Table 2.

Direction of Effects Reported for 6 Health Behaviors by Study (n=34).

					Health Behavior
Study			Physical activity	Nutrition	Sleep quality	Alcohol intake	Tobacco use	Body weight
Type of study	# of waves	d
Prospective longitudinal*
Avis et al., 1991bb	2	-	stablec			no change	stable
Eng et al., 2005 a	2	.03, .25, .06, -, .50	decreasedcd	worsened		increased	decreased	decreased
Janke et al., 2008a & 2008b	2	-	decreasede
Janke et al., 2008c b	2	.13	increasedd
Lee et al., 2005 b	2	-	increasedc	worsened		increased	increased	decreased
Schulz et al., 2001	2	.62, .50	increasedcg					decreasedh
Shahar et al., 2001	2	.97, 1.20		worsened				decreased
Tran, 2007	2	-	stablec	stable
Wilcox et al., 2003 b	2	.04, .19, .02, .02	increasedc	worsened			decreased	decreased
Williams, 2004	2	-	^		^		^	^
Postbereavement longitudinal*
Anderson, 1999	2	1.42			declined
Byrne & Raphael, 1997 a	3	.48			declined
Byrne et al., 1999 a	3	.43				increased
Caserta et al., 2001	2	.67	decreasedc
Chen et al., 2005	2	-	stablec	worsened	stable
McIntyre & Howie, 2002 b	3	-	stabled
Monk et al., 2009	14	1.41			declined
Nurriddin, 2008	3	.08, .51, .28, .02	decreasedc	worsened		increased	stable
Pasternak et al., 1992	2	1.11			declined
Quandt et al., 2000	5	-		worsened
Reynolds III et al., 1992	3	1.27			declined
Reynolds III et al., 1993	5	.46			stable
Wilcox & King, 2004	3	.38	decreasedbf

Type of study	Control group	d
Cross-sectional*
Barrett & Schneweis, 1980	no	.34		NR
Fitzpatrick et al., 2001 a	yes	.12	no differencecd
Johnson, 2002	yes	.92		worsened
Okun et al., 2011	yes	.15, .65, .01, -	decreasedc		declined	no difference	stable
Patterson, 1996; Patterson &	no	-	decreasedc
Carpenter, 1994			increasedd
Rosenbloom & Whittington, 1993	yes	1.75, 1.60		worsened				decreased
Schone & Weinick, 1998	yes	.92,.73, .70	decreasedc	worsened			increaseda
Welte & Mirand, 1993	yes	-				no difference
Wylie et al., 1999	no	-		worsened

NR=did not report results in reference to a control sample. Effect sizes were estimated for quantitative studies that reported means and standard deviations for outcome variables.

^*Results are reported for widowed elders across time.

^**Results are reported for widowed elders in reference to a control sample.

^-Did not report results that allowed for the estimation of effect sizes.

^{^}This study created a “health-risk” variable by summing 4 behavioral variables. Widowhood was associated with an increase in health-risk behaviors, but the effect of bereavement on each of these health behaviors was not reported.

^aSignificant effect reported for widowed men only.

^bSignificant effect reported for widowed women only.

^cSports/exercise activity.

^dLeisure/social activity.

^ePhysical activity and exercise only; elders increased their social activities across time.

^fClass-based exercise only; home-based exercise remained stable.

^gAmong caregivers (who experienced strain) only.

^hAmong noncaregivers only.