(dissenting). I dissent and would reverse finding the challenged regulation unconstitutional. Conceding arguendo, the existence of a compelling State need to monitor what the city describes as "essentially a pilot program in legal abortions”, I find absolutely no valid justification for the inclusion of the patient’s name and address on the termination of pregnancy certificate.
In the wake of Roe v Wade (410 US 113) and Eisenstadt v Baird (405 US 438) it cannot be seriously doubted that the protection of the due process clause of the Fourteenth Amendment extends to matters involving whether or not to have children (see, also, Doe v Bolton, 410 US 179; Loving v Virginia, 388 US 1; Griswold v Connecticut, 381 US 479; Skinner v Oklahoma, 316 US 535). Nevertheless, the city argues that the regulation here is not within the ambit of the constitutional protection enunciated in Roe and Eisenstadt because those cases apply solely to the decision to abort and are inapposite with respect to mere requests for information needed for record keeping. The fallacy of this argument is that it overlooks the lingering sensitivity of the abortion issue and the potential for stigmatization of women availing themselves of abortions which are now perfectly legal. In my view the right to privacy which the Supreme Court extended to a woman’s decision to abort necessarily extends to and includes her right to guard her identity from a centralized abortion registry.
While it is true that the Fourteenth Amendment’s embodiment of the right to an abortion is not absolute and may be subject to State regulation in appropriate instances (Roe v Wade, 410 US, at pp 154-155, supra), any such regulation must still pass constitutional muster. This principle obtains whether the regulation directly controls abortions (Roe v Wade, supra; Friendship Med. Center v Chicago Bd. of Health, 505 F2d 1141, cert den 420 US 997) or whether it affects abortions indirectly by mandating disclosure of name and address for a central file. To be valid any such enactment *246must be precisely and narrowly drawn so as to express only the legitimate State interests at stake (see, e.g., Griswold v Connecticut, 381 US, at p 485, supra; Aptheker v Secretary of State, 378 US 500, 508; Cantwell v Connecticut, 310 US 296, 307-308; Eisenstadt v Baird, 405 US, at pp 463-464, supra).
The respondents attempt to justify the disclosure requirement by listing several objectives which the regulation purports to accomplish. Despite this list, the city fails to show that the identity statistic is essential. Those objectives relating to continuing treatment and emergencies (Nos. 1 & 7 in majority opn, at p 238) are patently unpersuasive. Communication and time limitations inherent in a system functioning via the mails, render the use of termination certificates unworkable. Moreover, the termination certificates are totally useless to a treating agency unless the patient’s name is ascertained from some independent source, like the patient. Consequently, these objectives are better accomplished by the use of standard hospital files which are presumably more detailed and therefore more useful.
As far as the objective concerning the policing of abortion clinic operations (No. 2 in majority opn, at p 238) it is my view that this may best be accomplished by on-site supervision and inspection.1 It is highly doubtful whether those operating in an illegal or unethical manner will broadcast their illicit activity in the termination certificates. Additionally, the presence of a name is not a fact which helps to determine whether the procedures used were proper. If any impropriety or irregularity did appear on a certificate, the department could easily take that certificate to the appropriate facility and without using a name find the corresponding file and conduct the necessary investigation.2 Lastly, the objective relating to family planning and counseling for those undergoing repeated abortions (No. 6 in majority opn, at p 238) is highly suspect in view of the Supreme Court’s unequivocal pronouncement that this area is constitutionally cognizable as within the right of *247privacy (Roe v Wade, 410 US 113, supra; Doe v Bolton, 410 US 179, supra; see, also, Relf v Weinberger, 372 F Supp 1196).
The only valid objective presented by the city deals with the compilation of public health data to provide an empirical basis for future evolution in the area of legalized abortion (Nos. 3, 4 & 5 in majority opn, at p 238). Recognizing the necessity and desirability of these statistics the determinative question becomes whether this objective may be achieved without disclosure of names and addresses in a centralized data bank. The majority notes that the regulation is designed to provide information for (p 244) "narrow, well-defined and laudable governmental ends”. This formulation skirts the main thrust of appellants’ attack, which is that too much information is required. Clearly, the power to promulgate a form seeking vital statistics does not include the power to probe into private matters. It is my view that the same result would be accomplished if the regulation required the termination certificates to be compiled according to number.
Under this approach the patient’s file would receive a coded number which would also be placed on the termination certificate for identification. Since the cards would contain all the pertinent information required (except the name and address) the Board of Health could then analyze and correlate the data in order to test any hypothesis it desired. Moreover, in the event the name of a particular patient were needed for purposes of follow-up, counseling or further inquiry, that file could easily be extracted by number from the hospital or clinic utilized. Thus, it is very difficult to accept the majority’s bald conclusion that the inclusion of names and addresses on the certificate offers (p 239) "the most practicable means for researchers to retrieve a patient’s hospital records”.
Nor may the majority draw support from Planned Parenthood of Cent. Mo. v Danforth (392 F Supp 1362 [three-Judge District Court], probable jurisdiction noted 423 US 819). Although the statute under consideration there professed the need for statistical data and insured confidentiality, it did not establish a centralized list of names. Instead, that statute required that all such records be maintained in the permanent files of the health facility in which the abortion was performed. This is a far cry from the massive, open-ended compilation of names created by New York’s regulation and is obviously a more palatable approach.
Similarly, the majority’s attempt to distinguish Roe v Ingra*248ham (403 F Supp 931 [three-Judge District Court]) misses the mark. The court there was not concerned with the lack of confidentiality safeguards, as the majority intimates. Rather it focused on the magnitude of the intrusion noting that the existence of a computerized central file is qualitatively different from a name on file in one of a thousand pharmacies. Thus, the three-Judge court declared unconstitutional a New York State statute mandating the establishment of a centralized file of those using certain drugs. After discussing the State’s interest in preventing illicit drug trafficking, the court held that the minimal use derived from the names did not justify the centralized listing of thousands of names. The court opined that the presence on the form of the name of the person receiving drugs adds nothing.
This reasoning is especially applicable in the instant case where the operating facility is required to maintain individual files detailing the case history, and the inclusion of names on the certificate is of doubtful, if any, value. Here, as in Roe v Ingraham (supra), though both enactments implement concededly legitimate State needs, the centralized list serves no purpose in deterring illicit activity, nor materially advances the compilation of necessary statistics.
Considering the seriousness of the issue and the magnitude of the task of controlling abortion practices, it seems painfully obvious that the city should proceed in a manner calculated to assure accurate data. The present method is counterproductive. Even disregarding the appellants’ argument that the instant regulation has a chilling effect, it is clear from the record that women often use an alias to circumvent this regulation. Clearly a system of coded numbers would preclude resort to such measures and insure abortion patients the anonymity to which they are entitled.
Nor do I believe this conclusion should be altered by the recent addition of confidentiality safeguard to the regulation. While this measure is indeed welcome it does not totally dispel the specter conjured up by the existence of a centralized list of names growing at a rate of 500,000 per year. More significantly, however, the confidentiality provision cannot and does not provide a justification for the inclusion of names. It merely makes the present system less offensive. The proper solution—the one mandated by sound constitutional principles —would be for the city to compile the statistics without the use of names and addresses. Such an alternative system could *249utilize coded numbers or any other approach which would limit to the slightest degree the infringement on the patient’s right to keep her intimate decisions personal.
Where, as here, the information is of an intensely private nature and the State’s need for names and addresses is nonexistent, the centralization of such data is unwarranted and unconstitutional. Accordingly, I would reverse the order of the Appellate Division.
. The respondents argue that the addresses are necessary in order to determine accurately whether certain areas of the city need more clinical facilities. This, of course, could be accomplished by requiring the inclusion of the patient’s home area. The department has no need for the patient’s specific street address.
. At one point the city contends that the patient’s identity is required because without it an investigation into malpractice or use of improper procedures would be thwarted by virtue of surgeon anonymity. The obvious solution is to require the inclusion of the surgeon’s name on the certificate.