Determinants of Frequent Attendance of Outpatient Physicians: A Longitudinal Analysis Using the German Socio-Economic Panel (GSOEP)

Reviewer #1

This was an interesting study to   review and well designed. I found the use of services declined as one aged to   be interesting and I think this would warrant further research.

First, thank you very much to   review this manuscript. We really appreciate your comments.

Lines 240-241 "additional reflections"   of the spouse is not clear. Would you strengthen this statement so if is not   ambiguous?

Thank you for your comment.

We changed it accordingly:

A possible   explanation for the increased likelihood of becoming a frequent attender post   marriage could be an increased awareness regarding the personal health status   due to the additional health related advises provided by the spouse

I feel the conclusion could be   stronger by emphasizing the "so what" factor of the study's   findings.

Thank you for your comment.

Findings of this   study showed that need characteristics are particularly important for the   onset of frequent attendance. This might indicate that persons begin to use   health services frequently especially when medically indicated. Consequently,   finding ways to increase health may help to reduce the economic burden caused   by frequent attendance

Again, thank you for your helpful   comments. It helps to improve the quality of the manuscript.

Reviewer #2

I wish the authors could give us   some more information on how the SF-12 physical and mental component summary   scales were calculated.

Thank you for your comment.

We added:

Single items and subscales are computed to the two main dimensions   using norm-based scoring and factor analysis. Values of the two scales have a   range between 0 and 100 with a mean of 50 and a standard deviation of 10 [18,28].

Moreover, it seems that in the   reversed scales higher scores would indicated worse and not better health.

Thank you for your comment.

We changed it accordingly:

reversed scales   were used (MCS reversed ranging from 1.27 to 79.63, with higher values   corresponding to worse mental health; PCS reversed ranging from 9.20   to 79.60, with higher values corresponding to worse physical health)

It would also be good if the author   could explain the statistical model in more detail

Thank you for your comment.

We added:

Due to the dichotomous characteristic of the dependent variable a   logit regression model with person specific errors was applied. A logit   regression estimates the effect of several independent variables on the   unobserved probability of the binary dependent variable [31]. Using a   conditional likelihood function the person specific errors cancels out. Thus,   the estimator is consistent, even if it is correlated with time-invariant,   person-specific unobserved heterogeneity [33]. It should be noted that   individuals who have the same constant dependent result over the whole   observation period are excluded from the estimation. This can reduce the   dataset but does not evoke sample selection bias, because individuals do not   contain information for the estimation of the coefficients [31].

how they tested for   multi-collinearity as well as the application of the Hausman test

Thank you for your comment.

We added:

Multicollinearity   was not detected, according to low bivariate correlation values of the   included variables and below common threshold values of the Variance   Inflation Factor

The results of   the Hausman-Test (Stata command: hausman) strongly reject the null   hypothesis that a random effects (RE) model should be applied (p<0.001).   This result suggests the use of a FE estimator [35].

It would make the tables more   readable if the authors could use superscript for the footnotes.

Thank you for your comment.

We changed it accordingly. Please   see the Tables for further details.

I also do not understand why in the   discussion as well as in the abstract it is said that enabling   characteristics were not significantly associated with the outcome. There is   obviously a period effect for the year 2004 indicating that the introduction   of the practice fee might have reduced attendance of outpatient physicians.

Thank you for your comment.

We rearranged the sentence:

Except for the period effect of the introduction of the “practice fee”   [10], enabling characteristics were not associated with the outcome measure   in our study. That can be explained by the fact that individuals insured with   the statutory health insurance can easily consult a physician free of public   copayment.

And added a subordinate clause in   the abstract:

Enabling   characteristics were not significantly associated with the outcome measure,   expect for the onset of the “practice fee”.

A further limitation that the   authors could discuss is that hospitalizations were not considered in this   study.

Thank you for your comment.

We added:

Despite the rich   set of sociodemographic variables included in this analysis, the analysis   lacks evaluated need-factors, such as chronic diseases due to data   unavailability. A further limitation is that there is no differentiation in   the dataset between primary physicians and specialists. Regarding the   dependent variable, only a three-month period for physician visits was   captured. However, a three month time horizon was often used in previous   studies. Furthermore, it has been demonstrated that the time horizon is   suitable and the recall bias is negligible [37]. Moreover, future studies   are required to clarify the determinants of frequent hospital visits   longitudinally.

Again, thank you for your helpful   comments. It helps to improve the quality of the manuscript.