LARGE-SCALE COMPUTERIZATION - THE CURE FOR THE HEALTH-CARE CRISIS
Edition 4, May, 2006
by
Bruce Sundquist
bsundquist1@windstream.net
|
- - -The Only Cure? |
|
- - -The Permanent Cure? |
* Previous Editions: August, 2004// November 2004// April 2005// May 2006
Go to "Inefficiencies in the U.S. Health Care System"
~
Go to Home Page of this Web Site
Despite health care's huge and ever-increasing bite out of the U.S. GDP, stopgap and increasingly inefficient measures have characterized the bulk of the nation's response. Cost transfers, rather than efficiency improvements, appear to be the focus of these measures. This is in spite of many studies over some decades pointing out numerous inefficiencies in the system -inefficiencies largely ignored. This document focuses on the most promising mechanism for improving health care system efficiency - large-scale computerization of the system's information flow-, analysis- and storage processes by a single software source producing open-source software. Reasons why this has never been attempted are analyzed here. A few billion dollar's worth of one-time software development, plus a few tens of millions of dollars in annual management, maintenance-, development- and update costs thereafter, could produce direct benefits of roughly $470 billion dollars annually. Numerous opportunities for improving the quality of health care would also result. Other benefit categories, in total, could produce comparable benefits. Arguments here support the contention that large-scale computerization offers the only cure, and probably the only permanent cure, for the health care crisis. The shortfall of 200,000 doctors and 800,000 nurses projected for 2020 could be eliminated by this computerization.
INTRODUCTION
An analysis (Newhouse. 2004) offers compelling arguments for the contention that government subsidies are unlikely to even begin to keep up with the double-digit growth of health care costs, even in the near future. The combination of escalating health care premiums and competitive pressures from globalization is forcing employers to transfer a rapidly increasing portion of health care costs to employees. All this raises the likelihood of public anger, frustration and desperation forcing the US to follow the rest of the industrialized world into socialized medicine. At least it seems clear that, over the next decade, the public will come to the realization that the strategy of seeking symptomatic relief in terms of ever-increasing government subsidies for health care is fundamentally flawed. In a "charge-what-the-market-will-bear" environment, the response to subsidies is faster rates of price increases. This reality cannot help but shift the public focus to cost-containment, particularly containment strategies that involve little or no degradation in quality of health care.
All this is in spite of numerous studies over some decades pointing out numerous inefficiencies in the health care system -inefficiencies that have been largely ignored. Studies noted in the mass media in mid-2003 pointed to various computerization strategies as the most promising means of increasing health care labor utilization efficiency. Some would argue that computerization-related efficiency improvements are getting attention, noting that there are now about 10-20 mid-sized "information technology" (software) companies involved in health care system software, plus 300 or so small ones (J. Harvey. 2004). Arguments below contend that this is a tragically futile approach, since only large-scale computerization (defined below) involving system-wide, single-source software is capable of achieving major increases in efficiency. Others might argue that if large-scale computerization were such a great cost reducer, it would occur as part of the normal profit-maximization process. Reasons why this is not occurring, or even possible, are given below.
The futility of the current approach to health care system computerization is also clear from the extremely slow rate of progress in this area despite the huge amount of attention being given to the topic in the health care literature and the large number of companies involved in health care software. Dr. David J. Brailer, who President Bush appointed to be the national health information technology coordinator, expressed his frustration at the lack of progress by delivering a warning to the health care industry: take steps soon to make it happen, or the government will probably impose a solution (Lohr, 2005). The problem is that the overwhelming bulk of the health care computerization literature, the healthcare software industry, the health care industry itself and Dr. Brailer all have a common failing. They failed to first develop a concept-level picture of what a fully computerized health care system should look like. The current approach cannot possibly produce anything but a massive, confusing, non-functioning array of individual software systems, each of which is incompatible with dozens of other systems that information must pass to and from. Also countless gaps and overlaps are almost certain to occur between/ among the individual software systems. And the huge number of different "looks and feels" of different software packages will make life miserable for health care workers.
What is required is a single source for a total health care system software package, with the remainder of the health-care system software industry working as sub-contractors (See below.) But under present circumstances this is physically impossible. Imagine what might happen if Dr. Brailer ordered the hundreds of health care software companies to pick one of their number to serve as the "single source." Some entity in a leadership position must seek out that single source via some sort of competitive bidding process. That leadership entity does not currently exist, and it probably won't until Dr. Brailer or the Congress gets really frustrated. It may take some years before a huge pile of incompatible software packages is trash-canned and we start over with a single source. It may then take some more years to realize that we have a self-serving, indifferent, wildly expensive software monopoly on our hands and start over once again with a single-source developer of open-source software. And during those lost decades the health care crisis can only worsen (See the last few pages of this document.), with terrible repercussions for the physical and financial health of Americans. The purpose of this document is to develop a concept-level picture of what a fully computerized health care system should look like, and hopefully eliminate the nightmare scenario just alluded to.
Go to Table of Contents ~ Go to Reference List ~
Go (Return) to Home Page of this Web Site
CONTEXT
The computerization process and system proposed below would significantly change the day-to-day way of doing things for nearly everyone in the health care system. Much stress and turmoil would be expected during the changeover. For this reason, instituting additional efficiency improvements should be delayed until the large-scale computerization process has plateaued to a new, routine, way of life for the system. Large-scale computerization could demonstrate the benefits of efficiency improvements relative to the marginal and costly benefits to be expected from ever-increasing tax-payer-financed subsidies for health care that are unlikely to occur (Newhouse. 2004). This could trigger efforts to initiate these other major efficiencies. That is why large-scale computerization is broken out of the larger paper [Sundquist, 2004] into this separate document.
LARGE-SCALE COMPUTERIZATION
THE NEED FOR LARGE-SCALE COMPUTERIZATION Studies reported in the mass media in mid-2003 also noted significant declines in efficiency over the years. HMOs, Medicare, Medicaid and SCHIPs have been forced to set up increasingly elaborate rules and limitations on medical procedures and medications that they will cover when billed by direct providers of medical services. This has created major increases in the size and complexity of the paperwork burdens imposed on both health-care providers and insurers. But even worse, it has resulted in huge staffs at these insurers that do nothing but check on compliance with regulations. An October 2003 report in the mass media noted that the number of people on these staffs (Medicare + Medicaid + SCHIPs+ HMOs) now exceed the number of people providing direct health-care services. Thus for every hour of direct health-care service provided, at least another hour is spent making sure that everything done in direct service and every medication prescribed are within imposed limitations. All this is in spite of the fact that only a tiny fraction of bills from direct health service providers are examined, resulting in over $100 billion/ year (as of a decade ago) in losses due to fraudulent and questionable claims.
Why doesn't the health-care system just computerize its information flow-, analysis- and storage operations, the paperwork of direct health-care providers, and all those compliance analyses and fraud-seeking analyses of those growing hordes of HMO, Medicare, Medicaid and SCHIP compliance people? Doing so could reduce the direct-labor intensiveness of the health-care system significantly. Automating HMO compliance checking and Medicare/ Medicaid fraud/ abuse checking would be a significant part of this, as would reductions in the paperwork required of direct health care providers. All this ought to be occurring as a normal part of profit maximization, but it is not. There must be barriers that have, thus far, prevented computerization. Until these barriers can be identified and removed, recommending computerization of the health-care system's information flow, analysis- and storage systems is pointless. Below is an attempt to identify these barriers. Barrier removal is addressed after that.
Go to Table of Contents ~
Go to Reference List ~
Go (Return) to Home Page of this Web Site INTERFACES: THE "WHY" IN THE LACK OF LARGE-SCALE COMPUTERIZATION
The result of all this is that the impact of computers on health care system efficiency is still minimal. Nurses and doctors still spend about 40% of their time on paperwork and related activities, about the same percentage as before computers were in common use.
If all non-electronic interfaces and all software incompatibilities could be eliminated, and if all overlaps between competing software systems could be eliminated, and if all gaps between competing software systems could be eliminated, huge efficiency increases could be achieved by this idealization referred to here as large-scale computerization. One hospitalization can produce tens of thousands of separate observations, and that number is growing so rapidly that non-electronic information flow-, analysis- and storage processes are being severely challenged (McDonald. 1998). Discussing computerization issues with health-care system software people reveals numerous barriers that are impossible for any one person, or any one local health-care provider, or any one hospital or any one state- or national health-care organization to overcome. In other words, large-scale computerization cannot be achieved incrementally, e.g. as part of an individual hospital's efforts to maximize profits. Instead it requires well-planned, well-organized efforts on a national, system-wide scale. Computerizing one small part of a health-care operation cannot produce much, if any, increase in efficiency of information flow, analysis and storage. The result would most likely be a more laborious, messy, error-prone system plagued increasingly by interface proliferation. Connecting all the individually computerized components of the health-care information flow-, analysis- and storage system makes the problem worse. For then software incompatibility problems pop up. Software A in one component cannot "speak" to Software B in another component, so when someone connects everything together in hopes of reducing the number of interfaces that have been multiplying, chaos results. Also the electronic system will be found to be full of gaps and overlaps among software systems. To give an idea of the potential magnitude of this chaos, note that there are now about 10-20 mid-sized "information technology" (software) companies involved in health care system software, plus 300 or so small ones (J. Harvey. 2004). Huge amounts of information gets transferred from one hospital to another, from one medical office to numerous hospitals, and from doctors in one part of the US to doctors in other states or nations. In such an environment, chaos could easily reign in a sea of interfaces. Leadership: Adding translation software to facilitate electronic interface-crossing adds even more interfaces, more complexity, more errors, greater difficulty in upgrading software, more difficulty in correcting software errors, and intractable blame-games between software vendors when errors are discovered. A single-system approach is essential - not just in each medical office, billing department, hospital or in each region. This requires that the "leadership" of the health-care system seek out a single information flow-, analysis- and storage software system for the entire system from a single software source. At this time, even the necessary leadership does not exist. No wonder large-scale computerization has never happened. Progress and problems associated with computerization of health care systems were reviewed in a collection of articles in Health Affairs (Moran. 1998) (Kleinke. 1998) (Kelly.1998) (Kendall and Levine. 1998) (McDonald. 1998). Problems associated with the leadership vacuum, shifting public policies, and rapid technological advances were referred to repeatedly. Kendall and Levine argued that a new strategy and a new type of organization were needed for dealing with health care system computerization. They also argued that short-term concerns were undermining potential long-term gains. This analysis supports all three of these arguments. The health care system seems ripe for change. The need for national-scale (government-mandated) standards is being increasingly recognized, and a number of such standards have been developed (McDonald. 1998). Large-scale computerization would be impossible without these.
The obvious drawback to single-source software is that it gives someone a monopoly, enabling them to effectively hold the health care system for ransom, and to act accordingly in terms of prices, errors, responsiveness, innovativeness, commitment, etc. The monopoly problem is taken up after interface minimization is examined.
Go to Table of Contents ~
Go to Reference List ~
Go (Return) to Home Page of this Web Site
NON-ELECTRONIC INTERFACE MINIMIZATION Less than that brings out inefficiencies, opportunities for error, and interface proliferation. Feature (1) can be accomplished by every direct provider of health-care services having a little electronic "point-of-service interface device" (POSID) (perhaps similar to those "dyads" carried by trucking company delivery folk) that would probably hang from a belt into a pocket. Information would be efficiently transferred from the point of provider-patient contact into the POSID via efficient drop-down lists etc. on a screen so as to require only a few key taps or mouse-clicks to enter data. With a single set of specifications for POSIDs, the medical equipment industry would develop ways of directly transferring information from measurement devices directly to the POSID, eliminating more non-electronic interfaces. For safety, information transferred would need to also appear on the POSID's screen. POSIDs would plug into a desktop computer, e.g. at a central nursing station in, say, in a hospital wing, to upload and download information. Uploaded information would then be in a vast electronic world where it would never again have to be manually removed from electronic form until it is needed for a non-computerizable action. Nor would it be reentered into electronic form. It would automatically move to the patient's medical record file (electronic), food services, billing, administration, accounting, medical supplies storerooms, room assignments departments, primary care physicians, pharmacies, the patient's Email in-box, and medical specialists with few, if any, additional key-strokes or mouse-clicks. Under the current health-care system of multiple software vendors and incompatible software, every point-of-service health-care provider would currently need to carry one POSID for each software system that the electronic information is ever likely to encounter in its travels - dozens in today's world. Such a system would collapse of its own weight. Medical Records: Medical records of patients are currently stored in typically thick manila file folders, often filled with hundreds of pieces of paper. These files are typically stored in rooms full of people filing and removing them to supply the needs of primary care physicians, specialists etc. Photocopying costs can be huge and labor-intensive. When someone changes health-care providers, the original provider simply photocopies a few of the most recent records for the new health-care provider and sends the remaining data to an old limestone mine somewhere where its future accessibility is largely hypothetical. All that crudeness, mismanagement, inaccessibility and costliness would end with the new system. Information from POSIDS would be added electronically (no manual interface) to the electronic file containing the patient's medical records. These records would typically be stored on hard drives in medical offices, hospitals, etc. A patient could obtain such records on a 3.5 x 3.5-inch 1.44 MB diskette (cost: 20 cents). When a patient went to a hospital or specialist, they would avoid cumbersome, paper-oriented registration processes and simply hand their medical records diskette to the receptionist. Alternatively the receptionist could download the file from a computer in the medical office of the patient's primary care physician.
Today, when someone needs to check on, say, recent test trends he/she must currently request the manila medical records file folder. Then they must shuffle through it and form crude visual images of what the sought-after data are saying. Or they would find a chart laboriously prepared by some nurse (perhaps twice). All that would end with the new cheaper, faster system. A list of recent test results, diagnoses, etc. would take a few keystrokes or mouse clicks to pull up. To better visualize some data, a few more keystrokes would produce a software-prepared plot on a monitor screen. Small plot-data files, rather than huge plot files, could be transferred electronically to specialists for consultation. This would be possible because everyone would have the same plotting package (part of the overall software package), reducing the complexity and costs of transmitting graphical information by a factor of 100-1000. Nurses would be freed of the large amounts of labor required to "chart" patient data on paper (often twice) -paper that doctors now must fumble for through thick paper files of patients' medical records. The cost savings resulting from having patients' medical records in electronic form with a well-defined format would clearly be huge. The staggering costs of manually converting patients' medical records to electronic form would be largely eliminated (except for data obtained before computerization of the health care system's information flow-, analysis- and storage systems). The quality of medical care would increase by virtue of medical records being accessible far more quickly and in far more visually informative form.
Go to Table of Contents ~ Go to Reference List ~
Go (Return) to Home Page of this Web Site
MONOPOLY PROBLEMS The approach recommended here for dealing with monopoly problems is virtually unknown in the software industry, and hesitance might be expected from that quarter. However the software industry is likely to see the potential for broadening the scope of the industry into whole new areas where it is presently incapable of meeting the software needs of numerous large would-be customers. The health care system is but one example. (See below)
Rules for Rulers:
Some sort of "Computerization Committee" of the health care system (perhaps a new department within Medicare) would have to be created to provide leadership. The basic characteristics needed by such a committee have been described in some detail elsewhere (Kendall and Levine. 1998).
That committee would acquire the services of a single software company to generate the health-care system's complete information flow-, analysis- and storage software. Teams composed of software engineers and health care professionals would do the work since very few people possess both sets of skills. The following constraints would be needed to avoid a monopoly:
Upgrading Software:
The advantage of an open-source-code approach is that no one would get a monopoly, or the ability to hold the health care system up for ransom, or to be unresponsive. Health care professionals wanting an improvement to the software or a validation check for a suspicious output would simply request it of the "Computerization Committee". That committee would then put such requests out to the world of software companies, individual programmers, etc. for bids. Alternatively, these software people could submit the requested source-code patches ("updates") and related documentation updates to the Computerization Committee. That committee would judge the various submissions and send the winner a financial reward. Rewards too stingy would produce few future responses. Excessively generous financial rewards would likely produce unmanageably excessive responses. The original software company would be able to participate in that overall software development process. Because of its better understanding of the inner workings of the source codes, the original software company would win many of these development rewards. This advantage would translate into lower initial bids for the contract to do the health care system's information flow-, analysis- and storage software. The system would be quite "free-market" and have a high degree of economic efficiency. An efficient, responsive upgrade system would allow for future developments and improvements to the source code and its documentation indefinitely, without any monopoly problems.
Go to Table of Contents ~
Go to Reference List ~
Go (Return) to Home Page of this Web Site COSTS OF LARGE-SCALE COMPUTERIZATION Initial Software: The cost of the initial software to create proxies for the existing information flow-, analysis- and storage systems, and writing documentation, would probably be on the order of several billion dollars, based on numbers one hears of for costs of huge, complex software systems. Annual Software Management, Maintenance, Updates and Development Costs: These costs would probably be something on the order of some tens of millions of dollars annually. Workers would be under the management of the "Computerization Committee" (probably reorganized after the initial software development and documentation is complete). Most or all of the actual software work would be done by outside individuals and companies that would either work under contract on specific projects or submit software and documentation in hopes of winning financial rewards. These entities would be teams composed of software writers and people knowledgeable in the inner workings of various components of the health care system. Additional Computers and Memory: Like most US industries, health care is over-computered but under-computerized, so no additional investment in computers would be needed. No huge amounts of memory are required since there are no graphics involved - only words and numbers. Hard drives are readily available, and cheap, in the 50-100 gigabyte range. A single 1.44 MB diskette (costing about 20 cents) would be adequate for storing the medical records of almost any patient. The cost of hard drive storage is much less than 20 cents per 1.44 MB. POSIDs: The cost of a POSID would be on the order of $100, mass-produced, given typical costs of simple, hand-held computer devices (which is what a POSID would be). All point-of-direct-service health care providers (anyone who comes into direct physical contact with patients, or into direct conversations with patients on matters directly related to the practice of medicine) would need such devices. Roughly a third of all health care system personnel located in medical facilities would probably fall into this category. For a 10-year lifetime of a POSID, this sort of cost ($10 per health-care worker per year) suggests an annual system-wide POSID cost similar to the above-mentioned annual software management, maintenance, updating and developing costs.
Documentation:
Users' manuals would probably be required by roughly half of all health-care system personnel. The per-copy cost would be on the order of $20 up front and $2/ year for annual updates thereafter. The cost of programmers' manuals and validation documents would be negligible relative to the cost of users' manuals due to the far lower number of people who would need them. This documentation cost would be a negligible fraction of the annual cost of the average health-care system employee.
Go to Table of Contents ~
Go to Reference List ~
Go (Return) to Home Page of this Web Site BENEFITS OF LARGE-SCALE COMPUTERIZATION Direct Financial Benefits
Thus the total costs of health care should probably, for purposes of this analysis, be broken down into direct labor costs, direct capital costs, and indirect labor/ capital/ natural resource costs. This final category is roughly estimated as 20% of the total mix, suggesting that this total mix should be apportioned as 48% direct labor costs, 32% direct capital costs, and 20% indirect labor/ capital/ natural resource costs. Direct labor costs will be the focus of this analysis, since the final category appears to have little to do with the issues at hand.
Given the non-pharmaceutical portion of the health-care system's contribution to the US GDP of $1445 billion per year, the 48% direct labor component of this would be $694 billion per year. Doctors and nurses are said to spend about 40% or their time on paperwork activities. Assume about 30% of their time is spent on computerizable activities, and assume that this figure applies to all point-of-direct-service personnel are characterized by this figure. Such personnel compose an estimated 1/3 of health-care system direct labor. The remaining 2/3 of this labor also spends an estimated average of 30% of their time on computerizable activities (ranging from 100% for medical records people to perhaps 80% for people like book-keepers, billing departments, accountants and administrators to 60% for phlebotomists, 50% for secretarial people, receptionists, store-room people etc. to perhaps 10% for food-service personnel). This suggests a benefit of large-scale computerization of $208 billion per year (30% of $694 billion). Personnel reductions also imply reductions in the size of capital facilities. These savings are neglected here to be conservative. According to Government estimates, if the U.S. health-care community could create a secure network that would make health records and prescription orders available online nationwide, the US could save $140 billion/ year (J. Harvey. 2004). Some fraction of this amount overlaps the $208 billion per year estimate above, so it is neglected to be conservative. Savings in Compliance Checking: In addition, HMOs and Medicare and Medicaid have staffs of people checking invoices for compliance to regulations. The number of people performing such activities is estimated to be in excess of the number of health care system people providing point-of-direct-service to patients. Something on the order of 80% of this work is probably computerizable, given the far greater fraction of health care system data in electronic form, in standardized formats and more easily available. This would suggest an additional benefit of large-scale computerization of about $185 billion per year. Reductions in Billing Fraud and Abuse:
Large-scale computerization would permit a larger fraction of invoices received by HMOs, Medicare, Medicaid and SCHIPs to be checked for fraud and abuse (e.g. "upgrading"). That fraction today is only a few percent; it could be 100% and be far more automated, cheaper, sophisticated and thorough. This would be a result of much more of the health care system's stored information being readily accessible in electronic form and with standard formats. The risks of attempting fraud and other system abuses would increase dramatically, greatly reducing the number of attempts to commit it. This fraud and abuse costs medical insurers and taxpayers an estimated $150 billion per year (extrapolated from $100 billion about a decade ago). If only half of this fraud and abuse could be eliminated, the savings would be $75 billion per year. Reconciliation: The sum of the above three figures gives an estimated savings from large-scale computerization of roughly $470 billion per year. Clearly, just the above-mentioned direct financial benefits of computerization exceed the costs of computerization by several orders of magnitude. Spin-Off Benefits -- Foreign Health Care Systems
Spin-Off Benefits from Addressing Other Major Inefficiencies
Capital-Facility Utilization Efficiency:
Capital costs comprise something on the order of 40% of health care costs. The health care system is being changed by large-scale, new technologies for diagnosing and treating health problems. Cities and hospitals of all sizes tend to want one of each - regardless of their cost-effectiveness. Economic concerns get overruled by egos, politics and pressures from health-care professionals with financial investments in the technology. The less-than-dominant role of objectivity is due partly to the dulled enthusiasm for health care consumers to care about prices and efficiency, given all those subsidies, interposed insurers, and growing hordes of HMO compliance checkers alluded to above. As the health care system and the US economy grow more capital intensive, it becomes more important to use capital facilities more efficiently. But this is not occurring. Large-scale capital equipment sits idle on evenings, nights and weekends and even during normal working hours, so total capital costs must be heaped on prime-time users. Just a 10% increase in capital-facility utilization efficiency translates into savings of about $60 billion per year. Utilization efficiencies of large-scale new technologies could easily be increased by 100%. Further efficiencies could be achieved by procedural improvements.
Prescription Drug Pricing:
Prescription drug prices are well in excess of those that would exist in a free market system as a result of subsidies and the interposition of insurers between buyers and suppliers. An example illuminates. Blue Cross - Blue Shield, in the past, did not cover treatment of in-grow nails. The price charged then by doctors was about $7. This was a nearly perfect free-market, non-subsidized price reached by arm's length negotiations between willing buyers and willing sellers, both of the same, small, economic size (single individuals) with little likelihood of monopolistic pricing. Blue-Cross-Blue Shield later decided to insure such costs. Within a few years prices for this procedure had risen by a factor of about ten. This is not to say that prescription drug prices are a factor of ten above their free-market values, but it is suggestive of the strong effect of imposing insurers between buyers and suppliers in what would otherwise be a free market. Proxies could be designed to cause prescription drug pricing to take on more of a free market behavior and thereby, in the Adam Smith sense, make the system more efficient. The automobile collision insurance business faces the same problem for the same reason. One of its free-market proxies is its requirement that car owners obtain two bids on repair work. This proxy wouldn't work in prescription drug pricing, but other proxies might.
End-of-Life Health Care:
Expenditures on end-of-life health care are also far beyond what would be spent under free market conditions, and for the same reason as why prescription drug prices are far beyond free market levels. Proxies could be found to give these expenditures more of a free market character also. Financial benefits could be huge, since nearly two-thirds of all health-care dollars spent on each of us during our lifetimes is spent during the last six months of our lives (mid-1990s data). These expenditures are often on 11th hour heroics unsupported by objective medical analyses, often of questionable value, and frequently driven by irrational judgments on the part of the patient and/or the patients' kin, often supported by less-than-objective analyses on the part of doctors, surgeons and others. If just 25% of end-of-life (last 6 months) health care costs could be eliminated (presumably the least justifiable of these costs), total US health care costs could be reduced by about a sixth, i.e. by about $240 billion per year (in 2003), although with some overlap with the direct benefits noted above.
These three spin-off efficiency improvements could further reduce health care costs by several hundred billion dollars per year (Sundquist. 2004).
Health Care Quality Benefits
Error Reductions:
Studies have pointed out the huge costs of human errors in health care. Under large-scale computerization, processes prone to human error will be done increasingly by software. A software error, once detected, can be corrected, and this usually means that the error never happens again, unlike human errors that continue forever.
Workplace Environment Improvements:
Shortages of health care professionals are causing added stresses, increases in the workweek and reductions in health care quality. Large-scale computerization would reduce the need for direct health care labor by roughly 30%. The forecasted shortfall of 200,000 doctors and 800,000 nurses forecasted for 2020 could be eliminated by large-scale computerization of the health care system. (See Appendix D.)
It could also make careers of health care professionals more productive and more meaningful as they spend more of their time on the sorts of activities that motivated them to enter the field in the first place. Computerized Order Entry: Using electronic prescribing systems that avoid hand-written errors have been found to produce major benefits where it has been tried. One hospital reduced error rates by 55% over eight months. Rates of serious medication errors fell 88% over a four-year period in a subsequent study [Landro, 2003]. The National Academies' Institute of Medicine reported that more than 7000 deaths and as many as 7% of hospital admissions occur as a result of adverse drug effects and medication errors. The Institute said that up to 95% of these events could be avoided by using computerized physician-order-entry systems for prescriptions [Rundle, 2004]. Despite numerous studies supporting the value of information technology in reducing medical errors, only about 5% of doctors and about 19% of health-care organizations are using fully operational systems [Rundle, 2004]. Computerization of the health care system's information flow-, analysis- and storage systems would make physician-order-entry procedures for prescriptions far easier and cheaper to implement. Public Health Studies: With so much medical record information in electronic form and in a standard format, doing public health studies would change from slow, tedious, expensive processes to quick, cheap processes. A group doing a public health study would arrange to have its computer automatically dial up thousands of computers in medical facilities and interrogate millions of patient medical records - all without lifting a finger. The result: more public health studies, better public health studies and consequently, improvements in health care nationwide if not worldwide. Benefits of "Off-Proxy" Source Code Development Avoiding Unnecessary Care: As much as half of the care provided to Americans (including 17-32% of surgeries performed on Medicare patients) is unnecessary, including procedures that don't do any good, tests that are repeated, and drugs for which there is no evidence of benefit, according to studies cited by a 2001 report of the National Institute of Medicine, a government advisory group [Landro, 2003]. With the health care system burying essentially all the data needed to detect and eliminate these problems in paper records and randomly formatted files on isolated hard drives, these problems remain impossible to address in any cost-effective manner. Therefore they are certain to continue unabated. But with a far larger portion of health care system data in readily accessible, electronic form in files of standardized formats, low cost automated procedures could be developed to reduce or eliminate such problems. Avoiding Overuse of Antibiotics:
Antibiotics account for 15-20% of the average hospital's drug budget. Overuse has caused resistance to many antibiotics, leading to more medical complications and costs. Computerization of information-flow/ analysis/ storage makes such problems far easier to detect and prevent. One VHA study indicated that US hospitals could save more than $1 billion/ year in this way [Landro, 2003]. Disease Management: Many experts agree that the best opportunity to improve health care and reduce expensive complications is "disease management" - monitoring people with chronic conditions such as diabetes, congestive heart failure and coronary artery disease. These diseases are expected to cost $510 billion in 2003 and about double by 2020 [Landro, 2003]. An electronic record-keeping system in South Carolina for Medicaid patients found that hospitalization costs could be reduced by about two thirds for diabetes patients [Landro, 2003]. Preventive care and close monitoring for patients with congestive heart failure have been found by numerous studies to reduce hospital costs by more than 30% and cut re-hospitalizations by 50% [Landro, 2003]. Large-scale computerization of information flow, analysis and storage would make such disease management procedures easier, less expensive and probably of higher quality.
Opportunities for Impromptu Software Development:
A health care professional might encounter some repetitive paperwork procedure not covered by the existing software. They could then write (and test) a macro for doing the procedure automatically. The macro could then be submitted to a web site established by the Computerization Committee for information sharing among health-care providers. That would enable health-care providers worldwide to download the macro and use it to save millions of person-hours worldwide rather than a few hundred person-hours in the local medical facility where the macro authors work. In future updates of the main source code, the macro could be incorporated into that software, resulting in a financial reward to the authors. Efficiency enhancements of this nature are virtually impossible in the present environment.
Go to Table of Contents ~
Go to Reference List ~
Go (Return) to Home Page of this Web Site COLLATERAL BENEFITS OF LARGE-SCALE COMPUTERIZATION Collateral Benefits --
Plotting Software One major beneficiary would be education systems. Once mathematics is taught well, it will become a process of converting physical problems to mathematical form - functions -- that can then be manipulated and solved for some desired result by various tools (e.g. algebra, trigonometry, geometry, calculus, vector analysis, partial differential equations etc.). The ability of students to examine and manipulate mathematical functions using plotting software enables them to develop a far more intuitively understandable comprehension of a physical / mathematical problem and its solution. This makes science, math etc. more understandable, fun and useful. Students of math, science, engineering, architecture, medicine etc. would be able to use obsolete computers containing the open source plotting software to do all this at extremely low cost -- globally. When students take homework home, or enter the workplace, the identical plotting package would await them.
Collateral Benefits -- Applicability to Other Systems
Benefits Summary - A Cure? - The Only Cure? - The Permanent Cure?
The Only Cure?
Consider:
It would seem, therefore, that a compelling case could be made for concluding that large-scale computerization of the US health care system, using the strategy laid out above, is the only cure for the health care crisis.
The Permanent Cure?
Double-digit growth rates in health-care expenditures suggest that any "cure" for the health care crisis may lack permanence. The benefits of any efficiencies could be negated by future cost increases, turning any cure into just a temporary respite. To make a judgment on this issue, one must understand the nature and cause of these cost increases. Consider the following five effects.
(1) Subsidies for health care continue to grow, meaning more money in the pot for health care providers and prescription drug companies to consume via price increases. Subsidies also dull consumer concern over price growth. All this is part of why subsidies are so inefficient and why health care cost-growth is in the double digits. A compelling case has been made, however, for the contention that subsidy growth is rapidly grinding to a halt [Newhouse, 2004]. Thus the portion of the growth of health care expenditures that results from subsidy growth is likely to cease also.
(2) The US population is graying. The elderly require more health care per capita than younger people. This trend is temporary however, reflecting the slowing of the US population growth rate over the past few decades. Population "momentum" effects are not likely to last much longer. Fewer children mean more wealth available for financing health care, and less need for capital for expanding the health care system. Life spans are increasing, meaning that there are more people around to finance health care during the last six months of life (which consumes 2/3 of health care expenditures). Life-span growth and declining population growth rates thus counteract the temporary effects of population graying. This of course assumes that the length of the period of ill health at the end of life does not increase with life span. This has been found to be true [Mathers et al, 2001]. Health-adjusted life expectancy (each life-year weighted by a measure of health status) increases approximately one for one with life expectancy across countries. This implies that the 6-month length period of ill health at the end of life appears to be fairly constant, and as life expectancies increase, health status improves proportionately [Mathers et al, 2001].
(3) Two thirds of health-care expenditures occur in the last six months of life. As noted earlier, decisions on end-of-life health care expenditures have large components of irrationality on the parts of patients and their kin, and cost insensitivity borne of subsidies and insurance. Self-dealing by health care professionals probably also plays a role. The combination of diminishing subsidies and large-scale computerization would generate increasing public interest in tackling such inefficiencies.
(4) The health care system continually finds more creative ways to "game the subsidy system" and thereby extract more money from HMOs, Medicare, Medicaid and SCHIPs. Large-scale computerization would make it increasingly difficult for the health care system to "game the system," i.e. commit fraud and abuse. Thus the adverse effects of subsidies on health care expenditures generally would decrease as large-scale computerization effectively bestows more power on insurers and HMOs. Health care professionals have obvious financial motives for giving less weighting to cost-effectiveness in medical decisions than objectivity would recommend. This shift of power would give cost-effectiveness a more rational (objective) level of influence on treatment decisions. Granted, HMOs and insurers are not totally objective in regards to the issue of cost effectiveness in that increasing a patient's risk of death causes HMOs and insurers no expense, and not all costs of illnesses are borne by HMOs and insurers. But the shift of power would probably cause any over-weighing of cost-effectiveness by insurers and any under-weighing of cost-effectiveness by health care professionals to come closer to balancing out.
(5) Science and technology are continually developing more medical treatments for diseases. In a sense this increases the value of health care services, and, logically, would be expected to increase expenditures thereon. But on the other hand, a healthier population with increasing life spans creates more people to finance end-of-life health care and reduces the indirect costs of ill health. Also, many new technologies provide cheaper alternatives to current treatments of diseases. So, all in all, it is far from clear that the effects of science and technology are anything but a wash insofar as their effects on trends in per-capita health care expenditures are concerned.
The above five effects probably account for at least the bulk of the double-digit rate of growth of health care expenditures. Effects (1) (subsidy growth) and (2) (population graying) are likely to be short-lived. Effect (3) (inefficiencies in end-of-life care) would diminish from the combined effects of reduced subsidies and increasing public sensitivity to efficiency issues. Effect (4) (increasing fraud and abuse) is likely to diminish as large-scale computerization sifts medical decision-making powers closer to where the issue of cost-effectiveness gets the consideration it would receive in a totally objective decision-making environment. Effect (5) (science/ technology), may not be an effect at all. All this would suggest that the cure for the health-care crisis by large-scale computerization could be permanent.
Go to Table of Contents ~
Go to Reference List ~
Go (Return) to Home Page of this Web Site
CONCLUSIONS
REFERENCE LIST
APPENDIX A -- SUGGESTED IMPLEMENTATION PLAN Funding: Funding for the project could best be provided by the federal government, since the initial cost of the software and documentation (perhaps a few billion dollars) would be recouped by such agencies as Medicare and Medicaid within a few months of operation under the new health-care system.
Computerization Committee Makeup:
A wide variety of health-care system administrators and professionals (including "point-of-service" types - people who come into direct physical contact with patients or who speak directly with patients on matters directly related to the practice of medicine), plus some experts and administrators experienced in software engineering should make up the bulk of the Committee. Someone in the hand-held computer system would be desirable to deal with the development of POSIDs. Representation by patient groups would make sure the health care system people do not engage in system self-dealing at public expense, and to make sure the software people don't allow anyone to wind up with any sort of monopoly on any sort of software or hardware that would require long-term payments of royalties. Since the system would probably be used in foreign countries, foreign health care representation might also be desirable to make sure that all input/ output text is isolated (rather than sprinkled throughout the entire code) to facilitate translation into foreign languages, and to make sure that the sets of option parameters have sufficient breadth to cover the medical practices and policies of health care systems worldwide.
The basic characteristics needed by such a committee have been described in detail elsewhere (Kendall and Levine. 1998). Strategy: The strategy of the Computerization Committee ought to be to develop the basic "backbone" of the system fairly quickly, e.g. within a year or so. The "backbone" is defined here as:
Once the system's "backbone" is complete, everything would be in place to obtain large cost savings and health-care improvements. It would also get health-care system people acquainted with, and accustomed to, the basic processes of the new system. The remaining, not-yet-developed parts of the system would consist of numerous relatively independent components. These could be developed relatively independently of each other over a period of perhaps three additional years. This relative independence should make the remaining computerization straightforward.
Go to Table of Contents ~
Go to Reference List ~
Go (Return) to Home Page of this Web Site APPENDIX B ~ PRIVACY OF MEDICAL RECORDS APPENDIX C ~ VERSATILITY ISSUES IN LARGE-SCALE COMPUTERIZATION APPENDIX D ~ PERSONNEL ADJUSTMENTS
According to a study in the New England Journal of Medicine, about 126,000 nursing jobs in US hospitals are currently unfilled. It was also estimated that, by 2020, the US would have a shortfall of 200,000 doctors and 800,000 nurses (ABC News Online, 2005). The number of registered nurses in 2000 was 2,697,000 (Table 148 of Statistical Abstracts of the U.S. - 2004-2005). If 30% (809,000) of these positions could be eliminated through large scale computerization of the U.S. health care industry, that would eliminate the projected shortage of nurses in 2020. The number of professionally active doctors of medicine in 2002 was 768,500 (Table 149 of Statistical Abstracts of the U.S. - 2004-2005). If 30% (230,550) of these positions could be eliminated through large-scale computerization, that would eliminate the shortfall of doctors projected for 2020.
Go to Table of Contents ~
Go to Reference List ~
Go (Return) to Home Page of this Web Site
In 1965, nationwide health expenditures were 6% of the US Gross Domestic Product (GDP). By 1990 they had doubled to 12% and were forecasted to reach 18% by 2000. Fortunately a respite during 1992-1999 kept 2002 expenditures at 15% of GDP [Pear, 2004] -- but it was still a $1.55 trillion "industry" in the US in 2002 [McGinley, 2004]. The Center for Medicare and Medicaid Services stated (2/11/04) that total US health care spending in 2003 was $1.7 trillion - more than $5800 for every American, and more than 15% of the US GDP. By 2013 spending on healthcare is expected to reach $3.4 trillion/ year and exceed 18% of the GDP. After 2013, baby boomers start reaching retirement age. Far more rapid increases in health care spending can be expected then, according to Dan Crippen, former director of the Congressional Budget Office. Even today hospitals are going after unpaid bills with tough tactics such as placing liens on homes and assessing interest, fines and legal fees. One result: hospital- and medical bills are now the second leading cause of personal bankruptcy after unemployment [Unknown, 2004]. It seems unlikely that hospital- and medical bills will remain only the second leading cause of personal bankruptcy for much longer.
This paper was developed with the eventual shift in public focus from subsidies to efficiencies in mind. It is part of a larger paper [Sundquist, 2004] on four major health-care system inefficiencies: (1) lack of computerization, (2) Inefficient use of capital and capital facilities, (3) failure to keep prescription drug prices down to their free-market values, and (4) questionable end-of-life care. That paper also examines the economics of interposing government subsidies and health insurance between buyers and sellers of health care. With health-care costs rising at double-digit annual rates, little value is seen in seeking out nickel-and-dime cost-containment strategies. These only postpone the inevitable need to attack the major inefficiencies that prevent meaningful, permanent results.
Large-scale computerization of health care's information flow-, analysis- and storage processes has never been attempted. This is in spite of the fact that such computerization could produce direct savings of on the order of hundreds of billion dollars per year. Indirect savings in the health care system and in other parts of the US- and foreign economies could be similar. (See the analysis below.) In addition, the stage would be set for an endless string of health care quality improvements that would be far more difficult, costly or impossible today (See below). The analysis below explains why such a process has never been attempted, despite the potential benefits. This explanation tells us what must be done to achieve this end. Only a conceptual level treatment of large-scale computerization is presented in this document. Computerization on this scale would involve significant changes in the way virtually all health care professionals carry on their daily routines. Questions involving how to deal with the adjustments, turmoil, shifts in decision-making powers and other formidable human problems are left to those more familiar with the health care system.
The more recent studies of health-care system inefficiencies (noted in the mass media in mid-2003) pointed out numerous inefficiencies that all seem to have largely resulted from the health care system's failure to utilize the full potential of computers. Also, hand wringing has been on-going for decades concerning the huge amounts of paper work being done by health-care professionals. Some estimates indicate that 40% of the average nurse's time is consumed by paperwork, and time allocations by doctors are similar. These massive amounts of wasted time, money, talent and training are just a few of the effects of the failure to make full use of the potential of computers to lower costs, and increase the quality, of health care.
The U.S. health care system probably has the most complex, elaborate, labor-intensive, paper-intensive information flow-, analysis- and storage systems of any industry or public institution. Computers are everywhere, but they aren't used for much, relative to their potential because:
For an information flow system to work at all, there cannot be any need for information to cross interfaces between incompatible software systems, as noted above. But non-electronic interfaces must also be eliminated, or at least minimized, because they are so slow, labor-intensive, expensive and error-prone. So for optimal efficiency, the information flow system must have three key features:
Problems with moving information across interfaces explain why the health-care system has never computerized its information flow-, analysis- and storage systems beyond the mundane, with little in terms of information flow beyond local environments - even though the system possesses more than enough computer power to accomplish this (neglecting the POSIDs). To solve the interface problem, software incompatibilities must be eliminated. This requires a single source for all software in the health care system. But this creates monopoly problems as noted above.
This author, like everyone else, lacks the data to accurately estimate costs. But the decisions needed at this point do not require a high degree of accuracy because:
Benefits of computerization tend to be counted purely as direct financial benefits of increased labor utilization efficiency. But for large-scale computerization, numerous other categories of benefits must also be considered because they are very large. The analysis below attempts to cover all significant categories of benefits.
US health care expenditures were about $1700 billion/ year in 2003, but pharmaceuticals were about 10% of this figure (9.4% in 2000 but increasing faster than anything else). Presumably dental care, optometry and chiropractic care are included also. These might total 5% of these expenditures Since the pharmaceuticals, dental care, optometry and chiropractic care have no part in the present computerization issue, the question becomes: What fraction of the remaining $1445 billion/ year in health care costs can be saved through large-scale computerization in terms of increased labor utilization efficiency? Expenditures on health care excluding pharmaceuticals etc. in the US go to pay labor costs, capital costs and natural resource costs with the latter being generally negligible relative to the other two. Labor costs and capital costs tend to be proportioned about 60:40 for US industry as a whole, and the
health care system's labor component is probably close to this figure. However the labor- and capital costs associated with supply production, heating, utilities, and maintenance and depreciation on capital facilities need to be broken out, since these costs probably have little to do with inefficiencies in the health care system. They have no computerizable content, at least in the present context, so they should be considered as fixed.
Because the products of the single-source software developer would be source codes and manuals available to anyone (in order to avoid a monopoly by opening future code development to anyone), foreign countries could obtain the source code free. They could simply change the language in inputs and outputs, change parameters to reflect differences in health care policies and practices, and translate the manuals. US citizens would benefit indirectly, but probably enough to pay for a significant fraction of the software development costs. Foreign nations would further develop the software, and these updates would be freely available in the US. Joint software development projects could be undertaken. Citizens traveling outside their own country could have their medical records accessible by foreign doctors, and they could have these records updated to reflect medical treatments abroad.
The huge benefits of large-scale computerization could awaken the US public from its long-term neglect of opportunities for efficiency improvements in the health care system. If so, then other efficiency-improvement projects could be undertaken. In addition to labor utilization inefficiencies examined above, three other major inefficiencies in the health care system offer opportunities for substantial cost savings. These involve capital facility utilization, prescription drug pricing and end-of life care [Sundquist, 2004].
Many articles in the mass media over the past year have pointed out the potential for computerization in the health care system to improve the quality of health care. Senator Hillary Rodham Clinton submitted a bill in early 2004 titled "The Health Information for Quality Improvement Act". It proposed numerous health care quality improvements, to be achieved by computerization at a scale much smaller than that being proposed here. Below are a few other examples of possible improvements. These would be much easier to implement in a health care system that has undergone large-scale computerization that puts much more health care data in electronic form, in a standard formats and readily accessible by qualified individuals.
The initial version of the information flow-, analysis- and storage software would presumably consist of just the electronic proxies for each of the health care system's existing channels and means of information flow, analysis, and storage. This would be sufficient to achieve huge cost reductions and major improvements in the quality of health care. However this should be seen as just the beginning of cost reductions and quality-of-care improvements. With so much more of the system's information (data) in electronic form, in standard formats, and so much more accessible, all sorts of further source-code developments would facilitate programs not economically feasible under present-day conditions. It seems likely that these "off-proxy" developments would produce benefits that, over time, would be comparable to the initial proxy-based benefits. Below are brief descriptions of a few possible off-proxy developments.
Collateral benefits (those outside the health-care system) of large-scale computerization would be likely. Two examples are described below.
One component of the overall health care software system would be a plotting package capable of generating plots, graphs and charts. Sharing data in such formats becomes far easier, faster and cheaper when everyone has the same plotting software, as could be the case if the health-care system underwent large-scale computerization. Plotting software is available today, but only in compiled form. However the Computerization Committee would require source-code plotting software to avoid monopoly-related risks. Plotting software is useful in virtually every industry, business and public institution. Once the health care system's plotting software enters the public domain as a source code, such software becomes free to everyone.
The health care system is unique in terms of the size and complexity of its information flow-, analysis- and storage systems and in the complexity of its structure. This should not suggest, however, that other industries could not benefit from large-scale computerization using the same basic strategy as that outlined above. The health care system's experience in large-scale computerization, and its software, could guide other industries interested in cost reductions, efficiency improvements and enhanced output quality.
Some systems that come to mind include the legal-law-enforcement-judicial system, the library system, the electric power grid, the military, and many large industries (especially when suppliers and wholesale customers are included). This author's experiences have led him to conclude that the bulk (on the order of 60%) of what is now being done by humans in the "paper" portion of US business and industry could be done by large-scale computerization. Presently information flow between, and within, manufacturing, sales, marketing, customers, suppliers, personnel, administration, research, development, accounting, legal etc. in any given industry takes place via a hodge-podge of paperwork, phone calls, faxes, e-mails, group meetings, face-to-face encounters (often involving travel over large distances) and electronic (computer-based) means. The result of all this is interface proliferation with its associated expense, slowness and proneness to error - the same problems as those facing the health care system today.
A Cure? The roughness of these benefit analyses is obvious, but it would be impossible for these estimates to be so off base as to challenge the economic viability of large-scale computerization. If US taxpayers were to pay for large-scale computerization in the health care system, the average taxpayer would probably recoup his/her investment after their first visit to a doctor, or after the first hour or so of their first hospital stay. The sum of the benefits listed above appears to be a sufficiently large fraction of health care expenditures to justify declaring large-scale computerization as at least one cure for the health care crisis. The magnitude and scope of the benefits also seem to be sufficient to overcome any political resistance to the concepts and methodologies proposed.
Computerization of the information flow-, analysis-, and storage operations in the total US health care system by a single software source is essential for achieving large-scale efficiency improvements. Direct benefits, in terms of labor utilization efficiencies, could be on the order of $470 billion per year. Numerous improvements in quality of health care could also be achieved. Other benefit categories, in total, could produce comparable benefits. In total, these benefits could largely cure the health care crisis. Such large-scale computerization offers the only cure, and probably the only permanent cure, for the health care crisis.
ABC Online. 2005. West draining doctors from poor nations, ABC News Online, 10/27/05 http://www.abc.net.au/news/newsitems/200510/s1492420.html
James Harvey. 2004. Personal communication, 9/28/04 (JimPgh@aol.com)
John T. Kelly, 1998, "After The Chaos': Expected Benefits of Health Information Management", Health Affairs, 17(6) pp. 39-40.
David B. Kendall and S. Robert Levine, 1998, "Pursuing the Promise of An Information-Age Health Care System", Health Affairs, 17(6) pp. 41-43.
J. D. Kleinke, 1998, "Release 0.0: clinical information technology in the real world", Health Affairs 17(6) pp. 23-38.
Laura Landro. 2003. "Six Prescriptions to Ease Rationing in US Health Care", Wall Street Journal 12/22/03.
Steve Lohr, 2005, "Health Industry Under Pressure to Computerize", New York Times, 2/19/05.
C. D. Mathers, R. Sadana, J. A. Salomon et al, 2001, "Health life expectancy in 191 countries," The Lancet, 357(9269) pp. 1685-1691. (See David E. Bloom, David Canning, "Global Demographic Change: Dimensions and Economic Significance", Harvard Initiative for Global Health, Working Paper Series, Working Paper No. 1, April 2005, http://www.globalhealth.harvard.edu/WorkingPapers.aspx.).
Clement J. McDonald, 1998, "Need for Standards in Health Information", Health Affairs, 17(6) pp.44-46.
Laurie McGinley. 2004. "State and Local Programs Seek to Aid Uninsured", Wall Street Journal, 1/9/04 (reporting on a Health/ Human Services report in the journal Health Affairs on 1/8/04).
D. W. Moran, 1998, "Health information policy: on preparing for the next war", Health Affairs, 17(6) pp. 9-22.
Joseph P. Newhouse. 2004. "Financing Medicare in the Next Administration", New England Journal of Medicine, Oct. 21, 2004.
Robert Pear. 2004. "Health Spending takes 15% of GDP", New York Times, 1/9/04 (reporting on a Department of Health and Human Services report published in the journal Health Affairs on 1/8/04).
Rhonda L. Rundle. 2004. "WellPoint to Pay $30 Million for Doctors' Computers", Wall Street Journal, 1/15/04.
Bruce Sundquist. 2004. "Inefficiencies in the U.S. Health Care System -Identifying and fixing Them", Ed. 3, August 2004
http://home.windstream.net/bsundquist1/hci.html
Unknown. 2004. "Lawsuit Challenges Charity Hospitals on Care for Uninsured", Wall Street Journal, 6/17/04 p. B1.
Go to Table of Contents ~
Go (Return) to Home Page of this Web Site
Step 1:
Obviously Step 1 would be to form a "Computerization Committee" with the funding and power to obtain and administer the necessary agreement with a software vendor willing to produce non-copyrighted source code and related documentation at an agreed-upon price. The Computerization Committee should be structured so as to be able to resist any attempt at self-dealing by either the health care system or the software industry. Also it should have a strong commitment to achieving outstanding results in terms of both cost reductions and improvements in quality of care. A separate department within Medicare would seem to meet these qualifications. Whether the Computerization Committee should be charged with providing the necessary health-care system consultants for use by software engineers or whether this should be the task of the software vendor should be decided through negotiations between the Computerization Committee and the software vendor. The same sort of negotiations should resolve the question of who is to handle the training of health-care system people in the use of the software and POSIDs with the aid of the user's manuals. The Computerization Committee would also handle the maintenance, updating, management, development and distribution of the software system and its documentation on a perpetual basis. This in-perpetuity activity would be expected to require a fairly small annual cost relative to the initial investment in source code software and documentation. The Computerization Committee would also need to establish the specs for the POSIDs by working with the software vendor. (The software vendor might not be sympathetic with the need to keep the POSIDs lightweight and compact, and therefore assign an inordinate share of the software to reside in the
POSIDs.)
Much effort and expense have been expended in finding ways of keeping medical records private. HIPAA and other patient-privacy initiatives are terribly complex and require enormous planning and analysis to ensure that all compliance and governance mandates are met. So the question is bound to arise as to whether computerization of medical records will make patient privacy harder or easier to maintain. With a far better-organized electronic system for creating, storing and retrieving medical records, it would seem that creating procedures for protecting privacy could be easier than for the more haphazard current manual means of record creation, storage and retrieval. Of course there are hackers, but this is a problem faced by every industry, business and government entity. So whatever ways are developed to keep hackers at bay there ought to work in the health care system as well. Public health surveys may require computer access to millions of electronic personal medical records, but such surveys could be allowed only under tightly controlled procedures. Manually done public health surveys would seem to be far more vulnerable to breaches of privacy, besides being far more expensive and error-prone.
Having the entire health-care system doing all of its information flow-, analysis- and storage operations using the same software conjures up illusions of some "Big Brother" Computerization Committee deciding on medical practices, procedures and policies with no room for the knowledge and insights acquired by individual health-care providers over a lifetime. There is no reason why this should be tolerated. When individual health-care providers or medical offices or hospitals or other controllers of one or more computers download the initial software and subsequent updates, their first task would be to run through a large set of options and to specify parameter values that reflect that entity's practices and policies. Various medical institutions, e.g. the AMA, might recommend sets of parameters that might serve as default values. State or national laws might establish the value of some parameters, or lower- or upper limits on such parameters. The set of options ought to be broad enough that foreign health care institutions could fit the software to their practices.
This analysis suggests that the health care system could reduce its direct labor content by about 30% by computerizing its information flow-, analysis- and storage systems. The systems would be installed over a three-year period following the first year's development of its "backbone" that would produce significant savings starting at the end of the first year. But for about a year after the 3-year installation, the adjustments to the new system would require more than the normal amount of labor.