Analysis For Prevention of Latelife Depression

1
Analysis For Prevention of Late-life Depression

• Qin Yu
• Department of Biostatistics and Computational
  Biology
• University of Rochester

2
Research Context

• Context Late-life depression is associated with
  substantial societal costs through its disease
  burden and prognosis.
• Prevention may be an attractive means to generate
  health gains and reduce future costs.
• Objectives To target high-risk groups for
  depression prevention such that maximum health
  gains are achieved.
• Design Population-based cohort study over 4
  years.
• Participants 745 community residents aged 65 to
  97 years.

3
Analysis Methods

• Subjects and procedures
• At baseline, 745 people were interviewed aged
  from 65 to 97 years. These people were followed
  for 4 years, all the predictors are recorded.
• Two separate analyses
• limited sample to those without initial major
  depression, among 617 people at risk, 33 became
  depressed this estimate is equal to 4.4 cases
  per 100 person-years
• limited sample to those without initial major or
  minor depression, among 533 people at risk, 96
  became depressed.

4
Statistical Analysis

• The analyses were conducted in several steps.
• First Step
• Calculation of the absolute and relative risk of
  becoming a case for subjects with or without
  being exposed to a certain risk indicator.
• These risk indicators were dichotomized.

5
Statistical Analysis

• Second step
• The importance of risk indicators was
  quantified by computing the level of exposure to
  a risk factor in the population.
• Four risk indicators
• Exposure Rate (ER)
• The incidence rate ratio (IRR)
• Population attributable fraction (AF)
• The number needed to be treated (NNT)
• 
  
  
  

6

• Table 1a. Exposure Status (in Percent) in the
  Group at Risk and the incident Group limited to
  those without initial major depression.

7
(No Transcript)
8

• Table 1b. Exposure Status (in Percent) in the
  Group at Risk and the incident Group limited to
  those without initial major or minor depression.

9
(No Transcript)
10
Statistical Analysis

• To summarize
• We calculate the ER, IRR, AF, and NNT. Together,
  these indicators of impact allow us to select
  high-risk groups for which depression prevention
  is likely to be associated with the highest
  health benefit at the lowest cost.

11
Statistical Analysis

• Third Step
• High-risk group selection process was carried out
  by backward-stepping procedures
• We selected the smallest set of risk indicators
  in which each risk indicator has a unique and
  significant contribution to the prediction of
  depression.

12
Results

• Parsimonious Model
• Table 3a Parsimonious model for the incident
  Group limited to those without initial major
  depression.

13
Results

• Parsimonious Model
• Table 3b Parsimonious model for the incident
  Group limited to those without initial major or
  minor depression.

14
Comments

• The most promising risk indicator was selected
  from Table 3a and Table 3b, which was followed by
  consecutively selecting and adding risk
  indicators in such a way that the values for
  potential health benefit (IRR and AF) were kept
  as high as possible and the cost (ER and NNT) as
  low as possible.

15
Results

• Selecting a smaller set of risk indicators
• Limited sample to those without initial major
  depression

16
Results

• Selecting a smaller set of risk indicators
• Limited sample to those without initial major
  depression

17
Statistical Analysis

• The fourth step analyses for joint exposures.
• Assessed the potential health benefits when
  prevention efforts target people who are exposed
  to combination of multiple risk indicators.
• Took SSD as the starting point because this risk
  indicator is associated with the highest IRR and
  AF values and has the lowest NNT and ER.

18
Results
Table 4a Additive effect of Joint Exposures
limited to the sample without initial major
depression.

19
Results

• Table 4b Additive effect of Joint Exposures
  limited to the sample without initial major or
  minor depression.

20
Comments Quantifying the health gains

• Consider the risk profile in Table 4a and 4b,
  e.g. in row 11 of table 4b, the risk of becoming
  depressed is a factor 1.39 higher in female with
  SSD and has received high education.
• If the adverse effect of the joint exposure could
  be completely blocked, then the incidence rate in
  the population would be reduced by 20.4.
• The current incidence rate of 4.4 cases per 100
  person-years would then become 4.4 x
  (1-20.4)3.5 cases per 100 person-years. In
  every 1 million older adults, this would
  represent roughly 9,000 prevented cases per year.

21
Comments Quantifying the health gains

• Assuming that an intervention is 30 successful
  in avoiding new onsets, then 2,700 onsets will be
  avoided.
• Assume the excess costs of minor and major
  depression as at least 1,045 per person per half
  year. Preventing 2,700 onsets would thus result
  in a cost saving of at least 2.82 million for
  every million elderly people in the population.
• If the costs of the intervention do not exceed
  1,045 per avoided case, we can still use AF and
  NNT value to select interesting preventive
  interventions for further cost-effectiveness
  research.

22
Comments

• We aimed to answer the question as to whether it
  would be possible to target depression prevention
  where it generates the best health benefits at
  the lowest cost.
• This would also guide research efforts toward
  more promising research areas in preventive
  psychiatry.