[The author has visited India twice - once in the year 2000 and then in 2003 - to discuss ICT in the public health sector and has also studied the India Health Plan 2002 document. He understands why it has been so difficult to get started in this sector on a national basis in India. As of today, India's 14,000 public hospitals still rely massively on manual paper-based systems. But most of the private hospitals (some 10,000) do have fairly good ICT environments. The question is why? The answer is fairly obvious – economics, which unfortunately is not being followed by the ‘consultants’ or other advisors to the health sector, leading to the many ICT failures that have occurred, and which will continue to occur. This is a white paper prepared by him for the WHO Health Metrics Network (HMN) initiative on the subject: ‘The Economics of Health Informatics’ for developing countries. The paper addresses various problems faced by India in this sector, and also gives suggestions to improve the overall condition and make ICT measures more effective with the help of better patient administration systems in the hospitals.]  

OVERVIEW

This paper will hopefully lead to a better understanding as to how India can take the first successful step towards achieving e-Health and an Electronic Health Record (EHR) through an Affordable and Sustainable approach recognized by WHO, PAHO and CDC.

The computerization of India’s health sector, hospitals and clinics, is indeed a most challenging task. However it has to be undertaken and it is better to do it sooner than later. The patient benefits, and efficiencies, are many. The lives of citizens rely on good health care, and good health care relies on information, which computerization gathers in an accurate and timely fashion. A currently topical example is Disease Surveillance[1] (TB, HIV/AIDS, Injury, etc)

The failure rate of public health computerization is extremely high - as high as 75% - according to the BMJ (British Medical Journal, 2003). Would anyone fly in an airplane known to have a 50% failure rate! Why an HIS?

Why would any hospital, region, province, state, or country initiate a health IT program that from the outset has been proven to have staggering odds against ever being a success in a developing country?

There is no reasonable or rational response to this question. The possible motivators would include:

• Lack of knowledge on the part of government and hospital officials as to the reality of health IT systems in other developing countries.

• Too much reliance on outsiders (from developed countries) who are either incompetent or lacking in knowledge as to the resource realities of DCs

• Officials who are dishonest, and who are involved in bribes and payoffs

• Pressure from international financial institutions, which it appears do not evaluate the sustainability of funded projects for failures continue to happen time and time again.

INTRODUCTION

Here we are in early 2008 and it can be reasonably stated that most developing countries (151) are looking at computerizing their health sectors. This sector remains very far behind the commercial sector in the use of computerization in their own countries, and also very far behind health ICT (Information and Communications Technology) [or IT] in the developed countries.

These countries are as interested in achieving e-health as are the developed countries. E-health permits the electronic exchange of data between national health care facilities, a requirement now being addressed by the developed countries after almost thirty-five years since computerization was first initiated in the health sector. The in-place health IT systems in developed countries are now being replaced, as in the United Kingdom, where a national US$ 20 billion e-health initiative has been underway since the late 1990s. The ultimate goal is to have an EHR (Electronic Health Record), which is the repository of the many bits of patient data generated by numerous systems serving the patients of the country.

The following are a few definitions that will help to define the nature of the comments on health IT that are presented in this paper:

• Public Health: All comments relate to the public health sector – not private. In most countries, citizens can access both public and private health facilities.

• Developing Countries (DCs): Unless otherwise stated, all comments relate to DCs. A DC is often defined as having a GDP per capita of less than US$ 10,000 (PPP).

• University Hospitals: Such large hospitals, even in DCs, have a much higher per bed operating budget compared with other public hospitals, and accordingly are not included in any comparisons. They may have extensive HIS implementations.

• Private Hospitals: These hospitals, which are numerous, if not dominant, in many developing countries, are not relevant in this document.

• HIS: In all developed countries, the governments have attempted to achieve the implementation of a Hospital Information System in every health facility of the country, and have mainly achieved this objective. An HIS includes about 20 key application software modules, covering both the financial (accounting, etc.) and clinical (patient registration, etc.) areas of operation.

• PAS: A component of an HIS is the Patient Administration System, which is a suite of modules that constitute about 15% of an HIS. It can be described as the key starting point that can be built upon to achieve an HIS.

• National e-Health: The intent today in all countries is to have a national e-Health implementation, based on extensive computerization of each health facility. This leads to an EHR (Electronic Health Record), which enables informed treatment of patients at the point of care.

When Health Ministries in a developing country look at the many health-related applications, it usually boils down to a choice between two philosophical concepts as to how to best start in order to get this important initiative underway. The choices are:

• To start with an HIS, often from a single vendor, that involves the implementation of more than twenty application software modules over a period of a few years. A partial list of these many application software modules is enclosed. Or,

• To start with the PAS module, the core HIS module, and implement the PAS in all public hospitals and clinics within the country before moving on to the next application software module (Clinical Laboratory, Pharmacy, Radiology, etc.).

Let us take a pragmatic look at these two options for developing countries.

The HIS Approach

Since this is the desired end-objective, it is clearly a reasonable objective to set out at the outset the goal of achieving a Hospital Information System (HIS). As with most situations in life, there are arguments that can be raised for, and against, this starting option. We have observed over thirty years that this is not a good option, and present some background information in support of this stated position by asking some questions, and by providing answers.

1.      What is the chance of success when a country starts its health IT with an HIS?

Zero, or very close to zero. Certainly not recommended nor a good gamble. An HIS implementation plan, when compressed into a typical three to five year plan, does not have any chance of success, so should never be started.

2.      How and when did the developed countries start their health IT?

They started in the late 1960s and early 1970s with financial applications like G/L, Payroll, Billing – A/R, Inventory, etc. The advent of ‘micro-computers’ enabled this computerization as the few ‘mainframes’ in use in the late 1960s at the large university hospitals (500 beds and up) were beyond the budgets of the majority of hospitals, being the small and mid-sized hospitals.

3.      Did the DCs start their health IT with the HIS approach?

No, they did not ever start with the HIS approach. Such an approach was not available until about twenty years later - in the late 1980s.

4.      Why did the hospitals in DCs not start with an HIS before 1980?

This answer has multiple components, namely:

• The clinical application software modules (Laboratory, Pharmacy, for example) did not exist.

• Hospitals were seriously challenged just to implement the financials, most of which are integrated into other financial modules.

• Hospital budgets could not accommodate additional expenses, as the vast majority of hospitals were trying to implement computerization within 1% of their operating budget.

• In the early 1970s the available computers were too limited in capacity (10MB hard disk drive) to even hold a CPI (Central Patient Index) of 100,000 patients.

 5.      Over what period of time did the DCs implement their PAS/HIS initiatives?

Developed countries, and in particular the 1,000 Canadian hospitals, started to implement their PASs in 1978, and by 1990 only a few very small hospitals remained with completely manual paper-based systems for registration. The last Canadian hospital to implement a PAS did so in 1996, some 18 years after the first hospital, and 25 years after health IT started in Canada.

Once the PAS implementations were operational Canadian hospitals started to implement one-by-one, in no particular sequence, the various software components of an HIS, and this started in the early-1980s, and continues to the present time. A fair estimate of the number of Canadian hospitals that have a complete HIS (90%+ of the listed modules) operational by 2007 would be about 20%, or about 200 hospitals.

6.      Of Canadian hospitals with an HIS, how many have a single vendor solution?

It is not known that any of the 1,000 hospitals have all of their application software modules from a single vendor. Many hospitals do have a majority of their application software modules from a single vendor but they also have software modules from other specialty vendors.

7.      How is integration achieved with software modules from different vendors?

Through the use of middleware, which is being used by virtually all Canadian hospitals to achieve the desired integration. International experts (‘Integrating the Healthcare Enterprise’ speakers) in this field have stated, in many presentations, that in the USA today there are no entirely ‘single-vendor’ IT solutions operational in the country’s 5,500 public and private hospitals. There is also limited integration within a hospital.

8.      Why have developing countries failed in their many attempts with an HIS?

There are many factors contributing to such failures like:

• Cost (GDP per capita is much lower than in the developed countries)

• Speed (going too fast - time is needed for each module to be stabilized)

• Many modules are based on earlier modules, which must be operational first.

• Vendors are writing software ‘on-the-fly’, which is then not field-tested.

• Consultants often specify an HIS from their health IT knowledge in developed countries. The consultants then move on to new opportunities and assignments.

• The key government and hospital officials have limited, if any, background knowledge on which to build. In Canada, such officials now have over thirty years of experiences, some bad, and some good, on which to build and make informed decisions going forward.

 9.      What are some of the problems in contracting for an HIS with any vendor(s)?

There are many problems in this matter, some among them being:

• The module timelines will not easily be achieved, causing friction between hospital officials, department managers, and between vendors, if multiple vendors are involved. Often a particular application module relies on data being received from, or being sent to, another application module.

• It does not appear to be very feasible or acceptable, to stretch a contract implementation plan out over ten to twenty years (which is required). So, the powers that be attempt to put the HIS completely in place within three years or so, a timeline which has been shown to be impossible.

• Technology changes over time so the application software modules that are to be implemented beyond ten years may be quite different then, compared to now. This being the reality it is not wise to contract today for an application that will only be implemented ten years from now.

The PAS Approach

The core module within an HIS is the Patient Administration System, to which all subsequent clinical application software modules will be connected. Accordingly, it is the first software module to be implemented, and there need not be a fixed timeline before moving on to the subsequent application software modules.

1.      What is the chance of success when a country starts its health IT with a PAS?

Very high - close to one hundred percent (100%) - if managed adequately. The PAS implementation has no mandatory deadline, so the challenges that are encountered along the way can be attended to without involving the many other hospital departments, and having to interface to other software modules.

2.      What are the application software components of a PAS suite?

A PAS constitutes about 15% of a complete HIS. There are many components within a PAS, and the order of implementation can be changed to suit the availability of key personnel, financial resources, and to suit the priority of the different department managers. The key PAS application modules are:

• Central Patient Index (Master Patient Index)

• Registration

• Outpatients

• Billing and Accounts Receivable

• Accident and Emergency Department

• Health Records Abstracting

• Surveillance (Injury, TB, HIV/AIDS, etc.)

3.      When did most developed countries start to computerize their PASs?

In Canada, as in the USA and Europe, most hospitals started to computerize their operations around 1970 with the financials, and billing in USA hospitals. It was not until the very late 1970s that Canadian hospitals (1978) started to implement the Patient Administration System.

4.      Is it prudent to implement the PAS concurrently with other clinical modules?

No, it is not a suggested approach. It has been observed that this key core module, the PAS, can be successfully implemented best when there are no outside pressures to concurrently implement additional clinical software modules. These other modules must be integrated to the PAS, so the PAS must be very stable first before being connected to other modules. However, the PAS can be concurrently implemented with some financial modules, for the General Ledger (G/L) is often desired early on, and the PAS provides the G/L with certain data, and it is desirable that such data be verified by G/L account codes.

5.      How long would it typically take to implement PAS, and at what cost?

There are numerous factors involved in addressing this matter, among them:  

• The number of public hospitals and clinics in the country:

• The resources (personnel and financial) that the government applies to the initiative are very important, and often dependent on annual budgets.

• The continuity of the key people overseeing the initiative is critical.

• The initiative must have a high-level executive ‘cheerleader’ who takes control when problems occur - for there will always be problems.

• The status of the hospital infrastructure (air conditioning, wiring, etc.)

• In summary, in a country with 20-200 public hospitals it should be possible to implement the PAS successfully in all hospitals within 3 to 5 years.

• The cost of a PAS implementation in a specific country, from the software perspective, will vary based on the GDP per capita of the country. An example of the up front costs can be deduced by looking at the country’s current health budget. The national PAS license fee would constitute about 0.5% of this budget, amortized over 10 years. When the required training costs, plus the necessary hardware and on-going support fees are included, the annual PAS expenditure would be about 1% of the health budget. In Canada today about 3.0% of the health budget is spent on IT (for an HIS), or about US$ 60.00 per capita. Allocating 1/6th to the PAS component we see that about $10.00 is for the PAS. In Jamaica, the current PAS IT cost is in the range of $0.73 per capita, less than 10% of what Canada spends.

6.      What application software module comes next after a successful PAS implementation?

The choice as to the next application software module will vary greatly from country to country. It appears that the two most needed modules after the PAS are Pharmacy and Clinical Laboratory, in no particular order.

7.      How long might it take to achieve a complete HIS/EHR implementation?

The length of time required will vary greatly from country to country, and will be mainly dependent upon the resources available, as noted above. The minimum time is estimated at being about ten years, with the more likely time frame being in the order of twenty years. It may then be time to start all over, unfortunately, for the rate of technology change is rapidly accelerating. The use of ‘Open Systems’ technology will mitigate the risks involved.  

Annexure-I: Modular Software Components within an HIS  

Once the above fundamental PAS information modules are fully operational, progress can be made with the many financial and clinical modules required to provide for a fully integrated HIS. The main follow-on modules are: