Deborah McDonald appeals the district court’s1 order affirming the denial of disability insurance benefits. In her January 2002 application, McDonald claimed disability since May 2001 from, inter alia, a herniated cervical disc and back spasms. After a February 2008 hearing, an administrative law judge (ALJ) determined that McDonald’s cervical disc disease was severe, but not of listing-level severity; and that her allegations about her limitations were not totally credible. The ALJ further determined that McDonald had the residual functional capacity (RFC) for a significant range of sedentary work, which precluded her past relevant work, but— with the Medical-Vocational Guidelines as a framework, and based on the testimony of a vocational expert — she was not disabled. Having carefully reviewed the record, we affirm. See Draper v. Barnhart, 425 F.3d 1127, 1130 (8th Cir.2005) (standard of review).
We reject McDonald’s challenges to the ALJ’s credibility analysis. The ALJ gave multiple valid reasons for finding McDonald not entirely credible; and contrary to McDonald’s assertions, the ALJ considered the requisite credibility factors announced in Polaski v. Heckler, 739 F.2d 1320, 1322 (8th Cir.1984), and explained his credibility findings. See Goff v. Barnhart, 421 F.3d 785, 791-92 (8th Cir.2005) (ALJ need not explicitly discuss each Polaski factor so long as he acknowledges and considers them; this court will not disturb decision of ALJ who considers, but for good cause expressly discredits, claimant’s complaints of disabling pain).
We also find no error in the ALJ’s decision to discount Dr. Ralph Meehan’s 2001 opinion that McDonald was fully disabled. While a treating physician’s medical opinion is generally entitled to deference, a medical-source opinion that an applicant is unable to work is not the type of opinion to which the Commissioner must defer. See Ellis v. Barnhart, 392 F.3d 988, 994 (8th Cir. 2005). The ALJ’s interpretation of Dr. Meehan’s opinion as reflecting only McDonald’s inability to perform her past relevant work is supported by substantial evidence, especially because Dr. Meehan had recommended in 2002 that McDonald have a functional-capacity evaluation and try to work. We disagree with McDonald that the ALJ was required to contact Dr. Meehan for clarification, as the record was not undeveloped on McDonald’s RFC. To the extent McDonald is challenging the ALJ’s final RFC determination, the determination was based as required on the medical records, the observations of treating physicians and others, and on McDonald’s own description of her limitations, which were properly discredited in part. See Stormo v. Barnhart, 377 F.3d 801, 807 (8th Cir.2004) (RFC determination); Dolph v. Barnhart, 308 F.3d 876, 880 (8th Cir.2002) (while there is little doubt claimant has pain, issue is whether pain is so severe as to be disabling).
Finally, McDonald contends that the ALJ should have arranged for a myelogram in order to develop the record fully. We disagree. The ALJ arranged for and considered a post-hearing orthopedic evaluation; McDonald does not explain why she failed to contact neurosurgeon Luke Knox when she was ready to sched*694ule a myelogram as he had directed her to do in September 2001; and, in any event, we find sufficient evidence in the record to support the ALJ’s decision, see Haley v. Massanari, 258 F.3d 742, 749-50 (8th Cir. 2001) (ALJ may issue decision without obtaining added medical evidence if existing evidence provides sufficient basis for decision).
Accordingly, we affirm.