We interviewed 201 physicians for this study. Table 2 indicates the socio-demographic and professional characteristics of the participating physicians. The mean age of the physicians was 48.5±11.2 years and 53.5% were female. About half of the physicians (48.7%) were born and graduated in Israel and most of them (67.2%) defined themselves as secular.
The participants were experienced physicians, with a mean of 18.7±12.2 years of work and 16.2±11.0 years as PCPs. Among the sample, 84.2% were board-certified physicians of whom 85.3% were board certified family physicians, mostly employed in HMO clinics.
Relating to the physicians’ own family, 39.1% had dependent elderly family members, and about 34% had been caregivers for severely ill elderly family members with a mean duration of 44.3±57.2 months of caregiving.
Most of the physicians always (47.8%) or often (35.8%) identified that the caregiver in families with a severely ill or disabled member was at high risk for morbidity or mortality and recommended they receive preventive care.
The physicians’ recommendations for preventive care for the caregivers included four elements. Most of the physicians recommended treatment for the caregiver often (39.8%), or always (26.0%), to prevent a decline in their health. Regular physical activity was recommended always (41.7%) or often (38.5%) together with good sleep habits. Physicians advocated that the caregiver seek regular help from a social worker or a psychologist (35.7%), or frequently (31.1%),
Many of the physicians (60.5%) stated that they invited caregivers for a follow-up visit on their own initiative. Moreover, regular clinic appointments for the caregiver were initiated by 56.6% of the doctors over the prior six months.
The association of the study indices with socio-demographic characteristics varied. Female physicians were more likely to recommend preventive care than males (p= 0.04). Israeli medical graduates were more aware of preventive care than those who studied abroad, but abroad medical graduates were more likely to recommend preventive care than those who studied in Israel (p= 0.005). Specialists were more liable to make recommendations for preventive care than non-specialists, and family doctors were more likely to do so than their colleagues (p= 0.001).
Physicians who experienced being caregivers themselves were more likely to recommend preventive care than those who were not (p= 0.03). There were no statistically significant differences in physicians’ awareness, recommendations for preventive care, and in the treatment follow-up related to religiosity or place of work. (Table 1).
We conducted a multivariate linear regression analysis to identify the unique contribution of the study variables to the explanation of physicians’ awareness of the risks of caregiving (Table 3). Since the variable “duration of caregiving” for a sick family member was significantly associated with risk awareness in the univariate analyses, and only one third of the study physicians answered this item, it was added to the multiple regression analysis as a dichotomous variable referring to whether the physician himself was a caregiver for a severely ill family member. The model for the regression analysis on physicians’ awareness was found to be significant (F[3,172]=6.65, P<0.000). Only two variables made a unique contribution to physicians’ awareness: Board certified family physicians and physicians who graduated medical schools in Israel were more aware of the risks of caregiving compared to the others. In total, these disparities explained a relatively low percentage (11%) of the variance of this dependent variable. Surprisingly, being a caregiver for a sick family member did not contribute significantly to physicians’ awareness of the risks of caregiving.
Table 4 shows the multivariate linear regression analysis to identify variables that explain physicians’ recommendations for preventive treatment. These variables were introduced into this analysis in two steps. In the first step, we entered the socio-demographic and professional variables (sex, country of medical school graduation, specialist/non-specialist, primary caregiver). Only variables that were significantly associated with the dependent variable in the univariate analysis were considered.
Board certification emerged significant. Country of medical school graduation was considered to be of statistical significance (F=0.13, P=0.08).
In the second step, we added the variable “physician’s awareness of the risks of caregiving”. This variable explained another 8% of the variance of the dependent variable and was statistically significant (F[1,185]=17.6, P<0.000). Thus, physician’s awareness represents a substantial contribution to the 16.5% percentage of the stated variance. Three variables made a significant contribution: physician’s awareness of the risks of caregiving, country of medical school graduation and board certification (Table 4).
We did not conduct a multivariate analysis on “treatment & follow-up” because no independent variable was significantly associated with it in the univariate analysis.