ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846
-A A +A

Barber–Johnson Technique to Assess the Efficiency of India’s Apex Tertiary Care Hospitals

A 10-year Trend Analysis

LaxmiTej Wundavalli ( is Senior Resident at the Department of Hospital Administration, All India Institute of Medical Sciences, Delhi. Sidhartha Satpathy ( is Professor and Head at the Department of Hospital Administration, All India Institute of Medical Sciences, Delhi. D K Sharma ( is Medical Superintendent at the All India Institute of Medical Sciences, Delhi.

Scarcity of beds and long waiting lists is a challenge for public hospitals worldwide. This study uses the Barber–Johnson technique to analyse the efficiency of five hospitals of India’s apex medical institute. Four hospital efficiency indicators based on annual hospital statistical reports from 2005–06 to 2015–16 were studied and compared with the respective figures of other countries. Though the results reveal a high efficiency of the five hospitals, there is a high probability of poor health outcomes in future. There is thus a need to redesign processes and to innovate better care models to improve healthcare service delivery.


The authors would like to thank the reviewers for their valuable comments and suggestions.

The supply of hospital beds and their utilisation are fundamental to health economics and planning in all health systems, regardless of administration and financing methods. The addition or closing of hospital beds is one of the difficult and controversial issues in health planning and health politics (Tulchinsky and Varavikova 2000). Their supply is measured in terms of hospital beds per 1,000 population. However, this varies widely between and within countries. Currently, the bed to population ratio in Delhi is 2.76 despite increase in beds from 32,941 in 2004 to 49,969 in 2015 (Government of NCT of Delhi 2017). Hospital bed is a scarce resource.The cost of construction of a bed is usually equal to the cost of operating the bed over two to three years. The decision to build a bed obliges the health system to indefinitely fixed costs even if that bed is unused as a result of regulation or the reduced utilisation from professional or economic incentives (Tulchinsky and Varavikova 2000). Therefore, each bed should be efficiently and optimally utilised. Some of the traditional hospital performance metrics are hospital bed occupancy rate (BOR), turnover interval (TOI), turnover rate (TOR), average length of stay (ALS), and hospital throughput. These indicators are usually examined in isolation for lack of a convenient way of examining them together. B Barber and D Johnson described the eponymous Barber–Johnson diagram in 1973 to present information about patient length of stay, TOI, discharges and deaths per available bed, and percentage bed occupancy in one diagram (Barber and Johnson 1973).

Mayer et al (1994) and Premik et al (1999) compared the hospital utilisation indices for Croatia and Slovenia, respectively, in times of war and peace using the Barber–Johnson technique, which helped in evaluating decisions about hospital bed utilisation. Lastrucci et al (2016) highlighted the inefficiency of hospitals and the paradoxes between inadequate hospital bed density and low hospital bed utilisation in the Republic of Albania. Morera (2013) graphically compared the hospital bed management in eight hospitals of Costa Rica.Alternatively, the Pabón Lasso (1986) model, which is a modification of the method proposed by Massabot (1978), is used as a tool to measure efficiency of hospitals.

Studies using Barber–Johnson technique are scant reportedly due to poor publicity of the technique and the need for time-consuming precise technical drawings. The primary objective of the present study was to analyse the hospital bed efficiency at India’s apex medical institute where bed scarcity and bed waits are daily occurrences despite the increase in number of beds over the years in the form of speciality centres within the institute. The second objective was to compare the efficiency of hospital bed utilisation in India with other countries across the world using the performance of the institute as a reference.


The study was conducted at the All India Institute of Medical Sciences, New Delhi, which is a premier apex tertiary care medical institute and university in South Asia. The institute was established by an Act of Parliament and together with its constituent centres is one of the largest public sector, tertiary-level teaching hospitals in India. It has dual patient care roles as a specialised referral hospital and as a large general hospital. Over the years, the brand value of the institute has led to high patient load. The annual outpatient department attendance has risen from around 0.5 million in the early 1970s to 1.2 million in the late 1990s to more than 3 million in 2015–16. Nearly 55% of the outpatients come from outside Delhi. The cost to the inpatient is less than $1 per day and treatment is provided free of charge to those “below the poverty line” (Duran et al 2015).

This study analysed the hospital bed efficiency of five constituent centres of the institute: Centre A (main hospital) is a broad speciality hospital; Centre B is a super-speciality hospital providing services in cardiothoracic and neurosciences; Centre C is a tertiary-care teaching cancer hospital; Centre D is an apex tertiary care centre for ophthalmic sciences; and Centre E is an apex drug-dependence treatment centre. Annual hospital statistics from 2006–07 to 2015–16 were obtained from the annual reports of each centre (AIIMS nd). The parameters were then calculated using the formulae described below and then plotted on a standard Barber–Johnson diagram. The tool was chosen for its emphasis on capturing four different indicators simultaneously in a simple manner on a single diagram and to highlight its utility in India, where it is rarely used. As per the technique, initially the four parameters are calculated based on annual hospital statistics, such as the number of available beds (A), number of beds occupied (O), and the number of discharges and deaths (D) (Yates 1982).

The average length of stay is then calculated using the formula: ALS = [O × 365]/D. This effectively sums up all the occupied bed days during a specific period and divides that figure by the number of discharges and deaths during the same period.

The average turnover interval (TOI; expressed in days) = [(A – O) × 365]/D. This is the average time that beds are left empty between each discharge and admission. It might be better understood as the average length of emptiness.

The average number of discharges and deaths per available bed (DDPB; expressed as a number per year): DDPB = D/A. This becomes a measure of throughput or activity. Throughput is not merely a function of length of stay; it also includes the TOI.

The average bed emptiness (expressed as a percentage) = [(A – O) × 100]/A. This is a crude description of unused capacity.

Description of the diagram: In Figure 1 (i), the x axis is TOI and the y axis is ALS. Figure 1 (ii) adds the percentage emptiness lines that radiate from the origin. In Figure 1 (iii), the points on each axis are joined and the resultant diagonal lines show the number of discharges and deaths per available bed (throughput). The closer such lines are to the origin, the higher the throughput. Putting the three diagrams together produces Figure 1 (iv). Any point on the diagram gives four figures.

The obtained statistics were also compared using the Pabón Lasso model (Figure 2).


The bed complement and total number of admissions of each centre each year have been compiled in Table 1 (p 46).


The hospital statistics for ALS, TOI, deaths and discharges per bed, and percentage emptiness for each centre are depicted in Tables 2–4 and Tables 5 and 6 (p 47).

Centre A is the first teaching hospital of the institute that provides services in all broad specialities. Between 2006–07 and 2015–16, the number of admissions increased by 66.4%. However, the number of beds increased by 10.5% (110 beds) only. As per Figure 3, in 2006–07, ALS was five days, TOI was 1.6 days, with a throughput of 55 and 25% bed emptiness. In the 10th year, ALS reduced to 3.57 days, TOI to 0.47 days, and bed emptiness to 11%, with a throughput of 90. There is an average decrease of 4.2% per year in the ALS during the 10-year period.

Centre B is a super-speciality centre providing cardiothoracic and neurosciences services. Between 2006–07 and 2015–16, the number of admissions increased by 15.7% and the number of beds rose by 17.2% (62 beds) in Centre B. As shown in Figure 4 (p 46), in 2006–07, the ALS was seven days, the TOI was 1.5 days, with a throughput of 43 and 17% bed emptiness. In the 10th year, the ALS reduced to a little less than seven days, TOI to one day, and bed emptiness to 14%, with a throughput close to 50.

The number of beds in Centre C has been constant at 182 except for 2006–07. As shown in Figure 5 (p 46), in 2006–07, the ALS was close to 13 days, the TOI was 1.83 days, with a throughput of 25 and 13% bed emptiness. During the 10-year period, the ALS markedly reduced to 5.48 days and TOI decreased to 1.19 days, with a throughput of 55, while bed emptiness decreased to 18%. The TOI increased after the addition of beds in 2007, but has been gradually reducing to reach the present figure of 1.19 days.

Centres D and E are single-specialty centres and have homogeneous bed pools. Centre D is an apex tertiary care and referral centre providing ophthalmic services. The number of beds in the hospital has been constant. As shown in Figure 6 (p 46), in 2006–07, the ALS was close to five days, the TOI was 1.16 days, with a throughput of 59 and 19% bed emptiness. Over the next 10 years, the ALS reduced significantly to 2.55 days, TOI to 0.41 days, and bed emptiness to 14%, with a throughput of 123. There is an average change of 11.59% per year in ALS at this centre.

Centre E is a drug-dependence treatment centre at the national level. The number of beds in the centre has been constant. As per Figure 7, in 2006–07, the ALS was close to 9.5 days, the TOI was 7.48 days, with a throughput of 21 and 44% bed emptiness. In the 10th year, the ALS increased to 15.29 days, TOI reduced to 4.31 days and bed emptiness to 22%, with a throughput of 19.


Centre A: There is a left and downward shift towards the origin (Figure 3). The shift is more obvious in the later part when the TOI and ALS were reducing rapidly. The hospital attempted to increase the throughput by increasing the bed occupancy and reducing the ALS. Especially from 2012–13 to 2015–16, both the TOI and percentage emptiness are increasing beyond the international reference values, reflecting intensive capacity utilisation. As per international guidelines, the recommended reference values for TOI are between one and three days and for the BOR they are ≥75% (Lastrucci et al 2016).As per the National Beds Enquiry, United Kingdom (UK), 2001, the BOR must not exceed 82% (Department of Health 2001).By the end of 2016, the bed occupancy of main hospital (Centre A) was 88% with a throughput of 90. Further, since the “average” occupancy is measured at midnight, that is, the point of lowest occupancy in the 24-hour cycle, the true average occupancy may be higher than the quoted figure.

According to the European Hospital and Healthcare Federation, in 2014,Ireland had the highest average bed occupancy in Europe at 93.3%, followed by the UK (84.4%), Austria (82.7%), and Switzerland (82.6%), while the European Union (EU) average was 76.9%. In the last years, no consistent trends were registered among EU member states (World Health Organization Europe 2016). In England, hospital bed occupancy has risen from an average of 87.1% in 2010–11 to 90.3% in 2016–17. As per the data from the Organisation for Economic Co-operation and Development (OECD), in 2011, Israel had the highest rate of hospital bed occupancy at 98%, followed by Norway and Ireland (both at over 90%) (OECD 2013; Rosen et al 2015). It was also stated that these hospitals have fewer curative beds compared to other OECD countries. The European average bed occupancy is 77% and this roughly corresponds to the average occupancy in the United States (US) for hospitals with 1,000 beds (Jones 2011).The average occupancy rate for acute care beds in Canada in 2009 was 93% (Silversides and Sullivan 2013). In South Asia, the average BOR in Kabul’s national hospitals was 58% in 2012 (Health Economics and Financing Directorate 2012).A study of 98 secondary- and tertiary-level hospitals in Sri Lanka found that around 80% of individual hospitals had bed occupancies of more than 70% and around 50% hospitals had an occupancy of 80%–85% (Dalpatadu et al 2015). Indonesian hospitals had an average BOR of 60%–85% (Andayani et al 2015). The mean occupancy rate in Thailand was 82.9% (Tangcharoensathien et al 2015).Ironically, Vietnam, which also has the highest bed to population ratio of 25 in South East Asia, had a very high BOR of 99.4% in 2012 (Tran Thi Mai Oanh et al 2015).

A study by Boyle et al (2013) demonstrated that in a large acute teaching hospital with 650 beds, the probability of severe adverse events, that is, an in-hospital fall resulting in a fracture or an overdose of medication requiring treatment or a longer stay to rectify, increased significantly above 75% occupancy. Half of all falls resulting in a fracture occurred above 99% occupancy and half of medication events occurred above 98% occupancy. The probability of at least one severe event per day was 15% at 80% occupancy, 20% at 90% occupancy, and 28% at 100% occupancy (Boyle et al 2013).A study among patients securing admissions via the emergency departments of three tertiary care hospitals in Western Australia by Sprivulis et al (2006) demonstrated that high daily occupancy leads to a 30% increase in in-hospital deaths.As per Jones (2016),it takes a 10 percentage point increase in annual average occupancy to make a 1.5 percentage point increase in hospital deaths (including death within 30 days of discharge). Hospitals with the highest occupancy rates are farthest from the four-hour waiting time target of the National Health Service (NHS) (Nuffield Trust 2016).

Occupancy Rate and Its Impact

A review of average occupancy in British and American acute hospitals concluded that the “optimum” whole hospital occupancy, that is, across multiple constituent specialty bed pools, is around 72% for a 200-bed hospital, 78% for 500 beds, and 83% for 1,000 beds. An occupancy level of around 80% has been suggested to be a suitable balance between throughputs and turn-away for a large acute hospital, especially for those handling a high volume of elective surgery via a number of specialty specific bed pools. Irrespective of the occupancy specific to different sized hospitals, an absolute maximum occupancy (even during the winter months) in the range of 82% to 85% is required to maintain the level of hospital-acquired infection at the minimum possible (Jones 2011).Cunningham et al (2006) found a significant negative correlation between TOI and rates of Methicillin-resistant Staphylococcus aureus infections, the influence of TOI being greater than that of bed occupancy. Reassuringly, the crude infection rate of the hospital in our study has been relatively stable with a notable decrease in the last three years of research (Table 7). As per the factsheet of the World Health Organization (WHO), at any given time, the prevalence of healthcare–associated infection varies between 5.7% and 19.1% in low- and middle-income countries (WHO nd).

Average length of stay: The ALS for Centre A was 3.57 days in 2015–16. In most countries, ALS for all causes has fallen from an average of 9.2 days in 2000 to eight days in 2011. Since 1980, the ALS in acute care beds has been constantly reducing in all EU states (OECD 2013). Between 2000 and 2011, it dropped by one bed-day. As per OECD (2019), the ALS was 6.6 days and it ranged from 4.1 days in Turkey to as high as 16.2 days in Japan in 2017. The average for EU was 6.3 days in 2017. The ALS in an NHS hospital has fallen by more than 40% from 8.4 days in 1998–99 to 4.9 days in 2015–16. As per the National Health Performance Authority of Australia (2018), ALS for all public and private hospitals (major, large, medium, small, and children’s) decreased from 4.36 days (+/-2.32) in 2012–13 to 3.82 days (+/-1.99) in 2016–17. Between 2012–13 and 2016–17, ALS decreased from 3.4 days to 3.2 days in public hospitals (a decrease of 1.5% on average each year) and from 2.3 days to 2.2 days in private hospitals (a decrease of 0.9% on average each year) (Australian Institute of Health and Welfare 2018). In 2011–12, the ALS across Kabul’s national hospitals was 9.1 days. Excluding outliers—hospitals with ALS of 10 days or more—the ALS decreases to 4.1 days vis-à-vis the desired ALS of three days (Health Economics and Financing Directorate 2012). In 2009, the ALS for 89 of 98 public hospitals in Sri Lanka was 2.4 days (Dalpatadu et al 2015). Indonesia had an ALS of six–nine days in 2012 (Andayani et al 2015) and Vietnam of seven days in 2013 (Tran Thi Mai Oanh et al 2015).

In the US, the ALS decreased on average 0.2% per year between 2003 and 2012, and was estimated at 4.5 days in 2012 (Weiss and Elixhauser 2014). On the other hand, there is an average decrease of 4.2% per year in ALS in Centre A (main hospital) from 2006–07 to 2015–16 which indicates the extremely high pressure on beds in this centre. High pressure on hospital beds can encourage clinical staff to examine the patients in the ward to determine who is most able to go home in order to make space for an incoming admission. This process can lead to the discharge of the patients earlier than is clinically or socially desirable. The length of stay is also influenced by the clinical policies of the medical staff, the organisational and administrative habits of the doctors and hospitals, and the extra-hospital influences of social environment and availability of paramedical support services to which the patient is to be discharged (Barber and Johnson 1973).

There is a need to investigate the re-admission rates of the hospital under study and compare them with expected 30-day re-admission rates for select clinical conditions. However, Harrison et al (1995) showed that improving hospital efficiency by shortening the length of stay does not appear to result in increased rates of re-admission or number of physician visits within 30 days of discharge from hospital. Unlike increasing BORs, shorter stays have not been found to be related to adverse patient outcomes (Brownell and Roos 1995; Clarke and Rosen 2001; Clarke 2002; Westert and Lagoe 1995).

The indicators reflect that Centre A is working at a very high efficiency and has an acceptable crude infection rate. The Pabón Lasso model places it in sector 3, which is designated for the most efficacious hospital. However, the model misses the increasing workload which can have serious consequences on quality of care, staff burnout, and patient safety metrics. Paradoxically, in an attempt to accommodate larger number of patients, the percentage of turn-away increases with increasing occupancy (Jones 2001). Increasing throughput also increases turn-away (Jones 2002).

A strict referral system may help direct the resources of the hospital towards patients truly requiring tertiary care and towards improving the quality of care. However, this requires improving the care facilities across the hospitals in the country. The requirement of beds for each speciality may need to be evaluated separately to confirm a true bed shortage or a process-influenced shortage. Delayed discharges need to be minimised as the increase in number of days a bed is occupied unnecessarily might cause a rise in BOR. There is an urgent need to plan for future service demands and provide support to health professionals to prevent premature discharges due to bed shortages. Support for hospitals to improve the coordination of care across diagnostic and treatment pathways has also been suggested by Borghans et al (2012). Further, according to projections by the National Programme for Health Care of the Elderly, the elderly population will increase to 12% of the total population in India by 2025, 8%–10% of which is estimated to be bed-ridden, requiring utmost care (Ministry of Health and Family Welfare 2011). Given the anticipated rise in ageing population, which will lead to an increased demand for hospitalisation and longer lengths of stays, development of alternate models of care, such as community care services, might be required as part of national health policy.

Centre B: Figure 4 highlights a stark deviation in 2008–09, suggesting probable error in calculations in the absence of any major policy shift in hospital admissions or discharges. The Pabón Lasso model places this hospital in sector 3. As per the Barber–Johnson diagram, the trends have been relatively stable and optimal between 2012–13 and 2015–16. However, a proactive approach might help in view of the trends that suggest a future reduction in TOI and increase in bed occupancy beyond 85%.

Centre C: Centre C is an apex tertiary care cancer hospital where patients are referred from all over the country, especially north India. There has always been a great demand for both regular admissions and day-care admissions for chemotherapy, blood transfusions, etc (Table 4). The day-care beds were excluded for calculating the efficiency indicators as depicted in Table 8.

The trends depicted in Figure 5 highlight the pressure under which the hospital has been performing in the last 10 years. There is on an average 10% decrease in ALS per year. There is an average increase of 10.46% in throughput per year. Increasing throughput leads to increasing turn-away. At 83% occupancy, the turn-away is 1% for 100 beds (Jones 2001).An average occupancy of 72% has been recommended to be optimal for a multi-speciality hospital of 200 beds (Jones 2011). As such, Centre C may be described as performing most optimally, especially in the context of the demand. The Pabón Lasso model places this hospital in sector 3. The average percentage of emptiness has been 25.57% in the 10 years. While the average bed occupancy of some specialities is almost 100% throughout the year, the average occupancy of all the combined speciality bed pools decreases which is due to their being closed speciality pools.

The above statistics do not consider the day-care beds of the hospital, which is an important element of oncology services. Day-care admissions were 88% of the total admissions in 2006–07 compared to 80% in 2015–16. Though the total number of admissions on day-care beds increased by 67.9% (from 18,325 in 2006–07 to 30,764 in 2015–16), the day-care beds increased by only 36.3% (from 22 in 2006–07 to 30 in 2007–08, with the number remaining unchanged till 2015–16). The BOR is more than 300% in 2015–16 with a very high throughput of 1,025. Despite the increase in services provided within the constraints of space, humanpower, and other support services, there has also been a steady rise in the waiting lists over the years. Pressure on beds results in priority being given to urgent cases at the expense of the less urgent ones. If the less urgent cases cannot be admitted straightaway, they are placed on a waiting list for planned admission. Continued pressure on beds from urgent cases can block beds for planned admissions, and the waiting lists grow and waiting time lengthens. This has also resulted in patients lodging complaints with e-grievance portals initiated by the Government of India. This again leads to an ethical dilemma for clinicians and administrators as to who should be served first. Should the patient who has lodged a grievance be considered first, bypassing patients legitimately awaiting their turn as per the waiting lists, notwithstanding the premise of severity and/or salvagability of a case? Clinicians and hospital administrators are also found to express unhappiness regarding the performance of the hospital being assessed by the waiting lists rather than by their productivity.

At present, the day-care services operate for a maximum duration of 12 hours, six days per week. A potential solution may be to increase the service duration to 24 hours, utilising the same number of beds, or increase the number of day-care beds, which will require additional space and trained staff. Further, in the context of this hospital, space is at a premium.

Centre D: Internationally, eye hospitals have seen a decline in the ALS of patients owing to the shift towards day-care surgeries. Day surgery accounts for over 90% of all cataract surgeries in a majority of countries (OECD 2013).In 2016, the American Hospital Association (AHA) reported a day surgery rate of 67% of all procedures carried out in community hospitals only (AHA 2018).For cataract surgery, Iceland recorded the shortest length of stay in Europe at 1.1 days in 2005. In the same year, the ALS for cataract operations was 1.5 days in the UK and 1.7 days in Denmark. Luxembourg, Norway, and Sweden each recorded ALS of 1.8 days for cataract surgery in 2005. In comparison, Belgium and Spain had already reached a rate of 1.6 days in 2004. The ALS for non-cataract ophthalmic surgery also fell to its lowest-ever level in 2005: 2.1 days in the UK, 2.6 days in Norway, and 2.7 days in Denmark and Sweden. The Netherlands, Denmark, and Finland performed cataract surgery as a day case more than 96% of the time in 2005 (McGinn 2007). The relatively lower rate of drop in ALS in some countries has been explained by more advantageous reimbursement for inpatient stays, national regulations, obstacles to changing individual practices of surgeons and anaesthetists, and tradition (Castoro et al 2007). Figure 6 reflects a decrease in ALS in Centre D, which might reduce further with re-engineering of healthcare and hospital administrative processes.

Centre E: Traditionally, mental hospitals and drug-dependence treatment centres have relatively longer patient stays compared to general hospitals. As per Malone et al (2004), the median length of three cohorts of psychiatric admissions to a district general hospital psychiatric unit was approximately 15 days. The National Institute of Mental Health and Neurosciences(NIMHANS), India, recommends a stay of at least 21 days for drug-dependence treatment (NIMHANS De-Addiction Centre 2009).Internationally, as per the European Commission (2019), the ALS for mental and behavioural disorders was 24.65 days (SD: +/-0.86) between 2006 and 2016. The longest average stays for in-patients with mental and behavioural disorders were reported for Malta, the Czech Republic, and the UK, between 39 days and 45 days (Eurostat 2016).

As per the recommendations of the Royal College of Psychiatrists (2010), a BOR of 85% is seen as optimal. As per Jones (2013), all psychiatric hospitals with fewer than 100 beds should be operating below 85% average occupancy while larger hospitals should be limited to a maximum of 85% occupancy in order to protect both patients and staff from untoward incidents arising from “busyness.” The BOR for Centre E was 78% in 2015–16, which may be described as optimal. The corresponding turn-away of patients at this occupancy lies between 1% and 5%, which is an optimal balance between providing services and queuing. However, this may be significant in India where the number of beds in mental hospitals is 2.05 per 1,00,000 population compared to the global average of 17.5 (WHO 2014). The demand for these beds and barriers in utilisation, if any, may be explored to further review the requirement of beds and/or private rooms in this centre. Figure 7 also demonstrates the stark difference in hospital indicators between a broad specialty hospital (Figure 3) and an exclusive drug-dependence treatment centre. We might need to revisit the statistics calculated for Centre E.


The Barber–Johnson technique reveals a high efficiency of the constituent hospitals of the apex medical institute. However, if similar trends continue for the main hospital, there is a high probability of poor health outcomes. Presently, the day-care centre of the cancer hospital has an extremely high throughput (1,025) and BOR (>300%). A high throughput can lead to drop in standards of care.

The value proposition of the institute is to provide quality care at affordable rates which will be at stake if steps are not taken to redesign processes. Though institutes of similar standing have come up in the recent past and new institutes are being planned across the country, they are yet to cater to the healthcare demand substantially. It is necessary to explore the factors that are hindering the growth of these institutes as merely increasing beds without addressing quality and sophistication of care will not resolve the issue of healthcare needs. Another concern is the lack of an efficient referral system in India. It is envisaged by the government that the Ayushman Bharat (National Health Protection Scheme) will assuage the issue, including in terms of cost and provision of care, critique of the financial outlays for the scheme notwithstanding. However, to realise its purpose, the scheme will require not only an improvement of the health infrastructure but also an objective, robust monitoring and evaluation mechanism to upgrade the processes and outcomes of the empanelled hospitals. The anticipated increased availability, accessibility, and affordability of healthcare might lessen the load on the apex hospitals.

The results of this study highlight two perennial questions: (i) What is the ideal size of a hospital?; (ii) Do we simply need more hospital beds? The sizing of hospital beds attached to medical colleges in the country is currently based on the norms set by the Medical Council of India. The Indian Public Health Standards recommend bed numbers based on forecasts of admissions (which are based on demography) and estimated ALS. Though current literature does not have a precise answer for an ideal size, it does recognise that the need for hospital beds is a moving target. The question of whether we need more hospital beds is a wicked problem in terms of design thinking that is beyond the scope of this article. Internationally, the current debate is on reducing the number of hospitals and downsizing them, and innovating better models of care with minimal reliance on hospital beds (except for intensive care and acute care) as part of sustainability and transformation plans. The emphasis is on creation of a wide range of alternatives to hospital care and on developing more population health approaches using accountable care organisation-type models (Smith 2016).

In the context of this study, where the focus was on a single premier apex tertiary care institute, if optimal bed occupancy is considered as 82% and ALS as 6.6 days (mean calculated from OECD 2019), the efficient TOI is 1.5 and throughput is 45.

More detailed studies are required to analyse the efficiency of public hospitals ranging from district hospitals to community health centres in order to project countrywide desired values of hospital efficiency indicators using Barber–Johnson technique. The Indian Public Health Standards (2012) for district hospitals assume an ALS of five days and expect bed occupancy of at least 80%. However, there are no inputs on TOI and throughput. Data can be collected during monthly follow-up of hospital activity. The tool is simple to use and serves as an explanatory and comparative tool for multiple stakeholders. A comparison between the tables and figures in this article shows that the significance of the information is more readily discernible through the figures rather than the tables. Since the Barber–Johnson diagram requires plotting the length of stay and TOI using standards, it also circumvents the liability of graphs portraying value judgments by using scales of one’s convenience. To get further accurate results, it is also necessary to note periods of closure of wards due to engineering repairs and maintenance, infections, etc(Barber and Johnson 1973). It needs to be assessed if periods of vacation of consultants may alter length of stay of patients and their TOI.

In terms of the limitations, the diagram is only as valuable as the accuracy of the statistics calculated. All the variables in the diagram are expressed as an average. The distribution of the average of each variable should be understood before making policy changes based on the diagram. Fluctuations might occur in small specialities or if plotted for short periods, say, one month. Composite measurements of hetereogeneous activity might obscure the contribution of individual elements. The results of the indicators must also be understood as one aspect of the overall evaluation of a hospital’s performance and outcomes.


AHA (2018): Trendwatch Chartbook 2018: Trends Affecting Hospitals and Health Systems, Washington: American Hospital Association, 2018-aha-chartbook.pdf, viewed on 30 July 2019.

AIIMS (nd): Annual Reports, All India Institute of Medical Sciences, New Delhi,, viewed on 30 July 2019.

Andayani, Ni Luh Putu Eka Putri, et al (2015): “Public Hospital Governance in Indonesia,” Public Hospital Governance in Asia and the Pacific, D Huntington and K Hort (eds), Comparative Country Studies, Vol 1, No 1, Manila: World Health Organization Regional Office for the Western Pacific, pp 26–63, 9789290617112_eng.pdf?sequence=1, viewed on 30 July 2019.

Australian Institute of Health and Welfare (2018): “Australia’s Hospitals 2016–17 at a Glance,” Health Services Series No 85, Cat No HSE 204, Canberra: Australian Institute of Health and Welfare,, viewed on
2 August 2019.

Barber, B and D Johnson (1973): “The Presentation of Acute Hospital In-patient Statistics,” Hospital and Health Services Review, Vol 69, pp 11–14.

Borghans, I et al (2012): “Fifty Ways to Reduce Length of Stay: An Inventory of How Hospital Staff Would Reduce the Length of Stay in Their Hospital,” Health Policy, Vol 104, pp 222–33.

Boyle, J et al (2013): “Probability of Severe Adverse Events as a Function of Hospital Occupancy,” IEEE Journal of Biomedical and Health Informatics, Vol 18, pp 15–21.

Brownell, M D and N P Roos (1995): “Variation in Length of Stay as a Measure of Efficiency in Manitoba Hospitals,” Canadian Medical Association Journal, Vol 152, pp 675–82.

Castoro, C et al (2007): “Policy Brief—Day Surgery: Making It Happen,” World Health Organization on Behalf of the European Observatory on Health Systems and Policies, Copenhagen, 107831.

Clarke, A (2002): “Length of In-hospital Stay and Its Relationship to Quality of Care,” Quality & Safety in Health Care, Vol 11, pp 209–10.

Clarke, A and R Rosen (2001): “Length of Stay: How Short Should Hospital Care Be?” European Journal of Public Health, Vol 11, pp 166–70.

Cunningham, J B, W G Kernohan and T Rush (2006): “Bed Occupancy, Turnover Intervals and MRSA Rates in English Hospitals,” British Journal of Nursing, Vol 15, pp 656–60.

Dalpatadu, S et al (2015): “Public Hospital Governance in Sri Lanka: A Case Study on Processes and Performance,”, viewed on 2 July 2017.

Department of Health (2001): “Shaping the Future NHS: Long Term Planning for Hospitals and Related Services—Response to the Consultation Exercise on the Findings of the National Beds Inquiry,”, viewed on 2 June 2017.

Duran, A, et al (2015): “Public Hospital Governance in India,” Public Hospital Governance in Asia and the Pacific, D Huntington and K Hort (eds), Comparative Country Studies, Vol 1, No 1, Manila: World Health Organization Regional Office for the Western Pacific, pp 222–55, 208218/9789290617112_eng.pdf?sequence=1, viewed on 30 July 2019.

European Commission (2019): “European Core Health Indicators (ECHI) Data Tool,”, viewed on 2 August 2019.

Eurostat (2016): “Hospital Discharges and Length of Stay Statistics,”, viewed on 2 June 2017.

Government of NCT of Delhi (2017): “Economic Survey of Delhi 2016–17,”, viewed on 2 April 2017.

Harrison, M L et al (1995): “Discharging Patients Earlier from Winnipeg Hospitals: Does It Adversely Affect Quality of Care?” Canadian Medical Association Journal, Vol 153, pp 745–51.

Health Economics and Financing Directorate (2012): Cost Analysis of Kabul’s National Hospitals, Ministry of Public Health, Islamic Republic of Afghanistan, November, 325325.pdf, viewed on 30 July 2019.

Indian Public Health Standards (2012): Guidelines for District Hospital (101 to 500 Bedded),, viewed on 30 July 2019.

Jones, R (2001): “Bed Occupancy: Don’t Take It Lying Down,” Health Service Journal, Vol 111, pp 28–31.

— (2002): “New Approaches to Bed Utilisation: Making Queuing Theory Practical,” aches_Bed_Utilisation.pdf.

— (2011): “Hospital Bed Occupancy Demystified,” British Journal of Healthcare Management, Vol 17, pp 242–48.

— (2013): “Optimum Bed Occupancy in Psychiatric Hospitals,”, viewed on 2 June 2017.

— (2016): “Bed Occupancy and Hospital Mortality,” British Journal of Healthcare Management, Vol 22, pp 380–81.

Lastrucci, V, et al (2016): “The Barber–Johnson Technique for Assessing Hospitals Efficiency: The Case of the Republic of Albania,” European Journal of Public Health, Vol 26.

Malone, D, N A Fineberg and T M Gale (2004): “What Is the Usual Length of Stay in a Psychiatric Ward?” International Journal of Psychiatry in Clinical Practice, Vol 8, pp 53–56.

Massabot, N E (1978): “Interpretacion de los indicadoresquemiden la utilisacion de las camas hospitalarias,” Revista Cubana de Administracion de Salud, Vol 4, pp 47–58.

Mayer, V, et al (1994): “Barber–Johnson Analysis of Bed Utilisation in Hospital-based Health Care in the Croatian Republic During Peace and War: 1986–1991,” Lijec Vjesn, Vol 116, pp 65–69.

McGinn, P (2007): “European Ophthalmic Surgeons Maintain Relentless Momentum to Perform More Operations More Quickly,” Special OECD Report,

Ministry of Health and Family Welfare (2011): Operational Guidelines: National Programme for Health Care of the Elderly (NPHCE), Government of India, _NPHCE_final.pdf, viewed on
30 July 2019.

Morera Salas, M (2013): “Barber and Johnson Diagram for the Analysis of the Hospital Bed Management in Costa Rica,” Revista Costarricense de Salud Pública, Vol 22, pp 45–50,

National Health Performance Authority (2018): “My Hospitals,” Australian Institute of Health and Welfare, Australian Government,, viewed on 2 August 2019.

NIMHANS De-Addiction Centre (2009): “Alcohol, Drug Misuse and Other Addictions: What to Know and How to Get Help,” Bangalore: National Institute of Mental Health and Neurosciences,, viewed on 2 June 2017.

Nuffield Trust (2016): Understanding Patient Flow in Hospitals, London: Nuffield Trust.

OECD (2013): Health at a Glance 2013: OECD Indicators, Organisation for Economic Co-operation and Development, Paris: OECD Publishing,

— (2019): “Length of Hospital Stay (Indicator),” doi:10.1787/8dda6b7a-en, viewed on 2 August 2019.

Pabón Lasso, H (1986): “Evaluating Hospital Performance through Simultaneous Application of Several Indicators,” Bulletin of the Pan American Health Organization, Vol 20, pp 341–57.

Premik, M et al (1999): “Bed Utilisation Performances of Slovenian and Croatian Acute Hospitals Systems,” Studies in Health Technology and Informatics, Vol 68, pp 606–9.

Rosen, B, R Waitzberg and S Merkur (2015): “Israel: Health System Review,” Health Systems in Transition, Vol 17, pp 1–212, /Israel-HiT.pdf?ua=1.

Royal College of Psychiatrists (2010): “Looking Ahead—Future Development of UK Mental Health Services: Recommendations from a Royal College of Psychiatrists’ Enquiry,” Occasional Paper OP75, Royal College of Psychiatrists, London.

Silversides, A and T Sullivan (2013): “Hospital Crowding: Despite Strains, Ontario Hospitals Aren’t Lobbying for More Beds,”, viewed on 2 July 2017.

Smith, J (2016): Sustainability and Transformation Plans (STPs): What, Why and Where Next?, IPPR,, viewed on 15 October 2018.

Sprivulis, P, et al (2006): “The Association between Hospital Overcrowding and Mortality among Patients Admitted via Western Australia Emergency Departments,” Medical Journal of Australia, Vol 184, pp 208–12.

Tangcharoensathien, V et al (2015): “Public Hospital Governance in Thailand,” Public Hospital Governance in Asia and the Pacific, D Huntington and K Hort (eds), Comparative Country Studies, Vol 1, No 1, Manila: World Health Organization Regional Office for the Western Pacific, pp 138–85,, viewed on 30 July 2019.

Tran Thi Mai Oanh et al (2015): “Public Hospital Governance in Viet Nam,” Public Hospital Governance in Asia and the Pacific, D Huntington and K Hort (eds), Comparative Country Studies, Vol 1, No 1, Manila: World Health Organization Regional Office for the Western Pacific, pp 98–136, sequence=1, viewed on 30 July 2019.

Tulchinsky, T H and E A Varavikova (2000): The New Public Health: An Introduction for the 21st Century, San Diego: Academic Press.

Weiss, A J and A Elixhauser (2014): “Overview of Hospital Stays in the United States, 2012,” HCUP Statistical Brief #180, Agency for Healthcare Research and Quality, Rockville, MD,, viewed on 30 July 2019.

Westert, G P and R J Lagoe (1995): “The Evaluation of Hospital Stays for Total Hip Replacement,” Quality Management in Health Care, Vol 3, pp 62–71.

WHO (2014): “Mental Healthcare Delivery,” Global Health Observatory Data, World Health Organisation, _delivery/beds_hospitals/en/, viewed on 2 July 2017.

— (nd): “Health Care-Associated Infections: Factsheet,” World Health Organization, _fact_sheet_en.pdf, viewed on 2 July 2017.

World Health Organization Europe (2016): Comparative Key Figures, European Hospital and Healthcare Federation, European HFA Database,, viewed on 2 August 2019.

Yates, J (1982): Hospital Beds: A Problem for Diagnosis and Management? London: Heinemann Medical Books.

Updated On : 13th Sep, 2019


(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top