Self-Reported Needs for Care, Support and Treatment of Persons Who Frequently Visit Psychiatric Emergency Rooms in Sweden

Abstract

Aim: To investigate self-reported needs for care, support and treatment among persons who frequently visit psychiatric emergency rooms (PERs).

Design: A cross-sectional design. Qualitative and quantitative data were collected using an interview-based manual. Qualitative data were analysed using content analysis, whereas quantitative data were analysed using descriptive, non-parametric statistical tests.

Results: Persons who frequently visit PERs self-reported unmet needs for care, support and treatment in life domains such as health, socialisation, daytime activities, and emotional and financial security.

Conclusion: To meet the needs of persons who frequently visit PERs, close cooperation between concerned welfare actors should be implemented.

Introduction

Persons who frequently visit psychiatric emergency rooms (PERs) are a relatively small and heterogeneous group accounting for a disproportionately large number of visits (Chaput & Lebel, 2007; Pasic, Russo, & Roy-Byrne, 2005) and using a high proportion of allocated resources (Ledoux & Minner, 2006; Pasic et al., 2005). However, most studies in this research area cover associations between frequent use of PERs and the sociodemographic and clinical characteristics of the patients.

Adopting a psychiatric, person-centred perspective, emphasising a holistic approach including recovery and empowerment, it has been argued that patients’ self-reported needs for care, support and treatment should play a pivotal role when planning, organising, and providing psychiatric and mental healthcare (Barker, 1999). In the present study, self-reported needs for care, support and treatment among persons who frequently visit PERs are in focus.

Background

It is a well-known phenomenon that there are persons who frequently visit PERs (Schmidt, 2018; cf. Vandyk, Harrison, VanDenKerkhof, Graham, & Ross-White, 2013). Clinical predictors for frequent use of PERs have been found to be personality disorder (Bruffaerts, Sabbe, & Demyttenaere, 2005), schizophrenia (Aagaard, Aagaard, & Buus, 2014; Chaput & Lebel, 2007), substance use disorder (Aagaard et al., 2014; Ledoux & Minner, 2006), and/or anxiety disorder (Ellison, Blum, & Barsky, 1989; Schmidt, Ekstrand, & Bengtsson Tops, 2018). These are all results that may indicate health needs, and needs for accurate medical treatment. Interestingly, prior psychiatric hospitalisation has been shown to be another clinical predicting factor for frequent visits to PERs (Aagaard et al., 2014; Arfken, Zeman, Yeager, Mischel, & Amirsadri, 2002; Pasic et al., 2005). Also, social predictors such as male gender, being single/unmarried, unemployed, homeless, living alone or in sheltered housing (Aagaard et al., 2014; Boyer et al., 2011; Ledoux & Minner, 2006; Pasic et al., 2005; Saarento, Hakko, & Joukamaa, 1998; Sullivan, Bulik, Forman, & Mezzich, 1993) as well as having shortcomings in social interactions (Ledoux & Minner, 2006) and weak social support (Pasic et al., 2005) have been pointed out in previous research. Intriguingly, Aagaard et al. (2014) found—after having interviewed 15 persons who frequently visited PERs—that they viewed staff at PERs as an integrated and valued part of their social network. The same study also found that persons who frequently visit PERs had numerous contacts with other professional healthcare providers (Aagaard et al., 2014).

Yet another reason for the high frequency of PER visits may be the long-term effects of the deinstitutionalisation (Dencker & Gottfries, 1991; Fakhoury & Priebe, 2002), which has resulted in a reduction of psychiatric inpatient services, a downsizing of hospital beds, shorter hospital stays, service gaps between institutional and community services, and a lack of continuity of care between the different care settings (Aagaard et al., 2014; Brown, 2005; Schmidt, 2018; Vandyk et al., 2013). The above findings may indicate that persons who frequently visit PERs have needs for care, support and treatment interventions that could offer them the opportunity to become better socialised.

In a mental healthcare context, the concept of need has been suggested to broadly cover needs of care, support and treatment in various life domains such as health, social integration, private economy and everyday basic needs (Slade, Phelan, & Thornicroft, 1998). One Canadian study (Sirotich, Durbin, & Durbin, 2016), investigating the need profile of persons with mental health problems who frequently visited an emergency department, adopted a broad need perspective by using the Camberwell assessment of need tool (Phelan et al., 1995). In that study (Sirotich, Durbin, & Durbin, 2016), the most prevalent unmet needs were found in the life domains of psychological distress, alcohol/drugs use, money/benefits, and company. However, few studies have—from a broad point of view—investigated self-reported needs in persons who frequently visit PERs.

Aim

The aim of the study was to investigate self-reported needs for care, support and treatment among persons who frequently visit PERs.

Methods

The study has a cross-sectional design. Both qualitative and quantitative data were collected during a 6-month period from December 2015 to May 2016. Frequent visits were defined as five or more visits to the PER during a 12-month period. The definition was based on clinical experience of the staff at the involved PER and previous research within the field (Aagaard et al., 2014; Aagaard, Buus, Wernlund, Foldager, & Merinder, 2016; Schmidt et al., 2018). The PER was a hospital-based unit, located in the south of Sweden, that covered a catchment area of roughly 200,000 inhabitants (Statistics Sweden, 2015) and consisted of both urban and rural areas. The PER had open access 24 h/day, 7 days/week and used an open referral system and the Rapid Emergency Triage and Treatment System (RETTS-psy) for care assessment (SBU, 2010).

Participants

The participants were informed about the study by posters at the PER. Immediately after receiving triage assessment, the patients were orally informed about the study and invited to participate by triage staff including nurses and assistant nurses. Once a patient was asked to participate, a note was made in the registration system. Patients were eligible for participation if they had made at least five PER visits within the last 12 months and if they were aged ≥18 years. Exclusion criteria were if the patients, though fulfilling the inclusion criteria, could not give oral and written consent to participate, came to the PER with police assistance, suffered from acute psychosis or severe aggression, or were intoxicated.

In total, 177 persons frequently visited the PER during the study period, of whom 47 (26.6%) were omitted due to the exclusion criteria. In total, 130 persons were eligible to participate in the study. Eleven declined to participate and 38 were not asked to participate due to a high work load of triage staff. The final sample therefore comprised 81 persons, resulting in a response rate of 62.3%. A combined drop-out analysis with focus on variables age and gender found no significant differences between the participants and non-participants.

Data collection

Data were collected by triage staff in face-to-face structured interviews using a manual developed in collaboration with the third author, service user organisations, and PER professionals. The interview lasted around 30–45 min and all answers were written down by the interviewer. Triage staff received 1 day of training in how to perform interviews according to the manual. Though one triage nurse was given the main responsibility and collected the vast majority of the data (n = 50), in total, 16 triage staff were involved due to the stressful workplace situation.

The first part of the interview manual covered background characteristics such as gender, age, home municipality, and number of PER visits within the last 12 months. The next part of the manual contained the qualitative part of the study, including open questions about reasons for current and previous visits to a PER, and needs with which the PER helped them as well as how many contacts they had with other healthcare and social care services. In this part, the participants’ answers were written down mainly verbatim or at times in condensed meaning units. The third part of the interview manual contained instruments commonly used for assessing needs in persons with mental health problems. This part included the Camberwell Assessment of Need Short Appraisal Schedule (CANSAS) (Ericson, 2013; Phelan et al., 1995). CANSAS is a well-validated instrument (Phelan et al., 1995) for assessing persons’ needs of care, support and treatment within the last month in a psychiatric care context. It consists of 22 predefined need domains and permits to add supplemental domains. For the purpose of this study, the domain “dental care” was added (Bengtsson-Tops & Hansson, 1999). Need is assessed using a three-point rating scale: 0 = no need, 1 = met need, and 2 = unmet need; a rating of 9 was used when the participant did not know or did not want to answer questions in specific domains. The instrument produces three scores: total needs, total met needs, and total unmet needs.

Since CANSAS does not assess exposure to interpersonal violence, which is common among persons with mental health problems (Howard et al., 2010; Oram, Trevillion, Feder, & Howard, 2013) the four following dichotomous questions were added: Have you within the last year been exposed to (1) verbal threats of physical abuse; (2) verbal threats to be killed; (3) physical violence, and (4) sexual abuse/violation.

Furthermore, to collect more detailed data than CANSAS permits with regard to self-reported frequency and amount of alcohol and drug intake, the Swedish version of the Alcohol Use Disorders Identification Test (AUDIT) (Babor, Higgins-Biddle, Saunders, & Monteiro, 2001; Bergman & Källmén, 2002) and the Drug Use Disorders Identification Test (DUDIT) (Berman, Bergman, Palmstierna, & Schlyter, 2003) were used. The scales consist of 10 and 11 questions, respectively, resulting in score ranges of 0–40 and 0–44. AUDIT scores of eight in men and six in women indicate alcohol-related problems (Källmén, Wennberg, Berman, & Bergman, 2007). DUDIT scores of six in men and two in women indicate drug-related problems (Berman, Bergman, Palmstierna, & Schlyter, 2005). Both AUDIT and DUDIT are used in various psychiatric care contexts and have shown good psychometric standards (Hildebrand, 2015; Lundin, Hallgren, Balliu, & Forsell, 2015).

Finally, in order to better elaborate the persons’ self-reported social network than the CANSAS allows, the Interview Schedule for Social Interaction (ISSI) (Eklund, Bengtsson-Tops, & Lindstedt, 2007; Undén & Orth-Gomér, 1989) was used. The participants answered four questions related to the availability of social interaction by choosing one of six options ranging from “I have no-one” to “I have more than 15 persons”.

Analysis

Qualitative data

Eighty-one participants answered the open questions included in the qualitative part of the interview-based manual. The answers resulted in 405 text segments, with ranging segment length from 3–4 words up to 70–80 words. The answers were analysed using manifest content analysis (Graneheim & Lundman, 2004). Initially the first and third author read all the answers several times in order to get a sense of what was said in the interviews. Then the two authors met and discussed the content of the answers. In relation to the aim of the study, statements were extracted and openly coded. The codes were then compared to each other to find differences and similarities, and sorted into clusters. In the next step, the clusters were labelled into subcategories and finally into main categories. This procedure was conducted until the authors reached agreement on the manifest content of the data.

Quantitative data

In CANSAS, ratings of nine were recoded to zero. Thereafter, total met needs per person were calculated by counting the numbers of rating 1 in the different domains. Likewise, the total unmet needs were the sum of rating 2 in the domains. The sum of met and unmet needs was used as the total needs score.

For continuous variables (for example, age and number of visits) and variables with a natural rank order (for example, ISSI), associations between variables were investigated by using the Spearman rank order correlation coefficient (denoted r_s). Chi square tests including Bonferroni correction were used to test for differences in proportions concerning categorical variables. Concerning drop-out analysis, a Chi-square test was used to investigate differences in gender, and a Mann–Whitney U-test to investigate differences in age. Associations and differences were considered significant at p values of ≤.05. Quantitative data were transferred to and analysed in SPSS, version 23 (SPSS Inc., Chicago, IL) (Field, 2013).

Ethics

Ethical approval was obtained from the Ethics Board in Lund (Dnr. 2015/645). An informed consent form was first read out to the participants and thereafter signed by them. The consent included recognition that in the case of self-harm, harm to others, or maltreatment of children, the interviewer was obliged to make a report to the responsible physician or to social services. If participants were exposed to any interpersonal violence, contact information on support services was provided by the interviewer.

Results

Background characteristics

Of the 81 participants, 56.8% (n = 46) were men and most of them, 66.7% (n = 54), lived within the municipality where the PER was located. The mean participant age was 39.74 years (SD ± 15.61, range 20–82 years). The mean of visits was 12.63 (SD ± 14.37) while the median of visits was 8 (5–112).

Regarding additional contacts with healthcare and social care services, 50.6% (n = 41) of the participants reported having contact with social services, 34.6% (n = 28) reported enrolment with an open psychiatric healthcare unit, 21.0% (n = 17) reported contact with a primary care centre, and 9.9% (n = 8) reported having no contact with healthcare or social services. Nearly two-fifth of the participants (n = 30; 37.0%) reported multiple contacts with these services.

Self-reported needs in persons who frequently visit PER, the qualitative part of the study

The manifest content of the qualitative data resulted in the three main categories presented in Table 1.

Table 1. Categories and subcategories of the qualitative content analysis.

Subcategories	Categories
Need to talk to someone	Need to reduce acute suffering
Need for medication
Need for hospital admission
Need for a professional approach	Need to feel secure
Need for PERs to be accessible
	Need to have caring encounters with staff

Need to reduce acute suffering

Having a need to reduce acute suffering included for participants to find relief from troublesome psychiatric symptoms, such as anxiety/panic attacks, suicidal thoughts/actions, substance abuse, depression, self-harm, sleeping problems, and severe stress, and everyday problems that become obvious when trying to handle the current life situation. “Today, I have come to a point where I do not trust myself not to hurt myself impulsively, to prevent internal pain” (woman, 24 years, fifth visit).

The category included the three following subcategories: (1) need to talk to someone, (2) need for medication, and (3) need for hospitalisation.

The “need to talk to someone” was expressed by most of the participants. They needed someone to tell their stories to and to discuss their problems with. Having someone to talk to generated a sense of relief and was expressed in terms of “unburdening my heart” (man, 48 years, fifth visit) or “getting relief from my depression [through supportive counselling from here]” (man, 34 years, 21st visit). The “someone” to talk to could include triage staff, the physician at the PER or other visitors at the unit. Needing to reduce acute suffering also included “need for medication”. It involved the necessity to obtain immediate medication, for example analgesic, hypnotic or sedative medicine or new prescriptions of permanent psychotropic drugs such as neuroleptics and hypnotics. “Need for medication” also included to get individually adjusted medicine dosages of medication such as neuroleptics. Furthermore, needing to reduce acute suffering included “need for hospital admission”. A number of participants wanted to be hospitalised due to their troublesome acute symptoms of mental illness. However, being hospitalised was not always easy: “[I get help with] admission and counseling but usually they send me back home” (woman, 25 years, 20th visit).

Need to feel secure

The need to feel secure involved the sense of being safe with regard to the physical and social environments. The PER was experienced as a safe environment: “There are not that many knives here that you could hurt yourself with” (man, 23 years, 19th visit), or “I can come here instead of sitting alone and brooding; it’s reassuring to know you can land here when your thoughts are their worst” (man, 32 years, fifth visit).

The category included two subcategories: (1) need for a professional approach and (2) need for the PER to be accessible.

The “need for a professional approach” involved being guaranteed privacy and secrecy when talking to staff, and that there would be no interruptions in such conversations. It also involved meeting non-stressed staff with competence in assessing the participants’ needs. The need for a professional approach also included that staff responded by taking them seriously, and being assured that the triage process was fair regardless of what staff members were working. The need for a professional approach was sometimes communicated by pointing out what they were missing, for example: “Don’t feel I’m getting any help. Being unfairly treated by the doctors. Feeling like an experiment” (woman, 49 years, 11th visit) or “The staff are substandard, not all are well suited for working in psychiatry” (man, 27 years, sixth visit). The participants also expressed a “need for the PER to be accessible”. This involved an available PER being located nearby and that the PER was open day and night. It was expressed in terms of knowing “that you can come round the clock” (woman, 21 years, fifth visit) or that “you have somewhere to turn when everything feels hopeless” (man, 65 years, fifth visit).

Need for caring encounters with staff

The need for caring encounters with staff involved being cared for and understood as well as to feel welcomed, and being treated well in terms of kindness, humanity, and fairness: “Getting a smile makes me feel better” (man, 37 years, seventh visit); “They listen and understand. Don’t give you the cold shoulder” (woman, 24 years, fifth visit); or “[You] always feel welcome” (woman, 57 years, 16th visit).

Self-reported needs in persons who frequently visit the PER, the quantitative part of the study

CANSAS

According to CANSAS (23 need domains), the mean number of total needs was 9.52 (SD ± 3.85, range 2–18), met needs 2.89 (SD ± 2.00, range 0–8), and unmet needs 6.63 (SD ± 4.09, range 0–16).

Regarding the figures for total needs score, total met needs score, and total unmet needs score, there were no significant correlations with gender or age. However, a significant positive association between total needs and number of visits was found (r_s = 0.262, p = .018). While a positive significant association was also found between unmet needs and number of visits (r_s = 0.231, p = .038), no significant association between met needs and number of visits was found.

Table 2 presents the numbers of total needs, met needs, and unmet needs according to CANSAS. There are eight domains in which more than half of the participants assessed needs: “psychological distress”, “psychotic symptoms”, “daytime activities”, “intimate relationships”, “company”, “safety of self”, “money”, and “physical health”. The lowest proportions of self-assessed needs were in the domains “telephone”, “safety of others”, and “basic education”.

Table 2. Total needs, met needs, and unmet needs according to CANSAS (n = 81).

Needs	Total needs		Met needs		Unmet needs
	n	%	n	%	n	%
Psychological distress	76	93.8	17	21.0	59	72.8
Psychotic symptoms	63	77.8	27	33.3	36	44.4
Daytime activities	55	67.9	6	7.4	49	60.5
Intimate relationships	47	58.0	12	14.8	35	43.2
Company	46	56.8	12	14.8	34	42.0
Safety of self	46	56.8	8	9.9	38	46.9
Money	45	55.6	20	24.7	25	30.9
Physical health	41	50.6	16	19.8	25	30.9
Food	40	49.4	19	23.5	21	25.9
Dental care	37	45.7	14	17.3	23	28.4
Information about condition and treatment	35	43.2	12	14.8	23	28.4
Household skills	32	39.5	11	13.6	21	25.9
Sexual expression	32	39.5	7	8.6	25	30.9
Public transport	30	37.0	8	9.9	22	27.2
Social benefits	27	33.3	10	12.3	17	21.0
Self-care	25	30.9	11	13.6	14	17.3
Accommodation	24	29.6	7	8.6	17	21.0
Drugs	19	23.4	6	7.4	13	16.0
Alcohol	18	22.2	3	3.7	15	18.5
Child care	18	22.2	4	4.9	14	17.3
Basic education	7	8.6	2	2.5	5	6.2
Safety of others	6	7.4	1	1.2	5	6.2
Telephone	2	2.5	1	1.2	1	1.2

Note: Rounded numbers, thus the sum may differ from 100%.

Regarding met needs, the highest proportions were found in the domains “psychotic symptoms”, “money”, “food”, and “psychological distress”. Concerning unmet needs, the highest proportions were in the domains “psychological distress”, “daytime activities”, “safety of self”, “psychotic symptoms”, “intimate relationships”, and “company”, whereas the fewest met and unmet needs were reported in the domains “telephone”, “safety of others”, and “basic education”.

Besides needs related to symptoms of mental illness, the patients also assessed somatic needs, such as “physical health” and “dental care”, which were reported by 50.6% and 45.7% of the participants, respectively. “Safety of others” was one of the least reported need domains. In contrast, “safety of self” was assessed as a need by the majority and remained unmet in 82.6% of those who assessed it.

A significant association between gender and the CANSAS domain concerning “safety to self” was found (r_s = 0.231, p = .038), indicating that females have more needs or more unmet needs within this domain. Moreover, age was significantly negatively associated with domain “safety to self” (r_s = −0.316, p = .004), i.e., the younger the participants, the more often they experienced need in this domain. There was a significant positive association between age and the domain “dental care” (r_s = 0.226, p = .042). Regarding the number of visits, there were significant positive associations with the domains “accommodation” (r_s = 0.314, p = .004), “money” (r_s = 0.254, p = .022) and “drugs” (r_s = 0.256, p = .021). There was also a significant negative association between the need domain “physical health” and number of visits (r_s =  −0.258, p = .020), i.e., the more needs in this domain, the fewer the visits to the PER.

Interpersonal violence

In total, 41.8% of the participants¹ (n = 33) reported being exposed to some sort of interpersonal violence during the last year. More than a third (n = 29, 36.7%) reported being exposed to verbal threats, of which 22.8% (n = 18) were threats of being killed. More than a fifth reported being exposed to physical violence (n = 17, 21.5%) and 11.4% (n = 9) reported being sexually abused. Concerning the latter, significant differences in gender were found (p = .009), i.e., with one exception, all such reports were from women. Concerning death threats, there was a significant positive association between exposure and number of visits (r_s = 0.289, p = .010).

AUDIT/DUDIT

According to the AUDIT score,² 29.5% (n = 23, 16 men and 7 women) of the participants reported alcohol-related problems. According to the DUDIT score,³ 28.8% (n = 23, 14 men and 9 women) of the participants reported drug-related problems. A significant positive association between AUDIT and DUDIT score was found (r_s = 0.391, p ≤ .001). Moreover, total DUDIT score showed a significant negative association with age (r_s =  −0.275, p = .014), i.e., the younger the participants, the higher the DUDIT score.

Social interactions

Table 3 presents the frequencies of social interactions according to the availability subscale of ISSI. Nearly two-thirds of the participants (65.1%, n = 52) reported that no-one or up to two persons could visit them and feel at home, and more than half of the participants (58.2%, n = 46) reported having no-one or up to two people to whom they could speak openly. Furthermore, 46.3% (n = 37) reported knowing no-one or up to two people who had the same interests, and more than a third (35.1%, n = 28) reported meeting no-one or up to two persons to talk with during a week.

Table 3. Social interactions, %.

Availability of social interactions	No-one	1–2	3–5	6–10	11–15	>15
1. How many people do you know with the same interests as yours?^a	21.3	25.0	33.8	3.8	8.8	7.5
2. How many people do you know who you meet or talk to during a week?^a	6.3	28.8	36.3	18.8	2.5	7.5
3. How many friends, not relatives, can visit you and feel at home?^a	38.8	26.3	25.0	3.8	2.5	3.8
4. To how many, friends and family, can you speak openly?^b	29.1	29.1	26.6	11.4	2.5	1.3

Note: Rounded numbers, thus the sum may differ from 100%.

^a One missing value.

^b Two missing values.

No significant association between gender, or number of PER visits, and each of the four questions about social interaction was found. However, concerning how many friends, not relatives, that could visit the participants and feel at home (Question 3), there was a significant negative association with age (r_s =  −0.221, p = .048), i.e., the older the participants, the fewer friends to visit they had. Regarding how many people the participants could speak openly with (Question 4), there was a significant negative association with age (r_s =  −0.223, p = .048), i.e., the older the participants, the fewer people they had with whom they could speak openly.

Discussion

The aim of the study was to investigate self-reported needs for care, support and treatment among persons who frequently visit PERs.

The results indicate that persons who frequently visit the PER report needs in many life domains, involving physical and psychiatric health problems as well as financial, emotional, and social problems. The results are in line with a recent study in which case managers rated needs in persons with mental health problems who frequently visited emergency rooms (Sirotich et al., 2016). In that study, most of the participants had needs for care, support and treatment with regard to psychological distress, psychotic symptoms, safety to self, and substance abuse. Previous studies within the research area have reported similar results (Aagaard et al., 2014; Bruffaerts et al., 2005; Chaput & Lebel, 2007; Ellison et al., 1989; Ledoux & Minner, 2006). Furthermore, the present study found that the number of visits increased when participants had unmet needs, yet no such association was found with met needs. Those results could indicate the necessary nature of the visits made when needs remained unmet. A qualitative study by Vandyk et al. (2018) found similar results showing that each visit made by persons who suffered from mental health problems and visited the emergency department frequently, was needed, necessary and unavoidable.

In addition, the qualitative part of the present study contributed more in-depth knowledge about the participants’ current needs when attending PERs. They wanted to be relieved from their suffering as soon as possible. They also wanted to feel secure and to be accepted by staff as the suffering persons they were. Given the results from previous studies and especially the results from the present study, persons who frequently visit PER do so because they are in fact in need of care, support and treatment, and not out of a repetitive habit, as suggested by Aagaard et al. (2014).

Previous research has shown that persons who frequently visit PERs have weak social networks (Ledoux & Minner, 2006; Pasic et al., 2005). In the present study, the persons’ social networks were investigated in more detail, revealing that most participants had few available people to interact with in their daily lives. It was notable that the older participants had fewer interactions than did the younger ones. Also a high proportion of the participants reported unmet needs in the domains of company and intimate relationships. Accessible social network and trustful social interactions are of great importance for recovery and empowerment in persons with mental illness (Barker, 1999). It may therefore be suggested that community-based psychiatry services take their responsibility more serious to develop and offer support interventions through which persons who frequently visit PERs can become more socialised and integrated in society. For example, supported socialisation programmes have shown to increase social integration and inclusion as well as the sense of belonging in persons with mental illness (Sheridan et al., 2018).

Furthermore, the present study showed that more than two-fifths of the participants had been exposed to some type of interpersonal violence during the last year. Previous research has found interpersonal violence against persons with mental illness leads to poorer self-esteem, fear, anxiety, and worries, as well as problems in social relations, especially with regard to close relationships and reliance on others (Bengtsson‐Tops & Tops, 2007).

Although most of the participants in the present study reported one or more established contacts within the health and/or social services, high proportions of self-reported unmet needs were found in both the health and social domains. These results are in line with the findings in the study of Aagaard et al. (2014), which reported that persons who frequently visit PERs have numerous contacts with health and social care services. To some extent the results may be explained by shortcomings in the cooperation between the two service systems, and by the lack of an overall policy indicating how to best respond to the needs of these persons. Literature reviews covering interventions targeting persons who frequently visit primary and hospital-based somatic emergency departments have noted that interventions like case management have positive outcomes in terms of improved social and clinical outcomes and reduced visits and costs (Althaus et al., 2011; Soril, Leggett, Lorenzetti, Noseworthy, & Clement, 2015). However, transferring interventions from one context to another may be associated with difficulties, as cultural differences may exist among PERs, primary care centres, and somatic emergency departments. For example, Aagaard et al. (2014) found that persons who frequently visited PERs saw this as a supplementary or alternative solution. It has also been found that PER staff had difficulties in identifying alternative services for persons who frequently visit PERs (Arfken et al., 2002). For functional case management interventions, a key factor is that there is access to services that can meet the person’s needs (Hudon et al., 2017).

Conclusions

It may be concluded that frequent visitors to PERs have comprehensive problems in various life domains and complex intertwined need patterns that are important for leading life on equal terms as the rest of society. The high proportions of self-rated unmet health and social needs combined with their need to reduce acute suffering may reflect a difficult life situation from which they want to be relieved.

This study contributes both empirically and clinically to the research field. It is the first study conducted in a PER context that investigates self-reported needs in persons frequently vising PERs, nationally and internationally. Those results can be used to inform the clinical practice, for example, by adjusting the triage process for this group or suggesting a need-oriented clinical practice approach, i.e., a more person-centred approach when triaging those patients. It also furthers the existing literature by providing evidence of the non-habitual nature of the visits, which contrasts with the findings of Aagaard et al. (2014) in a Danish context, suggesting that context-specific differences need to be taken into account when performing and comparing different studies.

Limitations

The study has several limitations. A number of persons who frequently visited the studied PER (n = 47) had to be excluded due to their condition, even though they represented the target group, which raises questions about the participants’ and the study’s representativeness. Other eligible patients were not invited to participate (n = 38) due to the high workload of the triage staff, and 11 patients declined participation. However, the drop-out analysis found no differences in age and gender between those persons included and those not included in the study. Despite the number of participants who were not invited to participate, the response rate was considered high, which strengthens the validity of the study. That the data were collected in one PER might threaten the inference transferability of the study; however, the inference quality of this study was strengthened by the use of both qualitative and quantitative data. Furthermore, there may be a risk of lack of subjectivity with regard to AUDIT and DUDIT where the results were interpreted by the authors in line with the guidelines of the screening tools, while those participants identified as having drug and alcohol-related problems may not experience such problems themselves. To enhance the trustworthiness of the qualitative part of the study, the first and third authors analysed the qualitative data independently, and thereafter together. Quotations were used to validate the findings. For future research, a larger sample size could be considered, given the limited number of participants in this study.

Relevance to clinical practice

To more effectively meet the needs of persons who frequently visit PERs and thus reduce the repetitive pattern of their visits, it would be advisable to introduce a person-centred approach to their care, involving an overall policy of cooperating with other services involved in their life situation. Furthermore, it may be useful for clinical practice to use assessments tools that cover problems in basic health, social, emotional, and financial life domains. By doing so, a more holistic caring approach will be applied.

Acknowledgements

We thank the participants and staff from the psychiatric emergency room where the data were collected.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Research Platform for Collaboration for Health (Kristianstad University) and by Kristianstad University.

Notes

1 Two missing values.

2 Three missing values.

3 One missing value.