Health services in Trinidad: throughput, throughput challenges, and the impact of a throughput intervention on overcrowding in a public health institution

Background Throughput might be partially responsible for sub-optimum organisational and medical outcomes. The present study examined throughput and the challenges to ensuring optimum throughput in hospitals, and determined the effectiveness of a throughput intervention in reducing overcrowding in a public healthcare institution in Trinidad and Tobago. Methods First, a literature review of throughput and its processes in relation to improving hospital care was conducted. Second, the challenges to throughput in healthcare were reviewed. Data were also collected from print media, hospital records, and the central statistical office in Trinidad and Tobago to discuss throughput and describe the throughput status in hospitals. Finally, the effect of a throughput intervention on overcrowding was determined. The intervention was implemented over six months, from October 2010 to March 2011, and comprised three stages of a five-stage throughput process: transferring patients to a specific medical ward, bedside electrocardiograms (ECG), and promptly obtaining patient investigative reports and patient files. Results Problems with the throughput process led to prolonged delays or failures in obtaining lab reports, radiology services, ECGs, and pharmaceutical supplies, as well as inadequate social work services and other specialised services. During the throughput intervention, there was a reduction in overcrowding/overflow to 5–10 patients per day with a daily admission rate of 58. However, at post-intervention, there was increased overcrowding/overflow to 20–30 per day but fewer admissions (52 per day) i.e. similar to pre-intervention period. Additionally, there was an increase in bed complement in the department of medicine from 209 (2011) to 227 (2012). Overcrowding continued into 2016 and beyond: medical admissions in 2016 were 46.4 per day and the medical bed capacity was 327 (indicating a 44% increase in capacity from 2012). Conclusion Hospital throughput processes are currently suboptimum. Improving specific throughput processes or targeting the greatest primary constraints might help decrease overcrowding.


Background
Acceptable hospital performance is a much desired outcome by payers, patients, and providers. Trinidad and Tobago, though a small country, has high expectations for hospital performance because of its proximity to the United States of America (USA), the presence of trained professionals from the United Kingdom (UK) and USA, and a multitude of reports highlighting the poor performance of the health sector in terms of waiting times, overcrowding, and a lack of timely investigations and treatment. Such outcomes, however, are largely dependent on hospital efficiency, public policy [1,2], and resource utilisation [1]. According to some researchers, it 'would be difficult if not impossible' to improve outputs without increasing inputs [3,4]. Others believe that better outcomes can be achieved by using the 'convergence model' , which is defined as 'the integration of historically distinct disciplines and technologies into a unified whole that creates fundamentally new opportunities for life science and medical practice' [5]. Another instrument for improving outcomes is the use of throughput interventions, which are recognised as a 'critical success strategy' [6] and a major indicator of outcome quality [7]. Research on throughput in the context of accidents and emergencies in many hospitals has revealed that overcrowding in the emergency department is attributed to the inability to transfer patients to wards because of poor-quality inpatient care. However, throughput analyses of inpatients are rare. In Trinidad and Tobago, throughput studies have not been done in emergency settings or among inpatients. This study attempts to fill this gap by reviewing throughput and throughput challenges for inpatients in a tertiary health care institution in Trinidad; as well as determining the effectiveness of a throughput intervention for inpatient care in reducing overcrowding.

Throughput
Throughput, according to Little's law, is defined as the rate at which a business can produce a product or service in a given unit of time [8]. In the context of healthcare, it can refer to the number of patients served in a unit of time [9]. It can also refer to a product (e.g. number of surgeries or eye tests conducted) [10] or an organisational process (i.e. 'the cycling of patients through a hospital's physical resource base') [6]. Others define it as the sum total of actions (support services and operating systems) that are required to move a patient from admission to discharge [6]. Common throughput interventions target improvements in bed flow and lab report availability, more prompt ward electrocardiograms (ECG), and better clerical communication. While throughput is a major contributor to improving organisational outcomes, healthcare providers have largely ignored the optimisation of throughput processes [11]. The use of throughput processes combined with convergence technology such as electronic devices (iPods, smart phones) can further improve organisational outcomes and throughput times or length of stay (LOS) [12], which is defined 'as the time from patient arrival to discharge time' [12]. Well-designed throughput processes improve clinical outcomes and patient satisfaction, and decrease overcrowding. Indeed, merely increasing inputs by expanding capacity, staff, and bed occupancy [13] might not lead to the desired outcomes. Throughput partially determines high-quality healthcare [14], and encompasses the processes involved in patient flow from preadmission through discharge [6]. Throughput services (e.g. lab, radiology, pharmacy, cardiology, gastroenterology, neurology, attendant and clerical services) and prompt dispatch to the ward [15], which improves the efficiency of the emergency department, are also relevant for admitted patients. According to Press Ganey, negative organizational outcomes demonstrated by long waiting times, overcrowding, misplacement of patients, delayed surgeries, backlog, and cancellation of cases result from inadequate support or throughput services [16]. A study conducted by Pedroja found that poorly managed patient flow (i.e. overcrowding in emergency departments, intensive care units, or other hospital departments) results largely from 'support services such as laboratory and radiology being unable to keep up, resulting in physicians having less time to focus on individual patients' [17].
Throughput optimisation has its genesis in the theory of constraints (TOC), which is based on the ideology that a chain is only as strong as its weakest link. The TOC is about 'managing the flow of a good or a person through a system and not about managing the capacity within a system'. TOC views every organization as 'a chain of interdependent events (or processes) where the performance of each event (or process) is dependent upon the previous event' [18]. However, maximising the efficiency of a microsystem at the expense of the macrosystem decreases organisational outcomes [11] and should be avoided. Instead, the focus should be on improving the efficiency of primary constraints, and not non-constraints. A focus on the latter could lead to 'efficiencies syndrome' , whereby the efficiency of nonconstraints is increased, thus reducing the efficiency of the entire system [18]. Management of primary constraints is key to the attainment of good patient outcomes [19]. Organisational or patient outcomes have been shown to improve with throughput interventions such as patient flow logistics [20], pharmacist-facilitated medication reconciliation [21], hospital co-ordinator assistance of patients from admission to discharge [22], patient communication [23], patient flow [24,25], and bridging gaps identified through quality improvement models such as Six Sigma [26]. These are all represented in the revised model for hospital efficiency shown in Fig. 1. Emergency care depends heavily on emergency departures and partnering effectively with inpatient care providers to decrease emergency department boarding [27]. To improve hospital efficiency and patient outcomes, it may not be enough to choose any throughput interventionrather, the 'right throughput' intervention must be chosen. This is similar to the deletion of wasteful processes and the adoption of lean techniques to improve emergency department outcomes [28]. 'In a hospital environment characterized by increasing patient demand, constrained physical resources and a rising cost of capital, optimizing inpatient throughput is an essential operations management strategy' [6].

The case of Trinidad and Tobago
Trinidad and Tobago is a small twin-island state comprising an area of 1864 mile 2 and a population of 1.33 million people, of which 35.4% are East Indian, 34.2% African, and 22.8% mixed [29]. Public health services are free. The last few decades have seen increases in inputs (staff, material, and financial resources [30]) into the healthcare system. However, despite increases in the gross domestic product per capita, health budget per capita, healthcare personnel per capita, and investments in socio-economic parameters such as education, roads, telephones, and the Internet, national health indicators have shown no significant improvement [30]. Additionally, the increasing health budget per capita has done little to curb major illnesses such as ischaemic heart disease ( Table 1).
Trinidad and Tobago's health services have been plagued with inefficiencies over the last few decades, partly emanating from poor throughput processes. Reports from numerous commissions, hospital administrators, and healthcare providers have alluded to hospital inefficiencies and poor throughput processes. In 1970, the editorial of a daily newspaper reported that a full-scale enquiry into the concern was needed, since 'the crisis in the nation's health services seems to be getting dangerously close to the point of total collapse' [31]. Such headlines have not been uncommon in Trinidad and Tobago. In 2016, the Trinidad and Tobago Guardian reported that hospitals are at a 'crisis point' [32]. Furthermore, an enquiry into the tragic deaths of 14 people, led by former chief justice Sir Issaac Hyatali in 1992, reported that the deaths were due to a massive systems failure [33]. In 2008, the Gladys Gaffoor Commission of Enquiry identified a multitude of inefficient services and processes (e.g. lab, pharmaceutical, procurement, and other support services) [34]. In 2011, the Ramsoomair Enquiry at the San Fernando General Hospital (SFGH) identified a 'failure of medical staff to recognise the  Data retrieved from [79] massive blood loss in a timely manner, and lack of prompt and efficient intervention by both medical and nursing staff' [35]. Despite knowledge of the problems of waiting times for treatment, clinics, lab reports, and obtaining services (e.g. cardiology, radiology), no formal scientific studies on throughput have been conducted in Trinidad and Tobago The aim of this study was to determine the throughput challenges in a tertiary public health facility in Trinidad and Tobago and the effect of a throughput intervention on hospital overcrowding.

Methods
This was a descriptive, observational study conducted at SFGH, the only tertiary public health facility in South Trinidad. It serves approximately half a million people, predominantly Indo-Trinidadian nationals, from central and South Trinidad. At this institution, the total annual admissions amount to 52,252, of which medical admissions account for 17,245. Furthermore, the number of total clinic visits are around 178,184, with new visits accounting for 21,618 [36]. In 2010, the overall bed capacity was 654, with 209 assigned to the department of medicine. The present bed capacity is 745, of which 327 are assigned to the department of medicine. SFGH has suffered from chronic overcrowding for decades, and it persists to this day: as much as twenty to thirty medical patients are found at any given time sitting in chairs or lying in trolleys in corridors and hallways in the medical wards and/or the emergency department.
Data covering a period of about two decades were collected from customer relations, social and pharmacy services, and various commissions of enquiry. The waiting times in the radiology lab and clinic were obtained from the clinic appointment records. Further data were obtained from utilisation reports from the South West Regional Health Authority (SWRHA) and the media, and from observational analysis. Information on health budgets and populations were gathered from the central statistical office and the Ministry of Health of Trinidad and Tobago.
A throughput process intervention was conducted over 6 months (September 2010 to March 2011; Table 2). The intervention process involved transferring patients to a specific ward that would in turn allocate them to a bed (patient flow: Stage 0); recording ECGs at the patient's bedside (Stage 1); and the facilitation of lab services, files, and other clerical work by a ward clerk assistant attached to each medical team (Stage 2). Additionally, several minor infrastructural improvements such as additional medication trolleys and drug cupboards were made. Medication facilitation (i.e. collecting and transportation of medication from the pharmacy to the ward) was partially implemented because of a shortage of pharmacists and pharmacy assistants (Stage 3). Discharge facilitation (Stage 4) was not implemented because of time and resource constraints. Implementation of stages 0, 1, and 2 required the services of ECG technicians, ward clerk assistants attached to each medical team, and pharmacy assistants. Without this intervention, ECGs and blood investigations requested on a given day would only be obtained on the following day, the day after that, or even longer (especially if requested on a Friday). Based on the TOC, these were considered primary constraints affecting the treatment and recovery of patients. With the intervention in place, however, the ECG technicians performed the requested ECGs throughout the day at the patient's bedside. The ward clerk, as part of the medical team, assisted in a multitude of tasks, including the collection and sorting of patient reports before attaching them to patients' files. These files would then be available for doctors conducting early morning (and sometimes daytime) ward rounds. Ward clerks also kept a master sheet of the names and housing destinations of patients belonging to their teams. Ward clerk assistants also had several other tasks such as obtaining urgent reports and patients' files from the medical records department, and transporting blood samples or prescription sheets to the pharmacy. They provided an important link between the medical team and other support services. The objectives were to ensure prompt and easy availability of ECGs, lab reports, medications, and other support services. The ultimate aim of the intervention was to decrease overcrowding. During the intervention, data on patient overflow, admissions, average LOS (ALOS), and occupancy rate were collected daily. Further, patients' LOS was monitored on a daily basis. Support services, particularly radiology services, were continuously monitored and daily feedback was provided to all stakeholders, including the Minister of Health. A daily logbook for all activities was used to determine which processes were accomplished and which were not. Multidisciplinary team meetings were conducted intermittently with various staff, including pharmacists, nurses, doctors, clerks, attendants, and social workers. Data from similar months (e.g. January 2011 [during the intervention] and January 2012 [after the intervention]) were analysed to determine changes in patterns relating to overflow patients and ALOS (Table 3).
This study was conducted as part of an efficiency project, coordinated by the author, and driven and approved by the chief executive officer (CEO) of the SWRHA. Data are available to the public and healthcare providers, and can be obtained through the Freedom of Information Act. Simple descriptive analysis was undertaken and is represented in tables and graphs in the following section.

Throughput status
Sub-optimum throughput processes were identified in a number of services, such as pharmacy, laboratory, radiology, physiotherapy, attendant, medical social work, bereavement, and communication services; ECG, echocardiogram, and stress testing services; and endoscopy, colonoscopy, endocrine, and neurological tests. For instance, for pharmacy services in 2008, the percentage of prescriptions that were fully dispensed ranged from 39.3% in April to 55.6% in December; that of partially dispensed prescriptions was 19.4% in March and 39.6% in July; and that of prescriptions not dispensed was 6.8% in May and 38% in April (Table 4). When compared to the percentages obtained in 2016, there were no meaningful changes [37]. There were, however, additions to various services, such as oncology, pharmacology, and parenteral feeding services.
There are also recurring customer complaints. At the SFGH, the most common complaints were for obtaining lab reports (32%), delays in obtaining medical reports (16%), and misplaced medical files (14%) (Fig. 2) [38]. The problem of missing files was noted by the Commission of Enquiry into the Health Sector in 2008. The enquiry revealed that 'the records at most of the hospitals and at some clinics and health offices are very unsatisfactory. Very often, they are 'lost or cannot be traced' [34]. At the SFGH, delays in obtaining test results (32%) and delays in medical reports (16%) were the most frequent complaints in 2006 [38]. However, in 2015, waiting time for test results (28.9%), staff attitude (9.9%), and staff competency and credibility in relation to medical care (9.9%) were the most frequent complaints [39]. The top complaints on a national scale in 2006 were delays in obtaining test results (26%), delays in obtaining medical reports (13.3%), misplaced medical notes at clinics (11.57%), problems with the lab (7.1%), staff attitudes (5.5%), medical management (5.36%), ineffective communication (4.8%), postponement of surgical procedures (3.1%), nursing management (2.4%), and disappointment with clinic appointments (2.1%) [40]. The results also showed that processes for dealing with audits and feedback are inefficient. Complaints and feedback about monitoring and improving the healthcare delivery system do not appear to have been effectively utilised (Fig. 2). Furthermore, the 11th Annual Client Feedback Complaints and Commendations Report (2006-2007) revealed that 'there has been little involvement of management at the Facility and Regional Health Authority (RHA) levels in resolving complaints' [40].
Moreover, physiotherapy/radiology services continue to take days and sometimes weeks to obtain. Noninvasive cardiology services such as stress testing and echocardiography are severely compromised due to the lack of appropriate staff and resources. Many patients were not even given an appointment. Patient discharge facilitation, bereavement, and health information services were also found to be lacking. Important safety measures (e.g. isolation rooms, optimum space between beds and fire extinguishers) were either insufficient or virtually absent except for in the teaching hospital wing, which was opened in 2013.
Services that are not covered by the public healthcare system are covered by patients themselves privately, the Medical Social Work Department [34], the RHA voucher system, the Ministry of Health waiting list initiative, or  [41]. This funding is used for a multitude of tests and services, such as bone scans, blood tests, endoscopic retrograde cholangiopancreatography, the cystic fibrosis screen test, magnetic resonance angiogram (MRA), sestamibi scans, echocardiograms, stress tests, flowmetry tests, renography, nuclear scans, magnetic resonance imaging (MRI) scans, endoscopies, and colonoscopies. Delayed or absent services lead to poor quality care and increase the cost to healthcare providers. The Ministry of Health has developed a variety of guidelines to improve certain throughput processes, some of which are not used or will not be usable in the immediate or foreseeable future, including lab However, these systems and guidelines have also never been implemented or enforced. Other throughput processes at SFGH such as support services and waiting times, and how they compare with international benchmarks, are given in Tables 5 and 6. Importantly, dissatisfaction with support services, medication, hospitality, and management issues and customer complaints remained high [30] throughout the observation period.

Throughput intervention results
From 2009 to 2010, there were 42 daily medical admissions on average, and the ALOS was 5.2. During this period, which was about 1 year before the intervention, overcrowding at the institution manifested as excess patients (between 10 and 40 patients waiting for a bed daily)? Patients were held in the corridors, waiting rooms, overflow wards, and even sent to private nursing homes. In January 2011, during the intervention, there was 368 overflow or excess patients. However, in January 2012, about 1 year after the cessation of the intervention, there were 687. This indicated an increase of 317 (or 87%) 'excess patients' (see Table 3 and Figs. 3 and 4). The increase in overcrowding occurred even though medical admissions decreased from 58 per day in January 2011 to 52 per day in January 2012, and while the number of medical beds increased from 209 to 227. The ALOS remained stable, at about 4.9 days, in 2011 and 2012 but increased to 5.9 in January 2016. In 2016, the number of medical beds increased to 325 (a 44% increase from 2012), with 47.6 medical admissions per day. However, overcrowding persisted, with an average of 20 overflow patients per day awaiting beds each morning.

Discussion
Throughput at this institution is clearly suboptimum, with major gaps existing in important throughput processes, such as pharmacy, lab, and radiology services (Tables 4, 5, and 6). Throughput services, such as pharmacy [42], radiology [43], clerical [44], and lab services [45] as well as patient flow [46], are major determinants of hospital efficiency [47,48]. As such, throughput optimisation would naturally increase efficiency [49,50]. While increasing input is helpful, the failure to deal effectively with throughput might have contributed significantly to poor organisational outcomes such as quality of care or patient satisfaction [51], as was found in another study of SFGH patients, among whom there was a low rate of satisfaction (17.55%) with the hospital's support services [30]. Relatively low satisfaction rates were also found for treatment (30.42%), hospitality services (27.71%), and management issues (33.99%) [30]. In 2016, the quality department found that 58% of patients were initially dissatisfied with services, and that 39% (36 out of 92) remained dissatisfied even after their complaints had been resolved [39]. Poor throughput outcome indicators such as patient safety, satisfaction, quality of hospital care, readmission rates, and human resource indicators were identified in some studies [52][53][54]. Given the deficient throughput services, the intervention was designed to addressing the primary constraints for patient treatment [55], including bed flow, bedside ECG, lab report availability, and access to other radiological services. The results of the intervention reveal a dramatic improvement in throughput processes, resulting in a decrease in overcrowding. In fact, when the intervention was discontinued and systems returned to the pre-intervention conditions, overcrowding worsened dramatically. In particular, the addition of beds following the intervention led to a further increase in ALOS from 4.8 to 5.8 days (25% increase) and overcrowding, which  returned to pre-intervention [37] levels. This occurred despite the fewer admissions in general. The primary constraints or throughput processes that require improvement might be unique to a particular setting and time. Therefore, it is necessary to target the 'right throughput processes' based on the principles of the TOC [19] and the particular healthcare setting. The primary constraints identified should then be improved within a timely manner. Ensuring appropriate throughput using principles such as the TOC has proven effective, as found in studies implementing interventions such as patient flow [56], communication [57,58], pharmacist-facilitated discharge facilitation [21], and lab report availability [59].
The demand for better outcomes, despite limited resources and a complex environment, has been fuelled by advances in other specialities such as networking and global interconnectivity, technology, and process models. Convergence allows healthcare providers to use the contributions (e.g. operating principles) of multidisciplinary teams from various specialities (e.g. medical engineering, computer science) to improve healthcare [5]. Throughput can also benefit from the additional input brought by convergence, such as imaging, re-engineering, big data and health information technology, and nanotechnology [60]. In addition, according to the KPMJ Healthcare and Life Sciences Institute, operating processes or throughput must be adjusted to satisfy the customer base [61], and can be further facilitated with the right culture and environment [61].
Adding inputs, though helpful, is insufficient to increase outputs. It can even breed inefficiency [30,52]. One such intervention, as demonstrated in an earlier study on increasing bed capacity to solve the overcrowding problem, only led to an increase in ALOS [62] and no significant decrease in overcrowding. Another study reported that adding more staffed beds only exacerbated the problem [63]. In fact, increasing capacity might worsen the inappropriate use of resources [13,64]. Increasing inputs without optimising throughput (i.e. targeting the throughput processes with the highest constraints) increases costs and wastage, and decreases efficiency. Inadequate and inefficient hospital care is postulated to lead to poor services, patient dissatisfaction, and poor outcomes, which manifest as overcrowding [65,66], increased LOS [3,67], emergency department patient backup [68], delayed treatment [69], and heightened opportunities for errors [70]. These in turn increase costs and the risk of complications [71] and overall result in decreased hospital efficiency [72]. Furthermore, the backlog of patients in emergency departments leads to the inability to obtain appropriate beds, lost opportunities to treat, and increases in the rates of patients discharging against medical advice [73][74][75][76][77][78]. Improving hospital throughput not only improves inpatient care but frees up beds for accepting emergency patients, prevent emergency overcrowding, and improving emergency patient care.

Limitations
This was a single-centre study. Given that the centre has its own unique culture and value systems, extrapolation to the rest of the country and other parts of the world might not be appropriate. However, similar institutions with similar backgrounds are found worldwide and the challenges and findings might be relevant to them. Another limitation is that some of the data are the author's own observations and may be biased. In addition, the data are predominantly secondary data available to the public, and ethical approval for this study was not sought.

Conclusion
Throughput is a critical success strategy. The throughput processes at this public health institution fall short of expectations and international benchmarks. However, the use of an appropriate throughput intervention led to a decrease in overcrowding. The underlying principles of throughput and its impact on quality outcomes can likely be applied to any public service.