Outpatient Follow-Up Care and Risk of Hospital Readmission in Schizophrenia and Bipolar Disorder
Abstract
Objective:
Although outpatient care within 30 days of mental health hospital discharge is an established quality indicator, little is known about the clinical implications of not receiving such care. This study evaluated whether receipt of outpatient care within 30 days of discharge was associated with a reduced risk of readmission during days 31–120 postdischarge among adult inpatients with schizophrenia or bipolar disorder.
Methods:
Retrospective longitudinal cohort analyses were performed with Truven MarketScan Commercial (2010–2014) and Medicaid (2010–2013) databases. Among inpatients with schizophrenia (N=25,401) or bipolar disorder (N=46,375), overall and stratified associations were examined between receipt of an outpatient follow-up visit in the 30 days postdischarge and hospital readmission during days 31–120 postdischarge.
Results:
Receipt of a follow-up visit within 30 days of discharge was associated with a slightly lower adjusted odds ratio (AOR) of hospital readmission during days 31–120 postdischarge (schizophrenia, AOR=.88, 95% confidence interval [CI]=.81–.96; bipolar disorder, AOR=.91, CI=.85–.98). For patients with schizophrenia, the strongest observed inverse association of follow-up care with readmission risk was among inpatients whose index admissions were 13 to 30 days long (AOR=.73, CI=.61–.89). For patients with bipolar disorder, the strongest corresponding inverse association was among those in the manic phase of illness at the index discharge (AOR=.73, CI=.63–.86).
Conclusions:
Outpatient visits during the 30 days after discharge were associated with a lower hospital readmission risk during the following 90 days. Assertive hospital discharge planning to secure outpatient visits after hospital discharge is needed for these patient populations.
Receipt of outpatient follow-up care after hospitalization for a mental disorder is a recognized mental health service quality indicator (1,2). The National Committee on Quality Assurance (NCQA) has endorsed quality indicators for follow-up 30 and seven days after hospitalization for mental illness. Yet, rather than resting on empirical research, the rationale for these quality indicators rests primarily on the clinically intuitive concept that individuals whose mental disorders are sufficiently severe to warrant inpatient treatment benefit from timely follow-up with outpatient mental health care. Surprisingly few studies have examined the clinical benefits of continuing outpatient mental health care after hospital discharge (3–6).
Hospital readmission has been used as a measure of the effectiveness of mental health services. An analysis of administrative data of privately insured psychiatric inpatients showed that patients with follow-up outpatient visits within a year of hospital discharge were readmitted to the hospital approximately half as often as those without follow-up visits (5). However, because this analysis did not control for potentially confounding patient characteristics, the difference in readmission risk may have reflected a tendency for higher-risk patients not to receive outpatient treatment after hospital discharge. A Department of Veterans Affairs analysis of psychiatric inpatients with co-occurring substance use disorders that adjusted for several potentially confounding patient characteristics found that continuing psychiatric treatment within 30 days of hospital discharge was not related to risk of hospital readmission in the 90 days postdischarge (6). A controlled study of privately insured patients who were assigned to receive usual, enhanced, or intensive-case-management discharge planning after hospital discharge also reported no significant group differences in hospital readmission at 60 and 180 days following discharge (7).
In light of the paucity of research supporting an association between follow-up outpatient care and subsequent hospital readmission, we examined this issue in large populations of privately insured and Medicaid-insured psychiatric inpatients. To reduce the clinical heterogeneity of the study cohorts, we limited the study to adult inpatients with first-listed diagnoses of either schizophrenia or bipolar disorder, and we performed separate analyses of each diagnostic group.
To better understand risk factors for not receiving outpatient care within 30 days of hospital discharge, we also examined proportions of patients who did and did not receive such follow-up care stratified by demographic and clinical characteristics. Prior research indicates that about half of Medicaid inpatients with psychiatric disorders receive outpatient mental health care during the 30 days after hospital discharge (8). Although several patient characteristics have been associated with not receiving outpatient treatment after psychiatric hospitalization (7,9–11), little is known about whether risk factors for not receiving follow-up care vary across major psychiatric diagnostic groups (12).
Methods
Data Sources
The analyses used Truven MarketScan Commercial (2010–2014) and Medicaid (2010–2013) databases. The project was exempted from human subjects review by the New York State Psychiatric Institute’s Institutional Review Board.
Sample Selection
For the analysis of mental health outpatient visits within 30 days of hospital discharge, which was based on an NCQA HEDIS measure, patients were 18 to 64 years of age, had inpatient admissions of one to 30 days, had a primary discharge diagnosis of schizophrenia (ICD-9-CM code 295) or bipolar disorder (codes 296.0–296.1 and 296.4–296.8), were discharged home, and were enrolled in a health plan during the 120 days preceding admission and the 30 days following hospital discharge. For the analysis of hospital readmission in the 31–120 days following hospital discharge, patients were enrolled for 120 days after the index hospital discharge. Patients readmitted to the hospital within 30 days after the index hospital discharge were removed from the analysis. Patients entered the study cohort on the earliest date of meeting study eligibility criteria, and no patient contributed more than one episode. Separate analyses were performed for the schizophrenia and bipolar cohorts.
Independent Variables
Patient characteristics included age, sex, and payment source (commercial insurance or Medicaid). Information on race-ethnicity (black, white, Hispanic, and other) was available for Medicaid-insured patients. As in prior research (9), length of inpatient stay was partitioned by days into short (one to eight days), medium (nine to 12 days), and long (13 to 30 days). During the 120-day period before the index hospital admission, patients were classified by presence or absence of claims for a substance use disorder (ICD-9-CM codes 291–292 and 303–305). Also during this period, patients were classified by whether they had an inpatient or an emergency department visit with a first-listed mental disorder diagnosis and a visit to a mental health provider. On the basis of hospital discharge diagnoses, patients with schizophrenia were subclassified as having schizoaffective disorder (ICD-9-CM code 295.7) or other schizophrenic disorders (codes 295.0–295.6 and 295.8–295.9), and those with bipolar disorder were subclassified as depressed (codes 296.5 and 296.50–296.56), manic (codes 296.01–296.06, 296.4, and 296.40–46), mixed (codes 296.6 and 296.61–296.66), or other (codes 296.7, 296.80, and 296.89).
Other Covariates
Other covariates in the logistic regressions included a prescription for an antipsychotic, antidepressant, anxiolytic, mood stabilizer, or stimulant medication during the 120 days prior to hospital admission. Patients were also assigned Elixhauser comorbidity scores based on the number of treated comorbid general medical conditions in the 120-day preadmission period (13), excluding depression, psychosis, and substance use disorders. Patients were further classified as having a claim for depression (ICD-9-CM codes 296.2, 296.3, 296.9, 300.4, and 311) or an anxiety disorder (codes 300.0, 300.2, 300.3, 304, and 308.3) during the 120 days prior to hospital admission.
Dependent Variables
In the analysis of postdischarge follow-up, visits were defined based on HEDIS specifications for follow-up mental health visits during the first 30 days after the index hospital discharge (2). For the psychiatric hospital readmission analyses, readmissions were defined as inpatient episodes with first-listed diagnoses of mental disorders (ICD-9-CM codes 290–319) during the 90 days (days 31–120 after hospital discharge) after the first 30 days following the index hospital discharge. In a post hoc analysis of follow-up visits after the first seven days following hospital discharge, mental health readmission was considered during the eight to 97 days following hospital discharge.
Statistical Analysis
Demographic and health care service use characteristics of patients with and without mental health follow-up outpatient visits within 30 days were compared by using Wald chi-square tests. Rates of mental health follow-up outpatient visits within 30 days were calculated overall and stratified by each independent variable. Wald chi-square tests were used to test for differences in the rates across strata. Logistic regression models were used to examine associations between patient characteristics and receipt of a mental health follow-up outpatient visit within 30 days, controlling for sex, age, payer type, length of index admission, diagnostic subtype, comorbid mental disorders, use of mental health services, psychotropic medications, and Elixhauser score.
The proportion of patients who had a readmission in the 31- to 120-day period postdischarge was compared between groups who did and did not have a mental health follow-up outpatient visit within 30 days of discharge. Wald chi-square tests were conducted for the overall sample and stratified by patient characteristics. A series of logistic regression models was used to calculate associations between receipt of a 30-day mental health follow-up outpatient visit and hospital readmission overall and within each stratum, controlling for demographic and service use factors. We also conducted the 30-day outpatient follow-up analysis and 31- to 120-day readmission analysis by race-ethnicity in the Medicaid subsample. To assess the generalizability of the results, we performed corresponding analyses of the association between follow-up visits within seven days of discharge and hospital readmission within eight to 97 days of discharge.
Results
Outpatient Mental Health Follow-Up Visits: Patient Characteristics
Compared with patients with schizophrenia who did not receive a mental health follow-up outpatient visit within 30 days of discharge (N=9,157), those who received such a visit (N=16,244) were significantly more likely to be female, to have a longer inpatient stay, to have schizoaffective disorder subtype, and to have recently made a visit to a mental health care professional and were significantly less likely to have a substance use disorder (Table 1). For the bipolar cohort, those who received a mental health follow-up outpatient visit (N=32,329) were significantly more likely than those who did not (N=14,046) to be female, to be older, to be commercially insured, to have longer inpatient stays, to have bipolar disorder depressed subtype, and to have recently received mental health services and were significantly less likely to have a substance use disorder (Table 1).
| Schizophrenia | Bipolar disorder | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| 30-day follow-up visit (N=16,244) | No 30-day follow-up visit (N=9,157) | 30-day follow-up visit (N=32,329) | No 30-day follow-up visit (N=14,046) | |||||||
| Characteristic | N | % | N | % | pb | N | % | N | % | pb |
| Gender | <.001 | <.001 | ||||||||
| Female | 7,270 | 44.8 | 3,683 | 40.2 | 20,736 | 64.1 | 8,178 | 58.2 | ||
| Male | 8,974 | 55.2 | 5,474 | 59.8 | 11,593 | 35.9 | 5,868 | 41.8 | ||
| Age | .157 | <.001 | ||||||||
| 18–24 | 3,036 | 18.7 | 1,620 | 17.7 | 7,167 | 22.2 | 3,654 | 26.0 | ||
| 25–35 | 4,058 | 25.0 | 2,372 | 25.9 | 7,205 | 22.3 | 3,297 | 23.5 | ||
| 36–45 | 3,168 | 19.5 | 1,798 | 19.6 | 7,057 | 21.8 | 2,879 | 20.5 | ||
| 46–64 | 5,982 | 36.8 | 336 | 36.8 | 10,900 | 33.7 | 4,219 | 30.0 | ||
| Payment source | .072 | <.001 | ||||||||
| Commercial insurance | 5,988 | 36.9 | 3,272 | 35.7 | 23,919 | 74.0 | 8,947 | 63.7 | ||
| Medicaid | 10,256 | 63.1 | 5,885 | 64.3 | 8,410 | 26.0 | 5,099 | 36.3 | ||
| Length of index admission (days) | <.001 | <.001 | ||||||||
| Short (1–8) | 9,502 | 58.5 | 5,591 | 61.1 | 23,137 | 71.6 | 10,535 | 75.0 | ||
| Medium (9–12) | 3,365 | 20.7 | 1,887 | 20.6 | 5,636 | 17.4 | 2,139 | 15.2 | ||
| Long (13–30) | 3,377 | 20.8 | 1,679 | 18.3 | 3,556 | 11.0 | 1,372 | 9.8 | ||
| Subtype | ||||||||||
| Schizophrenia | <.001 | |||||||||
| Schizoaffective | 8,090 | 49.8 | 4,045 | 44.2 | ||||||
| Other | 8,154 | 50.2 | 5,112 | 55.8 | ||||||
| Bipolar disorder | <.001 | |||||||||
| Depressed | 14,565 | 45.1 | 6,223 | 44.3 | ||||||
| Manic | 7,672 | 23.7 | 3,293 | 23.4 | ||||||
| Mixed | 8,726 | 27.0 | 3,786 | 27.0 | ||||||
| Other | 1,366 | 4.2 | 744 | 5.3 | ||||||
| Substance use disorder in past 120 days | 4,265 | 26.3 | 2,686 | 29.3 | <.001 | 8,377 | 25.9 | 4,556 | 32.4 | <.001 |
| Use of mental health services in past 120 days | ||||||||||
| Inpatient | 1,192 | 7.3 | 629 | 6.9 | .164 | 3,084 | 9.5 | 1,215 | 8.7 | .002 |
| Emergency department | 4,459 | 27.5 | 2,417 | 26.4 | .068 | 7,336 | 22.7 | 2,985 | 21.3 | <.001 |
| Mental health professional | 14,023 | 86.3 | 5,557 | 60.7 | <.001 | 26,664 | 82.5 | 8,454 | 60.2 | <.001 |
| No use of mental health services in past 120 daysc | 1,094 | 6.7 | 1,812 | 19.8 | <.001 | 2,240 | 6.9 | 2,269 | 16.2 | <.001 |
a
Characteristics were ascertained at index admission.
b
Derived from chi-square tests
c
Mental health services include those for schizophrenia or bipolar disorder, services for comorbid mental disorders, and psychotropic medications.
Rates of Mental Health Follow-Up Outpatient Visits
Approximately two-thirds (64.0%) of patients with schizophrenia received 30-day mental health follow-up outpatient visits (Table 2). Among patients with schizophrenia, the adjusted odds of receiving a follow-up visit were associated with being female, younger, and commercially insured and having a longer inpatient stay. The likelihood of receiving a 30-day follow-up visit was strongly associated with having visited an outpatient mental health provider during the 120-day period prior to admission. Roughly one-third (37.6%) of patients with schizophrenia with no recent use of mental health care received follow-up outpatient visits. Having a comorbid substance use disorder was also associated with a lower likelihood of receiving a 30-day follow-up visit among patients with schizophrenia.
| Schizophrenia | Bipolar disorder | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Variable | N | % | pb | AOR | 95% CI | N | % | pb | AOR | 95% CI |
| Total | 16,244 | 64.0 | 32,329 | 69.7 | ||||||
| Sex | <.001 | <.001 | ||||||||
| Female (reference) | 7,270 | 66.4 | 1.00 | 20,736 | 71.7 | 1.00 | ||||
| Male | 8,974 | 62.1 | .90 | .85–.96 | 11,593 | 66.4 | .87 | .93–.90 | ||
| Age | .157 | <.001 | ||||||||
| 18–24 (reference) | 3,036 | 65.2 | 1.00 | 7,176 | 66.2 | 1.00 | ||||
| 25–35 | 4,058 | 63.1 | .87 | .80–.95 | 7,205 | 68.6 | 1.12 | 1.06–1.19 | ||
| 36–45 | 3,168 | 63.8 | .85 | .77–.93 | 7,057 | 71.0 | 1.17 | 1.10–1.24 | ||
| 46–64 | 5,982 | 64.0 | .81 | .74–.88 | 10,900 | 72.1 | 1.17 | 1.10–1.24 | ||
| Payment source | .072 | <.001 | ||||||||
| Commercial insurance (reference) | 5,988 | 64.7 | 1.00 | 23,919 | 72.8 | 1.00 | ||||
| Medicaid | 10,256 | 63.5 | .85 | .80–.90 | 8,410 | 62.3 | .58 | .56–.61 | ||
| Length of index admission (days) | <.001 | <.001 | ||||||||
| Short (1–8) (reference) | 9,502 | 63.0 | 1.00 | 23,137 | 68.7 | 1.00 | ||||
| Medium (9–12) | 3,365 | 64.1 | 1.10 | 1.02–1.18 | 5,636 | 72.5 | 1.21 | 1.14–1.28 | ||
| Long (13–30) | 3,377 | 66.8 | 1.22 | 1.13–1.31 | 3,556 | 72.2 | 1.15 | 1.07–1.23 | ||
| Subtype schizophrenia | <.001 | |||||||||
| Schizoaffective | 8,090 | 66.7 | .91 | .86–.96 | ||||||
| Other (reference) | 8,154 | 61.5 | 1.00 | |||||||
| Subtype bipolar disorder | <.002 | |||||||||
| Depressed (reference) | 14,565 | 70.1 | 1.00 | |||||||
| Manic | 7,672 | 70.0 | 1.07 | 1.02–1.14 | ||||||
| Mixed | 8,726 | 69.7 | 1.01 | .96–1.06 | ||||||
| Other | 1,366 | 64.7 | .86 | .77–.94 | ||||||
| Substance use disorder in past 120 days (reference: no) | 4,265 | 61.4 | <.001 | .69 | .65–.74 | 23,952 | 64.8 | <.001 | .66 | .63–.69 |
| Mental health services in past 120 days (reference: no) | ||||||||||
| Inpatient | 1,192 | 65.5 | .164 | .78 | .70–.87 | 3,084 | 71.7 | .002 | .92 | .86–1.00 |
| Emergency department | 4,459 | 64.8 | .069 | .90 | .84–.96 | 7,336 | 71.1 | <.001 | .98 | .92–1.03 |
| Mental health professional | 14,023 | 71.6 | <.001 | 3.32 | 3.04–3.62 | 26,664 | 75.9 | <.001 | 2.68 | 2.53–2.85 |
| No use of mental health services in past 120 daysc | 1,094 | 37.6 | <.001 | 2,240 | 49.7 | <.001 | ||||
a
The models controlled for sex, age, payment source, length of index admission, subtype, comorbid mental disorders, use of mental health services (except for no mental health care), psychotropic medications, and Elixhauser score. Adjusted odds ratios (AORs) represent the likelihood of having a follow-up visit within 30 days of hospital discharge in relation to the reference group, by characteristic.
b
Derived from chi-square tests comparing the rate of follow-up visits among patients with and without each characteristic.
c
Because this variable was constructed from other covariates in the models, a logistic regression limited to patients with no use of mental health services in the past 120 days was not fit. Mental health services include those for schizophrenia or bipolar disorder, services for comorbid mental disorders, and psychotropic medications.
Over two-thirds (69.7%) of patients with bipolar disorder received a 30-day mental health follow-up outpatient visit postdischarge. The adjusted odds of receiving a follow-up visit were associated with being female, older, and commercially insured and having a longer inpatient stay. The odds of receiving a follow-up visit were increased by receiving care from an outpatient mental health provider during the 120-day period prior to admission. In the bipolar cohort, having a comorbid substance use disorder was also associated with a lower likelihood of receiving a 30-day follow-up visit (Table 2).
Among patients enrolled in Medicaid, those with either schizophrenia or bipolar disorder who were white were more likely than their counterparts who were black to receive a 30-day follow-up visit (Table 3).
| 30-day follow-up visit | Readmission within 31–120 days | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Yes | No | 30-day follow-up visit | No 30-day follow-up visit | |||||||||
| Group | N | % | N | % | AORb | 95% CI | N | % | N | % | AORc | 95% CI |
| Schizophrenia | ||||||||||||
| White | 5,022 | 68.4 | 2,323 | 31.6 | 1.28 | 1.19–1.39 | 669 | 14.2 | 263 | 12.9 | .99 | .84–1.17 |
| Black | 4,235 | 59.3 | 2,902 | 40.7 | 1.00 | 501 | 12.6 | 320 | 12.4 | .92 | .78–1.08 | |
| Hispanic | 151 | 67.1 | 74 | 32.9 | 1.20 | .89–1.62 | 28 | 19.3 | 11 | 16.4 | 1.40 | .54–3.61 |
| Other | 848 | 59.1 | 586 | 40.9 | .93 | .82–1.05 | 108 | 13.9 | 74 | 15.1 | .87 | .61–1.24 |
| Bipolar | ||||||||||||
| White | 6,340 | 64.6 | 3,472 | 35.2 | 1.33 | 1.20–1.47 | 692 | 11.9 | 315 | 10.8 | 1.06 | .91–1.23 |
| Black | 1,313 | 54.9 | 1,077 | 45.1 | 1.00 | 152 | 12.6 | 112 | 12.0 | 1.03 | .77–1.37 | |
| Hispanic | 93 | 63.3 | 54 | 36.7 | 1.35 | .93–1.95 | 10 | 11.6 | 6 | 14.6 | .07 | .01–.88 |
| Other | 664 | 57.2 | 496 | 42.8 | 1.04 | .98–1.21 | 75 | 12.5 | 38 | 9.8 | 1.29 | .80–2.06 |
a
The models controlled for sex, age, payment source, length of index admission, subtype, comorbid mental disorders, use of mental health services (except for no mental health care), psychotropic medications, and Elixhauser score.
b
The adjusted odds ratios (AORs) represent the likelihood of having a follow-up visit within 30 days of hospital discharge among blacks versus other racial ethnic groups.
c
The AORs represent the odds of being readmitted within 31–120 days for individuals with a 30-day follow-up visit in relation to the odds for individuals without a 30-day follow-up visit.
Mental Health Follow-Up Outpatient Visits and Hospital Readmission
In the schizophrenia cohort, after adjustment for all other characteristics, receipt of a 30-day mental health follow-up outpatient visit was associated with significantly lower odds of readmission during that 90-day period (adjusted odds ratio [AOR]=.88) (Table 4). In multivariable models for patients with schizophrenia, the association between a 30-day follow-up visit and reduced risk of mental health hospital readmission was significant for younger adults (18–24 years), males, commercially insured patients, and patients with “other” schizophrenia subtypes at the index hospital discharge. The adjusted association between 30-day follow-up and reduced risk of readmission was also significant among patients without a comorbid substance use disorder.
| 30-day follow-up visit | No 30-day follow-up visit | ||||||
|---|---|---|---|---|---|---|---|
| Variable | N | % | N | % | p | AOR | 95% CI |
| Total | 1,921 | 13.0 | 1,013 | 13.1 | .831 | .88 | .81–.96 |
| Age | |||||||
| 18–24 | 413 | 15.2 | 220 | 16.4 | .317 | .80 | .67–.97 |
| 25–35 | 510 | 14.0 | 279 | 14.4 | .684 | .88 | .74–1.04 |
| 36–45 | 353 | 12.3 | 198 | 13.0 | .500 | .88 | .72–1.07 |
| 46–64 | 645 | 11.7 | 316 | 10.8 | .239 | .94 | .81–1.10 |
| Sex | |||||||
| Female | 838 | 12.6 | 384 | 12.5 | .797 | .87 | .76–1.00 |
| Male | 1,083 | 13.3 | 629 | 13.6 | .708 | .88 | .79–.98 |
| Payment source | |||||||
| Medicaid | 1,306 | 13.6 | 668 | 12.9 | .245 | .95 | .85–1.06 |
| Commercial insurance | 615 | 11.9 | 345 | 13.5 | .046 | .76 | .65–.88 |
| Subtype, index admission | |||||||
| Schizoaffective | 1,039 | 14.2 | 474 | 14.2 | .960 | .90 | .80–1.02 |
| Other schizophrenia | 882 | 11.9 | 539 | 12.3 | .455 | .86 | .76–.98 |
| Length of index admission (days) | |||||||
| Short (1–8) | 1,134 | 13.1 | 594 | 12.6 | .416 | .94 | .84–1.05 |
| Medium (9–12) | 395 | 13.0 | 214 | 13.4 | .739 | .87 | .72–1.05 |
| Long (13–30) | 392 | 12.9 | 205 | 14.7 | .102 | .73 | .61–.89 |
| Substance use disorder in past 120 days | |||||||
| Present | 601 | 15.6 | 366 | 16.2 | .545 | .87 | .75–1.01 |
| Absent | 1,320 | 12.1 | 647 | 11.8 | .635 | .88 | .80–.98 |
| Use of mental health services in past 120 days | |||||||
| Inpatient | 243 | 23.0 | 110 | 21.6 | .547 | 1.07 | .82–1.40 |
| Emergency department | 606 | 14.8 | 304 | 14.7 | .939 | .93 | .79–1.08 |
| Mental health professional | 1,753 | 13.7 | 674 | 14.4 | .238 | .93 | .84–1.02 |
| No use of mental health services in past 120 daysb | 168 | 8.4 | 339 | 11.1 | .002 | ||
a
Each row represents results from a separate logistic regression model restricted to inpatients within the group stratum. The models controlled for sex, age, payment source, length of index admission, subtype, comorbid mental disorders, use of mental health services (except for no mental health care), psychotropic medications, and Elixhauser score. AOR, adjusted odds ratio The AORs represent the odds of being readmitted within 31–120 days among individuals with a 30-day mental health follow-up visit in relation to individuals without a 30-day follow-up visit (reference group).
b
Because this variable was constructed from other covariates in the models, a logistic regression limited to patients with no use of mental health services in the past 120 days was not fit. Mental health services include those for schizophrenia or bipolar disorder, services for comorbid mental disorders, and psychotropic medications.
In the bipolar cohort, 30-day mental health follow-up outpatient visits were associated with lower adjusted odds of subsequent hospital care over the following 90 days (AOR=.91) (Table 5). In the adjusted analyses, the association between 30-day follow-up and reduced risk of readmission was significant for patients who were older (46–64 years), who were commercially insured, and who had the manic subtype diagnosis at index hospital discharge. The likelihood of readmission was also lower among patients with bipolar disorder who had a substance use disorder during the 120-day period prior to admission and those who had a mental health emergency department visit during the preadmission period.
| 30-day follow-up visit | No 30-day follow-up visit | ||||||
|---|---|---|---|---|---|---|---|
| Variable | N | % | N | % | p | AOR | 95% CI |
| Total | 2,859 | 10.1 | 1,164 | 10.5 | .296 | .91 | .85–.98 |
| Age | |||||||
| 18–24 | 761 | 11.9 | 322 | 10.6 | .068 | 1.01 | .87–1.16 |
| 25–35 | 615 | 10.1 | 243 | 9.9 | .734 | .97 | .83–1.15 |
| 36–45 | 592 | 9.7 | 231 | 10.5 | .289 | .93 | .79–1.11 |
| 46–64 | 891 | 9.2 | 368 | 10.8 | .008 | .77 | .67–.88 |
| Sex | |||||||
| Female | 1,819 | 10.0 | 658 | 10.1 | .785 | .92 | .83–1.01 |
| Male | 1,040 | 10.3 | 506 | 10.9 | .233 | .90 | .80–1.02 |
| Payment source | |||||||
| Medicaid | 929 | 12.1 | 471 | 11.0 | .091 | 1.06 | .94–1.20 |
| Commercial insurance | 1,930 | 9.4 | 693 | 10.1 | .072 | .83 | .76–.92 |
| Subtype, index admission | |||||||
| Depressed | 1,338 | 10.5 | 507 | 10.3 | .638 | .97 | .83–1.08 |
| Manic | 586 | 8.7 | 276 | 10.8 | .003 | .73 | .63–.86 |
| Mixed | 799 | 10.5 | 306 | 10.1 | .607 | 1.01 | .87–1.17 |
| Other | 136 | 11.4 | 75 | 12.6 | .464 | .85 | .61–1.18 |
| Length of admission (days) | |||||||
| Short (1–8) | 1,938 | 9.6 | 837 | 10.0 | .251 | .92 | .84–1.00 |
| Medium (9–12) | 542 | 11.0 | 199 | 11.7 | .451 | .89 | .74–1.06 |
| Long (13–30) | 379 | 12.1 | 128 | 12.0 | .907 | .92 | .73–1.15 |
| Substance use disorder in past 120 days | |||||||
| Present | 882 | 12.1 | 475 | 13.2 | .098 | .88 | .77–.99 |
| Absent | 1,977 | 9.4 | 689 | 9.2 | .499 | .94 | .85–1.04 |
| Use of mental health services in past 120 days | |||||||
| Inpatient | 507 | 19.0 | 201 | 22.0 | .048 | .82 | .67–1.00 |
| Emergency department | 711 | 11.0 | 314 | 13.1 | .005 | .81 | .70–.94 |
| Mental health professional | 2,502 | 10.7 | 786 | 11.8 | .010 | .92 | .84–1.00 |
| No use of mental health services in past 120 daysb | 357 | 7.3 | 378 | 8.4 | .033 | ||
a
Each row represents results from a separate logistic regression model restricted to inpatients within the group stratum. The models controlled for sex, age, payment source, length of index admission, subtype, comorbid mental disorders, use of mental health services (except for no mental health care), psychotropic medications, and Elixhauser score. The AORs represent the odds of being readmitted within 31–120 days among individuals with a 30-day mental health follow-up visit in relation to individuals without a 30-day follow-up visit (reference group).
b
Because this variable was constructed from other covariates in the models, a logistic regression limited to patients with no use of mental health services in the past 120 days was not fit. Mental health services include those for schizophrenia or bipolar disorder, services for comorbid mental disorders, and psychotropic medications.
Among the patients with schizophrenia, 40.7% (N=11,372) received a mental health follow-up outpatient visit within seven days of discharge. There was little difference in the risk of hospital readmission during the eight- to 97-day postdischarge period between patients with schizophrenia who received a seven-day follow-up visit (17.0%, N=1,930) and those who did not (16.4%, N=2,724). Among patients with bipolar disorder, 46.0% (N=23,298) received a seven-day follow-up visit. The proportion who were readmitted to the hospital during the eight- to 97-day posthospital period also did not significantly vary by presence (14.5%, N=3,369) or absence (14.1%, N=3,861) of an outpatient visit within seven days of hospital discharge.
Discussion
After the analyses accounted for several potentially confounding patient characteristics, patients with schizophrenia or bipolar disorder who received outpatient mental health care in the first month after hospital discharge had slightly lower odds of psychiatric hospital readmission during the following three months. These associations support a focus on improving the transition from inpatient to outpatient care (14).
A tendency for patients who received follow-up visits in the 30 days after hospital discharge to have a lower risk of subsequent readmission was prominent among patients with schizophrenia who had not received any mental health care in the 120 days prior to the admission, a group with an especially low rate of outpatient follow-up care. For these patients, follow-up mental health treatment appeared to be important in lowering the likelihood of readmission. Among patients with bipolar disorder, follow-up care was associated with a lower risk of readmission for patients who were older, who were admitted for a manic episode, and who had a mental health emergency department visit in the preadmission period. For these patients, the first postdischarge outpatient mental health visit may help consolidate a process of recovery that was initiated during the inpatient stay. In manic episodes, there is a high risk of treatment nonadherence (15) and symptom exacerbation after treatment discontinuation (16). Clinical instability may increase the importance of continuing care after hospital discharge.
In accord with prior research (6,17), longer inpatient stays were correlated with receipt of mental health follow-up outpatient visits within 30 days of hospital discharge for both diagnostic groups. Longer inpatient stays may permit more time for stabilization of clinical symptoms and may improve discharge planning. These patients may also have more severe conditions that increase the likelihood that they will be seen within 30 days of hospital discharge. Because the length of psychiatric hospital admissions has declined over the past several years (18), clinical staff involved in transition management may have less time to assist patients and their families with discharge planning.
In both diagnostic groups, a significant association between outpatient follow-up visits and short-term risk of hospital readmission occurred among commercially insured patients but not among Medicaid-insured patients. Variations in clinical presentation, patient resources, discharge readiness, or the effectiveness of available outpatient services may contribute to this difference. In other clinical contexts, Medicaid-insured psychiatric patients have been found to receive less intensive care than their commercially insured counterparts (19).
A corresponding significant inverse association was not observed between outpatient follow-up visits within seven days of discharge and subsequent hospital readmission. The reasons for the differences between the seven-day and 30-day analyses are unclear but may be related to differences in factors that govern readmissions during the two following time periods. By design, the analysis of outpatient visits within the first 30 days excluded patients who were readmitted within the first 30 days, whereas the seven-day analysis excluded a narrower group who were readmitted within the first seven days. Readmissions that follow quickly after hospital discharge may be more closely related to clinical readiness for discharge or the quality of discharge planning. In the general medical literature, factors such as living alone (20), low patient satisfaction with the inpatient physician (21), and not having a regular physician (22) have been associated with hospital readmission within 30 days of discharge.
Consistent with previous research (3,11,12), receiving mental health care in the months before hospital admission was the strongest predictor of outpatient mental health treatment continuity following hospital discharge. For patients with schizophrenia or bipolar disorder, established treatment routines may influence future treatment-seeking behavior. Among patients without existing outpatient treatment relationships, difficulties building new treatment relationships (23), long wait times for first appointments (24), stigma, and negative symptoms may complicate completion of outpatient mental health referrals. For high-risk patients, assertive discharge planning, engagement of family members, and shared decision making may facilitate new treatment contacts and coordinate the care transition (25).
Roughly a quarter of the inpatients with schizophrenia and bipolar disorder had received treatment for a substance use disorder during the 120-day period prior to their inpatient admission. After the analyses controlled for several patient characteristics, these patients were about one-third less likely than their counterparts without a comorbid substance use disorder to receive outpatient follow-up care within 30 days. Co-occurring substance use disorders among inpatients with other mental illnesses have previously been linked to lower odds of follow-up after psychiatric hospital discharge (12). In community practice, treatment for comorbid substance use disorders is not always offered, and when offered it is not always accepted. Among patients with schizophrenia or bipolar disorder, the strength of the associations between follow-up visits and risk of hospital readmission did not markedly vary by the presence of a comorbid substance use disorder.
In the Medicaid sample of patients with schizophrenia or bipolar disorder, white patients were more likely than black patients to receive follow-up outpatient care. This pattern is consistent with prior work with diagnostically mixed samples (8) and with a broad tendency for African Americans with severe mental illnesses to receive lower-quality care (26). There is a critical need to develop interventions that not only improve continuity of care but that narrow racial-ethnic disparities (14).
This study had several potential limitations. First, data were not available concerning many factors that may influence continuing care or the benefits of this care. For example, we were not able to measure inpatient staff preparation of patients for their first outpatient visit or availability of outpatient services (27) nor were we able to assess associations of outpatient visits with symptoms. Second, diagnoses were based on clinician judgment. Third, analyses were limited to nonelderly adult Medicaid and commercially insured patients with schizophrenia or bipolar disorder and may not generalize to uninsured populations (28), elderly patients, or those with other diagnoses. Fourth, because this exploratory study was based on a large number of observations, even modest associations of uncertain clinical and policy significance may be statistically significant. Fifth, our primary focus on outpatient follow-up within 30 days of discharge and hospital readmission over the subsequent 90 days excluded the substantial number of patients with hospital readmissions within the first 30 days of hospital discharge. Finally, residual confounding may have accounted for the tendency for patients who received continuing care to have lower odds of psychiatric hospital readmission.
Conclusions
Interruptions in service delivery during transitions from the hospital to outpatient psychiatric care were found to be common among Medicaid and commercially insured patients with schizophrenia and bipolar disorder. Improving transitions from inpatient to outpatient mental health care may help lower short-term risk of subsequent hospital readmissions for these patients. Some relatively simple interventions may improve follow-up care, including providing telephone case management (29), shortening the period between discharge and the first scheduled follow-up appointment (24), and arranging meetings with the outpatient clinician prior to discharge (30). Assertive discharge planning, engagement of family members, and shared decision making may also be necessary to ensure continuity of care following hospital discharge.
References
1.
HEDIS2014: Summary Table of Measures, Product Lines, and Changes. Washington, DC, National Committee for Quality Assurance, 2014. http://www.ncqa.org/Portals/0/HEDISQM/HEDIS2014/List_of_HEDIS_2014_Measures.pdf. Accessed Dec 21, 2015
2.
Inpatient Psychiatric Facility Prospective Payment System: Payment System Fact Sheet Series. Baltimore, Centers for Medicare and Medicaid Services, 2014. http://www.cms.gov/Outreach-and-Education/Medicare-Learning-Network-MLN/MLNProducts/downloads/InpatientPsychFac.pdf
3.
Carson NJ, Vesper A, Chen CN, et al: Quality of follow-up after hospitalization for mental illness among patients from racial-ethnic minority groups. Psychiatric Services 65:888–896, 2014
4.
Olfson M, Marcus SC, Doshi JA: Continuity of care after inpatient discharge of patients with schizophrenia in the Medicaid program: a retrospective longitudinal cohort analysis. Journal of Clinical Psychiatry 71:831–838, 2010
5.
Nelson EA, Maruish ME, Axler JL: Effects of discharge planning and compliance with outpatient appointments on readmission rates. Psychiatric Services 51:885–889, 2000
6.
Ilgen MA, Hu KU, Moos RH, et al: Continuing care after inpatient psychiatric treatment for patients with psychiatric and substance use disorders. Psychiatric Services 59:982–988, 2008
7.
Cuffel BJ, Held M, Goldman W: Predictive models and the effectiveness of strategies for improving outpatient follow-up under managed care. Psychiatric Services 53:1438–1443, 2002
8.
Stein BD, Kogan JN, Sorbero MJ, et al: Predictors of timely follow-up care among Medicaid-enrolled adults after psychiatric hospitalization. Psychiatric Services 58:1563–1569, 2007
9.
Olfson M, Mechanic D, Hansell S, et al: Prediction of homelessness within three months of discharge among inpatients with schizophrenia. Psychiatric Services 50:667–673, 1999
10.
Copeland LA, Zeber JE, Salloum IM, et al: Treatment adherence and illness insight in veterans with bipolar disorder. Journal of Nervous and Mental Disease 196:16–21, 2008
11.
Novick D, Haro JM, Suarez D, et al: Predictors and clinical consequences of non-adherence with antipsychotic medication in the outpatient treatment of schizophrenia. Psychiatry Research 176:109–113, 2010
12.
Lin CH, Chen MC, Chou LS, et al: Time to rehospitalization in patients with major depression vs those with schizophrenia or bipolar I disorder in a public psychiatric hospital. Psychiatry Research 180:74–79, 2010
13.
Elixhauser A, Steiner C, Harris DR, et al: Comorbidity measures for use with administrative data. Medical Care 36:8–27, 1998
14.
Horvitz-Lennon M, Volya R, Donohue JM, et al: Disparities in quality of care among publicly insured adults with schizophrenia in four large US states, 2002–2008. Health Services Research 49:1121–1144, 2014
15.
Lingam R, Scott J: Treatment non-adherence in affective disorders. Acta Psychiatrica Scandinavica 105:164–172, 2002
16.
Suppes T, Baldessarini RJ, Faedda GL, et al: Risk of recurrence following discontinuation of lithium treatment in bipolar disorder. Archives of General Psychiatry 48:1082–1088, 1991
17.
Wang HR, Song HR, Jung YE, et al: Continuity of outpatient treatment after discharge of patients with major depressive disorder. Journal of Nervous and Mental Disease 201:519–524, 2013
18.
Babalola O, Gormez V, Alwan NA, et al: Length of hospitalisation for people with severe mental illness. Cochrane Database of Systematic Reviews 1:CD000384, 2014
19.
Marcus SC, Bridge JA, Olfson M: Payment source and emergency management of deliberate self-harm. American Journal of Public Health 102:1145–1153, 2012
20.
Iloabuchi TC, Mi D, Tu W, et al: Risk factors for early hospital readmission in low-income elderly adults. Journal of the American Geriatrics Society 62:489–494, 2014
21.
Boulding W, Glickman SW, Manary MP, et al: Relationship between patient satisfaction with inpatient care and hospital readmission within 30 days. American Journal of Managed Care 17:41–48, 2011
22.
Hasan O, Meltzer DO, Shaykevich SA, et al: Hospital readmission in general medicine patients: a prediction model. Journal of General Internal Medicine 25:211–219, 2010
23.
Lecomte T, Spidel A, Leclerc C, et al: Predictors and profiles of treatment non-adherence and engagement in services problems in early psychosis. Schizophrenia Research 102:295–302, 2008
24.
Compton MT, Rudisch BE, Craw J, et al: Predictors of missed first appointments at community mental health centers after psychiatric hospitalization. Psychiatric Services 57:531–537, 2006
25.
Vigod SN, Kurdyak PA, Dennis CL, et al: Transitional interventions to reduce early psychiatric readmissions in adults: systematic review. British Journal of Psychiatry 202:187–194, 2013
26.
Horvitz-Lennon M, Zhou D, Normand SLT, et al: Racial and ethnic service use disparities among homeless adults with severe mental illnesses receiving ACT. Psychiatric Services 62:598–604, 2011
27.
Heggestad T: Operating conditions of psychiatric hospitals and early readmission: effects of high patient turnover. Acta Psychiatrica Scandinavica 103:196–202, 2001
28.
Yanos PT, Lu W, Minsky S, et al: Correlates of health insurance among persons with schizophrenia in a statewide behavioral health care system. Psychiatric Services 55:79–82, 2004
29.
Kolbasovsky A, Reich L, Meyerkopf N: Reducing six-month inpatient psychiatric recidivism and costs through case management. Care Management Journals 11:2–10, 2010
30.
Boyer CA, McAlpine DD, Pottick KJ, et al: Identifying risk factors and key strategies in linkage to outpatient psychiatric care. American Journal of Psychiatry 157:1592–1598, 2000
Information & Authors
Information
Published In
Cover: New York City Views. Skaters Near 59th Street, Samuel H. Gottscho. Museum of the City of New York. Photo credit: Museum of the City of New York/Art Resource, New York City.
History
Received: 28 October 2016
Revision received: 18 January 2017
Revision received: 3 April 2017
Revision received: 2 May 2017
Accepted: 12 May 2017
Published online: 3 July 2017
Published in print: December 01, 2017
Keywords
Authors
Funding Information
Sunovion10.13039/100009655: Outpatient follow-up care and risk of rehospitaliz
Dr. Marcus reports receipt of financial support from Alkermes, Allergan, Ortho-McNeil Janssen, and Sunovion. Dr. Chuang is an employee of Vertex Pharmaceuticals. Dr. Ng-Mak is an employee of Sunovion Pharmaceuticals. Dr. Olfson’s work on this study was supported by a grant from Sunovion Pharmaceuticals to Columbia University Medical Center.Metrics & Citations
Metrics
Citations
Export Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
For more information or tips please see 'Downloading to a citation manager' in the Help menu.
View Options
View options
PDF/EPUB
View PDF/EPUBLogin options
Already a subscriber? Access your subscription through your login credentials or your institution for full access to this article.
Personal login Institutional Login Open Athens loginNot a subscriber?
PsychiatryOnline subscription options offer access to the DSM-5-TR® library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.
Need more help? PsychiatryOnline Customer Service may be reached by emailing [email protected] or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).