IEEE TRANSACTIONS ON INFORMATION TECHNOLOGY IN BIOMEDICINE, VOL. 9, NO.

2, JUNE 2005

283

Clinical Knowledge Management Using

Computerized Patient Record Systems:
Is the Current Infrastructure Adequate?
Daniel P. Lorence and Richard Churchill

AbstractThe proliferation of technology in health care,

spurred by environmental factors encouraging the adoption of
computerized patient records (CPRs), has led to a widely held
perception of fully computerized patient information systems
as the industry norm. To test the validity of this assumption,
using data from a national survey of certified health information
managers, we examined the CPR technology adoption rates
reported by health information managers, assessing variation
across practice settings, regions, and organizational types. Results
show that significant nonadoption, and regional variation, exists
in the implementation of CPRs. Overall, nonuniform diffusion of
computerized health information technology was found, despite
national mandates that promote and at times require uniform
adoption. A significantly greater number of hospital-based patient records were computerized, compared to clinics and other
practice settings. Managers were frequently found to maintain
duplicate CPRs and paper-based patient record systems, even
after the initial implementation period. Nonuniform regional CPR
adoption and redundant paper-based systems were found to be
a common practice in medical systems, due in part to cultural
factors, mistrust of computerized data, and lack of technology
training and knowledge.
Index TermsComputerized patient record (CPR), medical
record, paper-based, regional variation.

I. INTRODUCTION

SERIES of recent industry-wide health-care quality

reports suggest that the availability of information technology, along with growing availability of decision support
resources, has for the most part kept pace with health information technology requirements. Whether the adoption of
this technology, and associated management practices, have
likewise kept pace has to date received relatively little attention
[1], [2]. Given this environment, this study seeks to determine:
Do information managers readily adopt computerized patient
records (CPRs) in place of traditional paper-based records?
Does adoption vary across organizational and regional boundaries, and if so how does variation affect the abandonment of
paper-based systems?
The proliferation of information in health care and evidence-based medicine is central to transforming health-care
data from a paper-based recording medium into an integration
Manuscript received November 25, 2003; revised October 5, 2004.
D. P. Lorence is with the Department of Health Policy and Administration,
School of Information Sciences and Technology, Pennsylvania State University,
University Park, PA 16802 USA (e-mail: DPL10@psu.edu).
R. Churchill is with The Virtual Management Institute, Gettysburg, PA 17325
USA (e-mail: rchurch@vminetwork.org).
Digital Object Identifier 10.1109/TITB.2005.847153

of systems-based medical decision-making, disease management, and public health assessment. The potential opportunities,
while promising, present significant challenges [3]. Greater
demands will likely be made on current information systems
for more reliable and timely data quality assessment, affecting
clinical and administrative implementation of evidence-based
medicine practice [4]. Greater reliance on data likewise will
promote the adoption and diffusion of enterprise-wide decision-making methods. While automated information systems
are well adapted to such methods, paper medical records may
not be as compatible with high-speed multisite information
environments.
At the administrative level, the pressure to implement more
timely and comprehensive measures of system-based data has
never been greater. Recent U.S. regulatory actions, such as The
Health Insurance Portability and Accountability Act of 1996,
known as HIPAA, mandates major paperwork reduction and
data standardization practices in patient record documentation
[5], [6]. Strong financial, regulatory, and quality incentives exist
for health information managers to replace paper-based record
systems with automated computer-based technologies [7], [8].
Given this mandate, we surveyed a national group of health information managers regarding their adoption of CPR systems.
Our study sought to determine: 1) the prevalence of CPR adoption level by health-care organizations and 2) the extent of any
significant regional and practice variation in adoption patterns.
II. METHODS
Data from a nationwide survey of 8700 health-care information managers were used to examine the organizational
and environmental characteristics of CPR adoption in a variety of health-care settings. Selected adoption characteristics
were compared across practice settings, geographic areas, and
selected health-care demographic characteristics. The survey
data included measures of the accuracy via automated systems,
and organizational characteristics relevant to data acquisition.
The sample population targeted was selected from a database
of health information managers, certified as Registered Health
Information Administrators or Registered Health Information
Technicians, as of March 30, 2002. Location and demographic
information was provided by the Foundation for Record Education, and contained current and historical information on all
credentialed health information professionals in the U.S. The
survey was designed to provide the health-care industry with
more timely and frequent practice information about health

1089-7771/$20.00 2005 IEEE

284

IEEE TRANSACTIONS ON INFORMATION TECHNOLOGY IN BIOMEDICINE, VOL. 9, NO. 2, JUNE 2005

information managers. No comprehensive study of this kind has

been available previously. The data included on the population
surveyed were obtained primarily from membership renewal
forms and an annual member profile of all active members.
Preferred mailing address data were obtained from the member
population from those members with changes in address or
professional status.
The main questionnaire was developed from past surveys and
focus group results. Managers from a variety of practice settings and work roles were asked questions that were generally relevant to the profession overall, and included: hospitals
and medical centers, group practices, ambulatory care clinics,
managed care offices, long-term care and rehabilitation facilities, colleges, and universities, consulting firms, government
agencies, software product companies, pharmaceutical companies, and self-employed health information management professionals. Topics were pretested with broad groups of practicing
information managers.
Survey announcements appeared in professional publications
and meetings. A preprinted questionnaire was mailed to all credentialed health information managers with identifiable mailing
addresses. An instruction letter accompanied the form and explained the purpose of the survey and instructions for completion. At the six-week point, a second mailing was sent to
nonrespondents. Follow-up on specific issues identified after
the second mailing was accomplished as a series of more focused studies, reported elsewhere. The questionnaires were processed by an independent testing and research firmthe National Computer Systems of Minneapolis, MN. Confidentiality
standards were maintained in this study. Data were entered via
a computerized scanning system and released only in aggregate
form, without individual respondents identified within the reported results. Respondents were asked to report the: 1) percentage of patient record information in computerized form, and
2) percentage of computerized records maintained in redundant
paper form.
We assessed the adoption of computerized records and
reported use of redundant paper records by examining adoption rates which occurred during the originally announced
HIPAA implementation timeframe. As a preliminary analysis,
we employed one-tailed t-tests to examine whether adoption
preferences differed significantly across regions and selected
demographic groups. Coefficients from preliminary estimated
multivariate regression models and means for the control
variables were used to obtain adjusted scores by region, organizational type, or respondent grouping. Probability sampling
weights were calculated as the inverse of sampling probabilities
and used to adjust all regressions for distortions in analytical
samples used when evaluating preliminary cross-group sampling probabilities. To minimize covariate discrepancies related
to respondents reported adoption and level of computerization,
we employed controls of baseline practice setting characteristics of respondents surveyed within logistic regression models.
Further controls for demographic characteristics of location
settings were also employed. To adjust for differences in the
dependent variables, we obtained bivariate logistic regression
measures, minimizing clustering effects across practice settings
and organizational types. We controlled for variation in the

TABLE I
OVERALL PERCENTAGE OF TOTAL PATIENT RECORD
INFORMATION CAPTURED BY COMPUTER

TABLE II
PERCENTAGE OF TOTAL PATIENT RECORD INFORMATION
CAPTURED BY COMPUTER, BY PRACTICE SETTING

dependent variables across computerization levels by comparison with the variation observed during pretest and with like
variation between adopters and nonadopters. We examined
these differences using covariate-adjusted average probabilities, including computerized record adopters and nonadopters,
using raw reported incremental computerization levels as well
as quartile computerization categories. Logistic coefficients
and levels of covariance were then subjected to a Wald test and
examined for significance of observed differences.
III. RESULTS AND DISCUSSION
The survey obtained data from 7151 health information managers, for an 82.2% response rate. Respondents were from a variety of practice settings and job titles and excluded students.

LORENCE AND CHURCHILL: CLINICAL KNOWLEDGE MANAGEMENT USING CPR SYSTEMS

285

TABLE III
PERCENTAGE OF PATIENT RECORD INFORMATION CAPTURED BY COMPUTER, BY REGION

Table I shows that adopters of a significant (greater than 25%)

electronic patient record remain a minority. 51.9% of respondents have less than 25% of their patient information in computerized record format.
Table II shows the mean percentage of current total health
record information captured by computer for hospital respondents (40.3%) are significantly different than the percentages for
clinic (33.6%) and other (34.1%) practice setting respondents.
Table III shows the percentage of Pacific region respondents
(41.3%) is likewise significantly different from the percentages
for Middle Atlantic (35.5%), West North Central (33.4%), South
Atlantic (36.9%), West South Central (34.8%), and Mountain
(35%) regions regarding the average percentage of patient information captured by computer.
Table IV shows the mean percentages for organizations that
have merged within the past two years (41.5%) are also significantly different when compared to the percentages for those that
have not merged (36%), related to the percentage of patient information captured by computer.
Managers were also asked: What percentage of your computerized records are also maintained on paper? As shown in
Table V, nearly 70% of respondents reported maintaining substantial (over 75%) duplicate paper record systems, even after
initial testing and implementation of the CPR.
Generally, dual system maintenance was consistent across
demographic and market characteristics, though there existed
some variation across practice settings regarding dual systems.
Table VI highlights the fact that hospital setting (78.8%) respondents report a significantly different average percentage for
computerized records that are also maintained on paper, compared to clinic (71.9%) and Other (65.8%) practice setting
respondents.
There likewise existed some variation across regional settings
regarding dual systems. Seen in Table VII, when reporting the

TABLE IV
PERCENTAGE OF PATIENT RECORD INFORMATION
CAPTURED BY COMPUTER, BY MERGER STATUS

percentage of computerized records that are also maintained

on paper, Pacific region (78.9%) respondents were significantly different from the West South Central region (72.2%)
respondents.
Table VIII likewise shows the percentage of paper duplication
of the CPR for organizations that have merged. While moderate
variation was seen in the 0% (no CPR) category, little difference
was found at other levels, when compared to the percentages for
those that have not merged.

286

IEEE TRANSACTIONS ON INFORMATION TECHNOLOGY IN BIOMEDICINE, VOL. 9, NO. 2, JUNE 2005

TABLE V
PERCENTAGE OF COMPUTERIZED RECORDS ALSO MAINTAINED ON PAPER

TABLE VI
PERCENTAGE OF COMPUTERIZED RECORDS ALSO
MAINTAINED ON PAPER, BY PRACTICE SETTING

The availability of information technology is highly valued

by health-care policymakers, payers, and patients. Actual adoption of such tools into practice does not always follow, however. Relatively low overall rate of adoption existed here, with
significant variation across geographic and organizational characteristics. Nonuniform diffusion of computerized health-care
technology was the norm, despite national mandates that promote, and eventually require, uniform adoption. Most respondents have not adopted CPRs, and adoption varied significantly
across selected key demographic indicators.
Why are adoption practices still varied? Are implementation
costs in certain areas too great to be reasonable and affordable
in other areas? Are state and local regulatory measures more
stringent than federal standards in some areas? Do managers not
take HIPAA requirements seriously, believing that enforcement
is delayed? When electronic migration does occur, traditional

paper records may still be preferred as a fail-safe in healthcare organizations. Regardless of the motivation, the approach
taken by providers to adopt electronic evidence-based medicine
is likely to be an important factor in ensuring that developments
are met with an appropriate response.
Of note here was the lack of any strong consensus regarding
quality of information within the CPR. Only 51% of respondents
who use computerized records indicate that the overall accuracy
of their patient data is better or much better than data maintained
on paper records. 37% reported accuracy to be the same as with
paper records. This may explain the prevalence of backup, redundant paper-based systems reported in this study. Adoption
of the CPR likewise appeared to have little positive effect on
file space requirements. Among more advanced adopters of the
CPR (those with 50% or greater of their record data computerized), about 70% indicate an anticipated increase in file space
requirements in the next two years. Overall, about 17% of the
high adopters predicted their file space needs would increase by
31% or greater. In contrast, about 81% of all respondents indicated an increase in file space needs in the near future.
Regional variation in the adoption of system-based patient
records, and shortages of trained health information professionals, suggests an urgent need for new training programs
within the health-care professions. If there is insufficient specialist staff to cope with an increased demand, then this may
have to be met from other sources. The need for more focused
areas of information management expertise in the field of data
quality remains one of the greatest challenges faced by the
health-care industry today. Involving health-care professionals
already familiar with paper-based information management
would be a likely first step, serving to develop increased
knowledge of data processes and problems among health-care
professionals. The environment now provides far greater opportunities for electronic migration, with the challenge at the
primary care level being the identification of low-cost scalable
solutions to assess and improve the quality and accessibility of
medical data.
Many health professionals are likely to experience challenges
in implementing evidence-based medicine in their respective organizations, both as a consequence of unfamiliarity with datadriven medical decision-making, and because of the influence
that data will have on specialized fields of medicine. Professionals may have to adapt by acquiring new knowledge and by
adopting new roles, anticipated at several levels of service. Information systems developers will, thus, need to adopt a more
flexible approach to teamwork and incorporate evidence-based
medicine into practice. Traditional resistance to a data-driven
model of care may persist due to its reliance on impersonal authority, but the evolution of health care toward evidence-based
models suggest that medical decisionmaking will grow beyond
the experience and judgment of any given provider.
A. Action Needed
A number of studies to date have recommended potential
areas for further research related to computerization of medical
records. A landmark Institute of Medicine report, for example,
outlined a number of technological barriers to adoption, citing
a need for further maturation of a few emerging technologies,

LORENCE AND CHURCHILL: CLINICAL KNOWLEDGE MANAGEMENT USING CPR SYSTEMS

287

TABLE VII
PERCENTAGE OF COMPUTERIZED RECORDS ALSO MAINTAINED ON PAPER, BY REGION

TABLE VIII
PERCENTAGE OF COMPUTERIZED RECORDS ALSO MAINTAINED ON PAPER

improved decision making [13]. Standards are also needed for

the development of more secure CPR systems [14].
IV. CONCLUSION

such as hand-held computers, voice-input or voice-recognition

systems, and text-processing systems [9]. These were thought
to be critical to further development of state-of-the-art CPR systems, and lacked adequate testing in real-world conditions. In
addition to further development of necessary technology infrastructure, more complete technology standards have been proposed which must be developed, tested, and implemented before the CPR can realize its full potential at both the macro
(e.g., epidemiological) and micro (e.g., physician office) levels
[10][12]. Standards to facilitate the exchange of health-care
data have likewise been suggested so that clinical data may be
transmitted on networks or aggregated and analyzed to support

As early as 1991, consideration of the various barriers to CPR

development, the interest and resources of individuals and organizations able to effect change, and the concerns of individuals
who would be affected by implementation of CPRs prompted
a national U.S. summit [15] to identify eight critical activities
to help advance CPR development: 1) identification and understanding of CPR design requirements; 2) development of standards; 3) CPR and CPR systems research and development; 4)
demonstrations of effectiveness, costs, and benefits of CPR systems; 5) reduction of legal constraints for CPR uses as well as
enhancement of legal protection for patients; 6) coordination of
resources and support for CPR development and diffusion; 7)
coordination of information and resources for secondary patient
record databases; and 8) education and training of developers
and users. Despite these recommendations, relatively few advances have taken place in any of these areas. What is needed is
positive action to bring the U.S. up to the level of computerization currently existing in most of the developed world.
REFERENCES
[1] Doing what counts for patient safety; federal actions to reduce medical
errors and their impact, Report of the Quality Interagency Coordination
Task Force, Washington, DC, Feb. 2000.
[2] R. Kadas, The computer-based patient record is on its way. HMOs, the
economy and HIPAA will drive adoption, Healthcare Informatics, vol.
19, no. 2, pp. 5758, Feb. 2002.
[3] J. Broadhurst, W. Caswell, L. Deneault, T. Ng, N. Rauscher, and F.
Shiner, Using quality improvement tools to improve healthcare internal
quality audits, in 7th Ann. Quality Audit Conf., 1998, pp. 162182.
[4] D. Ballou and H. Pazer, Modeling data and process quality in multiinput
multioutput information systems, Manage. Sci., vol. 31, pp. 150162,
Feb. 1985.

288

IEEE TRANSACTIONS ON INFORMATION TECHNOLOGY IN BIOMEDICINE, VOL. 9, NO. 2, JUNE 2005

[5] S. Harrington, HIPAA, HIPAA, hurrah, Health Manage. Technol., pp.

3138, Apr. 2001.
[6] E. Lanser, Capitalizing on HIPAA compliance, Healthcare Executive,
pp. 711, May 11, 2001.
[7] Health Insurance Portability and Accountability Act of 1996, 104th
Congress, Public Law 104-191, 1996.
[8] D. Brailer, Interview with national health information technology coordinator David Brailer, Brit. Med. J., vol. 329, no. 7471, pp. E328E329,
Oct. 16, 2004.
[9] To Err Is Human: Building a Safer Health System, Institute of
Medicine, National Academy Press, Washington, DC, 1999.
[10] D. Sarudi, HIPAA early adopters, Hospitals Health Netw., vol. 75, no.
2, pp. 3841, 2001.
[11] D. Sands, D. Rind, and C. Safran, Dissemination of electronic patient
records using primary care referrals as a vector for change, Medinfo,
vol. 10, no. 1, pp. 685689, 2001.
[12] D. Bates, M. Ebell, E. Gotlieb, J. Zapp, and H. Mullins, A proposal for
electronic medical records in U.S. primary care, J. Amer. Med. Informatics Assoc., vol. 10, no. 1, pp. 110, Jan./Feb. 2003.
[13] C. Safran and H. Goldberg, Electronic patient records and the impact
of the internet, Int. J. Med. Informatics, vol. 60, no. 2, pp. 7783, Nov.
2000.
[14] Fifth Annual Survey of EHR Trends and Usage, Medical Records Institute, Aug. 2003.

[15] R. Dick and E. Steen, The Computer-Based Patient Record: An Essential

Technology for Health Care. Washington, DC: Nat. Acad. Press, 1991.

Daniel P. Lorence is an Assistant Professor at Pennsylvania State University, University Park, PA, with
appointments in Health Policy and Administration
and in Information Science and Technology.

Richard Churchill is a Senior Research Fellow at The Virtual Management

Institute, Gettysburg, PA, specializing in information science and technology
applications in health-care delivery and medical practice.