Identifying patients with cost-related medication non-adherence: a big-data approach

Figures & data

Table 1. Respondent characteristics and cost-related medication non-adherence.

	Not reporting CRN, n (%)	Reporting CRN, n (%)	p-value
Characteristics	422 (100)	137 (100)
Age: mean (SD)	63.8 (15.6)	58.2 (13.9)	<0.001
Gender: Male	160 (38)	53 (38)	0.87
Female	262 (62)	84 (61)
Race: White	32 (8)	11 (8)	0.88
African-American	366 (88)	119 (87)
Other	20 (4)	7 (5)
Ethnicity: Hispanic	13 (3)	3 (2)	0.59
Non-Hispanic	409 (97)	134 (98)
Insurance status			0.55
Dual eligible	194 (46)	67 (49)
Non-dual eligible	228 (54)	70 (51)
Education			0.86
Less than high school	83 (20)	26 (19)
High school or more	339 (80)	111 (81)
Comorbidities			0.03
Anemia: Yes	89 (21)	39 (28)
No	326 (77)	92 (67)
U/K or refused	7 (2)	6 (1)
Asthma: Yes	123 (29)	58 (42)	<0.01
No	296 (70)	77 (56)
U/K or refused	3 (1)	2 (1)
Arthritis: Yes	218 (52)	86 (63)	0.03
No	188 (45)	47 (34)
U/K or refused	16 (4)	3 (2)
Back problem: Yes	163 (39)	67 (49)	0.04
No	257 (61)	68 (50)
U/K or refused	2 (<1)	2 (1)
Cancer: Yes	40 (9)	13 (9)	0.29
No	380 (90)	123 (90)
U/K or refused	2 (<1)	1 (1)
Depression: Yes	97 (23)	52 (38)	<0.01
No	318 (75)	83 (61)
U/K or refused	7 (2)	2 (1)
Diabetes: Yes	163 (39)	53 (39)	0.22
No	253 (60)	79 (58)
U/K or refused	6 (1)	5 (4)
Digestive problem: Yes	102 (24)	63 (46)	<0.01
No	315 (75)	70 (51)
U/K or refused	5 (1)	4 (3)
Heart trouble: Yes	177 (42)	55 (40)	0.26
No	238 (57)	77 (56)
U/K or refused	7 (2)	5 (4)
Hypertension: Yes	315 (75)	103 (75)	0.32
No	101 (24)	32 (23)
U/K or refused	6 (1)	2 (1)
HIV or AIDS: Yes	13 (3)	5 (4)	0.28
No	407 (96)	131 (96)
U/K or refused	2 (< 1)	1 (1)
Kidney disease: Yes	116 (27)	39 (28)	0.37
No	297 (70)	94 (69)
U/K or refused	9 (2)	4 (3)
Liver problem: Yes	24 (6)	13 (10)	0.12
No	389 (92)	119 (87)
U/K or refused	9 (2)	5 (4)
Stroke: Yes	65 (15)	24 (18)	0.23
No	349 (83)	111 (81)
U/K or refused	8 (2)	2 (1)

U/K, unknown.

Table 2. Patterns of cost-related medication non-adherence behaviors.

No fill or refill	Delay filling	Skip doses	Split doses	n (%)
+	+	+	+	30 (22)
+	+	+		2 (1)
+		+	+	2 (1)
+	+			29 (21)
+		+	+	2 (1)
+		+		2 (1)
+			+	3 (2)
+				17 (12)
	+	+	+	6 (4)
	+	+		5 (4)
	+		+	2 (1)
	+			21 (15)
		+	+	6 (4)
		+		7 (5)
			+	3 (2)
87	97	60	54	137 (100)

Table 3. The association between ordering of Rx on receipt of SS payments and cost-related medication non-adherence.

Occurrences of ordering of Rx upon receipt of SS payment	no CRN n (%)	CRN n (%)
0	176 (42)	41 (30)
1	111 (26)	46 (34)
2	67 (16)	22 (16)
3	31 (7)	15 (11)
4	18 (4)	10 (7)
5	9 (2)	0 (0)
6	4 (1)	2 (1)
7	3 (1)	0 (0)
8	3 (1)	0 (0)
10	0 (0)	1 (1)
Total	422 (100)	137 (100)

p = 0.04 by Chi-square test. The cut-off date is 2 days upon receipt of SS payment.

Table 4. Odds ratio of timed ordering medication in predicting cost-related medication non-adherence in multivariate logistic regression analysis.

	Model I. Age only: Odds ratio (p-value)	Model II. Age + ordering of Rx when receiving SS payment Odds ratio (p-value)	Model III. Age + ordering of Rx when receiving SS + chronic conditions Odds ratio (p-value)
Age (mean +1 SD reporting CRN)	2.35 (< 0.01)	2.25 (< 0.01)	2.22 (0.03)
1 or more ordering of Rx when receiving SS	–	1.55 (0.03)	1.47 (0.08)
Anemia	–	–	1.17 (0.50)
Asthma	–	–	1.25 (0.31)
Arthritis	–	–	1.36 (0.17)
Back problem	–	–	0.95 (0.83)
Depression	–	–	1.47 (0.09)
Digestive	–	–	2.25 (<0.01)
Area under Receiver Operating Curve	0.57	0.61	0.70

The cut-off date is 2 days upon receipt of SS payment. The presence of any one or more of the four-type CRN behaviors was considered CRN.