Recurrently Readmitted Inpatient Psychiatric Patients: Characteristics and Care
William H. Sledge
Christine L. Dunn
Case Vignette
Pam is a 33-year-old mother of five children, divorced, with the diagnoses of mood disorder not otherwise specified (NOS), alcohol dependence, borderline personality disorder, and psychotic disorder NOS. She has had three hospitalizations in 3 months and has a history of cutting and impulsivity and chaotic personal relationships leading to the loss of custody of her children. The plan was to have involved her in a women’s shelter program developing toward independent housing and in Alcoholics Anonymous (AA) and to organize her treatment at the local community mental health center. She was readmitted to the hospital approximately 6 weeks from the index hospital discharge for cutting and suicidal ideation but following discharge was lost to all efforts at follow-up.
“Recidivist,” “frequent flyer,” “revolving door,” and “high user” are some of the not-so complimentary terms that have been used to refer to patients who are recurrently admitted to the hospital. This phenomenon is hard on everyone. There is no indication that patients find it a satisfactory way of engaging the mental health system (despite some older reports that these patients like to be in the hospital).1,2 Patients’ families and loved ones (when they are in the patients’ lives) find it disruptive and alarming, and health care professionals frequently range from being annoyed to being substantially troubled. When employers of the patients are involved, they do not stay involved long. In general, it is not an effective way of getting mental health services to patients who need the services.
This pattern of use has been a preoccupation among health care professionals for some time, and that concern is represented in a robust literature that spans several eras of psychiatry in the last half of the 20th century. Geller,3 one of the most creative and passionate investigators on the recurrent readmission (RR) phenomenon, provided a historical perspective in his account of a state hospital that spanned a 100 years during the last part of the 19th century and the first part of the 20th century. He suggests the RR phenomenon may be a nasty side effect of deinstitutionalization and the fragmented mental health care that resulted in many states.3,4 (See also, Chapter 1.)
This chapter will examine these and other perspectives as the authors review briefly some of the literature addressing the phenomenon of the recurrently admitted (to psychiatric hospitalization) patient. The authors will describe some of their work at the Yale-New Haven Psychiatric Hospital (YNHPH) in attempting to understand and intervene with this problem. The focus here will be on those patients who are recurrently admitted to the hospital. Although these patients may be a subset of those with persistent and severe mental illness, one of the repeated questions of
the literature is to what degree the group of recurrently readmitted inpatient psychiatric patients (RRIPP) is coincident with the broader category of persistently and severely mentally ill and if there might be subcategories of RRIPP that are not typical of patients with severe and persistent mental illness.
the literature is to what degree the group of recurrently readmitted inpatient psychiatric patients (RRIPP) is coincident with the broader category of persistently and severely mentally ill and if there might be subcategories of RRIPP that are not typical of patients with severe and persistent mental illness.
Selected Review of the Literature
The RRIPP phenomenon has been explored and given different meanings over the years. Each era has imposed its particular issues on the phenomenon. The last 40 years have seen three substantial English language reviews5, 6, 7 and one German review.8 The authors will refer to these reviews and others as they develop a variety of themes and perspectives.
DEFINITION
What exactly is taken to constitute RR has varied in the literature, rendering comparisons among studies difficult and meta-analyses almost impossible. Not only have investigators used different rates of readmissions as the criterion, but also other measures have been suggested, such as days in hospital, time between hospitalizations, community tenure, and complex indices taking into consideration not only variables such as number of hospitalizations and rate, but also age at first hospitalization and associated costs. Byers and Cohen9 reported the only study the authors of this chapter found that explored the relationships between some of these various outcomes. They found that readmission within 1 year correlated with number of days in the community at the 0.69 level and with the number of days to first readmission at 0.89, and that the number of days in the community correlated with number of days to first admission at the 0.71 level. These findings suggest that these different definitions are not too discordant from one another. Without a consensus for the definition of RR, naturally it will not be possible to get a consensus on who the RRIPP are. Nonetheless, a common picture does emerge when one puts together all the available evidence. The standard of using the rate of admission of three (or more) within 18 months has been more widely adopted than any other standard.
SALIENCE OF RECURRENT READMISSION
As the RR phenomenon emerged as a research focus in the 1970s, some authors asserted that these returnees were “failures” on the part of the hospital, either because the hospital could not effect appropriate psychological or situational change, or the hospital was unable to provide adequate aftercare.5 More narrowly, DiScipio and Sommer10 considered readmission within 30 days of a therapeutic failure, reflecting either an error in judgment or failure to provide proper aftercare. However, several authors questioned the use of the readmission rate as a measure of the quality of care, at least of the hospital. Voineskos and Denault,11 Solomon and Doll,12 and Erickson and Paige13 all agreed that it is a fallacy to use the readmission rate as a measure of effectiveness of the hospital because the hospital did not control the actions of patients once they left, did not control the outpatient services, and did not have the ability to predict who might be vulnerable to return.
From 1955 to 2000, the population of state hospitals for the mentally ill went from 558,000 to 55,000,4 varying, of course, by region and state. As this massive cutback of inpatient services was coming to a close, an emphasis emerged in the public sector on the organization and management of community-based treatments. This heralded a time of professional preoccupation with rehospitalization rates, the adequacy of community services, and the high cost of rehospitalization. Managed care or rather “managed cost” policies were prevalent in the practice of inpatient psychiatry. Some studies directly addressed the effects of managed care, such as the reduction of length of stay in the hospital. For example, Appleby et al.14 examined whether patients who had a short length of stay were more likely to be readmitted. Length of stay was statistically significantly related to time to relapse after the effect of the number of previous admissions was taken into consideration. However, the small size of the effect raised questions about clinical significance. The authors argued that some patients need to stay in the hospital longer in order to recover sufficiently to remain out of the hospital because of the finding that the likelihood for readmission within 30 days was significantly higher in those with brief hospitalizations.
PREDICTION
Major themes of the literature have been the questions: Who are the RRIPPs? Can RR be predicted? And how did these patients end up using so much inpatient care? Many studies attempt to predict or show associations with RR even when the primary focus of the study is on some narrowly defined issue. Rather than discussing them individually and at length, the authors have summarized the results in Table 13.1 of studies that attempt to answer the questions in the preceding text. References in Table 13.1 are listed chronologically and include reports since 1970 that include a definition of RR and an inferential statistical approach to estimating the probability of variables’ being associated with membership in the RR group. But because the definitions of RR differ widely, these results are not easily comparable. The authors’ discussion in this section will take up measures of demographic variables, clinical features of symptoms and diagnoses, and attitudes toward clinical care and their illness. Later special issues and treatment and prevention approaches will be addressed.
Demographically, on first look the findings seem diverse and at times inconsistent among studies. Klinkenberg and Calsyn,7 the review most oriented to research method, emphasized the shortcomings of the research designs and methods as a major cause of this variation through the use of underpowered studies and the failure to use multivariate statistics to control for multicollinearity. Nevertheless, summarizing from the table of 31 selected reviews spanning 33 years from 1973 to 2006, the authors find the following:
No studies reported female gender more likely to be associated with RR; 4 reported male gender more likely to be associated with RR; and 30 found no difference or reported no gender findings.
Race barely ever distinguished RRIPP from non-RRIPP.
Youth was associated with RR status in nine studies, and no studies reported any other age-group more likely to be in the RR status.
Being unmarried was associated with RR status in three studies, and being married was not associated with RR status in any study.
Unemployment was positively associated with RR status in three studies.
Educational level was not a factor in any study.
In general, taking the studies in Table 13.1 together, one could say that there was a trend for patients to be male, younger, unmarried, and unemployed. However, it is important to note that almost all these associations fell out when there was control for one or two other variables such as a history of recurrent hospitalizations. None of these demographic variables have the specificity and sensitivity to be considered predictors of the RR.
Patient attitudes toward care have received a fair amount of attention, especially compliance.15, 16, 17, 18, 19, 20 Of the authors’ selected group of references, five reported a highly significant correlation between noncompliance with medication and frequent rehospitalization.
Diagnostically and clinically, psychosis is the more prevalent condition with nine references (four for schizophrenic diagnosis and five additional with psychotic symptoms). Yet other conditions were found in some studies, as follows: four for substance abuse and three for affective disorders. In addition to categoric diagnostic and symptom complex clinical characterizations, some studies reported correlations with RR through severe symptoms2 or chronicity.2 However, the main feature that was associated with RR status was former hospitalization: 13 studies reported an association and when multivariate statistics were used to control for multicollinearity, other associations usually dropped out of statistical significance.
Some, though, have found little difference between RRIPP and other patients. Lucas et al.21 did a 6.5-year study of 193 inpatients who were heavy users of inpatient care and compared them with a control group of 400 inpatients. Heavy users were diagnostically and demographically similar to ordinary inpatients, despite the fact that they used services at roughly three times the rate of other inpatients in terms of health care costs.
Social issues have also been given strong consideration in the literature. Kent and Yellowlees22 determined that social factors contributed to 39% of admissions. These social factors included relationship problems such as dysfunction and conflict in relationships, acute loneliness, social isolation, and lack of adequate continuous social supports, from either professionals or natural social networks. Another social issue of importance is the association of RR with homelessness in some studies.23 Casper24 found homelessness and a history of arrest to be important contributions to readmission.
TABLE 13.1 SUMMARY OF CITATIONS PREDICTING RECURRENT PSYCHIATRIC HOSPITALIZATION | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Related posts:
The Ins and Outs of 200 Years of Psychiatric Hospitals in the United States
Psychosocial Approaches in Inpatient Psychiatry
Legal Issues in Inpatient Psychiatry
Psychiatric Education on the Inpatient Unit
The Patient with Borderline Personality Disorder
Evaluation and Management of Substance Use Disorders on the Inpatient Psychiatric Unit
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
