A Systematic Review of the Impact of Admission to a Psychiatric Hospital on Quality of Life in Patients With a First Episode Psychosis

ABSTRACT

Introduction

Quality of Life (QoL) is an important outcome measure for people with a first episode psychosis (FEP). Despite high numbers of patients being admitted to psychiatric hospitals following an FEP, the impact of hospital admission on outcome measures such as QoL is under‐investigated. This study aimed to systematically review evidence of the impact of hospital admission on QoL in patients with an FEP.

Methods

A systematic review was conducted in line with the PRISMA Guidelines. PsychINFO, Medline, CINAHL, and EMBASE were searched from 2000 to 2022, examining the impact of QoL for people with an FEP following a hospital admission. Quantitative, qualitative, and mixed methods studies were included in the search. Study quality was assessed using the Mixed Methods Assessment Tool. Data was analysed using narrative synthesis.

Results

Seven hundred and eight two abstracts were found. Fifty‐three full texts were reviewed. Data was extracted for eight articles. Across the eight studies, four discrete QoL measures were employed. Results showed that QoL scores were consistent over time, with little statistical change from baseline to follow‐up. Across studies, participants reported low average QoL scores. Across studies, information on the treatment components and duration of hospital admission were scarce.

Conclusion

There is a relative paucity of studies that have examined QoL in relation to hospital admission. QoL scores remained consistent over time, suggesting that QoL is relatively well preserved at the first episode. Further research is needed on the impact of inpatient‐based interventions in FEP.

1. Introduction

1.1. Quality of Life (QoL) and First Episode Psychosis

The World Health Organisation defines Quality of Life (QoL) as an ‘individuals' perceptions of their position in life in the context of the culture and value systems in which they live and in relation to their goals, expectations, standards and concerns’ (Mas‐Expósito et al. ²⁰¹¹, 11). This broad‐ranging concept incorporates the persons' physical health, psychological state, level of independence, social relationships, personal beliefs and their relationships to their environment.

QoL is an important area in the treatment and prognosis of First Episode Psychosis (FEP). A high QoL conveys both good clinical and personal recovery from illness (Åsbø et al. 2022). Research has highlighted a decline in QoL as a factor in the onset of psychosis (Marwaha et al. 2016; Guedes de Pinho et al. ²⁰¹⁸). Similarly, whilst someone may reach symptomatic remission, this does not mean that they experience a better QoL as cognitive, affective or social/relationship elements of their life may continue to suffer (Lambert et al. 2010). Due to a historical overfocus on clinical recovery, Lambert and others have called for QoL assessments to be a standard part of the development of the definition of remission and recovery in psychosis (Lambert et al. 2006).

There are a variety of factors which are protective or negatively impact QoL in FEP. Socially, in FEP populations, racial inequities, single marital status, and being in receipt of welfare or unemployment have been related to poor QoL (Kaikoushi et al. 2021). Clinically, poor insight, a diagnosis of bipolar or schizophrenia, and a history of involuntary hospitalizations are also related to poorer QoL (Kaikoushi et al. 2021).

Studies with FEP participants reported that social support and a high level of satisfaction with friends and intimacy were independently associated with a higher quality of life (Guedes de Pinho et al. ²⁰¹⁸). Zendjidjian et al. (2014) found that satisfaction with care was related strongly to treatment adherence and QoL. Critically, poor QoL can have a negative impact on treatment outcomes in psychosis, whereby a reduction in a patient's QoL can reduce treatment adherence (Lambert et al. 2006).

1.2. Hospital Admission and First Episode Psychosis

Whilst the course of psychosis is heterogeneous and unpredictable (Di Michele and Bolino ²⁰⁰⁴), a recent meta‐analysis reported estimates that between 30% and 90% of people with a psychosis have a psychiatric hospital admission, with up to 55% of patients requiring a hospital readmission, over a 7‐year follow‐up (Ajnakina et al. 2019). There are some reported benefits to hospital admission, such as reducing risk of suicide (Shrivastava et al. 2016), and promoting healthy cognitive functioning (Bergh et al. 2016). However, admission to hospital may also disadvantage patients experiencing an FEP, due to its traumatic nature (Berry et al. 2013), the loss of independent skills, or prolonged unemployment (Tomar et al. 2003). Repeated hospital admissions have resulted in lower reported health related QoL (Herrman et al. 2002).

1.3. Research Aims

Despite a high number of people being admitted to hospital with an FEP, research has typically focused on positive symptomatic remission or the traumatic nature of psychiatric admission. Little is known about the impact of psychiatric admission on QoL in people with an FEP. To our knowledge, this is the first study to synthesise evidence on the impact of psychiatric admission on QoL in patients presenting with an FEP.

2. Materials and Methods

2.1. Research Design

This study utilised a systematic review methodology in line with the Prisma Guidelines. The extracted data was analysed using a narrative synthesis approach.

2.2. Search Strategy

Four electronic databases (i.e., PsychINFO, Medline, CINAHL and EMBASE) were searched to identify studies pertaining to the research question. Other search strategies, including forward and backward searches and hand‐searching bibliographies of related articles, systematic reviews and grey literature were also undertaken. The search strategy combined variations of main terms relating to hospital (e.g., ‘hospital admission, ‘hospitalisation’, ‘inpatient’, and ‘psychiatric hospital’), quality of life and first episode psychosis. Further, each search string was adjusted according to the specific requirements of each database (See Table 1 for key term combinations). This systematic review was registered with PROSPERO database (ref: CRD42022350678).

TABLE 1.

Search term variations & combinations.

Group	Key search terms (in title or abstract)
1	Quality of life or QOL or health‐related quality of life or HRQOL or s‐QOL or subjective QOL or global functioning or level of daily activities or ADL or level of social activities
2	First episode psychosis or FEP
3	Psychiatric hospital* OR psychiatric admission OR psychiatric inpatient OR hospitalisation
4	1 AND 2 AND 3

2.3. Inclusion and Exclusion Criteria

2.3.1. Inclusion Criteria

The systematic review search included studies that utilised quantitative, qualitative, or mixed method designs. The eligibility criteria comprised an admission to a psychiatric hospital for treatment of an FEP, as well as an assessment of QoL. Studies evaluating elements of QoL consisting of questionnaires, interviews, or patients' feedback were included. Only peer‐reviewed papers were included.

2.3.2. Exclusion Criteria

Only FEP samples were included. Studies that examined QoL in mixed patient groups, non‐differentiated or non‐FEP samples were excluded, due to the conflicting impact of chronic illness, multiple admissions or repeated trauma. Studies that assessed QoL in relation to other areas of psychosis (e.g., intervention studies) where no raw data on QoL was presented were excluded from the review as it was not possible to access the QoL scores. Conference and poster presentations were excluded, as were studies not published in English. Studies published before the year 2000 were excluded, due to the significant international changes in care systems, related to psychiatric admission, and treatment of an FEP, around this time (e.g., Handerer et al. ²⁰²¹; Healy et al. ²⁰⁰¹; Hotopf et al. ²⁰⁰⁰).

2.4. Screening Procedure

The titles and abstracts of the studies were screened independently by two researchers (CG/GM) according to the inclusion/exclusion criteria. In the case of disagreements, full articles were read, and a consensus was reached. In the case of full text review, any conflicts on inclusion or exclusion were resolved between the two researchers, with the inclusion of a third researcher (KG), as necessary. The details of studies at all stages of the screening process were recorded in the PRISMA flow diagram (See Figure 1).

FIGURE 1.

Prisma flow diagram of included studies.

2.5. Data Extraction

Two researchers independently extracted data from the included studies. A data extraction form was developed, piloted, and adjusted accordingly. This data extraction form was used to facilitate extraction of the following variables: (a) author and publication information, (b) sample population characteristics, (c) the quality of life measure used, (d) outcome measures of interest, and (e) potential bias/study quality (See Table 2).

TABLE 2.

Table of characteristics.

Authors	Sample (N) presenting with FEP	Country	Method	Findings as reported by the authors	Quality appraisal	Criteria not met
Addington et al. 2012	N = 606	Canada	Quantitative	Psychiatric admission was significantly related to lower s‐QoLa at baseline, 12 months and 24 months.	6/7	It was unclear if the sampling strategy appropriately addressed the research question.
			Prospective longitudinal
	18–65 years		Heinrich QoL Scale (HQLS)
	Male: Female (65%: 35%)		Assessment at admission, 12 months, 24 months
Górna et al. 2008	N = 74	Poland	Prospective longitudinal	At 1 month post discharge, s‐QoLa was significantly lower than the general population. There was no significant change on s‐QoLa at 12 months or 3–6 years.	6/7	The risk of nonresponse bias was unclear.
	Mean age: 24.7		WHOQOL‐BREF
	Male: Female (62%: 38%)		Assessment at 1 month post discharge, 12 months and 4–6 years
Górna et al. 2010	N = 86	Poland	Quantitative	Both at 1 month and 12 months, people not experiencing suicidal thoughts prior to admission, scored higher on s‐QoLa domains	6/7	The risk of nonresponse bias was unclear.
			Prospective longitudinal
			Pre‐Post Design
	Mean age: 25.5		WHOQOL‐BREF
	Male: Female (60%: 40%)		Assessment at 1 month post discharge and 12 months
Melle et al. 2010	N = 252	Norway	Quantitative	Change in general s‐QoLa from baseline to 2 year follow up was associated with change in depressive symptoms, and o‐QoLb (daily activities and social contacts).	6/7	The risk of nonresponse bias was unclear.
			Prospective longitudinal	There were no changes to o‐QoLb over the follow‐up period.
			Pre‐Post Design
	Mean Age: 30		Lehman's Quality of Life Interview
	Male: Female (56%: 44%)		Assessment at baseline and 24 months
Mwesiga et al. 2022	N = 94	Uganda	Cross‐sectional	Cognitive impairment was associated with both lower s‐QoL and functional impairment.	7/7
	Age range: 18–59 years		Quantitative
	Median Age: 26 years		WHOQOL‐BREF
	Male: Female (25%: 75%)		Assessment at single‐time point at admission
Renwick et al. 2012	N = 146	Ireland	Cross sectional	Improved s‐QoLa was predicted by admission. Improved social relations QOL domain was predicted by education level. Reduced QoLa domains were predicted by higher depression scores.	7/7
			Quantitative
	Mean age: 32.7 years		WHOQOL‐BREF
	Male: Female (60%: 40%)		Assessment at single‐time point at first referral
Srihari et al. 2015	N = 117 ST: 57 STEP: 60	USA	RCT	Participants in STEP programme improved on s‐QoLa, but not significantly compared to the control ST group.	7/7
			Standard Treatment vs. STEP programme
	Mean Age: 22.6 year		Heinrich QoL Scale (HQLS)
	Male: Female (81%:19%)		Assessment at Baseline, 6 months, 12 months
Thorup et al. 2010	N = 280	Denmark	RCT	There were no significant differences in QoLb between the ST group and the ‘OPUS’ group concerning the nine life domains, the ‘global QoL’ or the visual ‘quality‐of‐life‐ladder’	7/7
	ST: 140		Standard Treatment vs. OPUS trial
	‘OPUS’: 140
	Mean Age: 26.2		Lancashire Quality of Life Profile (LQoLP)
	Male: Female (68%: 32%)		Assessment at Baseline and 24 months

Abbreviations: O‐QoL = Objective Quality of Life; S‐QoL = Subjective Quality of Life.

2.6. Critical Appraisal/Risk of Bias Assessment

To determine the methodological quality of the included studies, the Mixed Method Appraisal Tool (MMAT), a series of quality indicators developed by Hong et al. (2018), was employed. These indicators assessed the appropriateness of the study design, population, characteristics of the sample, intervention results, analyses, and conclusions. Each indicator was rated one of three scores: ‘yes’, ‘no’ or ‘unclear’. The MMAT recommends not calculating an overall score from the ratings of each criterion (Hong et al. 2018) or excluding studies that are of a low methodological quality. As per guidelines, a more detailed presentation of the ratings of each criterion is described, highlighting the range of papers in a literature, as well as specific areas of strength or weakness. Two reviewers independently (CG, GM) matched the indicators against the included studies before converging to discuss any disagreement with a third reviewer (KG).

2.7. Data Synthesis

A narrative synthesis (Popay et al. 2006) was chosen as the method of data synthesis due to the existing heterogeneity of sample populations, study designs, contexts, and data analyses. This review followed the general framework of a narrative synthesis, as outlined in the guidance documents supporting this synthesis (Popay et al. 2006).

3. Results

3.1. Search Results

The initial search yielded 782 papers, of which 770 were identified through the four databases: PsychINFO, Medline, CINAHL and EMBASE. A further 12 studies were identified through other sources (i.e., grey literature, bibliographies). Following an initial removal of duplicates (n = 58), 724 papers remained. Screening of the titles and abstracts indicated that 671 papers did not meet the criteria and, therefore, were excluded. Fifty‐three papers were left for full‐text review. Following review, eight studies were included in the narrative synthesis. See Figure 1. for a PRISMA flow chart which provides an overview of the screening process referred to above, and reasons for study exclusion.

3.2. Study Characteristics

3.2.1. Study Designs

3.2.1.1. Admission to Hospital

Studies recruited participants through multiple referral pathways. Five of the studies investigated QoL following hospital admission in patients presenting with FEP. One study investigated QoL and its impact on cognitive functioning in a FEP population and two studies presented raw data on QoL in a patient group receiving standard treatment versus a specialised intervention for psychosis.

Table 2 presents an overview of outcomes examined across the included studies. Six of the eight studies examined hospitalisation and quality of life in patients presenting with FEP longitudinally. Two studies were cross‐sectional. Only one paper (Górna et al. 2008) followed the same cohort of participants over three different time periods.

3.2.2. Sample Recruitment

All eight studies recruited participants at point of admission. Three studies also recruited from out‐patient sites (e.g., community clinics: Melle et al. ²⁰¹⁰; Thorup et al. ²⁰¹⁰; Srihari et al. ²⁰¹⁵). Of note, with the exception of Mwesiga et al. (2022), all other studies were conducted in western countries.

3.2.3. QoL Measures

Six studies utilised a questionnaire to assess quality of life; four employed the World Health Organisation Quality of Life Measure—Brief form (WHOQOL‐BREF: Mas‐Expósito et al. ²⁰¹¹); two used the Heinrich Quality of Life Scale (HQLS: Heinrichs et al. ¹⁹⁸⁴; Addington et al. ²⁰¹²; Srihari et al. ²⁰¹⁵). These provided subjective measures of QoL. Two interview measures, Lancashire QoL Profile (Oliver et al. 1997; Thorup et al. ²⁰¹⁰) and the Lehman QoL Interview (Lehman 1988; Melle et al. ²⁰¹⁰), provided objective QoL.

3.2.4. Sample Characteristics

The total number of participants across the eight studies was 1391, of which Górna et al. (2010) recruited the least number of participants (N = 74) and Addington et al. (2012) included the largest sample size (N = 606). All but one study reported a higher number of males (averaging 64.6%). Four studies reported participants' mean ages as being in their 20s (Srihari et al. 2015; Górna et al. ^{2008, 2010}; Thorup et al. ²⁰¹⁰).

3.3. Narrative Synthesis

The data yielded three main findings: (i) four of the eight studies reported on QoL scores over time and of these, three found no statistically significant changes; (ii) a specialised intervention designed to improve QoL did not result in any significant improvements compared to treatment as usual (TAU); (iii) on average participants scored in the low average range for QoL across the majority of studies.

When taken together these findings suggest that QoL is relatively well preserved at FEP when considered in relation to significant symptomatic and functional decline that typically precedes an admission to hospital (Alvarez‐Jimenez et al. ²⁰¹²). This is encouraging and suggests a relative resilience which indicates that QoL may not be related to symptomatic and functional presentations but to other underlying factors such as social support, vocational engagement and depressive symptoms as previous research has suggested (Alvarez‐Jimenez et al. ²⁰¹²). However, current treatment options do not specifically focus on QoL and do not appear to have a specific impact with little change over time or across treatment allocations. These findings suggest that treatments could take a greater QoL focus with the potential benefit of preserving or improving QoL.

3.3.1. Quality of Life Over Time

The first finding was that QoL scores remained relatively stable over time, whether short‐term follow‐up or long‐term follow‐up. There were very few clinically or statistically significant differences across time points in the four studies that presented longitudinal data. Górna et al. (2010) showed very little variation across the QoL domains in the 12 months between data collection. The one paper that did measure objective QoL, Górna et al. (2008) reported that after one month of discharge patients' objective QoL was relatively low. Furthermore, during the first year of observation subjective QoL did not change significantly in any of the domains analysed at group level, nor did it at 4–6 year follow up (Górna et al. 2008). Melle et al. (2010) did find a moderate but statistically significant change in QoL score at two‐year follow up.

3.3.2. Standard Treatment vs. Specialised Intervention

The two RCTs included found no significant differences in QoL in patients between an enhanced early intervention in psychosis specialised intervention and TAU, with neither leading to a change in QoL scores (Thorup et al. 2010; Srihari et al. ²⁰¹⁵). Both studies reported greater improvements in other areas of functioning in the specialised interventions group: reduced hospitalisation; better engagement with outpatient services; engagement with vocational training but not on subjective QoL scores.

3.3.3. Consistency of Scores Across Studies

Although QoL scores appeared consistent across each of the studies, the scores differ based on the measure employed and cut‐off scores. It was perhaps a surprising finding that overall scores appeared only slightly below average given that hospital admission can be considered a proxy measure for illness severity or acuteness or risk. The results could be interpreted in one of two ways: (1) that participants themselves felt that hospital admission and the treatment was a positive addition to their recovery and wellbeing, or that (2) participants may not subjectively assess their QoL scores as low as they lack the insight to appraise QoL neutrally in an acute state of psychosis (López‐Moríñigo et al. ²⁰¹⁴; Leonhardt et al. ²⁰¹⁷).

3.3.4. Admission to Hospital

None of the papers make any reference to the length of admission any participant underwent, nor did any of the papers outline clearly what treatment was received by any participant or group of participants receiving TAU. Renwick et al. (2012) outlined what admission to hospital may involve for participants in their study. However, similar to the RCTs, the authors make reference to medication and also the wider team that may be involved in treatment but do not explicitly state what interventions patients received.

3.4. Study Quality and Potential Bias

All eight studies were reviewed against the MMAT quality appraisal tool (Hong et al. 2018). The research question was clearly stated in all studies. Each study was appropriate in subject selection. Whilst all studies stated their data collection methodology and were deemed appropriate, three of the studies who gathered follow up data (Górna et al. 2008, 2010; Addington et al. ²⁰¹²) were not clear in how this follow up data was obtained. Whilst each study noted attrition, no study gave a specific attrition rate. Only Górna et al. (2008), Thorup et al. (2010) and Srihari et al. (2015) reported reasons behind participants' attrition, for whom no data was included in the follow up timepoints. Further, the completeness of data as well as control for confounding variables were evident and deemed acceptable in four of the studies (Górna et al. 2008; Thorup et al. ²⁰¹⁰; Srihari et al. ²⁰¹⁵; Mwesiga et al. ²⁰²²). The remaining 4 studies failed to sufficiently address missing data and confounding variables. The statistical methods and analyses of each included study were deemed appropriate. Each of the studies could potentially be replicated. No study reported any ethical issues, and seven of the eight studies explicitly stated how ethical approval was granted.

4. Discussion

The aim of this study was to provide a systematic review of existing evidence evaluating the impact of psychiatric admission on QoL in people with FEP. The search found only eight studies reported on QoL in patients with FEP following admission to hospital, indicating a relative dearth of research in this area. Psychiatric admission did not have a significant positive or negative effect on participants' QoL over time. Equally, treatment allocation between specialist and standard care did not have a significant impact on QoL. The majority of patients had low average QoL at admission to hospital and this remained stable over time.

These findings in the context of the typical functional decline that often precedes a hospital admission are encouraging and suggest QoL may be relatively well preserved in FEP despite a period of acute illness. Psychiatric admission may provide social support, work, training or education, and intervention for depressive symptoms, all of which have been found to be supportive factors for protecting QoL (Alvarez‐Jimenez et al. ²⁰¹²). A focus on QoL at assessment and when designing interventions would support both the continued preservation and could potentially lead to interventions with a focus on QoL improvement in this cohort.

4.1. What Is Treatment as Usual During Psychiatric Admission?

None of the included papers described what defined standard care during psychiatric admission. Core aspects of the admission such as duration, pharmacological and psychosocial interventions were not described. Whilst this may be assumed to be obvious, it is highly unlikely to be universally similar or consistent over time. Inpatient admission is a common part of treatment pathways for people with FEP and can play a fundamental role in safety and wellbeing; however, within the literature, the ‘active ingredient’ or ‘negative side‐effects’ appear to be hiding in plain sight. This marks a substantial and important absence of knowledge. One of the studies in this present review outlined the difficulty in gathering baseline data in this population.

To further establish the efficacy of psychiatric admission, greater methodological robustness is warranted (e.g., studies that report pathways to admission, voluntary or involuntary status, the use of seclusion, restraint, non‐voluntary medication, reporting effect sizes and more studies that report follow‐up data), as well as the inclusion of clear outline of hospital intervention components.

4.1.1. Quality of Life Measures

Four discrete QoL measures were employed across studies presented in this systematic review. All four measures have been deemed valid and reliable. These findings all support the idea that recovery from FEP should not be based on one single feature, such as symptom remission (Lee et al. 2022). It should reflect what is meaningful for an individual, supporting further investment in QoL research.

However, the complexity of using QoL measures was demonstrated across the studies reported in this review. Renwick et al. (2012) noted a difficulty with the assessment of subjective QoL during the acute phase of illness as the presence of positive symptoms of psychosis may yield an inaccurate representation of QoL. Further, the researchers posited that subjective appraisal of QoL over time may be impacted by downward readjustment of expectations or adaptation to poorer conditions (Renwick et al. 2012). Melle et al. (2010) reported in their paper that changes in subjective QoL scores were associated with affective states and that objective QoL scores were associated with social activities and perceived general health.

Pursuing a more comprehensive understanding of the QoL measures being used, determining time of optimal administration, and confirming how they are best administered may allow a shift to more QoL‐focused interventions. Moreover, any collateral or objective measures used should form part of future studies to contribute to the knowledge in this important area, strengthening the knowledge base on personal and clinical recovery.

4.2. Quality Assessment

Overall, the quality assessment identified that the studies were at moderate risk of bias. Purposive sampling was employed in all studies, which is appropriate to the population being studied but also a concern as all participants presenting to services were invited to participate. Further, six reviewed studies identified a potential for missing data and information bias as a concern. Four studies reported on participants being recruited from inpatient and outpatient settings but did not offer clear data on the source of inpatient status. No study provided information on the effect sizes.

4.3. Strengths and Limitations

Several strengths support the robustness of this review. The study was pre‐registered and PRISMA Guidelines were followed. Two researchers undertook methodological processes (e.g., data screening) independently. There are also several limitations. Only a small number of studies were identified which impacts the generalisability of the findings. The identified studies had a range of QoL measurements and methodologies. The complexity of measuring QoL during an acute phase of illness was identified even within the small number of studies in this area.

5. Conclusions

This systematic review presents preliminary evidence that psychiatric hospital as an intervention for patients with FEP is an under‐researched and not a clearly documented area of treatment. QoL measures were shown to be appropriate and effective as an outcome measure in this population. The narrative synthesis reported three findings: subjective QoL scores remained consistent during admission; specialised intervention did not result in improved QoL scores, and subjective QoL scores were relatively well‐preserved. Future studies should aim to address the methodological limitations outlined and bring a greater research focus to this common intervention.

Funding

This work was supported by Health Service Executive.

Ethics Statement

As systematic review, ethical approval was not needed for this study.

Consent

As systematic review, participant consent was not needed for this study.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.