Underuse of self-controlled designs in pharmacoepidemiology in electronic healthcare databases: a systematic review

A Novel Randomized Trial Protocol for Evaluating Wound Healing Interventions

Background: Randomized controlled trials using complete healing as an endpoint suffer from poor statistical power, owing to the heterogeneity of wounds and their healing trajectories. The Food and Drug Administration (FDA) has recently consulted with expert groups to consider percentage area reduction (PAR) of the wound over a 4-week period as a valid intermediate endpoint, creating the opportunity for more powerful study designs. Methods: A within-subject controlled study design comparing the PAR of venous leg ulcers (VLU) in patients over 4 weeks receiving different interventions. Twenty-nine patients received multilayer compression over 4 weeks, followed by neuromuscular electrostimulation (NMES) of the leg muscle pump in addition to compression for a further 4 weeks. Paired comparison was then made of PAR between the two phases. A second cohort of 22 patients received only multilayer compression throughout both 4-week phases. Results: Patients randomized to NMES saw a significant increase in healing rate compared with compression alone, whereas patients receiving compression only saw no significant change in healing rate throughout the course of the study. Conclusions: Intermittent NMES of the common peroneal nerve significantly accelerates the healing of VLU. It is well tolerated by patients and deserves serious consideration as an adjuvant to compression therapy. PAR is a useful metric for comparing the performance of wound healing interventions, and the self-controlled trial design allows sensitive discrimination with a relatively small number of subjects over a reasonably short trial period. The study is reported according to the CONSORT reporting guidelines. Clinical Trial Registration: NCT03396731 (ClinicalTrials.gov).

Self-Harm and Suicide Attempts Preceding and Following Electroconvulsive Therapy: A Population-Based Study

OBJECTIVE

The beneficial effect of electroconvulsive therapy (ECT) on suicidality has been documented in clinical trials, whereas naturalistic studies on the topic are scarce and restricted to individuals with mood disorders. Here, based on population-based data from Danish registers, we aimed to investigate the course of self-harm and suicide attempts preceding and following ECT across 4 major mental disorders. This was done to examine whether data from the real-world clinical setting are compatible with the positive results from clinical trials.

METHODS

We identified all patients diagnosed with unipolar depression (n = 8843), bipolar disorder (n = 2713), psychotic disorder (n = 2692), or personality disorder (n = 2085) who received ECT for the first time in the period from 2008 to 2019, as well as age-, sex-, diagnosis-, illness duration-, and admission-matched comparison groups not receiving ECT. A mirror-image model was used to examine whether the number of incidents of self-harm/suicide attempts changed following ECT (paired t test).

RESULTS

There were substantial and statistically significant reductions in the number of incidents of self-harm/suicide attempts when comparing the month leading up to and the month following initiation of ECT for all diagnostic groups (unipolar depression: reduction, 83% [P < 0.001]; bipolar disorder: reduction, 72% [P < 0.001]; psychotic disorder: reduction, 82% [P < 0.001]; personality disorder: reduction, 83% [P < 0.001]). The analog results for the comparison groups not receiving ECT suggested that these reductions in self-harm/suicide attempts were partly mediated by a protective effect of admission.

CONCLUSIONS

Data from the real-world clinical setting are compatible with results from clinical trials with regard to the protective effect of ECT on suicidality.

The use of stimulants in depression: Results from a self-controlled register study

OBJECTIVE

To investigate the effectiveness of stimulants in patients with depression, by using naturalistic outcome measures, such as psychiatric admissions, psychiatric bed-days and incidents of intentional self-harm or suicide attempts.

METHODS

Via linkage of the Danish nationwide health registers, we identified all patients with a diagnosis of depression initiating stimulants, including methylphenidate, modafinil, amphetamine, dexamphetamine or lisdexamphetamine, from 1995 to 2012. We used a mirror-image model to test whether redemption of a stimulant prescription was associated with a reduction in psychiatric admissions, inpatient days and incidents of intentional self-harm or suicide attempts. Specifically, the number of these outcomes in the 2 years leading up to redemption of a stimulant prescription was compared to the two subsequent years. Similar outcomes were used in a reverse mirror-image model to investigate the effect of stimulant termination.

RESULTS

A total of 3354, 935 and 105 patients diagnosed with depression redeemed prescriptions for methylphenidate, modafinil or amphetamine/dexamphetamine/lisdexamphetamine, respectively. Initiation of methylphenidate was not associated with a significant change in psychiatric admissions (mean: -0.02 admissions, p = 0.11) or inpatient days (mean: 0.13 days, p = 0.74). Similar findings were made for modafinil and the amphetamines. In addition, no clinically relevant change in psychiatric admissions or inpatient days was found after termination of a stimulant. After initiation of methylphenidate, the incidents of self-harm or suicide attempts were reduced by 54%, from 68 to 31 events (p = 0.004). No significant change in incidents of self-harm or suicide attempts were found for modafinil or the amphetamines.

CONCLUSION

This nationwide study, using naturalistic outcomes, does not support the use of stimulants in patients with depression. However, the use of methylphenidate was associated with a 54% reduction in incidents of self-harm or suicide attempts, indicating that methylphenidate may potentially be useful in patients with depression with suicidal- or self-harming behaviour. However, further studies are needed, before any firm conclusions can be made.

Seasonal influenza vaccine and Guillain-Barré syndrome: A self-controlled case series study

OBJECTIVE

To evaluate the risk of Guillain-Barré syndrome (GBS) following seasonal influenza vaccination based on French nationwide data.

METHODS

All cases of GBS occurring in metropolitan France between September 1 and March 31 from 2010 to 2014 were identified from the French national health data system. Data were analyzed according to the self-controlled case series method. The risk period started 1 day after the patient received vaccine (D1) until 42 days after vaccination (D42). The incidence of GBS during this risk period was compared to that of the control period (D43-March 31). The incidence rate ratio (IRR) was estimated after adjusting for seasonality and presence or not of acute infections.

RESULTS

Between September and March, of the 2010/2011 to 2013/2014 influenza vaccination seasons, 3,523 cases of GBS occurred in metropolitan France and were included in the study. Among them, 15% (527 patients) had received influenza vaccination. A total of 140 patients developed GBS during the 42 days following influenza vaccination. The crude risk of developing GBS was not significantly increased during the 42 days following influenza vaccination (IRR, 1.02; 95% confidence interval [CI], 0.83-1.25; p = 0.85). This result remained nonsignificant after adjustment for calendar months and the incidence of acute gastrointestinal and respiratory tract infections (IRR, 1.10; 95% CI, 0.89-1.37; p = 0.38). In contrast, the risk of GBS was fourfold higher after acute respiratory tract infection (IRR, 3.89; 95% CI, 3.52-4.30; p < 0.0001) or gastrointestinal infection (IRR, 3.64; 95% CI, 3.01-4.40; p < 0.0001).

CONCLUSIONS

No association between seasonal influenza vaccination and GBS was shown during the 42 days following vaccination.

Antidepressant use during pregnancy and risk of congenital heart defects: A case-time-control study

PURPOSE

We estimated the association between maternal antidepressant (AD) use in early pregnancy and risk of congenital heart defects.

METHODS

We applied a case-time-control design with the aim of controlling for confounding from time-invariant factors and compared the results of the design to results from a cohort design in a population of 792 685 singletons born alive in Denmark during 1995-2008. In the case-time-control design, we identified children diagnosed with a congenital heart defect in the first 5 years of life (cases) and compared maternal AD use in the risk period (the first 3 months of pregnancy) and the reference period (gestational months 5-7). A nondiseased control group was included to adjust for time trends of exposure. In the cohort design, we identified children whose mothers redeemed at least one AD prescription in the first 3 months of pregnancy (the exposed) and two other groups including the unexposed children with maternal AD prescriptions in the 12 months before pregnancy. We applied conditional logistic regression and logistic regression to compute odds ratios (ORs) and 95% confidence intervals (CIs).

RESULTS

The case-time-control OR for any congenital heart defect was 1.03 (95% CI, 0.61-1.73), which was similar to the OR (1.09, 95% CI, 0.88-1.35) from the cohort design when we compared the exposed children with the unexposed children with maternal AD use before pregnancy.

CONCLUSIONS

The case-time-control design provided results similar to the cohort design when the cohort design had a better confounder control strategy. We discussed the strengths and drawbacks of case-time-control design.

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

Background

Observational studies are widely used in pharmacoepidemiology. Several designs can be used, in particular self-controlled designs (case-crossover and self-controlled case series). These designs offer the advantage of controlling for time-invariant confounders, which may not be collected in electronic healthcare databases. They are particularly useful in pharmacoepidemiology involving healthcare database. To be valid, they require the presence of some characteristics (key validity assumptions), and in such situations, these designs should be preferred. We aimed at describing the appropriate use and reporting of the key validity assumptions in self-controlled design studies.

Methods

Articles published between January 2011 and December 2014, and describing a self-controlled study design involving electronic healthcare databases were retrieved. The appropriate use (fulfilment of key assumptions) was studied in terms of major (abrupt onset event, rare or recurrent event, and intermittent exposure) and minor assumptions (those for which the design can be adapted).

Results

Among the 107 articles describing a self-controlled design, 35/53 (66%) case-crossover studies, and 48/55 (87%) self-controlled case series fulfilled the major validity assumptions for use of the design; 4/35 and 14/48 respectively did not fulfill the minor assumptions. Overall, 31/53 (58%) case-crossover studies and 34/55 (62%) self-controlled case series fulfilled both major and minor assumptions. The reporting of the methodology or the results was appropriate, except for power calculation.

Conclusions

Self-controlled designs were not appropriately used in34% and 13% of the articles we reviewed that described a case-crossover or a self-controlled case series design, respectively. We encourage better use of these designs in situations in which major validity assumptions are fulfilled (i.e., for which they are recommended), accounting for situations for which the design can be adapted.

Electronic supplementary material

The online version of this article (doi:10.1186/s12874-016-0278-0) contains supplementary material, which is available to authorized users.