****************************************************************************************;
*   Program:		combined_comorbidity_score_code.sas									;
*																						;
*   Purpose:  		To compute combined comorbidity scores based on ICD-9 codes			;
*																						;
*	Description: 	This code takes in a list of ICD-9 codes for						;																	
*																						;
*   Citation: 		Gagne JJ, Glynn RJ, Avorn J, Levin R, Schneeweiss S. A combined		;
* 					comorbidity score predicted mortality in elderly patients better	;	 
*					than existing scores. Journal of Clinical Epidemiology 2011 Jan 	;
*					3 [Epub ahead of print]												;
*																						;
*	Definitions: 	datain = input datset												;
*                 	ICD9variable = ICD-9 codes											;
*																						;
*	Comments:		After merging the final data file (combinedcomorbidityscore) with	;
*					the analytic data file, patients with missing combined comorbidity	;
*					score values (i.e. combinedscore = .) should be set to 0 as these	;
*					patients did not have any of the component conditions.				;	
*																						;
****************************************************************************************;


data conditions;
    set datain(rename=(ICD9variable = ICD));

	length disease $25;
	disease = 'nopoints';

	if substr(ICD,1,3) = '196' or substr(ICD,1,3) = '197' or
	substr(ICD,1,3) = '198' or substr(ICD,1,3) = '199'
								then disease = 'metastatic_romano';

	if ICD = '40201' or ICD = '40211' or ICD = '40291' or
	substr(ICD,1,4) = '4293' or substr(ICD,1,3) = '425' or
	substr(ICD,1,3) = '428'		then disease = 'chf_romano';

	if substr(ICD,1,4) = '3310' or substr(ICD,1,4) = '3311' or 
	substr(ICD,1,4) = '3312' or	substr(ICD,1,3) = '290'		
								then disease = 'dementia_romano';

	if ICD = '40311' or ICD = '40391' or ICD = '40412' or
	ICD = '40492' or substr(ICD,1,3) = '585' or 
	substr(ICD,1,3) = '586' or substr(ICD,1,4) = 'V420' or
	substr(ICD,1,4) = 'V451' or substr(ICD,1,4) = 'V560' or
	substr(ICD,1,4) = 'V568' 	then disease = 'renal_elixhauser';

	if '260' <= substr(ICD,1,3) <= '263'
								then disease = 'wtloss_elixhauser';

	if substr(ICD,1,3) = '342' or substr(ICD,1,3) = '344'
								then disease = 'hemiplegia_romano';

	if substr(ICD,1,4) = '2911' or substr(ICD,1,4) = '2912' or
	substr(ICD,1,4) = '2915' or substr(ICD,1,4) = '2918' or
	substr(ICD,1,4) = '2919' 				or 
	'30390' <= ICD <= '30393' 	or 
	'30500' <= ICD <= '30503' 	or substr(ICD,1,4) = 'V113' 
								then disease = 'alcohol_elixhauser';

	if '140' <= substr(ICD,1,3) <= '171'	or
	'174' <= substr(ICD,1,3) <= '195'		or
	substr(ICD,1,4) = '2730'				or
	substr(ICD,1,4) = '2733'				or
	substr(ICD,1,5) = 'V1046'				or
	'200' <= substr(ICD,1,3) <= '208'	
								then disease = 'tumor_romano';

	if ICD = '42610' or ICD = '42611' or ICD = '42613' or
	'4262' <= substr(ICD,1,4) <= '4264' or 
	'42650' <= ICD <= '42653' 			or 
	'4266' <= substr(ICD,1,4) <= '4268'	or
	substr(ICD,1,4) = '4270' or substr(ICD,1,4) = '4272' or
	ICD = '42731' or ICD = '42760' or substr(ICD,1,4) = '4279' or
	substr(ICD,1,4) = '7850' or substr(ICD,1,4) = 'V450' or 
	substr(ICD,1,4) = 'V533'
								then disease = 'arrhythmia_elixhauser';

	if substr(ICD,1,4) = '4150' or substr(ICD,1,4) = '4168' or
	substr(ICD,1,4) = '4169' or substr(ICD,1,3) = '491'  or
	substr(ICD,1,3) = '492'  or substr(ICD,1,3) = '493'  or
	substr(ICD,1,3) = '494'  or substr(ICD,1,3) = '496'
								then disease = 'pulmonarydz_romano';

	if '2860' <= substr(ICD,1,4) <= '2869'	or
	substr(ICD,1,4) = '2871'				or
	'2873' <= substr(ICD,1,4) <= '2875'
								then disease = 'coagulopathy_elixhauser';

	if '25040' <= ICD <= '25073' 			or
	'25090' <= ICD <= '25093'	then disease = 'compdiabetes_elixhauser';

	if '2801' <= substr(ICD,1,4) <= '2819'	or
	substr(ICD,1,4) = '2859'	then disease = 'anemia_elixhauser';

	if '2760' <= substr(ICD,1,4) <= '2769'
								then disease = 'electrolytes_elixhauser';

	if ICD = '07032' or ICD = '07033' or ICD = '07054' or
	substr(ICD,1,4) = '4560' or substr(ICD,1,4) = '4561' or
	ICD = '45620' or ICD = '45621' or
	substr(ICD,1,4) = '5710' or substr(ICD,1,4) = '5712' or 
	substr(ICD,1,4) = '5713' or
	'57140' <= ICD <= '57149' or substr(ICD,1,4) = '5715' or
	substr(ICD,1,4) = '5716' or substr(ICD,1,4) = '5718' or
	substr(ICD,1,4) = '5719' or substr(ICD,1,4) = '5723' or
	substr(ICD,1,4) = '5728' or substr(ICD,1,4) = 'V427'
								then disease = 'liver_elixhauser';

	if '4400' <= substr(ICD,1,4) <= '4409' or 
	substr(ICD,1,4) = '4412' or	substr(ICD,1,4) = '4414' or 
	substr(ICD,1,4) = '4417' or substr(ICD,1,4) = '4419' or 
	'4431' <= substr(ICD,1,4) <= '4439' or
	substr(ICD,1,4) = '4471' or substr(ICD,1,4) = '5571' or 
	substr(ICD,1,4) = '5579' or substr(ICD,1,4) = 'V434'
								then disease = 'pvd_elixhauser';

	if '29500' <= ICD <= '29899' or
	ICD = '29910' or ICD = '29911'	
								then disease = 'psychosis_elixhauser';

	if substr(ICD,1,3) = '416' or substr(ICD,1,4) = '4179'
								then disease = 'pulmcirc_elixhauser';

	if substr(ICD,1,3) = '042' or substr(ICD,1,3) = '043' or
	substr(ICD,1,3) = '044'		then disease = 'hivaids_romano';

	if substr(ICD,1,4) = '4011' or substr(ICD,1,4) = '4019' or
	ICD = '40210' or ICD = '40290' or ICD = '40410' or ICD = '40490' or
	ICD = '40511' or ICD = '40519' or ICD = '40591' or ICD = '40599'
								then disease = 'hypertension_elixhauser';

	if disease ^= 'nopoints';
run;

proc sort nodupkey data = conditions;
    by uniqueidentifier disease;
run;

*Applying the weights;
data conditionweights;
	set conditions;

	weight = 0;

	if disease = 'metastatic_romano'		then weight = 5;
	if disease = 'chf_romano'				then weight = 2;
	if disease = 'dementia_romano'			then weight = 2;
	if disease = 'renal_elixhauser'			then weight = 2;
	if disease = 'wtloss_elixhauser'		then weight = 2;
	if disease = 'hemiplegia_romano'		then weight = 1;
	if disease = 'alcohol_elixhauser'		then weight = 1;
	if disease = 'tumor_romano'				then weight = 1;
	if disease = 'arrhythmia_elixhauser'	then weight = 1;
	if disease = 'pulmonarydz_romano'		then weight = 1;
	if disease = 'coagulopathy_elixhauser'	then weight = 1;
	if disease = 'compdiabetes_elixhauser'	then weight = 1;
	if disease = 'anemia_elixhauser'		then weight = 1;
	if disease = 'electrolytes_elixhauser'	then weight = 1;
	if disease = 'liver_elixhauser'			then weight = 1;
	if disease = 'pvd_elixhauser'			then weight = 1;
	if disease = 'psychosis_elixhauser'		then weight = 1;
	if disease = 'pulmcirc_elixhauser'		then weight = 1;
	if disease = 'hivaids_romano'			then weight = -1;
	if disease = 'hypertension_elixhauser'	then weight = -1;

	keep uniqueidentifier disease weight;
run;

*Summing the weights;
data combinedcomorbidityscore;
	set conditionweights(keep = uniqueidentifier weight);

	by uniqueidentifier;
		if first.uniqueidentifier then combinedscore = 0;
	   		combinedscore + weight;
 		if last.uniqueidentifier then output;

keep uniqueidentifier combinedscore;
run;

*Note: patients not included in the final data set (combinedcomorbidityscore) did not
have any of the component conditions. Therefore, be sure to set their combined comorbidity
score values to zero;