Background

Hospitals provide a wide range of services necessary to meet the increasing demand for health care services and are an integral component of any health delivery system. However, inappropriate utilisation of high-cost but unnecessary or ineffective tests and medications in hospitals remains a significant challenge in many health systems [1]. Several studies documented improper hospital service use, which can be defined as “a hospital admission to provide care that could have been given in a less complex healthcare environment and at a lower cost” [2]. For example, it was previously shown that up to one-third of days of care [3,4,5] and diagnostic tests [6, 7], and one-fifth of all hospital admissions [8] could be inappropriate or unnecessary, negatively impacting patients’ physical and mental well-being, and driving up overall health care costs. Hence, eliminating inappropriate utilisation and waste is essential given the existing shortage of financial and human resources.

Advances in medical technology and, consequently, aggressive marketing to health care providers, direct-to-consumer advertising, political pressure from advocacy organisations, defensive medical decision making, fragmentation and discontinuity of care within and between health and social sectors - all can become the cause of healthcare overutilisation [9, 10]. Cost containment strategies can limit healthcare-related expenditure by eliminating inappropriate use of health care services while ensuring the continuous improvement of the quality of care. For example, one could consider controlling demand or supply for care, altering provision structures or hospital performance, cost-sharing, managed care, reference pricing, and generic substitution [11]. Another strategy is fostering hospital mergers and networks that may speed up restructuring through economies of scale at relatively small hospital sizes. However, creating a dominant position in the local hospital market may have an anticompetitive effect [12].

With the rising demand for healthcare services, hospitals can apply innovative methods to increase their efficiency [4]. This can be achieved by strengthening operational efficiency and targeting more significant healthcare expenditure cases. A range of measures can be used for this purpose: reducing duplication of services, decreasing the use of expensive inputs, decreasing the length of stay for inpatient care, reducing the number of long-stay beds, and reducing medical errors [13,14,15]. Another approach would be implementing measures that could rebalance services provision across the health system, improve allocative efficiency, and centralise administrative functions. Such measures could include shifting the provision of care from the hospital into the community, improving care coordination, strengthening preventative care, increasing the use of day surgeries, providing appropriate levels of acute care at home (hospital at home), and facilitating the discharge of patients who have to stay in hospitals longer [16, 17]. One could also consider implementing initiatives that lower management expenses and enhance administrative efficiency, such as simplifying managerial procedures; introducing uniform standards, distribution strategies and the availability of real-time consumer and provider information; improving electronic mechanisms of lodging, processing, and reimbursement of payments and claims; and outsourcing member management systems and other back-office services [18, 19].

Most importantly, besides the cost-saving and improving operational, allocative, and administrative efficiency, reducing inappropriate utilisation could eliminate potential iatrogenic effects of unnecessary services while improving healthcare quality. However, previous studies primarily focused on evaluating the effectiveness of interventions in reducing a specific service, while studies that would provide a clear overview of the utilisation management strategies for adult hospital services are still lacking. Hence, our study aimed to identify the initiatives intended to manage adult hospital services and reduce unnecessary hospital use among the general adult population.

Methods

We conducted a systematic review of published studies investigating initiatives intended to manage adult hospital services and reduce unnecessary hospital use among the general adult population.

Inclusion criteria

Studies were included if they reported using intervention in a general population aimed to reduce relevant primary outcomes (i.e., hospital services and/or emergency department (ED) use, frequency of hospital admissions, LOS, and use of diagnostic tests) compared to care as usual or different intervention. There were no time restrictions, but the publication language was restricted to English only.

Exclusion criteria

We excluded studies that targeted adult patient populations only with a specific medical condition (e.g., diabetes, asthma, cardiac failure, or cancer) or children to increase homogeneity and comparability between studies.

Search strategy

Five bibliographic databases (PubMed, ProQuest, Scopus, Web of Science, Ovid/Medline) were searched until March 2020. To capture a broad range of primary outcomes, in addition to utilisation management and utilisation review, we included the following search terms: concurrent review, prospective review, retrospective review, pre-admission review, pre-admission review, pre-certification, pre-admission certification, pre-admission certification, pre-admission authorisation, pre-admission authorisation, pre-admission testing, pre-admission testing, prior authorisation, same-day admission, physician profiling, provider profiling, physician financial incentives, demand management, case management, discharge planning, second surgical opinions, second opinions, step therapy, therapeutic substitution, closed formulary, utilisation. We additionally searched the references of included studies for other potentially essential studies.

Study selection, data extraction, and synthesis

Results from the bibliographic databases were merged, and duplicates removed. Two reviewers (LD and RKh) independently screened the search results by title, abstract and performed a full-text review. Disagreements were resolved by discussion and consensus with a third reviewer (HJ). We extracted the following information from the studies included in the review: type of intervention, study design, details of the intervention, and effects on primary outcomes (hospital services and ED use, admissions, LOS, use of diagnostic tests) and secondary outcomes (readmissions and costs). This review follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses [20].

Assessment of the methodological quality

We used an adapted version of the Quality Assessment Tool for Quantitative Studies (developed by the Effective Public Health Practice Project [21] to assess the methodological quality of the included studies (Appendix). The tool contains 19 items in eight key domains: (1) study design; (2) blinding; (3) representativeness in the sense of selection bias; (4) representativeness in the sense of withdrawals/drop-outs; (5) confounders; (6) data collection; (7) data analysis; and (8) reporting. Studies can have between six and eight component ratings, with each component score ranging from 1 (low risk of bias; high methodological quality) to 3 (high risk of bias; low methodological quality). An overall rating for each study was determined based on the component ratings. For example, if eight ratings have been given, a rating of ‘strong’ was attributed to those with no weak ratings and at least five strong ratings, ‘moderate’ to those with one weak rating or fewer than five strong ratings, and ‘weak’ attributed to those with two or more weak ratings. To minimise the risk of bias, assessments were completed independently by two reviewers (LD and EK). The ratings for each of the eight domains and the total rating were compared, and a consensus was reached on a final rating for each included article.

Data Analysis

Descriptive analyses were used to describe all studies that met the inclusion criteria, focusing on study design, participants, interventions and outcomes.

Results

The results of the screening process are shown in Fig. 1. After removing duplicates, 2261 papers were screened by title and abstract for possible inclusion in the review. The full text of 264 articles was obtained and assessed for eligibility. Of them, 56 selected papers were eligible for review. After screening references of included papers, we identified additional nine papers. Sixty four studies [22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85] met the eligibility criteria and were included in the final review.

Fig. 1
figure 1

PRISMA flow diagram

Full size image

Characteristics of the selected studies

Included studies were published between 1982 and 2020, conducted mostly in the USA (n = 34) [22,23,24, 29,30,31,32, 37, 39, 40, 42, 43, 45, 47, 49, 56, 57, 60, 63, 65, 67,68,69,70,71, 73,74,75, 77, 78, 81, 82, 84, 85], Canada (n = 4) [26, 35, 55, 61], Australia (n = 4) [38, 41, 59, 83], UK (n = 3) [36, 64, 72], Sweden (n = 3) [62, 66, 76], and one each in the Netherlands [52], Korea [44], China [53], Taiwan [27], Singapore [54], and Bahrain [34]. All studies focused on the general adult population; however, some focused on specific broader subgroups with psychiatric problems [29, 45, 54, 83], comorbid conditions [49, 77], psychosocial problems (e.g., problems with housing, medical care, substance abuse, mental health disorders, or financial entitlements) [70], uninsured [30, 31, 43, 68], patients with chronic medical conditions [27, 46, 49, 61, 67], or older patients [41, 43, 47, 49, 64, 66, 67, 76]. The duration of the study follow-up ranged from one month to seven years (Table 1).

Table 1 Study characteristics
Full size table

Fourteen studies (21.9%) were randomized controlled trials [22, 23, 43, 47, 49, 52, 53, 55, 62, 66, 69, 70, 73, 74], three were multicenter research trials [36, 63, 76], two were quasi-experimental studies [31, 67], four were controlled before-and-after studies [30, 68, 72, 85], twenty-one studies (32.8%) were non-controlled before-and-after studies (NCBA) [24, 27,28,29, 32, 35, 37,38,39, 41, 42, 50, 54, 56,57,58,59,60,61, 75, 78], three were time-series studies [26, 34, 44], three were case-control studies [64, 65, 84], one was a prospective cohort study [77], one was longitudinal study, six were retrospective cohort studies [25, 33, 79,80,81,82], and four were cross-sectional studies [40, 45, 71, 83]. While, in two studies were not stated type of design [48, 51]. Fourty studies (59.7%) can be categorized as assessing interventions targeted at the patient journey during hospital stay or medical center-based interventions [22,23,24, 26, 27, 29, 30, 34, 37, 39, 40, 42, 44, 45, 49, 54, 56, 57, 59,60,61,62,63, 65, 69, 70, 72, 74, 75, 78, 81,82,83, 85]; four evaluated interventions aimed at discharge [41, 47, 55, 76], Not; and 13 examined community-based interventions [31, 35, 38, 43, 46, 52, 53, 64, 66,67,68, 73, 77].

Methodological quality assessment

In the overall assessment, the methodological quality of only one reviewed study (1.5%) was rated as ‘strong’, while seven (11%) and 56 (87.5%) articles were rated as ‘moderate’ and ‘weak’, respectively (Appendix). In terms of study design, 21 studies (32.8%) were rated as ‘strong’. The remaining 13 studies (20.3%) scored ‘moderate’ and 30 studies (46.9%) scored ‘weak’. We were able to rate 39 studies for representativeness relating to withdrawals and drop-outs: 25 (64.1%) studies rated as ‘weak’, four (10.3%) as ‘moderate’, and ten (25.6) as ‘strong’. With respect to confounders, 11 (17.2%) studies were rated as ‘strong’, six (9.4%) as ‘moderate’, and 47 (73.4%) as ‘weak’. There were 23 studies (35.9%) rated as ‘weak’ for their data collection because the authors did not provide sufficient information on the validity or reliability of their collection methods. There were 37 papers (57.8%) rated as ‘moderate’ and four papers (6.3%) rated as ‘strong’. Based on the data analysis of each reviewed study, 36 (56.3%) of the reviewed studies were rated as ‘strong’, while 12 (18.8%) and 16 (25.0%) were rated as ‘moderate’ and ‘weak’, respectively. The reporting quality of the reviewed articles was also analysed. Out of the 64 articles included, 36 studies (56.3%) were rated as ‘strong’, 21 studies (32.8%) and seven studies (10.9%) were rated as ‘moderate’ and ‘weak’, respectively.

Nine broad utilisation management methods

We identified nine broad utilisation management methods: care plan, case management, care coordination, utilisation review, clinical information system, physician profiling, consultation, education, and discharge planning. The findings related to these nine methods are described below in Table 2, using sub-categories of the following main types of interventions: non-organisational interventions aiming to reduce hospital utilisation, organisational interventions to reduce hospital utilisation, and interventions at the discharge stage of the patient journey.

Table 2 Reported measures and outcomes
Full size table

Prehospital advanced life support drug treatment

These interventions focused on access to primary care, medical and social resources. For example, two studies [31, 68] evaluated interventions that aimed to improve access to primary care. Studies suggest that improving access to primary care centres is associated with fewer ED visits [31, 68], fewer inpatient hospital days than controls [31], but report no difference in inpatient admissions between groups [68]. One retrospective cohort study examined the effect of prehospital advanced life support drug treatment in reducing subsequent hospital utilisation by the medical patients receiving such drugs [35]. There was a significant decrease in admissions in the drug intervention group driven by chest pain patients and improved prehospital field conditions for all chief complaints. Care plan and case management were the main interventions related to prehospital advanced life support drug treatment.

Two comparative cohort studies examined the impact of patient care plans on service utilisation [38, 77]. Sweeney et al. [77] compared patient-centred management to usual case management for patients who had a life-limiting diagnosis with multiple comorbid conditions. Among the patient-centered management, inpatient admissions reduced by 38%, inpatient hospital days by 36%, and emergency department visits by 30%. Grimmer-Somers et al. [38] found that a holistic community-based program using a care plan for frequent ED attendees had significant improvements in client health and decreased crisis emergency department and inpatient admissions.

Case management

Primary care case management

Case management is “a collaborative process that assesses, plans, implements, coordinates, monitors, and evaluates the options and services required to meet an individual’s health needs using communication and available resources to promote quality and cost-effective outcomes” [50]. Eight studies focused on using case management interventions based outside the hospital. Five studies reported a decrease in hospital utilisation [45, 46, 64, 66]. Three studies found no significant difference between groups in neither ED visits nor hospital admissions [43, 67, 73].

Hospital-based case management

Of 23 studies evaluating case management interventions, 12 focused on case management as an ED-initiated or medical centre-based intervention for frequent hospital utilisers. Six comparative cohort studies observed a decrease in the mean or the median number of ED visits than the controls [30, 72] or before the case management [27, 39, 57, 61]. One study reported an increase of 2.79 median ED visits post-intervention [59]. This study included primarily patients with substance abuse or psychiatric problems underlying the ED visits, suggesting case management may be less effective in reducing ED utilisation in this population. One RCT reported no significant difference in the median number of ED visits following CM [74]. In contrast, two RCTs reported a decrease in the number of ED visits [62, 70] and hospital days [64] among those in the intervention group. Two studies have examined changes in hospital admissions or LOS, found a significant decrease in the number of admissions [29], hospital readmissions [54] and LOS.

Care coordination

Two studies examined the impact of care coordination programs on ED visit rate amongst frequent ED users [49, 56]. The randomised controlled pilot study by Koehler et al .[49] found that hospital-based care coordination using extra care bundle comprising three interventions (medication counselling, enhanced discharge planning, and phone follow-up) targeting high-risk older people compared to usual care was successful in reducing 30-day post-discharge hospital readmission or emergency department visit rates. The comparative cohort study by Murphy et al. [56] implemented a multidiscipline ED-care coordination program using a regional hospital information system capable of sharing patients’ individualised care plans between ED providers. The study reported a significant decrease in ED visits 12-months following the intervention.

Utilisation Review

The utilisation review program consists of several different review activities: pre-admission authorisation (prospective review), concurrent review (during the patient stay), retrospective review (relying on medical records), prospective review. One study investigating a pre-admission review program found a decrease in hospital admissions by approximately 12% [81]. Of eight studies that examined the effect of concurrent review on the LOS, five studies found a decrease in hospital LOS [26, 34, 63, 82, 84]. Another study that examined the effect of utilisation review on patterns of health care use found that the referrals for a second opinion have reduced the number of procedures performed in the review group. However, there was no significant difference between the groups during the study period in terms of rates of admission to medical-surgical, substance abuse, or psychiatric units, average LOS, the percentage of those who received pre-admission testing, or the rates of use of home care following utilisation review activities [65].

A retrospective analysis of utilisation management programs has concluded that pre-admission review rarely denies requests for admission, and nearly one-third of patients approved by pre-admission review for inpatient care requested approval for continued stay through concurrent review [82]. One multicenter trial examined the effect of utilisation management strategies on the use of a radiological test [36]. There was a consistent reduction from 29.4 to 13.3 X-rays per 100 operations after introducing the new request form and concurrent review. Two studies that evaluated the effectiveness of a prospective review program in reducing blood component utilisation reported that the implementation by the blood bank staff of a prospective review of orders for blood products resulted in a significant decrease of 38.8% and 31.4% in the use of fresh frozen plasma and platelets, respectively [40], as well as a total reduction inpatient medical costs realised as a result of cancelled orders [71]. Due to the importance of drug utilisation, this type of utilisation review has been categorised as a primary intervention.

Drug utilisation review

Three studies focused on drug utilisation review interventions. One study reported a significant decrease in the number of antibiotic treatment courses and the percentage of patients receiving any antibiotic following implementing an antibiotic order form for all inpatient antibiotic orders in the hospital [32]. The second study reported a significant decrease from 40% to 20% of patients using benzodiazepines after drug utilisation review activities in an inpatient setting [83]. Another retrospective cohort study examined the effect of implementing a drug utilisation management program and evidence-based guidelines on the appropriate use of drugs and found that implementing a drug-utilisation management program using clinical pharmacists was associated with a decrease in inappropriate epoetin prescribing and significant cost savings [24].

Clinical information system

A clinical information system is a computer-based system encompassing clinical or health-related information, distinguished from administrative information systems by the requirement for data entry or data retrieval by clinicians at the point of care. Some areas addressed by clinical information systems are clinical decision support, electronic medical records, physician’s order entry, telemedicine, problem lists, summary reports, results review, nursing protocols and care plans, and alerts and reminders. Recently, interests have been focusing on medical errors with monitoring and managing variation in practice [86]. Electronic medical records and physician’s order entry systems, and clinical decision support are the primary interventions related to clinical information systems.

Electronic Medical Record

One before-after analysis of an intervention targeting ED frequent users reported that the use of health information technologies to identify the most frequently visiting patients and easy access to individualised care plans through the EMR to all healthcare providers resulted in a significant reduction in the number of ED visits, labs ordered, total ED contact time, and ED charges [75].

Physician’s order entry system

A physician’s order entry system is a subsystem of a hospital information system. One prospective time series study reported that the number of stat lab tests and overall LOS at six months after physician’s order entry implementation decreased significantly compared with the pre- physician’s order entry system period [44]. Using a randomised controlled design, Shea et al. [69] demonstrated that a computer-generated informational message directed to physicians as an intervention resulted in reduced LOS in an inpatient setting. According to Bates et al. [22], 69% of potentially redundant diagnostic tests were cancelled in response to reminders following the introduction of a clinical information system that included a physician’s order entry system.

Clinical decision support

A clinical decision support system is a computer-based application that analyses data and provides knowledge and person-specific information to aid physicians and other health providers in clinical decision making [87]. One study that evaluated real-time clinical decision support intervention observed improved blood utilisation. After implementing clinical decision support system, the percentage of patients transfused outside the guidelines decreased to 35% [37].

Physician profiling

Physician profiling is a cost-containment strategy whereby the patterns of health care provided by a practitioner or other provider (e.g., hospital) for the defined population are compared to other norms - profiles of other physicians or practice guidelines - based on practice [88]. A quasi-experimental study with control groups found that LOS at the profiled site decreased by an additional third of a day in the profiling year than at the non-profiled sites [85].

Consultation

The randomised controlled trials by Bree et al. [24] implemented mandatory radiology consultation whereby each radiology examination required prior approval. This intervention did not observe differences in inpatient imaging use following the mandatory radiology consultation.

Discharge planning

Discharge planning refers to developing a plan to treat the patient’s medical needs after leaving the inpatient department to contain costs and improve patient outcomes. Discharge planning should ensure that patients leave the hospital at an appropriate time in their care and that, with adequate notice, the provision of post-discharge services is organised [89]. We identified three studies that focused on interventions at the discharge stage of the patient journey [41, 47, 55]. All three studies that examined the effect of discharge planning on LOS in hospital and readmission rates compared with usual care found a decrease in hospital LOS for those allocated to discharge planning. There were lower readmission rates in the discharge planning group for older participants with a medical condition at three months of discharge [41, 47].

Early supported discharge

Discharge planning typically involves a greater degree of care provision and support following discharge than discharge planning interventions. Early supported discharge or early home-supported discharge may include discharge planning but aims specifically to accelerate discharge from the hospital with continued support in a community setting, typically at the same intensity that would have been provided had the patient remained in hospital. These interventions are usually provided by multidisciplinary teams, including doctors, nurses, and therapists. Still, the degree of coordination and whether they are driven by hospital outreach or community teams can vary [89].

Post-discharge case management

Two RCTs have examined the effectiveness of case management provided after patients are discharged from the hospital regarding the utilisation of hospital services by these patients. One study found a significant reduction in hospital admissions, bed-days and attendances at the out-patient department [53]. In contrast, the second study did not find significant differences between groups for readmission, care utilisation, quality of life, or psychological functioning [52].

Cost outcome

Of all included studies, 23 studies provided cost-related outcomes. Six studies reported savings after implementing utilisation review programs [24, 37, 40, 81, 84] or a computerised physician order entry system [22]. One study reported cost savings from reduced days of hospitalization [29]. Ten studies reported significantly reduced hospital charges [30, 31, 56, 62, 64, 67, 68, 77] or ED costs after the intervention [43, 75]. One randomised controlled trial of 96 patients observed a trend toward reduced total healthcare cost in the experimental group, but the difference was not statistically significant [73]. Two studies reported a mixed effect - one reported a significant decrease in ED and medical inpatient costs but no apparent change in the cost of medical out-patient, psychiatric inpatient, psychiatric emergency, or ambulance services [57]. The other found a significant decrease in ED costs. However, no difference was reported for inpatient services, psychiatric emergency services, out-patient services, physicians’ fees, or total hospital costs, with the cost of case management included [70]. Also, one study reported program costs with no assessment of net costs or savings [38].

Education

Developing education programs for patients, families and health care providers (i.e., nurses or physicians) is considered the primary intervention in many countries [49, 67, 77, 90]. The goal of the education programs is to provide health care providers with the principles of utilisation management.

Discussion

Our review identified nine utilisation management methods, including care plan, case management, care coordination, utilisation review, clinical information system, physician profiling, consultation, education, and discharge planning. Of all interventions reported in the reviewed studies, case management strategy was the most frequently examined. Disease management is considered an effective strategy for dealing with frequent hospital users with specific diseases (e.g., congestive heart failure or diabetes). Whereas disease management focuses on particular illnesses, case management is focused on optimising multidisciplinary treatment. We identified several models of case management, such as brokerage [54], assertive community treatment [46], intensive case management [29, 39], clinical case management [57, 70], and different case management models (i.e., strengths-based case management, generalist case management, rehabilitation).

Our findings suggest that interventions aimed to increase primary care accessibility and case management effectively reduce ED visitation [31]. Though mostly uneven in methodological rigour, studies indicate that pre-admission review for hospitalisation is highly effective in reducing hospital admissions. The implementation of utilisation management interventions increased out-patient visits, possibly reflecting the link of frequent hospital users to other services. Overall, studies that focused on interventions during the patient stay in the hospital (e.g., concurrent review) and interventions at the discharge stage of the patient journey (e.g., discharge planning) effectively reduce the LOS. However, the limited evidence showed that mandatory radiology consultation interventions were ineffective in reducing inpatient imaging use. As a good outcome, introducing the clinical information systems (e.g., physician’s order entry system) reduced LOS. Such automated access to patient records improved the efficiency of information exchange among physicians across the continuum of care. Clinical decision support systems, which consisted of interruptive best practice alerts at the physician’s order entry system, also significantly improved blood utilisation. We found that interventions directed towards supply, such as physician profiling, were associated with decreased LOS without adversely affecting physician satisfaction. However, such reductions were also observed among control groups in ED visit numbers [30, 70, 73, 74], hospital admissions [66, 70, 73] and LOS [70]. Case or care management and utilisation review interventions were consistently reported to reduce hospital costs, and no studies reported increases in hospital costs following the intervention.

There were several limitations to this review. First, there is marked heterogeneity among reviewed studies. Second, in an attempt to focus on the literature concerning the general adult frequent user populations, studies were excluded that did not examine a general population (e.g., pediatric, individuals with asthma, cancer, diabetes, and cardiovascular disease) or focused on a specialised out-patient care setting.

Conclusion

To ensure the delivery of efficient and effective health care, to reduce the misuse of inpatient and outpatient services, the use of utilisation management strategies in hospitals is unavoidable. The use of relevant strategies and interventions allows for avoiding unintended consequences emanating from the financial incentives and disincentives on health care professionals’ decisions around care and service delivery.