Table. Indicators of short- and medium-term success of rehabilitation.

Parameter	Change during rehab (T1−T0) SES or ± percentage point p-value	Measuring point	Predictive value* for
			Returning to work OR [95% CI]	Physical quality of life (SF-12, PCS) β [95% CI]	Mental quality of life (SF-12, MCS) β [95% CI]
Physical endurance abilitykörperliche Leistungsfähigkeit
Endurance training load in ergometer training (5 Watt)	1.03; < 0.001	T0 TΔ	1.09 [1.04; 1.14] –	0.15 [0.03; 0.27] 0.13 [0.01; 0.24]	– –
Subjective well-being/quality of life
Depression (PHQ-9; points)	0.42; < 0.001	T0 TΔ	–	–	−0.41 [−0.60; −0.22]
Subjective well-being (WHO-5; 10 points)	0.71; < 0.001	T0 TΔ	– –	– –	– 0.36 [0.03; 0.69]
Physical quality of life (SF-12; 10 points)	0.53; < 0.001	T0 TΔ	1.41 [1.20; 1.66] –	4.26 [3.44; 5.08] 2.98 [2.21; 3.75]	– –
Mental quality of life (SF-12; 10 points)	0.49; < 0.001	T0 TΔ	1.48 [1.25; 1.75] 1.30 [1.06; 1.60]	0.48 [0.03; 0.92] –	1.20 [0.46; 1.93] –
Physical health (IRES-24; 2 points)	0.43; < 0.001	T0 TΔ	– –	1.18 [0.54; 1.81] 0.78 [0.13; 1.43]	– –
Mental health (IRES-24; 2 points)	0.57; < 0.001	T0 TΔ	– –	– –	2.90 [1.91; 3.88] 2.54 [1.67; 3.42]
Self-assessed health prognosis (good/very good)	+7.8; < 0.001	T0 T1	– –	1.11 [0.05; 2.16] 1.58 [0.49; 2.68]	– –

Parameters are shown that are useful for mapping both the short-term (changes of at least 5 percentage points for categorical variables or standardized effect sizes of > 0.35 [interpretation analogous to Cohen‘s d] for metric variables with proven statistical significance) as well as the medium-term rehabilitation success (predictive power for return-to-work and / or quality of life at six months after discharge) of cardiac patients proved to be suitable. In the domains „cardiovascular risk factors“ (e.g., smoking behavior, blood pressure) and „social medicine“ (such as self-?assessed work ability prognosis or desire for retirement), no parameters were identified that could be used to operationalize both short-term and medium-term rehabilitation success.

* The predictive power was determined in multivariable regression models with backward selection (only the effects in the final models are given). Parameters were included in the models based on their respective characteristics at the start and end of rehabilitation, or as the difference between the two measurement times. Variables considered in the models were age, gender, life situation, level of education, type of rehabilitation (follow-up rehabilitation / general application procedure), diagnoses, comorbidities, experience of stress, smoking behavior, motivation for lifestyle change, blood pressure, cholesterol, body mass index, fasting glucose, physical performance, maximum exercise capacity, endurance load in ergometer training, 6-min walking distance test, pain-free walking distance, desire for retirement, self-assessed occupational prognosis, work capacity, subjective health, depression, anxiety, physical stress, well-being, physical / mental quality of life, subjective physical and mental health and pain, self-assessed health prognosis.

95% CI, 95% confidence interval; ß, standardized correlation coefficient; IRES-24, Indicators of Rehabilitation Status-24; MCS, mental component summary; OR, odds ratio; PCS, physical component summary; PHQ-9, Patient Health Questionnaire-9; SES, standardized effect size; SF-12, Short-Form Health Survey; T0, start of rehabilitation; T1, end of rehabilitation; TΔ, rehabilitation course (T1 - T0); WHO-5, World Health Organization Well-Being Index.