The COPD Severity Score: A Dynamic Prediction Tool for Health-Care Utilization

Figures & data

Table 1 COPD Severity Score Items (n = 267)

Items	Score	No (%)
Respiratory symptoms (maximum 7 points)
Dyspnea on exertion, current
None	0	58 (22)
Hurrying on level ground or walking uphill	1	80 (30)
Walking with peers on level ground	2	44 (16)
Walking at own pace on level ground	3	85 (32)
Dyspnea during the past 14 d or nights
None	0	61 (23)
1 to 2 d or nights	1	34 (13)
3 to 6 d or nights	2	58 (22)
7 to 13 d or nights	3	38 (14)
Every day or night	4	76 (28)
Systemic corticosteroid use (maximum 5 points)
Ever used	1	75 (28)
Long-term use in past year	3	21 (8)
Used in past 2 wk	1	13 (5)
Other medication use (maximum 10 points)^†
Metered-dose inhaler in past 2 wk
Short-acting ß-agonists	1	104 (39)
Long-acting ß-agonists	1	53 (20)
Inhaled corticosteroids	1	66 (25)
Anticholinergic^‡	1	66 (25)
Nebulizer use, past 2 wk
Short-acting ß-agonists	1	40 (15)
Ipratropium bromide	1	16 (6)
Oral medications
Theophylline, past 2 wk	1	17 (6)
ß-Agonists, past 2 wk	1	2 (1)
Antibiotics for lung condition, past 12 mo
One to two courses	1	48 (18)
Three or more courses	2	26 (10)
Hospitalization/intubation/oxygen use (maximum 13 points)
Hospitalized for COPD, past 5 yr	3	83 (31)
Intubated for COPD, past 5 yr	5	9 (3)
Home oxygen, current day or nighttime use	5	31 (12)

*At least three times per week for at least 3 months during the past 2 years.

^†Combination medications were counted under both categories. For example, the combined preparation of ipratropium bromide and albuterol sulfate was counted as both short-acting ß-agonist and an ipratropium bromide metered-dose inhalers.

^‡Anticholinergic medications included ipratomium bromide and tiotropium bromide. There were no subjects taking tiotropium bromide until the final year of interviewing, in 2004, at which point 2 subjects reported tiotropium usage.

Table 2 Characteristics of 267 adults with COPD

	Mean ± SD or
Characteristic/Measurements	N(%)
Age at interview, years	65.2 ± 6.1
Male gender	99 (37)
Race/ethnicity
White, non-Hispanic	227 (85)
Latino / Hispanic	13 (5)
African American	21 (8)
Asian / Pacific Islander	1 (< 1)
Other	5 (2)
Education^†
High school or less	145 (54)
Some college	76 (28)
College or graduate school	46 (17)
Cigarette Smoking
Never smoked	48 (18)
Current smoker	82 (31)
Former smoker	136 (51)
Comorbidities
Coronary Artery Disease	57 (21)
Congestive Heart Failure	43 (16)
Diabetes	61 (23)
Sleep Apnea	23 (9)
COPD Severity Score (baseline)	7.8 ± 6.3

Table 3 Logistic regression prediction models of health-care utilization for respiratory problems

	Derivation Cohort			Internal Validation
	Hospitalization	Emergency Department Visit	Two or More Outpatient Visits	Hospitalization	Emergency Department Visit	Two or More Outpatient Visits
Model 1: Demographics, comorbidities, & tobacco history^†
Model P-value^‡	.018	.63	.46	.046	.051	.0052
C-index**	80%	64%	59%	79%	73%	70%
Model 2: Model 1 + COPD Severity Score
Odds ratio for COPD Severity Score^§	1.66 (P <. 001)	1.77 (P <. 001)	1.91 (P <. 001)	1.99 (P =. 001)	1.54 (p =. 001)	1.67 (P <. 001)
Model P-value^‡	<. 0001	.0005	<. 0001	.0003	.0012	<. 0001
C-index**	87%	82%	77%	91%	82%	78%
P-value for C-index difference between Model 1 & Model 2^¶	.03	.0031	< .0001	.048	.021	.029

*Derivation cohort used baseline data from 2002 to predict outcomes reported one year later in 2003. Internal validation used baseline data from 2003 to predict outcomes reported one year later in 2004.

^†Covariates in Model 1 consist of age, race, educational attainment, tobacco history, and medical comorbidities (heart failure, coronary artery disease, diabetes, and sleep apnea). Covariates chosen to maximize C-index of Model 1.

^‡Based on likelihood ratio test of global null hypothesis. Model 1 has 8 df. Model 2 has 9 df.

**C-index is area under the receiver operator characteristics curve.

^§Odds ratios expressed per ½ standard deviation increase in COPD Severity Score (P-value).

^¶P-value from χ² comparing C-index.

Table 4 The prospective impact of baseline and annual change in COPD Severity Score

Variable (Covariates Not Shown)	Hospitalization OR (95% CI) P-value	ED Visit OR (95% CI) P-value	Two or More Outpatient Visits OR (95% CI) P-value
COPD Severity Score^†	1.84 (1.48 – 2.29)	1.70 (1.40 – 2.07)	1.76 (1.49 – 2.07)
	P <. 001	P <. 001	P <. 001
One-year change in	1.40 (1.01 – 1.92)	1.46 (1.11 – 1.91)	1.55 (1.22 – 1.97)
COPD Severity Score^†	P =. 043	P =. 007	P <. 001

Data analyzed using generalized estimating equation repeated measure analysis with robust standard errors examining two waves of data. Factors from one wave of data used as prospective predictors of subsequent wave of outcomes.

*Covariates included in multivariate model were: age, race, educational attainment, tobacco status, and self-reported medical comorbidities (congestive heart failure, coronary artery disease, diabetes, and sleep apnea).

^†Odds ratios expressed per ½ standard deviation increase in COPD Severity Score.

Figure 1 Change in COPD Severity Score over time.

Figure 2 COPD severity score as a predictor of utilization outcomes. COPD severity score is modeled at the average value of the model covariates: age, race, education, tobacco history, and medical comorbidities (heart failure, coronary artery disease, diabetes, and sleep apnea).