In this appeal, the defendants in a personal injury action challenge the trial .court’s computation of the attorney’s fees awarded to plaintiffs’ counsel after a judgment was obtained in an amount greater than 25% of the plaintiffs’ demand. We reverse because the trial court erred by enhancing the initial lodestar figure by application of both a results obtained factor and a contingency risk multiplier.
No challenge is made to the trial court’s determination of the number of hours reasonably expended and the reasonable hourly rate. However, after multiplying the number of hours by the hourly rate and arriving at a lodestar figure of $38,380, the trial court ádjusted this amount upward by 50% because of the results obtained. The trial court then applied a contingency risk multiplier of 1.5 to the enhanced figure of $57,570 and thereby increased the fee to $86,355. The defendants do not challenge the trial court’s decision to *381apply a contingency risk multiplier in this case, nor do they challenge the selection of 1.5 as the amount of the multiplier. They do, however, challenge the trial court’s upward adjustment of the initial lodestar figure by 50% which was based on a results obtained factor. We agree that the trial court erred in this regard.
In a contingency fee case, the results obtained is one of several factors that the court must consider in determining whether or not to apply a multiplier. As the supreme court made clear in Standard Guaranty Insurance Co. v. Quanstrom, 555 So.2d 828 (Fla.1990), the application of a multiplier is not mandatory in contingency fee cases. The trial court must determine whether a multiplier is necessary by considering the following factors:
(1) whether the relevant market requires a contingency fee multiplier to obtain competent counsel; (2) whether the attorney was able to mitigate the risk of nonpayment in any way; and (3) whether any of the factors set forth in Rowe are applicable, especially, the amount involved, the results obtained, and the type of fee arrangement between the attorney and his client. Evidence of these factors must be presented to justify the utilization of a multiplier.
Id. at 834.
Once the court determines that a multiplier is necessary, the trial court must then examine the likelihood of success at the outset and select a multiplier from the range of 1 to 2.5 in accordance with the guidelines set out in Quanstrom. From our review of this two-step procedure, we conclude that the “results obtained” is a factor that is subsumed within the enhancement of a fee by application of a contingency risk multiplier and, therefore, may not be used again as an independent basis to enhance the lodestar figure in those cases where a contingency risk multiplier is applied. However, as explained in Florida Patient’s Compensation Fund v. Rowe, 472 So.2d 1145 (Fla.1985), “[t]he ‘results obtained’ may provide an independent basis for reducing the fee when the party prevails on a claim or claims for relief, but is unsuccessful on other unrelated claims.” Id. at 1151.
Appellee argues that “results obtained” may be used to enhance as well as decrease the lodestar fee and cites to our statement in Lee County v. Tohari, 582 So.2d 104 (Fla. 2d DCA 1991), that “[i]n Quan-strom and Rowe, the supreme court recognized that the lodestar fee can be adjusted upward or downward for both a contingency risk factor and a ‘results obtained’ factor.” This dicta is an interpretation of the supreme court’s statement in Rowe that “[o]nce the court arrives at the lodestar figure, it may add or subtract from the fee based upon a ‘contingency risk’ factor and the ‘results obtained.’ ” Lee County involved the award of attorney’s fees in an eminent domain action. Having now examined the supreme court’s statement in the context of a contingency fee case, we believe that a more precise interpretation of this language as applied to contingency fee cases is that once the court arrives at the lodestar figure, it may add to the fee based upon a contingency risk factor or subtract from the fee based upon the results obtained.
We reach this conclusion for three reasons. First, a multiplier as set forth in Quanstrom, by definition, can never produce a reduced fee. Second, immediately following the sentence in Rowe that we are examining, the supreme court explained how the lodestar figure is enhanced by a multiplier and then how the lodestar fee may be reduced based on results obtained when a party prevails on only a portion of the claims made in the litigation. No explanation was given regarding an increase based on results obtained. Finally, the supreme court concluded its opinion by summarizing the proper computation of an attorney’s fee and stated that, when appropriate, the trial court may “adjust the fee on the basis of the contingent nature of the litigation or the failure to prevail on a claim or claims.” Id. at 1151-1152.
Because the trial court erred by enhancing the lodestar figure based on the results obtained, we reverse and remand with directions that the trial court compute the plaintiffs’ attorney’s fee by applying the undisputed multiplier of 1.5 to the undisputed *382lodestar fee of $38,380 and award a fee in the amount of $57,570.
We also certify the following question:
IN A CONTINGENCY FEE CASE, MAY THE LODESTAR FIGURE BE ADJUSTED UPWARD BY APPLICATION OF BOTH A RESÚLTS OBTAINED FACTOR AND A CONTINGENCY RISK MULTIPLIER?
Reversed and remanded with directions.
CAMPBELL, A.C.J., and PATTERSON, J., concur.