events preceding the initial psychotic disorder diagnosis or how best to identify patients at risk for developing psychotic symptoms come from settings outside of the United States. Studies from Canada, the United Kingdom, and Denmark examining patterns of healthcare use prior to a new psychotic disorder diagnosis found increased primary care use, as compared to the general population, with more frequent use in the several months before the new diagnosis.10–12 Within primary care, predictors of future psychotic disorder diagnosis include patients with increasing suicidal behavior, particularly among young men, and multiple affective symptoms.10,13,14 Other studies suggest that approximately one-half to two-thirds of individuals who present with a first episode of psychosis had prior contact with psychiatric care.15,16 In the United States, there are well documented challenges to obtaining psychiatric care as well as limited information on the extent to which individuals receive psychiatric care before their psychotic disorder diagnosis. Moreover, use of inpatient psychiatric care or mental health-related emergency department (ED) visits in the United States are more likely in the period before a first psychotic disorder diagnosis as compared to a first depression diagnosis.17
While knowledge gleaned from other countries can be valuable in characterizing healthcare utilization prior to a new psychotic disorder and identifying potential factors that predict risk, these results may not be representative of the United States experience. Several factors including population differences, how healthcare is financed, and delivery system differences (eg, fragmentation of care across delivery sites in the United States and the potential association between both risk factors and the disease itself with such fragmentation) limit the applicability to US populations. Some studies have used large, national datasets to examine service utilization prior to a diagnosis, however, these have been limited by other factors including limited patient age range and diagnosis.15,17,18 Other studies have examined enrollees within a health maintenance organization (HMO), however, these could have limited generalizability since individuals self-select to be in an HMO system and the systems may have different access pathways to psychiatry than non-HMO systems.19 Thus, identifying individuals who may have unrecognized psychotic symptoms, and understanding the patterns of care preceding a first psychotic disorder diagnosis at a population level in a US cohort with varying insurance types and access to healthcare services is critical to inform US efforts towards early detection and intervention services.18,20
While it is possible to identify the first healthcare encounter for individuals with new psychotic disorder diagnoses in administrative datasets, there may be a disconnect between when symptoms develop and when the diagnosis is given or billed for through insurance. This may result from reasons ranging from poor access to care to clinical uncertainty around diagnosis. Prior studies suggest that non-psychotic psychiatric disorders, such as depression and anxiety, might be associated with receiving a subsequent psychotic disorder diagnosis.20,21 These diagnoses could represent precursor illnesses, manifestations of unrecognized psychotic symptoms, or treatment for prevalent cases without a claims diagnosis. Further, some historical diagnoses may represent a prodromal period or early warning signs of an emerging psychotic disorder, eg, non-specific changes in thoughts or behaviors. These could manifest as non-specific injuries or accidents in the period before receiving a psychotic disorder diagnosis and, with recognition, could aid in earlier identification and treatment. Examination of the patterns of care, including treatment patterns, leading up to an initial psychotic disorder diagnosis with attention to these factors could help better identify patients with emerging psychosis, and eventually lead to tools such as population-level prediction algorithms or detection of area- or system-based supply and demand mismatches.
In this observational cohort study, we describe how frequently patients receive medical or psychiatric care in the years preceding their first psychotic disorder diagnosis. Leveraging the comprehensive data capture and ability to track individuals across insurance changes and delivery settings within the Massachusetts All-Payer Claims Database (APCD), we employ a population approach that includes nearly all Massachusetts residents aged 15–35 years of age. We examined the patterns of care and treatment over the preceding 4-year period among all individuals who received their first psychotic disorder diagnosis in 2016 and had continuous enrollment data across the study period. We also examined factors associated with filling a prescription for a psychotropic medication, and specifically antipsychotic medication, before a psychotic disorder diagnosis.
Methods
Population and Data
The 2012–2016 Massachusetts All-Payer Claims Dataset (APCD) is a state-wide, multi-payer dataset with medical and pharmacy claims containing information about all services paid through health insurance. The dataset contains information on Massachusetts residents and includes group commercial, individual commercial, Medicare Advantage, and Medicaid data, but does not include other payers (eg, fee-for-service Medicare, TRICARE).22 Using the APCD, we identified all individuals aged 15–35 with a psychotic disorder diagnosis in 2016 (the index diagnosis) who had no prior psychotic disorder diagnoses in the preceding 48 months. The age range was chosen based on previous literature about age of onset of psychotic disorders and the time period was determined based on data availability and prior studies showing that greater lookback windows reduce mis-identification of non-incident cases.23 The APCD contains an enrollment file that allows tracking of unique individuals over time regardless of changes in insurance type. Given that it is not possible to differentiate between loss of insurance coverage, departure from the state, or death using this dataset, we required that all individuals were present in the enrollment file for the full 48 months prior to their psychotic disorder diagnosis in 2016. We defined psychotic disorders to include mood disorders with psychotic features for this cohort and did not include psychotic disorders secondary to substance misuse or other medical disorders (Supplementary table S1). We collected demographic and clinical characteristics for the cohort at the time of index diagnosis. Details of the cohort construction are available in a published article.23 The Mass General Brigham Institutional Review Board approved this study.
Healthcare Utilization
Using medical and pharmacy claims, we identified the location of the index psychotic disorder diagnosis (ED visit or observational stay, inpatient admission [psychiatric vs non-psychiatric], or outpatient visit). We assessed all clinical encounters during the 48 months prior to the month of the new 2016 psychotic disorder diagnosis.
Diagnoses
We examined diagnoses associated with clinical event types, categorizing these as mental health and non-mental health based on ICD-9 and ICD-10 codes given that this version of the APCD did not permit specialty classification based on board certification. Non-psychotic psychiatric disorders, such as depression, anxiety, and attentional difficulties, might be associated with receiving a subsequent psychotic disorder diagnosis and some patients may present with psychotic symptoms, but not receive psychotic disorder diagnoses immediately (eg, do not meet time-based criteria of symptoms for at least six months, clinician is unsure of underlying disorder). For these reasons, we characterized the types of non-psychotic disorder diagnoses received in the 48 months prior to the index psychotic disorder diagnosis.
We hypothesized that some early warning signs of risk for psychosis could manifest as non-specific injuries or accidents in the period before receiving a psychotic disorder diagnosis. During this period, individuals may be symptomatic even if this is not recognized by treating clinicians, particularly for those with limited access to psychiatric care. We identified individuals with history of injury, accident, or trauma diagnoses as these may indicate psychotic symptoms or an early signal of a psychotic disorder. We compared the frequency of these diagnoses over time to an age/sex matched cohort as a reference. This cohort consisted of individuals aged 15–35 without any mental health disorder diagnoses across the study period who were engaged with the healthcare system, having at least one clinical encounter during 2016.
Psychotropic Medication
We examined psychotropic medications dispensed over time, including time between first prescription fill and new psychotic disorder diagnosis. We hypothesized that individuals would receive increasing antipsychotic prescriptions in the years prior to a new psychotic disorder diagnosis. These patterns could represent treatment for an appropriate or related indication (eg, bipolar disorder, current episode manic without psychotic features) prior to subsequent development of a psychotic disorder. Alternatively, individuals could receive these treatments for psychotic symptoms without receiving the diagnosis. To examine this, we identified the proportion of individuals with antipsychotic prescriptions who also had an FDA indicated non-psychotic disorder diagnosis code for use of antipsychotics.24
Analyses
We plotted the accumulation of events, diagnoses, and medications over time for the cohort. We assessed access to care by examining if individuals had any healthcare utilization prior to the index diagnosis, then examined utilization by site of service. We also examined prior psychiatric care including outpatient visits, ED visits with mental health diagnoses, and inpatient psychiatric hospitalizations by calculating mean use and utilization rates by 12-month period. We examined characteristics associated with receiving non-specific injury or accident diagnoses in the 48 months prior to a new psychotic disorder diagnosis using a logistic regression model. We also used a logistic regression model to examine the characteristics associated with having historical antipsychotic use in the preceding 48 months. In both logistic models, we included sex, age category, index diagnosis location and insurance, and baseline comorbidities; we report odds ratios and corresponding 95% confidence intervals for these models.
Sensitivity Analyses
In sensitivity analyses, we repeated these analyses using a narrower definition of psychosis, ie, new schizophrenia spectrum disorder diagnoses (Supplementary table S1), to examine for impact of heterogeneity in diagnoses. To understand the degree of increased healthcare utilization and psychotropic prescribing in the early psychosis population, we examined clinical events in an age/sex matched cohort of individuals without any mental health disorder diagnosis across the study period. To improve comparability between those with psychotic disorder diagnoses and those without any history of mental health diagnoses, we matched to individuals who received healthcare services at least once in 2016.
Results
There were 2505 individuals in our cohort (ie, those with a new psychotic disorder diagnosis in 2016 who had continuous enrollment across 48 months and no prior psychotic disorder diagnoses; Supplementary table S2), approximately 146 per 100 000 individuals aged 15–35 in the database. The average age at time of diagnosis was 24.7 years and 51.7% were male. More than half of individuals had Medicaid at the time of their psychotic disorder diagnosis, and a higher proportion of those who had historical antipsychotic medication prescription fills (66.1%) were covered by Medicaid than those who had not (55.3%; table 1).
Table 1.
Characteristics of the Cohort of Individuals with a New Psychotic Disorder Diagnosis in 2016
| Characteristic (Percentage or Mean ± Standard Deviation) | Full Cohort | Antipsychotic Prescription Filled in 48 Months Prior to a New Psychotic Disorder Diagnosis | |
|---|---|---|---|
| N = 2505 | Yes (N = 1143) | No (N = 1362) | |
| Male | 51.7 | 54.1 | 49.7 |
| Average age (years) | 24.7 ± 6.0 | 25.1 ± 6.0 | 24.4 ± 6.0 |
| Age category, % | |||
| 15–17 years | 13.1 | 11.0 | 14.8 |
| 18–29 years | 58.8 | 58.6 | 58.9 |
| 30–35 years | 28.1 | 30.4 | 26.3 |
| Index diagnosis location, % | |||
| ED/observational | 36.2 | 35.2 | 37.0 |
| (a) Admitted to the hospital within 7 days of the ED/observational visit* | 8.9 | 6.6 | 10.8 |
| (b) Discharged from a psychiatric hospital within 30 days of the ED/observational visit* | 6.3 | 5.2 | 7.3 |
| Inpatient, psychiatry | 4.4 | 5.2 | 3.7 |
| Inpatient, non-psychiatry | 8.3 | 9.0 | 7.7 |
| Outpatient | 51.1 | 50.7 | 51.5 |
| Index diagnosis insurance, % | |||
| Medicaid | 60.2 | 66.1 | 55.3 |
| Commercial, group | 29.2 | 23.9 | 33.6 |
| Commercial, individual | 2.7 | 2.9 | 2.6 |
| Other | 7.9 | 7.2 | 8.5 |
| Baseline comorbidities, % | |||
| Anxiety disorders | 52.8 | 65.9 | 41.8 |
| Depression | 48.3 | 60.6 | 37.9 |
| ADHD, conduct disorders | 26.6 | 34.2 | 20.2 |
| Bipolar disorder | 33.1 | 52.3 | 17.0 |
| Personality disorders | 5.4 | 9.5 | 2.0 |
| Asthma | 19.5 | 22.5 | 17.0 |
| Obesity | 14.7 | 17.1 | 12.6 |
| Autism | 5.3 | 8.5 | 2.6 |
| Epilepsy | 4.6 | 5.5 | 3.8 |
| Learning disabilities | 1.8 | 2.2 | 1.5 |
| Other ID/DD disorders | 9.0 | 12.3 | 6.2 |
| Hypertension | 9.2 | 12.3 | 6.6 |
| Hyperlipidemia | 7.7 | 10.0 | 5.8 |
| Diabetes | 4.7 | 6.4 | 3.2 |
Notes: (1) Characteristics obtained at time of index diagnosis. (2) Baseline comorbidities were defined using the Chronic Conditions Data Warehouse ICD-9 codes during the two-year period of 2013Q4–2015Q3, prior to the new psychotic disorder diagnosis in 2016. (3) ED, emergency department; ADHD, attention-deficit/hyperactivity disorder; ID/DD, intellectual or developmental disabilities. (4) Other insurance category includes Medicare, uncategorized commercial, mixed, missing, and unidentified.
*Proportion calculated based on the total cohort.
Clinical Events
Almost all individuals in the cohort had a history of at least one outpatient visit (96.8%) and 84.0% had a mental health-related outpatient visit in the 48 months prior to their psychotic disorder diagnosis (table 2). In the year prior to their diagnosis, 68.9% of individuals had a mental health-related outpatient visit with an average of 24.2 (standard deviation = 49.4) total outpatient visits and 11.0 (standard deviation = 27.7) mental health-related outpatient visits within that year.
Table 2.
Cumulative Healthcare Utilization in the Months Preceding a New Psychotic Disorder Diagnosis
| Healthcare Utilization Measure | During the X Months Preceding a New Psychotic Disorder Diagnosis in 2016 (N = 2505) |
|||
|---|---|---|---|---|
| X = 12 | X = 24 | X = 36 | X = 48 | |
| Proportion with any utilization (%) | ||||
| Any inpatient stay | 14.2 | 19.7 | 24.6 | 28.3 |
| Any psychiatric inpatient stay | 5.6 | 8.5 | 10.3 | 11.5 |
| Any ED visit/observational stay | 55.1 | 67.1 | 73.4 | 79.0 |
| Any ED visit/observational stay with a mental health diagnosis | 33.0 | 43.2 | 48.9 | 53.3 |
| Any outpatient visit | 91.0 | 94.4 | 96.4 | 96.8 |
| Any outpatient visit with a mental health diagnosis | 68.9 | 77.2 | 81.4 | 84.0 |
| Utilization rate (inpatient days or visits per patient per 12 months, not conditional on any utilization): | ||||
| Inpatient days | 1.5 | 1.2 | 1.1 | 1.0 |
| Psychiatric inpatient days | 0.8 | 0.7 | 0.6 | 0.5 |
| ED visits/observational stays | 1.7 | 1.4 | 1.3 | 1.2 |
| ED visits/observational stays with a mental health diagnosis | 0.6 | 0.6 | 0.5 | 0.5 |
| Outpatient visits | 24.2 | 22.6 | 21.6 | 20.5 |
| Outpatient visits with a mental health diagnosis | 11.0 | 11.6 | 11.3 | 10.7 |
Notes: (1) The preceding period also included the days in the index diagnosis month but preceding the index diagnosis day. (2) Utilization rates not conditional on any utilization. (3) Outpatient visits capped at most one visit per day and restricted to those started and ended on the same day. (4) Psychiatric inpatient stays were identified by service and billing National Provider Identifiers (NPIs) and linked taxonomy codes. (5) Mental health diagnoses included non-psychotic mental, behavioral, and neurodevelopmental disorders, or intentional self-harm/suicide attempt (F00–F99, excl. F20–F25, F28–F29, F30.2, F31.2, F31.5, F31.64, F32.3, F33.3; X60–X84; T14.91). (6) In this Table, the proportions are cumulative across time periods, whereas the rates are not and are calculated per each 12-month period.
As expected, ED utilization was high. More than half of individuals (55.1%) had an ED visit in the year prior to their psychotic disorder diagnosis and approximately one-third (33.0%) had a mental health-related ED visit in that time (table 2). In the 48 months prior to the psychotic disorder diagnosis, more than one quarter (28.3%) of individuals had at least one inpatient hospitalization and 11.5% had a psychiatric hospitalization.
While almost all individuals in the age/sex matched comparison group (95.6%) had an outpatient visit in the 4 years prior to 2016, other healthcare utilization was lower compared with those with a new psychotic disorder diagnosis (ie, 12.9% had an ED visit and 2.7% had an inpatient hospitalization in the year prior to 2016, average number of outpatient visits was 3.9; Supplementary table S3).
Prior Diagnoses and Early Signals of Disease
Across the 48 months prior to a new psychotic disorder diagnosis, 88.6% of patients received a non-psychotic disorder related mental health diagnosis. The majority had historical diagnoses of anxiety or depressive disorders (each at 59.6% in the 48 months prior to a new psychotic disorder diagnosis, table 3). Bipolar disorders and other/unspecified mood disorders without psychotic features were also common (30.5% and 37.6% in the 48 months prior to a new psychotic disorder diagnosis, respectively). Further, 28.1% of individuals received an ADHD diagnosis in the 48 months prior to their first psychotic disorder diagnosis.
Table 3.
Cumulative Frequency of Non-psychotic Mental, Behavioral, and Neurodevelopmental Disorder Diagnoses and Antipsychotic Medication Prescriptions Dispensed in the Months Preceding a New Psychotic Disorder Diagnosis
| Percentage of Patients During the X Months Preceding a New Psychotic Disorder Diagnosis in 2016 (N = 2505) | ||||
|---|---|---|---|---|
| X = 12 | X = 24 | X = 36 | X = 48 | |
| Diagnosis category (ICD-10 F-code) | ||||
| Anxiety disorders (F40–F41) | 43.4 | 51.6 | 56.6 | 59.6 |
| Depressive disorders, non-psychotic (F32–F33, excl. F32.3, F33,3) | 42.9 | 51.7 | 56.2 | 59.6 |
| Adjustment disorders (F43) | 23.6 | 31.3 | 36.3 | 40.1 |
| Other/unspecified mood disorders (F34–F39, F99) | 19.3 | 28.5 | 34.0 | 37.6 |
| Bipolar disorders, non-psychotic (F30–F31, excl. F30.2, F31.2, F31.5, F31.64) | 21.2 | 25.5 | 28.7 | 30.5 |
| ADHD (F90) | 17.8 | 22.6 | 25.7 | 28.1 |
| Substance use disorders due to use of tobacco (F17) | 14.5 | 20.8 | 24.2 | 26.6 |
| Other | 28.5 | 35.8 | 41.4 | 45.4 |
| Any of the above diagnosis categories | 76.8 | 83.5 | 87.1 | 88.6 |
| Only one of the above diagnosis categories | 19.4 | 15.5 | 13.3 | 11.3 |
| Psychotropic medication category | ||||
| Antipsychotic | 36.6 | 40.6 | 43.4 | 45.6 |
| Antidepressant | 44.5 | 50.8 | 54.9 | 58.0 |
| Anti-anxiety | 35.6 | 41.7 | 45.2 | 48.6 |
| Subcategory: antipsychotic, long acting | 0.9 | 1.0 | 1.0 | 1.0 |
| Mood stabilizer | 24.0 | 27.7 | 30.7 | 32.5 |
| ADHD | 18.2 | 20.8 | 23.0 | 24.6 |
| Subcategory: ADHD, stimulants | 17.3 | 19.8 | 22.0 | 23.7 |
| Sub-subcategory: ADHD, stimulants, amphetamine | 13.5 | 15.2 | 16.8 | 17.9 |
| Any of the above medication categories | 66.5 | 71.5 | 74.4 | 77.0 |
| Only one of the above medication categories | 17.0 | 16.5 | 15.3 | 15.3 |
Notes: (1) The preceding period also included the days in the index diagnosis month but preceding the index diagnosis day. (2) ICD-9 to ICD-10 codes conversions were based on a crosswalk developed by the CMS and processed by the NBER. (3) ICD-10 categories were based on groupings by the WHO. (4) Diagnoses of substance use disorders (F10–F19 and corresponding ICD-9 codes) other than those due to use of tobacco (F17 and corresponding ICD-9 codes) were redacted from the data. (5) Other F-code categories include: Adjustment disorders (F43); Delirium (F05); Dementia (F01–F03); Developmental disorders (F80–F89); Disorders with onset in childhood (F91–F98); Eating disorders (F50); Gender identity disorders (F64); Impulse disorders (F63); Intellectual disabilities (F70–F79); OCD (F42); Organic syndromes (F04, F06–F09); Other behavioral syndromes associated with psychological distress/physical factors (F52–F59); Other stress disorders (F44–F48); Personality disorders (F60, F68–F69); Sexual disorders (F65–F66); Sleep disorders (F51). (6) ADHD, attention-deficit hyperactivity disorder. (7) Note, please see Supplementary table S8 for the percentage of patients who had non-psychotic disorder diagnoses with FDA-approved indications for antipsychotics preceding the new psychotic disorder diagnosis. Indications include: depression without psychosis (F32 excl. F32.3, F33 excl. F33.3), bipolar disorder without psychosis (F30 excl. F30.2, F31.0, F31.1, F31.3, F31.4, F31.6 excl. F31.64, F31.7–F31.9), severe childhood behavioral problems (F90–F98), generalized non-psychotic anxiety (F41.1), and agitation; irritability associated with autistic disorder (R45.1; F84.0).
We also found that 68.1% received a diagnosis consistent with injury/poisoning/trauma or accident during the preceding 48 months before their initial psychotic disorder diagnosis (figure 1, Supplementary table S4). The comparison cohort of individuals with no prior mental health disorder diagnoses had lower rates (42.3%) of a diagnosis consistent with injury/poisoning/trauma or accident during the preceding 48 months preceding their first clinical encounter in 2016 (Supplementary table S5).
Fig. 1.
Cumulative frequency of specific medical diagnoses in the years preceding a new psychotic disorder diagnosis. (1) The preceding period also included the days in the index diagnosis month but preceding the index diagnosis day. (2) ICD-9 to ICD-10 codes conversions were based on a crosswalk developed by the CMS and processed by the NBER. (3) ICD-10 categories were based on groupings by the WHO. (4) V-codes, included in both ICD-9 and ICD-10 but with different definitions, were omitted in the last quarter of 2015, due to the ICD version transition and a lack of reliable ICD version indicator in the MA APCD.
Individuals who received their index diagnosis in the outpatient setting had lower odds of having a non-specific injury or accident diagnosis as compared to individuals who received their index diagnosis in the ED or observational setting (Supplementary table S6).
Psychotropic Prescriptions
Of the 2505 individuals in our cohort, 77.0% had a prescription fill for a psychotropic medication (antipsychotic, antidepressant, anti-anxiety, mood stabilizer, or stimulant/ADHD treatment) in the 48 months prior to their psychotic disorder diagnosis. In the year immediately preceding their diagnosis, approximately two-thirds (66.5%) received a psychotropic prescription.
Antidepressants were the most commonly prescribed psychotropic medication with 58% of individuals filling a prescription for an antidepressant in the 48 months prior to their diagnosis and 44.5% filling a prescription in the year prior (table 3). Medications to treat ADHD were least common with only one quarter (24.6%) of individuals filling a prescription in the 48 months prior to their psychotic disorder diagnosis (table 3).
Nearly half of patients (45.6%) received an antipsychotic in the 48 months prior to their diagnosis and 36.6% in the year prior. The average duration between time of first observed antipsychotic dispensed and time of receiving a new psychotic disorder diagnosis was 28.1 months (standard deviation 19.2 months; Supplementary figure S1 and table S7). Of those individuals who received an antipsychotic prescription prior to their psychotic disorder diagnosis, the majority had a corresponding non-psychotic disorder diagnosis code with an FDA indication for treatment with antipsychotic medication (76.6% in the year prior, 88.8% in the 4 years prior); of those who did not receive an antipsychotic prescription, 55.7% and 68.2%, respectively, had such a code (Supplementary table S8).
Table 4 displays the characteristics associated with the odds of filling an antipsychotic medication prescription prior to a first psychotic disorder diagnosis. Males had higher odds of receiving an antipsychotic medication prior to receiving a psychotic disorder diagnosis. Further, individuals with Medicaid at time of their psychotic disorder diagnosis had higher odds of filling an antipsychotic medication prescription prior to their psychotic disorder diagnosis compared to individuals with group commercial insurance.
Table 4.
Association Between Patient Characteristics and Having any Antipsychotic Medication Prescription Dispensed in the 48 Months Preceding a New Psychotic Disorder Diagnosis
| Characteristic | Odds of Having an Antipsychotic Medication Prescription Dispensed Prior to a New Psychotic Disorder Diagnosis (N = 2505) |
|---|---|
| Odds ratio (95% CI) | |
| Male | 1.45 (1.21, 1.75) |
| Age category (reference group: 15–17 years) | |
| 18–29 years | 1.31 (1.00, 1.73) |
| 30–35 years | 1.29 (0.95, 1.77) |
| Index diagnosis location (reference group: ED/observational) | |
| Inpatient, psychiatry | 1.42 (0.90, 2.23) |
| Inpatient, non-psychiatry | 1.17 (0.83, 1.64) |
| Outpatient | 1.11 (0.91, 1.34) |
| Index diagnosis insurance (reference group: commercial, group) | |
| Medicaid | 1.36 (1.11, 1.68) |
| Commercial, individual | 1.51 (0.88, 2.57) |
| Other | 0.99 (0.69, 1.41) |
| Baseline comorbidities | |
| Anxiety disorders | 1.54 (1.26, 1.88) |
| Depression | 1.38 (1.12, 1.69) |
| ADHD, conduct disorders, and hyperkinetic syndrome | 1.11 (0.90, 1.38) |
| Bipolar disorder | 3.59 (2.92, 4.41) |
| Personality disorders | 2.27 (1.42, 3.63) |
| Asthma | 0.98 (0.78, 1.23) |
| Obesity | 1.00 (0.77, 1.30) |
| Autism | 2.15 (1.37, 3.39) |
| Epilepsy | 0.80 (0.52, 1.23) |
| Learning disabilities | 0.73 (0.37, 1.45) |
| Other ID/DD disorders | 1.43 (1.02, 2.01) |
| Hypertension | 1.13 (0.81, 1.58) |
| Hyperlipidemia | 1.26 (0.89, 1.79) |
| Diabetes | 1.03 (0.66, 1.63) |
Notes: (1) Baseline comorbidities were defined using the Chronic Conditions Data Warehouse ICD-9 codes during the 2-year period of 2013Q4–2015Q3, prior to the new psychotic disorder diagnosis in 2016. (2) ED, emergency department; ADHD, attention-deficit/hyperactivity disorder; ID/DD, intellectual or developmental disabilities.
Sensitivity Analyses
Results were similar when examining these patterns for individuals with new onset schizophrenia spectrum disorders.
Discussion
In a cohort of individuals with new onset psychotic disorders, we find high levels of healthcare utilization, mental health diagnoses, and psychotropic medication use in the years before their first psychotic disorder diagnosis. Nearly everyone in the cohort received some outpatient care suggesting at least basic engagement in medical care in Massachusetts among insured individuals. Further, the majority (84%) of individuals had some psychiatric care in the 4 years prior to their diagnosis, with a mean of 24 outpatient visits per year and 11 mental health-related outpatient visits per year in the year prior the first psychotic disorder diagnosis. We also find that, even for individuals in our age/sex matched cohort without a mental health diagnosis, the vast majority (>95%) had at least one clinical encounter suggesting at least some form of engagement with the healthcare system. Overall, nearly half (46%) of the cohort with a new psychotic disorder diagnosis received an antipsychotic medication in the 48 months preceding their first psychotic disorder diagnosis. This pattern suggests that a psychotic disorder diagnosis arises for most individuals in a context of evolving psychopathology and ongoing psychiatric treatment, and not in isolation.
While studies have examined healthcare utilization prior to a first psychotic disorder diagnosis, many describe healthcare delivery systems in countries with universal healthcare and often restrict to individuals with new onset schizophrenia spectrum disorder diagnoses rather than all psychotic disorder diagnoses.10,11,15,25 These non-US studies have demonstrated that visits to primary care increase in the years prior to a new schizophrenia spectrum disorder diagnosis and that EDs are the most common referral source to psychiatric services. One study in Denmark demonstrated high rates of mental health diagnoses (primarily psychotic disorder diagnoses) prior to a new schizophrenia diagnosis and high rates of psychotropic use prior to the schizophrenia diagnosis with more than half (52%) receiving antipsychotic medication.25 While we find similar rates of utilization and psychotropic medication use, our findings describe patterns of care prior to any psychotic disorder diagnosis, rather than a first schizophrenia spectrum disorder diagnosis, among a cohort with continuous insurance coverage for at least 48 months.
One prior study examining healthcare utilization in the 3 years before a new onset psychotic disorder found lower clinical event rates than our study (eg, one-third of individuals had a mental health-related ED visit in the 3 years prior compared to 49%, albeit including observational stays, in our study). Further, they found that only 40% received a psychotropic medication prescription in the year prior to a diagnosis, with antidepressants the most commonly prescribed.17 Notably, we find increased use of healthcare services and higher rates of psychotropic medication prescription for several years prior to a new onset psychotic disorder and at higher rates than an age/sex matched cohort. While the average time between the first antipsychotic prescription and psychotic disorder diagnosis was nearly 2.5 years, approximately three quarters of individuals had other recorded diagnoses that are also indications for treatment by antipsychotic medications. Notably, we find that males and individuals with Medicaid (compared to group commercial insurance) had higher odds of receiving a prescription prior to their psychotic disorder diagnosis.
There are several potential reasons that nearly half of the patients with a new psychotic disorder diagnosis had received an antipsychotic medication before receiving the diagnosis. First, antipsychotic medications have multiple FDA-approved indications. In our sample, 89% of those receiving the medication before the diagnosis had another diagnosis that represented an FDA-approved indication. The protracted time period between first antipsychotic medication fill and psychotic disorder diagnosis could represent treatment for another FDA-approved condition such as a mood disorder potentially with progression to a psychotic disorder, or use for an off-label indication. Other studies examining healthcare utilization and psychotropic medication use preceding a schizophrenia spectrum disorder diagnosis have largely attributed prior receipt of antipsychotic medications to indicate treatment for other behavioral health-related comorbidities (eg, non-schizophrenia psychotic disorders, bipolar affective disorder).18,26 Second, physicians frequently prescribe antipsychotic medications for other, non-FDA-approved indications, ie, off-label prescribing. Third, high levels of clinical uncertainty early in the disease course for patients with psychotic symptoms, which when combined with the stigma of diagnosis and perceptions about the disease course could increase the clinical threshold required to make the diagnosis or delay coding. Further, individual clinicians could vary in their threshold for making a diagnosis, which also could be higher than their threshold for treatment of symptoms, such as with an antipsychotic medication, suggesting that some individuals may receive treatment for psychotic symptoms prior to receiving a billing diagnosis for a psychotic disorder (ie, treatment of prevalent cases). In this study, we were unable to differentiate between these possibilities; moreover, the possibilities are not mutually exclusive. Despite this, understanding who ultimately receives a psychotic disorder diagnosis and when is critical to limit delays in treatment or referral into early intervention services, particularly for those who are not yet diagnosed with a psychotic disorder. Further, varying patterns of treatment with antipsychotic medication in the period leading up to a diagnosis may also have implications for acceptance into early intervention programs and warrants further investigation.
We also find that more than two-thirds (68.1%) of individuals received a diagnosis consistent with injury/trauma or accident in the 4 years prior to their new psychotic disorder diagnosis which is more than 1.5 times the rate in an age/sex matched cohort with no history of mental health diagnoses. While we do not know whether individuals in our cohort were experiencing psychotic symptoms during these events, these findings suggest that the time leading up to a psychotic disorder diagnosis is a risky period and may represent unrecognized levels of behavioral dysregulation.
Our study has some limitations. First, while we used a broad definition to identify patients with psychotic disorders, we are unable to identify the timing of symptom onset, rather we can identify the timing of diagnosis using the diagnostic codes contained within health insurance claims. We are also unable to account for heterogeneity in clinical practice patterns (ie, timing and threshold to use a diagnosis code indicating psychotic symptoms) which could impact timing of diagnosis. Together, these factors may help explain why the average age of our cohort is older than in prior studies which may be able to identify timing of symptom onset more easily. Further, diagnostic uncertainty early in the disease course or limited clinical experience or knowledge regarding the identification of patients with emerging psychotic symptoms may influence which diagnosis codes are used and suggest a potential area for additional clinician education. We attempted to examine this by identifying events that may be early signals of disease, however, these data do not provide information on whether patients were experiencing psychotic symptoms at that time. Given that these events may indicate unrecognized psychosis and that the vast majority of individuals engaged in mental health treatment prior to their psychotic disorder diagnosis, it is possible that many individuals have at least some access to mental health care, but this level might not be sufficient to diagnosis or treat the psychotic symptoms. We were also unable to examine co-occurring substance use and substance use disorders, which are known risk factors for psychotic symptoms.27 Lastly, this study included only insured individuals, thus the findings may not generalize to the uninsured. Most individuals also received some mental health care before their psychotic disorder diagnosis, which suggests that simple definitions of access might not be sufficient, eg, those based on being insured or having mental health care, and the quality of the mental health care or expertise of the clinician could be relevant.28–30
In conclusion, in this cohort of Massachusetts residents with a new psychotic disorder diagnosis, nearly all individuals have had at least some engagement with the healthcare system, with almost 90% having received care for a mental health condition, and nearly half having received an antipsychotic medication in the 48 months preceding their psychotic disorder diagnosis. The majority of these patients also received medical care for injuries or trauma suggesting that at least some of these patients could have experienced psychotic symptoms without receiving a formal diagnosis. Given these findings, there appear to be informative patterns of psychiatric and non-psychiatric care preceding the initial psychotic disorder diagnosis that could help identify patients at risk for having psychotic symptoms. Future predictive models will need to differentiate between pre-disease signals (eg, trauma or injury related events) and delays in receiving a formal diagnosis, within the context of our current level of clinical knowledge, to reduce duration of untreated psychosis and promote faster access to specialized first-episode care.
Supplementary Material
Contributor Information
Nicole M Benson, Psychotic Disorders Division, McLean Hospital, Belmont, MA, USA; Mongan Institute, Massachusetts General Hospital, Boston, MA, USA; Department of Psychiatry, Harvard Medical School, Boston, MA, USA.
Zhiyou Yang, Mongan Institute, Massachusetts General Hospital, Boston, MA, USA.
Vicki Fung, Mongan Institute, Massachusetts General Hospital, Boston, MA, USA; Department of Medicine, Harvard Medical School, Boston, MA, USA.
Sharon-Lise Normand, Department of Health Care Policy, Harvard Medical School, Boston, MA, USA; Department of Biostatistics, Harvard TH Chan School of Public Health, Boston, MA, USA.
Matcheri S Keshavan, Department of Psychiatry, Harvard Medical School, Boston, MA, USA; Department of Psychiatry, Beth Israel Deaconess Medical Center, Boston, MA, USA.
Dost Öngür, Psychotic Disorders Division, McLean Hospital, Belmont, MA, USA; Department of Psychiatry, Harvard Medical School, Boston, MA, USA.
John Hsu, Mongan Institute, Massachusetts General Hospital, Boston, MA, USA; Department of Medicine, Harvard Medical School, Boston, MA, USA; Department of Health Care Policy, Harvard Medical School, Boston, MA, USA.
Conflict of Interest
Dr. Benson volunteers for the Epic Systems Behavioral Health Subspecialty Steering Board. Dr. Hsu reports consulting with CHA, USC, Columbia University, Community Servings, Delta Health Alliance, Alta Med. Dr. Fung consults for Headspace, Inc. Dr. Öngür received honoraria from Neumora Inc. and Guggenheim LLC for scientific presentations. Drs. Yang, Keshavan, and Normand have no financial relationships with commercial interests.
Funding
This work was supported by the National Institutes of Health (P50MH115846 and R01MD010456).
References
- 1. Schoenbaum M, Sutherland JM, Chappel A, et al. Twelve-month health care use and mortality in commercially insured young people with incident psychosis in the United States. Schizophr Bull. 2017;43(6):1262–1272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Kane JM, Robinson DG, Schooler NR, et al. Comprehensive versus usual community care for first-episode psychosis: 2-year outcomes from the NIMH RAISE Early Treatment Program. Am J Psychiatry. 2016;173(4):362–372. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. National Institute of Mental Health. Fact Sheet: First Episode Psychosis. https://www.nimh.nih.gov/health/topics/schizophrenia/raise/fact-sheet-first-episode-psychosis.shtml. Updated August 2015. Accessed November 3, 2020.
- 4. George P, Ghose SS, Goldman HH, et al. Growth of coordinated specialty care in the United States With changes in federal funding policies: 2014–2018. Psychiatr Serv. 2022;73(12):1346–1351. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Cascio MT, Cella M, Preti A, Meneghelli A, Cocchi A.. Gender and duration of untreated psychosis: a systematic review and meta-analysis. Early Interv Psychiatry. 2012;6(2):115–127. [DOI] [PubMed] [Google Scholar]
- 6. Penttila M, Jaaskelainen E, Hirvonen N, Isohanni M, Miettunen J.. Duration of untreated psychosis as predictor of long-term outcome in schizophrenia: systematic review and meta-analysis. Br J Psychiatry. 2014;205(2):88–94. [DOI] [PubMed] [Google Scholar]
- 7. Levi L, Bar Haim M, Burshtein S, et al. Duration of untreated psychosis and response to treatment: an analysis of response in the OPTiMiSE cohort. Eur Neuropsychopharmacol. 2020;32:131–135. [DOI] [PubMed] [Google Scholar]
- 8. Fridgen GJ, Aston J, Gschwandtner U, et al. Help-seeking and pathways to care in the early stages of psychosis. Soc Psychiatry Psychiatr Epidemiol. 2013;48(7):1033–1043. [DOI] [PubMed] [Google Scholar]
- 9. Turner M, Smith-Hamel C, Mulder R.. Pathways to care in a New Zealand first-episode of psychosis cohort. Aust N Z J Psychiatry. 2006;40(5):421–428. [DOI] [PubMed] [Google Scholar]
- 10. Chen Y, Farooq S, Edwards J, et al. Patterns of symptoms before a diagnosis of first episode psychosis: a latent class analysis of UK primary care electronic health records. BMC Med. 2019;17(1):227. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Norgaard H, Sondergaard Pedersen H, Fenger-Gron M, et al. Increased use of primary care during 6 years of prodromal schizophrenia. Acta Psychiatr Scand. 2016;134(3):225–233. [DOI] [PubMed] [Google Scholar]
- 12. Schoer N, Rodrigues R, Reid J, et al. Patterns of primary care use prior to a first diagnosis of nonaffective psychotic disorder in Ontario, Canada: Modeles d’utilisation des soins de premiere ligne avant un premier diagnostic de trouble psychotique non affectif en Ontario, Canada. Can J Psychiatry. 2021;66(4):406–417. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Sullivan SA, Hamilton W, Tilling K, Redaniel T, Moran P, Lewis G.. Association of primary care consultation patterns with early signs and symptoms of psychosis. JAMA Netw Open. 2018;1(7):e185174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Sullivan SA, Kounali D, Morris R, et al. Developing and internally validating a prognostic model (P risk) to improve the prediction of psychosis in a primary care population using electronic health records: The MAPPED study. Schizophr Res. 2022;246:241–249. [DOI] [PubMed] [Google Scholar]
- 15. Anderson KK, Fuhrer R, Wynant W, Abrahamowicz M, Buckeridge DL, Malla A.. Patterns of health services use prior to a first diagnosis of psychosis: the importance of primary care. Soc Psychiatry Psychiatr Epidemiol. 2013;48(9):1389–1398. [DOI] [PubMed] [Google Scholar]
- 16. Rietdijk J, Hogerzeil SJ, van Hemert AM, Cuijpers P, Linszen DH, van der Gaag M.. Pathways to psychosis: help-seeking behavior in the prodromal phase. Schizophr Res. 2011;132(2–3):213–219. [DOI] [PubMed] [Google Scholar]
- 17. Simon GE, Stewart C, Hunkeler EM, et al. Care pathways before first diagnosis of a psychotic disorder in adolescents and young adults. Am J Psychiatry. 2018;175(5):434–442. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Wallace A, Isenberg K, York W, et al. Detecting schizophrenia early: prediagnosis healthcare utilization characteristics of patients with schizophrenia may aid early detection. Schizophr Res. 2020;215:392–398. [DOI] [PubMed] [Google Scholar]
- 19. Simon GE, Coleman KJ, Yarborough BJH, et al. First presentation with psychotic symptoms in a population-based sample. Psychiatr Serv. 2017;68(5):456–461. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Mennigen E, Bearden CE.. Psychosis risk and development: what do we know from population-based studies? Biol Psychiatry. 2020;88(4):315–325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Bolhuis K, Lang U, Gyllenberg D, et al. Hospital presentation for self-harm in youth as a risk marker for later psychotic and bipolar disorders: a cohort study of 59 476 Finns. Schizophr Bull. 2021;47(6):1685–1694. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Center for Health Information and Analysis. Overview of the Massachusetts All-Payer Claims Database. https://www.chiamass.gov/assets/docs/p/apcd/APCD-White-Paper-2016.pdf. Updated September 2016. Accessed November 6, 2021.
- 23. Benson NM, Yang Z, Weiss M, et al. Identifying diagnoses of schizophrenia spectrum disorder in large data sets. Psychiatr Serv. 2022;73:1210–1216. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Christian R, Saavedra L, Gaynes BN, et al. Future Research Needs for First- and Second-Generation Antipsychotics for Children and Young Adults [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2012Feb. (Future Research Needs Papers, No. 13.) Appendix A, Tables of FDA-Approved Indications for First- and Second-Generation Antipsychotics. Available from:https://www.ncbi.nlm.nih.gov/books/NBK84656/?report=classic.Accessed August 9, 2023. [PubMed] [Google Scholar]
- 25. Rohde C, Hojlund M, Gasse C, et al. Psychopharmacological treatment patterns prior to a schizophrenia diagnosis: a Danish nationwide study. Schizophr Res. 2022;246:268–276. [DOI] [PubMed] [Google Scholar]
- 26. Stralin P, Hetta J.. First episode psychosis: register-based study of comorbid psychiatric disorders and medications before and after. Eur Arch Psychiatry Clin Neurosci. 2021;271(2):303–313. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Di Forti M, Quattrone D, Freeman TP, et al. ; EU-GEI WP2 Group. The contribution of cannabis use to variation in the incidence of psychotic disorder across Europe (EU-GEI): a multicentre case-control study. Lancet Psychiatry. 2019;6(5):427–436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Rabinowitz J, Bromet EJ, Lavelle J, Severance KJ, Zariello SL, Rosen B.. Relationship between type of insurance and care during the early course of psychosis. Am J Psychiatry. 1998;155(10):1392–1397. [DOI] [PubMed] [Google Scholar]
- 29. KFF. State Health Care Snapshots. Massachusetts. https://www.kff.org/statedata/election-state-fact-sheets/massachusetts/. Updated 2023. Accessed March 2, 2023. [Google Scholar]
- 30. Compton MT, Ramsay CE, Shim RS, et al. Health services determinants of the duration of untreated psychosis among African-American first-episode patients. Psychiatr Serv. 2009;60(11):1489–1494. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.