BROWN, J.

The circuit court reversed the State of Wisconsin Medical Examining Board's findings that Alonzo R. Gimenez, M.D., endangered the health of his patients. The circuit court concluded that the Board had not mentioned or discussed certain elements which it believed to be necessary components of Board factfinding. The Board appeals, claiming that it need only set forth facts which, in the Board's opinion, support a finding that the physician endangered a patient's health. We disagree with the Board and hold that with every charge of endangering a patient's health, there are five elements which the Board must discuss seriatum in a written decision. The Board is compelled by law to find the facts either supporting or not supporting each element. We disagree, however, with the circuit court's remedy of dismissing all charges against Gimenez. The proper remedy is to remand the case to the Board with directions to craft a decision conforming with the law. We affirm in part, reverse in part and remand with these directions.

Gimenez is a general surgeon who has been practicing medicine in the Berlin area since 1955. Starting in 1965, Gimenez worked in partnership with Dr. David Sievers until Sievers retired in 1987. After Sievers left the practice, Gimenez was forced to take on a greater caseload.

The incidents which formed the basis for the Board's investigation took place during the period when Gimenez was seeing more patients owing to his partner's retirement. The four patients which the Board was concerned about had a range of diseases, including an infected appendix, cancer of the bladder, cancer of the colon and cardiovascular problems. Because we are focused on the Board's procedures, and not its substantive decision that Gimenez acted improperly, we need not provide any further details of these patients' illnesses and Gimenez's treatment.

The Board initiated proceedings against Gimenez in February 1991. The administrative law judge submitted his proposed decision to the Board on August 14, 1992. The Board held oral arguments that October, made some modifications to the ALJ's recommendations and issued its final decision in November 1992. The Board found that Gimenez's treatment of these four patients was in violation of § 448.02(3), STATS., and WIS. ADM. CODE § MED 10.02(2)(h) which prohibit physicians from engaging in conduct which threatens the health and safety of their patients.

In November 1992, Gimenez filed his appeal of the Board's decision with the circuit court. See § 227.52, STATS., 1993-94, amended, 1995 Wis. Act 27, § 6233. In July 1995, the court ruled that the Board's decision was "arbitrary" and "not sustained by the record" and therefore set aside the Board's decision. In particular, the circuit court was concerned by the "lack of any findings to the ultimate material facts."

The Board now claims that the circuit court erred and asks us to reinstate its original decision. We will apply the same standards that the circuit court did and independently review the Board's decision. See Gibson v. State Public Defender, 154 Wis.2d 809, 812, 454 N.W.2d 46, 47-48 (Ct.App. 1990).

Our review of the Board's decision would ordinarily be governed by the "substantial evidence test." Under this test, we would only determine if its findings are reasonably supported by the evidence. See id. at 812-13, 454 N.W.2d at 48. Here, however, the principal issue pertains to whether the Board fulfilled its duties under § 227.47(1), STATS., to adequately support its decision with written findings. This issue involves a question of law on which we owe no deference to the Board. See Sauk County v. WERC, 165 Wis.2d 406, 413, 477 N.W.2d 267, 270 (1991); but compare Kelley Co. v. Marquardt, 172 Wis.2d 234, 244-45, 493 N.W.2d 68, 73 (1992) (describing how courts may defer to administrative interpretations of law when the statute pertains to specialized, technical matters).

In Gilbert v. Medical Examining Bd., 119 Wis.2d 168, 349 N.W.2d 68 (1984), the supreme court reviewed the legal standards used to define whether a physician's choice of treatment constituted "unprofessional conduct" because it posed a threat to his or her patient or to the public. See § 448.02(3), STATS.; WIS. ADM. CODE § MED 10.02(2)(h). In its review of the specific finding that the physician had engaged in "unprofessional conduct," the court addressed five separate issues.

The court began by describing the specific patient's condition and the course of treatment that the physician provided. See Gilbert, 119 Wis.2d at 175-77, 349 N.W.2d at 70-71. It then explained that the Board was required in these cases to establish what the minimum standards of treatment involved. See id. at 191-92, 196, 349 N.W.2d at 78, 80. The court then described that the Board must show how the physician's treatment decisions departed from these standards. See id. at 193, 349 N.W.2d at 79. In addition, the court discussed why the Board must also demonstrate that the physician's course of treatment created "risks and negative results which are unacceptable to other physicians." See id. Finally, the court noted that the Board must also explain what "different course of treatment" the physician could have taken to avoid creating an unreasonable risk for the patient. See id. at 197, 349 N.W.2d at 81.

Accordingly, we conclude that a reasonable reading of Gilbert is that the supreme court set out a five-pronged test to guide the Board in its determination of whether a physician improperly treated a patient. Again, these five elements are:

(1) what course of treatment the physician provided;

(2) what the minimum standards of treatment required;

(3) how the physician's treatment deviated from the standards;

(4) how the treatment created an unacceptable level of risk; and

(5) what course of treatment a minimally competent physician would have taken.

See generally id. at 196, 349 N.W.2d at 80. Nonetheless, what concerns us in this case is not only the Board's substantive conclusion that there was sufficient evidence on these five elements; we are also concerned with whether the Board's written decision provides an adequate explanation of why it believed it had sufficient evidence on all these elements.

We will therefore turn to the Board's written decision to see what it provides. As we noted above, the Board set out its conclusion regarding Gimenez's treatment of these patients in an eighteen-page report. After a brief introduction, the report contains thirteen pages outlining how Gimenez treated each patient. This section is labeled "Findings of Fact" and thoroughly details each patient's symptoms and test results. It also describes what Gimenez did to treat each patient.

The report then sets forth the Board's "Conclusions of Law." In this one-page section, the Board identifies the reasons why it believed that Gimenez had endangered the health and safety of these patients. Below we have set out the specific finding with regard to Gimenez's treatment of Patient I, which is typical (in form) of the Board's findings in regard to his other three patients:

With regard to his treatment of Patient I, Respondent violated sec. MED 10.02(2)(h), Wis. Admin. Code and sec. 448.02(3), Wis. Stats. by delaying an exploratory laparotomy beyond April 24, 1988 in the presence of one or more diagnosable abdominal abscesses. Respondent did not violate any rule or statute by performing a diverticulectomy during the exploratory laparotomy.

We will now describe why this explanation is insufficient as a matter of law.

The Board defends these findings with the claim that its entire eighteen-page written decision, when taken as a whole, provides enough information to cover each of the five Gilbert factors. Indeed, even if we could not reach such a conclusion, the Board contends that it "is not even required to state detailed evidentiary findings of fact."

It rests these positions on a statement from State ex rel. Harris v. Annuity Pension Bd., 87 Wis.2d 646, 275 N.W.2d 668 (1979). There, the court faced a challenge to a pension board which had denied benefits after finding that the pensioner had died from naturally occurring cardiovascular illness, not suicide brought on by job-related stress. See id. at 649-50, 275 N.W.2d at 670. More specifically, the Board cites the finding of that administrative panel which stated only that "Dr. Harris did not commit suicide by means of an overdose of secobarbital on December 6, 1965." Id. at 650, 275 N.W.2d at 670. The Board thus contends that its findings with regard to Gimenez are "at least as definite and instructive as those approved by the supreme court in Harris."

We acknowledge that the Board's four written findings regarding Gimenez's alleged errors in treatment, such as the phrase "delaying an exploratory laparotomy beyond April 24, 1988 in the presence of one or more diagnosable abdominal abscesses," are remarkably similar in form to the one which was allegedly "approved" by the Harris court. Still, we do not accept the Board's basic premise that the Harris court ever intended to set out a litmus test which courts should use to measure the quality of administrative findings concerning any medical matter. Imposition of such a limited requirement would not coincide with the supreme court's explanation of why the legislature provided for judicial review of agency decisions.

The Gilbert court explained that the primary purpose of providing judicial review of administrative rulings is to prevent the exercise of "arbitrary, unreasonable or oppressive conduct by the Board." Gilbert, 119 Wis.2d at 191, 349 N.W.2d at 78. But to insure against such improper behavior, the reviewing court must understand what the Board did. As a result, the administrative rules require the Board to set out its reasoning. See § 227.47, STATS.

In addition, the Gilbert court recognized that judges have little expertise in medical matters. It therefore cautioned that the Board needs to carefully outline every aspect of why it reached its stated conclusions. See Gilbert, 119 Wis.2d at 205, 349 N.W.2d at 84. Thus, the Board's present position, that there exists some minimum acceptable standard of stating findings of fact, would not serve the policy goal of protecting against irrational factfinding because the appropriate detail of written discourse depends upon the issues involved in each case.

While in Harris the court held that the pension board's cryptic finding was sufficient, the board in that case only had to address one narrow issue: was the cause of death natural cardiovascular disease or was it work-related stress? See Harris, 87 Wis.2d at 649, 275 N.W.2d at 670. But in this case, as in all cases involving allegations of mistreatment, the Board needs to cover five different elements. Its decision needs to explain the very complex issue of why Gimenez misinterpreted the information he received about his patients and why he made an error in judgment when he selected a course of treatment. While the Board's terse written findings, which are full of technical jargon, might reveal everything necessary to a medical expert, they will not survive judicial review because the courts are not equipped with the expertise to adequately interpret them.

In sum, we hold that the Board must provide a plain and thorough written decision that summarizes its findings. This decision must separately identify the five Gilbert elements and discuss the evidence which relates to each element. The decision must also provide details of why the evidence supports the Board's findings.

Having concluded that the Board's "Conclusions of Law" are inadequate, we now turn to the appropriate remedy. Again, the Board seems to suggest that we can uphold its decision in this specific case because we should be able to discern a sufficient explanation of the Gilbert factors if we review its eighteen-page decision in its entirety. We disagree.

The first section of the complete report does contain a good explanation of exactly how Gimenez treated these patients. Thus, the Board's decision satisfies the first Gilbert requirement. Moreover, some of the Board's other findings could be construed as meeting the fifth requirement, which requires the Board to describe the better course of treatment. For instance, the statement that Gimenez erred when he "administer[ed] heparin to [Patient VI] without having thoroughly investigated the nature and extent of bleeding in her gastroninestinal tract" suggests that Gimenez would have acted correctly if he would have first investigated whether the patient was bleeding before he administered Heparin. Nonetheless, the Board's written explanation still leaves this court to "draw inferences from the record as to how a minimally competent physician would have proceeded." Cf. Gilbert, 119 Wis.2d at 205, 349 N.W.2d at 84. As judges, not experts voiced in medicine, we cannot confidently draw these necessary inferences and fill in the gaps that exist in the Board's decision.

While we are unable to uphold the Board's decision in its current form, we nonetheless disagree with the circuit court's conclusion that the Board's failure to fully document its reasoning warrants dismissal of the entire investigation. Instead, we will remand this case to the circuit court with directions that it further remand the case to the Board and direct the Board to reconsider the charges against Gimenez in light of this opinion.

We base our choice of remedy on precedent which favors remand over dismissal. First, in Heine v. Chiropractic Examining Bd., 167 Wis.2d 187, 192-93, 481 N.W.2d 638, 641 (Ct.App. 1992), this court held that the Chiropractic Examining Board did not provide a sufficient explanation of why it departed from the hearing examiner's disciplinary recommendations. In addition, we also held that the proper remedy was to remand the matter to the board and give it an opportunity to better explain why it deviated. Id. at 194, 481 N.W.2d at 642. We reasoned that outright reversal of the board and reinstatement of the hearing examiner's findings was inappropriate in light of § 227.57(5), STATS., as this section provides that remand to the agency is appropriate when the agency makes a procedural error.

Furthermore, we are equally aware of the supreme court's command in Wurtz v. Fleischman, 97 Wis.2d 100, 108, 293 N.W.2d 155, 159 (1980), that the court of appeals should not make its own findings. In that case, the supreme court set out a rule to govern those situations when this court is confronted with inadequate findings made by the trial court. The supreme court instructed the court of appeals that it should not simply apply the existing facts to the appropriate standard. In those situations, the supreme court held that such matters should be remanded to the trial court for reconsideration. See id. Although the Wurtz decision specifically dealt with the inadequacy of a trial court record, the principle also applies to the situation when this court (or a circuit court) faces an inadequate administrative record.

Our decision to remand this case dictates that we not address the parties' arguments about the sufficiency of the evidence placed before the Board. Nor will we address Gimenez's complaints about the sanction the Board imposed against him. After the Board reconsiders the case, it may decide that Gimenez did not act in an unprofessional manner with respect to some, or possibly all, of his patients.

In the interests of efficient judicial administration, however, we will briefly address one issue discussed by the parties in their respective arguments about the sufficiency of the evidence. Gimenez's principal complaint on this issue is that the Board did not have "expert testimony which unequivocally indicat[ed] the deviation below minimum standards and the causation of an unacceptable risk." According to Gimenez, the Gilbert decision requires the Board to meet this "unequivocal" standard.

After reviewing Gimenez's arguments, we are concerned that he may be operating under a misconception regarding the amount and quality of proof that Gilbert actually requires. Of course, the court did use the term "unequivocally" to describe the quality of the required proof. Gilbert, 119 Wis.2d at 197, 349 N.W.2d at 81. This term, nonetheless, must be read in context.

When the Gilbert court examined the evidence placed before that board, it acknowledged that it was derived from an expert who testified to a "reasonable degree of medical certainty." See id. at 200, 349 N.W.2d at 82. But if the term "unequivocal" is taken literally, as Gimenez seems to suggest it should be, then the Board could never reach a decision because all the medical testimony it receives is only conclusive to a "reasonable degree of medical certainty."

Thus, we emphasize that the supreme court's use of the term "unequivocal" means only that the Board must rely on evidence from a qualified medical expert who is able to testify on the factor at issue. See id. Since the evidence that the Board uses in its decision-making will necessarily be curtailed by the general inability of experts to testify with any greater conviction than beyond a "reasonable degree of medical certainty," the Board may rely on such testing as a basis for its findings so long as the board finds it credible.

By the Court. — Order affirmed in part; reversed in part and cause remanded with directions.