Patient satisfaction and acute psychiatric inpatient treatment

Abstract

Background: Patient satisfaction (PS) with treatment is one of different outcome- and quality measures used by health care providers worldwide to improve service. We report from a study of patients admitted to the Department of Acute Psychiatry at the Oslo University Hospital where we investigated PS and difference between genders, days of hospital stay, diagnostic groups, voluntary—and involuntary admitted patients according to hospital records and perceived voluntary—and involuntary admittance.

Materials and methods: All admitted patients during a 9-month period in 2014 were asked to participate by written consent. We used The Psychiatric Inpatient Questionnaire (PIPEQ), a self-report survey validated for assessment post-discharge. Analyses were conducted for a general dimension of PS and individual questions. A user representative was a part of the study from the beginning.

Results: A total of 357 patients were asked and 256 consented. Results show that 68% were over all satisfied and 14% dissatisfied. Highest PS was found for cooperation with relatives and lowest for influence on choice of treatment and medication. We found no significant difference in PS between men and women, but patients with a personality disorder and with short stay were less satisfied. PS was significantly less for those perceiving involuntary admission regardless of legal status.

Conclusion: The PIPEQ gives important input of patient’s experience with the delivery of care. Answers range from very much satisfied to not at all depending on what was asked for. Exploring PS provides valuable information for quality improvements for different patient groups.

Keywords:

• Patient satisfaction
• acute psychiatry
• hospitalization
• involuntary admission
• perception

1. Introduction

Satisfaction with treatment correlates with clinical outcomes, adherence to treatment and patient safety and is one of different outcome and quality measures used by health care providers to improve service and allocate resources [1]. Satisfaction with care is suggested as one of several key indicators by the World Health Organization and other international organizations in their promotion of health for all [2–5]. With the emergence of the ‘patient`s health system’ and the increasing role of consumers in health policy decisions, assessing patient satisfaction (PS) is important [6] and it is implemented as a mandatory key indicator of the deliverance of the health service by governments of both low- middle and high income countries [7–12].

Patient satisfaction is a multidimensional concept [13,14] with expectations of treatment and outcome, experience with different aspects of delivery of care and information as core elements. Satisfaction is influenced by personal factors, former experience, and time since discharge [14–16]. High PS is not necessarily the same as a high quality health care [17,18]. Both qualitative and quantitative methods are applied to measure satisfaction, such as self-administrated forms, telephone or face-to-face interviews and lately also by electronic questionnaires [19,20]. Many different scales are in use and different medical disciplines have developed their own adjusted instruments [8,20,21].

Assessment of patients satisfaction with the mental health service started in the 1980s [21,22]. The mental health services differ from other medical specialties in its worldwide use of coercion and involuntary treatment, and assessment of PS may seem somewhat contra intuitive. The mental health service is under change in a joint global effort to reduce coercion and improve service [23,24]. Systematically exploring patient’s satisfaction may indicate whether these changes reach out and represent improvements for the patients [25]. In the last 30 years, the Norwegian health care system has prioritized mental health service both in terms of increased financial investments and improvement of quality [26,27]. The first standardized national survey of patient`s experience and satisfaction with inpatient treatment was carried out in 2005 [28]. As the national surveys have been of short duration spanning over a few days to weeks and with few participants from each hospital, the potential for using patient experience for quality improvements has been limited. Although regular assessment of PS is mandatory for every hospital, they are now often conducted as electronic surveys and little specific feedback is given directly to the health care professionals executing the services on an everyday basis.

Despite representing challenging clinical settings, acute psychiatric services must provide high quality treatment both from a medical perspective and from the patient’s point of view [29,30]. For many patients, admission to an acute psychiatric ward is their first encounter with psychiatric health care, either as a solo admission or as the first of several and dissatisfaction with treatment could delay future help seeking [31]. Patients are vulnerable as their mental state is in disarray, many are involuntary admitted and some perceive they are involuntarily admitted while they are not [32–34]. Both groups have shown less satisfaction with treatment than those voluntarily admitted.

Most studies exploring PS in all patients following an acute psychiatric admission are more than 10 years old [35–38]. Studies of recent date either include a small sample of the population [39] or selected groups of patients such as those with psychosis [31,36,37,40], first-time admissions [41,42], those involuntary admitted [43] or difference in care provided [44]. To the best of our knowledge, there are no studies of recent date including all patients admitted to an acute psychiatric ward.

The Department of Acute Psychiatry at Oslo University Hospital (OUS) participated in the first national survey in 2005. Results were only presented on a regional level where OUS was one of eight participating hospitals represented by different inpatient wards, from open—to forensic, which made the results hard to interpret at the individual ward level [28]. The Department of Acute Psychiatry thus aimed to conduct an on-site survey in order to explore how patients evaluated their stay and what contributed to satisfaction. Together with the user representative, we aimed to assess PS throughout several months with the same instrument used in 2005 [28]. We investigated potential differences in satisfaction between men and women, days of hospital stay, diagnostic groups, voluntary—and involuntary admitted patients according to hospital records and perceived voluntary—and involuntary admittance.

2. Methods

2.1. Participants and procedures

All patients admitted to the Department of Acute Psychiatry at the OUS in the study period were eligible to participate, except those with a diagnosis of mental retardation and those who did not understand or speak Norwegian. The survey was conducted during 9 months from 1st of April until the 31st of December in 2014. In 3 months prior to the study, we conducted a pilot study and prepared the wards for the main survey. Included in the study were those who made a written informed consent and above the age of 18. During the study period, a total of 650 patients were admitted. A total of 357 were asked to participate and 256 consented. Patients admitted more than once during the study period, were only asked during the first stay. The acute ward is part of the sectorized health system in Norway and served a population of 202.000 over 18 years in 2014. Patients are admitted 24/7 from either a primary physician, the central acute medical service or psychiatric outpatient clinics. Upon admission, all patients are assessed in a reception ward and either allocated to other wards for further inpatient treatment or discharged within a few days.

Voluntary (VA) and involuntary admission (IA) in the Norwegian health care system is regulated by the Mental Health Care Act from 2001 and amendments from 2007 [45]. A specialist in psychiatry must evaluate the patient who has been involuntarily referred to within 24 hours. Involuntary admission may be in terms of involuntary observation or involuntary treatment. In the cases of observation a new evaluation must be made within 10 days and changed either to VA or IA. IA can be appealed by the patient to the local Judical Committee. The patient is assigned a lawyer free of charge for representation in the appeal. Coercive measures such as seclusion, isolation, use of restraints and pharmacological treatment againts the patient`s consent are regulated by different sections of the Mental Health Act and can also be appealed.

The study was approved by the Data Protection Authority at the OUS (2014/1047). User representatives were involved in the planning stage preparing the outline of the study, participated in formulating the written information and consent, analysing results and preparing the manuscript. Patients were informed by a brochure describing the survey upon admission, they were verbally informed during the stay and was asked for consent before discharge. Nurses were responsible for reminding the patients during the last days of stay to complete the survey. Unless the patient specifically asked for assistance, the survey was answered by the patient with no personal attending and left in a closed box in order to secure anonymity.

2.2. Assessments

Information regarding sex, age, length of stay, diagnosis and legal status upon admission was retrieved from the hospital records for all patients who consented. For all patients admitted during the year of the survey, we used the hospital clinical database to withdraw information regarding diagnoses, legal status and an average length of stay, but these data were only available as background information and were not applicable for statistical analyses (Table 1). The ICD-10 is the official diagnostic system used in Norway and the Mini Diagnostics Interview is used for diagnostic evaluation [46]. We only report the primary diagnosis at discharge. We organized the different diagnoses into the following six groups: 1: F20-29 Psychotic disorders; 2: F30-39 Affective disorders; 3: F40-48 Neurotic disorders; 4: F10-19 Drug-related disorders; 5: F60-69 Personality disorders; 6: F00-09 and F99 Other disorders. Global assessment of Functioning scale (GAF) is scored upon admittance and at discharge to give an overall measurement of severity of symptoms and functioning. The OUS uses the split version of the GAF scale that reports on symptoms (GAF-S) and functioning (GAF-F) independently [47]. The wards organize regular in-house training for all clinicians in GAF assessment.

Table 1. Sociodemographics and clinical data.

	All admissions	Consented	Not consented
	N = 875	N = 256	N = 91
Sex (women) %	51	50	46
Age
Mean (SD)		39.6 (15.1)	42.7 (15.3)
Civil status %
Married		15
Co-habitation		8
Single		66
Other		11
Diagnosis %
F 20-29 Psychotic disorders	28	28	54*
F 30-39 Affective disorders	19	33	15*
F 40-48 Neurotic disorders	10	11	10
F 10-19 Drug related disorders	16	10	5
F 60-69 Personality disorders	6	8	4
F 00-09 and F 99 Other disorders	21	10	12
GAF Mean (SD)
GAF-S admittance (n = 192)		38.5 (10.9)	38.8 (13.2)
GAF-F admittance (n = 189)		40.2 (11.2)	38.3 (11.7)
GAF-S discharge (n = 139)		52.8 (10.7)	45.4 (11.8)*
GAF-F discharge (n = 142)		50.8 (11.4)	44.8 (13.4)*
Days of stay (range) mean (SD)
All (0–173 days)	12.0 days	18.4 (23.0)	18.6 (29.6)
Reception ward (0–19 days)	1.6 days	1.5 (1.5)	1.3 (1.1)
Other wards (1–173)	23.4 days	26.8 (20.8)	27.8 (33.6)
Discharge from reception ward %	50	36	40
Referred to inpatient treatment %	50	64	60
Legal status of referral %
Voluntary admission (VA)	50	55	46
Involuntary admission (IA)
Involuntary observation	32	25	34
Involuntary treatment	18	20	20
Legal status of admission %
Voluntary admission (VA)	66	64	58
Involuntary admission (IA)
Involuntary observation	5	6	8
Involuntary treatment	29	30	34
PIPEQ: Perception of admittance %
Admitted at own free will (n = 246)
YES		72
NO		28

GAF-S: Global Assessment of Functioning, Symptoms; GAF-F Global Assessment of Functioning, Functioning; N, n number; PIPIEQ: Psychiatric Inpatient Experience Questionaire; SD: standard deviation.

*p ≤ 0.05.

We used the the Psychiatric Inpatient Experience Questionaire (PIPEQ) which is a self-report survey developed and validated to measure PS and experience post discharge [28,48]. PIPEQ is developed by the former Norwegian Knowledge Centre for the Health Services, commissioned by the Norwegian Directorate of Health and was originally a postal survey. The questionaire is part of the Program for Norwegian Measurement of Patient Reported Experience intending to provide external indicators at an institutional level. The PIPEQ consists of 26 questions (Q) of which 17 address different aspects of general satisfaction, experience- and information of treatment. The remaining nine questions address the patients right to access their medical record, right to complain about treatment, opinion on mental health in general, perception of voluntary and involuntary admission and sociodemographic information regarding age, sex, civil status and mother language. The 17 Qs regarding satisfaction are all scored on a 5-point likert scale: very dissatisfied = 1, dissatisfied = 2, neutral = 3, satisfied = 4 and very satisfied = 5. All Qs have an option for answering “Do not know” or “Not applicable to me”. Table 2 displays the 17 questions with corresponding scores shown as mean, standard deviation (SD), number of missing (%),”Do not know/Not applicable to me” (%) and finally number (%) of highest score in order to assess the ceiling effect [49]. A potential ceiling effect was defined as ≥50% of the participants scoring 5, which would have indicated limited variance and skewed data.

Table 2. PIPEQ: Question descriptions and characteristics.

	N	Missing (%)	Do not know (%)	Mean^a (SD)	Ceiling^b (%)
Q1: In general, how much are you satisfied by the treatment opportunities provided by this institution?	254	0.8	1.6	3.9 (1.2)	37.4
Q2: In general, how helpful was the treatment you received at this institution?	242	1.2	4.7	3.5 (1.4)	23.7
Q3: Has your condition improved or worsened in comparison to when you were admitted?	253	1.6	4.7	4.1 (1.1)	47.9
Q4: In your opinion, has the medical doctor/medical staff prepared you to life after the discharge	248	3.5	5.1	3.5 (1.0)	14.5
Q5: Did you have enough time for discussing your problems with your treating doctor/medical staff	241	3.5	2.7	4.0 (0.9)	32.5
Q6: How did the medical staff understand your situation?	245	4.7	2.5	3.7 (1.1)	27.3
Q7: Did you get the opportunity to tell your doctor /medical staff what was important for your situation?	227	5.9	3.5	3.9 (0.9)	30.6
Q8: In your opinion, did the treatment correspond with your condition?	249	3.1	5.6	3.4 (1.2)	19.7
Q9: Did you have an influence on the choice of treatment?	247	3.9	11.0	3.1 (1.2)	12.3
Q10: Did you have an influence on the choice of medication?	223	5.5	8.3^c	3.0 (1.2)	12.6
Q11: How would you describe information about the treatment possibilities?	250	2.7	10.0	3.6 (1.1)	22.3
Q12: How would you describe the quality of information you got for your illness /diagnosis?	245	4.7	11.9	3.6 (1.2)	26.0
Q13: How was the medical staff cooperating with your relatives?	246	3.9	19.1^d	4.0 (1.1)	38.2
Q14: How would you describe the quality of the activities provided by the institution?	249	3.1	19.8	3.5 (1.3)	21.1
Q15: How do you think now; was the admission to the hospital necessary or not?	250	2.7	3.6	3.9 (1.3)	45.0
Q16: Were the treatment given against your will during your stay at the hospital?	247	3.4	5.3	3.7 (1.4)	43.8
Q17: Considering your plans, you ready for the discharge now?	246	4.3	8.6	3.9 (1.2	38.4

Highlighting of the eight items included in the main dimension.

^aBased on exclusion of answer category ‘Do not know’ and ‘Do not have relatives’; ^b% of a score of 5; ^cDo not use medication; ^dDo not have relatives.

The underlying factor structure of the PIPEQ has been explored in a national sample where one main dimension was identified [28]. As missing data and ‘Do not know’ answers reduced the data set to n ≤ 150, our sample did not meet the general assumption of 10 cases per variable [50], and factor analysis was therefore not performed. We used the main dimension from the national sample consisting of eight items (Q4,Q5,Q6,Q7,Q8,Q9,Q11,Q12), computed a sum score (PIPEQ-8) and used this to explore differences between groups. We calculated the internal consistency of the original 17 Qs of the PIPEQ and PIPEQ-8 using Cronbacs`s α with a satisfaction level at >0.7.

In accordance with the national survey from 2005, we present the numbers of satisfied and dissatisfied by merging the 5 scores into 3 as follows: 1 + 2 (defined as dissatisfied), 4 + 5 (defined as satisfied) and 3 (defined as neutral) [28].

In order to compare potential differences in satisfaction according to legal status and percieved voluntary and involuntary admission, we used the hospital records of IA and VA and the scores YES or NO for the question ‘Were you admitted of own free will’ (Table 1).

The cut-off for defining short and long stay was set at 3 days (mean days of stay at the reception ward + 1 SD).

2.3. Statistical analysis

Descriptive data are represented with mean, standard deviation (SD), and mean differences, with 95% confidence intervals (CI). Questions answered with ‘Do not know’ and ‘Not applicable to me’ are treated as missing data in the analysis. Differences between groups where testet using Pearson’s chi-squared test. Since neither the sum score of PIPEQ-8 nor the individual 17 Qs of the PIPEQ was normally distributed we applied bootstrap replications for statistical tests and confidence intervals in the analysis of difference in scores between diagnostic groups, those of short and long stay and between hospital records of - and those perceiving involuntary - and volunatry admittance. Confidence intervals were calculated using percentile boostrap intervals based on 10,000 bootstrap resamplings. Significant p value was set at p ≤ 0.05 for individual analysis. Bonferroni correction was added when analyzing the difference between the 6 diagnostic groups (0.05/6) and set to p ≤ 0.01 and when analyzing difference for the individual 17 Q (0.05/17) set to p ≤ 0.003. We used SPSS version 25 for statistical anlyses.

3. Results

3.1. Who consented to participate?

Table 1 shows the sociodemographic data of all patients admitted during the study period and the study population of those, consenting and not-consenting. Fifty-three percent (n = 347) of the 650 patients admitted during the study period were asked to participate; of these 74% (n = 256) consented. Reasons for not being asked to participate were not recorded, and the study includes 39% of all patients admitted during the 9-month period. There were no significant differences between those consenting/not consenting regarding age (t = −1.70; p = 0.092), sex (χ² = 0.482; p = 0.328) or VA/IA admissions (χ² = 0.757; p = 0.685). Among those consenting, there was a higher number of patients with a diagnosis within the affective spectrum (F30-39) and fewer with psychosis spectrum diagnosis (F21–29) at discharge (χ² = 22.65, p < 0.001). There were no differences upon admission regarding GAF-F and GAF-S between those consenting/not consenting, but at discharge those consenting had a higher GAF-S (t = 3.319; p = 0.003) and GAF-F (t = 2.537; p = 0.012) (Table 1). Among the 256 participants consenting, 36% were discharged from the reception ward and had a mean stay of 1.5 days. For those referred to further inpatient treatment, the mean day of stay was 26.5 with no difference between those consenting/not consenting. Table 1 shows that the study population is fairly similar compared to the patient population not consenting and for patients admitted during the whole year in terms of diagnoses, voluntary/involuntary treatment and length of stay.

3.2. Results from the PIPEQ

The description of the 17 PIPEQ Qs is displayed in Table 2. The % of missing answers is mostly below 5 and ‘Do not know’ answers range between 1.6 and 11.9%. The mean score range between 4.1 (Q3: have your conditioned improved or worsened) and 3.0 (Q10: influence on choice of medication). No Qs reached the ceiling effect with a score of 5 (most satisfied) ≥ 50%. The mean score of the PIPEQ-8 = 29.45 (SD 4.52) with Chronbachs α = 0.88 with all item-total correlations between 0.86 and 0.87. The Chronbachs α of the original 17 Qs from the PIPEQ = 0.85, with item-total correlations between 0.16 and 0.70. The Qs with low item-total correlation are not included in the PIPEQ-8.

Figure 1 displays the % of satisfied for 17 Qs of the PIPEQ in descending order. The highest % of satisfied was 74 for Q13 (cooperation with relatives) and the lowest was 34 for Q10 (influence on choice of medication).

Figure 1. PIPEQ: Percent satisfied, neutral and dissatisfied with treatment.

Two of the general questions were concerned with information. A total of 54% responded that they had not received information regarding the possibility to make formal complaints about treatment or legally challenge admission and 72% answered that they had not received information about the right to access their medical record.

3.3. Differences in satisfaction between gender, diagnosis and length of stay

We analysed the difference in satisfaction between gender, diagnostic groups and length of stay using the sum score of PIPEQ-8 and for the individual 17 Qs. We found no significant difference in mean score between the genders neither for PIPEQ-8 or any of the 17 Qs. We found a difference between the diagnostic groups [F (5,152) = 4.176; p = 0.001] with post hoc comparison indicating that those with a personality disorder had a lower mean PIPEQ-8 score than all other diagnostic groups [group 5: Personality disorders = 20.57 (CI 13.21–27.93); group 1: Psychotic disorders = 29.05 (CI 27.21–30.90); group 2: Affective disorders = 30.30 (CI 28.98–31.62); group 3: Neurotic disorders = 31.06 (CI 29.95–33.17); group 4:Drug related disorders = 29.11 (CI 26.45–31.76); group 6: Other = 29.61 (CI 26.14–33.08)]. Analysing the individal Qs, we found no difference between the diagnostic groups, but a trend indicating that those with affective disorders (F30-39) had a mean value higher than all others on most Qs and those with a personality disorder (F60-69) the lowest (not shown). We did not find any difference in mean score between short and long stay for the PIPEQ-8 [short stay: 28.46 (CI 26.22–30.57); long stay: 29.74 (28.82–30.66); p = 0.776], but at the individual Q level those of short stay were less satisfied with how helpful they perceived the treatment they received (Q2) [short stay: mean = 3.20 (CI 2.99–3.46) vs. long stay: mean = 3.69 (CI 3.50–3.86); p = 0.002] and if condition improved (Q3) [short stay: mean = 3.75 (CI 3.50–3.98) vs. long stay: mean = 4.29 (CI 4.14–4.44); p < 0.001].

3.4. Difference in satisfaction for VA, IA and perception of admission against own will

As can be seen in Table 1, a total of 45% (n = 113) were referred to IA and 35% (n = 89) continued IA. Out of 246 responses to the question ‘If stay at hospital was of free will’ 70 (28%) answered NO (=admission against own will). After cross-matching with the hospital records, it turned out that 21 of the 70 where in fact VA. The same lack of consistency was found in 24 out of 156 patients answering YES (= admission of free will), when in fact these 24 patients were according to hospital records IA. This mismatch between perception and legal status was found across all diagnostic groups. We created groups according to patients answering YES or NO and patients being VA or IA and compared the level of satisfaction across all items. There was no significant difference in mean score for the PIPEQ-8 between VA and IA [VA: 29.54 (CI 28.079; IA:29.27 (CI 28.08–30.40); p = 0.776] or for those perceiving VA or IA (voluntary: 30.00 (CI 28.91–31.06); involuntary: 28.52 (CI 26.47–30.52); p = 0.213]. We found no differences in any of the Qs when comparing VA and IA, but the patients perceiving themselves as VA (YES) scored higher on Q1, Q6, Q10 and Q11 compared to those who perceived themselves as IA (NO) (Table 3).

Table 3. Significant difference in PIPEQ scores according to either perception of- or legal records of voluntary and involuntary admission.

	Perception of admission			Hospital records
Questions from the PIPEQ	YES	NO	^ap Value	Voluntary admissions (VA)	Involuntary admissions (IA)	^ap Value
	Mean (CI )	Mean (CI)		Mean (CI)	Mean (CI)
Q1: General satisfied treatment	4.09 (3.9–4.3)	3.44 (3.2–3.8)	<0.001	4.02 (3.8–4.2)	3.63 (3.4–3.9)	0.010
Q6: Understanding situation	3.89 (3.7–4.1)	3.26 (2.9–3.6)	0.001	3.81 (3.6–4.0)	3.53 (3.3–3.8)	0.062
Q10: Influence of medication	3.16 (3.0–3.4)	2.46 (2.1–2.7)	<0.001	3.11 (2.9–3.3)	2.79 (2.5–3.1)	0.058
Q11: Information treatment	3.72 (3.5–3.9)	3.15 (2.8–3.5)	0.003	3.66 (3.5–3.9)	3.40 (3.2–3.7)	0.146

YES: admission of own will; NO: admission against own will; CI: 95% confidence interval. ^aBonferroni correction for 17 Q, sign level defined as p ≤ 0.003.

4. Discussion

4.1. Main findings

To the best of our knowledge, this is the first PS survey in ten years exploring different aspects of satisfaction, including all patients admitted to an acute psychiatric ward. Recent studies have been limited to include selected groups [37,39,43,51–53] or comparing service without a general report of satisfaction [43,44]. Sixty-eight percent report that they are generally satisfied with the treatment. Cooperation with relatives was the single one area with the highest satisfaction, in contrast patients were least satisfied with the influence on the choice of treatment and medication. This study also found that certain patient groups are less satisfied than others, in particular those who perceive themselves as being IA, patients with a personality disorder and patients admitted for less than three days. We also found that patients received little information regarding their rights to make a formal complaint and the right to access their own medical records. This study shows that the PIPEQ is a valuable tool in terms of understanding different aspects of patient satisfaction for acute psychiatric patients and has the potential to be an important tool for quality improvement.

4.2. Satisfaction accross time

When comparing our results with the results from the national survey conducted in 2005 [28], we find that general satisfaction (Q1: 68% vs. 54%) and satisfaction regarding cooperation with family (Q13: 74% vs. 48%) have increased. On the other hand, we found no improvement regarding the influence on the choice of medication (Q10: 34% vs. 31%). Involving the family has been an area of priority both at our ward and on a national level [54] and our results may indicate that targeting certain focus areas for improvement increases PS. The recent national focus on ‘Treatment without medication’ may also have the potential to improve patients’ influence [55].

4.3. Areas for improvement

In line with other studies, this study shows that patient’s own perception of legal status differ from the hospital records and that those perceiving IA were less satisfied, regardless of actual legal status [33,34,56,57]. This discrepancy indicates that it is not the legal status per se that causes dissatisfaction, which is an important finding in the discussion of IA for mental health patients. Different aspects have been shown to contribute to the perception of being admitted against own will such as the severity of illness, previous experience, lack of ownership to treatment plans, physical coercion, feeling of disrespect and insufficient information about legal rights [32,34,36,38,58,59]. Others have shown that systematic improvement of information routines increases the patient’s satisfaction with information [38] and our results underline the importance of asking patients about their understanding of the legal status.

Although involuntary admissions clearly challenge human rights, one may argue that perception of admission against own free will is equally important; as the underlying sources to the patient’s misinterpretation of their legal statuses such as disease severity and feeling of disrespect cannot be legally challenged. Results from our study and others highlight the importance of targeting this perception [60].

Patients diagnosed with a personality disorder and those of short stay were less satisfied than others. Patients diagnosed with a personality disorder are a vulnerable group with much suffering [61] who are often discharged or discharge themselves after a few days [62]. Future follow-up studies need to focus on these patients as reduced satisfaction may delay help seeking and adherence towards treatment [63].

4.4. The use of the PIPEQ

We found that the responses to individual Qs of the PIPEQ ranged from satisfied to dissatisfied across different groups indicating that the PIPEQ is able to reflect how patients assess our service. Group differences for the main dimension from the national survey, the PIPEQ-8, supported our findings at the item level and provide validity to our results.

We found high internal consistency for the PIPEQ-8 and also for the 17 Qs of the PIPEQ, with the latter pointing to a possible redundancy of questions [64]. As we were unable to conduct a factor analysis, we were not able to identify the underlying dimensions of the PIPEQ. However, the strong item-total correlation of the PIPEQ-8 supports the finding of one main dimension found in the national sample. Refining the PIPEQ with a reduction of Qs may improve response rate and factor analysis is warranted to understand underlying dimensions of PS in the acute psychiatric setting. Since first developed, the PIPEQ has undergone revisions that should be included in future studies [65].

4.5. Strengths and limitations

The strength of the study is that there were few missing answers, and of those asked 70% consented. Only 39% of all patients admitted were asked to participate which limits the representability of our results. Measures were taken in order to encourage participation by providing written and oral information, reminding the wards regularly of the survey and having the user representative giving motivational talks. However, acute wards are demanding clinical settings, with sudden discharges and a high turnover of staff which altogether makes it challenging to reach all eligible patients. Nonetheless, our sample corresponds to the inclusion rate in other studies [33,38,43,44] and is considered a satisfactory response rate [66]. Also, patients consenting were not different from patients not consenting and they were similar to those admitted during the whole year regarding gender, age, days of stay and legal status. Results may be inflated in a positive direction as more patients with affective disorders than psychotic disorders consented and those consenting had a better functioning as assessed with the GAF-F at discharge. Finally, using the conservative Bonferroni method adjusting for multiple testing when analysing differences between groups may have resulted in Type II errors.

5. Conclusions

Surveys of patients satisfaction provide important knowledge of patients perspective and perceptions that is not easy to gain in the day to day clinical situation. It provides important input for quality improvement that have implications both at the administrative and clinical level. Summarizing the results from our survey, we conclude that information on legal status regarding admission is the number one area for improvement. In addition, increasing patient influence on treatment and medication should be of high priority. User representation in the survey and interpretation of results has been invaluable and provides a more solid validation of the results. Regular surveys at the ward level are encouraged in order to follow-up on improvements.

Acknowledgements

The authors thank all the patients who answered the PIPEQ and a special thanks Kate Storm in preparing the figures.

Disclosure statement

The study was publicly funded and did not receive any financial or other support from the pharmaceutical industry. None of the authors declare any conflict of interest related to the present work.

Additional information

Notes on contributors

Ann Færden

Ann Færde, MD, PhD.

Brita Bølgen

Brita Bølgen, BA Crim.

Lars Løvhaug

Lars Læhaug, RN, PMH.

Christian Thoresen

Christian Thoresen, Cand Psychol, PhD.

Ingrid Dieset

Ingrid Dieset, MD, PhD