Effect of New York state's do-not-resuscitate legislation on in-hospital cardiopulmonary resuscitation practice

Effect of New York State’s Do-Not-Resuscitate Legislation on In-Hospital Cardiopulmonary Resuscitation Practice RUSSELL S. KAMER,M.D., EILEENM. DIECK,M.D., JOHNA. MCCLUNG,M.D., PATRICIA A. WHITE,M.D., STEVEN L. SIVAK,M.D. Valhalla, New York

E

PURPOSE: On April 1,1988, New York State enacted legislation governing the withholding of cardiopulmonary resuscitation (CPR). Suggestions that the mandated protocol for withholding CPR is too cumbersome and will result in an increase in CPR attempts led us to study the effect of the new law on in-hospital resuscitation practice. PATIENTS

AND

METHODS:

we

retrOSpe&Vely

ffective April 1, 1988, New York became the first state to legislate the conditions governing the withholding of cardiopulmonary resuscitation (CPR) [l]. The state legislature enacted these strict regulations in response to a grand jury’s finding that a New York hospital had devised a policy of secret “do-notresuscitate” (DNR) orders not recorded in the medical record and issued without consulting the patient or family [2]. In addition to addressing the problem of withholding CPR without consent, the legislation was meant to ensure the issuance of a DNR order when medically appropriate and consistent with the patient’s wishes [3]. Although the law is intended to benefit patients by protecting their autonomy, some have speculated that the law may actually harm patients. It has been suggested that the legislation may lead to more attempts at resuscitation because the mandated protocol for withholding CPR is perceived as too cumbersome [4]. This study examined whether New York State’s DNR law has changed in-hospital resuscitation practice. Have resuscitation attempts increased? Are patients more involved in the DNR decision? Has the documentation of the decision to withhold CPR improved?

re-

viewed the charts of 245 adult in-patients at a county teaching hospital who died during threemonth periods before and after the law took effect. RESULTS: There was a Stati&ica.Ry nonsignificant decline in the frequency of CPR attempts at the time of death, from 59 (50%) of 119 patients in 1987 to 57 (45%) of 126 patients in 1988. Use of explicit written “do-not-resuscitate” (DNR) orders increased significantly from 13 (22%) of 60 patients who died without CPR in 1987 to 64 (93%) of 69 patients in 1988. Patient and family involvement in decisions to withhold CPR was common before the law and did not change significantly. CONCLUSION: Although changing the way DNR decisions are documented, the legislation resulted in no significant change either in the frequency of CPR or in the degree to which patients are involved in these decisions.

Summary of the Law Every patient is presumed to consent to CPR, unless there is consent to the issuance of a DNR order. If the patient has decision-making capacity, the attending physician must obtain the patient’s consent in ;:;eyesence of two witnesses before issuing a DNR If the patient lacks decision-making capacity, two physicians must certify, in writing, their opinions concerning the cause and probable duration of the patient’s incapacity. Notice of this determination must be given to the patient and the surrogate decisionmaker (legal guardian or-next of kin). Then, the surrogate may consent to a DNR order only after the written determination by two physicians that: (1) the patient has a terminal condition; or (2) the patient is irreversibly comatose; or (3) resuscitation is medically futile; or (4) resuscitation would impose an extraordinary burden in light of the expected outcome. For acute-care hospitalizations, the DNR order must be reviewed every three days.

PATIENTS AND METHODS From the Section of General Internal Medicine, Department of Medicine, Westchester County Medical Center, and the Alfred E. Smrth institute on Human Values in Medical Ethics, New York Medical College, Valhalla, New York. This work was presented in part at the International Congress on Ethics and Medicme, Stockholm, Sweden, September 1989. Requests for reprints should be addressed to Russell S. Kamer, M.D., Department of Medicine, Westchester County Medical Center, Valhalla, New York 10595. Manuscript submitted August 3. 1989, and accepted in revised form December 6. 1989. I

108

February

1990

The American

Journal

of Medicine

Volume

This study was conducted at the acute-care hospital of the Westchester County Medical Center, a major teaching hospital of New York Medical College. This facility is a tertiary referral center and comprises 476 nonpsychiatric beds. Physician staffing is provided by full-time faculty, and part-time private physicians, in concert with house officers in training. On April 1, 1988, a resuscitation protocol identical to state law in 88

DO-NOT-RESUSCITATE

all of its details was instituted. Previously, there was no formal DNR policy. Two groups of patients were evaluated in this retrospective study. The first group included all adult inpatients who died during a three-month period (October 1,1987, to December 31,1987) prior to the implementation of the DNR law. This group was compared with adult inpatients who died during the same three months one year later, after the law took effect. Charts were reviewed for the following information: age; sex; service (medical or surgical); diagnosis; presence of a DNR order on the order sheet; documentation of DNR status in the progress notes; who (resident or attending physician) spoke to the patient or family about withholding CPR; and hospital course. After implementation of the law, a subsample of cases in which a surrogate decision-maker had consented to a DNR order on behalf of the patient was reviewed for attending physician’s certification of the nature of the patient’s lack of decision-making capacity, concurring physician’s certification, surrogate decision-maker’s written consent to a DNR order, and witnesses’ signatures. Notations such as “No heroic measures” and “Supportive care only” were interpreted as evidence of DNR status in the progress notes; however, only explicit directives to withhold CPR were counted as evidence of a DNR order on the order sheet. Deaths were classified as cancer-related if the patient had a diagnosis of cancer (e.g., a patient dying of sepsis during chemotherapy for leukemia would be classified as having a cancer-related death). The chi-square test with Yates’ correction was used for categorical data, and p CO.05 was considered significant. Confidence intervals were calculated by standard formulas.

LEGISLATION

/ KAMER ET AL

TABLE I Frequency of CPR at Time of Death before and after Legislation, by Diagnosis

Diagnosis

Number (%) of Resuscitation Attempts 1987 1988

Change, % (95% confidence interval)

Cancer Other All patients

21(37) 38 (60) 59 (50)

-9 (-28 to 7) -l (-18 to 16) -5(-17to8)

16(28) 41(59) 57 (45)

TABLE II Documentation of the Decision to Withhold CPR

DNR Status Written On Order sheet Progress note ororder sheet

Charts (%) 1987 1988

Change, 46 (95% confidence interval)

93 99

71 (59 to 83) 1 (-4 to 5)

z;

1 TABLE Ill Involvement of Patients and Families in the Decision Not to Resuscitate for Patients Who Died without a Resuscitation Attempt

Decision Discussed With Patient Family, not patient Patient orfamily

Patients (96) 1987 1988 16 2 97

;:

Change, % (95% confidence interval) 3 (-10 to 15) -2 (-15 to 10) 0 (-5 to 6)

RESULTS The medical records of 245 of 258 patients who died during the study periods were available for analysis. The study group included 132 (54%) men and 113 (46%) women. The mean age was 61.4 years (range: 22 to 98 years). Table I shows a nonsignificant decline in the use of CPR after the DNR law was implemented. Table II displays a substantial and significant (p
Thirty-six cases in which a family member consented to a DNR order after implementation of the law were further reviewed. The attending physician documented the nature of the patient’s lack of decisionmaking capacity in 33 (92%) of the charts, and a concurring physician’s opinion was documented in 22 (61%) of the charts. The patient’s next-of-kin gave written informed consent in the presence of two witnesses in 34 (94%) cases; in the other two (6%) cases, only one person witnessed the surrogate’s signature. Reasons given for the patient’s incapacity included deteriorating physiologic condition (67%), anoxic encephalopathy (14%), brain metastases (II%), cerebrovascular accident (6%), and senile dementia (3%). There was no significant change in the relative involvement of house staff and attending physicians in the discussion of DNR status with the patient or family. In 1987, only the house staff discussed DNR status in 16 cases (28%), only the attending physician in 36 (62%), and both in six (10%). In 1988, only the house staff discussed DNR status in 18 cases (27%), only the attending physician in 41 (61%), and both in eight (12%).

COMMENTS Our data suggest that, although improving documentation in the medical record, the enactment of DNR legislation in the State of New York appears to

February

1990

The American

Journal

of Medicine

Volume

88

109

DO-NOT-RESUSCITATE

LEGISLATION

/ KAMER

ET AL

have had little effect on the frequency of CPR and on the degree of patient or family involvement in the DNR decision at our institution. It is clear that the number of patients for whom DNR orders are explicitly written on the order sheet has increased rapidly and substantially. This can be construed as a benefit, given that it reduces any likelihood of miscommunication between physicians, nursing staff, and the patients or their surrogates. Notwithstanding, specific legislation may not be required to accomplish this goal. A similarly rapid, although not quite so pronounced, increase in order sheet documentation was observed by Quill et al [5] for elderly patients at another New York State facility solely in response to the institution of a comprehensive hospital DNR policy. A more gradual increase in the use of explicit DNR orders following the creation of a hospital protocol was noted earlier in this decade by Bedell and co-workers [6]. With regard to the issue of miscommunication, it is of particular note that two patients were resuscitated despite chart documentation suggesting DNR status, one before and one after the legislation. It is of further note that in each case, there was no formal DNR order extant. This suggests that explicit DNR orders are more likely to be complied with and do indeed represent an improved form of communication. Consultation with either the patient or family before withholding CPR was common before the law and did not change significantly afterwards. The Medical Society of the State of New York issued guidelines in 1982, six years before the legislation, directing physicians to obtain the consent of the patient or his or her family before issuing a DNR order [7]. Although lacking legal authority, these guidelines became the standard of care in the state. The newly legislated procedures are similar to the pre-existing medical society guidelines in requiring consent to a DNR order. As such, it is not surprising that the legislation has had little effect on the way in which this consent is obtained. In a similar fashion, the legislation appears to have had no meaningful impact on the patient’s opportunity to take an active role in CPR discussions. In only a minority of cases was the patient able to be involved in the decision to withhold resuscitation, both before and after the enactment of the law. Our finding of active patient involvement in the decision to forgo CPR in only 13% to 16% of cases studied is similar to the findings of previous investigators who found percentages ranging from 14% to 28% [6,8-111. There appears to have been little change in the common practice of deferring decisions regarding CPR status until later in the hospital course, at which time the patient’s deteriorating physiologic status often renders him or her incapable of making an informed judgment. Above and beyond a basic unwillingness to discuss the matter earlier, there may be other reasons for this relatively consistent observation. Patients with primary neurologic or cardiovascular disease may initially present with a favorable prognosis that militates against a DNR order. Only when serious hemodynamic or cerebrovascular instability ensues might these individuals become candidates for the withholding of resuscitation. This may account for the 67% of patients for whom “deteriorating physiologic condition” was identified as responsible for the patient’s lack of 110

February

1990

The American

Journal

of Medicine

Volume

capacity. Given that there is no provision for a DNR order that is limited only to certain physiologic conditions, it may be impossible to issue an order earlier in the hospital course. Regardless of the reason, the legislation does not appear to have engendered a genuine increase in patient autonomy in our patient population. Concern that the complexity of the statute would increase CPR attempts appears to be unfounded. In fact, our data revealed a statistically nonsignificant trend towards fewer CPR attempts overall. The trend was seen only in the subgroup of cancer deaths and may reflect a clinical recognition that this subpopulation may also be less likely to survive to hospital discharge following CPR [12]. This finding may be explained by the fact that doctors may be more willing to withhold CPR now that the legality of such decisions has been clarified. In this regard, it is important to realize that DNR legislation was considered necessary in New York in part because the state’s highest court limited the right of surrogates to refuse life-sustaining treatment on behalf of an incompetent patient, even if the patient is terminally ill [13]. Even after enactment of the law, the Court of Appeals has continued to limit the power of a surrogate to refuse treatment other than CPR [14]. Our results suggest a reasonably high rate of compliance with the details of the law. In cases in which the patient lacked decision-making capacity, the attending physician nearly always certified in writing an opinion as to the cause of the patient’s incapacity. In the majority of these charts, a second physician’s concurring opinion was documented; however, this was absent in a significant minority. With regard to this, it should be noted that, although the statute provides both civil and criminal immunity to those professionals who comply with it, there are no sanctions specified against any individual who does not. Under these circumstances, the statutory requirements are handled as a matter of hospital policy, rather than as state law, and noncompliance becomes a matter requiring auditing and correction by the institutional quality assurance department in a fashion virtually identical to other hospital policies. This study was limited by four primary factors. First, this investigation was conducted in an acutecare tertiary referral center, and its results cannot necessarily be extrapolated to other hospitals or nursing homes. Second, it is a retrospective evaluation that relies solely on the hospital chart for its data. A more detailed evaluation of the actual decision-making process involved in generating a DNR order was beyond the scope of this study. Third, the number of patients studied was limited. Our study lacks the power to detect small changes. Finally, this investigation focused only on mortalities and the time period immediately prior to these mortalities. No attempt was made to analyze the number of previous resuscitative efforts carried out before this time, nor was there any attempt made to evaluate the effect of the legislation on those patients electing DNR status who did not die. No formal DNR policy existed at our hospital before the enactment of the current legislation. As a result, the striking similarity of our data to that previously reported solely in response to the creation of a formal institutional policy suggests that the existence of an effective hospital policy may obviate the need for spe88

DO-NOT-RESUSCITATE

cific legislation in other states. A more general statute that simply mandates the creation of such a policy may achieve nearly identical results. Furthermore, strict adherence to current Joint Commission on Accreditation of Healthcare Organizations requirements for such policies may render even this limited legislative initiative redundant. In conclusion, the enactment of DNR legislation in New York State appears to have resulted in a substantial increase in the number of formal DNR orders appearing on the order sheet at our institution. It has not been associated with an increased number of resuscitative efforts; rather, a trend to the contrary has been observed. The frequency of discussing the order with either the patient or his or her surrogates has not changed significantly. As such, the law does not appear to increase patient autonomy so much as it validates the legitimacy of surrogate decision making. In view of these findings, this legislation may be of more import as a means of providing a mechanism for substituted judgment, rather than as a means of altering outcome. Further study of the impact of this legislation at other medical institutions in New York State will be of assistance in clarifying our observations.

LEGISLATION

/ KAMER

ET AL

REFERENCES 1. Article 29-B, Statute 413-A. The State of New York Public Health Law, effective April 1, 1988. 2. Sullivan R: Hospital’s data faulted in care of terminally ill. The New York Times 1984 March 21: 61. 3. New York State Task Force on Life and the Law: Do not resuscitate orders: the proposed legislation and report of the New York State Task Force on Life and the Law, 2nd ed. New York: New York State Task Force on Life and the Law, 1988; 1819. 4. Rosner F, Hotchkiss EJ: Must we always offer the optlon of CPR? The law I” New York (letter). JAMA 1988; 260: 3129. 5. Quill TE, Stankaitis JA. Krause CR: The effect of a community hospital resuscltation oolicv on elderlv oatients. NY State J Med 1986: 86: 622-625. 6. B$dell iE. Pelle D:Maher PL, Cleary PD: Do-not-resuscitate orders for critlcally ill patlents in the hospital: how are they used and what is their impact? JAMA 1986; 256: 233-237. 7. Sullivan R: Standards issued on resuscitation of dying patients. The New York Times 1982 Sea 19: 1. 8. Jonsson Pi, LcNamee M, Campion EW: The ‘do not resuscitate’ order: a profile of its changing use. Arch Intern Med 1988; 148: 2373-2375. 9. Evans AC. E&ody BA: The do-not-resuscitate order in teaching hospitals. JAMA 1985; 253: 2236-2239. 10. Zimmerman JE, Knaus WA, Sharpe SM, et&The use and implications of do not resuscitate orders in intensive care units. JAMA 1986; 255: 351-356. 11. Schwartz DA, Reilly P: The choice not to be resuscitated. J Am Genatr Sot 1986; 34: 807-811. 12. Taffet GE, Teasdale TA, Luchi RJ: In-hospital cardiopulmonary resuscitation. JAMA 1988; 260: 2069-2072. 13. Matter of Storar, 52 NYPd 363 (1981). 14. Matter of O’Connor, 72 NYPd 517 (1988).

February

1990

The American

Journal

of Medicine

Volume

88

111