OPINION

BethEnergy Mines, Inc. (BethEnergy) petitions for review of the decision of the U.S. Department of Labor’s Benefits Review Board (BRB) affirming an award of benefits to David Cunningham under the Black Lung Benefits Act (the Act), 30 U.S.C. § 901 et seq. BethEnergy first argues that Cunningham’s claim was not timely filed, but we conclude that the company has waived that issue. BethEnergy also raises evidentiary challenges to the award of benefits, which we reject. We therefore affirm the award of benefits and deny BethEnergy’s petition for review.

I.

David Cunningham is a retired coal miner. He worked underground for 28 years, beginning in the early 1950s. In 1982 the mine where Cunningham worked shut down, and he was out of work for five years before being called back in 1987. He last worked as a mine mechanic, a job that required heavy labor. Cunningham quit in 1991 after shortness of breath made it difficult for him to do his job. He was 58 years old at the time.

Cunningham filed his first claim for black lung benefits with the Department of Labor in 1986, while he was out of work. A claims examiner in the Department’s Office of Workers’ Compensation Programs (OWCP) denied the claim later that year, and Cunningham made an untimely attempt to appeal. The adjudication of Cunningham’s first claim is therefore final. He filed his second.claim for benefits in July 1997; that claim, which BethEnergy *883contested, made its way to a hearing before an ALJ. The ALJ found that Cunningham had established a material change in condition, the existence of a totally disabling respiratory or pulmonary condition, and the existence of pneumoconiosis (by chest x-ray and medical opinion evidence) that was a contributing cause of his total disability. The ALJ thus awarded benefits. BethEnergy appealed the award to the BRB, and Cunningham cross-appealed the ALJ’s determination of the onset date. The BRB affirmed the award of benefits, but remanded for reconsideration of the issue of onset date. On remand the ALJ issued a second order awarding benefits and establishing an earlier onset date. BethEnergy again appealed to the BRB, which affirmed the ALJ. BethEnergy now petitions this court for review, asserting that Cunningham’s claim is untimely and, in the alternative, that the ALJ erred in finding the existence of coal workers’ pneumoconiosis and total disability due to pneumoconiosis. In considering the petition, we “undertak[e] an independent review of the record to determine whether the ALJ’s findings of fact were supported by substantial evidence.” Island Creek Coal Co. v. Compton, 211 F.3d 203, 207 (4th Cir. 2000) (internal quotation marks omitted). The legal conclusions of the ALJ and the BRB are reviewed de novo. Id. at 208.

II.

BethEnergy first argues that Cunningham’s 1997 claim must be rejected as untimely because it was not filed within three years after he learned that he had been medically diagnosed as totally disabled due to pneumoconiosis. See 30 U.S.C. § 932(f); 20 C.F.R. § 725.308(a) (“[a] claim” must be filed “within three years after a medical determination of total disability due to pneumoconiosis which has been communicated to the miner.”). We conclude that BethEnergy has waived the statute of limitations argument because it stipulated at the first hearing before the ALJ that Cunningham’s claim was timely. Administrative waiver is a “flexible” concept. Rana v.United States, 812 F.2d 887, 889 n.2 (4th Cir. 1987). When a litigant presents an issue to us that is “within the purview of agency expertise,” without having first properly presented the issue to the agency, we will normally regard the issue to be waived. Rana, 812 F.2d at 890. This makes sense because one of the reasons for administrative waiver is to encourage the exhaustion of administrative remedies. Thom v. Itmann Coal Co., 3 F.3d 713, 717 (4th Cir. 1993). Exhaustion is important because “it is normally desirable to let the agency develop the necessary factual background upon which decisions should be based,” and “judicial review may be hindered by the failure of the litigant to allow the agency to make a factual record.” McKart v. United States, 395 U.S. 185, 194, 89 S.Ct. 1657, 23 L.Ed.2d 194 (1969). Moreover, “practical notions of judicial efficiency come into play as well” because if a litigant “is required to pursue his administrative remedies, the courts may never have to intervene.” Id. at 195. Adequate administrative consideration also gives the agency “a chance to discover and correct its own errors.” Id. Finally, the reasons for requiring “administrative issue exhaustion” as a prerequisite to appellate review are most compelling when “the parties are [competing] in an adversarial administrative proceeding” instead of participating in a proceeding that is not adversarial. Sims v. Apfel, 530 U.S. 103, 109-10, 120 S.Ct. 2080, 147 L.Ed.2d 80 (2000).

In black lung cases, developing a factual record and making findings of fact is delegated to the ALJ. The respect afforded to an ALJ’s factual findings is reflected by *884the rule that we (and the BRB) affirm the ALJ’s factual findings so long as they are supported by substantial evidence. Convp-ton, 211 F.3d at 207 (appellate review standard); 30 U.S.C. § 932(a), incorporating 33 U.S.C. § 921(b)(3) (BRB review standard). Moreover, ALJ hearings in black lung cases are adversarial. Dep’t of Labor v. Triplett, 494 U.S. 715, 733-34, 110 S.Ct. 1428, 108 L.Ed.2d 701 (1990) (Marshall, J„ concurring in the judgment). In this case, rather than participating in the development of a factual record on the timeliness of Cunningham’s claim, BethEnergy stipulated that the claim was timely at the beginning of the first hearing before the ALJ. The ALJ then relied on the stipulation to limit the hearing to consideration of contested issues, which did not include any statute of limitations issue. BethEnergy compounded the problem by not raising the timeliness issue when the case was first appealed to the BRB and by failing to ask the ALJ to reopen the record on this issue on remand. BethEnergy finally raised the timeliness issue on its second appeal to the BRB, when the BRB applied its rule that the statute of limitations does not apply to second (or duplicate) claims like Cunningham’s. See Faulk v. Peabody Coal Co., 14 BLR 1-18 (1990); Andryka v. Rochester & Pittsburgh Coal Co., 14 BLR 1-34 (1990).

We conclude that BethEnergy waived the timeliness argument because its stipulation prevented the agency from adequately addressing it. We are hindered in our review of the BRB’s decision because the Board never had an opportunity to reexamine its legal rule on timeliness in the context of all of the facts that would bear on whether Cunningham’s claim was filed on time. See McKart, 395 U.S. at 194-95. We would be in a better position to consider the limitations issue with respect to a duplicate claim if we had a developed set of facts. BethEnergy does assert that a couple of documents that happen to be in the record would support a finding that Cunningham’s claim was not timely filed. However, it is the ALJ’s job in the first instance, not ours, to say what those documents mean (specifically, whether they constitute a medical determination of total disability due to pneumoconiosis that was communicated to Cunningham). See Compton, 211 F.3d at 207. BethEner-gy’s failure to bring these documents to the ALJ’s attention in either hearing supports our conclusion that BethEnergy has waived this issue. Indeed, BethEnergy suggests that if we apply the statute of limitations to Cunningham’s claim and “question still exists ... the case should be remanded for the purpose of examining Mr. Cunningham about the timeliness of his 1997 application.” Pet. Br. at 31. The company’s suggestion illustrates the uncertain position we would be in if we considered the timeliness issue: we would be reviewing an agency decision without a fully developed record or findings of fact. We may invoke waiver here because the inadequacy of the record was caused by the stipulation of BethEnergy, a represented party, during adversary proceedings. Cf. Sims, 530 U.S. at 110 (holding that the issue exhaustion requirement is more appropriately invoked when the administrative proceedings were adversarial); Rana, 812 F.2d at 889 n. 2 (discussing how pro se litigants are held to a more lenient waiver standard).

Considerations of judicial efficiency are also relevant. McKart, 395 U.S. at 195. If BethEnergy had explored whether Cunningham’s claim was timely filed during the proceedings before the ALJ — for instance by questioning Cunningham about when he was told he was totally disabled by pneumoconiosis — we would know how the facts stack up for each side, and we would be in a better position to decide *885whether the company could assert a valid limitations defense. Instead, we are being asked to intervene now — in a case that has been litigated for seven years — to address the BRB’s timeliness rule just because BethEnergy might be able to show in a third hearing before the ALJ that Cunningham’s claim was not timely filed. See id.

Finally, BethEnergy argues that we should not hold the timeliness issue waived because it would have been futile to raise it before the agency in light of the BRB’s rule. We do not agree. We simply do not know whether the BRB would have reconsidered its rule if BethEnergy had developed a compelling factual record showing that Cunningham was advised of a medical determination of total disability more than three years before he filed his current claim. This is especially so because the Director of OWCP takes the position that the BRB’s rule is incorrect. Moreover, the two circuits that have considered the limitations issue have avoided application of the BRB’s timeliness rule. See Wyoming Fuel Co. v. Director, O.W.C.P., 90 F.3d 1502, 1507 n. 7 (10th Cir. 1996); Sharondale Corp. v. Ross, 42 F.3d 993, 996 (6th Cir. 1994); see also Tenn. Consolidated Coal Co. v. Kirk, 264 F.3d 602, 606-07 (6th Cir. 2001) (affirming an award of benefits where the BRB did not attempt to apply its rule).* We do not believe, in other words, that the BRB had a full opportunity to evaluate its rule in the context of Cunningham’s facts, and that is a factor in favor of finding a waiver. See McKart, 395 U.S. at 195. In any event, regardless of whether the BRB would have revisited its rule, BethEnergy was still “required to exhaust the fact-finding mechanisms” of the agency so that we could consider the issue in the context of a developed set of facts. Dooley v. Ploger, 491 F.2d 608, 615 (4th Cir. 1974).

In sum, because the policy reasons for requiring administrative exhaustion indicate that this is an appropriate case to find waiver, and because we appear to have an inadequate factual record as a result of BethEnergy’s stipulation to timeliness in the adversary proceedings before the ALJ, we conclude that BethEnergy has waived its statute of limitations defense.

III.

Because the statute of limitations defense is waived, we turn to BethEnergy’s substantive challenges to the BRB’s decision to affirm the ALJ’s award of benefits. To receive an award of black lung benefits the claimant must prove by a preponderance of the evidence that “(1) he has pneumoconiosis; (2) the pneumoconiosis arose out of his coal mine employment; (3) he has a totally disabling respiratory or pulmonary condition; and (4) pneumoconiosis is a contributing cause to his total respiratory disability.” Milburn Colliery Co. v. Hicks, 138 F.3d 524, 529 (4th Cir. 1998). BethEnergy makes two main arguments for vacating the BRB’s affirmance of the award. First, the company argues that the ALJ erred by giving the most weight to the x-ray interpretations of “dually *886qualified” physicians. Second, it argues that the ALJ failed to follow Island Creek Coal Co. v. Compton, 211 F.3d 203 (4th Cir. 2000), when he did not weigh together all the evidence in the record bearing on the issue of whether Cunningham has pneumoconiosis. For the reasons that follow, we reject all of BethEnergy’s arguments directed at the ALJ’s substantive decision to award benefits, and we affirm that award.

A.

BethEnergy first argues that the ALJ erred by mechanically deferring to the opinions of “dually qualified” x-ray readers. The term “dually qualified” physician is not defined in the regulations, but the ALJ here used that term to refer to physicians that were both B-readers and board-certified in radiology. A B-reader is a physician who has passed a specific examination testing his or her ability to classify x-rays for the presence and severity of pneumoconiosis. See 20 C.F.R. § 718.202(a)(l)(ii)(E). A board-certified radiologist is certified “in radiology or diagnostic roentgenology by the American Board of Radiology, Inc. or the American Osteopathic Association.” Id. § 718.202(a)(l)(ii)(C). The regulations provide that in evaluating conflicting x-ray reports, “consideration shall be given to the radiological qualifications of the physicians” that interpreted the x-rays. Id. § 718.202(a)(1) (emphasis added). The regulations do not establish a hierarchy of radiological qualifications, but by defining the terms B-reader and board certified radiologist, they suggest that those are relevant qualifications. See also id. § 718.102(c) (generally requiring x-ray reports to include the qualifications of the reader, but specifically requiring board certified radiologists and B-readers to indicate those qualifications).

In Cunningham’s case there were 18 recent x-rays, read a total of 27 times. Thirteen of the readings were positive for pneumoconiosis. The ALJ refused to consider three positive readings because the reader, who was dually qualified, acknowledged that she could not accurately assess the presence of pneumoconiosis due to Cunningham’s “vascular congestion.” The ALJ then rejected three more readings because there was no record of the identity of the readers or the degree of pneumoco-niosis found. He also gave little weight to another reading because the reader failed to classify the film. Finally, the ALJ noted that the record contained two negative readings of a film from 1986, but he gave these readings “little weight” because Cunningham experienced significant exposure to coal mine dust after the x-ray was taken. The ALJ then evaluated the radiological qualifications of the physicians reading the remaining x-rays, giving the “most weight” to dually qualified physicians and “great weight” to the interpretations of B-readers. Applying this evaluation of the relative qualifications of the readers, the ALJ concluded that two of the x-rays were positive for pneumoconiosis because they had been read as such by dually qualified physicians and there were no negative readings of those x-rays. He also concluded that a third x-ray was positive for pneumoconiosis because it was read as positive by two dually qualified readers and negative by the other readers, all of whom were only B-readers. Three physicians read the final x-ray as negative; two of the physicians were B-readers and the third was neither a B-reader nor a board certified radiologist. The ALJ concluded that the x-ray evidence as a whole, “and in particular the positive readings by the most qualified readers of record,” established the existence of pneumoconiosis by a preponderance of the x-ray evidence.

The regulations require ALJs to consider the radiological qualifications of the *887physicians when x-ray readings conflict. 20 C.F.R. § 718.202(a)(1). Although the dually qualified physicians and B-readers have both passed examinations testing their proficiency in classifying x-rays for pneumoconiosis, the dually qualified physicians have board certifications in radiology or diagnostic roentgenology that the physicians that are only B-readers lack. As a result, the ALJ’s finding that these dually qualified physicians were more qualified than the other x-ray readers was supported by substantial evidence. See, e.g., Zeigler Coal Co. v. Director, O.W.C.P., 326 F.3d 894, 898-99 (7th Cir. 2003) (affirming award of benefits where ALJ gave greater weight to similar dually qualified physicians). Moreover, the ALJ did not mechanically defer to readings by dually qualified physicians. Rather, he gave careful consideration to several relevant factors and even rejected three readings by a dually qualified physician. We conclude that the ALJ did not err in his treatment of the x-ray readings made by dually qualified physicians.

B.

BethEnergy also argues that we should vacate the benefits award under Island Creek Coal Co. v. Compton, 211 F.3d 203, 209-11 (4th Cir. 2000), because the ALJ should have weighed all of the relevant evidence together before deciding that Cunningham has pneumoconiosis. A miner may prove the first element of his claim, that he has pneumoconiosis, by any of the four methods described in 20 C.F.R. § 718.202(a). We consider only the two methods, chest x-ray evidence, see id. § 718.202(a)(1), and physician opinion evidence, see id. § 718.202(a)(4), used in this case. After careful consideration of the conflicting evidence in each category, the ALJ found by a preponderance of both the x-ray evidence and the physician opinion evidence that Cunningham has pneumoconiosis. The ALJ did not explicitly weigh together the x-ray evidence with the physician opinion evidence. However, in evaluating the physician opinion evidence, the ALJ did point out that the opinions he found most credible were “in accordance with the objective medical evidence of record.” J.A.734.

In Island Creek Coal Co. v. Compton the ALJ found that the claimant did not establish that he had pneumoconiosis by a preponderance of the x-ray evidence. However, the ALJ found that a preponderance of the physician opinion evidence did establish the existence of pneumoconiosis. 211 F.3d at 207. The ALJ then awarded benefits without weighing the conflicting x-ray and physician opinion evidence. Id. at 208. The BRB approved, reasoning that benefits can be awarded so long as an ALJ finds the existence of pneumoconiosis by a preponderance of the evidence in one of the categories. Id. We concluded, however, that the BRB’s legal analysis was erroneous because the Act “states that ‘in determining the validity of claims ... all relevant evidence shall be considered.’ ” Id. (quoting 30 U.S.C. § 923(b)) (alteration omitted). We further reasoned that “weighing all of the relevant evidence together makes common sense. Otherwise, the existence of pneumoconiosis could be found even though the evidence as a whole clearly weighed against” finding that a miner has pneumoconiosis. Id. at 209.

This case is not like Compton because the ALJ here found that the evidence in each of the two categories offered, x-ray and physician opinion, established pneumo-coniosis. Assuming, however, that it is error under Compton when an ALJ fails to weigh two categories of evidence together after each one leads to a finding of pneu-moconiosis, we conclude that the error is harmless in this case. “[W]hen a mistake of the administrative body is one that clearly had no bearing on ... the substance of [the] decision reached,” that er*888ror is harmless. Mass. Trustees of E. Gas & Fuel Assocs. v. United States, 377 U.S. 235, 248, 84 S.Ct. 1236, 12 L.Ed.2d 268 (1964). Because the ALJ found that the most credible x-ray and opinion evidence each showed the presence of pneumoconio-sis, it would be irrational for him to conclude on remand that “the evidence as a whole ... weighed against” finding that Cunningham has pneumoconiosis. Compton, 211 F.3d at 209. In sum, because any failure by the ALJ to weigh the evidence together in this case could not have influenced the outcome, we will not remand for application of Compton.

C.

We have considered BethEnergy’s other arguments — that the evidence is insufficient to establish that Cunningham’s disability was due to pneumoconiosis and that the ALJ failed to explain adequately his reasons for crediting evidence of pneumo-coniosis and disability due to pneumoconio-sis. As to these arguments, we affirm on the reasoning of the BRB. Cunningham v. BethEnergy Mines, Inc., BRB Nos. 00-0841 BLA and 00-0841 BLA/A.

IV.

The award of benefits is affirmed, and the petition for review is denied.

PETITION FOR REVIEW DENIED.

concurring in part and concurring in the judgment.

I concur in parts I, III (with the exception of the first sentence), and IV of the majority opinion. I concur only in the judgment, however, as to Part II. Unlike the majority, I do not believe that BethEn-ergy waived the issue of timeliness because, given the Benefit Review Board’s (BRB) pronouncement that the three year statute of limitations does not apply to duplicate claims, see Faulk v. Peabody Coal Co., 14 BLR 1-18 (1990), it would have been futile for BethEnergy to have raised the issue. Furthermore, the BRB addressed BethEnergy’s argument that the statute of limitations did apply to duplicate claims, and it rejected that argument in continuing to adhere to Faulk.

I agree with the Director, who’s interpretation of regulations promulgated under the Black Lung Benefits Act is owed substantial deference, see Clinchfield Coal Co. v. Harris, 149 F.3d 307, 309 (4th Cir. 1998), that “[t]he plain language of the statute and the implementing regulation demonstrate that the limitations period applies to all claims filed by a miner.” (Federal Respondent’s Br. at 9.) I concur in the judgment, however, because, applying the standard advocated by the Director, Beth-Energy has failed to show that Cunningham did not file his claim within three years of a “medical determination of total disability due to pneumoconiosis which ha[d] been communicated to the miner.” 20 C.F.R. § 725.308(a) (2003).