H O S P I T A L

D I S C L O S U R E

Hospital Disclosure Practices: Results Of A National Survey

Most hospitals disclose harm to patiehts at least some of the time, this 2002 survey finds.
by Rae M. Lamb, David M. Studdert, Richard M.J. Bohmer, Donald M. Berwick, and Troyen A. Brennan
ABSTRACT: New patient safety standards from JCAHO that require hospitals to disclose to patients aii unexpected outcomes of care took effect 1 July 2001, In an early 2002 survey of risk managers at a nationally representative sample of hospitals, the vast majority reported that their hospitai's practice was to disclose harm at least some of the time, although only one-third of hospitals actually had board-approved policies in place, iVlore than half of respondents reported that they wouid always disclose a death or serious injury, but when presented with actuai ciinical scenarios, respondents were much less likely to disclose preventabie harms than to disclose nonpreventabie harms of comparable severity. Reluctance to disclose preventable harms was twice as likely to occur at hospitals having major concerns about the maipractice implications of disclosure.

ELLING PATIENTS ABOUT UNANTICIPATED OUTCOMES of Careis an es-

tabUshed ethical expectation for physicians and nurses.' However, decisions about the appropriateness, timing, and content of disclosure have traditionally remained a private matter, left to the preferences of individual clinicians and health care institutions. Advances in informed-consent law and patients' rights over the past thirty years appear to have had httle demonstrable impact on providers' willingness to disclose information about errors and adverse outcomes.^ Today providers face new constraints in the area of disclosure. The Institute of Medicine's (IOM's) 1999 report. To Err k Human, prompted calls for greater transparency in health care.^ In July 2001 the Joint Commission on Aecreditation of Healthcare Organizations (JCAHO) responded by introducing new patient safety standards, including a requirement that all unanticipated outcomes of care be disclosed.*' Although the requirement itself does not specify the need to disclose poor

Roc hamb is a hedtk correspondent for Radio New Zealand in Wdlif^cm. Da^

Studdert is an assistant

Health, in Boston. Rkhani Bohnir is an assistant professor of technology and operations management at the Harvard Business School Don Berwtcfe is presidmt anichie/execultve affiixr ofd\e Institute/or Healthcfifc Improvement in Boston. Troy Brermcm is aprofessor of medicine at Harvard Mcdiall School and a professor of law aripMichealihattiitHayardSAoolofPikiclkdllk

HEALTH AFFAIRS - Votumt 22. Nlnr

I N F O B M A T I O N

outcomes, JCAHO has clarified that accredited organizations must tell patients when harms occur to them in the course of treatment.' To investigate how hospitals are dealing with this standard, we surveyed risk managers from a nationally representative sample of hospitals. We sought information on how and what hospitals were disclosing six months after the JCAHO standards took effect. We also sought to gauge the importance of several potential barriers to disclosure, including fear of litigation.

Study Methods
Suivey design. We developed the survey instrument through extensive consultation vvith physicians, risk managers, senior hospital administrators, patients, and experts in patient safety and quahty improvement. A draft version was pretested on chief medical officers and risk managers at four different hospitals to determine validity and the type of respondent best able to answer the questions. We judged risk managers to be the most kno^vledgeable and appropriate respondents. The survey comprised three sections.'^ The first section asked respondents about their institutional policies and practices related to disclosure. We defined disclosure as 'honestly telling patients or their families about unexpected harm that occurs as a result of treatment or care, not directly because of a patient's illness or underlying condition." Section two elicited specific information on disclosure practices, including respondents' propensity to disclose harms of varying severity levels, the elements commonly included in disclosures (for example, explanation, apology, acknowledgement of harm, and undertaking to investigate), and actions that commonly accompany a disclosure (for example, pay costs of associated care, pay compensation, and provide details of support groups). We also asked respondents to estimate the likelihood that each of four different clinical scenarios would be disclosed at their institution. The scenarios mixed combinations of severity and preventabiHty.' Section three elicited information about the actual frequency of disclosure, trends in disclosure practice, barriers to disclosure, and perceptions of malpractice risk. Sampling and administration. We used a stratified random sampling approach to select 500 hospitals from the American Hospital Association (AHA) database of 1,218 medical/surgical hospitals with 200 or more beds. One stratum was based on facility size, with half of the sample coming from hospitals with 200-399 beds and the other half from hospitals with 400 or more beds. A second stratum, based on region, ensured a representative geographic spread. Exclusion of Veterans Affairs (VA) hospitals and several other ineligible institutions resulted in a sample of 493 hospitals (245 hospitals with 200-399 beds; 248 hospitals with 400 or more beds).^ We derived sampling weights to allow adjustment of the survey results to represent the larger sample of AHA hospitals vvlth 200 or more beds. Finally, using AHA contact information, we mailed the survey to risk managers at each of the sampled hospitals in January 2002. Two weeks

ch/April 2003

H O S P I T A L

D I S C L O S U R E

later a second copy of the survey was sent to nonrespondents, with intensive telephone follow-up. Analysis. We used existing descriptive data on the sampled hospitals from the AHA's annual survey for 2000 to categorize them by ownership (for-profit, not-forprofit, and government), whether or not they were academic medical centers (AMCs), number of admissions, and region (using the four census regions). We created a binary variable separating the hospitals situated in the sixteen states with mandatory reporting laws in place at the time of our survey from those in states without such laws.' We used the STATA statistical package to conduct weighted analyses of the survey response data. We calculated descriptive statistics summarizing disclosure poUcies, practices, and experience. We also used chi-square tests to test for statistically significant differences in responses to the clinical scenarios based on respondents' malpractice perceptions and behefs. Finally, we used logisric regression to investigate factors associated with divergent reporting practices among respondents.

Results
Of the 479 surveys mailed, we received 338 replies, yielding 245 usable responses for analysis: a completion rate of 51 percent.'^ The hospital characteristics for this respondent group closely resembled those of the nonrespondents with two exceptions: For-profit hospitals were underrepresented among respondents (p = .05) and AMC hospitals were overrepresented (p < .001) (Exhibit 1). Disclosure policies and practices. Approximately one in three hospitals had board-approved disclosure policies in place, and nearly half were in the process of developing a formal policy (Exhibit 2). The remainder had no disclosure policy Nevertheless, 54 percent of respondents reported that it was routine practice at their hospital to tell patients or their famihes when a patient had been harmed by care. Another 44 percent reported that such disclosures occurred some of the time, leaving only five respondents who said that their hospitals did not disclose harms. With respect to the types of harms generally disclosed, 65 percent of hospitals reported always disclosing death or serious injury. A smaller propordon always disclosed in the case of serious, short-term harms (Exhibit 2). The most common elements of disclosures were an explanation, an undertaking to investigate the Incident, an apology, and an acknowledgement of harm. Relatively few respondents reported that a typical disclosure included a declaration of responsibihty for the harm or a promise to share investigation results with the patients or their families. However, thirty-seven respondents (17 percent) indicated that disclosures at their hospitals routinely included all six of the elements we queried them about. The majority of hospitals also met the costs of health care associated with the harm, but few paid compensation or provided details of outside support groups, regulatory agencies, or lawyers.

HEALTH A f FAIRS - Votume 23,

I N F O R M A T I O N

EXHIBIT 1 Characteristics Of Sampled Hospitals, Respondents And Nonreapondente

RMpo!utentft<n-242) CharactarMtcs Mean number of beds 200-399 400 ov more Mean number of admissions Less than 20.000 20,000 or more iHospital ownership'^ Government (excluding federal) For-profit Not-for-profit Academic medical centers" Mandatory reporting laws Region Northeast Midwest South West Number Percant^ Nonnspondents (n 237)* NuRibw Percsirt'

117 125

48% 52 57 43 15 6 79 57 49 27 27 32 14

125 112 134 103 31 30 176 92 119 52 50 97 38

53% 47 57 43

13? 105 37 15 190 138 118 66 65 78 33

13 13 74 39 50 22 21 41 16

SOURCE: Authors' analysis. NOTES: For respondents, the mean rRirriber of beds was 437; for n on respondents, 439, For respondents, the mean number of admissions was 20,254; for nonrespondents, 20,321, Numbers may not add to 100 because of rounding. 'Three respondents are missing from the respondentscolumn. and ttiey appear in the notirespondents column. The rerftoval of code numbers on three surveys made it impossiOle to match these responses with the hospital characteristics, "Weighted. "p < .05 In chi-square test for difference between respondents and non respondents, "p < 0 1 in chi-square test for difference between respondents and non respondents.

Seventy percent of respondents said that the number of disclosures made in their hospitals had Increased in the past two years, but half still reported fewer than five disclosures per 10,000 annual admissions (Exhibit 3). The mean frequency vi'as 24.6 disclosures per year across all respondent hospitals. This corresponds to a mean of 7,4 disclosures per 10,000 admissions. Medta coverage. News media coverage did not appear to affect hospitals' willingness to tell patients about harm (Exhibit 3). Seventy-nine respondents indicated that their hospitals had experienced harms to patients that were reported prominently by the news media. Of these, the vast majority said that the publicity had no impact on their organization's future v^^ngness to disclose. Utigatlon risk. Respondents were divided in their beliefs about whether disclosure increased, decreased, or did not alter the disclosing clinician's or institution's probability of being sued (Exhibit 3). Nonetheless, fear of medical malpractice litigation was still the most commonly cited institutional barrier to developing and implementing disclosure policies, followed by staff opposition. Dteclosing preventable harms. Exhibit 4 summarizes respoases to the four clinical scenarios we presented. Although most respondents indicated that their

ch/April 2003

H O S P I T A L

D I S C L O S U R E

EXHtBfT2 Hospttal Disclosure Policies And Practices

Number of hMpttatswrfth Policy/practice
Policy status Established Under development None Practice Routinely disclose Sometimes disciose Do not distose Tipes of harms disciosed Death/serious injury Always Freq uentiy/sometimes Never Serious shott-term harm Always Frequently Never Elements of typical disclosure" Explain Undertake to investigate Apoiogize Acknowiedge harm Promise to share Investigation resuits Take responsibility for harm Actions foiiowing typicai disclosure" Pay compensation Provide details of support groups Provide regulatory agencies'ctetaiis Provide information about lawyers SOURCE: Auttiot^' analysis. ' Weighted. The frequencies and percentages may not sum to 245 and 100 percent, respectiveiy, because of a smali number of missing values and "don't know" responses. Do not sum to 100 percent because respondents had muitipie options.

Percent of hospital* wttli practtce" 36% 44 19 54 44 2

86 111 44 132 108 5

159 84 1 90 154 0 224 212 161 160 94 76

hospitals were likely to disclose serious and minor harms when those harms were preventable, the overall propensity to disclose preventable harms was lower than the propensity to disclose nonpreventable harms of corresponding severity (for serious harms, 90 percent versus 94 percent; for minor harms, 80 percent versus 97 percent). Indeed, specific within-hospital comparisons across the injury scenarios showed that more than half of hospitals were less likely to disclose the preventable harms than the nonpreventable ones. Exhibit 4 also contrasts responses to the different clinical scenarios across two groups of respondentsthose who believed that disclosure increased the risk of litigation and/or who cited malpractice concerns as an institutional barrier to disclosure, and those who reported no such malpr^tice concerns. A significantly smaller proportion of the group concerned about malpractice was likely to dis'

HEALTH AFFAIR.S - Volumt 22. Nnii.ei

I N F O R M A T I O N

EXHIBIT 3 Hospitals' Experiences With And Beliefs About Disclosure

Number of Percent Of

Experlence/bllef
Mean number of disclosures <5 per 10,000 admissions >5 per 10,000 admissions Trend in numbef of disclosures over past 2 years Increased Decreased No change Adverse publicity about disclosed harms Impact (of publicity) on future willingness to disclose" Less More No impact Main barriers to disclosure'^ Malpractice fears Staff opposition Fear of scaring patiente Physicians/hospital had different malpractice insurers Cost concerns Beliefs about impact of disclosure on malpractice r isk Increased Unchar.ged Decreased

hospitals

126 119 170 2 71 79 2 18 58

51% 49

70 1 29 28 3 21 74

187 126 63 54 21

92 63 80

37 25 33

SOUfti Authors' analysis. NOT: The mean number of disclosures was 7.4. "Weighted. The frequencies and percentages may not sum to 245 and 100 percent, respectively, because of a small number of missing values and 'dmi't know" responses. "This applies oniy to tbe 79 respondents who reported publicized harm. " Do not sum to 100 percent because respondents had multiple ootions.

close preventable serious harms (p = .02). In addition, the group of institutions worried about litigation was significantly more likely to exhibit a preference for disclosing nonpreventable harms over preventable ones (p = .02). Multivariate analysis confirmed that malpractice concerns were associated with hospitals' being significantly less likely to disclose preventable harms than nonpreventable ones (ockls ratio 2.03, p = .03)." No other hospital characteristics had a statistically significant association with reluctance to disclose preventable injury.

Discussion
The explosion of public interest in medical error following the 1999 IOM report galvanized attention on consumers' expectations around disclosure of medical error. Several years later it is timely to ask what has changed for patients, if one measure of the report's success is the extent to which hospitals own up to error, then our study provides some encour^^ii^ results. Virtually all (98 percent) of the respondents in our study reported disclosing harms to patients at least some of the

March/Apri! 2003

EXHIBIT 4 Hospitals' Willingness To Disclose Harms, By Level Of Concern About Malpractice

Percent of hospitals not concerned about malpractice* (n - 1 6 8 )

Disclosure re^Hmse to harm scenarios Likeiy to disciose preventabie serious harni Likeiy to disclose preventable minor harm Likely to disdose nonpreventabie serious harm Likeiyto disclose nonpreventabie minor harm Less likely to disclose preventable than nonpreventabie harms'"

Number of hospHals (N-245)

Percent of hospitals

Percwrtof ho^ritals concerned about mslfKactlce '* (n - 77)

219 195 232 236 129 94 97 53 90 95 65 96 98 47'=

SOURCE: Authors' analysis, ' Weighted. ' Consists crt responderrts wtio reported Oelief that disclosure increased their hospitars probability of being sued and noted concerns about lawsuits as a main barrier to developing and implementing a disclosure policy. ' p < ,05 in adjusted Pearson's chi-square test for difference with "concerned at>out malpractice' group. "Consists of respondents who reported that they were likely to disclose the scenario Involving nonpreventabie serious harm or nonpreventabie minor harm, but unlikely to disclose the corresponding scenarios involving preventable harm.

time, and 80 percent had disclosure policies in place or under development. The fact that 44 percent of surveyed hospitals v^^ere in the process of developing disclosure policies at the time of our survey suggests that the IOM's message, together with the patient safety initiatives it has sparked at JCAHO and other agencies, is driving substantial reform. Follow-up investigation of the final form and content of the many institutional policies that were budding at the time of this survey would add greatly to our knowledge in this area. Disclosure frequency. Our study also suggests that there is still a long way to go before serious harm is consistently and thoroughly disclosed to patients. For example, our respondents reported considerably fewer disclosures than wouid be expected from epidemiologic estimates of general rates of iacrogenic Injury. Leading studies of medical injury from Utah/Colorado and New York, which were used as the basis of the IOM estimates, found that adverse events occurred in 2.9 percent and 37 percent of hospitalizations, respectively.'^ These rates imply 290-370 potentially disclosable harms per 10,000 admissions. Adjusting the estimates of adverse events to include only the most serious incidents suggests approximately 44-66 medical injuries per 10,000 admissions that should be disclosed. Only two hospitals in our study were in this range. In fact, only sixteen hospitals (less than 10 percent) reported making more than twenty disclosures per year. Alternative explanations certainly exist for the gap between these estimates and the number of disclosures reported in our survey. Specifically, the risk manag-

HEALTH AFFAIRS - Volume 22. Numbei

I N F O R M A T I O N

"The openness ofproviders about error depends on reforms of the malpractice system that can mitigate the fear it generates."
ers we surveyed may not be aware of all disclosures in their hospitals, and rates of injury in our national sample of hospitals may be louver than those previously identified in New York, Utah, and Colorado. Nevertheless, it seems likely that disclosure of the most serious events would come to risk managers' attention. Hence, the relatively low rate of disclosure reported in our study raises questions about both the extent to which harms are recognized by hospital staff and the frequency with which known harms are disclosed. The malpractice barrier. Hospitals' heightened reluctance to disclose preventable harms raises further questions. Of all hospital and respondent characteristics we examined, fear of litigation was most strongly associated with this reluctance. The challenge that the medical malpractice environment poses for patient safety efforts has been well documented, as have the fears of physicians." Some conclude that the cultural change necessary for major safety improvements cannot occur against the backdrop of a litigation system that induces secrecy and silence, and they stress the need for malpractice reform.^** A different, and increasingly prominent, twist on the malpractice issue is that clinicians' and hospitals' perceptions about litigation risk may be worse than the reality." The experience of the VA Medical Center in Lexington, Kentucky, where a proactive disclosure policy has reportedly not resulted in higher hability payments at the institutional level, is widely cited."^ There is also growing anecdotal evidence from some nongovernmental hospitals, such as the Dana Farber Cancer Institute in Boston, that their policies to disclose have not been accompanied by a big increase in lawsuits.'' Another Massachusetts hospital. Sturdy Memorial, says that it found owning up to error a positive experience.'^ These reports find support in studies suggesting that patients who are dealt with openly and honestly are less likely to sue.'^ Our findings do not contradict any of this. However, they do suggest that regardless of whether or not providers' concerns about malpractice are well founded, litigation fears continue to pose a serious obstacle to transparency about patient injury. Important breakthroughs in the openness of providers about error may thus depend on reforms of the malpractice system that can mitigate the blame, guilt, and fear it generates." Malpractice reforms are a top priority for policymakers at the federal and state levels.^' However, these proposals tend to focus on the immediate problem of claims volume and award size, not the more fundamental issue of the barriers the system creates for advances in the patient safety arena. Adveise publicity. There is ongoing debate about the impact of adverse publicity on patient safety advances, with some commentators arguing that it provides

Marct/April 200!

important impetus and others that it may cause inertia,^ Our survey results suggest that such coverage has Uttle effect in the area of disclosure. Only two respondents indicated that thdr hospital's willingness to disclose had decreased as a result of prominent reporting of cases of harm. Study limitations. There are several Hmitations to our study. A survey completion rate of 51 percent introduces the possibility of nonresponse bias. The similarity in hospital characteristics between respondents and nonrespondents provides some comfort in this regard, although for-profit hospitals and non-AMCs were underrepresented among respondents, and their disclosure behavior may have differed systematically.^^ However, the fact that a substantial proportion (27 percent) of the ninety-three respondents who said that they would not complete the survey cited legal concerns as the reason bolsters rather than undercuts our findings about the impact of litigation fears on willingness to disclose. Second, the findings may also be limited because the survey is focused on risk managers. Our decision to target this group was based on a recognition that risk managers in many hospitals are at the center of efforts to develop formal written disclosure processes and policies to comply with the new JCAHO standards.^"* Although physicians have not always viewed hospital risk managers as advocates of disclosure, a previous survey of 650 risk managers suggested that their personal support of disclosure is consistent with that of other managers and may even exceed the willingness of their organizations to disclose.^' Finally, we used only four injury scenarios to measure the willingness to disclose preventable versus nonpreventable harm. These findings should be explored further by testing a wider range of possible clinical events.

HE RESULTS OB THIS SURVEY give some cause for optimism; A large proportion of hospitals appear to be telling patients about harms caused by medical care. Moreover, the far-reaching impact of the IOM report and the JCAHO standard are evident in the sizable number of hospitals that are in the process of developing disclosure policies. However, it is clear that the spread and execution of such policies and practices fail short of the standards that would be expected in a therapeutic model based on partnership and patient empowerment.^' There is stiU marked variation in the types of harm that hospitals are prepared to disclose and how they handle such disclosure. Malpractice concerns appear to be the most prominent foil to aspirations of openness. As malpractice insurance costs spiral for physicians in a number of states and pundits herald a fresh set of malpractice "crises," the litigation barrier looks set to g

HEALTH AFFAIRS - Volu

I N F O R M A T I O N

The authors thank Shimon Shaykcvichforpro^mming si^port; Meghan Mattinofor datxi entry work and other icgistical help; the chitf medical officers, risk managers, and others who offered expertise in the survey development; and the respondents who took time to participate. This work was conducted while Rae Lamb was a 2001-2002 Harkncss Fellow in Health Policy, bdsed joinrl)' at the Harvard School 0/Public Healtfi and lihe Institute fov Healthcare Improvement and supported by the Commonwealth Fund. David Studdert was supported inparthyGrantm}.K02HSlUS5fromthe/^ncyforHea}thcarc'Researchand^uality.Theviewspresentcdherc are tftoseo/rfie authors and not necessarily those ofthe Commonwealth Fund.

HOTES 1. American Medical Association, ""Code of Medical Ediics; Current Opimons," 31 July 2002, www.ainii' assn.o]^aina/pub/category/2503.html (30 December 2002); Artierican College of Physicians, "Ethics Manual, Fouith Edinon," Annuls ofMcma\ Medicine 128, no. 7 (1998): 576-594; and American Nurees Association, "Code of Ethics for Nurses with Interpretive Statements," 2001, www.nursingworld.org/ethics/code/ etliicscodeI50.htm (30 December 2002). 2. T.A. Brennan, Jusi Doctorir^ Medical Ethics in the Ubcral State (Berkeley: University of California Press, 1991); M.L. Millenson, DemandingMedical Excellence: Doctors and Accountability in thelr^ormationA^ (Chicago: Univecsity of Chicago Piess, 1997); and L.M. Peterson and T.A. Brennan, "Medical Ethics and Medical Injuries: Takii^ Our Duties Seriously,7ouma( ofCUnicd Ethics 1, no. 3 (1990): 207-211. 3. L T . KohnJ.M, Corrigan, andM.S. Donaldson, e6s.,ToErrIsHuman:Buildii^aSaferHcalthSyitem (Washington: National Academy Press, 1999); Institute of Medicine, Crossir^ the Quality Chasm: A New Health System for tk Vwmty-fint Cemury (Washii^^n: National Academy Press, 2001); and White House Press CSice, -Medical Errors Remarks by President Chnton," Press Release, 7 December 1999. 4. Joint Commission on Accreditation t^ Healthcare Organisatioiis, Standard RI.1.2.2,1 July 2001. 5. JCAHO, JCAHO press conference transcript, 2 July 2001. Atlanta, Georgia, 6. The Insntutionai Review Board at the Harvard School of Public Health approved boiii the survey and the study design. 7. The four clinical scenarios were (1) preventable event with serious outcome (a patient suffei^ a catastrophic hemorrhagic cerebrovascular accident following heparin pump failure that causes heparin overdose); (2) nonpreventable event with serious outcome (patient widi no kno\\'n drug alleigies dies as a result of anaphylactic shock secondary to arapidlhn); (3) preventable event with minor outcome (coumaditi overdose leads to elevated prothrombin time and two additional, otherwise unnecessary, hospital days); and (4) nonpreventable event vidth minor outcome (patient wdth no drug allergies develops rash and itchi i ^ secondary to antibiotic reaction). B. VA hospitals were excluded from both strata. We also removed seven hospitals from the sampled group: Six were in Puerto Rico, and one had specially requested exclusion. 9. J. Rasenthal et aL, State R^rtir^ 0/ MaJical Errors and Adverse Events: Results 0/ a Fifty-Stare Survey (Portland, Maine: National Academy for State Health Policy, 2000). Utah has a Patient Safety Senrinel Event Reporting Rule R380-200-3, effective .15 October 2001. 10. Of tlK 338 total replies we received, ninety-three were refusals to complete the full survey, with twenty-five of these d d n g as their primary reason that they or others at their institution had legal concerns about doing so. An additional eighteen of the refusers said they were too busy, fifteen said that they did not participate in surveys, and ten said it was too early in the development of their disclosure pohcies to answer our questions. 11. The dependent variable in this multivariate model was propensity to disclose preventable harm, as determined by responses to the clinical scenarios. The independent variables were msdpractice cca:icems, AMC status, ownership, number of beds, reporting law environment, and region. 12. EJ. Thomas et aL, "Incidence and Types of Adverse Evetits and Negligpnc Care in Utah and ColOTado," Medical Care 38, no. 3 (1999): 261-271; aad T,A. Brennan et aL, "Indcknce of Adverse Events and Negligence in Hospitalized Patients: Results of the Harvard Medical Practice Study I," New Er^Jand Journal of Medicine 324, no. 6 (1991): 370-376. 13. B.A. liar^, "Enor in Medicine: Legal Impediments to US. Reform," Journalf^HeaitfiRjItrics, Policy andLaw 24, no. 1 (1999): 27-58; L. Gostin, "A Public Healdi ^ ^ r o a c h to Reduci:^ EITOT: Medical Malpractice as a Baina:,'-Joumdof^AmmcanMedical Association 283, no. 13 (2000): 1742-1743; A.W. Wu, "Handling Hospital Errors: Is Disclosure the Best Defenser Atmh t^ internal Mdicint Bl, no. 12 0999)-. 970-972; M.

Maich/Aptil 200J

H O S P I T A L

D I S C L O S U K E

Hingorani, T. Wong, and G. Vafidis, "Patients' and Doctors' Attitudes to Amount of Infonnation Given after Unintended Injury during Treatment: Cross Sectional, Questionnaire Survey," British Medmlpurmd 318, no, 7184 (1999): 640-641; and A.W. W u et al,, "Do House Officers Leam from Tlieir M i s t a k e s ^ l / f t American tvkdical Association 264, no. 4 (1991): 2089-2094. 14. T.A. Brennan, "The Institute of Medicine Report on Medical ErrorsCould It Do Harm?" ^ Joumaio/fitaiicira: 542.no. 15 (2000): 1123-1125; andD.M. Studdert andXA. Bmman, "No-Fault Corrqsensa' don for Medical Injuries: The Prospect for Error Prewntion," Jourml oftk American Medical AisocMm 286, no. 2 (2001): 217-228. 15. G. Porto, "Disclosure of Medical Error: Facts and d]hcies" jourmlc^HmlAcare Risk Maru^ment 21, no. 4 (2001): 67-76: C. Vincent, Clinical Risk Management: Enhancing Patient Safety (London: BMJ Books, 2001), 476-478; and M.B. Kapp, "Legal Anxieties and Medical Mistakes," Joumaf o/General Internal Me^dm 12, no. 12 (1997): 787-788. 16. S. Kraman and G. Hamm, "Risk Management: Extreme Honest}' May Be the Best Policy," Amah oflntcrml Medicine 31. no. 12 (1999): 963-967 17 J. Tieman. "Enforcing a New Openness," Modem Healthcare 31,no. 26 (2001): 4-5; andj. Conway,"Q:eaiing a Culture of Safety: Challenge Your Mental Models 'Cause It Ain't Necessarily So!" (Speech at Anierican Society for Healthcare Risk Management conference, Boston, 29 October 2001). 18. D. Pietro et al., "Detecting and Reporting Medical Errors: Why the Dilemma?" Briti^MedicalJountd 320. no. ;^37 (2000): 794-7%. 19. A. Witman, D. Park, and S. Hardin, "How Do Patients Want Physicians to Handle Mistakes? A Survey oE Internal Medicine Patients in an Academic Setting," Archives oflnterml Medicine 156, no. 22 (1996): 25652669; W Le^'inson, "Physician-Patient Communication: The Relationship with Malpractice Claims among Primary Care Physicians and Surgeons," journal of the American Medical Association 277, no, 7 (1997): 553-559; and G.B. Hickson et al., "Factors That Prompted EamiHes to File Medical Malpracdce Claims following Perinatal Inj'uries,'' joumai of the American Medical Association 267, no. 10 (1992): 1359-1^3. 20. Studdert and Brennan, "No-Fault Compensation for Medical Injuries." 21. See, for example, U.S. House of Representatives, "Help Efficient, Accessible, Low Cost, Timely Health Care (HEALTH) Act of 2002; H.R. 4600,107th Cong,, 2d sess. (25 April 2002); and "Healthcare Ser\1ces Malpractice Act," H.B. ]8O2,186di Leg., Reg. Sess. (Pa. 2002), Public Law 154, No. 13, Sec. 5104. 22. M.L. Millenson, "Pushing the Profession: How the News Media Turned Patient Safety into a Priority." Quality and Safety in Health Care 11. no. 1 (2002): 57-63. 23. For example, if low frequency of disclosure (for instance, as a proportion of all harms) were correlated with low quality, the gap we identified between disclosures and actual adverse events may be even ^leatet, ^veri some evidence of relatively low quality of care in for-proiit hospitals. See E.J. Thomas et al., "Hospital Ownership and Preventable Ad\'erse E\'ents," journal of Gcncml Imenud Medicine 15, no. 4 (2000): 211-219; PJ. Devereaux et al., "A Systematic Review and Meta-Analysis of Studies Comparing Mortality Rates of Private For-Profit and Private Not-for-profit Hospitals," Canadian'MedicalAssociatimJmmal 166 no 11 (2002): 1399-1406; and Y.C. Shen. "The Effect of Hospital Ownership Choice on Patient Outcomes after Treaonent for Acute, Myocardial Infarction," ]mrmi t^Hedrfi Economics 21, no. 5 (2tX)2); 901--922. 24. Amori, "President's Message"; American Society for tfcalthcare Risk Management. ASHRM conference agenda. 2001; and M. Ott, "Key Considerations on Drafting a Policy on Disclosure of Unanticipated Outcomes," Journal offferitfecarcRisfeJVtoiagemcnt 21, no. 4 (2001): 27-31. 25. Porto. "The Risk Mans^er's Role"; and D. Finkelstein et aL, "When a Physician Harros a Patient by Medical Error: Ethical, Legd, and Risk Management Considerations," loumo! of Clinical Ethics 8, no 4 (1997)330-335. 26. E.J. Emanuel and L.L. Gorman. "Four Models of the Physician-Patient Relationship," Jbiima! c^Oic American Medical AsiOciaHm 267, no. 16 (1992): 2221-2226. 27. T. Gorman, "Physicians Fold under Malpractice Fee Burden," Los A J ^ Times, 4 March 2002; K. Hiindiey, "Prognosis for Trouble," St fttersk^g Times, 11 March 2002; M. Freedman, "The Tort Mess." Forbes. 13 May 2002; andJ.B. Treaster. "New York Doctors Eadng Bigjump in Insurance Rates." New York Times, 22 March 2002.

HEALTH AFFAIRS - Volume 22, Number 3