Table 1. Description of data used to study the effects of various risk factors on the duration of hot flashes in US women aged 45–54.

Parameter	Level	Women with hot flashes	Women with observed end to hot flashes	Mean duration of hot flashes (range) in years	Mean time to peak severity (range) in years	p-value* for duration	p-value† for time to peak severity
Race	White	421	164	2.32 (1–25)	2.75 (0–20)	0.06	0.08
	Non-White	224	73	2.93 (1–29)	3.5 (0–14)
Education	Graduated College	404	156	2.32 (1–29)	2.96 (0–20)	0.03	0.38
	Did Not Graduate College	241	81	2.83 (1–26)	2.98 (0–20)
Ever Pregnant	Yes	574	208	2.54 (1–29)	2.93 (0–20)	0.21	0.52
	No	72	30	2.23 (1–18)	3.25 (0–16)
Menopause Status‡	Pre	339	133	2.23 (1–25)	2.53 (0–15)	<0.001	0.16
	Peri	483	80	2.58 (1–29)	3.07 (0–20)
	Post	236	24	3.88 (1–12)	4.12 (0–18)
Smoking‡	Current	151	18	3.89 (1–29)	3.05 (0–20)	0.002	0.92
	Former	369	86	2.52 (1–26)	2.95 (0–20)
	Never	467	134	2.31 (1–25)	2.97 (0–19)
BMI‡	<25	376	95	2.48 (1–29)	2.42 (0–19)	<0.001	0.02
	25–30	314	72	2.36 (1–26)	3.33 (0–20)
	>30	366	71	2.68 (1–25)	3.35 (0–19)
Alcohol Consumed‡	≥12 Drinks/ Year	536	171	2.44 (1–29)	2.47 (0–20)	<0.001	<0.001
	<12 Drinks/ Year	291	48	3.12 (1–25)	4.78 (0–20)
Leisure Activity‡	Much Less	51	9	1.22 (1–3)	4.33 (0–14)	0.03	0.82
	Less	190	39	1.92 (1–14)	2.82 (0–20)
	As Much	251	59	1.78 (1–12)	3.13 (0–19)
	More	176	36	1.86 (1–26)	3.61 (0–11)
	Much More	66	19	1.16 (1–3)	1.81 (0–19)
Work Activity‡	Much Less	116	25	1.92 (1–13)	3.54 (0–11)	0.76	0.21
	Less	244	56	1.96 (1–14)	3.12 (0–14)
	As Much	235	53	1.57 (1–26)	3.1 (0–20)
	More	91	16	2.12 (1–7)	3.22 (0–19)
	Much More	24	8	1.62 (1–6)	2.18 (0–12)
Estradiol Levels‡^#	<42	493	110	2.2 (1–13)	3.11 (0–20)	<0.001	0.10
	≥42	427	128	2.77 (1–29)	2.79 (0–20)
Progesterone Levels‡**#	<0.2	497	92	2.35 (1–13)	3.28 (0–19)	<0.001	0.09
	≥0.2	442	146	2.6 (1–29)	2.67 (0–20)
Testosterone Levels‡**	<0.34	442	95	3.04 (1–29)	3.07 (0–20)	0.18	0.96
	≥0.34	483	143	2.15 (1–26)	2.9 (0–19)
Total		647	238	2.50 (1–29)	2.97 (0–20)

^*p-values for duration are based on an accelerated failure time model with exponential distribution and interval censoring.

^†p-values for time to peak severity are based on a linear regression model with baseline data for all covariates.

^‡These parameters are time-dependent, and may vary throughout the study. Therefore, the sum of the number of women in each level of a parameter may be greater than the total number of women in the study.

^**These parameters are included as continuous variables in the analysis; levels for this table were chosen based on the median value in the sample.

^#These parameters were log-transformed for statistical analysis.