WILKINS, C.J.

In Herridge v. Board of Registration in Medicine, 420 Mass. 154 (1995) ( Herridge I), we considered Dr. Herridge's appeal from a decision of the defendant Board of Registration in Medicine (board) to suspend for three years the doctor's right to renew his license to practice medicine in the Commonwealth. The board had ruled professional discipline should be imposed because in 1985 the doctor, a psychiatrist, initiated "inappropriate conversation of a sexual nature . . . with a patient during treatment, and a brief sexual relationship with this patient which occurred shortly after the physician-patient relationship terminated." Id. at 155.

Discipline was imposed based on the board's conclusion that the doctor had "sexualized" the doctor-patient relationship, and, therefore, it was unethical for him to engage in a sexual relationship with the patient after the professional relationship had terminated. Id. at 165. The doctor did not deny engaging in sexual relations with the patient on two or three occasions after the doctor-patient relationship had ended. At that time, it appears that ethical standards did not prohibit sexual relations between a psychiatrist and a former patient. Id. Under the standards applicable in 1985, the sexual relationship would have been unethical only if the doctor used the professional relationship as an enticement to sexual contact.

The case against the doctor depended substantially on the patient's testimony. In our earlier opinion, we noted that the hearing officer who submitted a recommended decision to the board, which it accepted, had rejected significant portions of the patient's testimony but accepted others. The rejected testimony, if believed, would have warranted a finding that the doctor-patient relationship existed when the doctor and the patient had sexual relations at the patient's apartment in October and November, 1985. In our first opinion, we stated that, "[i]n passing on the credibility issue, the board could not properly ignore the fact that it was declining to credit significant portions of the patient's testimony, some of which, as has been noted, was susceptible of an inference of fabrication aimed at establishing that sexual relations occurred during the physician-patient relationship." Herridge I, supra at 163-164. We vacated the board's decision and remanded the matter to the board for additional proceedings. Id. at 167. We did so because, "[i]n the face of conflicting evidence, and inconsistencies in the patient's testimony, the board could not choose to rely on portions of that testimony, reject other, significant portions of it, and fail to explain its reasons for doing so. The board was required to confront the problems in the patient's testimony and to provide `an explicit analysis of credibility and the evidence bearing on it.'" Id. at 164-165, quoting Morris v. Board of Registration in Medicine, 405 Mass. 103, 107, cert. denied, 493 U.S. 977 (1989).

The board then recommitted the matter to the original hearing officer who prepared a proposed amended final decision. The board adopted the proposal, which reaffirmed the board's previous position. Although we shall refer to the decision as the board's decision, the decision was that of the hearing officer to whom the credibility of witnesses was committed. The doctor again appealed, and a single justice of this court reserved and reported the matter to the full court.

1. We have reviewed the board's amended decision and have considered the various grounds on which the doctor now challenges the board's credibility determinations.

Any suggestion that the patient's testimony was consistent and generally trustworthy is belied by the fact the board disbelieved parts of her testimony that, if believed, would have been very damaging to the doctor's position. The amended decision presents no sufficient explanation as to why the board accepted some of the patient's damaging testimony as credible, while rejecting other significant portions of it. The amended decision is unpersuasive on the issues remanded to it for consideration. We set forth examples of why the amended decision fails to convince us that there was substantial evidence justifying the imposition of professional discipline.

a. In its original decision the board rejected the patient's testimony that she had had a final office appointment with the doctor in February, 1986. The existence of such an appointment would have been significant because the admitted sexual relations between the doctor and the witness had occurred prior to that date, and thus there would have been improper professional conduct during the doctor-patient relationship. In our first opinion, we said that the board's finding that the doctor-patient relationship had ended on October 8, 1985, required the board "to reject a considerable portion of the patient's testimony bearing on the duration of the physician-patient relationship." Herridge I, supra at 161.

In its amended opinion, the board treated the matter of the alleged February, 1986, appointment more fully, suggesting that the patient may have recalled an appointment in some other winter. The amended opinion erroneously states that the patient "could recall no details" or "specifics" about the appointment, pointing out that, if it had occurred, she would have recalled details because the session would likely "have been memorable since it occurred three months after the termination of her sexual relationship with the [doctor]." In fact, the patient testified several times that the appointment occurred in February after an October, 1985, dinner engagement with the doctor, and after the sexual relations in October and November, 1985. She testified to how the appointment may have been set up and the room in which it took place, and stated that the doctor and she discussed her mood and medications. We thus have the former patient testifying to details of an appointment that the board found never occurred, and the board denying the existence of the detailed false testimony about the appointment, and thus in this respect failing "to provide a thorough and reasoned explanation for the decision to credit the patient's testimony." Herridge I, supra at 156.

If the board believed the doctor that there had been no February, 1986, appointment, thereby finding the patient's testimony not credible, it was incumbent on the board to explain why it believed the patient, and not the doctor, in other circumstances.

b. The credibility issue is compounded by the fact that the board in its original decision rejected the patient's testimony that a social dinner engagement occurred before her final office visit, and stated that her recollection was "somewhat confused." The board had credited the doctor's testimony that the dinner occurred after the patient's last (October 8, 1985) appointment. In its amended decision, however, the board eliminates the reference to the patient's "confusion" and provides no explanation as to why it concluded that the patient's testimony was not credible on this point but acceptable elsewhere.

c. The patient testified that the doctor continued to consult with her by telephone about her use of medication through February, 1986. In its original decision, the board made no findings on this point. On remand, the board made new findings. It found that there had been telephone conversations after the last (October 8, 1985) appointment. The doctor testified that he typically made a record of patient telephone calls, and that he only had a record of a May, 1985, telephone call from the patient. He also testified that he did not treat patients over the telephone, but that patients could telephone him if they liked. The board characterized the doctor's testimony as "evasive and inconsistent," but it does not explain why, and, on its face, it appears to be neither. The board found that the doctor's "testimony does not negate the evidence that he told his patients to call between sessions, that [the patient] did so" (emphasis supplied), and that they talked about her treatment. The issue before the board was not treatment by telephone between sessions, but rather whether there was treatment by telephone after the October 8, 1985, office appointment. The patient gave no details about the number of and the reasons for the telephone calls. The record fails to show any basis for concluding that there were treatment-related telephone calls after October 8, 1985, that would support the patient's alleged perception that a doctor-patient relationship continued after that date.

d. The hearing officer's task on remand from the board was to explain the credibility determinations in her original recommended decision. It is not clear that her task included the making of new findings. In her original decision, the hearing officer declined to decide whether a kissing incident occurred during one of the patient's appointments, as testified to by the patient. In our first opinion, we noted that the board could not ignore the fact that it had declined to credit testimony of the patient that "was susceptible of an inference of fabrication aimed at establishing that sexual relations occurred during the physician-patient relationship." Herridge I, supra at 163-164. On remand, the board, for the first time, made a finding that some physical contact occurred during the October 8, 1985, appointment. The patient's original complaint did not allege any physical intimacy occurred in the doctor's office. In her testimony, the patient gave arguably inconsistent statements about the circumstances in which the kissing incident assertedly occurred. In its amended decision, however, the board states that the patient never said where the doctor was sitting when the alleged kissing occurred and that she was not asked about it. That statement is patently incorrect. See Administrative Record, Volume II, pp. 193, 258. Moreover, the board has mischaracterized the doctor's testimony about the physical characteristics of his office.

The fact that the patient did not fabricate other instances of physical intimacy does not, contrary to the board's conclusion, establish that the patient was truthful about the kissing incident. Contrary to the board's view, her testimony about this incident was not consistent. She testified that it occurred at a time that the board rejected. The board's reliance on the "tone and quality" of her complaint letter to bolster a finding of the patient's general credibility is curious because that letter made no reference to the kissing incident. Moreover, it is doubtful that the tone and quality of a complaint letter can have much bearing on whether subsequent testimony under oath is credible.

2. Our review of the board's attempted resolution of the conflicting testimony on the issues that we have discussed, and on others the discussion of which would unnecessarily prolong this opinion, leads us to the firm conclusion that the board failed "to confront the problems in the patient's testimony and to provide `an explicit analysis of credibility and the evidence bearing on it.'" Herridge I, supra at 165, quoting Morris v. Board of Registration in Medicine, 405 Mass. 103, 107 (1989). The patient's testimony was essential to the proof of facts justifying suspension of the doctor's right to renew his license to practice medicine in the Commonwealth. Herridge I, supra at 166. The board has failed to provide an adequate explanation as to why it has treated parts of the patient's testimony as credible when it has rejected as untrue so much of that testimony. The amended decision does not adequately follow the instructions we gave in our review of the board's original decision.

We do not offer the board another chance to explain its credibility determinations. If the task was ever capable of performance, it is not now. The amended final decision of the Board of Registration in Medicine is vacated. The board should enter an order dismissing the proceeding that is the subject of this appeal.

So ordered.