MEMORANDUM AND ORDER
LeRoy Carhart (Carhart), a medical doctor, seeks a preliminary injunction regarding enforcement of Nebraska’s “partial-birth” abortion law. After an evidentiary hearing, we conclude that a preliminary injunction should be issued.
We agree with the first argument raised by Carhart: Nebraska’s ban on the “partial-birth” abortion procedure has the effect of subjecting Carhart’s patients to an appreciably greater risk of injury or death than would be the case if these women could rely on him to perform his variant of the banned procedure on nonviable fetuses when medically advisable. Such a ban, therefore, is an “undue burden” to women seeking abortions, and it violates the Due Process Clause of the Fourteenth Amendment.
We do not reach the merits of Carhart’s second argument that the statute is unconstitutionally vague. We note, however, that a similar vagueness argument was recently adopted in a persuasive opinion. Evans v. Kelley, No. 97-CV-71246-DT, slip op. at 48-67, 85-88 n. 38, — F.Supp. -,--- -, ----n. 38 [1997 WL 471906] (E.D.Mich. July 31, 1997) (granting preliminary and permanent injunction) (Evans). It is unnecessary to reach this question given our resolution of Carhart’s first argument, and “[a]s the U.S. Supreme Court has repeatedly cautioned, federal courts should avoid unnecessary and broad constitutional adjudication.” Id., slip op. at 85-86 n. 38, — F.Supp. at---n. 38 (citations omitted). We believe it particularly appropriate to avoid this issue until we have the benefit of a trial on the merits.1
Our reasons for the foregoing findings and conclusions are set forth in the following memorandum and order.2 We turn to those reasons now.
I. FINDINGS OF FACT
A. The Parties
1. Plaintiff LeRoy Carhart, M.D., practices medicine and surgery in Nebraska and *510performs abortions in Bellevue, Sarpy County, Nebraska. (Filing 1, Compl, at 3; Filing 9, Stenberg and Thomas Answer, at 2; Ex. 16, Carhart Curriculum Vitae, at 1, 5.)
2. Carhart received his Doctorate of Medicine in 1973; completed his internship at Malcolm Grow USAF Hospital at Andrews Air Force Base, Maryland, in 1974; and completed his general surgery residency at Hahnemann Medical College and Hospital in Philadelphia, Pennsylvania, and Atlantic City Medical Center in Atlantic City, New Jersey, in 1978. Carhart is a retired lieutenant colonel in the United States Air Force who served as chief of general surgery, chief of emergency medicine, and chairman of the department of surgery at Offutt Air Force Base in Nebraska from 1978 to 1985. As part of his duties at Offutt, Carhart supervised 20 to 25 other physicians, including obstetricians and gynecologists. (Tr. 193:25-194:5.) Carhart has been an assistant professor in the surgery departments of both Creighton University School of Medicine and the University of Nebraska Medical Center. (Ex. 16, Carhart Curriculum Vitae, at 2-4.) Since 1985 Carhart has operated a general medical practice with a specialized abortion facility.3 (Tr. 82:14-21.) He performs 800 abortions each year. (Tr. 83:3.) Carhart has never attempted to become certified by a medical specialty board and currently has no hospital privileges. (Tr. 139:2-25.) He is licensed to practice medicine in eight states. (Ex. 16, Carhart Curriculum Vitae, at 5.)
3. Defendant Don Stenberg is attorney general of the State of Nebraska, and defendant Deb Thomas is director of the Nebraska Department of Health and Human Services Regulation and Licensure. (Filing 1, Compl., at 3-4; Filing 9, Stenberg and Thomas Answer, at 2.)
4. Defendant Mike Munch is the elected county attorney for Sarpy County, Nebraska, and is responsible for the enforcement of criminal law within Sarpy County. (Filing 1, Compl., at 3-4; Filing 11, Munch Answer, at 1.)
B. Legislative Bill 23
5. On June 3, 1997, the Nebraska Unicameral passed Legislative Bill 23 (“LB 23”) with an emergency clause making it effective' upon the governor’s signature on June 9, 1997. (Ex. 6.)
6. Legislative Bill 23 prohibits “partial-birth abortions” in the State of Nebraska “unless such procedure is necessary to save the life of the mother whose life is endangered by a physical disorder, physical illness, or physical injury, including a life-endangering physical condition caused by or arising from the pregnancy itself.” LB 23 § 3(1).
7. Legislative Bill 23 defines “partial-birth abortion” as follows:
Partial-birth abortion means an abortion procedure in which the person performing the abortion partially delivers vaginally a living unborn child before killing the unborn child and completing the delivery. For purposes of this subdivision, the term partially delivers vaginally a living unborn child before killing the unborn child means deliberately and intentionally delivering into the vagina a living unborn child, or a substantial portion thereof, for the purpose of performing a procedure that the person performing such procedure knows will kill the unborn child and does kill the unborn child.
LB 23 § 2(9).
8. Legislative Bill 23 makes the “intentional and knowing performance of an unlawful partial-birth abortion” a Class III felony, as well as grounds for automatic suspension and revocation of an attending physician’s license to practice medicine in Nebraska. LB 23 § 3(2) & (4).
9. “Partial-birth abortion” is not a recognized medical term. (Tr. 88:18-89:6, Carhart Test.; Tr. 216:3-13, Hodgson Test.)
C. Abortion Procedures
10. Carhart performs abortions in a clinic setting from a gestational age of 3 weeks *511until fetal viability,4 with gestational age being measured from the first day of a woman’s last menstrual period, as verified by ultrasound. (Tr. 83:9-84:5; 141:20-22.) Of the 800 women on whom Carhart performed abortions in 1996, 200 were past their 14th week of pregnancy. (Tr. 83:1-3; 185:14-24.) As far as he knows, Carhart is the only abortion provider in Nebraska who performs elective abortions past 16 weeks’ gestation. (Tr. 132:10-18.)
11. If a woman wants an abortion after viability and the abortion is not medically indicated, Carhart refers the patient elsewhere. (Tr. 87:13-22.) If a patient comes to him for an abortion and “there is any concern of fetal viability,” Carhart does not use his own judgment to determine viability, but instead insists on a specific referral from the patient’s physician identifying fetal flaws, stating that the fetus is not viable, and stating that the patient needs an abortion. (Tr. 174:4-16.)
12. Carhart performs abortions on patients whose health, rather than life, would be preserved by having an abortion, such as those with severe renal disease, severe diabetes that has required hospitalization, and hyperemesis gravidarum, a condition characterized by constant vomiting throughout pregnancy such that the pregnant woman loses a good portion of her body weight. Carhart has also performed abortions on patients who indicated that if abortion had not been an option for them, they would have considered attempting a self-induced abortion or suicide. (Tr. 133:16-134:11.)
13. Carhart selects the abortion procedure he will use on various patients based on gestational age and other medical factors. (Tr. 84:6-12.)
14. The parties have stipulated to the admission of Exhibit 7, which is a portion of the American Medical Association’s (AMA’s) “Report of the Board of Trustees on Late-Term Abortion.” The board of trustees prepared and submitted the report to the AMA’s board of delegates in May, 1997, in response to the passage of a 1996 resolution by the delegates calling for the AMA to conduct a study of late-term pregnancy termination techniques. (Filing 13 ¶ 2; Tr. 326:7-327:16.) Hereinafter, Exhibit 7 shall be referred to as the “AMA report.”5
1. Suction Curettage or Vacuum Aspiration
15. The AMA report indicates that suction curettage, or vacuum aspiration, is the most common means of inducing abortion from the 6th through the 12th week of gestation. (Ex. 7, at 7:29-30.) The AMA report describes this procedure as follows:
Prior to the procedure a pelvic examination is done to determine the size and position of the uterus. A speculum is used to visualize the cervix, a local anesthetic such as a paracervical block is administered, and the cervix is then dilated using rigid dilators (e.g., the Pratt dilator). Osmotic dilators may be used prior to the procedure. Once the cervix is sufficiently dilated, a suction tube is inserted and rotated inside the uterus to loosen and remove the contents. The suction tube may be attached to a suction machine or syringe. A curette may be used to scrape the endometrium, thereby ensuring the removal of any remaining tissue. These procedures are typically performed on an outpatient basis.
(Ex. 7, at 7:31-37 (footnotes omitted).)
16. Carhart uses curettage6 with vacuum aspiration from approximately the 12th *512through the 15th week of gestation. Carhart stated that “[a]t about the 12th or 13th week, we usually use curettage with vacuum aspiration up through the 15th, 16th week, after that 16th week, generally, it’s a dilation and evacuation procedure.” (Tr. 84:20-23.) At another point, Carhart stated that the “14th to 15th week” was the latest he would use the curettage with vacuum aspiration procedure. (Tr. 93:12-14.) While Carhart attempts to use curettage with vacuum aspiration in the 14th and 15th weeks of gestation, “[m]ost of the time by the 15th week, it doesn’t work,” and he must “use a mechanical forcep to actually grasp the fetus and remove it.” (Tr. 93:14-17.)
17. Carhart uses ultrasound while conducting the vacuum aspiration procedure. Ultrasound is “an adaption of the sonar developed by the Navy in the 1930’s to find the different densities of substances, usually in comparison with a known substance.” (Tr. 93:22-25, Carhart Test.) Ultrasound is used to display a picture of fetal tissue and cartilage on a television screen. (Tr. 94:1-5, Car-hart Test.)
18. When the vacuum aspiration procedure is used, a fetus can come through the suction tube, or cannula, intact or dismembered. Carhart uses cannulas ranging from 5 to 16 millimeters. (Tr. 95:10-24; 96:2-8.) If the tube becomes clogged during this procedure, Carhart must remove the tube to “de-clog” it, at which time the uterus will expel its contents into the vaginal cavity. (Tr. 155:18-20.)
19. 'The fetus is not dead before Carhart begins the vacuum aspiration procedure, and the entire fetus comes through the cannula alive in many instances. (Tr. 96:15-18; 97:16-17.)
2. Dilation and Evacuation (D & E)
20. According to the AMA report, the most common procedure for inducing abortion early in the second trimester of pregnancy, or in the 13th through 15th weeks of gestation, is dilation and evacuation, or D & E. (Ex. 7, at 8:1-10.) The AMA report describes the D & E procedure at 13 through 15 weeks’ gestation as follows:
Ultrasonography is used prior to the procedure to confirm gestational age, because the underestimation of gestational age can have serious consequences during a D & E procedure. D & E is similar to vacuum aspiration except that the cervix must be dilated more widely because surgical instruments are used to remove larger pieces of tissue. Osmotic dilators7 are usually used. Intravenous fluids and an analgesic or sedative may be administered. A local anesthetic such as a paracervical block may be administered, dilating agents, if used, are removed, and instruments are inserted through the cervix into the uterus to remove fetal and placental tissue. Because fetal tissue is friable and easily broken, the fetus may not be removed intact. The walls of the uterus are scraped with a curette to ensure that no tissue remains.8 In pregnancies beyond 14 weeks, oxytocin is given intravenously to stimulate the uterus to contract and shrink.
(Ex. 7, at 8:10-19 (footnotes omitted).)
21.According to the AMA report, the D & E procedure is also used from 16 to 24 weeks’ gestation, with the following variations:
Dilation and evacuation procedures performed in the mid-to late-second-trimester involve the preoperative use of laminaria or osmotic dilators (rather than surgical dilators) which are inserted in the endoeervical canal in order to dilate the cervix. The procedure is usually performed under local anesthesia, using sedation and para-cervical block. Intracervical vasopression is often used to minimize bleeding, and high dose oxytocin is administered intravenously prior to the procedure. Fetal tis*513sue is extracted through the use of surgical instruments, followed by extraction of placental tissue and subsequent curettage. Because the fetus is larger at this stage of gestation (particularly the head), and because bones are more rigid, dismemberment or other destructive procedures are more likely to be required than at earlier gestational ages to remove fetal and placental tissue. Some physicians use intrafetal potassium chloride or digoxin to induce fetal demise prior to a late D & E (after 20 weeks), to facilitate evacuation.
(Ex. 7, at 8:28-38 (footnotes omitted).)
22. Carhart uses the D & E procedure “after that 16th week” of gestation, combined with prostaglandin to aid in cervical dilation and other medication to cause the uterus to contract. (Tr. 84:24-85:16.)
23. Carhart’s method of mechanically removing the fetus from the uterus during D & E involves using ultrasound in order to observe the fetus and the surrounding area; using vacuum aspiration or forceps to rupture the membranes; and then grasping a portion of the fetus in order to bring it out of the uterus with long-handled forceps. (Tr. 100:1-14.) After the membranes are ruptured and the uterus begins to contract, a fetal extremity will often prolapse through the cervical os such that Carhart must dismember the extremity. He then attempts to bring the feet or skull down, dismembering the remaining arm or other extremities in the process. (Tr. 110:6-9; 118:9-12.)
24. When Carhart performs a D & E, he inserts an instrument inside the uterus, grabs a portion of the fetus, pulls it through the cervical os, and dismembers various fetal parts by the traction created between the instrument and the cervical os. (Tr. 116:2-10.) The tearing of fetal parts from the fetal body is accomplished by means of traction at the cervical os.9
25. Carhart’s description of this procedure is consistent with that of Dr. Jane Hodgson, founding fellow of the American College of Obstetrics and Gynecology, past president of the Minnesota Ob/Gyn Society, and author of 50 to 100 published articles on abortion. (Tr. 196:16-24; 197:11-13; Ex. 14, Hodgson Curriculum Vitae, at 2.)10
26. Through the 19th week of gestation, ultrasound confirms, by indicating a fetal heartbeat, that the fetus is “invariably” alive when Carhart performs a D & E, and via ultrasound Carhart has observed fetal heart activity with “extensive parts of the fetus removed.” (Tr. 100:15-22; 109:4-23; 110:12-14; 119:2-9; 195:9-11.) Dr. Hodgson also testified that the fetus may still have a heartbeat while extremities are being removed. (Tr. 211:4-7.)
3. Intact Dilation and Evacuation (Intact D & E or D & X)
a. The Procedure
27. In an effort to minimize perforation of the uterus or cervix by instruments used during a D & E or from piercing caused by fetal parts, some physicians use “a form of D & E that has been referred to in the popular press as intact dilation and extraction (D & X).” (Ex. 7, at 8:40-42.)
28. For any abortion “over 15 weeks[’]” gestation, Carhart intends and prefers to remove the fetus intact by using the intact D & E procedure, although it is “usually the 19th- and 18th- and early 20-week fetuses that come out intact.” (Tr. 100:9-10; 101:3-5; 184:1-23.)
*51429. Citing the American College of Obstetricians and Gynecologists’ January, 1997, statement on intact dilation and extraction, the AMA report describes the intact D & X as “deliberate dilation of the cervix, usually over a sequence of days; instrumental conversion of the fetus to a footling breech; breech extraction of the body excepting the head; and partial evacuation of the intracranial contents of a living fetus to effect vaginal delivery of a dead but otherwise intact fetus.” (Ex. 7, at 8:42-46.)
30. In contrast to the AMA’s description of the intact D & E or D & X procedure, Carhart does not perform instrumental conversion of the fetus to a footling breech, but removes the fetus headfirst or feet first, depending on how the fetus is positioned. (Tr. 102:19-24; 111:3-6; 156:18-157:19.) Carhart prefers the feet-first presentation because less dilation is required and “that’s the absolute safest scenario.” (Tr. 192:18-193:2.)
31. When Carhart performs this procedure, he drains the amniotic fluid before beginning the evacuation procedure in order to avoid amniotic fluid embolus, which he views as a serious and common cause of maternal death or complications. (Tr. 102:1-6.) If possible, Carhart then attempts to grasp and divide the umbilical cord of the fetus, which is the structure that transports arterial and venous blood between the fetus and the placenta, giving the fetus its only source of oxygen. If he divides the cord, the fetus will usually die within 6 to 10 minutes. (Tr. 111:11-20; 189:20-24.) There are instances in which Carhart cannot divide the cord because he is unable to reach it due to fetal position or spontaneous protrusion of a fetal part through the cervical os, preventing access to the cord. (Tr. 111:21-25; 190:17-25.)
32. If Carhart succeeds in dividing the umbilical cord, he does not wait for fetal death to occur before continuing the procedure because once the membranes have been ruptured and drugs administered to induce contractions, each minute of delay causes maternal blood loss. (Tr. 190:2-16.)
33. When the fetus is presented feet first, Carhart, using forceps, pulls the feet of the living fetus from the uterus into the vaginal cavity and then pulls the remainder of the fetus, except the head, into the vaginal cavity to a point where the base of the fetal skull is lodged in the uterine side of the cervical canal. (Tr. 110:19-22; 112:2-12.) At that point, the size of the head will not permit him to pull it through the cervical canal into the vaginal cavity. To decompress the fetal skull and evacuate the contents in order to pull it through the cervical canal, Carhart uses an instrument to either tear or perforate the skull to allow insertion of a cannula and removal of the cranial contents. Sometimes he will crush the skull rather than pierce it in order to reduce the size of the skull. (Tr. 104:2-6; 112:13-16.) Brain death occurs sometime during this two- to three-second reduction procedure, but fetal heart function may continue for several seconds or minutes after the fetus’s skull is decompressed. (Tr. 112:17-113:7.)
34. While he intends to remove the fetus intact for any abortion performed past 15 weeks’ gestation, only about 5 or 10 percent of the fetuses Carhart aborts are delivered totally intact due to softness of the fetal tissue such that it is easily fragmented. Car-hart normally cannot perform this procedure before the 16th week of gestation because the fetal body parts tear apart during the process. (Tr. 115:9-18; 184:10-19.)
35. Some patients have requested that Carhart perform an intact D & E for personal reasons, and some physicians have asked to have the fetus as intact as possible for genetic study when the entirety of fetal deformities is unknown. (Tr. 123:9-25.)
b. Fetal Death
36. After the 20th week of gestation, Car-hart attempts to induce fetal death 48 to 72 hours before beginning the abortion procedure with an ultrasound-guided intracardiac fetal injection of digoxin and lidocaine, both of which reduce and stop cardiac activity. (Tr. 119:10-25.) Carhart attempts to inject the drugs through the mid-line of the maternal abdomen where there are fewer blood vessels and he is less likely to encounter colon or small bowel contents. (Tr. 186:14-23.) Carhart attempts to inject the fetal thoracic cavity or heart, whereupon fetal *515death will occur within 15 to 20 minutes, but sometimes is able to inject only the amniotie sac, causing fetal death approximately 24 hours later. (Tr. 187:1-6.)
37. Carhart attempts to induce fetal death in this manner to achieve softness and compression of the fetal tissue and skull and to provide mental comfort to his patients. (Tr. 120:2-19.)
38. Carhart does not attempt to induce fetal demise in this manner during the 16- to 20-week time frame because the waiting time between the injection and performance of the procedure is only 12 to 24 hours, and not much fetal tissue change occurs; Carhart finds that many of his patients in the earlier stages of pregnancy are more apprehensive of the fetal injection; and in the earlier stages of pregnancy, the uterus is smaller and the risks of the needle penetrating the bowel and of missing the fetus and injecting the medication into maternal circulation are greater. (Tr. 120:20-122:3; 188:2-11.) Furthermore, Carhart sees patients for whom digoxin and lidocaine are medically contraindicated during any part of their pregnancy, such as those who have seizure disorders,11 heart disease, or who are already taking either medication such that an injection would exceed the maximum recommended dosage. (Tr. 122:20-123:4.)
39. Dr. Hodgson testified that in all forms of abortion, the point at which fetal demise occurs is “extremely variable.” (Tr. 217:17-18.) In her opinion, lack of fetal heartbeat is the best available measure for determining fetal demise, and fetal death by that measurement can be quite protracted. (Tr. 219:16-25.)
C. Benefits of the Intact D & E or D & X
40. The AMA report states that “[t]his procedure may minimize trauma to the woman’s uterus, cervix, and other vital organs. Intact D & X may be preferred by some physicians, particularly when the fetus has been diagnosed with hydrocephaly or other anomalies incompatible with life outside the womb.” (Ex. 7, at 8:46-49.)
41. Carhart’s intent to remove the fetus intact for any abortion performed past 15 weeks’ gestation is aimed at reducing the chances of maternal complications or death. (Tr. 100:9-10; 101:7-102:6; 124:16-125:2.) Intact removal of the fetus lowers maternal complications by preventing sharp fragments, such as pieces of long bone or skull fragments, from passing through the cervical os without some kind of covering or protection. When the fetus is removed intact, its bones are covered by fetal tissue, causing less trauma to the cervix. (Tr. 101:3-16; 131:1-7.)
42. Carhart also stated that intact removal of the fetus minimizes the risk of damage to maternal structures from repeated use of instrumentation in the uterine cavity. (Tr. 107:18-108:13; 131:1-23.) The more times Carhart must enter the uterus with an instrument, the more the complication rate multiplies. (Tr. 179:13-22.) The intact D & E or D & X procedure involves fewer insertions of forceps or other foreign objects into the uterus than a D & E resulting in dismemberment of the fetus. (Tr. 179:23-180:2.)
43. Performing the intact D & E or D & X procedure also allows a more accurate assessment of whether the uterine cavity has been emptied. Fetal and placental debris remaining in the uterus — as is possible with a D & E involving dismemberment — can cause infection, greater bleeding, and risk of absorption of the fetal tissue into the maternal bloodstream, as explained in more detail below. (Tr. 165:4-10; 183:4-24.)
44. Dr. Hodgson described leaving fetal parts in the uterus as a potentially “horrible complication” that can cause infection and often results in perforation of the uterine wall by bony splinters. (Tr. 211:19-24.)
45. Carhart’s method of intact removal of the fetus and evacuation of the contents of the fetus’s brain when it is reachable through the cervical os directly outside the uterus *516also helps prevent “disseminated intravascular coagulopathy” (DIC), that is, the absorption into the mother’s bloodstream of fetal brain, skin, and blood tissue through the blood sinuses or cavities in the uterine wall, thereby causing the mother’s own coagulation factors to stop working. According to Carhart, DIC is another cause of maternal death or complications, with the risk of such a complication being less than 1 in 1,000. (Tr. 101:16-25; 102:7-14; 158:13-159:14.) Compression of the fetal skull also enables Carhart to obtain as little cervical dilation as possible in order to reduce other maternal complications, such as incompetent cervix, at a later date. (Tr. 103:2-17.)
46. In Carhart’s opinion, performing an intact D & E is much safer than performing a D & E that does not result in removal of an intact fetus. (Tr. 108:14-18.) Although she has not intentionally12 performed an intact D & E, Dr. Hodgson believes the procedure is a technological advance that has received favorable reports from those who are performing the procedure. (Tr. 212:8-22.) Dr. Hodgson believes the D & X procedure is “an advance in technology” because by removing the fetus intact there is “less instrument manipulation,” which means, “of course, the higher your safety.” (Tr. 212:4-22.)
d. Number of Intact D & Es or D & Xs Performed by Carhart in Which Fetal Life Could Not be Terminated Before Delivery
47. In 1996, Carhart had 180 to 190 patients with a live fetus in útero for whom it was medically inadvisable to inject digoxin and lidoeaine to terminate the life of the fetus and on whom he began what he intended to be an intact D & E. (Tr. 189:1-12.) Of those 180 or 190 patients, Carhart removed approximately 20 fetuses intact, and 10 of those 20 presented themselves feet first. (Tr. 191:10-22.)
4. The Haskell D & X
48. In Women’s Medical Professional Corp. v. Voinovich, 911 F.Supp. 1051 (S.D.Ohio 1995), the court considered the constitutionality of an Ohio law banning the use of the so-called “Dilation and Extraction” (D & X) procedure. One of the plaintiffs, Dr. Martin Haskell, used a method of abortion from 20 to 24 weeks’ gestation characterized by various parties as the D & X procedure13 and performed as follows:
On the first and second days of the procedure, Dr. Haskell inserts dilators into the patient’s cervix. On the third day, the dilators are removed and the patient’s membranes are ruptured. Then, with the guidance of ultrasound, Haskell inserts forceps into the uterus, grasps a lower extremity, and pulls it into the vagina. With his fingers, Haskell then delivers the other lower extremity, the torso, shoulders, and the upper extremities. The skull, which is too big to be delivered, lodges in the internal cervical os. Haskell uses his fingers to push the anterior cervical lip out of the way, then presses a pair of scissors against the base of the fetal skull. He then forces the scissors into the base of the skull, spreads them to enlarge the opening, removes the scissors, inserts a suction catheter, and evacuates the skull contents. With the head decompressed, he then removes the fetus completely from the patient.
Id. at 1066 (footnotes omitted). Dr. Haskell routinely cuts the umbilical cord before penetrating the skull with scissors. Id. at 1066 n. 17.
5. Labor Induction
49. Labor induction, an alternative to D & E and what has been described above as the intact D & E or D & X, may also be used to induce abortion during the 16th to 24-th week of gestation. Labor may be induced by *517use of hypertonic solutions such as urea or saline, or prostaglandin. (Ex. 7, at 9:1-4.)
50. The use of saline to induce labor requires insertion of a needle through the abdomen and injection of the amniotic sac with a concentrated salt solution, which causes fetal demise and induces uterine contractions. Over a period of several hours, the uterine contractions cause dilation of the cervix and expulsion of the contents of the uterus. (Ex. 7, at 9:5-8.)
51. Urea is a nitrogen-based solution that causes fetal demise when injected into the amniotic sac and is typically followed by administration of prostaglandin to induce uterine contractions which will expel the contents of the uterus. (Ex. 7, at 9:8-12.)
52. Carhart does not perform saline or prostaglandin induction, and he does not perform inductions at all during the second trimester of pregnancy because he believes induction is medically contraindicated as a method of abortion. (Tr. 86:2-7; 125:3-126:1; 126:18-19.) Carhart believes induction during the second trimester is too uncontrolled — that is, the procedure can take over a week to complete; women have reactions to the drugs used during the procedure; and a segment of the population cannot undergo induction because of medical conditions such as hypertension, heart disease, or diabetes. (Tr, 125:6-126:1.) Carhart is aware of other physicians in Nebraska who perform abortions by induction, but only for maternally or fetally indicated abortions. (Tr. 126:2-4; 132:1-9.)
53. According to Stanley K. Henshaw, deputy director of research at the Alan Guttmacher Institute (AGI) in New York, statistical studies with which he is familiar indicate that abortions by induction are performed primarily in a hospital setting.14 (Tr. 74:20-25; 76:20-22.) The Centers for Disease Control and Prevention (CDC) recognizes the accuracy of the data AGI collects. (Tr. 46:8-21.)
6. Hysterotomy and Hysterectomy
54. Other abortion procedures available, but not routinely used, during 16 to 24 weeks’ gestation are hysterectomy and hysterotomy. According to the AMA report, “maternal mortality and morbidity associated with these procedures are significantly greater than those associated with other procedures used to induce abortion.” (Ex. 7, at 9:14-16.)
55. Hysterotomy is major surgery and must be performed in a hospital setting. General anesthesia or anesthesia administered by epidural or spinal injection is necessary. The procedure consists of surgical delivery of the fetus through an incision in the abdomen and uterine wall, after which the fetus is removed, the umbilical cord cut, and the placenta removed. (Ex. 7, at 9:16-20.)
56. Hysterectomy “is appropriate in cases in which there is a preexisting pathology, such as large leiomyomas or carcinoma in situ of the cervix.” (Ex. 7, at 9:20-22.)
57. Carhart does not perform hysterotomy or hysterectomy as methods of abortion, but he is aware of other physicians in Nebraska who perform abortions by these methods when an abortion is maternally or fetally indicated. (Tr. 86:2-9; 127:17-23; 132:1-9.)
D. Dr. Riegel
58. Defendants presented Dr. Christopher Riegel, an obstetrician, gynecologist, and infertility specialist from Dallas, Texas, as an expert witness to challenge the testimony presented by Drs. Carhart and Hodgson, discussed in detail above. Dr. Riegel received his medical degree from the University of Texas in 1987 and completed a one-year internship in pediatrics at Children’s Medical Center in Dallas in 1988 and a four-year internship and residency in obstetrics and gynecology at Parkland Hospital at the *518University of Texas before entering private practice in 1992. Dr. Riegel became board certified in obstetrics and gynecology in 1995. (Tr. 232:13-234:9.)
59. Dr. Riegel does not perform abortions due to moral objections and claims to be familiar with abortion procedures, risks, complications, and contraindications only from reading technical bulletins and textbooks and by becoming, during his service as chief resident in labor and delivery, “acquainted with what goes on” in medically indicated induction abortions of fetuses having malformations incompatible with life. His experience consists of attending at abortions that were already in progress when he arrived to complete the procedure. (Tr. 236:7-8; 238:12-239:8; 274:15-20; 275:16-22; 277:4-17.) Dr. Riegel has never observed a D & E abortion because he has “chose[n] not to be associated with them.” (Tr. 276:11-20; 276:23-277:3.) Dr. Riegel testified that the intact D & E or D & X procedure is not medically recognized and in fact “does not exist.” (Tr. 295:17-297:25.)
60. Dr. Riegel testified it is generally medically accepted that a fetus is dead if its heart stops beating (Tr. 236:14-20); injection of potassium chloride or digoxin into the heart of a fetus to cause fetal demise is generally done at 19 weeks’ gestation or more (Tr. 244:8-245:4); in the hands of a skilled operator, the risk of perforating the maternal bowel or injecting the mother with digoxin or potassium chloride during such a procedure is “inconsequential” (Tr. 245:9-22); the risk of complications for a mother with a preexisting seizure or heart disorder from using such an injection to cause fetal demise is nonexistent, “rare,” “low,” or “minuscule” (Tr. 247:22-249:22); there is no maternal medical advantage to injecting a substance into a fetus while it is in the uterus in order to kill the fetus before removing it (Tr. 250:3-22); there is no maternal medical advantage to partially delivering the fetus alive, killing it, and then completing delivery, with the latter scenario involving the “blind” use of a sharp instrument in the vagina accompanied by risk of damage to the urethra, bladder, vaginal wall, cervix, and uterus (Tr. 235:13-18; 256:6-257:24; 292:4-5); no type of fetal tissue or amniotic fluid should enter the mother’s bloodstream (Tr. 259:4-9); beginning at 16 weeks’ gestation, Dr. Riegel uses induction instead of D & E to remove dead fetuses from the uterus, and if the patient has heart disease, diabetes, renal disease, or a prior Cesarean section, he refers her to a high-risk obstetrician (Tr. 264:17-19; 266:1-24; 272:13-15; 293:15-23); since 1988 Dr. Riegel has delivered by induction 10 to 20 demised fetuses at 16 weeks’ gestation or later (Tr. 303:24-304:5); and in his geographical area of practice, it is generally accepted that amniocentesis at 13 to 18 weeks’ gestation is performed only by high-risk obstetricians (Tr. 279:24-25; 284:25-285:5).
61. With regard to induction by prostaglandin injection, a process which requires hospitalization, Dr. Riegel described the side effects of nausea, vomiting, diarrhea, and fever that accompany the injections the patient receives every three hours and the 8- to 36-hour time frame the induction procedure requires. (Tr. 306:5-307:23.)
E. Credibility of Witnesses Giving Medical Testimony
62. Having observed the witnesses’ demeanor and listened to their testimony, the court finds the testimony of Drs. Carhart and Hodgson to be credible based on their extensive training, experience, and knowledge of abortion procedures, whereas Dr. Riegel’s testimony regarding abortion procedures is generally not credible. Putting aside each witness’s views about the morality of abortion, the court’s judgment of the strength of the witnesses’ medical testimony is based solely on general indicia of credibility such as demeanor, candor, and each witness’s training, experience, and knowledge about medicine in general and abortion in particular.
63. Dr. Riegel has never observed or performed a D & E. He has only completed medically indicated induction abortions of fetuses having malformations incompatible with life in situations where the abortion was already in progress when he arrived on the scene. This limited experience occurred when he was a medical resident. Furthermore, as the AMA report proves, (Ex. 7, at 8), Dr. Riegel was poorly informed regarding use of the intact D & E or D & X procedure *519when he stated that the procedure — which is, after all, the focus of this case — simply “does not exist.” (Tr. 297:22-25.)
F. Statistical Risks and Frequency Associated With Abortion Procedures
64. Overall maternal mortality rates for the D & E, labor induction, and hysterectomy/hysterotomy methods of abortion at 13 weeks’ gestation or later are 51.6 per 100,000 abortions for hysterectomy/hysterotomy, 7.1 per 100,000 abortions for labor induction, and 3.7 per 100,000 abortions for D & E. Mortality rates resulting from labor induction and D & E performed at 16 to 20 weeks are 7.9 and 6.5 respectively, and 10.3 and 11.9 respectively at 21 weeks or more. (Ex. 7, at 10, tbl. 4: Maternal Mortality Rates for Induced Abortion Procedures at 13 Weeks’ Gestation or Later, U.S., 1974-1987.) The AMA report summarizes these statistics as follows: “Maternal mortality rates, overall, are higher for labor induction than D & E (7.1 and 3.7, respectively), but mortality rates resulting from labor induction and D & E are comparable for induced abortions performed at 21 weeks or more (11.9 and 10.3).” (Ex. 7, at 9:39-41; 11:28-30.)15 Dr. Hodgson stated that induction beyond weeks 18 to 20 “can be made as safe” as the D & E procedure resulting in dismemberment. (Tr. 212:1-3.)
65. While concluding that more systematic research is needed on complication rates associated with various abortion procedures performed at 13 weeks’ gestation and beyond, the AMA report cited statistics from the “best available national data on complications” collected during the 1970s by the Joint Program for the Study of Abortion, sponsored by the Population Council and the CDC. According to the report, the complication rate associated with vacuum aspiration was 2 per 1000 procedures; D & E had a complication rate of 7 per 1000 procedures; labor induced by saline or prostaglandin injection had rates of 21 and 25 per 1000 procedures respectively; and abortion methods involving major surgery had the highest rate of complications. (Ex. 7, at 10:20-26; 10:43-48; 11:30-31.)
66. Forcing a woman to wait until post-20-weeks’ gestation so that Carhart may be able to achieve fetal demise prior to performing an abortion drastically increases the risk of complications. According to the studies with which Carhart is familiar, abortion at the earliest stage is 35 times safer than childbirth as far as complications are concerned. By the 20th week, the risk of complications from abortion approximately equals that of childbirth. (Tr. 135:18-136:9, Carhart Test.; Tr. 59:9-10, Henshaw Test.)
67. The complication rate from abortion increases by about 20 percent for each week of gestation past eight weeks. (Tr. 59:11-14, Henshaw Test.)
68. According to his analysis of data from the Nebraska Department of Health and Human Services, Dr. Henshaw opined that D & E is the prime method of abortion performed in Nebraska after the first trimester of pregnancy. Of the 5,214 reported abortions in Nebraska in 1996, 5,161 were performed by suction curettage, the prime method for first trimester abortions; 550 were performed by “sharp curettage” which were “most likely D & E’s”; no inductions were performed; and 49 abortions recorded as “other” were most likely drug-induced abortions prior to eight weeks’ gestation. (Tr. 49:4-50:9; Ex. 9, Neb. 1996 Statistical Report of Abortions at 2, 4 (tbls.l, 5).)
69. Nationwide approximately 86 percent of abortions past 15 weeks are performed by curettage, or D & E, while induction is performed in less than 10 percent of abortions past 15 weeks. (Tr. 50:15-51:9;' Ex. 17, 1992 Abortion Surveillance Report, tbl. 16, reported legal abortions, by weeks of gestation and type of procedure — U.S., 1992 (Centers for Disease Control 1992).)
70. The number of curettage, or D & E, procedures reported in Nebraska and across the nation also includes the intact D & E or *520D & X,16 so there is “no way to quantify how many of the D & X abortions are done based on the statistics collected by the State [of Nebraska] or the CDC.” (Tr. 52:1-5, Henshaw Test.)
71. According to a study by Dr. Henshaw, only 25 abortions, or .4 percent of the total, were performed in Nebraska hospitals in 1992, and only three Nebraska hospitals reported performing abortions that year. Two of the three indicated a referral must be made by an attending physician, and one hospital did not answer the question, leading Dr. Henshaw to conclude that a woman cannot arrange an abortion directly with a Nebraska hospital. Although no such information was requested in the 1992 survey, one Nebraska hospital reported that abortions were permitted only for therapeutic reasons. (Tr. 55:10-56:19.)
G. Fear of Prosecution
72. Carhart fears prosecution under LB 23 because the Nebraska Attorney General’s Office and the State Department of Health are both aware of, and have investigated, his performance of abortions in the past. (Tr. 136:10-137:22; Ex. 20, Letter from Att’y Gen.’s Office to Carhart of 11/22/95; Ex. 21, Letter from State Dep’t of Health investigator to Carhart att’y of 10/9/96.)17
II. CONCLUSIONS OF LAW
Carhart asserts for himself and his female patients that Nebraska’s ban on “partial-birth abortions” is invalid under the Due Process Clause of the Fourteenth Amendment because it prevents Carhart’s patients from choosing, with his advice, a safe and desired method of terminating a pregnancy before viability.
To evaluate this argument, we must first address issues of jurisdiction, ripeness, and standing. Next we must clarify which abortion procedures Carhart does and does not perform. We must also decide whether Car-hart’s challenge to Nebraska’s law is “facial” or “as applied.” Then we must set forth the standards governing the granting of preliminary injunctions. Finally, we must apply the preliminary-injunction standards to the facts presented at the evidentiary hearing. We turn now to these tasks.
A. Jurisdiction, Ripeness, and Standing
We have jurisdiction because Carhart attacks the constitutionality of a state statute under the United States Constitution. As a result, federal-question jurisdiction exists under 28 U.S.C. § 1331. See, e.g., Women’s Med. Prof'l Corp. v. Voinovich, 911 F.Supp. 1051, 1058 (S.D.Ohio 1995), appeal argued, Nos. 96-3157 & 96-3159 (6th Cir. May 2, 1997) (district court had federal-question jurisdiction over action brought by doctor which challenged constitutionality of state law prohibiting “D & X” abortions) (WMPC).
In addition, this case is ripe because Carhart faces a significant risk of criminal prosecution and license revocation regarding at least 10 to 20 “D & X” procedures performed annually on nonviable fetuses. Moreover, Carhart tries to do the banned procedure for each of his patients between the 16th and 20th week of pregnancy, and he performs approximately 190 abortions during that time. (Tr. 185-186, 189.) The ban therefore has a direct and immediate impact upon Carhart and about 190 of his patients.
The State of Nebraska has admitted that 10 to 20 abortions performed by Carhart in 1996 clearly would have been prohibited by Nebraska’s law. (Tr. 366.) Moreover, Carhart’s abortion practices have been the subject of previous investigation by the defendants. Consequently, Carhart may sue for pre-enforcement review of the law. Id. (citing Doe v. Bolton, 410 U.S. 179, 188, 93 S.Ct. 739, 745-46, 35 L.Ed.2d 201 (1973)).
Finally, Carhart has the necessary standing to raise both his own rights and the rights of his patients. Id. To the extent that Nebraska’s ban prohibits him from practicing medicine in a safe and effective manner be*521cause it requires him to subject his patients to increased medical risk, Carhart has a strong personal stake in the argument. Moreover, Carhart has third-party standing to assert the interests of his patients because (1) a unique fiduciary-like relationship exists between doctor and patient; and (2) the pregnant women who are the doctor’s patients have significant obstacles to bringing suit on them own, such as a desire for privacy and the likelihood that their claims would be mooted by the time-sensitive nature of pregnancy and abortion. Id. (citing Planned Parenthood v. Casey, 505 U.S. 833, 112 S.Ct. 2791, 120 L.Ed.2d 674 (1992); Singleton v. Wulff, 428 U.S. 106, 96 S.Ct. 2868, 49 L.Ed.2d 826 (1976)). See also Evans, No. 97-CV-71246-DT, slip op. at 44, — F.Supp. at - (“[T]he Supreme Court ... [has] explicitly held that doctors who perform abortions are entitled to third-party standing to assert the rights of women seeking abortions.”) (citations omitted).
B. What Carhart Does and Does Not Do
Carhart does not intentionally perform abortions on viable fetuses.18 On the contrary, when he has a concern about viability Carhart requires an independent medical assessment to decide whether the fetus is viable. “Viability” occurs when “there is a reasonable likelihood of the fetus’ sustained survival outside the womb, with or without artificial support.” Colautti v. Franklin, 439 U.S. 379, 388, 99 S.Ct. 675, 682, 58 L.Ed.2d 596 (1979). See also Casey, 505 U.S. at 870, 112 S.Ct. at 2816-17 (viability is “the time at which there is a realistic possibility of maintaining and nourishing a life outside the womb, so that the independent existence of the second life can in reason and all fairness be the object of state protection that now overrides the rights of the woman”).
Moreover, while Carhart performs a similar procedure, he does not do the precise procedure pioneered by Dr. Martin Haskell commonly (although not medically)19 thought of as “partial-birth abortion.” Compare WMPC, 911 F.Supp. at 1066 (describing the Haskell “D & X” procedure) with Report of Am. Med. Ass’n at 8 & n. 19 (Ex. 7) (describing “intact dilation and extraction” or “D &' X”). The procedure performed by Dr. Haskell has four components: (1) deliberate dilation of the cervix, usually over a number of days; (2) intentional conversion of the living fetus to a “footling breech” (feet-first position) using instruments; (3) pulling the living fetus, except the head, which is too large to pass through the cervix, intact from the uterus through the cervix using instruments; and (4) evacuating the contents of the fetal skull (causing brain death but perhaps not cessation of heart function), thereby compressing the skull and allowing the intact but compressed skull to be pulled through the cervix. WMPC, 911 F.Supp. at 1066. When these four steps are accomplished, vaginal delivery of a dead but otherwise intact fetus takes place and the abortion is complete. Id. Dr. Haskell typically uses this procedure from 20 to 24 weeks. Id. at 1065-66.
In contrast to the Haskell procedure, Car-hart takes the living fetus “as he finds it.” In other words, Carhart performs the same procedure as Dr. Haskell, but he does not convert the fetus to a “footling breach” if it is not naturally in that position.20 If the living *522fetus presents “feet first,” Carhart extracts the fetus’s torso from the uterus and then drains (or crushes) the skull, which is too large to pass through the cervix unless the skull size is reduced. He performs this procedure approximately 10 to 20 times a year. He tries to do the procedure for every patient between the 16th and 20th week of. pregnancy. Carhart performs about 190 abortions during this period.
Thus, Carhart’s procedure is the same as Dr. Haskell’s if the fetus presents “feet first.” Other doctors also use this variation of the Haskell procedure. WMPC, 911 F.Supp. at 1067 n. 20 (Dr. Doe Number Two).
If the fetus does not present “feet first,” Carhart performs a D & E procedure, that is, dismemberment of the living fetus by (1) inserting an instrument into the uterus; (2) grasping a part of the fetus, such as a limb; (3) pulling the fetal part through the cervix; and (4) shearing the part from the body by means of traction where the end of the cervix (cervical os) is the leverage point. When the “D & E” procedure is used, the living fetus is removed from the uterus in parts rather than intact.
Furthermore, and again in contrast to Dr. Haskell, Carhart does his variation of the “partial-birth” abortion procedure primarily on living fetuses around the 16th week to the 20th week rather than between the 20th week and the 24th week. Carhart normally does not attempt the D & X procedure before the 16th week because the fetal body tears apart during the process. After about the 20th week, Carhart induces fetal death by injection before removing the fetus from the uterus unless the woman suffers from a seizure disorder or heart disease or is taking medication that contraindicates use of the injection.
To accomplish fetal death, about 48 to 72 hours before removal of the fetal body, Car-hart surgically pierces the woman’s abdomen and injects the fetus with a lethal substance intended to stop its heart. Carhart does not induce fetal death before about the 20th week.
We summarize these findings as follows:
1. Carhart does not do abortions on viable fetuses.
2. Carhart is primarily at risk for prosecution between the 16th week and the 20th week when, save for its head, he places a living fetus into the vaginal cavity before reducing its skull because: (a) Carhart does not normally use the D & X procedure before the 16th week since the fetal body tears apart when pulled from the uterus; and (b) after the 20th week Carhart normally kills the fetus by injection before beginning the D & X procedure.
3. Carhart’s D & X procedure varies from Dr. Haskell’s procedure in two respects: (a) Haskell attempts to manipulate every fetus to obtain a “feet-first” presentation while Carhart will take the fetus “feet first” if doing so is easier; and (b) Carhart normally uses the procedure on living fetuses between about the 16th week and the 20th week, while Haskell uses the procedure between the 20th week and the 24th week.
C. “Facial” Or “As Applied” Challenges
For purposes of his preliminary injunction request, Carhart limited his challenge to how the law is “applied” to him and his patients. The “practical effect of holding a statute unconstitutional as applied’ is to prevent its future application in a similar context, but not to render it utterly inoperative. To achieve the latter result, the plaintiff must succeed in challenging the statute ‘on its face.’ ” Ada v. Guam, Soc’y of Obstetricians & Gynecologists, 506 U.S. 1011, 1012, 113 S.Ct. 633, 633, 121 L.Ed.2d 564 (1992) (Scalia, J., dissenting from denial of certiorari).
This concession occurred during a telephone conference between the undersigned and counsel on July 24,1997. As a result, we need not decide whether Nebraska’s law is facially invalid in “‘a large fraction of the eases in which [the law] is relevant.’ ” Planned Parenthood, Sioux Falls Clinic v. Miller, 63 F.3d 1452, 1456 (8th Cir.1995) (quoting Casey, 505 U.S. at 895, 112 S.Ct. at 2829-30), cert. denied sub nom. Janklow v. *523Planned Parenthood, Sioux Falls Clinic, — U.S. -, 116 S.Ct. 1582, 134 L.Ed.2d 679 (1996). On the contrary, the inquiry is limited to Carhart and his patients.
If the ban is invalid with regard to Car-hart, it is invalid for “a large fraction of the cases in which [the law] is relevant.” The law applies to every woman who seeks an abortion from Carhart between the 16th and 20th week of pregnancy; therefore, the ban directly impacts about 190 of the 800 women who seek abortions from him annually.
As long as he limits his challenge to how the law is “applied” to his situation, a favorable ruling for Carhart will not invalidate the Nebraska law generally. In contrast, such a ruling will only preclude enforcement of the Nebraska law against Carhart and his patients (and other similarly situated doctors and their patients). For example, a favorable ruling for Carhart would not preclude enforcement of the ban on post-viability abortions since Carhart does not perform such abortions and his “as-applied” challenge does not extend to that portion of the law. Ada, 506 U.S. at 1012, 113 S.Ct. at 633.
D. Preliminary Injunction Standards
In Dataphase Systems, Inc. v. C L Systems, Inc., 640 F.2d 109, 113 (8th Cir.1981) (en banc), the United States Court of Appeals for the Eighth Circuit required district courts confronted with a request for a preliminary injunction to evaluate (1) the threat of irreparable harm to the movant; (2) the state of the balance of the harm and injury that granting the injunction will inflict on other parties litigant; (3) the probability that the movant will succeed on the merits; and (4) the public interest. “No single factor in itself is dispositive; in each case all the factors must be considered to determine whether on balance they weigh towards granting the injunction.” Calvin Klein Cosmetics Corp. v. Lenox Lab., Inc., 815 F.2d 500, 503 (8th Cir.1987).
E. Unnecessary Maternal Medical Risk is an “Undue Burden”
Carhart claims Nebraska’s “partial-birth” abortion ban prohibits his female patients from choosing, with his advice, a safe and desired method of terminating pregnancy before viability. Therefore, Carhart argues that the Nebraska law as applied is unconstitutional in that it violates his patients’ Fourteenth Amendment Due Process right to choose whether to have an abortion before viability. Roe v. Wade, 410 U.S. 113, 93 S.Ct. 705, 35 L.Ed.2d 147 (1973), as modified by Casey, 505 U.S. at 833, 112 S.Ct. at 2796.
After applying and carefully weighing the Dataphase factors, we find and conclude that Carhart is entitled to a preliminary injunction. Nebraska’s ban on Carhart’s “D & X” procedure has the effect of subjecting Car-hart’s patients to an appreciably greater risk of injury or death than would be the case if these women could rely upon Carhart to perform his variant of the banned procedure on nonviable fetuses when medically advisable.
1. Probability of Success on the Merits
a. The Law
Carhart is likely to succeed on the merits. Nebraska’s ban on “partial-birth abortions” requires women who seek abortions on nonviable fetuses to take appreciable medical risks that are not necessary to preserve their lives or health. Nebraska’s law thus subordinates maternal life and health to the life and health of a nonviable fetus. As a result, Nebraska’s ban imposes a constitutionally unacceptable “undue burden” on Carhart’s patients. Casey, 505 U.S. at 878, 112 S.Ct. at 2821. Simply put, we cannot constitutionally allow the life or health of a woman to be made subservient to the state’s, otherwise profound interest in a nonviable fetus.
An abortion law is invalid if it places an “undue burden” in “the path of a woman seeking an abortion before the fetus attains viability.” Id. See also Planned Parenthood, Sioux Falls Clinic, 63 F.3d at 1463 (applying Casey and holding, among other things, that parental notification provisions of abortion law that did not include bypass options other than bypass options for abused and neglected minors was unconstitutional). An abortion law may create an “undue burden” if it “has the purpose or effect of placing a substantial obstacle in the path of a woman seeking an abortion of a nonviable fetus.” Casey, 505 *524U.S. at 877, 112 S.Ct. at 2820. In other words, the law may be invalid because (1) the legislature intended to place a substantial obstacle in the path of a woman seeking abortion, or (2) the effect of the law is to place a substantial obstacle in the path of a woman seeking abortion. Id.
Of course, not all burdens are “undue.” Thus, state regulations “which do no more than create a structural mechanism by which the State, or the parent or guardian of a minor, may express profound respect for the life of the unborn are permitted, if they are not a substantial obstacle to the woman’s exercise of the right to choose.” Id. Likewise, “[regulations designed to foster the health of a woman seeking an abortion are valid if they do not constitute an undue burden.” Id. at 878, 112 S.Ct. at 2821. In short, the “very notion that the State has a substantial interest in potential life leads to the conclusion that not all regulations must be deemed unwarranted.” Id. at 876, 112 S.Ct. at 2820.
Nevertheless, it is beyond dispute that “[unnecessary health regulations that have the purpose or effect of presenting a substantial obstacle to a woman seeking an abortion impose an undue burden.” Id. at 878, 112 S.Ct. at 2821. The Supreme Court and lower federal courts have consistently held that abortion regulations that impose medically unnecessary health risks on women are invalid. See, e.g., Thornburgh v. American College of Obstetricians & Gynecologists, 476 U.S. 747, 768, 106 S.Ct. 2169, 2182-83, 90 L.Ed.2d 779 (1986) (state law not susceptible to construction that did not “ ‘require the mother to bear an increased medical risk in order to save her viable fetus’ ” was unconstitutional) (citation omitted)21; Colautti, 439 U.S. at 400, 99 S.Ct. at 688 (state law requiring a doctor to use the technique that provided the best opportunity for the viable fetus to be aborted alive was unconstitutional because it was uncertain whether the law permitted “the physician to consider his duty to the patient to be par-amount to his duty to the fetus”); Planned Parenthood of Mo. v. Danforth, 428 U.S. 52, 79, 96 S.Ct. 2831, 2845-46, 49 L.Ed.2d 788 (1976) (state law that “forces a woman and her physician to terminate her pregnancy by methods more dangerous to her health than the method outlawed” was unconstitutional); Jane L. v. Bangerter, 61 F.3d 1493, 1502-1504 (10th Cir.1995) (Utah law which required post-viability abortion to be conducted in a manner that gave the unborn child the best chance of survival unless the method would cause “grave damage to the woman’s medical health” impermissibly burdened a woman’s right to seek abortion under Casey because it required a woman to bear an “ ‘increased medical risk’ ” in order to save the life of a viable fetus) (quoting Thornburgh, 476 U.S. at 769, 106 S.Ct. at 2183), rev’d in part on other grounds sub nom. Leavitt v. Jane L., — U.S.-, 116 S.Ct. 2068, 135 L.Ed.2d 443 (1996) (per curiam) 22; WMPC, 911 F.Supp. at 1070 (D & X procedure “appears to pose less risk to maternal health than any other alternative”; therefore, the ban had the “effect of placing a substantial obstacle in the path of women seeking pre-viability abortions” and was “an undue burden and thus unconstitutional under Casey ”).
b. Greater Maternal Medical Risk
Keeping in mind that Carhart normally does the D & X procedure on living fetuses between about the 16th and 20th weeks of pregnancy, Nebraska’s ban on “partial-birth” abortions regarding nonviable fetuses has the “effect”23 of subjecting his *525patients to an appreciably greater risk of injury or death than would be the case if these women could rely upon Carhart to do his variant of the banned procedure when medically advisable. The result is an “undue burden” under Casey.
(1) Statistically, the D & X Procedure is at Least as Safe as the D & E Procedure and Appreciably Safer than All Other Forms of Abortion
The data suggest that the D & X procedure, a variant of the D & E procedure, is appreciably safer than all other forms of abortion during the relevant gestational time. According to the AMA report, there are four alternative procedures that “can be used to induce abortion” between “the 16th to 24th week of gestation,” “includ[ing] dilation and evacuation [¶] & E] (which may or may not be preceded by induced fetal demise), dilation and extraction (D & X), labor induction, hysterotomy and hysterectomy.” (Ex. 7, at 8.) The “D & X” procedure (used by Carhart and others) is “a form of D & E.” (Id.)
When measured by maternal death, for the period between 13 weeks and plus or minus 21 weeks, the “D & E” procedure is measurably safer than labor induction, hysterotomy, and hysterectomy, the other abortion procedures customarily used during this time. (Id. at 10, tbl. 4: Maternal Mortality Rates for Induced Abortion Procedures at 13 Weeks’ Gestation or Later, U.S., 1974-1987) (Total: “D & E,” 3.7 deaths 24; Induction, 7.1 deaths; Hysterectomy/ Hysterotomy, 51.6 deaths). The same is true for the 16th week to the 20th week. (Id.) (16-20 weeks: “D & E,” 6.5 deaths; Induction, 7.9 deaths; Hysterectomy/Hysterotomy, 103.4 deaths).
Moreover, the “risk of complications” is as important as the mortality risk. The complication 25 rate for the D & E procedure is only 7 per 1,000 compared to between 21 and 25 per 1,000 for induction procedures, and a much greater risk is associated with other procedures such as hysterectomy. (Id.)
According to Dr. Henshaw of the Guttmacher Institute,26 because the D & X is a variant of the D & E, no separate statistics comparing the D & E with the D & X are available. (Tr. 52, 70, 78-79.) Nevertheless, because “the D & X is one type of D & E,” the “only reason to single that [procedure] out for special interest is because it has special political significance.” (Tr. 79:4-7.) The difference between the two procedures “isn’t really a medical issue, but it is a political issue.” (Tr. 79:1-2.) Consequently, we believe the overall mortality and complication rates for the D & X procedure are at least as good as the D & E and much better than other forms of abortion.
(2) Medical Evidence Establishes That the D & X Procedure is Appreciably Safer for Women Than the D & E Procedure
The credible medical evidence establishes that the D & X procedure used by Carhart is appreciably, safer than the D & E procedure “may minimize trauma to the woman’s uterus, cervix, and other vital organs” because it reduces “uterine or cervical perforation either from instruments used during the D & E, or through piercing by fetal parts.” (Ex. 7, at 8.)27 Moreover, Dr. Hodgson, a very *526credible board-certified physician who has performed more than 30,000 abortions, delivered at least 5,000 babies, and is a founding fellow of the American College of Obstetricians and Gynecologists, believes the D & X procedure is “an advance in technology” because by removing the fetus intact there is “less instrument manipulation” and “of course, the higher your safety.” (Tr. 212:4-22.)
Furthermore, Carhart’s testimony convinces us that the D & X procedure he uses is appreciably safer than performing a D & E because:
1. Intact removal of the fetus reduces the possibility of complications from the passage of sharp bony fragments through the cervix and is therefore less traumatic to a woman’s cervix and uterus. (Tr. 101, 124-25, 131, 158);28
2. By removing the fetus intact, as opposed to removing pieces from the uterus as required by the D & E, the D & X procedure reduces the use of instruments in the uterus which in turn lessens the risk of complications from tearing or perforating the uterus with instruments. (Tr. 108, 131, 158, 179-80);
3. By using this procedure, Carhart prevents disseminated intravascular coagulopathy (DIC) and amniotic fluid em-bolus, both of which are among the most common causes of maternal mortality and complications. (Tr. 101-02, 131);
4. By removing the fetus intact, Carhart reduces the likelihood of retained fetal parts, (Tr. 183), which Dr. Hodgson testified is a “horrible complication.” (Tr. 211:21.)
Carhart’s testimony is consistent with testimony received by the court in WMPC, 911 F.Supp. at 1069.29 In that case, Dr. George Goler, Ohio section chief of the American College of Obstetricians and Gynecologists, testified that he “views [the D & X procedure] as an improvement over the traditional D & E procedure” because it causes “less trauma to the maternal tissues by avoiding the break up of bones, and the possible laceration caused by their raw edges” and because it causes “less blood loss.” Id. (parentheses omitted). In addition, Dr. Haskell testified that his use of the D & X procedure resulted in no serious complications in approximately 1.000 procedures, while out of approximately 1.000 other procedures where he used the D & E technique, two patients had serious complications. Id.
Carhart’s testimony is also consistent with the testimony received by the court in Evans. Evans, No. 97-CV-71246-DT, slip op. at 30-31, •— F.Supp. at---. There the court heard testimony from various doctors, including Dr. Mark Evans (plaintiff), a board-certified obstetrician, gynecologist, clinical geneticist, and vice-chief of obstetrics at a hospital; Dr. Dennis Christensen (plaintiff), a board-certified obstetrician and gynecologist and assistant clinical professor at the University of Wisconsin (where he teaches abortion procedures); Dr. Carolyn Westoff (plaintiffs expert), a board-certified obstetrician, gynecologist, and epidemiologist; Dr. Curtis Cook (defense expert), a board-certified obstetrician and gynecologist; and Dr. Timothy Johnson (court-appointed expert), a board-certified obstetrician and gynecologist with a sub-certification in the area of maternal-fetal medicine. Id., slip op. at 5-13, — F.Supp. at---.
“Comparing a conventional D & E to an intact D & E, Drs. Westoff, Doe, Evans, Christensen, Johnson and Cook all agree that the intact procedures reduce risks associated with conventional D & Es,” id. at 30, at --, for four reasons:
1. “The intact procedure reduces the risk of uterine perforation and cervical lac*527erations because removing the fetus intact, rather than in pieces, entails less instrumentation of the uterus and minimizes the passage of bony fetal fragments through the cervix and vagina.”
2. “Removing the fetus intact also reduces the risk of a free-floating head, an uncommon, but significant complication of a conventional D & E.”
3. “And,” Dr. Westoff testified, “[RJemoving the fetus intact is often quicker than dismembering it, the woman has less operative time, which means less risk of hemorrhage, less total bleeding and less risk of infection.”
4. “Further,” Dr. Evans opined, “[T]he intact procedure would be superior to a conventional dismemberment procedure for a woman for whom induction is contraindicated and who has a presumed fetal abnormality for which- a fetal autopsy would be very instructive.”
Id. at 30-31, at---.
(3) Inducing Fetal Death by Injection has Appreciable Maternal Risks But No Maternal Benefits and is Not Always Possible
While there is no maternal health benefit in attempting to kill a fetus in útero by injection before the 20th week of pregnancy, there are appreciable maternal health risks in attempting such a procedure. Consequently, if Carhart has to ensure fetal death by injection before the 20th week in order to do a D & X procedure, he must subject his patients to an appreciable risk of harm for no maternal benefit or else perform more risky procedures such as the D & E.
The reason doctors use an injection to kill the fetus in útero after the 20th week is that the size of the fetus makes it difficult to extract from the uterus. (Tr. 120,178-79.)30 If they kill the fetus and wait two to three days, the body of the fetus shrinks and becomes less rigid and more malleable. (Tr. 120.) Thus, the medical benefit to the mother of killing the fetus by injection after 20 weeks is reduced fetal size and increased malleability that results in less difficulty (including less dilation of the cervix) removing the fetus from the uterus. (Id.) However, before the 20th week, when the fetus is smaller, there is less need to obtain shrinkage and malleability, and shrinkage and malleability are, in any event, insignificant. (Id. at 120-21.)
On the other hand, killing the fetus by injection requires a surgical procedure that carries risks to maternal life and health. (Tr. 121-22; 186-88.) To kill a fetus using this method, a needle is inserted through the mid-line of the woman’s abdomen.31 (Tr. 186.) The doctor then attempts to guide the needle to the fetal heart by ultrasound. (Id.) The doctor then tries to pierce the fetal heart or surrounding amniotic sac with the needle to inject a solution that will stop the fetal heart.,, (Id. at 186-87.) During this procedure, the doctor must avoid the maternal bowel and uterine wall not only to avoid damage to those structures, but also to avoid injecting a potentially dangerous drug into the maternal bloodstream. (Tr. 121-22; 188.) Such a procedure is particularly difficult when the fetus is less than 20 weeks because “you have a much smaller target and you have more of a chance of getting medication into the maternal circulation than you do into the fetal circulation.” (Tr. 121:25-122:3.)32
*528The defense expert, Dr. Riegel, corroborated some of Carhart’s testimony about the advisability of using fetal injections to cause death. For example, Dr. Riegel testified there was no maternal medical advantage to injecting a substance into the fetus while it is in the uterus in order to kill the fetus prior to removing it. (Tr. 250.) Moreover, Dr. Riegel testified that physicians generally restrict themselves “to 19 weeks and up ... if people are going to use that particular route to stop the heart of a fetus.” (Tr. 244:24-255:1.)33 Still further, in the part of Texas where Dr. Riegel practices, only “high risk obstetricians” conduct the related but less risky amniocentesis procedure between 13 and 18 weeks. (Tr. 285.)
In addition, some women cannot be given an injection to kill the fetus under any circumstances. For example, injecting medication to stop the fetal heart when the mother suffers from a heart condition is inappropriate. (Tr. 122-23.) Consequently, if we required that Carhart use an injection to kill the fetus when'not medically advisable in order to do a D & X procedure (whether before or after 20 weeks), either his patients would be subjected to the unnecessary risk of the injection or Carhart would have to perform more risky procedures such 'as the D & E.
(4) There is No Maternal Health Benefit in Forcing the Woman to Wait Until After the 20th Week So That the Fetus Can Be Killed by Injection, and There is Appreciable Risk to Maternal Health in Doing So
Assuming the patient does not have a heart condition or similar problem precluding fetal injection, in order to do the D & X procedure with a dead fetus, Carhart could foi’ce a woman to wait for an abortion until after the 20th week and then undergo the injection procedure to kill the fetus in útero. There is no maternal medical benefit to this strategy, and there is appreciable risk in doing so.
For example, Dr. Henshaw testified that “the rate of complications increases by about 20% for each additional week of gestation from about eight weeks on.” (Tr. 59:12-14.) In this vein, the AMA report confirms that the “risk of maternal death increases with gestational age.” (Ex. 7, at 9.) In fact, before 21 weeks the risk of maternal death due to abortion is measurably less than the risk of maternal death from childbirth,34 but by 21 weeks the risks are similar. (Id.)
(5) Cutting the Umbilical Cord and Waiting for the Fetus to Die before Completing the D & X Procedure Carries Appreciable Maternal Risks and No Maternal Benefits and Is Not Always Possible
Although Carhart tries to cut the umbilical cord during the D & X procedure, with the result that the fetus eventually dies, he does not wait the 6 to 10 minutes it takes for the fetus to die before completing the D & X procedure because doing so increases maternal blood loss. (Tr. 190.) There is no maternal benefit in waiting for the fetus to die in this circumstance, and there is an appreciable risk to maternal life and health in doing so. If foi’ced to cut the cord and wait for the fetus to die, Carhart must either abandon the D & X procedure for the more risky D & E or subject his patient to unnecessary blood loss. Moreover, there are times when Car-hart cannot reach the cord to cut it before vaginal delivery of the fetal body. (Tr. Ill, 190.) In this circumstance, he would be forced to use the more risky D & E procedure or violate the law by completing the D & X procedure.
c. Defendants’ Arguments are Not Persuasive
The defendants assert various arguments to convince us that maternal risk is irrelevant, nonexistent, or insignificant. These arguments merit a brief response. Moreover, it is helpful to state the one defense not asserted by the defendants.
*529The defendants do not argue that the “medical exception” to Nebraska’s ban saves the statute from Carhart’s challenge. Given the plain meaning of the medical exception, they have no basis for such an argument.
The law prohibits Carhart from performing his variant of the D & X procedure “unless such procedure is necessary to save the life of the mother whose life is endangered by physical disorder, physical illness, or physical injury, including a life-endangering physical condition caused by or arising from the pregnancy itself.” LB 23 § 3(1). This exception does not apply to Carhart’s patients because:
1. Risks to maternal health are not considered;
2. Despite the fact that abortions are safer than childbirth until the 21st week, mortality risks associated with one abortion procedure being more dangerous than another are not considered since such risks do not arise out of a “physical disorder, physical illness, or physical injury, including ... pregnancy itself.”
As the defendants do not contend that the medical exception protects Carhart and saves the statute, we need not pursue the issue any further.
The defendants do argue that Roe and Casey do not apply to Nebraska’s law because “the Supreme Court has never recognized a constitutional right to kill a partially bom human being.” (Defs.’ Br. Opp’n Pl.’s Mot. Prelim. Mot. [sic ] at 3 (emphasis added).) We reject this argument because there is no precedent for it.35
Roe and Casey categorized fetuses as viable or not viable. No case with which we are familiar uses the “partially born human being” category as a construct for constitutional analysis. It is our job to fairly apply the precedents of the Supreme Court whether we agree with them or not. See, e.g., Dronenburg v. Zech, 741 F.2d 1388, 1396 n. 5 (D.C.Cir.1984) (Bork, J.). Accordingly, we decline the defendants’ invitation to ignore Roe and Casey.
Next the defendants argue that there is no risk to maternal health or life if Carhart is precluded from using his variant of the D & X procedure. Using “statistical significance” as their point of attack, defendants argue that no “statistically significant” difference in mortality rates exists for the D & E procedure (which, they agree, includes the D & X procedure) compared to the induction procedure. (Defs.’ Post-Hr’g Br. Opp’n Pl.’s Mot. Prelim. Inj. at 10-12.)36 Consequently, the defendants argue that Carhart’s patients can always use the equally safe induction method of abortion as an alternative to the D & X procedure.37
Among other reasons, we reject this argument because:
1. The statistical analysis in the defendants’ brief was not presented as evi*530dence through the testimony of a reputable analyst such as Dr. Henshaw;
2. The statistical analysis in the defendants’ brief is flawed because it does not employ accepted methodology for determining statistical significance;
3. The statistical-significance argument is limited to mortality rates and ignores complication rates, see, e.g., Evans, No. 97-CV-71246-DT, slip op. at 28, - F.Supp. at-(Dr. Westoff explained that “inductions have higher rates of infection, bleeding, cervical lacerations, amniotic fluid embolus and disseminated intravascular coagulation (‘DIC’)”);
4. “Statistical significance” is not normally used by doctors to express risk assessments for individual patients (Ex. 7, at 10 (tbl. 4 pertaining to maternal mortality and text pertaining to maternal morbidity));
5. The statistical-significance argument ignores the testimony of doctors like Carhart, Hodgson, Goler (WMPC), Haskell {WMPC), Westoff {Evans), Evans {Evans), Christensen {Evans), Johnson {Evans), and Cook {Evans) that the D & X is safer than the D & E; thus, even if the D & E procedure and the induction method were equally safe, the D & X appears to be safer than either of them;
6. The statistical-significance argument ignores the AMA report’s recognition that the D & X “procedure may minimize trauma to the woman’s uterus, cervix, and other vital organs,” (Ex. 7, at 8), and that “maternal mortality during second-trimester abortions is lower38 for dilation and evacuation procedures than for labor induction methods.” (Ex. 7, at 11);
7. The statistical-significance argument ignores the fact that inductions are contraindicated for certain women because of medical conditions like hypertension, heai't disease, or diabetes. (Tr. 125:6-126:1.) See also Evans, No. 97-CV-71246-DT, slip op. at 28-29, -F.Supp. at-- — — (Drs. Johnson, Evans, and Westoff testified that inductions are either “absolutely” or “relatively” contraindicated for (a) women with a prior classical Caesarean scar; (b; women with a prior hysterotomy; (c) women with renal disease; (d) women with cardiovascular disease; and (e) cases of fetal abnormality where the size of the fetal skull “can increase the risk of vaginal delivery”).
In a related argument, the defendants point to Dr. Riegel’s testimony suggesting the D & X procedure is dangerous to maternal health because it involves the use of sharp instruments to pierce the fetal skull and because the procedure is “blind.” Among other reasons, we reject this argument because Dr. Riegel is not competent to testify about procedures he has neither observed nor performed.
Finally, the defendants argue that a law denying a woman a “marginally ‘safer’ ” abortion is nevertheless constitutional because the increased risk must pose “a significant threat to the mother” in order to be unconstitutional. (Defs.’ Post-Hr’g Br. Opp’n PL’s Mot. Prelim. Inj. at 4-9.) We reject this argument for two reasons.
Initially, women deciding whether to seek abortions are confronted with a substantial obstacle to choosing abortion (undue burden) if they are required by the state to accept an increased risk of death or serious injury which, though small, is nevertheless real. The cases we have cited in part II. E.l.a. stand squarely for this proposition.
Moreover, we reject the premise of the defendants’ argument; that is, we are unwilling to find and conclude on the record before us that the increased risk to maternal life and health is “small” or “marginal.” In particular we point out that for women who die or suffer serious complications because they cannot have the safest available procedure to abort them nonviable fetuses, the increased risk cannot honestly be considered insignificant.
*5312.Irreparable Harm to Plaintiff
Plaintiff and his patients will suffer irreparable injury absent an injunction. It is well established that a statute that “interfered with the exercise of [abortion providers’] constitutional rights and the rights of [their] patients supports a finding of irreparable injury.” Planned Parenthood of Minn. Inc. v. Citizens for Community Action, 558 F.2d 861, 867 (8th Cir.1977). This is particularly true where, as here, the statute imposes appreciable risks on the lives and health of a doctor’s patients. Death is irreparable.
3.Harm to Defendants
We do not underestimate Nebraska’s profound interest in nonviable fetuses. However, any reasonable analysis of the balance of harm to the defendants and the plaintiff and his patients favors the plaintiff and his patients. Nonviable fetal life cannot constitutionally be considered superior to maternal life or health. More practically, one defense attorney candidly39 acknowledged that “we know there are a lot of problems,” including “great proof problems,” with enforcing the statute as written. (Tr. 43:4-11.) These statements amount to a concession that the defendants will suffer little or no injury if the statute is preliminarily enjoined.
4.Public Interest
We recognize that the Nebraska legislature passed LB 23 with near unanimity. We are also aware that a federal court must proceed cautiously when asked to interfere with the democratic processes of state government. Nevertheless, the public interest would not be served by failing to enjoin a legislative scheme such as this that carries with it real risks to maternal health and safety. We are not the only court to arrive at the conclusion that injunctive relief is justified in cases dealing with “partial-birth” abortion statutes. See, e.g., Evans, No. 97-CV-71246-DT, — F.Supp. - (preliminary and permanent injunction); Little Rock Family Planning Servs. v. Jegley, No. LR-C-97-581 (E.D.Ark. July 31, 1997) (temporary restraining order); Planned Parenthood of S. Ariz., Inc. v. Woods, No. CIV 97-385-TU C-RMB (D.Ariz. July 18, 1997) (same); Causeway Med. Suite v. Foster, No. 97-2211 , (E.D.La. July 14, 1997) (same); Rhode Island Med. Soc’y v. Pine, No. 97-416L (D.R.I. July 11,1997) (same); Midtown Hosp. v. Miller, No. 1:97-CV-1786-JOF (N.D. Ga. June 27,1997) (same); WMPC, 911 F.Supp. at 1094 (preliminary injunction).
III. CONCLUSION
We know three things with certainty. We know the turmoil the abortion controversy causes the citizens of Nebraska. We also know as unelected judges that we are not “free to invalidate state policy choices with which we disagree.” Casey, 505 U.S. at 849, 112 S.Ct. at 2806. Finally, and despite the first two points, we know we are not free to “shrink from the duties of our office.” Id. To this end, we have ruled as we believe the law, interpreted by the Supreme Court, requires — nothing less and nothing more.
IT IS ORDERED that:
(1) Plaintiffs request for a preliminary injunction is granted as provided herein;
(2) Defendants, and each of them, including their agents, servants, and employees, are preliminarily enjoined from enforcing LB 23 against Plaintiff regarding his performance of D & X abortions on nonviable fetuses;
(3) The court determines under Federal Rule of Civil Procedure 65 that a bond in the amount of $105.00 is sufficient and directs Plaintiff to post such a bond.