This action involves a class challenge of the Social Security Administration’s regulations, policies, and practices as they relate to the role of pain in the evaluation of disability claims. The named plaintiffs and intervenors sought class relief on the ground that agency decision makers have been systematically denying benefit claims based on disabling pain by improperly requiring objective corroboration of the severity of pain. The district court granted summary judgment for plaintiffs, holding the Secretary’s regulations misleading, overbroad, and invalid on their face. Luna v. Bowen, 641 F.Supp. 1109, 1113 (D.Colo. 1986). The court also held that the Secretary’s policy and practice in evaluating pain is inconsistent with the Social Security Dis*162ability Reform Act of 1984, 42 U.S.C. § 423(d)(1)(A) (Supp.III 1985). Id. at 1120. The Secretary has appealed both holdings and the remedy imposed. We reverse.1
I.
The district court held the Secretary’s regulations facially invalid because they require the claimant to show “objective” medical evidence of a pain-producing impairment before a decision maker may find a claimant disabled by pain. The court asserted that since the statute permits benefit awards based on “subjective” psychological evidence of pain-producing impairments, the Secretary’s apparent exclusive reliance on “objective” physiological evidence is invalid. The court recognized that the regulations discuss both physiological and psychological sources of pain; however, it concluded that certain passages referring only to “objective” findings misleadingly imply that only the physiological evidence really counts. Both parties essentially agree that this holding is unsupportable.
Although the district court correctly determined that both physical and mental impairments can support a disability claim based on pain, Turner v. Heckler, 754 F.2d 326, 330 (10th Cir.1985), the court incorrectly concluded that objective evidence is limited to concrete physiological data. The term “objective” in this context refers instead to any evidence that an examining doctor can discover and substantiate. Both physiological and psychological medical evidence is objective, because each is amenable to external testing.2 The regulations list both types of evidence as equivalent medical bases for the evaluation of pain, and the Secretary argues that it simply used the term “objective” as a short-hand way of referring to them. 20 C.F.R. §§ 404. 1528, 416.928 (1987).
The Secretary’s position is not novel. No circuit has noted any ambiguity in this area. See, e.g., Cotton v. Bowen, 799 F.2d 1403 (9th Cir.1986); Avery v. Secretary of Health and Human Services, 797 F.2d 19 (1st Cir.1986); Landry v. Heckler, 782 F.2d 1551 (11th Cir.1986); Foster v. Heckler, 780 F.2d 1125 (4th Cir.1986); Nieto v. Heckler, 750 F.2d 59 (10th Cir.1984); Green v. Schweiker, 749 F.2d 1066 (3d Cir.1984). In addition, we have recognized that psychological abnormality shown by “ ‘medical signs and findings, established by medically acceptable clinical or laboratory diagnostic techniques’ ” must be given as much consideration as similarly demonstrated evidence of physical impairments. Teter v. Heckler, 775 F.2d 1104, 1106 (10th Cir. 1985) (quoting 42 U.S.C. § 423(d)(5)(A). (Supp.III 1985)). Our language is tantamount to a holding that objective medical evidence can be physiological, psychological, or both.
In light of our clear pronouncements on the issue, we conclude that the Secretary’s use of the term “objective” in his regulations to refer to all measurable medical findings will not confuse lower level decision makers, and therefore this phrasing does not render the regulations facially invalid. Accordingly, we reverse the district court’s holding to the contrary.
II.
The district court additionally held that decision makers improperly rely primarily on objective medical evidence after finding a pain-producing impairment. This conclusion presents a more vexing problem. It hinges in part on the court’s erroneous belief that the Secretary does not treat psychological evidence as objective. We, therefore, cannot know whether the court would have held the Secretary’s practice improper if the court had correctly con*163strued the scope of the phrase “objective evidence” in the regulations.
Because the court’s definition of objective was incorrect, we cannot sustain its holding. See Part III, infra. Nevertheless, the parties have presented us with a legal issue as to the interpretation of the governing statute. In order to provide a proper framework for this case on remand, we will address that legal argument.
Neither the statute, the regulations, nor other agency rulings and instructions clearly describe how much weight a decision maker must give to subjective allegations of pain. Although we find little meaningful guidance in the agency’s documents, some objective showing is certainly a prerequisite. We have recognized the statute requires that a pain-producing impairment, whether psychological or physiological in origin, must be proven by objective medical evidence before an agency decision maker can find a claimant disabled by pain. Frey v. Bowen, 816 F.2d 508, 515 (10th Cir.1987); Nieto, 750 F.2d 59. Other courts unanimously agree. See, e.g., Cotton, 799 F.2d 1403; Avery, 797 F.2d 19; Landry, 782 F.2d 1551; Foster, 780 F.2d 1125; Green, 749 F.2d 1066. The issue in this case, however, is what the decision maker must do after finding that a pain-producing impairment exists.
In his brief, the Secretary describes the subsequent process as follows: If a pain-producing impairment is demonstrated by objective medical evidence, the decision maker must consider the relationship between the impairment and the pain alleged. “[T]he impairment or abnormality must be one which ‘could reasonably be expected to produce’ the alleged pain.” Brief for the Appellant at 16 (emphasis added). At this stage, the decision maker takes the subjective allegations of pain as true in determining whether they are reasonably related to the proven impairment. He does not evaluate the claimant’s credibility. If the nexus between impairment and pain alleged is insufficient, the claimant cannot receive benefits based upon disabling pain. If an appropriate nexus does exist, the decision maker must then consider all the evidence presented to determine whether the claimant’s pain is in fact disabling. Id. at 16-17. This evidence includes the medical data previously presented, any other objective indications of the degree of the pain, and subjective accounts of the severity of the claimant’s pain. Only at this point may the decision maker decide whether he believes the claimant’s assertions of severe pain.
No one appears to dispute this formulation of the decision maker’s role. The precise legal issue before us, however, is to determine the necessary nexus between the impairment shown and the pain alleged. In other words, what is the standard of reasonableness that the Secretary must use to determine whether the impairment is one that could “reasonably” be expected to produce the alleged disabling pain. In not focusing on the relationship between the impairment proven and the pain alleged, the parties’ arguments appear to pass in the night. When we reexamine them in light of the above framework, the dispute becomes more clear.
The Secretary attempts to truncate our inquiry by asserting that Congress intended to codify the agency’s existing policies regarding pain in amending 42 U.S.C. § 423(d)(5). The agency, the Secretary claims, has continued to determine whether a particular impairment can reasonably be expected to produce the alleged pain in exactly the same fashion as it did before the statute codified its regulations. Thus, agency decision makers must be acting consistently with the legislative mandate. This argument begs the question. Although Congress may have intended to adopt the agency’s existing policies, the Secretary must nevertheless show that its particular statements and instructions lead to individual decisions that are in conformity with the statute. Avery, 797 F.2d at 21.
Plaintiffs argue that decision makers are improperly relying primarily on objective medical data. In essence, plaintiffs claim the Secretary is finding that a proven impairment is not sufficiently related to the pain alleged in cases where a sufficient relationship does exist under the Act. Plaintiffs argue that once a pain-producing *164impairment has been proven, the Secretary, in primarily relying on objective medical evidence to determine whether an impairment is “reasonably” likely to produce disabling pain, places too great a burden on the claimant to demonstrate a causal relationship between the impairment established and the severity of the pain alleged.
The Secretary, however, has repeatedly maintained that all denials are based on a determination that the medical evidence was inconsistent with the subjective allegations. In so doing, the Secretary must be equating consistency with the statutory requirement of a reasonable relationship in order to comply with the statutory mandate. His articulation, however, does not respond to the issue before us because it does not define the controlling statutory standard that decision makers are applying.
The issue here is clearly one of degree. The statute requires the claimant to present objective medical evidence that shows an impairment
“which could reasonably be expected to produce the pain or other symptoms alleged and which, when considered with all evidence required to be furnished under this paragraph (including statements of the individual or his physician as to the intensity and persistence of such pain or other symptoms which may reasonably be accepted as consistent with the medical signs and findings), would lead to a conclusion that the individual is under a disability.”
42 U.S.C. § 423(d)(5)(A). Under this standard, a claimant certainly must demonstrate more than a potentially pain-causing impairment before the decision maker must weigh all other evidence. For example, an impairment that would be expected to cause pain in the lower extremeties would not be reasonably expected to produce disabling pain in the upper body. See Avery, 797 F.2d at 21.
The more difficult issue is how tight a relation the statute requires between the impairment and the allegation. Sufficient evidence of congressional intent exists to convince us that the statute requires only a loose nexus between the proven impairment and the pain alleged. Accordingly, if an impairment is reasonably expected to produce some pain, allegations of disabling pain emanating from that impairment are sufficiently consistent to require consideration of all relevant evidence. See Frey, 816 F.2d at 515 (“the law has never required and does not now require that medical evidence identify an impairment that makes the pain inevitable”); Nieto, 750 F.2d at 62 (decision maker abuses discretion in rejecting medical opinions on basis of objective test results alone); accord Cotton, 799 F.2d at 1407 (“ ‘Congress clearly meant that so long as the pain is associated with a clinically demonstrated impairment, credible pain testimony should contribute to a determination of disability’ ”) (quoting Howard v. Heckler, 782 F.2d 1484, 1488 n. 4 (9th Cir.1986)); Avery, 797 F.2d at 21 (statute permits awarding disability based on pain where subjective statements consistent with objective findings convince decision maker to find the claimant disabled even though objective evidence alone was insufficient); MacGregor v. Bowen, 786 F.2d 1050, 1054 (11th Cir.1986); Green, 749 F.2d at 1070 (“dismissal of subjective symptomology on the basis of an absence of direct medical evidence is at odds with ... the new statute”).3 For example, an impairment likely to produce some back pain may reasonably be expected to produce severe back pain in a particular claimant.
*165In amending § 423(d)(5), Congress intended to codify existing policy. The Secretary’s clearest elaboration of existing policy in regard to the nexus between the objective impairment and the alleged pain is found in a settlement agreement quoted in Polaski v. Heckler, 751 F.2d 943, 948 (8th Cir.1984), vacated and remanded on other grounds, — U.S. —, 106 S.Ct. 2885, 90 L.Ed.2d 974 (1986). In this settlement, the Secretary asserted that she was clarifying and restating the policies that her decision makers should have been (and presumably were in most cases) following all along. Id. at 949. In this clarification of existing policy, the Secretary stated
“Symptoms can sometimes suggest a greater severity of impairment than is demonstrated by objective and medical findings alone_ [DJirect medical evidence of the cause and effect relationship between the impairment and the degree of claimant’s subjective complaints need not be produced.... The absence of an objective medical basis which supports the degree of severity of subjective complaints alleged is just one factor to be considered in evaluating the credibility of the testimony and complaints. The adjudicator must give full consideration to all of the evidence presented relating to subjective complaints....”
Id. at 948 (emphasis added). According to the Secretary’s own account, the absence of an objective medical basis for the degree of severity of pain may affect the weight to be given to the claimant’s subjective allegations of pain, but a lack of objective corroboration of the pain’s severity cannot justify disregarding those allegations.
By requiring consideration of evidence other than objective medical data, Congress recognized that the severity of pain is inherently subjective. A relatively expansive interpretation of “consistent” or “reasonably expected” is the only logical reading of the statute, because a more restrictive reading would render meaningless that part of it requiring consideration of all evidence. If objective medical evidence must establish that severe pain exists, subjective testimony serves no purpose at all. A proper reading of the statute appropriately recognizes that two patients with the same impairment may be affected to totally differing degrees.
In amending section 423 and codifying an objective evidence requirement, Congress certainly intended to help alleviate the tremendous administrative burden borne by the social security system in determining who is in fact disabled by pain. Because one cannot conclusively prove the severity of an individual’s pain through medical test results, however, Congress stopped short of requiring medical evidence of severity. Rather, the decision maker must consider all the evidence presented that could reasonably produce the pain alleged once a claimant demonstrates a pain-causing impairment. Clearly, Congress believed that this scheme would reduce the administrative burden without permitting the Secretary to deny benefits automatically to those with objectively proven impairments that usually (but not invariably) cause only non-disabling pain.
Requiring consideration of evidence beyond laboratory and test results whenever a loose nexus is established between the pain-causing impairment and the pain alleged does not force decision makers to decide cases based solely on their evaluation of the credibility of the wholly subjective statements of the claimant. Although pain is subjective in that no one can know how much of it a particular nerve stimulus causes another person, many indicators give us some idea how much pain another is feeling. Anyone who has visited the dentist knows that one who is able to have his teeth filled without novocain is less sensitive to pain than one who requires a shot of the drug. Similarly, those with identical impairments may require vastly different painkillers. The one who needs only aspirin surely feels less pain than the one requiring a much stronger drug. In previous cases, we have recognized numerous factors in addition to medical test results that agency decision makers should consider when determining the credibility of subjective claims of pain greater than that usually associated with a particular impairment. For example, we have noted a claimant’s persistent attempts to find relief for his pain and his willingness to try any treatment prescribed, regular use of *166crutches or a cane, regular contact with a doctor, and the possibility that psychological disorders combine with physical problems. See Turner, 754 F.2d at 331; Nieto, 750 F.2d at 62. The Secretary has also noted several factors for consideration including the claimant’s daily activities, and the dosage, effectiveness, and side effects of medication. Polaski, 751 F.2d at 948. Of course no such list can be exhaustive. The point is, however, that expanding the decision maker’s inquiry beyond objective medical evidence does not result in a pure credibility determination. The decision maker has a good deal more than the appearance of the claimant to use in determining whether the claimant’s pain is so severe as to be disabling.
III.
Although the Secretary’s regulations are consistent with this interpretation of the statute, plaintiffs argue that individual decision makers are nevertheless systematically disregarding congressional intent. Cf. Bowen v. Yuckert, — U.S.—, 107 S.Ct. 2287, 2299, 96 L.Ed.2d 119 (1987) (O’Connor, J., concurring). The district court relied on more than the regulations to reach its conclusion that the Secretary’s policy and practice in evaluating pain is inconsistent with the Disability Reform Act. The court considered internal agency memoranda and statements of officials as well as the decisions in a few actual cases. The district court believed that agency decision makers were actually denying benefits to those who should be entitled to assistance under the statute. The court’s mistaken definition of objective evidence, however, infects its assessment of the record. Consequently, we must reverse the court’s conclusion that the Secretary is overemphasizing objective evidence in evaluating pain in violation of the Act. It is not our province to examine the record anew for the parties and the trial court in order to determine whether in practice agency decision makers are not properly applying the statute.4 See McKay v. Hammock, 730 F.2d 1367, 1374 (10th Cir.1984) (en banc) (“We decline to perform Rule 56 functions for the parties and the trial court_”).
The judgment of the district court is reversed and remanded for further proceedings.
. Because of our resolution of the merits of this appeal, we need not address the class remedy issues raised by the Secretary.
. Subjective evidence, conversely, consists of statements by a claimant or other witnesses on his behalf that are not based on information which an impartial medical expert can evaluate either from examining the claimant himself or from evaluating the claimant’s test results or examination reports. In other words, subjective evidence is that which a decision maker must evaluate solely on the basis of the credibility of the witness.
. In affirming the Secretary's denial of benefits in a few cases, some courts have cited legislative history to support the statement that the statute requires either "objective medical evidence to confirm the severity of the alleged pain arising from th[e] condition or [the condition] must be of a severity which can reasonably be expected to give rise to the alleged pain.” Duncan v. Secretary of Health and Human Services, 801 F.2d 847, 853 (6th Cir.1986) (quoting S.Rep. No. 466, 98th Cong., 2d Sess. 24, reprinted in 130 Cong.Rec. S26221 (daily ed. May 22, 1984)); Landry v. Heckler, 782 F.2d 1551, 1553 (11th Cir.1986) (same). Adoption of the Senate Report’s language, however, does not resolve the issue. Because the same impairment can cause differing levels of pain in different people, the question is the range of pain that any given impairment can be reasonably expected to produce.
. A decision to grant summary judgment in a class action based on improper agency practice must be supported by evidence that decision makers are not deciding cases as the statute requires. Certain internal memoranda do appear to overemphasize objective test results. For a sample of the conflicting messages conveyed by the agency’s rulings and instructions, see the Appendix to Avery v. Secretary, 797 F.2d 19, 25 (1st Cir.1986). In addition, agency officials have made various statements that indicate decision makers are encouraged to place too much emphasis on medical data. See, e.g., Luna v. Bowen, 641 F.Supp. 1109, 1121 (D.Colo. 1986) (Statement of Patricia M. Owens, Associate Commissioner for Disability, to the Atlanta Regional Commissioner for the Social Security Administration, December 18, 1984). Conversely, however, other internal documents stress full consideration of subjective descriptions of pain, and agency officials have made statements which indicate that decision makers should give the proper weight to evidence other than lab and test results. See, e.g., Rec., vol. I, doc. 3, attach. F, at 1. (Deposition of Jean Hall Hinck-ley, SSA’s Acting Associate Deputy Commissioner for Disability until June 25, 1984). The question is not what agency officials say, but rather what decision makers are doing. A thorough examination of internal agency communications and decisions in a substantial, representative sample of cases is necessary to make this determination on a class basis.