Abstract

Summations

• Relational continuity of care (CoC) is much called for in mental health care for patients with serious mental illness (SMI).

• This is a systematic review investigating the associations between measures of CoC and different clinical and social outcomes.

• We found some evidence that having higher relational CoC may be advantageous for care of patients with SMI, especially in association with premature deaths, suicide, emergency department visits and quality of life.

Limitations

• The results are mainly based on studies using observational study designs.

• Meta-analysis was not possible due to large methodological heterogeneity.

• Residual confounding affects the certainty of the evidence and cannot be ruled out completely.

Background

Serious mental illness (SMI) is a term that usually refers to persons diagnosed with schizophrenia, bipolar disorder or major depression [1]. Patients with SMI have a reduced quality of life [2], high rates of comorbidity [3], shortened life expectancy [4], and high rates of unplanned care needs [5]. SMI often goes with substantial functional impairment in domains like employment, housing and social integration.

The wide scope of psychiatric symptoms and social difficulties connected with SMI means that it is important for persons with SMI to have good access to health providers who can give comprehensive treatment and support over time. Interventions for patients with SMI must be based on good relationships between them and staff [6]. Continuity of care (CoC) has for many years been stated as an essential component of health care for patients with SMI [7].

There has been a steady increase in studies about the benefits of CoC. A number of qualitative studies have shown that CoC is highly appreciated by patients [8]. Quantitative studies focused on the association between CoC and clinical outcomes have shown more diverse results [9]. To date, evidence has been mixed on whether relational continuity improves outcomes for people with SMI.

There may be several different reasons for the diversity of research results in this field. One reason may be that the concept CoC is multidimensional and used with different implications, both regarding the process of care and from the perspective of the patients [10]. This has been suggested in theoretically based studies as well as in factor analysis derived studies [11].

Despite the importance placed on providing continuity, its definitions differ [12]. Another reason may be that the methods used for studying the question vary considerably between studies. A third reason may be a variation concerning what parts of the health system that are included in the studies. An example of this is that some studies only study effects on mental health care utilization whereas other studies also cover somatic health care including emergency department visits.

There seems to be consensus that CoC is a construct that is broadly defined as the long-term delivery of care over time which is coordinated among services to appropriately meet patient’s current needs [13]. Continuity is perceived by both the patients and staff as care that is comprehensive, consistent and connected [12]. There is also consensus that CoC is a multi-dimensional construct [14], especially within mental health services that usually involves more complex, integrated and coordinated care pathways [12].

A distinction between three different types of continuity has been a commonly used point of departure for several researchers in this field [14]. The first type called informational continuity focuses on systems for longitudinal information about past events and personal circumstances that may be important for current care givers. The second type called management continuity (sometimes called coordination continuity or treatment continuity) focuses on the ability of the health care system as a whole to provide and coordinate treatment modalities responsive to the changing needs of the patient. The third type called relational continuity (sometimes called interpersonal continuity) focuses on the longitudinal relationship between a patient and one or more health care providers that extends beyond specific episodes of illness [15]. The relationship may be with a few persons or a multidisciplinary team, which is often the case in mental health care. In certain treatment modalities, such as different forms of psychotherapy, continuity with the therapist is of special importance [6]. Relational continuity is often considered a cornerstone for organizing modern health care, which may counteract the risks of fragmentation that may be a consequence of high specialization [13]. Both patients and professionals tend to endorse the importance of CoC, which nevertheless not often is reached for various reasons. It is well known that discontinuity of care is a major source of patient dissatisfaction and disengagement [16, 17].

Since CoC is a multidimensional construct with different focus, a considerable number of instruments have been used in medical research [12]. There are several instruments specifically developed for measuring relational continuity. The following instruments have been used in the studies that were included in this systematic review: Alberta Continuity of Services Scale-Mental Health (ACSS-MH) [18], Continuity of Care Index (COC Index) [19], Continuity of Care – User Measure (CONTINU-UM) [20], Modified Modified Continuity Index (MMCI) [21], Sequential Nature of Provider Continuity (SECON) [22], Usual Provider of Care (UPC) [23]. Some of these instruments were developed within primary or somatic care, whereas others were specifically developed to assess continuity in mental care services [12]. Some of the studies used specific measures, for example the number of staff that a patient has met or the number of unique prescribers of drugs over a specific time period. Common variables have included aspects of relationship duration, density of visits, dispersion of providers, or sequence of providers [23]. Detailed information on the methods used in the reviewed studies is provided in Additional file 5.

The aim of this systematic review was to investigate treatment outcomes, including effects on resource use and costs, associated with receiving higher relational continuity of care for patients with SMI.

Methods

The review was conducted at the Swedish Agency for Health Technology Assessment and Assessment of Social Services, SBU, following a protocol pre-registered on the International prospective register of systematic reviews (PROSPERO: CRD42020196518). Results of the other study population (asthma/COPD) mentioned in the protocol have been reported separately [24]. It was not possible to perform a meta-analysis due to considerable heterogeneity in the included studies. The goal was, thus, to perform a synthesis without meta-analysis to summarize clinical outcomes associated with higher relational continuity of health care for patients with SMI. The certainty of the evidence was assessed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system [25], aiming at being supportive when going from evidence to recommendations [26].

Material

We identified 8 712 unique references that were screened for inclusion of which 56 articles were read in full text. Thirty-seven of these were excluded for different reasons. Nineteen articles fulfilled the inclusion criteria, of which two were later excluded because of an unacceptable risk of bias. Of the 17 included articles [9, 29–44], 16 were based on observational data, mainly retrospective cohort studies, and one was a prospective natural experimental study. Two articles were based on the same study population but reported different outcomes [29, 42]. Of the 17 included articles, eleven were assessed to have moderate risk of bias and six to have high risk of bias. The identification, selection and outcome of risk of bias assessments of included studies is shown in Fig. 1 (flow chart). The risk of bias assessments for included studies and reasons for exclusion of the other studies are shown in Additional files 3 and 4, respectively. A summary of characteristics of the included studies and ratings of evidence are presented in Table 1. Detailed information about the studies included is provided in Additional file 5.

Open in a separate window

Fig. 1

PRISMA flow chart

Table 1

Summary of study characteristics of the included studies. Results and further details are available in Additional file 5

Author Year	Study type Country	Population n	Measure of exposure (continuity)	Outcome(s)	Overall risk of bias
Adair et al. 2005	Prospective cohort study Canada, Alberta	SMI n=486 65% mood disorders 35% psychotic disorders	Alberta Continuity of Services Scale (ACSS-MH)	EQ-5D	High
Adnanes et al. 2019	National cross-sectional survey Norway	SMI or non-SMI patients receiving specialist outpatient psychiatric treatment n=155 SMI n=835 non-SMI	Perception of COC using the CONTINUUM measure	QOL measured with the Manchester Short Assessment of Quality of Life (MANSA) questionnaire	Moderate
Bindman et al. 2000	Prospective cohort study England, south London	SMI n=100 91% psychotic illness	Contact with number of community “keyworkers” over a period of time	BPRS (Brief Psychiatric Rating Scale) HoNOS (Health of the Nation Outcome Score) GAF (Global Assessment of Functioning)	Moderate
Catty et al. 2013	Prospective cohort study England	Psychotic disorders n=180	User rated overall “experienced” continuity using the CONTINUUM measure	BPRS (Brief Psychiatric Rating Scale GAF (Global Assessment of Functioning) Quality of Life - MANSA and SEQoL	High
Chien et al. 2000	Register study USA, Maryland	Schizophrenia n=351	Continuity of care (COC) Usual provider continuity (UPC) Sequential continuity (SECON)	Medicaid costs Lehman Quality of Health	Moderate
Conti et al. 2012	Register study Italy, Lombardy	SMI n=11 797 42% schizophrenia 8% bipolar disorder 50% depression	Continuity of Care defined as receiving at least one psychiatric contact every 90 days	Lack of persistence with pharmacological treatment defined as a gap of at least 30 days between subsequent medication fills	Moderate
Desai et al. 2005	Register study from the VA health system USA	SMI with or without alcoholism, substance abuse or PTSD n=121 933	Continuity of outpatient of care measured as the number of two-month periods in the six months after discharge in which the patient had at least two outpatient visits for his or her primary discharge diagnosis (range=0–3)	Suicide	Moderate
Farley et al. 2011	Register study USA, North Carolina	Schizophrenia n=7 868	Number of unique prescribers of medication for schizophrenia	Adherence to medication measured by the medication possession ratio (MPR) from Medicaid data	Moderate
Giacco et al. 2018	Prospective natural experiment Follow-up 1 year Belgium, England, Germany, Italy, Poland	Psychiatric in-patients 41% psychotic disorders 49% mood disorders 18% other disorders n=7 302	Personal continuity, i.e., a patient is under the care of the same psychiatrist for in- and out- patient treatment; or specialization, i.e., a patient is under the care of different psychiatrists for in- and out-patient treatment.	Readmission to hospital within one year following the index admission	Moderate
Hoertel et al. 2014	Observational study France	Patients who visited a psychiatrist at least twice over six months n=114 515 diagnosis available for n=2 863 17% schizophrenia 29% major depressive disorder 10% bipolar disorder 38% personality disorder	COC Index	All-cause mortality	Moderate
Kaltsidis et al. 2020	Retrospective observational study Canada, Quebec	Patients visiting emergency department for mental health reasons n=210	Exploring predictors for ED visits based on predisposing, enabling and needs factors One of the enabling factors was a regular source of care over the 12 months prior to the research interview at the emergency department	Number of visits to the emergency department for mental health reasons over the 12 months prior to the index visit	Moderate
Macdonald et al. 2019	Register study England, south London	Schizophrenia or delusional disorders 83% schizophrenia 17% delusional disorders n=5 552	Modified Continuity Index (MMCI) measuring the number of teams caring for the patient over time	Health of the Nation Outcomes Scales (HoNOS)	Moderate
Mitton et al. 2005	Observational study Alberta, Canada	SMI n=486 65% mood disorders 35% psychotic disorders	Alberta Continuity of Services Scale (ACSS-MH)	Quality of Life (EQ-5D) Community functioning (Multnomah Community Ability Scale) Administrative data concerning service use events and health care costs	Moderate
Puntis et al. 2016	Prospective cohort study England	Patients with psychosis diagnosis who were to be discharged from compulsory hospital treatment n=323	Average gap between face-to-face contacts Number of 60-day gaps without contact Number of different mental health professions seen Number of care coordinators Number of psychiatrists	Readmission to hospital Time to readmission Number of days in hospital	Moderate
Ride et al. 2019	Observational study England	SMI 53% schizophrenia and other psychoses 35% bipolar disorder 12% both diagnoses n=19 324	Family physician relational continuity measured by: COC Continuity of Care UPC Usual Provider Care SECON Sequential Continuity	Emergency presentations Unplanned admissions for SMI Ambulatory care-sensitive conditions (ACSC)	Moderate
van der Lee et al. 2016	Retrospective register-based cohort study The Netherlands	Schizophrenia n=7 392	The number of follow-up years of elective psychiatric care	Acute treatment events Inpatient care and somatic care Medical costs	High
Watkins et al. 2016	Retrospective cohort study USA	Co-occurring mental illness (schizophrenia, bipolar disorder type 1, PTSD or depression) and substance use disorder n=144 045	Continuous care over time defined as receiving≥1 visit each quarter over a one-year period from any type of provider	Mortality after 12 and 24 months after the end of the observation period (main outcome) Avoidable excessive mortality number	Moderate

Open in a separate window

Notes: COC=Continuity of Care; ED=emergency department; EQ-5D=Euro Quality of Life 5 dimensions

Ten of the studies were performed in Europe (five from England, one each from France, Italy, the Netherlands and Norway, and one using data from four European countries) and seven studies in North America (three from Canada and four from the US). In total, the studies cover around 300 000 patients with SMI. Some of the studies were large register-based studies, whereas others were fairly small clinical studies.

Results

The studies were quite different with respect to exposure as well as outcomes and have used different ways of analysing the material statistically. The outcomes were categorized into summarized outcomes as follows: mortality/suicide (two studies), hospitalizations (four studies), emergency department visits (three studies), costs (two studies), symptoms and functioning (three studies), adherence to pharmacotherapy (two studies) and quality of life (four studies). Summarized results and evidence ratings are presented in Table 2.

Discussion

The results in this systematic review indicate that higher relational continuity of care for patients with SMI may prevent premature deaths and suicide, may lower the number of ED visits, and may contribute to better quality of life compared to patients receiving lower levels of relational continuity of care. The certainty of the evidence was assessed as low, but the results show a high internal consistency with regard to the direction of the effects. The results were consistent across studies performed in different countries and health systems, across different methods used for measuring CoC, as well as across different analytic approaches. The certainty of the results related to risk of hospitalization, costs, symptoms and functioning, and adherence to drug treatment was too low for any reliable conclusions to be drawn. The general conclusion is nevertheless that higher relational continuity of care has been shown to have beneficial effects for patients with severe mental illness, and none of our results indicate an opposite relationship.

This systematic review was restricted to studies published in the English language from year 2000 and onward in international peer reviewed journals. This may include certain shortcomings in the review’s conclusions due to the risk of overlooking studies published in other languages or studies published before 2000. The study protocol included both controlled and observational studies. There were, however, no controlled studies available, which means that the review is entirely based on register or cohort studies, prospective or retrospective. This means that it is difficult to draw reliable conclusions regarding causality between exposure and outcome, which has affected the grading of the certainty of the results. It has therefore been of great importance for this review to consider in detail possible confounders and how these were dealt with by the different authors. Nevertheless, it cannot be ruled out that the results may be somewhat skewed due to residual confounding.

There was a considerable heterogeneity in the design of the studies. Nearly half of them were register studies which may contain some problems when it comes to the analytical choices made for both exposure and outcome. There were also some differences in how the exposure of continuity had been operationalized, measured, and classified in the analyses. This made it difficult to compare the results of the studies and furthermore to perform sound meta-analyses.

Many observational studies based on medical registers have a high quality with respect to coverage and variable definitions, and studies based on register data are most often large size studies on real everyday care. This review included 17 studies judged to be of satisfactory quality and relevant to the research question, and the total population included nearly 300 000 patients, which makes the conclusions reasonably reliable.

In most of the studies included in this review, illness severity and co-morbidity were potential confounders considered to a certain extent, but it cannot be ruled out that residual confounding remains. If all confounders relevant to illness severity and co-morbidity were also included, the effect results would probably be lowered which may be interpreted as a weakening of the association.

Serious mental illnesses are states of unhealth that can lead to major negative consequences in many fields of human life at the same time. The use of pharmaceutical agents is often needed but not always very successful. Some of these illnesses come and go over time, as in bipolar disorder, with quite varying needs in terms of health care, both with respect to quantity and quality.

These characteristics of SMI make it challenging to find ways within mental health care to help these patients in all different areas. SMI is a field where patients’ complex health care needs demand agile and flexible work over time. A relational continuity of care is probably one prerequisite for successful comprehensive health care that can be individually tailored towards each individual patient. This is especially important considering the continuous reorganizations of mental health care where less attention to continuity of care may imply worse clinical outcomes [39, 45].

It is clear that patients with SMI wish for better continuity of care than is the case today.⁷ This systematic review has shown that there is a potential for the development of even better relational continuity. It can therefore be seen as an ethical imperative that the health care system can offer good therapeutic relations that last over time with specific care givers to address the complex needs of patients with SMI.

The state of the art regarding the importance of CoC for patients with SMI gives hope but at the same time demands further development. Given the considerable heterogeneity in research methods, it would be very helpful for further research to establish better consensus with regard to more precise terms and better measures in the area. As patients with severe mental illness have multifaceted care needs and often have contacts with several caregivers, it would also be beneficial for future research to develop methods for studying the importance of relational continuity with teams. This was recently highlighted in a study where the use of care plans improved continuity and clinical outcomes [42].

Furthermore, since most studies so far have been designed as observational studies, future studies based on experimental designs should be attempted to heighten the current level of evidence. This is concluded in several earlier systematic reviews, e.g. in Puntis et al. [13], but is still relevant, as is the need for more consistent assessments of relational continuity of care.

Conclusions

In summary, this systematic review provides some evidence that higher relational continuity of care for persons with SMI may prevent premature deaths and suicide, may lower the number of ED visits and may contribute to better quality of life compared to patients receiving lower levels of relational continuity of care. Even though the certainty of this evidence is assessed as being low, the concordance across studies is high. The finding that no studies indicate an inverse relationship between relational CoC and clinical outcomes is also notable. The heterogeneity of scientific methods is still a large problem in this area and should be addressed going forward through international cooperation in broad multinational studies. As noted in earlier studies we also conclude that there is a lack of good quality randomized controlled studies which in the future could add to the certainty of evidence in the area. Observational studies, especially register studies, however, have some advantages in comparison with experimental studies as they may target larger patient populations and outcomes of “real world” medical practice [46, 47].

Since the certainty of the combined evidence was assessed as low in this review, the evidence-base must be considered as remaining inconclusive. The finding that no studies indicate an inverse relationship between relational CoC and clinical outcomes is, however, notable and should be taken into account when organizing mental health care for patients with SMI. Since SMI is a chronic life-long condition, it is important to tailor the content of the treatment program individually, which includes raising awareness of the importance of relational continuity in healthcare provision.

Acknowledgements

Author contributions

PL JB JL MKE SE ME IE LH LA planned the study. MKE performed literature searches. PL and JB screened abstracts, PL JB SE ME IE LH and LA performed assessments of relevance and risk of bias in pairs, independently. PL JB extracted data, which was verified by IE, LH and LA. PL JB MKE SE ME IE LH LA performed synthesis and evidence gradings of results. IE wrote the first draft of the report with support from PL and JB. All authors read and contributed to the final draft manuscript. All authors had access to all data/assessed studies, and all agreed to the decision of submission. All authors read and approved the final manuscript.

Funding

Open access funding provided by Örebro University. This review was performed within the Swedish HTA organisation SBU – Swedish Agency for Health Technology Assessment and Assessment of Social Services. PL, JB, JL and MKF participated in the work within their positions at SBU while IE, LH, LA, SE and ME participated as external experts in the area and were funded as such by SBU. There was no external funding for the project.

Open access funding provided by Örebro University.

Data Availability

All data generated or analysed during this study are included in this published article and its supplementary information files.

Declarations

Footnotes

References