Efficacy of day care treatment against readmission in patients with schizophrenia: A comparison between out-patients with and without day care treatment
Abstract
Abstract The present study examined the efficacy of day care (DC) treatment against readmission to a mental hospital. Subjects were 65 out-patients with chronic schizophrenia after discharge. Day care treatment was defined as positive if patients were under DC treatment constantly for 4 months or more with a frequency of at least one visit per week. Information regarding demographic and disease-related factors was obtained from medical records at the time of discharge. Logistic regression analysis was used to calculate odds ratios (OR) and 95% confidence intervals (CI) with adjustment for confounding variables. The modestly preventive efficacy of DC treatment against readmission was observed within 2 years after discharge (adjusted OR 0.52; 95% CI 0.12–2.35). The present findings suggest that DC treatment may be preventive against readmission for schizophrenic out-patients.
INTRODUCTION
Day care (DC) treatment has been established as rehabilitation for patients with mental disorders. However, in Western countries, previous studies are inconsistent with regard to the relationship between DC treatment and its effectiveness for rehabilitation.1–7 Regarding the readmission rate after DC treatment, several studies1–3 have reported that DC treatment is not associated with a decreased readmission rate, whereas some have supported the relationship between DC treatment and a decreased readmission rate.4,5 Guy et al.1 showed that patients with DC treatment required a shorter period of hospitalization than those without DC treatment in spite of the similar readmission rate between these two groups.
In Japan, there are few studies that have examined the relationship between DC treatment and its preventive efficacy against readmission to hospital based on certain clinical evidence.8–10 These studies are consistent in that DC treatment was preventive against readmission as long as the follow-up period was within 2 years. However, the studies failed to adjust for other confounding factors that could have some effect on the relationship between DC treatment and readmission rate, such as socioeconomic factors, drug treatment and patients’ comprehension about their mental disorders. Indeed, it was reported that social class had a great effect on the prognosis of schizophrenia.11 The present study examined the efficacy of DC treatment using multivariate analysis among patients with schizophrenia in a rural Japanese district.
METHODS
Subjects
There were 163 patients with schizophrenia who were discharged from Okumura Hospital over the period January 1990–August 1998. All met the criteria of ICD 10 schizophrenia. Only the last hospitalization was counted if subjects had had multiple hospitalizations during the period in question. We excluded patients who changed hospital or died (n = 75) or dropped out from attending the hospital (n = 23) within 2 years after discharge. Thus, 65 patients remained in the analysis. Furthermore, to assess prognosis 3 years after the discharge, we used a subgroup (n = 59) whose information on relevant factors was ascertained within 3 years after discharge. Of 65 patients, 19 (29%) were under DC treatment. Their ages ranged from 18 to 75 years, with a mean (±SD) of 41.9 ± 13.6 years. Fifty-four percent were males (data not shown).
Day care treatment
The DC institution in Okumura Hospital was established and received authorization from the Ministry of Welfare in July 1994. The treatment program is composed of several therapies, such as recreation therapy, occupational therapy and social skills training. Patients can choose any programs in consultation with their physicians in charge. The institution is equipped with comedical staff, such as nurses, occupational therapists, psychiatric social workers and clinical psychologists. The doctors of the hospital also practice at the DC institution.
The mean age of patients who received DC treatment was 44.1 years (2000). During the period from January 1996 to December 2000, the percentage of males and patients with schizophrenia was 59 and 71%, respectively. Patients had DC treatment on average 2.3 days per week and 14 patients on average per day used the institution during the period in question. The patients in the DC institution receive medical advice from the attending physician once a week concurrently with assessment of the DC treatment. Therefore, in the present study, DC treatment was defined as positive if the patients were under DC treatment constantly for 4 months or more with frequency of at least one visit per week. The present study used subjects whose beginning of treatment was within 1 month after discharge.
Demographic and disease-related factors
Information on demographic and disease-related factors was obtained from medical records at the time of discharge. These factors were age, sex, period of last hospitalization, admission type, presence of public assistance and dose of medicine. Age and sex were ascertained as demographic factors. They are the most basic demographic factors and could also be confirmed by the medical records of the hospital.
Other disease-related factors were checked as much as possible against the information contained in the medical records. Hospitalization period is considered to have a considerable effect on adaptability of the patients to society, because their social skills are considered to be occasionally harmed by a long stay in the specific environmental situation of the mental hospital. The type of admission was classified as either voluntary admission (defined as admission with the patients’ agreement) and compulsory admission (defined as admission with the guardians’ agreement but without the patients’ agreement). These classifications of admission type reflect the patients’ comprehension about their mental status. To assess the economic status of the subjects, the presence of public assistance was also ascertained. The economic status, which is probably connected with the patients’ quality of life, also plays an important role in the prognosis of the disease. Although familial support for the patients is also considered to be important, we did not take it into account in the analysis because the number of family members was recorded only at the time of admission and may have changed at the time of discharge. The dose of major tranquilizers was converted into that of chlorpromazine (mg/day) according to the method of Inagaki et al.12 Naturally, medication has a considerable effect on the prognosis of the disease.
Statistical analysis
Analysis of variance (ANOVA), Chi-squared test and unpaired t-test were used in the analysis of the distribution of demographic or disease-related factors. Because there may have been some bias with 23 patients dropping out of the study, a comparison was made between those who dropped out of the study and the remaining subjects with regard to demographic and disease-related factors.
Logistic regression analysis was used to make adjustments for potential confounding variables. Age, hospitalization period and dose of medicine were adjusted for as continuous variables. The other indicator variables were treated as covariates and adjusted odds ratios (OR) and 95% confidence intervals (CI) were obtained from the corresponding logistic regression coefficients and their standard errors. Two-sided P < 0.05 was considered statistically significant. All computations were performed using the SAS software package version 6.04 (SAS Institute, Cary, NC, USA).
RESULTS
Table 1 shows variances in demographic and disease-related factors among three groups: (i) patients who were readmitted (n = 14); (ii) out-patients (n = 51); and (iii) patients who dropped out of the study (n = 23). The patients who dropped out of the study tended to be on lesser doses of medication when they were discharged from hospital. The other factors did not differ among the three groups.
| Factor | Readmission (n = 14) | Out-patients (n = 51) | Drop outs (n = 23) | P |
|---|---|---|---|---|
| Mean age (years) | 41.6 ± 3.6 | 41.9 ± 1.9 | 44.7 ± 2.8 | NS |
| Males (%) | 57.1 | 52.9 | 69.6 | NS |
| Hospitalization period (days) | 684 ± 366 | 615 ± 192 | 599 ± 285 | NS |
| Voluntary admission (%) | 57.1 | 66.7 | 69.6 | NS |
| Public assistance (%) | 21.4 | 15.7 | 21.7 | NS |
| Dose of medicine* (mg) | 753 ± 161** | 642 ± 84 | 392 ± 126 | 0.15 |
| Day care treatment (%) | 21.4 | 31.4 | 0 | NS*** |
- Data are the mean ± SEM where appropriate.
- * Converted into the dose of chlorpromazine.
- ** P = 0.08 compared with the drop outs.
- *** Comparing the group who was readmitted with the out-patients group.
To check selection bias among patients receiving DC treatment, the distribution of demographic and disease-related factors was compared between patients with and without DC treatment (Table 2). There were no clear (significant) differences regarding these factors, except that patients with DC treatment were given higher doses of major tranquilizers than those not undergoing DC (P = 0.08).
| Factor | DC (+) (n = 19) | DC (−) (n = 46) | P |
|---|---|---|---|
| Mean age (years) | 38.3 ± 3.1 | 43.3 ± 2.0 | NS |
| Male (%) | 57.9 | 52.2 | NS |
| Hospitalization period (days) | 655 ± 329 | 619 ± 170 | NS |
| Voluntary admission (%) | 73.7 | 60.9 | NS |
| Public assistance (%) | 21.1 | 15.2 | NS |
| Dose of medicine*(mg) | 888 ± 202 | 574 ± 76 | 0.08 |
- Data are the mean ± SEM where appropriate.
- * Converted into the dose of chlorpromazine. DC (+), patients receiving day care treatment; DC (−), patients not undergoing day care treatment.
The relationship between DC treatment and readmission rate is presented in Table 3. It was found that DC treatment tended to be modestly preventive against readmission within 2 years after discharge (OR 0.52; 95% CI 0.12–2.35). In the 59 patients whose information regarding prognosis after 3 years of discharge was ascertained, DC treatment tended to be more preventive against readmission risk (OR 0.46; 95% CI 0.08–1.97). The relationship was slightly attenuated (OR 0.73; 95% CI 0.19–2.84) when the prognosis of 3 years after discharge was assessed.
| No. readmitted | Age-adjusted | Fully adjusted* | ||||
|---|---|---|---|---|---|---|
| Present | Absent | OR | 95% CI | OR | 95% CI | |
| 2 years | ||||||
| DC (−) | 11 | 35 | 1.00 | 1.00 | ||
| DC (+) | 3 | 16 | 0.58 | 0.14–2.43 | 0.52 | 0.12–2.35 |
| 2 years** | ||||||
| DC (−) | 11 | 30 | 1.00 | 1.00 | ||
| DC (+) | 3 | 15 | 0.53 | 0.12–2.26 | 0.46 | 0.08–1.97 |
| 3 years | ||||||
| DC (−) | 13 | 28 | 1.00 | 1.00 | ||
| DC (+) | 5 | 13 | 0.86 | 0.25–3.02 | 0.73 | 0.19–2.84 |
- * Adjusted for age, sex, hospitalization period, type of admission, public assistance and dose of medicine.
- ** Subjects were restricted to those whose information regarding prognosis was ascertained within 3 years (n = 59).
- OR, odds ratio; CI, confidence interval; DC (+), patients receiving day care treatment; DC (−), patients not undergoing day care treatment.
Furthermore, the hospitalization period before and after having DC treatment was compared in four patients whose hospitalization periods were available both before and after receiving DC treatment. The hospitalization periods became shorter (approximately 3 months) after DC treatment than before, but this difference did not reach statistical significance (P = 0.49; data not shown).
DISCUSSION
Selective effect of DC treatment on patients
As previous studies9,13 have pointed out, the efficacy of DC treatment could be confounded by other clinical or social factors and, therefore, the relationship between DC and readmission risk may be obscured without taking these factors into account. Put another way, it could be supposed that patients undergoing DC treatment may be biased in the process of selection for DC treatment. However, in the present study, demographic and disease-related factors did not vary between patients with and without DC treatment, except that the patients with DC treatment tended to be more medicated than those without DC treatment. As a rule, taking medicine regularly is more important for patients who are under treatment with high doses of medication than those taking low doses and patients with high doses of medication may be chosen for DC treatment because DC treatment helps them to keep taking medicine. The observation that DC patients in the present study were taking higher doses of medicine may be due to this selective effect. However, even after adjusting for the dose of medicine, the preventive tendency of DC treatment was still observed.
Overall preventive efficacy of DC treatment
In the present study, DC treatment was shown to be modestly preventive against readmission. The efficacy was attenuated in the prognosis 3 years after discharge. These findings are consistent with those of previous reports, which demonstrated that DC treatment is preventive against readmission within 2 years after discharge, but not for 3 years or longer. Takeda et al.13 reported that the preventive efficacy of DC treatment could last for 4 years or more. However, as suggested by Tsukahara et al.,9 that study analyzed periods of attending the hospital before and after receiving DC treatment in patients who had multiple hospitalizations and the period after DC treatment included that of the patients whose readmission had not occurred. This study design may affect the period in which DC treatment was considered to be effective in that study.13
Assessment of patients who dropped out of the study
As shown in Table 1, 23 of 88 subjects (26%) dropped out from attending the hospital within 2 years after discharge and the prognosis of these patients could not be ascertained. Comparison between patients who dropped out and other groups revealed that the patients who dropped out received less medication than others. This suggests that the patients who dropped out had relatively slight symptoms that vanished during the first several months and, consequently, they did not have to attend the hospital any more. If they had received DC treatment, the efficacy of DC treatment would have been overestimated because such patients are considered not to need readmission. The results may have be reversed if the patients had continued attending the hospital without undergoing DC treatment.
Study limitations
The limitations of the present study include a lack of assessment for social or clinical factors, such as social support, specificity of the community, substance abuse, compliance of medication and the patients’ physical condition, which have some effect on readmission risk.14–16 In particular, familial support was considered very important in previous studies.5,10 However, it should be noted that the mean age of our study subjects was not less than 40 years. Old parents are not necessarily helpful and occasionally even impose a burden on the patients. Schizophrenia often gives rise to induced psychosis in patients’ family members, such as children and spouses, thereby having a bad influence on the patients’ mental status. Thus, it is difficult and not practical to objectively evaluate familial support.
Another limitation of the present study is that the decision for readmission may depend on individual psychiatrists. However, we defined brief criteria for admission as follows: patients with self-injuring behavior (including suicide), those with violent or impulsive actions, those with a bad compliance with their medication and those beyond their family's care for other reasons. The judgment for admission type did not vary among the psychiatrists because the necessary procedures for admission conformed strictly to the law.
Finally, the sample size is too small to clearly indicate a protective effect of DC treatment. Although two studies1,17 have reported a significant reduction of the hospitalization period in patients undergoing DC treatment, there was no significant improvement regarding the hospitalization period with DC treatment in the present study. This result was mainly caused by a wide standard deviation in hospitalization periods caused by the small sample size.
CONCLUSIONS
In summary, the present findings suggest that DC treatment may be preventive against readmission in schizophrenic patients, independent of other relevant factors. However, some methodological problems have resulted in difficulties in interpreting the data. Further large-scale studies are warranted regarding DC treatment and its preventive efficacy against readmission.
ACKNOWLEDGMENTS
The authors thank Dr Atsume Okumura, the owner of the hospital, and all comedical staffs of the DC institution. Ms Kumi Gouto made a considerable contribution by preparing the data set in the present study.
