The Impact of Once-Monthly Paliperidone Palmitate on Healthcare Utilization Among Patients With Schizophrenia Treated in an Integrated Healthcare System: A Retrospective Mirror-Image Study

This retrospective cohort study examined the HRU during 12 months before and after the initiation of PP1M in patients with schizophrenia receiving care at an integrated healthcare system in the USA. In terms of characteristics of the study cohort, similar to previous reports [36, 37], the majority of the patients initiated on PP1M were male and non-Hispanic black, possibly because of the higher risk of antipsychotic non-adherence in these patients [13, 38]. The results showed that initiation of PP1M was associated with a reduction in acute HRU (inpatient and ER use) in the overall study cohort as well as the subset of patients with prior relapse. The reduction in acute HRU after PP1M initiation is likely to be due to better control of schizophrenia symptoms associated with improvement in medication adherence compared to previous treatment with oral antipsychotics. In addition to reduction in schizophrenia-related psychotic events requiring an ER visit and/or hospitalization, better control of schizophrenia symptoms also likely positively impacts self-management behaviors with regard to physical and mental comorbidities common in these patients, thereby leading to improvement in overall physical and mental health and, in turn, reduction in mental health-related and all-cause acute healthcare use [39]. Our findings are consistent with previous similar studies on this topic [29‐31]. For example, in a previous study of Veteran Affairs patients transitioning from orally administered paliperidone/risperidone to PP1M, Patel et al. found that the mean number of visits reduced for all-cause (2.3 to 1.5, p < 0.05), mental health-related (1.5 to 0.8, p < 0.05), and schizophrenia-related inpatient visits (0.6 to 0.3, p < 0.05) from 12 months before to 12 months after transitioning to PP1M. The mean length of stay also reduced for all-cause (28.1 to 14.0 days, p < 0.05), mental health-related (27.1 to 13.8 days, p < 0.05), and schizophrenia-related (13.2 to 5.7 days, p < 0.05) inpatient visits [30]. Other studies have compared HRU between patients on PP1M and those on oral antipsychotics and have reported reduced acute HRU in patients on PP1M compared to those on oral antipsychotics [23‐28]. Manjelievskaia et al. studied multistate Medicaid beneficiaries with schizophrenia and found lower proportions of patients with all-cause inpatient visits (25.6% vs. 33.9%, p < 0.001) and ER visits (54.7% vs. 65.8%, p < 0.001) during the 12 months after initiation of PP1M compared to patients initiated on oral antipsychotics [27].

As expected and consistent with the previous studies [24, 30], the mean number of outpatient visits increased significantly from pre to post PP1M initiation. PP1M treatment necessitates a physician outpatient visit each month as opposed to oral antipsychotic treatment in which routine visits are usually spaced every 3–6 months. The increase in the frequency of outpatient visits along with PP1M administration costs adds to the routine schizophrenia treatment costs in patients with PP1M. However, it has been reported that the increase in routine schizophrenia treatment costs in patients on PP1M is offset by the reduction in total costs due to reduced acute care utilization [23, 29, 31]. Taken together, our findings combined with those from other similar studies indicate that clinical and economic benefits are associated with PP1M.

The extent of reduction in the rates of acute HRU after PP1M initiation was higher in the subgroup of patients with prior schizophrenia relapse during the baseline period compared to the overall study cohort. For example, the proportion of patients with schizophrenia-related 30-day readmissions reduced nearly 89% from the pre- to post-index period, from 11.5% to 1.3%, in patients with prior relapse compared to about an 83% reduction from 8.6% to 1.4% in the overall cohort. The mean number of mental health-related ER visits decreased 60% from the pre- to post-index period, from 2.0 to 0.8, in patients with prior relapse compared to an approximately 53% reduction from 1.5 to 0.7 in the overall cohort. While the direct comparison of the findings between patients with and without a relapse was beyond the scope of this study, the higher relative reduction in acute HRU after PP1M initiation in patients with a relapse during the baseline period suggests that the use of PP1M could be particularly beneficial in patients with a recent relapse, who are likely to be highly non-adherent to their oral antipsychotic regimen.

Despite the likely clinical and economic benefits, the use of PP1M and other LAIs remains low in patients with schizophrenia with prescription rates less than 20% [40]. In a multisite nationwide observational study conducted in the USA, only about 12% of the patients who were non-adherent on oral antipsychotics were prescribed an LAI [41]. Patient factors such as feeling of coerciveness and/or stigma, high cost, and inconvenience due to the need to travel to the clinician’s office; physician factors such as limited knowledge and experience with LAIs, personal bias against the use of needles, and increase in workload associated with the start of a new treatment; and health system factors such as requirement of prior authorization from payors and the need for large amount of resources (budget, storage, and staff) have been reported to be the barriers to LAI use [42‐45]. Some of the proposed solutions to overcome these barriers include introduction of LAIs early in the treatment course, psychological interventions to address the fear of needles, shared decision-making approach with provision of accurate and updated information to the patients, provision of better education regarding LAIs during training and residency, and easier access from the insurance companies [43‐45]. These approaches should be implemented in routine clinical practice to increase the use of PP1M and other LAIs in appropriate patients.

There are a few limitations in this study. Data entry errors are possible in the EMRs and hence there could be inaccuracies. Prior use of oral antipsychotics was determined on the basis of prescription orders written by the physicians, not on the actual use of the medication. Also, while the inclusion criterion of one healthcare visit within the system every 6 months during the study period maximized the likelihood that patients were not lost to follow-up, it is possible that some patients completed some but not all visits within the system. Mirror-image studies are prone to expectation bias, which occurs as a result of patients’/providers’ expectations of a certain outcome when a new treatment is started [46]. Comparative analyses involving a control group of patients on oral antipsychotics were not conducted because of the difficulty in identifying a control group comparable in characteristics. Prior analyses conducted by some of the authors of the study suggested that patients initiated on PP1M usually have more severe schizophrenia and are usually less adherent to their oral antipsychotic regimen prior to PP1M initiation compared to patients who stay on oral antipsychotics. Also, imbalance between the groups is possible because of differences in factors such as family and social support, neighborhood disadvantage, employment, and lifestyle behaviors including exercise, diet, smoking, and alcohol use, information regarding which is not available in the EMRs. Therefore, a pre–post one-group study design, in which patients act as their own controls, was deemed appropriate for this study. A limitation of this study design is that it does not account for variation in patient characteristics over time. However, it could be expected that major changes in patient characteristics would not have occurred during the observation period of 24 months in this study. The impact of antipsychotic polypharmacy prior to and during PP1M use on the study outcomes was not examined as part of this study. Data on important outcomes in patients with schizophrenia such as symptom severity, cognition, quality of life, and treatment-emergent adverse events were not available. Finally, the findings of this study were from one healthcare system in the southeastern USA, primarily North Carolina, and therefore the findings may not be generalizable to other settings.