[This article is an abstracted version of the white paper titled ‘The Economics of Health Informatics’ prepared by the author for the WHO Health Metrics Network (HMN) initiative. It addresses issues pertaining to costs of implementing the HIS and PAS systems in developed and developing countries such as Canada and Jamaica respectively, replication of HIS in public and private hospitals of developing countries, and highlights a few lessons from ICT based health initiatives functioning in different countries.]

The computerization of your country’s public health sector, hospitals and clinics, is indeed a most challenging undertaking. However, it has to be undertaken and it is better to get started sooner than later. The patient benefits, and the efficiencies too are many. The risk of doing nothing is no longer an option. Tourists are a very important source of income to most countries, and the tourists expect efficient and modern health systems in countries where they care to vacation.

In the developed countries (about 40) there is a long history of health informatics that dates back to the early 1970s which was when Canada started to implement on-site computers in hospitals. However, in the developing countries today there is very little public health computerization. The World Health Organisation (WHO) has recognized this unfortunate situation and published in 2006 a manual to be used by developing countries for the start of an Electronic Health Record (EHR) system. In addition, the Health Metrics Network (HMN), a division of the WHO, published the second edition of the Frameworks and Standards for Country Health Information Systems in 2008.

The biggest challenge for the consultants and the advisors of the health sector today is to acquire a better understanding of the economics of e-health. ICT failures in the health sector of developed and developing countries can be avoided provided the role of economics in getting ‘started’ on the road to a national Health Information System (HIS) and EHR is recognized and understood.

E-health permits the electronic exchange of data between national health care facilities, a requirement being addressed by the developed countries now after almost thirty-five years since computerization was first initiated in the health sector. The ultimate goal is to have an EHR, which is the repository of the many bits of patient data generated by numerous systems serving the patients of the country.

 1. What are the known costs of an HIS/PAS in developed/developing countries?

 The following figures are very close approximations of the costs of health ICT:

§  $263.00 - Annual per capita cost of a National Health Information Network[1] (NHIN) in the USA

§  $110.00 - Annual per capita cost of an HIS in the USA

§  $87.00 - Annual cost per patient at the Veterans Administration EHR implementation in the USA

§  $74.00 - Annual per capita cost of an HIS in Canada

§  $12.40 - Annual per capita cost of a PAS in Canada

§  $1.00 – Annual per capita cost of PAS in Jamaica

The costs of an HIS or a PAS or an EHR in the developed countries is extremely high compared to the cost to be supported (by the health budget) of an HIS, PAS or EHR in a typical developing country. For instance, Jamaica (GDP per capita of US$3,400) experiences an annual per capita sustainable cost much lower (about 80% lower) than Canada ’s for their PAS implementation. Were a developing country, such as Jamaica (with a population of 2.7 million), to attempt to implement an HIS under an initial contract, say over three to five years, it would have to commit about US$200,000,000 per year (based on the Canada cost of $74.00 per capita annually). This is clearly unsustainable out of an annual health budget of about US$400 million – leaving virtually no funds for staff or anything else.      

2. Can an HIS from a developed country function in the public hospitals of a developing country?

In the developed countries health ICT started in the early 1970s, and has continued to be enhanced every year since then through the addition of new application software modules – often noted as now numbering thirty or more modules. This has been a very expensive process, and very challenging to the budgets of the hospitals.

Since the GDP per capita in the developed countries is typically about ten times greater, on average, than the GDP per capita in a developing country (see table below), it can be understood why an HIS cannot just be implemented at about 10% of the cost already being experienced in all of the developed countries.

GDP per capita of selected developed and developing countries

Source: WHO, 2006

While the GDP per capita as noted may be on average about ten times higher in the developed countries, the per capita spending on health is often much larger than ten times more in developed countries than in developing countries (see table below). According to the WHO, Canada spends $2,402 (PPP) per capita annually on public health whereas India’s per capita spending on public health is $19 (PPP) – a factor of about 125:1. Individual countries allocate different percentages of their government budgets to the health sector – resulting in very large discrepancies between countries in what funding is available for public health, and accordingly for health informatics in public facilities.

Annual public health spending per capita in selected countries

Source: WHO

Keeping the above budgetary considerations (economics) in mind, it is not possible to simply duplicate an HIS implementation from a developed country in a developing country. It might function partially for a short while but will soon become unsustainable. An example of this is the case of the HIS implementation in Limpopo Province, South Africa in 1998. Limpopo with a population of 5 million is a relatively poor province (GDP per capita of about US$2,500) in a developing country SA (GDP per capita US$4,960). The per capita annual public health expenditure in Limpopo, is about US$157 – compared to US$2,410 in Canada, and US$83 in Jamaica.

The Limpopo HIS implementation started in 1998, and by 2002 it was cancelled, resulting in a reported loss of about US$50 million – the largest reported health ICT loss in a developing country.[3] The application software modules were provided by Medicom, a software firm located in Bangalore, India. This firm had apparently successfully implemented the same HIS modules in the Al Jazira Hospital, the Mafraq Hospital , and Madinat Zayad Hospital in the UAE, as well as several other hospitals in the region. However, there are significant economic differences between hospitals in the developed country of UAE (GDP per capita - US$23,770) and the developing Limpopo Province in South Africa (GDP per capita - US$2,500.) – a factor of almost 10 to 1.

3. Can a private hospital HIS operate in a public hospital in a developing country?

India is an example where ICT has been in place in the private hospitals for years – however not yet (2009) in the public hospitals. During business visits to India in 2000 and 2003 the author observed that the private hospitals (about 10,000) in India appeared to be relatively well computerized. On the other hand, the public hospitals (about 14,000 in number) were still dependent on manual paper-based systems – in a country that has an international reputation for ICT excellence.

In order to compare the sustainability of ICT costs in each environment, let us contrast the ICT costs at the private Bumrungrad International Hospital (Bumrungrad IH) in Bangkok , Thailand with the public hospitals of Thailand , Jamaica and Canada . The Bumrungrad International Hospital is the largest private hospital in South Asia with 554 beds. It is also a corporation listed on the stock exchange, where it is reported that the revenues in 2007 were 9.4 billion bhat – the equivalent of about US$268 million. Bumrungrad IH is today highly computerized, with a complete HIS and EHR, to serve the more than one million patients from 190 countries visiting it every year. The annual expenditure on ICT would be in the same range as a large teaching hospital anywhere, which is typically around 3.5% of the operating budget. In the case of Bumrungrad, this would be about US$8 million per year or about US$14,400 per bed annually. This contrasts sharply with other jurisdictions as shown in the following table (all figures are best estimates):

Annual ICT spending (per hospital bed) in selected jurisdictions

Accordingly, the Bumrungrad IH, Thailand, is in the same cost range for an HIS as an average public hospital in Ontario, Canada – a developed country with a high GDP per capita. Jamaica and Thailand are also in the same much lower ICT cost range – and they are both developing countries. Thailand could not afford to spend US$14,400 per bed annually on ICT in the public hospitals, and does not use the Bumrungrad HIS system in any public hospital – instead, the 50 computerized public hospitals in Thailand utilize the license-free Linux based operating system.

Were Thailand to implement the Bumrugrad HIS, using the Canadian annual HIS cost of US$74 per capita, Thailand (population 63 million) would incur an annual HIS expenditure of about US$4.6 billion – or about 35% of the country’s annual public health budget of about US$13 billion.

4. Some lessons learned from the Limpopo, Jamaica ’s PAS and UK NHS CfH (Connecting for Health) initiatives

A brief comparison between the failed HIS ICT attempt in Limpopo, and the successful ICT ‘start’ with a PAS in Jamaica – both developing countries close in size and economic capability reveals several key lessons:

Where to start?

Do not start with an HIS implementation – as in Limpopo

Do start with a PAS (1/6 of an HIS) implementation – as in Jamaica          

How fast should a country proceed to achieve an HIS?

Do not contractually set a three (or four, or five) year end-date – Limpopo

Proceed with a single module - PAS, and get it operational – Jamaica

Economics: What should the financial commitment be at the outset?

Do not start at 2.5% of the annual health budget – Limpopo

Start at under 1% of the annual health budget – Jamaica .

When should the next clinical application (after PAS) module be implemented?

Not concurrently with the implementation of other modules – Limpopo

After the implementation of the PAS – Jamaica (10 years later). Jamaica has now in 2009 implemented a Laboratory Information System.

Should the country contract for all software modules (over 20) from a single vendor?

No. This is an old (1980s and 1990s) concept – as tried in Limpopo .

Both IBM and Unisys exited forever the vendor marketplace in 1995 – and both had offered the ‘single-vendor’ ALL software and hardware approach.

With the many problems today in the UK NHS CFH initiative, hospital trusts are moving to the multi – vendor ‘best-of-breed approach’ as being the safest and most flexible approach in achieving an HIS over time – as in Jamaica.

Comparison of the economics of ICT implementation in two developing countries (with reference to Canada)