FIGURE 13
Quarterly Absence Hours and Differences—Low-Risk Diagnosis Cohort
Quarter 1 Quarter 2 Quarter 3 Quarter 4 Quarter 5 Quarter 6
Quarter 7 Quarter 8
Mean Quarterly Absence Hours
Diagnosis Cohort: Cases of Depression
Diagnosis Cohort: Controls
P Value
Raw Difference
Diagnosis Cohort—Cases Minus Controls
Statistically Significant (5%) Results
Diagnosis Cohort—Cases Minus Controls
14.77
8. 34
<0.001
17. 47
9. 50
<0.001
17. 63
9. 33
<0.001
19. 73
9. 31
<0.001
21. 65
10. 14
<0.001
21. 21
10. 28
<0.001
22.98
10. 21
<0.001
19. 29
9. 16
<0.001
6. 44
7.97
8. 30
10. 43
11. 51
10.93
12.77
10. 13
6. 44
7.97
8. 30
10. 43
11. 51
10.93
12.77
10. 13
FIGURE 14
Quarterly Absence Hours and Differences—Low-Risk Rx Cohort
Quarter 1 Quarter 2 Quarter 3 Quarter 4 Quarter 5 Quarter 6
Quarter 7 Quarter 8
Mean Quarterly Absence Hours
Rx Cohort: Cases of Antidepressant Usage
Rx Cohort: Controls
P Value
Raw Difference
Rx Cohort—Cases Minus Controls
Statistically Significant (5%) Results
Rx Cohort—Cases Minus Controls
14. 42
8. 46
<0.001
15. 43
9. 72
<0.001
14. 37
10. 15
<0.001
19. 16
10.77
<0.001
20.98
10. 58
<0.001
19.81
10. 53
<0.001
19. 10
10. 39
<0.001
17. 70
9. 61
<0.001
5.96
5. 71
4. 22
8. 39
10. 40
9. 28
8. 72
8. 10
5.96
5. 71
4. 22
8. 39
10. 40
9. 28
8. 72
8. 10
in the low-risk group also was higher than in the high-risk group. Again,
there was a general upward trend in the
magnitude of the differences as we got
closer to the diagnosis date, with some
fluctuations.
possible explanation for the increasing
costs could be a deepening depression
that leads to a significant adverse medical event that finally triggers a diagnosis
of depression.
What do our findings mean for employers? To put our study in perspective,
consider an employer with 1,000 employees. We estimated that people with
undiagnosed depression cost such an employer nearly $132,560 over a two-year
Implications for Employers and
Insurers
Our study clearly shows that a significant differential exists in excess health
care and absenteeism costs between depressed people prior to their diagnosis
and comparable non-depressed insureds.
As we noted above, the low-risk group
had a larger difference in absence hours
between cases and controls than the
high-risk group. One explanation could
be that the high-risk group tends to go
through its available work leave faster
because of other medical needs. Some
absences that are the result of other co-morbid conditions may be covered by
short-term disability, which was not fully
captured in our data. It’s possible we may
not be seeing the complete picture in our
data for absences in the high-risk group.
We also noticed that the magnitude of
excess costs increased during the two
years prior to diagnosis to a highest level
in the quarter right before diagnosis. One
14
12
10
8
6
4
2
0
period. We developed our estimate based
on the following information:
The estimated prevalence of depres-
sion in the United States is approximately
9. 5 percent.
Based on our analysis of the MarketScan nationwide claims data for
commercially insured working-age people, we saw a prevalence of diagnosed/
treated depression to be 5.0 percent.
This would indicate a prevalence of
FIGURE 15
Significant Differences Between Cases and Controls:
Low-Risk Group, Hours Absent
Diagnosis Cohort
Rx Cohort
Quarter 1
Quarter 2
Quarter 3
Period Prior to Diagnosis
Quarter 4 Quarter 5
Quarter 6
Quarter 7
Quarter 8
JUL | AUG. 12 CONTINGENCIES 69
