Medication adherence as a predictor of 30-day hospital readmissions

Figures & data

Table 1 Patient characteristics

Patient characteristics	Medication adherencea						P-valueb
	Low (N=94)		Intermediate (N=141)		High (N=150)
Unplanned readmission, n (%)							0.020
Not readmitted	75	(79.8%)	113	(80.1%)	136	(90.7%)
Readmitted	19	(20.2%)	28	(19.9%)	14	(9.3%)
Age, median (Q1, Q3)	68.0	(56, 78)	67.0	(56, 80)	69.0	(59, 81)	0.66
Female, n (%)	50	(53.2%)	70	(49.6%)	70	(46.7%)	0.61
Ethnicity, n (%)							0.61
Hispanic	8	(8.5%)	15	(10.6%)	11	(7.3%)
Non-Hispanic	86	(91.5%)	126	(89.4%)	139	(92.7%)
Race, n (%)							0.047
Asian	6	(6.4%)	6	(4.3%)	13	(8.7%)
Black or African American	30	(31.9%)	30	(21.3%)	23	(15.3%)
White	58	(61.7%)	103	(73.0%)	111	(74.0%)
Others	0	(0.0%)	2	(1.4%)	3	(2.0%)
Marital status, n (%)							0.049
Divorced or legally separated	12	(12.8%)	15	(10.6%)	12	(8.0%)
Married, significant other or domestic partner	30	(31.9%)	69	(48.9%)	81	(54.0%)
Single	29	(30.9%)	32	(22.7%)	28	(18.7%)
Widowed	23	(24.5%)	25	(17.7%)	29	(19.3%)
Primary insurance, n (%)							0.86
HMO	9	(9.6%)	17	(12.1%)	13	(8.7%)
Medi-Cal	1	(1.1%)	2	(1.4%)	3	(2.0%)
Medicare	73	(77.7%)	104	(73.7%)	109	(72.7%)
PPO	10	(10.6%)	15	(10.6%)	23	(15.3%)
Self-pay/others	1	(1.1%)	3	(2.1%)	2	(1.3%)
Source of index admission, n (%)							0.34
Home	91	(96.8%)	130	(92.2%)	145	(96.7%)
MD office	2	(2.1%)	5	(3.5%)	3	(2.0%)
Transfer from another facility	1	(1.1%)	6	(4.3%)	2	(1.3%)
Type of index admission, n (%)							0.76
Elective	2	(2.1%)	2	(1.4%)	4	(2.7%)
Emergency	92	(97.9%)	139	(98.6%)	146	(97.3%)
Discharge from an oncology service, n (%)	10	(10.6%)	27	(19.1%)	21	(14.0%)	0.18
Length of stay of the index admission, median (Q1, Q3)	5.0	(3.0, 7.0)	5.0	(3.0, 8.0)	5.0	(3.0, 8.0)	0.59
Number of hospital admissions in the past year, median (Q1, Q3)	1.0	(0.0, 3.0)	1.0	(0.0, 2.0)	1.0	(0.0, 2.0)	0.07
Number of PTA medications, median (Q1, Q3)	20.0	(16, 27)	20.0	(15, 25)	20.0	(16, 24)	0.74
Number of procedures during index admission, median (Q1, Q3)	2.0	(1.0, 4.0)	2.0	(1.0, 4.0)	2.0	(1.0, 4.0)	0.91
Hemoglobin level at discharge, median (Q1, Q3)	10.7	(9.3, 12.1)	10.6	(9.3, 11.8)	10.7	(9.6, 12.0)	0.38
Serum sodium level at discharge, median (Q1, Q3)	138.0	(135, 140)	138.0	(135, 141)	138.0	(136, 141)	0.52
Serum creatinine level at discharge, median (Q1, Q3)	1.0	(0.8, 2.0)	1.1	(0.8, 1.8)	1.0	(0.8, 1.6)	0.18
Primary diagnosis on admission, n (%)							0.46
Arrhythmia	3	(3.2%)	2	(1.4%)	6	(4.0%)
Asthma/COPD	6	(6.4%)	3	(2.1%)	5	(3.3%)
Bleeding	3	(3.2%)	4	(2.8%)	8	(5.4%)
Bronchitis/pneumonia	4	(4.3%)	4	(2.8%)	9	(6.0%)
Diabetes mellitus	5	(5.3%)	4	(2.8%)	3	(2.0%)
Heart failure	11	(11.7%)	17	(12.1%)	12	(8.0%)
Hypertension	6	(6.4%)	2	(1.4%)	1	(0.7%)
Infection	13	(13.8%)	31	(22.0%)	27	(18.0%)
Ischemic heart disease	3	(3.2%)	10	(7.1%)	12	(8.0%)
Malignant neoplasm	2	(2.1%)	6	(4.3%)	6	(4.0%)
Pancreatitis	3	(3.2%)	4	(2.8%)	7	(4.7%)
Renal failure	4	(4.3%)	8	(5.7%)	9	(6.0%)
Venous thromboembolism	1	(1.1%)	2	(1.4%)	5	(3.3%)
Othersc	30	(31.9%)	44	(31.2%)	40	(26.7%)
Primary diagnosis at 30-day readmission, n (%)							0.42
Arrhythmia	1	(5.3%)	0	(0.0%)	0	(0.0%)
Asthma/COPD	2	(10.5%)	0	(0.0%)	0	(0.0%)
Bleeding	0	(0.0%)	1	(3.6%)	0	(0.0%)
Bronchitis/pneumonia	0	(0.0%)	1	(3.6%)	1	(7.1%)
Diabetes mellitus	0	(0.0%)	2	(7.1%)	1	(7.1%)
Heart failure	1	(5.3%)	3	(10.7%)	1	(7.1%)
Hypertension	2	(10.5%)	0	(0.0%)	0	(0.0%)
Infection	4	(21.1%)	8	(28.6%)	8	(57.1%)
Ischemic heart diseases	1	(5.3%)	1	(3.6%)	0	(0.0%)
Malignancy	0	(0.0%)	1	(3.6%)	0	(0.0%)
Pancreatitis	2	(10.5%)	2	(7.1%)	0	(0.0%)
Venous thromboembolism	0	(0.0%)	2	(7.1%)	0	(0.0%)
Othersd	6	(31.6%)	7	(25.0%)	3	(21.4%)
Comorbidity, n (%)
Atrial fibrillation	34	(36.2%)	53	(37.6%)	62	(41.3%)	0.68
COPD	17	(18.1%)	20	(14.2%)	17	(11.3%)	0.33
Diabetes mellitus	43	(45.7%)	58	(41.1%)	68	(45.3%)	0.71
Ischemic heart diseases	43	(45.7%)	52	(36.9%)	55	(36.7%)	0.30
Heart failure	39	(41.5%)	53	(37.6%)	52	(34.7%)	0.56
Malignant neoplasm	13	(13.8%)	20	(14.2%)	18	(12.0%)	0.84

Notes:

a For the assessment of medication adherence, we used MMAS-4. Use of the ©MMAS is protected by the US and International copyright laws. Permission for use is required. A license agreement is available from Donald E Morisky, MMAS Research (MORISKY), 294 Lindura Court, Las Vegas, NV 89135-1415; [e-mail] © 2007 Donald E Morisky. All rights reserved. MMAS, Morisky Medication Adherence Scale, and Morisky are the trademarks of Donald E Morisky and may be used only with permission.

b For categorical variables, we used chi-square tests. For nonnormally distributed variables, we used the two-sided Kruskal–Wallis rank sum test.

c Other diagnoses at admission included anemia, bile duct obstruction, arthritis, calculus, cervical spondylosis, cholangitis, fluid overload, fracture, complications of transplanted organ, complications of implanted device, constipation, degeneration of lumber or lumbosacral intervertebral disk, dehydration, dementia, dislocation of prosthetic joint, disorder of autonomic nervous system, displacement of cervical intervertebral disk, electrolyte imbalance, encephalopathy, epilepsy, fecal impaction, gastritis, gastromy complications, gout, hallucinations, headache, hemiplegia, inguinal hernia with obstruction, intestinal obstruction, migraine, orthostatic hypotension, osteoarthrosis, pain, poisoning, pressure ulcer, prosthetic joint repair, sickle-cell crisis, spasm of muscle, sprain of neck, systemic lupus erythematosus, systemic sclerosis, psychosis, ulcerative colitis, and vascular insufficiency of intestine.

d Other diagnoses at 30-day readmission included acute vascular insufficiency of intestine, altered mental status, cerebral infarction, electrolyte imbalance, fracture, hemiplegia, hepatic encephalopathy, complications due to renal dialysis device, hypersensitivity angitis, fall, pain, and unspecified disorder of stomach and duodenum.

Abbreviations: COPD, chronic obstructive pulmonary disease; HMO, Health Maintenance Organization; MD, Doctor of Medicine; MMAS-4, 4-item Morisky Medication Adherence Scale; PPO, Preferred Provider Organization; PTA, prior to admission.

Table 2 Patient readmitted within 30 days vs not readmitted

Patient characteristics	Not readmitted (N=324)		Readmitted (N=61)		P-valuea
Age, median (Q1, Q3)	68.0	(56.0, 80.0)	66.0	(59.0, 81.0)	0.98
Female, n (%)	155	(47.8%)	35	(57.4%)	0.17
Ethnicity, n (%)					0.43
Hispanic	27	(8.3%)	7	(11.5%)
Non-Hispanic	297	(91.7%)	54	(88.5%)
Race, n (%)					0.26
Asian	18	(5.6%)	7	(11.5%)
Black or African American	69	(21.3%)	14	(23.0%)
White	232	(71.6%)	40	(65.6%)
Others	5	(1.5%)	0	(0.0%)
Marital status, n (%)					0.93
Divorced or legally separated	33	(10.2%)	6	(9.8%)
Married, significant other or domestic partner	152	(46.9%)	28	(45.9%)
Single	73	(22.5%)	16	(26.2%)
Widowed	66	(20.4%)	11	(18.0%)
Primary insurance, n (%)					0.28
HMO	36	(11.1%)	3	(4.9%)
Medi-Cal	4	(1.2%)	2	(3.3%)
Medicare	240	(74.1%)	46	(75.4%)
PPO	40	(12.3%)	8	(13.1%)
Self-pay/others	4	(1.2%)	2	(3.3%)
Source of index admission, n (%)					0.35
Home	306	(94.4%)	60	(98.4%)
MD office	10	(3.1%)	0	(0.0%)
Transfer from another facility	8	(2.5%)	1	(1.6%)
Type of index admission, n (%)					0.79
Elective	7	(2.2%)	1	(1.6%)
Emergency	317	(97.8%)	60	(98.4%)
Discharge from an oncology service, n (%)	43	(13.3%)	15	(24.6%)	0.023
Length of stay of the index admission, median (Q1, Q3)	5.0	(3.0, 7.0)	6.0	(4.0, 10.0)	0.003
Number of hospital admissions in the past year, median (Q1, Q3)	1.0	(0.0, 2.0)	1.0	(0.0, 3.0)	0.14
Number of PTA medications, median (Q1, Q3)	20.0	(16.0, 25.0)	20.0	(16.0, 25.0)	0.92
Number of procedures during index admission, median (Q1, Q3)	2.0	(1.0, 4.0)	2.0	(1.0, 5.0)	0.97
Hemoglobin level at discharge, median (Q1, Q3)	10.7	(9.5, 12.1)	10.3	(9.3, 11.3)	0.09
Serum sodium level at discharge, median (Q1, Q3)	138.0	(136.0, 141.0)	137.0	(135.0, 140.0)	0.044
Serum creatinine level at discharge, median (Q1, Q3)	1.1	(0.8, 1.7)	1.1	(0.9, 2.4)	0.33
Medication adherence, n (%)					0.020
Low	75	(23.1%)	19	(31.1%)
Intermediate	113	(34.9%)	28	(45.9%)
High	136	(42.0%)	14	(23.0%)
Medication adherence collapsed, n (%)					0.005
Low/intermediate	188	(58.0%)	47	(77.0%)
High	136	(42.0%)	14	(23.0%)
Primary diagnosis on admission, n (%)					0.63
Arrhythmia	10	(3.1%)	1	(1.6%)
Asthma/COPD	11	(3.4%)	3	(4.9%)
Bleeding	14	(4.6%)	0	(0.0%)
Bronchitis/pneumonia	15	(4.6%)	2	(3.3%)
Diabetes mellitus	9	(2.8%)	3	(4.9%)
Heart failure	36	(11.1%)	4	(6.6%)
Hypertension	8	(2.5%)	1	(1.6%)
Infection	56	(17.3%)	15	(24.6%)
Ischemic heart diseases	21	(6.5%)	4	(6.6%)
Malignant neoplasm	11	(3.4%)	3	(4.9%)
Pancreatitis	9	(2.8%)	5	(8.2%)
Renal failure	18	(5.6%)	3	(4.9%)
Venous thromboembolism	7	(2.2%)	1	(1.6%)
Othersb	98	(30.2%)	16	(26.2%)
Primary diagnosis at 30-day readmission, n (%)
Arrhythmia			1	(1.6%)
Asthma/COPD			2	(3.3%)
Bleeding			1	(1.6%)
Bronchitis/pneumonia			2	(3.3%)
Diabetes mellitus			3	(4.9%)
Heart failure			5	(8.2%)
Hypertension			2	(3.3%)
Infection			20	(32.8%)
Ischemic heart diseases			2	(3.3%)
Malignant neoplasm			1	(1.6%)
Pancreatitis			4	(6.6%)
Venous thromboembolism			2	(3.3%)
Othersc			16	(26.2%)
Comorbidity, n (%)
Atrial fibrillation	132	(40.7%)	17	(27.9%)	0.06
COPD	40	(12.3%)	14	(23.0%)	0.029
Diabetes mellitus	144	(44.4%)	25	(41.0%)	0.62
Ischemic heart diseases	134	(41.4%)	16	(26.2%)	0.026
Heart failure	125	(38.6%)	19	(31.1%)	0.27
Atrial fibrillation	132	(40.7%)	17	(27.9%)	0.06

Notes:

a For categorical variables, we used Fisher’s exact and chi-square tests. For nonnormally distributed variables, we used the two-sided Kruskal–Wallis rank sum test.

b Other diagnoses at admission included anemia, bile duct obstruction, arthritis, calculus, cervical spondylosis, cholangitis, fluid overload, fracture, complications of transplanted organ, complications of implanted device, constipation, degeneration of lumber or lumbosacral intervertebral disk, dehydration, dementia, dislocation of prosthetic joint, disorder of autonomic nervous system, displacement of cervical intervertebral disk, electrolyte imbalance, encephalopathy, epilepsy, fecal impaction, gastritis, gastromy complications, gout, hallucinations, headache, hemiplegia, inguinal hernia with obstruction, intestinal obstruction, migraine, orthostatic hypotension, osteoarthrosis, pain, poisoning, pressure ulcer, prosthetic joint repair, sickle-cell crisis, spasm of muscle, sprain of neck, systemic lupus erythematosus, systemic sclerosis, psychosis, ulcerative colitis, and vascular insufficiency of intestine.

c Other diagnoses at 30-day readmission included acute vascular insufficiency of intestine, altered mental status, cerebral infarction, electrolyte imbalance, fracture, hemiplegia, hepatic encephalopathy, complications due to renal dialysis device, hypersensitivity angitis, fall, pain, and unspecified disorder of stomach and duodenum.

Abbreviations: COPD, chronic obstructive pulmonary disease; HMO, Health Maintenance Organization; MD, Doctor of Medicine; PPO, Preferred Provider Organization; PTA, prior to admission.

Table 3 Multiple logistic regression model results for 30-day readmission, including adherence

Association	HOSPITAL model with adherence				HOSPITAL model
	Estimate	95% confidence limits		P-valuea	Estimate	95% confidence limits		P-valuea
		Lower limit	Upper limit			Lower limit	Upper limit
Hemoglobin level at discharge (<12 g/dL)	1.336	0.635	2.810	0.445	1.343	0.643	2.804	0.432
Discharge from an oncology service	2.219	1.078	4.566	0.031	2.210	1.086	4.494	0.029
Sodium level at discharge (<135 mEq/L)	1.646	0.807	3.359	0.171	1.764	0.872	3.567	0.114
Procedure during hospital stay (any ICD-9-CM-coded procedure)	1.098	0.579	2.082	0.773	1.130	0.601	2.127	0.704
Index admission type: nonelective	3.383	0.946	12.107	0.061	3.294	0.925	11.732	0.066
Hospital admissions during the previous year	0.676	0.295	1.551	0.356	0.721	0.317	1.636	0.433
Length of stay ≥5 days	1.974	1.062	3.669	0.032	1.838	0.996	3.393	0.052
Medication adherence (low/intermediate vs high)	2.545	1.323	4.902	0.005

Notes: c-statistic =0.70 and 0.65 for the models with and without medication adherence, respectively.

a Wald chi-square test.

Figure 1 Selection process for study inclusion.

Abbreviation: SNF, skilled nursing facility.