Accounting Principles and Measuring Health by shrakdoc



the need for smoking prevention versus cessation—both are                               fore, a GAAP-style system for medicine has to be relatively
important.”6 Comprehensive strategies have been and should                              flexible and responsive. The information must also be reli-
continue to be the focus of global tobacco control efforts.                             able and comparable between sites and therefore must in-
Lawrence O. Gostin, JD                                                                  volve building consensus, similar to the process for GAAP,                                                               because a potential major pitfall of a standardized report-
O’Neill Institute for National and Global Health Law                                    ing system is the ability to misrepresent under the veil of
Georgetown University Law Center                                                        standardization. Finally, to set up a verification system, de-
Washington, DC                                                                          tailed cost-benefit and cost-effectiveness analyses would be
Financial Disclosures: None reported.                                                   necessary to avoid causing net harm by pulling limited re-
1. World Health Organization Framework Convention on Tobacco Control. http:             sources away from other opportunities.
// Accessed Decem-
ber 18, 2007.                                                                           Dev Jayaraman, MD, FRCP
2. Levy DT, Chaloupka F, Gitchell J. The effects of tobacco control policies on smok-
ing rates: a tobacco control scorecard. J Public Health Manag Pract. 2004;10            McGill University Health Center
(4):338-353.                                                                            Montreal, Quebec, Canada
3. Hopkins DP, Briss PA, Ricard CJ, et al. Reviews of evidence regarding interven-
tions to reduce tobacco use and exposure to environmental tobacco smoke. Am J           Howard Rivenson, PhD
Prev Med. 2001;20(2)(suppl):16-66.                                                      Harvard School of Public Health
4. Task Force on Community Preventive Services. Strategies for reducing expo-           Boston, Massachusetts
sure to environmental tobacco smoke, increasing tobacco-use cessation, and re-
ducing initiation in communities and health-care systems: a report on recommen-         Financial Disclosures: None reported.
dations of the Task Force on Community Preventive Services. MMWR Recomm
Rep. 2000;49(RR-12):1-11.                                                               1. Pronovost PJ, Miller M, Wachter RM. The GAAP in quality measurement and
5. Framework Convention Alliance. Priorities for the Second Session of the FCTC         reporting. JAMA. 2007;298(15):1800-1802.
Conference of the Parties. May 3, 2007.              2. Zelman WN, McCue MJ, Millikan AR, Glick ND. Financial Management of Health
/fca-2007-cop-cop2-priorities-en.pdf. Accessed December 18, 2007.                       Care Organizations: An Introduction to Fundamental Tools, Concepts, and
6. US Department of Health and Human Services. The Health Benefits of Smok-             Applications. Malden, MA: Blackwell Publishing; 2003:37-74.
ing Cessation. A Report of the US Surgeon General. Rockville, MD: Office on Smok-       3. Kane NM, Magnus SM. The Medicare Cost Report and the limits of hospital
ing and Health; 1990.                                                                   accountability: improving financial accounting data. J Health Polit Policy Law. 2001;
                                                                                        4. Bullen HG, Crook K. Revisiting the concepts: a new conceptual framework project.
Accounting Principles                                                          Accessed January 15,
and Measuring Health Care Quality                                                       5. Ding Y, Jeanjean T, Stoloway H. Why do National GAAP differ from IAS? the
                                                                                        role of culture. Int J Account. 2005;40:325-350.
To the Editor: In their Commentary, Dr Pronovost and col-
leagues1 discussed generally accepted accounting prin-                                  In Reply: In response to Drs Jayaraman and Rivenson, we
ciples (GAAP) and quality measurement. Although we agree                                recognize that measuring health care quality is exceed-
that a standardized reporting system using GAAP as a model                              ingly complex, and standards for reporting will evolve. Rather
is a good idea in theory, there are several cautionary les-                             than offer a prescriptive solution, the intent of our Com-
sons from a relatively simple system like accounting to keep                            mentary was to start a long overdue discussion. Although
in mind when compared with the complexities of quality                                  we support public reporting of quality, the data must be ac-
assurance measurements in medicine.                                                     curate to be helpful.
   Information conveyed in external reports may lack de-                                   Presently, the reporting of quality measures is like the Wild
tail required by internal reports, and vice versa. Therefore,                           West. Without national or international standards as guide-
the level of detail to be included in a GAAP-style report must                          lines for health care organizations publicizing quality mea-
be specified for either external or internal audiences. It is                           sures, there is little assurance that the information re-
not unusual for basic financial statements that consist of a                            ported is based on science. Such measures may misinform
balance sheet, income statement, statement of changes in                                consumers.
equity, and statement of cash flows (often 4 pages) to be                                  The contrast between reporting financial performance and
followed by a footnote section that can be more than 20 pages                           reporting quality performance is dramatic. Using GAAP is
long.2 Companies may also have to comply with several ex-                               only one component in a structured system that includes
ternal standards that differ from GAAP, such as one for the                             trained professionals to produce financial statements, as well
Securities and Exchange Commission and another for Medi-                                as mechanisms to audit and hold businesses accountable if
care.3 Even in accounting, where the data are relatively                                they misrepresent their financial performance. The entire
straightforward, there are lively debates about the superi-                             financial market is predicated on the accuracy of financial
ority of GAAP vs the International Accounting Standards                                 statements that allow investors to estimate the value of se-
used outside of the United States and Canada4,5; a similar                              curities.
trans-Atlantic debate about performance indicators for medi-                               Consumers should have confidence that publicly re-
cine would likely be even more contentious.                                             ported quality measures are accurate. This will require find-
   The demands of substantive relevance and timeliness are                              ing the proper balance between informing and overloading
likely to be harder to satisfy in medicine than in finance be-                          consumers with information. For measures that have na-
cause technology and treatments change rapidly. There-                                  tional standards, health care organizations could report mini-
764 JAMA, February 20, 2008—Vol 299, No. 7 (Reprinted)                                               ©2008 American Medical Association. All rights reserved.

                                         Downloaded from at University of Toronto, on February 25, 2008

mal details. For measures that lack national standards, such                         nursing facilities. Thus, the original 303 sepsis-related “hos-
as those developed by individual hospitals, more informa-                            pitalizations” were incorrectly reported and are more accu-
tion is needed to make biases transparent. A framework1                              rately described by the broader term sepsis events, which in-
based on the Users’ Guides to the Medical Literature2 has                            clude events in a hospital setting (146 events) and events
been developed to help health care organizations evaluate                            in other settings (157 events).
the validity of their quality measures. Rather than provide                             With correction of the title and text of the article (see Cor-
guidance on a specific measure, the framework provides guid-                         rection in this issue) by changing “hospitalizations for sep-
ance on estimating and reporting biases in quality and safety                        sis” to “sepsis events,” the results in the original manu-
measures. Such a framework can initiate discussion of data                           script are accurately described by this broader definition.
elements that should be reported for quality measures. It                            The original analyses were correct for this definition of sep-
must be coupled with additional research to identify evi-                            sis events, and the results of the published article remain
dence-based quality standards, determine the best strategy                           unchanged, as does the conclusion that statin use was as-
for enforcing standardized reporting, and evaluate the costs                         sociated with a reduced risk of sepsis in this cohort.
and benefits of reporting.                                                              On behalf of our coauthors, we apologize to the JAMA read-
Peter J. Pronovost, MD, PhD                                                          ers and editors for these errors in our study and for any con-                                                                    fusion or inconvenience caused by publication of this in-
Johns Hopkins University School of Medicine                                          correct information.
Baltimore, Maryland
                                                                                     Rajesh Gupta, MD
Marlene Miller, MD, MSc                                                              Laura C. Plantinga, ScM
National Association of Children’s Hospitals                                         Johns Hopkins University
   and Related Institutions                                                          Baltimore, Maryland
Alexandria, Virginia
                                                                                     Neil R. Powe, MD, MPH, MBA
Robert M. Wachter, PhD                                                     
University of California, San Francisco                                              Johns Hopkins Medical Institutions
Financial Disclosures: Dr Pronovost reported receiving grants from the Michigan      Welch Center for Prevention, Epidemiology,
Health and Hospital Association, the New Jersey Hospital Association, Rhode Is-         and Clinical Research
land Quality Institute, and MCIC Inc (a liability insurer) to measure and improve    Baltimore, Maryland
safety and honoraria from various hospitals to speak about measuring quality. Dr
Wachter reported receiving stock options from Hoana Medical (Honolulu, Ha-           Financial Disclosure: None reported.
waii) and Intellidot (San Diego, California) and honoraria from Google for serving   1. Gupta R, Plantinga LC, Fink NE, et al. Statin use and hospitalization for sepsis
on their scientific advisory boards and reported serving on the board of directors   in patients with chronic kidney disease. JAMA. 2007;297(13):1455-1464.
of the American Board of Internal Medicine. Dr Miller did not report any disclo-

1. Pronovost PJ, Berenholtz SM, Needham DM. A framework for health care or-
ganizations to develop and evaluate a safety scorecard. JAMA. 2007;298(17):
2. McAlister FA, Straus SE, Guyatt GH, Haynes RB; Evidence-Based Medicine Work-
ing Group. Users’ guides to the medical literature, XX: integrating research evi-    Inaccurate Classification: In the Original Contribution entitled “Statin Use and Hos-
dence with the care of the individual patient. JAMA. 2000;283(21):2829-2836.         pitalization for Sepsis in Patients With Chronic Kidney Disease” published in the
                                                                                     April 4, 2007, issue of JAMA (2007;297(13):1455-1464), the location of the sep-
                                                                                     sis events was incorrectly classified. The title should read, “Statin Use and Sepsis
                                                                                     Events in Patients With Chronic Kidney Disease.” Throughout the article (in the
Correction: Inaccurate Classification and                                            abstract, text, tables, and page headers), the phrases “hospitalization for sepsis”
Information Reported in a Study of Statin Use                                        and “sepsis-related hospitalizations” should read “sepsis events.” In addition, on
and Sepsis in Patients With Chronic Kidney Disease                                   page 1455, in the abstract, the main outcome measure should read, “Sepsis events
                                                                                     were determined from United States Renal Data System administrative files.” This
                                                                                     sentence should also replace the one on page 1456, in the right column, third para-
To the Editor: We are writing to inform the readers and edi-                         graph, the third sentence. On page 1457, in the first paragraph, the first com-
tors of JAMA about errors in a study of statin use and sepsis                        plete sentence should read, “Only episodes in which the primary event was sepsis
                                                                                     were included in our analysis to avoid including cases in which infection was ac-
in patients with chronic kidney disease published in the April                       quired as a secondary phenomenon.” On page 1463, in the middle column, the
4, 2007, issue of JAMA and for which we were the principal                           first complete sentence should read, “Although this method has limitations, ad-
                                                                                     ministrative information has been used widely in other observational studies to
investigators.1 These errors were detected in the process of                         determine outcomes.” See also related letter in this issue.
reviewing our coding algorithms for health care utilization
                                                                                     Data Errors: The Review article entitled “Mortality in Randomized Trials of Anti-
data from which we identified sepsis events.                                         oxidant Supplements for Primary and Secondary Prevention: Systematic Review
   In our article, we had identified sepsis events by using                          and Meta-analysis” published in the February 28, 2007, issue of JAMA (2007;
validated International Classification of Diseases, Ninth Re-                        297[8]:842-857) contained data errors. On page 842 in the “Data Synthesis” sec-
                                                                                     tion of the abstract, the lower confidence limit for the “multivariate meta-
vision (ICD-9) billing codes (038.0-038.9, septicemia; 790.7,                        regression analyses showed that low-bias risk trials” that read “1.05” should have
bacteremia) in United States Renal Data System adminis-                              read “1.04.”
trative files. However, we inaccurately identified the source                        On page 844, in the first paragraph of the “Results” section, the sentence describ-
administrative files as hospitalization billing files only. The                      ing the types of designs used in the study trials that read “Forty trials used parallel-
                                                                                     group design, 26 factorial design (23 trials 2 2; 2 trials 2 2 2; 1 trial half
files we received actually included billing data from other                          replicate of 2     2     2    2), and 2 crossover design,” the parenthetical break-
treatment settings as well, including outpatient and skilled                         out of the factorial design that attributed “23” to the 2 2 design, should have read

©2008 American Medical Association. All rights reserved.                                                (Reprinted) JAMA, February 20, 2008—Vol 299, No. 7             765

                                        Downloaded from at University of Toronto, on February 25, 2008

To top