Hospital staff at most psychiatric clinics in Stockholm experience that patients who self-harm have too long hospital stays, with ensuing detrimental effects

Abstract

Background

Previous research on patients who self-harm has indicated potential negative effects from long hospital stays. Yet, such care has been reported to occur regularly. We conducted this questionnaire study to investigate how hospital staff, who treat self-harming patients, experience the relation between lengths of stay and self-harm behaviour, and the motives for non-beneficial hospital stays.

Methods

The respondents of the questionnaire were nurses and mental health workers employed at public inpatient wards in Stockholm, treating patients who self-harm. The questionnaire contained questions with fixed answers and room for comments. A total of 304 questionnaires were distributed to 13 wards at five clinics, and the response rate was 63%. The data were analysed with descriptive statistics and qualitative descriptive content analysis.

Results

The results show that most staff experienced that more than a week’s stay either increased (57%) or had no effect (33%) on self-harm behaviour. Most respondents at most clinics considered the stays to be too long at their wards, and that the stays could be reduced. The respondents recognized several reasons for non-beneficial hospital stays, like fear of suicidal behaviour and doctors’ fear of complaints. Patients appearing as demanding or fragile were thought to be given more care than others. The respondents’ comments confirmed the majority’s experience of detrimental effects from longer hospital stays.

Conclusions

A majority of the health care staff experienced that patients who self-harm often receive too long hospital stays, with detrimental effects, and they had experienced several non-medical reasons for such care.

Keywords:

• Self-harm
• borderline personality disorder
• medical ethics
• inpatient care
• suicide prevention

Background

In 2019, the psychiatric leadership board of the county of Stockholm (public psychiatric care of Region Stockholm) put together a team of psychiatric professionals with the assignment to investigate how inpatient treatment of patients who self-harm could be improved and become more uniform within the Region. In order to map the current status at public inpatient wards in Stockholm, which treat patients who self-harm, a questionnaire was addressed to hospital staff at all relevant public wards in the county of Stockholm. In Stockholm, the vast majority of inpatient care is conducted at public clinics, partly due to tradition and partly because compulsory care is not allowed at private clinics, and these public inpatient clinics serve a population of about 2.3 million inhabitants [1].

Self-harm is an umbrella term which is described by NICE guidelines as any act of self-poisoning or self-injury, irrespective of the patient’s motivation [2], a definition we follow/apply in this article even though some studies we refer to include only non-suicidal self-injuries (a term more commonly used in the USA). In clinical practise, patients usually referred to as suffering from self-harm behaviour have underlying problems from young age with handling emotional regulation and interpersonal relations and have developed dysfunctional coping strategies such as chronically recurring suicidal/parasuicidal/self-destructive behaviour in response to inner and outer stressors [3–5]. This group of patients is targeted in the present questionnaire study.

Since psychiatric diagnostics can differ quite a lot depending on the caregiver’s own views and interpretations of diagnoses [6,7], we have chosen to describe the behaviour to the respondents, rather than limiting the study to one specific diagnosis. However, diagnoses such as borderline personality disorder [8] and neuropsychiatric disorders are common for patients who self-harm [9–11], and the patients treated in hospital for self-harm are mostly young women [12–14]. Most previous studies on patients who self-harm have been on patients with borderline personality disorder – the only disorder in DSM-V where self-harm is one of the inclusion criteria – and treatment for self-harm is often the same as for borderline personality disorder, like dialectic behavioural therapy (DBT) [15,16].

As with patients with borderline personality disorder, patients who more generally fall under the concept of self-harm do not seem to benefit from longer hospital stays. Much clinical experience and several studies point to iatrogenic effects of inpatient care; the self-harm behaviour seems to become more consolidated or reinforced in the inpatient care setting [17]. Restrictions in the patients’ care environment and freedom of movement have not shown to reduce self-harm behaviour and have been positively correlated with increase of self-harm behaviour during inpatient care [18,19]. Iatrogenic effects from inpatient care could partly be explained by contagion of self-harm behaviour at wards and interpersonal conflicts between patients and mental health workers [20,21]. When it comes to borderline personality disorder, there is quite a large body of experience indicating that inpatient care is ineffective and potentially harmful when used as a measure against self-harm behaviour, including suicidality [15,22,23]. Instead, clinical recommendations, such as NICE guidelines, focus on strengthening the patient’s autonomy, providing psychological treatments and outpatient care [24].

Aim

The aim of this study was to investigate the health care staff’s experiences concerning inpatient care of patients who self-harm, including lengths of stay, effects on self-harm behaviour, whether the hospital stays can be reduced, the motives for non-beneficial hospital stays, and whether the patients’ appearance could affect the amount of care given.

Method

Population

In autumn 2019, a questionnaire survey was sent to nurses and other mental health care workers at public inpatient clinics within the municipality of Stockholm (Region Stockholm). All regular staff who was on duty during a period of about two weeks received the questionnaire. The wards involved in the study treat patients who suffer from self-harm behaviour, including frequently recurring suicidal actions. The study was meant to include all public wards treating the patients of interest, which were 14 wards. In total, 13 wards from five different clinics in Stockholm participated in the study. One smaller ward, with 15 potential respondents, chose not to participate due to lack of interest. The ward that chose not to participate is a mixed ward for patients with need for longer hospital stays for social planning. Three hundred and four questionnaires (the number of staff which was reported to be on regular duty during the period of the study) were distributed to 13 different wards at five different clinics, and 192 were answered, leaving a response rate of 63%.

Survey questions

The questions concerned whether patients with self-harm behaviour, including chronically fluctuating suicidality, benefit from hospital stays longer than a week, whether the patients are treated too long or short at the relevant ward, whether the hospital stays could be shortened at the relevant ward without lowering quality of care, if there are non-medical reasons for prolonged hospital stays, and whether the patients’ personal behaviour and appearance could affect the amount of care given. Each question had fixed response alternatives. There was room for comments to each question. See Supplementary Appendix I for a translated version of the questionnaire.

Data analysis

The data from the fixed response alternatives was collected in an Excel file and analysed with descriptive statistics for categorical data. For the proportion p and number of respondents n, it was calculated that np > 5 and n(1-p)>5, which is the premise for assuming a normal distribution of data. Therefore, a normal distribution of data could be assumed and a 95% confidence interval (CI) was calculated.

The respondents’ comments were analysed using qualitative descriptive content analysis to extract subcategories, categories, and themes [25]. First, the comments were read repeatedly to get an overall impression of the content. Next, meaning units and phrases expressing thoughts relating to the overall research questions were identified. Meaning units expressing similar ideas were sorted into subcategories. Related subcategories were then reduced into categories. Finally, related categories were synthesized into overarching themes [25–27]. The analysis was made inductively, with no predetermined categories.

Results

The results show that 57% (n = 188, CI 50–64%) of staff considered that more than a week’s stay had detrimental effects on self-harm/suicidality for the patients who self-harm, while 33% (CI 26–39%) experienced no effects and 11% (CI 6–15%) thought that the patients got better from their self-harm/suicidality from more than a week’s stay (Figure 1). The majority at all clinics but one considered the hospital stays to be too long at their wards (72%, n = 141, CI 65–80%), and thought that the stays could be reduced without infringing on care quality (80%, n = 134, CI 73–87%) (see Figure 2). The only clinic where the majority of respondents were satisfied with the lengths of stay (78%), and did not think they could be decreased (70%), has a working method with short hospital stays that are less than a week (3–5 days) for the relevant patient group.

Figure 1. Respondents from all clinics, n = 188.

Figure 2. Respondents from all clinics except one (which practiced hospital stays which were less than a week long); n = 141.

The respondents recognized several reasons for prolonged stays, like fear of suicidal behaviour, and also clearly non-medical reasons like doctors’ fear of complaints and lack of housing for the patients (Figure 3). The results further show that patients’ behaviour and appearance could affect the amount of care they were provided with – demanding patients and patients who appeared sensitive or fragile were thought to receive more care than others (Table 1).

Figure 3. Respondents from all clinics, n = 188.

Table 1. Respondents from all clinics.

A. If the patient is perceived as a demanding person, then	Percentage of respondents confirming the statement, CI 95%, n = 186
1) the patient usually receives more care interventions	56 (CI 49–63)
2) it has no effect on care interventions	34 (CI 27–41)
3) the patient usually receives less care interventions	10 (CI 6–14)
B. If the patient is perceived to have an attractive appearance, then	Percentage of respondents confirming the statement, CI 95%, n = 188
1) the patient usually receives more care interventions	13 (CI 8–17)
2) it has no effect on care interventions	85 (CI 80–90)
3) the patient usually receives less care interventions	2 (CI 0–5)
C. If the patient is perceived as fragile/sensitive, then	Percentage of respondents confirming the statement, CI 95%, n = 188
1) the patient usually receives more care interventions	56 (CI 49–63)
2) it has no effect on care interventions	36 (CI 29–43)
3) the patient usually receives less care interventions	7 (CI 4–11)
D. If the patient is perceived as an unpleasant person, then	Percentage of respondents confirming the statement, CI 95%, n = 174
1) the patient usually receives more care interventions	11 (CI 7–16)
2) it has no effect on care interventions	59 (CI 52–66)
3) the patient usually receives less care interventions	30 (CI 23–36)

The respondents’ comments (Table 2) confirmed the majority’s experience of detrimental effects from longer hospital stays, such as hospitalization and increase of self-harm behaviour. Many respondents saw a need for more structure during inpatient care, with the aim to focus on the patient’s own responsibility and active participation, shorter hospital stays for the patients, non-medical interventions, and better cooperation with the outpatient clinic. In addition, medically irrelevant factors were observed to affect how much inpatient care was given.

Table 2. Qualitative analysis of the respondents’ comments.

Themes	Categories	Subcategories
There are mostly negative effects from longer hospital stays and inpatient care does not reduce self-harm behaviour	Longer hospital stays (more than a week) entails hospitalization, increase of self-harm and other negative effects, from but there can be exceptions	Inpatient care can make the patients hospitalized, regressive, taking on an illness identity, gradually feel worse and lose their abilities to function in everyday life, why shorter stays of a few days are better than longer stays
		Self-harm is not reduced and sometimes increases during inpatient care and is contagious between patients at the ward
		Longer hospital stays can be beneficial in single cases or when trying out new treatments
Inpatient care needs more structure, with focus on short hospital stays, the patient’s active participation, outpatient care and non-pharmaceutical interventions	There is a need for more structured inpatient care content, with specified goals, expectations and limitations	Need to emphasize the patient’s own responsibility to participate constructively in the care (for example, by care contracts), need to draw up a care plan at admission, have a fixed discharge date to keep the hospital stay short, and need for the staff to work uniformly
	Need for increased continuity in the care offered, better cooperation between caregivers, and more education on management of self-harm to inpatient staff	Need for increased continuity of caregivers, more outpatient care, better cooperation between inpatient and outpatient care and between social and medical authorities
		Need for more inpatient competence and practise of non-medical techniques in order to help the patients adequately
Fear of suicidal behaviour and several medically irrelevant considerations affect the amount of care offered	Problems with housing, fear of suicidal behaviour after discharge, and detrimental effects of hospital stay, are factors that lead to prolonged hospital stays	Problems to find proper housing for the patient leads to prolonged hospital stays
		Prolonged hospital stays due to fear of suicidal behaviour, increase of suicidal communication at discharge, and hospitalization
		Lack of regular doctors at the ward lead to prolonged hospital stays – temporary doctors don’t dare to discharge the patient
	The patient’s ability to attract attention, evoke sympathy, be compliant, etc., affects the amount of offered care and the outcome of the care given	Patients perceived as demanding get more attention and care than other patients, while quiet patients get less attention and can therefore get less care
		Attractiveness can make the staff misjudge the patient’s psychiatric abilities
		The patients’ willingness to cooperate and if they raise sympathy or not, can affect the outcome and amount of care

Discussion

Too long hospital stays with detrimental effects

Our results show that staff at most psychiatric inpatient clinics in the Stockholm area find the hospital stays to be too long for patients who self-harm, and experience detrimental effects for the patients due to this fact. Such aspects of care are problematic not only because of the negative effects for the majority of patients who self-harm, but also because extensive hospital resources are used for patients who are not perceived to benefit from such care. There is reason to believe that the respondents’ observation of detrimental effects from longer hospital stays is correct, since it is buttressed by previous research done on inpatient care for patients who self-harm, when used for suicide-protective reasons [15,17,22,23,28].

Several reasons for non-beneficial hospital stays

The respondents recognised several reasons for prolonged non-beneficial hospital stays for patients who self-harm; for example, doctors’ fear of litigation/complaints, fear of suicidal behaviour after discharge and social factors like the patients’ lack of housing.

Some of these reasons can seem paradoxical, since hospital care for these patients has no proven effect against self-harm behaviour, which includes suicidality, and the hospital care itself can even increase such behaviour [15,17,22,23,28,29]. The reasons for these detrimental effects of inpatient care can be several. As mentioned previously, conflicts and contagion of self-harm behaviour at the wards could be partial explanations [20,21]. Another explanation could be that inpatient care works as an avoidant coping mechanism for the patient, by temporarily alleviating the patient’s inner pain by conferring the responsibility for problem-solving and emotional regulation to others [28]. However, avoidant coping mechanisms, which can consist of behaviours such as self-harm or seeking inpatient care to escape emotional distress, are very short-term solutions and risk increasing the inner pain in the long term through a rebound effect, risk increasing the use of the avoidant coping mechanisms over time through reinforcement, and withhold the patient from learning more constructive methods to handle distress and adversities on her own [28]. This can be compared to addiction, where the drug temporarily removes inner discomfort but increases anxiety over time and increases the need for higher doses of the drug to alleviate the discomfort [28]. Also, the self-harm behaviour can become directly positively reinforced by inpatient care, by giving the patient extra care and affirmation as a response to her suicidality [28].

Drawing on the above, prolonged inpatient care for patients who self-harm can become counterproductive and is advised against in several clinical guidelines [17,24]. Considering these potentially detrimental effects of inpatient care, one may question the overall adequacy of using inpatient care as a solution, for instance, when the patient lacks housing or is discontented with current housing.

When it comes to fear of suicidal behaviour after discharge and caregivers’ fear of litigation, one may speculate if the widespread use of suicide risk assessments is a partial explanation. Suicide risk assessment is often mandatorily used in psychiatry to predict and prevent suicides. However, suicide risk assessment has such low positive predictive value and limited sensitivity that the clinical value is limited in individual cases and does not justify highly interfering interventions such as admission to inpatient care [30–32]. Also, admitting a patient assessed as having high suicide risk to inpatient care has not shown to reduce the incidence of suicides over time and admission to hospital itself has been argued to possibly play a causal role in a proportion of in-patient suicides [32,33].

These unrealistic expectations from society when it comes to psychiatrists’ ability to predict and prevent suicides may explain why many psychiatrists experience work-related anxiety related to handling suicidality, sometimes to the extent that it affects their clinical practise negatively [34,35]. As a consequence, in order to avoid criticism for negative outcomes which one cannot predict with enough certainty, it may be tempting for psychiatrists to resort to restricting the patients’ autonomy in an attempt to control their suicidality, for example by liberal use of hospital care as a suicide protective measure – even if such measures are poorly supported when it comes to preventing suicides in general [28,33], and for preventing suicides among patients with self-harm behaviour in particular [15,17,22,23,28,29].

Drawing from own clinical experience, extensive use of medical interventions, inpatient care and compulsory care rarely meets criticism from either colleagues, investigative authorities or relatives to patients who self-harm. The opposite, however, is often questioned and met with scepticism. The rule of thumb at the psychiatric emergency unit is often “it is better to compulsorily admit too many than too few”. This further suggests a strongly held belief among psychiatrists that hospital care, medicalisation and compulsory care are life-saving measures – even if most experience supports the opposite when it comes to patients who self-harm – and that such measures will keep the doctor safe from litigations and complaints (which probably is correct but an unreasonable justification for these measures).

The inclination to detain patients who self-harm is also questionable from a legal aspect. To detain a patient under the Mental Health Act in Sweden, one of the prerequisites is that the patient must suffer from a “severe psychiatric disorder”. Self-harm behaviour or personality disorder by themselves do not qualify as “severe psychiatric disorders” from a legal point of view, unless combined with a psychotic state, severe depression, mania, confusion or other mental health states of similar severity (Supplementary Appendix II, [36,37]).

Suggestions on better inpatient care structure

The respondents recommend several tangible changes in working methods at the wards, like giving the patients care plans with fixed discharge dates, short-term care (a few days), and information to the patients what is expected of them during inpatient care. Several of the suggested changes are in line with NICE recommendations for borderline personality disorder [24].

Assertiveness or fragile appearance may render more psychiatric care

The way the patients’ own behaviour and appearance affect the amount and quality of care is rarely mentioned in the context of inpatient psychiatric care, even if interpersonal interactions are known to affect people’s feelings and willingness to help each other. However, as this study shows, such factors are probably not without importance – even if it does not answer to what extent those factors matter. This subject should be further investigated, since the results imply that care is not always prioritized according to patients’ needs and best interests, but according to factors that should not matter when deciding on care – like the patient’s assertiveness and ability to attract attention and evoke sympathy.

These findings can be understood from normal psychological mechanisms. For example, a fragile appearance suggests that the patient has less abilities than others and needs more help. However, the reasons for the patient’s fragile appearance can be more complex than that. In borderline personality disorder, active passivity and regressive behaviour are common features [7,15,38], suggesting that the patient can appear as less able than she actually is. Reinforcing such traits by taking over agency from the patient, could have negative consequences for the patient’s ability to manage her own emotions and adversities in the future [24].

One can easily picture how assertive patients can elicit different responses among caregivers. One conceivable consequence is that the demanding patient makes the caregiver annoyed and reluctant to comply with the patient’s demands. However, most respondents thought that demanding patients received more care than others, suggesting that the caregivers prefer to comply with the patients’ demands. There is some weak support for the occurrence of what could be called the “decibel factor”, that is, that demanding patients get more attention in general (those with the loudest mouths get most attention) [39]. In such case, the motives for the caregiver could be to minimize the risk of complaints and to avoid potential conflicts.

The results above can implicate medically unsupported extra care for assertive or fragile-looking patients, entailing a more unequal care and in some cases a risk of inadvertently decreasing the patients’ abilities to take care of themselves in the future. An increased awareness with the medical staff on such phenomena could help prevent them from occurring.

Strengths and limitations

A strength of the study is the, by today’s standards, relative high response rate (63%).

This study was confined to the municipality of Stockholm, which is a limiting factor when it comes to generalisability of the results. However, we have reason to believe that the problems with unequally distributed long hospital stays with detrimental results for patients who self-harm, which are described in this study, also occur in other parts of Sweden. This assumption is based on our own clinical experience and unpublished results, previous studies, and current statistics on compulsory care in Sweden [7,17,40,41].

Conclusions

This questionnaire study confirms previous research done on patients who self-harm, indicating that that inpatient stays for more than a week often have detrimental effects on self-harm/suicidal behaviour. Yet, the majority of respondents experience that patients who self-harm receive too long stays at their wards. The motives reported for such non-beneficial hospital stays range from doctors’ fear of complaints to patients’ lack of housing. Also, the patients’ own behaviour and appearance can affect the amount of care offered. To improve the inpatient care, the respondents suggest a more structured inpatient care, including care plans with fixed discharge dates to keep the hospital stays short, better cooperation with the outpatient clinic and more non-medical interventions during inpatient stays.

Ethics approval and consent to participate

According to 3–6 §§ and 13 § of the Swedish Ethical Review Act concerning research on human subjects [42], an ethics approval and an informed consent is needed if the research concerns sensitive personal data, personal data concerning delinquency, physical interventions on research subjects, is performed with potentially harmful methods, concerns biological material from humans or physical interventions on deceased humans. Since this study concerned the healthcare staff’s experiences regarding compulsory treatment of BPD patients, including no sensitive personal data, and no patients were involved, the project was not of sensitive nature and did not need to be ethically reviewed according to the Swedish Ethical Review Act [42]. All participants in this study were informed in the cover letter that their participation was anonymous and voluntary (see Appendix I). We confirm that all methods were carried out in accordance with relevant guidelines and regulations.

Acknowledgements

The authors acknowledge the psychiatric leadership board of the county of Stockholm, which prompted this work in order to improve the psychiatric inpatient care in Stockholm. The authors acknowledge all healthcare staff who accepted to participate in the study. Finally, the authors thank all the unit managers at the wards, who accepted their wards to participate in the study and helped with handing out and collecting the anonymous questionnaires.

Disclosure statement

The authors declare that they have no competing interests.

Data availability statement

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Additional information

Notes on contributors

Antoinette Lundahl

Antoinette Lundahl, M.D., works at Norra Stockholms Psykiatri, S:t Görans hospital, as a senior consultant in psychiatry at a ward for patients with anxiety- and personality disorders. She is also a PhD student at Stockholm Centre for Healthcare Ethics at Karolinska Institutet.

Gert Helgesson

Gert Helgesson is a professor in medical ethics at the Department of Learning, Informatics, Management and Ethics (LIME) at Karolinska Institutet, Stockholm.

Niklas Juth

Niklas Juth is an associate professor at LIME, Karolinska Institutet. Helgesson and Juth have prior experience of the methods used in this study.