Skip to main content

Table 2 Fulfilment of the validity assumptions for the use of a self-controlled design

From: Self-controlled designs in pharmacoepidemiology involving electronic healthcare databases: a systematic review

 

Case-crossover

Self-controlled case-series

N = 53

N = 55

Major assumptions:

 Type of studied exposure

  Intermittent/transient

34 (64)

49 (89)

  Sustained

11 (21)

5 (9)

  Unclear

8 (15)

1 (2)

 Event onset

  Abrupt

52 (98)

53 (96)

  Insidious

1 (2)

2 (4)

  Rare event

46 (87)

48 (87)

  Recurrent event (for self-controlled case series only)

-

42 (76)

  Rare and/or recurrent event (for self-controlled case series only)

-

55 (100)

All major validity assumptions fulfilleda

35 (66)

48 (87)

Minor assumptions:

Same opportunity of exposureb

34/35 (97)

-

 Time trend in exposure

  Yes

8/35 (23)

-

   In case of exposure time trend, the design is adapted

5/8 (63)

 

  No

5/35 (14)

-

  Unclear

22/35 (63)

-

Absence of exposure time trend (or design adapted)

32/35 (91)

 

 Independence of consecutive events

  Yes

-

5/48 (10)

  No

-

17/48 (35)

   For non-independent recurrences, the design is adapted

 

16/17 (94)

  Unclear

-

14/48 (29)

  Not applicable (for non-recurrent events)

-

12/48 (25)

Independence of consecutive events when recurrent (or design-adapted)c

-

47/48 (98)

 Event-independent exposure

  Yes

-

5/48 (10)

  No

-

38/48 (79)

   For non-event-independent recurrences, the design is adapted

 

32/38 (84)

  Unclear

-

5/48 (10)

Event-independent exposure (or design-adapted)d

-

42/48 (88)

 The event affects the short-term mortality probability (censoring of the observation period)

-

 

  Yes

-

12/48 (25)

   For event-dependent censoring, the design is adapted

 

4/12 (33)

  No

-

36/48 (75)

Absence of censoring the observation period (or design-adapted)e

-

40/48 (83)

All minor validity assumptions fulfilled for use of a self-controlled design (among articles with valid major assumptions)f

31/35 (89)

34/48 (71)

All major and minor validity assumptions fulfilled for use of a self-controlled design (among all articles)

31/53 (58)

34/55 (62)

  1. Data are reported as n (%)
  2. aFor case-crossover design, a transient or intermittent exposure, an abrupt onset event, and a rare event. For self-controlled case series, a transient or intermittent exposure, an abrupt onset event, and a rare and/or recurrent event. Unclear exposures were considered intermittent
  3. bSame opportunity of exposure during case and control time periods
  4. cThis assumption is fulfilled for studies examining non-recurrent events, recurrent events with independence between consecutive events, or non-independent recurrences when the design is adapted, or recurrent events for which independence is unclear
  5. dThis assumption is fulfilled for studies examining exposure whose probability is not affected by previous events, with an adapted design when probability of exposure is affected by a previous event, or with unclear event-independent exposures
  6. eThis assumption is fulfilled for studies examining outcomes that do not censor the observation period by affecting the short-term mortality probability or with an adapted design when the observation period can be censored by the outcome
  7. fFor case-crossover design, the opportunity for exposure should be the same during the case and control time periods, and there should not be any time trend in exposure. For self-controlled case series, two consecutive events should be independent if they are recurrent; the probability of further exposure should not be affected by a previous event; and the event should not affect the short-term mortality probability nor censor the observation period. In case of violation of one of these assumptions, the design should be adapted