Hostname: page-component-586b7cd67f-tf8b9 Total loading time: 0 Render date: 2024-11-24T18:27:54.588Z Has data issue: false hasContentIssue false

Association between the timing of consultation-liaison psychiatry interventions and the length of stay in general hospital

Published online by Cambridge University Press:  13 November 2019

Hélène Vulser*
Affiliation:
Doctor, Paris Descartes Faculty of Medicine, University of Paris; and Department of Psychiatry, European Georges-Pompidou Hospital, AP-HP.Centre, Université de Paris, France
Victoire Vinant
Affiliation:
Doctor, Department of Psychiatry, European Georges-Pompidou Hospital, AP-HP.Centre, Université de Paris, France
Victoria Lanvin
Affiliation:
Doctor, Department of Psychiatry, European Georges-Pompidou Hospital, AP-HP.Centre, Université de Paris, France
Gilles Chatellier
Affiliation:
Professor, Paris Descartes Faculty of Medicine, University of Paris; and Department of Medical Informatics, Biostatistics and Public Health Department, European Georges-Pompidou Hospital, AP-HP.Centre, Université de Paris, France
Frédéric Limosin
Affiliation:
Professor, Paris Descartes Faculty of Medicine, University of Paris; Department of Psychiatry, European Georges-Pompidou Hospital, AP-HP.Centre, Université de Paris; and Inserm U1266, Institute of Psychiatry and Neuroscience of Paris, France
Cédric Lemogne
Affiliation:
Professor, Paris Descartes Faculty of Medicine, University of Paris; Department of Psychiatry, European Georges-Pompidou Hospital, AP-HP.Centre, Université de Paris; and Inserm U1266, Institute of Psychiatry and Neuroscience of Paris, France
*
Correspondence: Hélène Vulser, Department of Cognitive and Clinical Neuroscience, Central Institute of Mental Health, J5, 68159 Mannheim, Germany. Email: [email protected]
Rights & Permissions [Opens in a new window]

Abstract

Background

Psychiatric comorbidities are frequent in patients admitted in general hospital and are associated with greater lengths of stay (LOS). Early consultation-liaison psychiatry (CLP) interventions may reduce the LOS but previous studies were underpowered to allow subgroup analyses and have generally not considered the severity of the condition for which patients were admitted (‘disease severity’).

Aims

To investigate the association between the timing of CLP interventions and LOS in a general hospital.

Method

We retrospectively included 4500 consecutive patients admitted in non-psychiatric wards of a university hospital between 2008 and 2016 who had a first CLP intervention. We used general linear models to examine the association between the referral time, defined as log(days before the consultation)/log(LOS), and log(LOS), adjusting for age, gender, year of admission, place of residence, main psychiatric diagnosis, admission to the intensive care unit (ICU), main physical condition and disease severity.

Results

Referral time was associated with log(LOS) (β = 0.31; P <0.001), notably for older patients (β = 0.43; P <0.001) and those admitted to the ICU (β = 0.50; P <0.001), but not for those with psychotic disorders (β = −0.20; P = 0.10). The association was confirmed when considering the expected LOS for each patient. For instance, for an expected LOS of 10 days, a CLP intervention on day 3 compared with day 6 was associated with a reduction of the actual LOS of 2.4 days.

Conclusions

Earlier CLP interventions were associated with a clinically significant shorter LOS in a large population even after adjusting for disease severity. Early CLP interventions may have benefits for both patients and health-related costs.

Type
Paper
Copyright
Copyright © The Authors 2019

Comorbid psychiatric disorders are frequent in patients admitted to general hospitals, with an estimated prevalence around 30%.Reference Hansen, Fink, Frydenberg, Oxhøj, Søndergaard and Munk-Jørgensen1 Several studies have observed that patients with psychiatric comorbidities had a longer length of stay (LOS).Reference Bourgeois, Kremen, Servis, Wegelin and Hales2,Reference Levenson, Hamen and Rossiten3 The costs of hospital admissions have come under high scrutiny and a growing interest has been shown in demonstrating the effectiveness of consultation-liaison psychiatric (CLP) interventions. Thus, beyond their clinical impact on patients, CLP interventions might convey a significant economic benefit by attenuating this increase in LOS. As early as 1941, Billings et al Reference Billings, McNary and Rees4 reported that 5 years after setting up a CLP service at the University of Colorado Hospital, the hospital LOS for medical in-patients with psychiatric comorbidities was reduced from 28 to 15 days. Since then, some studies have investigated the relationship between CLP interventions and shorter LOS. In 1981 and 1991, two studies investigated the LOS of elderly patients with a hip fracture and found that CLP interventions were associated with a reduced LOS.Reference Levitan and Kornfeld5,Reference Strain, Lyons, Hammer, Fahs, Lebovits and Paddison6 In 1992, a first randomised controlled trial addressed this question, but no evidence for a reduction in LOS was found.Reference Levenson, Hamer and Rossiter7 More recently, some evidence of cost-effectiveness for proactive CLP interventions was found in observational and quasi-interventional studies.Reference Desan, Zimbrean, Weinstein, Bozzo and Sledge8,Reference Sledge, Gueorguieva, Desan, Bozzo, Dorset and Lee9 The need to conduct studies with robust methodology that would also reflect the real-life operation of CLP servicesReference Wood and Wand10 led to another approach, consisting in investigating the association between the timing of CLP interventions and the LOS. If CLP interventions do indeed reduce the LOS, then having a psychiatric consultation earlier in the hospital stay may have a greater impact than having the consultation later in the stay.Reference Lyons, Hammer, Strain and Fulop11 Studies with different approaches have suggested that early psychiatric consultation is associated with early discharge.Reference Wood and Wand10Reference Sockalingam, Alzahrani, Meaney, Styra, Tan and Hawa16 However, study samples were too small to allow subgroup analyses and the majority did not take into account the severity of the condition for which patients were admitted (‘disease severity’), which is a critical determinant of the LOS. This large-scale study aimed to confirm the association between the timing of CLP intervention and the LOS in a large population of in-patients of a general university hospital, to investigate the contribution of the disease severity to this association, to explore the extent to which this association may vary across subgroups and to estimate the clinical significance of such an association.

Method

Participants

The study was conducted in a 712-bed acute metropolitan general university hospital in Paris, France (the Georges-Pompidou European Hospital). All consecutive in-patients referred for the first time to a consultation-liaison psychiatrist between 1 January 2008 and 31 December 2016 were included in the study. Patients were not included if they had previously met a consultation-liaison psychiatrist in the hospital, because of a difference in referral pathway, as psychiatrists may decide to plan a consultation even if not asked by the medical team in charge of the patient. For each patient, the main psychiatric diagnosis was added to their computerised medical record by the psychiatrist after the consultation.

Eligibility criteria were: ≥18 years of age, living in France, having a main psychiatric diagnosis according to ICD-10 codes F0–F6 made during the hospital admission (F0, Organic, including symptomatic, mental disorders; F1, Mental and behavioural disorders due to psychoactive substance use; F2, Schizophrenia, schizotypal and delusional disorders; F3, Mood disorders; F4, Neurotic, stress-related and somatoform disorders; F5, Behavioural syndromes associated with physiological disturbances and physical factors; F6, Disorders of adult personality and behaviour) and having no missing data for the selected variables. Exclusion criteria were: a length of stay <2 days or a psychiatric consultation on the day of admission.

We assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on medical research and with the Helsinki Declaration of 1975, as revised in 2008. All procedures were approved by the local ethics committee (CERHUPO) under the allocated reference number 17-7-1. For research based on data collected in the context of the administration of care and treatment, French law does not require written informed consent but only that the patient is informed of their right to oppose to the use of their data.

Data collection

Data were obtained from the computerised medical records. Collected data included age, gender, admission date, discharge date, date of the first psychiatric consultation, place of residence, type of ward, main psychiatric diagnosis, main non-psychiatric diagnosis and an indicator of disease severity.

The LOS was calculated as the whole number of days from admission to discharge. The number of days before the psychiatric consultation was calculated as the whole number of days from admission to the first CLP consultation. Year of admission was derived from the date of admission and categorised into nine classes, from 2008 to 2016. Place of residence was categorised into three classes: Paris, Paris area and elsewhere in France. Type of ward was taken to be where the patient had a bed at the time of the psychiatric consultation and categorised into two classes: intensive care unit or not. The main psychiatric diagnosis was recorded by the psychiatrist according to the seven ICD-10 codes from F0 to F6. The main non-psychiatric diagnosis was identified by the Groupe Homogène de Malades (GHM), the French version of the diagnosis-related group (DRG) and categorised into two classes: medical or surgical. We used the Groupe Homogène de Séjour (GHS) tariff, a French version of the DRG-based tariff derived from the GHM, as a proxy of disease severity. GHS tariffs were extracted from the annual national databases for public health facilities. For each patient, we used the GHS tariff corresponding to their GHM diagnosis and to the admission year. The GHS was preferred to the GHM as this variable is continuous, whereas the GHM contains a thousand of different categories and was thus not usable for statistical analyses.

Statistical analysis

Statistical analysis was carried out using the R Statistics software (http://cran.r-project.org, version 3.0.2 for Windows).

Our principal aim was to investigate the associations between the timing of CLP intervention and the LOS. As previously reported in the literature,Reference Lyons, Hammer, Strain and Fulop11Reference Handrinos, McKenzie and Smith13,Reference Wood, Wand and Hunt15 as patients who are in the hospital longer will obviously be more likely to receive psychiatric consultations later, it would be inaccurate to merely study this association. To control for this, Lyons et al Reference Sockalingam, Alzahrani, Meaney, Styra, Tan and Hawa16 suggested using the ratio of duration of hospital stay before the consultation to the LOS as an indicator of the timing of the consultation, and because the LOS and the number of days before the consultation are both positively skewed, they applied logarithmic transformations to both. We therefore used the following formula, described by Lyons et al and used by several authors since,Reference Lyons, Hammer, Strain and Fulop11Reference Handrinos, McKenzie and Smith13,Reference Wood, Wand and Hunt15 to calculate the referral time: referral time = log(number of days before the consultation)/log(LOS).

Preliminary analyses were performed to identify factors associated with referral time. To this aim, all variables (i.e. age, gender, place of residence, type of ward, main psychiatric diagnosis, main non-psychiatric diagnosis and disease severity) were included in the same general linear model, taking referral time as the dependent variable.

Then, we first assessed the associations between referral time and log(LOS) using three general linear models, taking log(LOS) as the dependent variable. Model 1 included only age and gender as covariates. Model 2 further included all the covariates except disease severity (i.e. year of admission, place of residence, type of ward, main psychiatric diagnosis and main non-psychiatric diagnosis). Model 3 included all covariates plus disease severity measured by the GHS tariff. Next, we tested for the presence of an interaction between referral time and each covariate by including corresponding interaction terms in the fully adjusted model (model 3). Should these interactions be significant, we planned to further examine the association with log(LOS) in stratified analyses.

As the association between referral time and log(LOS) may differ between patients with short, intermediate and long LOS, stratified analyses were performed for patients with LOS <5 days, 5–30 days and >30 days using the same covariates as model 3. The cut-offs of 5 and 30 days were drawn from the literature.Reference Desan, Zimbrean, Weinstein, Bozzo and Sledge8Reference Wood and Wand10,Reference Handrinos, McKenzie and Smith13

Finally, exploratory analyses were carried out to take into account the theoretical LOS of each patient, given the severity of their disease. To this aim we used the national database of average LOS (ALOS). For each patient, we extracted the ALOS corresponding to their GHM (the French version of the DRG-based tariff) and to the admission year, and calculated a relative stay index (RSI). The RSI was calculated as the actual number of days spent in hospital (LOS) divided by the expected number of days spent in hospital (ALOS):Reference Davis, Milne, Parker, Hider, Lay-Yee and Cumming17 RSI >1 indicates that the LOS was higher than expected, whereas RSI <1 indicates that the length of stay was less than expected.Reference Davis, Milne, Parker, Hider, Lay-Yee and Cumming17 The association between the ratio ‘number of days before the consultation divided by ALOS’ (n days/ALOS) (the main exposure) and RSI (outcome) was investigated using the same covariates as in model 2.

Finally, to assess the clinical significance of these results we aimed to predict the adjusted RSI on the basis of the timing of the first CLP consultation. To this aim, patients with a LOS between 5 and 30 days were divided in ten deciles of the ratio n days/ALOS. These deciles were used as the main exposure in a model assessing their association with RSI (outcome) using the same covariates as in model 2.

Results

The final study population consisted of 4500 participants (2298 women, 51%) with a mean age of 59.83 years (s.d. = 19.29). The study population selection is described in Fig. 1. Table 1 displays the characteristics of the study participants.

Fig. 1 Flowchart of the study population.

ICD-10 codes for: F7, Intellectual disabilities; F8, Pervasive and specific developmental disorders; F9, Behavioral and emotional disorders with onset usually occurring in childhood and adolescence.

Table 1 Characteristics of the study population (n = 4500)

F0, Organic, including symptomatic, mental disorders; F1, Mental and behavioural disorders due to psychoactive substance use; F2, Schizophrenia, schizotypal and delusional disorders; F3, Mood disorders; F4, Neurotic, stress-related and somatoform disorders; F5, Behavioural syndromes associated with physiological disturbances and physical factors; F6, Disorders of adult personality and behaviour; GHS tariff, the treatment tariff based on the Groupe Homogène de Séjour (the French version of the diagnosis-related group), used as proxy of disease severity.

In preliminary analyses, factors associated with shorter referral time were: younger age, admission to the ICU, more recent year of admission, main psychiatric diagnoses F1, F2 or F3, and lower disease severity (supplementary Table 1 available at https://doi.org/10.1192/bjp.2019.233). Patients with psychotic disorders (F2) were especially likely to have a shorter referral time than those with another main psychiatric diagnosis (supplementary Fig. 1).

In the main analyses, after adjustment for age and gender, referral time was significantly associated with log(LOS) (β = 0.52; P <0.001). This association remained roughly similar following adjustment for all covariates except disease severity (model 2: β = 0.47; P <0.001) and slightly decreased after further adjustment for disease severity (model 3: β = 0.31; P <0.001) (Table 2). Associations between referral time and LOS in the univariate model and in eight bivariate models adjusted for each covariate separately are presented in supplementary Table 2.

Table 2 Associations between referral time (exposure) and logarithm of length of stay (outcome) in multivariate modelsa (n = 4500)

β, estimated parameter (GLM coefficient) of the association between referral time (exposure) and logarithm of length of stay (outcome) or between each covariate and logarithm of length of stay; Ref., reference category; F0, Organic, including symptomatic, mental disorders; F1, Mental and behavioural disorders due to psychoactive substance use; F2, Schizophrenia, schizotypal and delusional disorders; F3, Mood disorders; F4, Neurotic, stress-related and somatoform disorders; F5, Behavioural syndromes associated with physiological disturbances and physical factors; F6, Disorders of adult personality and behaviour; GHS tariff, the treatment tariff based on the Groupe Homogène de Séjour (the French version of the diagnosis-related group), used as proxy of disease severity.

a. Model 1 was adjusted for age and gender; model 2 is model 1 further adjusted for smoking status; model 3 is model 1 further adjusted for disease severity.

The interactions between referral time and age (P = 0.002), type of ward (P = 0.02) and psychiatric diagnosis (P = 0.002) were significant. Stratified analyses revealed a more important association between referral time and log(LOS) in older patients (β = 0.43, P <0.001 in the 1763 patients ≥60 years, compared with β = 0.22, P <0.001 in the 1717 patients <60 years) and in patients admitted to the ICU (β = 0.50, P <0.001 in the 826 patients admitted to the ICU, compared with β = 0.25, P <0.001 in the 3674 patients not admitted to the ICU). Considering psychiatric diagnoses, referral time remained significantly associated with log(LOS) in patients with an F0, F3 or F4 code (respectively: n = 490, β = 0.30, P <0.001; n = 1175, β = 0.17, P = 0.01; n = 1799, β = 0.49, P <0.001). In patients with an F1, F5 or F6 code, we may lack sufficient power to reach significance (respectively: n = 301, β = 0.22, P = 0.10; n = 60, β = 0.06, P = 0.86; n = 250, β = 0.35, P = 0.35). Finally, in patients with an F2 code (i.e. psychotic disorders), there was no association between referral time and log(LOS) (n = 425, β = −0.20, P = 0.10).

Stratified analysis according to the LOS indicated a significant association between referral time and log(LOS) in patients with a LOS ≤30 days. The strongest association was observed in patients with a LOS <5 days (n = 569, β = 0.11, P <0.001), followed by patients with a LOS between 5 and 30 days (n = 3025, β = 0.09, P <0.001). We only observed a trend for an association in patients with a LOS >30 days (n = 906; β = 0.10; P = 0.06).

In exploratory analyses taking into account the theoretical LOS, similar results were obtained, with a significant association between the ratio n days/ALOS and the RSI (β = 1.10; P = 0.03). When patients were divided in ten deciles of this ratio, we observed that in patients for whom the CLP intervention occurred after 40% of the ALOS (i.e. in deciles 6–10), the ALOS was significantly higher than expected (Table 3). For instance, for a patient with an ALOS of 10 days, a CLP intervention at day 3 (i.e. in the 4th decile of the ratio) would be associated with an ALOS of 9.8 days (95% CI 8.6–11.0), compared with 12.2 days (95% CI 11.0–13.4) for an intervention at day 6 (in the 7th decile) (i.e. a LOS reduction of 2.4 days) (Table 3).

Table 3 Adjusted relative stay index (RSI) according to deciles of the ratio of the number of days before the consultation to the average length of stay (n days/ALOS)

Ref., reference category.

a. Adjusted for age, gender, year of admission, place of residence, type of ward and main non-psychiatric diagnosis.

b. Bold figures indicate significant results at P <0.05.

Discussion

This study confirms and extends previous literature. First, the long LOS for our population (21.48 days, compared with 5.31 days in the same period for patients in the hospital not referred to a consultation-liaison psychiatrist) confirmed the increased LOS associated with mental disorder.Reference Bourgeois, Kremen, Servis, Wegelin and Hales2,Reference Levenson, Hamen and Rossiten3 This longer LOS is partly explained by exclusion from the study of patients with a LOS <2 days. But it is also certainly partly explained by a greater disease severity, which could be a consequence of the comorbid mental disorder (e.g. because of a lack of ambulatory care), one of its causes (e.g. stress-related disorders) or both, testifying to a greater biopsychosocial complexity.Reference Consoli, Lemogne, Cole, Consoli and Limosin19 The relatively high ALOS and mean GHS tariff for our population (€8122.98, compared with €2030.96 for non-referred patients in the same period) confirm a high mean disease severity in our population. Other factors, such as a greater difficulty providing optimal care, could contribute to longer LOS for patients with psychiatric comorbidities. The high RSI in our study supports this hypothesis.

Our results also confirm and extend previous findings showing an association between earlier CLP interventions and shorter LOS in general hospitals.Reference Wood and Wand10Reference Sockalingam, Alzahrani, Meaney, Styra, Tan and Hawa16 Only one study, focusing on patients with an organic mental disorder, did not find any association between the timing of the CLP consultation and the LOS.Reference Hales, Polly and Orman20 Several hypotheses may explain the association between earlier CLP interventions and shorter LOS. First, as mental disorders may interfere with medical care and recovery, early treatment of the psychiatric condition may improve later care.Reference Daumit, Pronovost, Anthony, Guallar, Steinwachs and Ford21 In physically ill patients with depression, Aoki et al Reference Aoki, Sato and Hosaka22 reported that the LOS could be shortened by early effective antidepressant treatment. Second, early psychiatric interventions may lead to a better organisation of secondary care projects and reduce the delay of discharge or transfer. Third, the association may be explained by earlier CLP interventions in patients requiring a rapid transfer to a psychiatric ward. Previous studies have reported that patients with suicidal ideation and/or behaviour had earlier psychiatric consultations.Reference Kishi, Meller, Kathol and Swigart12,Reference Handrinos, McKenzie and Smith13 Furthermore, in a previous study from our group, we observed that, in 225 patients admitted to the Georges-Pompidou European Hospital after a suicide attempt by jumping, admission to a psychiatric unit on discharge from the hospital, which concerned 18% of the patients, tended towards an association with a shorter LOS in the hospital.Reference Todorov, Vulser, Pirracchio, Thauvin, Radtchenko and Vidal23

In the present study, we also found that the association between the timing of the first CLP intervention and the LOS only slightly decreased and remained significant after adjustment for disease severity. To our knowledge, only one previous study also adjusted for the severity of the disease and it found an attenuated but still significant association between the date of the request for a CLP consultation and the LOS in 814 patients.Reference Sockalingam, Alzahrani, Meaney, Styra, Tan and Hawa16 By demonstrating that early CLP interventions are associated with shorter LOS, independently of disease severity, the present results support the notion that CLP interventions may be cost-effective.

We also identified subgroups in which the association between earlier CLP interventions and shorter LOS was stronger. We first observed that the association was stronger in older rather than younger patients. This may be due to the relative underestimation of psychiatric comorbidities in older patients,Reference Andreas, Schulz, Volkert, Dehoust, Sehner and Suling24 which may result in a longer LOS unless properly recognised by the medical team. This interpretation is consistent with the longer referral time observed for older patients in the present study as well as in other studies.Reference Handrinos, McKenzie and Smith13,Reference Wood, Wand and Hunt15 However, a stronger association between earlier CLP interventions and shorter LOS in younger patients has also been reported.Reference Wood, Wand and Hunt15 We also found a stronger association in patients admitted to the ICU. This result might be because patients admitted to the ICU who had an early CLP intervention were on average less seriously ill (had a lower disease severity). Indeed, patients with more serious medical conditions in an ICU are more likely to have a tracheal tube and/or to be sedated, which often delays psychiatric assessment. This may notably apply to patients admitted to an ICU after a suicide attempt by self-poisoning, whose non-psychiatric medical condition may rapidly improve so that a transfer to a psychiatry ward is urgently indicated. Psychiatric interventions in response to suicide attempts may indeed partly account for the shorter referral time observed after admission to the ICU. However, patients having received a CLP intervention on the day of their admission or with a LOS <2 days were not included in our study.

Considering psychiatric diagnoses, we did not find that earlier CLP interventions were associated with shorter LOS in all categories of psychiatric diagnosis. Regarding substance use disorders, personality disorders or behavioural syndromes such as eating or sleep disorders, we may lack sufficient power to reach significance as these diagnoses were present in only a small proportion of our sample. Regarding psychotic disorders, this might be explained by the high proportion of patients who had a first CLP consultation early in their hospital stay, thus leaving fewer patients for comparisons of later consultations. For example, in the present study, 57% of patients with a psychotic disorder had a consultation in the first 3 days after admission (compared with 36% of patients with other psychiatric diagnoses, χ² = 71.16; P <0.001). Earlier CLP consultations for patients with psychotic disorders have been previously reportedReference Ormont, Weisman, Heller, Najara and Shindledecker14 and might be due to a greater vulnerability to the stress of the hospital admission, psychiatric symptoms that are more specific (i.e. positive symptoms), behaviours that could confuse non-psychiatric care teams or antipsychotic medications that these teams are not used to managing.

Stratified analysis according to LOS indicated that the association was observed in patients with short (<5 days) and intermediate (between 5 and 30 days) stays, while a simple trend was observed in patients with long stays (>30 days). Regarding patients with long stays, the result was consistent with the literatureReference Sledge, Gueorguieva, Desan, Bozzo, Dorset and Lee9 and can be explained by the multiple other factors that may co-occur in prolonged stays.

Strengths and limitations

The present study has several strengths. To our knowledge, it is the largest to address the association between the timing of CLP interventions and LOS in a general hospital, the first to investigate this association in subgroups and the first to estimate its clinical significance. Specifically, using the logarithm of the LOS as our main outcome, we also estimated, in terms of days, the reduction in the LOS that could be expected for any given timing of the first CLP intervention. This approach is a necessary step to convince stakeholders that CLP interventions could be cost-effective. Other strengths are the inclusion of consecutive in-patients who met a consultation-liaison psychiatrist for the first time, the use of the date of the consultation and not the date of the request for a consultation and adjustment for disease severity. The severity of the disease was the only covariate for which the statistical adjustment resulted in a substantial reduction in the association between earlier interventions and a shorter LOS.

Some limitations should also be acknowledged. First, owing to the cross-sectional design of this study, no causal conclusions could be drawn. Second, variables such as socioeconomic status were not considered. Third, we did not use structured clinical interviews to assess psychiatric diagnoses and suicide risk assessment was not systematically recorded. Fourth, the readmission rates of the patients were not available. Although no previous study investigating the association between the timing of CLP interventions and LOS has taken into account this last parameter, it has been suggested as an important point to consider in future studies,Reference Wood, Wand and Hunt15 as shorter LOS have been associated with higher readmission rates.Reference Carey and Lin18

Implications and future research

Our results suggest that earlier CLP interventions in general hospitals are associated with substantially shorter LOS, notably in patients with LOS <30 days. This association remained significant after adjustment for disease severity and, although the magnitude of the association varied, it remained significant across subgroups, except in patients with psychotic disorders, who are generally referred earlier. These results provide indirect evidence for the cost-effectiveness of CLP interventions. Further studies should assess ways of improving the timeliness of psychiatric interventions and the clinical- and cost-effectiveness of such timely psychiatric interventions.

Acknowledgements

We thank Mrs Ilham Cherrak, who collected the data.

Supplementary material

Supplementary material is available online at https://doi.org/10.1192/bjp.2019.233.

Footnotes

Declaration of interest: H.V. reports grants from Servier, outside the submitted work. F.L. reports personal fees and non-financial support from Janssen, personal fees and non-financial support from Lundbeck, personal fees and non-financial support from Servier, outside the submitted work. C.L. reports personal fees and non-financial support from Janssen-Cilag, personal fees and non-financial support from Lundbeck, non-financial support from Otsuka Pharmaceutical, personal fees and non-financial support from Boehringer Ingelheim, outside the submitted work.

References

1Hansen, MS, Fink, P, Frydenberg, M, Oxhøj, M, Søndergaard, L, Munk-Jørgensen, P. Mental disorders among internal medical inpatients: prevalence, detection, and treatment status. J Psychosom Res 2001; 50: 199204.CrossRefGoogle ScholarPubMed
2Bourgeois, JA, Kremen, WS, Servis, ME, Wegelin, JA, Hales, RE. The impact of psychiatric diagnosis on length of stay in a university medical center in the managed care era. Psychosomatics 2005; 46: 431–9.CrossRefGoogle Scholar
3Levenson, JL, Hamen, RM, Rossiten, LF. Relation of psychopathology in general medical inpatients to use and cost of services. Am J Psychiatry 1990; 147: 1498–503.Google ScholarPubMed
4Billings, EG, McNary, WS, Rees, MH. Financial importance of general hospital psychiatry to hospital administrator. Hospitals 1937; 11: 40–4.Google Scholar
5Levitan, SJ, Kornfeld, DS. Clinical and cost benefits of liaison psychiatry. Am J Psychiatry 1981; 138: 790–3.Google ScholarPubMed
6Strain, JJ, Lyons, JS, Hammer, JS, Fahs, M, Lebovits, A, Paddison, PL, et al. Cost offset from a psychiatric consultation-liaison intervention with elderly hip fracture patients. Am J Psychiatry 1991; 148: 1044–9.Google ScholarPubMed
7Levenson, JL, Hamer, R, Rossiter, L. Psychosocial interventions in chronic medical illness. An overview of outcome research. Gen Hosp Psychiatry 1992; 14(suppl 6): 43S–9S.CrossRefGoogle ScholarPubMed
8Desan, PH, Zimbrean, PC, Weinstein, AJ, Bozzo, JE, Sledge, WH. Proactive psychiatric consultation services reduce length of stay for admissions to an inpatient medical team. Psychosomatics 2011; 52: 513–20.CrossRefGoogle Scholar
9Sledge, WH, Gueorguieva, R, Desan, P, Bozzo, JE, Dorset, J, Lee, HB. Multidisciplinary proactive psychiatric consultation service: impact on length of stay for medical inpatients. Psychother Psychosom 2015; 84: 208–16.CrossRefGoogle ScholarPubMed
10Wood, R, Wand, APF. The effectiveness of consultation-liaison psychiatry in the general hospital setting: a systematic review. J Psychosom Res 2014; 76: 175–92.Google ScholarPubMed
11Lyons, JS, Hammer, JS, Strain, JJ, Fulop, G. The timing of psychiatric consultation in the general hospital and length of hospital stay. Gen Hosp Psychiatry 1986; 8: 159–62.CrossRefGoogle ScholarPubMed
12Kishi, Y, Meller, WH, Kathol, RG, Swigart, SE. Factors affecting the relationship between the timing of psychiatric consultation and general hospital length of stay. Psychosomatics 2004; 45: 470–6.CrossRefGoogle ScholarPubMed
13Handrinos, D, McKenzie, D, Smith, GC. Timing of referral to a consultation-liaison psychiatry unit. Psychosomatics 1998; 39: 311–7.CrossRefGoogle ScholarPubMed
14Ormont, MA, Weisman, HW, Heller, SS, Najara, JE, Shindledecker, RD. The timing of psychiatric consultation requests: utilization, liaison, and diagnostic considerations. Psychosomatics 1997; 38: 3844.CrossRefGoogle ScholarPubMed
15Wood, R, Wand, APF, Hunt, GE. Relationship between timeliness of contact and length of stay in older and younger patients of a consultation-liaison psychiatry service. BJPsych Bull 2015; 39: 128–33.CrossRefGoogle ScholarPubMed
16Sockalingam, S, Alzahrani, A, Meaney, C, Styra, R, Tan, A, Hawa, R, et al. Time to consultation-liaison psychiatry service referral as a predictor of length of stay. Psychosomatics 2016; 57: 264–72.CrossRefGoogle ScholarPubMed
17Davis, P, Milne, B, Parker, K, Hider, P, Lay-Yee, R, Cumming, J, et al. Efficiency, effectiveness, equity (E3): evaluating hospital performance in three dimensions. Health Policy 2013; 112: 1927.CrossRefGoogle ScholarPubMed
18Carey, K, Lin, MY. Hospital length of stay and readmission: an early investigation. Med Care Res Rev 2014; 71: 99111.CrossRefGoogle Scholar
19Consoli, SM. Critères d’évaluation et aspects médico-économiques. In Psychiatrie de Liaison (eds Lemogne, C, Cole, P, Consoli, S, Limosin, F). Lavoisier, 2018.Google Scholar
20Hales, RE, Polly, S, Orman, D. An evaluation of patients who received an organic mental disorder diagnosis on a psychiatric consultation-liaison service. Gen Hosp Psychiatry 1989; 11: 8894.Google ScholarPubMed
21Daumit, GL, Pronovost, PJ, Anthony, CB, Guallar, E, Steinwachs, DM, Ford, DE. Adverse events during medical and surgical hospitalizations for persons with schizophrenia. Arch Gen Psychiatry 2006; 63: 267–72.CrossRefGoogle ScholarPubMed
22Aoki, T, Sato, T, Hosaka, T. Role of consultation-liaison psychiatry toward shortening of length of stay for medically ill patients with depression. Int J Psychiatry Clin Pract 2004; 8: 71–6.CrossRefGoogle ScholarPubMed
23Todorov, L, Vulser, H, Pirracchio, R, Thauvin, I, Radtchenko, A, Vidal, J, et al. Suicide attempts by jumping and length of stay in general hospital: a retrospective study of 225 patients. J Psychosom Res 2019; 119: 3441.CrossRefGoogle ScholarPubMed
24Andreas, S, Schulz, H, Volkert, J, Dehoust, M, Sehner, S, Suling, A, et al. Prevalence of mental disorders in elderly people: the European MentDis_ICF65+ study. Br J Psychiatry 2017; 210: 125–31.CrossRefGoogle ScholarPubMed
Figure 0

Fig. 1 Flowchart of the study population.ICD-10 codes for: F7, Intellectual disabilities; F8, Pervasive and specific developmental disorders; F9, Behavioral and emotional disorders with onset usually occurring in childhood and adolescence.

Figure 1

Table 1 Characteristics of the study population (n = 4500)

Figure 2

Table 2 Associations between referral time (exposure) and logarithm of length of stay (outcome) in multivariate modelsa (n = 4500)

Figure 3

Table 3 Adjusted relative stay index (RSI) according to deciles of the ratio of the number of days before the consultation to the average length of stay (ndays/ALOS)

Supplementary material: File

Vulser et al. supplementary material

Tables S1-S2 and Figure S1

Download Vulser et al. supplementary material(File)
File 918.3 KB
Submit a response

eLetters

No eLetters have been published for this article.