In this prospective study, we focused on psychiatric readmissions in one hospital district in Finland. From the available administrative and clinical data, we examined the associations of patient-, treatment- and ward-level factors with readmission. The most prevalent diagnoses were psychotic disorders (43%) and depressive and other mood disorders (35%). Approximately one-third of the treatment periods lasted less than one week. Readmission was relatively common: 11% of the patients returned within 30 days, and a third returned within a year. These figures are in line with previous studies from Finland [5].
There were slightly more women (52%) than men (48%), which is in accordance with a previous national study [5]. There were no major differences in readmission between women and men; however, men were less likely than women to have a readmission within 30 days. A previous study suggested no difference between women and men in Finland [3]. In general, clinical and biopsychosocial factors may explain the difference between women and men. Women have a greater tendency to seek treatment when experiencing symptoms [21]. It has also been suggested that cultural assumptions related to both gender and health are produced in social interaction with the environment; therefore, gender roles might guide treatment seeking among women and men [22]. However, the interaction analysis suggested that women with a psychotic disorder might have a lower likelihood of readmission within 30 days than women without a psychotic disorder. As this association was not statistically significant, this question remains to be examined in future studies.
We found that in some models, those in the oldest age group (> 70 years) had a lower likelihood of readmission within a year while no association with age was found for 30-day readmission or multiple readmissions. This is in line with previous studies carried out in an adult population in Finland [3] and internationally [10, 14, 18], in which younger patients have been more likely to return to the hospital than older patients. However, we also found that patients with disorders of psychological development and behavioural and emotional disorders with onset usually occurring in childhood and adolescence were less likely to experience readmission within 30 days. In addition, frequent readmissions were also less likely in this group, where the patients typically are children and adolescents. Our findings may reflect the effort to implement treatment in outpatient care whenever possible so that the child’s or young person’s connection to the family and other support networks is not broken. However, it is important to acknowledge that in child and adolescent psychiatry, there has been a shortage of resources for years, which can mean that you cannot reach a hospital even if there is a need.
We found that neurotic, stress-related and somatoform disorders were associated with readmission within 30 days. The examination of neurotic, stress-related and somatoform disorders is new, as previous research has mostly focused on severe mental disorders and depression. The few studies including these disorders have suggested mixed findings [10]. Many stress-related and anxiety disorders, such as posttraumatic stress disorder (PTSD), are relatively severe; however, generalized anxiety disorder (GAD) is considered a milder disorder. PTSD can be long-lasting and requires intensive treatment. Typical symptoms of GAD include excessive worry about one’s own symptoms, which can increase help-seeking behaviour.
Psychotic disorders, such as schizophrenia, were associated with readmission within one year. Schizophrenia has been associated with the likelihood of readmission in previous studies [3, 9,10,11]. This reflects the severity of psychotic disorders and the need for inpatient treatment for these disorders.
Having a behavioural syndrome associated with physiological disturbances and physical factors or disorders of adult personality and behaviour was associated with readmission. The few previous studies on personality disorders have suggested an association with readmissions [10] whereas we are not aware of previous studies on eating disorders. In the present study, eating disorders were the most prevalent disorders in this group of mental disorders. Eating disorders include anorexia nervosa, bulimia nervosa and binge eating disorder. The most severe, complex, and long-lasting cases of eating disorders are treated in specialized hospital care [23]. Approximately 70% of the patients recover, but only a quarter recover within a year, which suggests a long-lasting recovery process with several readmissions [23]. Of the personality disorders, the prognosis of patients with emotionally unstable personality disorders is quite good; more than half of them no longer fulfil the diagnostic criteria after five years [24]. However, this and other personality disorders, such as paranoid, schizoid, dissocial, anxious (avoidant), or dependent personality can weaken the treatment efficacy for other disorders, such as depression [25]. All diagnoses were recorded in our dataset, so it is possible that personality disorders were coded as secondary diagnoses.
Comorbid mental disorders were neither directly associated with readmissions nor was there any interaction between comorbidity and ward overload as a predictor of readmission. The lack of evidence is in line with previous research which suggests that the association between psychiatric comorbidity and readmission is inconclusive [10]. Comorbid somatic disease was not associated with readmission either, which is in contrast with previous findings showing an increased risk of readmission associated with somatic comorbidity [3, 13, 15]. In our study, a small negative association was found only within one year. It is possible that somatic comorbidity can play a role, for example, in older patient groups, which could be examined with larger datasets.
Neuromodulation treatment was associated with readmission within 30 days, and among those with at least one readmission, it was associated with frequent readmissions. In a previous study with a Chinese sample, ECT treatment was associated with a lower likelihood of readmission [14]. However, in China, ECT treatment is also provided for treatment groups other than those with depression, e.g., for those with schizophrenia, and in that study, the effect of ECT on readmission was shown among schizophrenia patients [14]. In our study, readmission can be related to planned readmissions; the treatment procedure included new, agreed-upon neuromodulation sessions.
In line with previous findings [10, 11], we found that a history of psychiatric hospitalizations was consistently associated with readmission and frequent readmissions. Having hospitalizations in the past may be a proxy for several disease-, patient- and environmental-related factors affecting hospitalization for mental disorders. However, the length of the index period was not associated with readmission. A systematic review suggested mixed findings on this issue [10]. The authors concluded that among those with schizophrenia or substance use disorders, longer inpatient treatment could protect against relapse. However, future research could examine in more detail the role the duration of treatment plays in different diagnostic groups. Some studies have applied natural experimental designs to assess the impacts of system-level decisions to shorten hospital treatments. These studies have found that readmissions increased after the shortening of treatment periods [12].
We did not find associations with ward overload, as expressed by the bed occupancy rate. Previous studies focusing on hospital capacity have usually utilized the number of patients or the number of beds to represent hospital capacity rather than bed occupancy, as in our study, and the findings have been mixed [9, 12]. Other studies have focused on personnel resources, and in those studies, better resources in the hospital district [9] or unit [16, 17] were associated with a lower risk of readmission, although in one study, the area-level number of mental healthcare personnel per inhabitant was not associated with readmission [15]. Overload in the hospital ward can have many effects on the treatment process. For example, in long-term overloaded wards, access to treatment can become more difficult. This might have had an impact on our results, as the difficulties in admissions may have led to fewer readmissions in our patient population. In further studies, it would be important to examine the link between ward overload and access to acute psychiatric inpatient care.
The strengths of our study include the use of large, routinely collected data from an entire hospital district, which enabled us to examine readmissions in detail. Restricting the study to one hospital district excluded the possibility of regional differences affecting the findings [5]. However, there are also several limitations. Since the data available for this study were limited to inpatient care records, we had not, for example, the possibility to consider factors that were related to ward personnel, such as nurse-to-patient ratio [16, 17] and to assess post-discharge characteristics. The quantity and quality of outpatient services after inpatient treatment are important topics for future research [26], as a previous study in Finland showed that those without outpatient treatment contact were at the greatest risk of readmission [5]. For some hospital wards, the ward overload variable covered more than one sub-unit which may had led to some inaccuracy in the estimates. Our study did not assess several patient characteristics, such as disorder-specific factors, commitment to treatment, and economic circumstances [26]. The prevalence of somatic diseases was 10%, which suggests that they are under recorded in psychiatric medical records. Therefore, we did not conduct interaction analyses between psychiatric and somatic diseases. This could be an important topic for future studies [27]. Although we did not find any meaningful interactions between the assessed factors, future studies could continue this line of research by identifying vulnerable groups. In addition, we did not have information on whether the readmission was planned. In Finland, approximately 80% of readmissions are unplanned [5]. Finally, our findings are generalizable to one hospital district in Finland.
link
