/*This code was written by Janel Hanmer in August 2005

All predictions require the subscales from the SF-36 (physical function, general health, etc).
The subscales are presumed to be named PF36, GH36, etc.
The HUI2 prediction by Nichol et al requires age in years.  */


/*put the location of your dataset here */
data sf36;
	set 'C:\ . . .';
run;

/*Nichol MB, Sengupta N, Globe D. Evaluating quality adjusted life years: Estimation of the Health Utility Index 
(HUI) from the SF-36. Med Decis Making. 2001;21:19-26.

Age is in years.

For HUI2, used SF-36 version 1 presumed because of year of data collection (1992). 
There is an appendix in the paper for confidence intervals around estimates. 

This algorithm uses T-scores, not the scale score.  T-scores for version 1 are calculated using:

pf_z=(pf36-84.52404)/22.89490;
	rp_z=(rp36-81.19907)/33.79729;
	bp_z=(bp36-75.49196)/23.55879;
	gh_z=(gh36-72.21316)/20.16964;
	vt_z=(vt36-61.05453)/20.86942;
	sf_z=(sf36-83.59753)/22.37642;
	re_z=(re36-81.29467)/33.02717;
	mh_z=(mh36-74.84212)/18.01189;

	pf_t = pf_z*10 + 50;
	rp_t = rp_z*10 + 50;
	bp_t = bp_z*10 + 50;
	gh_t = gh_z*10 + 50;
	vt_t = vt_z*10 + 50;
	sf_t = sf_z*10 + 50;
	re_t = re_z*10 + 50;
	mh_t = mh_z*10 + 50;
*/

data sf36;
	set sf36;
	Nichol_HUI2 = .0009*GH_t + .00046*RP_t + .0043*BP_t + .0042*MH_t + .0018*PF_t + .0018*VT_t + .0015*RE_t + 
				.0015*SF_t - .0006*AGE + .045;
run;


/*Fryback DG, Lawrence WF, Martin PA, et al
Predicting Quality of Well-being Scores from the SF-36, Results from the Beaver Dam Health Outcomes Study.
Med Decis Making 1997;17:1-9 */

data sf36;
	set sf36;
	Fryback_QWB= .59196 + .0012588*PF36 - .0011709*MH36 - .0014261*BP36 + .00000705*GH36*RP36 + 
				.00001140*PF36*BP36 + .00001931*MH36*BP36;
run;