Using heart rate profiles during sleep as a biomarker of depression

BMC Psychiatry

Table 7 Algorithm sensitivity across the testing sample depression group stratified by potential confounding factors

	Stratification	Sensitivity (%)^a	n	χ ²(df)	p
Body Mass Index	Normal	61.9	21	10.5 (2)	.005
	Overweight	83.3	36
	Obese	96.7	30
Psychotropic Medication Use	Yes	87.0	69	4.1 (1)	<.050
Psychotropic Medication Use	No	66.7	18	4.1 (1)	<.050
Depression Severity	Mild	83.9	31	0.2 (2)	.927
	Moderate	80.6	31
	Severe	84.0	25
Psychiatric Comorbidity	Yes	75.9	29	1.5 (1)	.229
Psychiatric Comorbidity	No	86.2	58	1.5 (1)	.229
Cardiovascular Disease or Related Risk Factor	Yes	83.9	31	0.0 (1)	.838
Cardiovascular Disease or Related Risk Factor	No	82.1	56	0.0 (1)	.838
Cardiovascular Medication Use	Yes	82.6	23	0.0 (1)	.982
Cardiovascular Medication Use	No	82.8	64	0.0 (1)	.982
Smoking Status	Non-Smoker	82.5	63	0.0 (1)	.930
Smoking Status	Current Smoker	83.3	24	0.0 (1)	.930
AHI	< 5	79.7	69	2.2 (1)	.141
AHI	≥ 5	94.4	18	2.2 (1)	.141
Sleep Onset Latency (min)	≤ 15	82.2	45	0.0 (1)	.891
Sleep Onset Latency (min)	> 15	83.3	42	0.0 (1)	.891
REM Onset Latency (min)	≤ 164	80.0	40	0.7 (1)	.390
REM Onset Latency (min)	> 164	87.2	39	0.7 (1)	.390
%REM	≤ 13	90.9	44	4.1 (1)	<.050
%REM	> 13	74.4	43	4.1 (1)	<.050

^aThe algorithm’s ability to accurately detect cases of depression (i.e. true positive rate)
AHI Apnea-Hypopnea Index, REM Rapid eye movement

ISSN: 1471-244X