Is There Any Evidence To Show That For Adults With Treatment-Resistant OCD, Antipsychotic Augmentation Is More Effective Than Placebo Augmentation In Reducing The Symptoms Of OCD?
This paper documents literature conducted in the identification of the current research in regard to the efficacy of antipsychotic augmentation towards the management of treating resistant obsessive-compulsive disorder (OCD) in adults. The effectiveness, in this case, is determined by changes in YALE-BROWN obsessive compulsive scale n(Y-BOCS) ranging from the baseline to post-intervention. Based on existing evidence, the Y-BOCS scale has recorded validity towards the determination of obsessive-compulsive symptoms as well as in the indication of ‘wellness’ among individuals with OCD.
With the free school education in place in kenya, it emerged that a high number of Turkana county children did not go to school for lack of food. A group of students from this region resolved to support a food program through a Charity Shop that would offer free essay resources to university students.
Touched with this story, support to show that poor Turkana child that we care for their education and a better future.
Literature Search Strategy
A literature search was undertaken by identifying numerous keywords obtained from abstract databases within studies published between the years 2012 to 2017. The following databases were searched; PsycINFO, Cochrane, Medline (Ovid) and, CINAHL complete. In order to refine the search towards the most recent articles, research was only done for the past five years and for research findings which have been published in the last NICE review (NICE, 2013). The exclusion criterion was based on the relevancy of the title of the research article. Similarly, an examination of the abstract of each paper was undertaken to determine any potential studies within the research topic. In order to ascertain the relevancy of the paper, a preliminary exclusion criterion was undertaken. In this case, six studies were considered relevant after careful evaluation. However, one article was disregarded on the basis that same author was involved in the publishing of the content and shared a similar research question based on a meta-analysis. In this case, Dold, Aigner, Lanzenberger, and Hasper, (2013) was considered instead of Dold, Aigner, Lanzenberger, and Hasper, (2015). The table below documents the articles identified per search terms within their respective databases documenting the number of articles selected after the preliminary exclusion criteria.
|Antipsychotic and augmentation and obsessive compulsive disorder||Antipsychotic and augmentation and OCD||Antipsychotic and treatment resist and obsessive compulsive disorder||Antipsychotic and treatment and obsessive compulsive disorder|
You buy food to a poor school kid whenever you buy these crafts
Following the identification of the articles from different databases that contained the selected keywords, the authors undertook a preliminary exclusion criterion as follows:
|Preliminary Exclusion Criteria|
|Studies investigating OCD in adolescents or children|
|Studies which have no specifications on whether the participants were treatment resistant|
|Studies which do not include a control group receiving a placebo|
|Entries with duplicates|
|Studies not published in English|
|Papers with a low focus on the effectiveness of antipsychotic augmentation|
|Studies that look at OCD as comorbid or focus on other disorders|
Among the five selected articles that met the exclusion criteria, a critical appraisal skills program was used as a screening framework to determine research studies with the best evidence. All the three articles that were identified constituted randomized control trials. In this case, Simpson, Foa, Liebowitz, Huppert, Cahill, Maher, and Weaver, (2013) failed to meet the inclusion criteria due to an unstable dose in antipsychotic augmentation group. This is because the risperidone dose was elevated to 0.5mg based on a weekly dosage to titrate with 4mg at the end of the 8th week that marked the end of the study period. Moreover, despite taking a keen focus on the placebo pill condition, a comparison group which had undergone a cognitive behavior therapy (CBT) was included. Therefore, this aspect failed to meet the criteria since it constituted of other interests in the study apart from our research question. As such, since they conducted 17 sessions at two times on a weekly basis, it would not be representative of the cognitive behavior therapy as listed in the National Health Service. Based on this exclusion criteria, Wheaton, Rosenfield, Foa, and Simpson, (2015) was excluded since it employed data from Simpson et al. (2013) However, it was used in analyzing the moderators as well as the predictors of positive treatment outcomes. Simpson et al. (2013) did not employ pure randomization but rather computer-generated randomization in order to balance different parts within the study. The results from Simpsons et al. (2013) indicate that 86% of those who participated in the study were retained since there was no significant difference that existed in the attrition groups. Further, the study also measured medication adherence with blood tests in order to ensure that the participants were concordant and that a sufficient dose had been metabolized. Based on the evaluation of Sayyah, Boostari, Ghaffari, and Hoseini (2012) on the final RCT, the augmentation of aripiprazole passed the validity screening. To evaluate the results accordingly, the population criteria used in the treatment of resistant OCD. In this case, each population was categorized as one where at least a single trial of selective serotonin
You buy food to a poor school kid whenever you buy these crafts
reuptake inhibitor (SSRI) in its maximum tolerable dose is taken for a period of three months. On the contrary, though, this does not meet the NICE criteria employed in the treatment of resistant OCD (NICE, 2006). In this case, it was considered for 12 weeks full trial insufficient response to SSRI and cognitive behavior therapy (CBT) which included the prevention of exposure-response above 10 therapy hours. In this case, full trial clomipramine monotherapy and SSRI was used. Most notably, sufficient response towards a 12-week trial that was previously carried on SSRI was collected for sampling purposes as they observed and documented treatment resistance. However, these studies did not document the methodology used in the random assignment of participants, rather, the allocation and double-blind. Attrition levels of 19.05% for the placebo and 16.67% for aripiprazole were recorded; this was within acceptable levels of less than 20%. Wegmann (2012) however, notes that any levels of attrition can be problematic when sample sizes are small. For this case, however, the levels of attrition did not differ significantly among different groups.
Through the literature undertaken, two meta-analysis papers were identified where both met the CASP screening criteria. As such, Dold et al. (2015) and Veale, Miles, Smallcombe, Ghezai, Goldacre, and Hodsoll, (2014) examined RCTs through making comparisons of antipsychotic augmentation with placebo by employing Y-BOCS score towards the measurement of the effectiveness of interventions. A YBOCS score exceeding 16 and a maximum tolerable dose of SSRI exceeding 8 weeks must be met by sampled individuals (CASP, 2017). Dold et al. (2015) evaluated studies that had an insufficient response to SSRI on the population with little restrictions on the length of the trial. Rather, a subgroup analysis was used to determine the impact of the length of the SSRI monotherapy trial and the meta-regression analysis in accordance with the Y-BOCS baseline score.
Following the evaluation of the three studies, a CASP criterion was undertaken using detailed questioning. Sayyah et al. (2012) met the criteria for RCT toolkit for blinding to condition. In this case, the blind to condition was achieved as evidenced by medical doctors and patients in the study in meeting inter-rater reliability for Y-BOCS ratings. Despite the fact that the research paper fails to state that experimenters were unaware, the authors note that the list was secured until the end of the trial when it was availed to the experimenters. Sayyah et al. (2012) do not find any significant differences between the study conditions in terms of age, gender, baseline Y-BOCS, and marital status. The study, however, failed to account for the differences in the study groups as it appertains the level of education, comorbid disorders, the age of onset, and the duration of illness. Benatti, Camuri, Dell’Osso, Cremaschi, Sembira, Palazzo, and Carpiniello, (2016) finds that worth noting is that these factors may have an impact on prognosis. In as much as the authors reported on the number of participants taking SSRI doses, they failed to include a significance analysis. Rather, the dosage was reported as a range among the participants. On the other hand, the authors failed to undertake a covariance analysis of the confounding variables. In essence, this may have failed to account on the extent to which these groups were controlled. On the better side, the procedure adopted was suitable for this kind of study since equal treatment was undertaken on both groups. As such, the procedure and experimental conditions were similar while identical pills of aripiprazole and placebo were used.
For a study to be reliable, a wide scope of literature review must be undertaken. In this case, the systematic reviews undertaken delved on a wide scope of literature by incorporating multiple databases that were published and unpublished in clinical trial registries and international studies (FaraoRe, 2008). The two studies contacted pharmaceutical companies and employed similar strategies on obtaining evidence as per the NICE guidelines. The two studies used unpublished studies in their meta-analysis, for that reason, it was hard to account for the impact of the unpublished data that was used in the study. In essence, this may have a significant effect on the experimental hypothesis Joober, Schmidt, Annable, and Eodska, (2012) note. 13 of the studies passed the exclusion criteria among 14 studies that met the meta-analysis and inclusion criteria. In this case, McDougle, Goodnran, Leckman, Lee, Heninger, and Price, (1994) was excluded on the basis that it used atypical antipsychotics as a result of higher rates due to adverse effects resulting from older antipsychotics. The RCT that was conducted by Sayyah et al. (2012) met the inclusion criteria in both studies; however, Dold et al. (1994) excluded Diniz et al. (2011) as a result of increased SSRI dose within the experimental group.
You buy food to a poor school kid whenever you buy these crafts
The grades of recommendation assessment, development, and evaluation (GRADE) system are highly effective in evaluating the quality of medical studies. For that reason, Veale et al. (2014) used the GRADE system in order to assess the quality of the study as well as the risk of bias in every study. The results obtained were very low in quality ratings emerging at negative two (-2) and consistency at zero. The small sample size was identified as a major factor towards the low-quality ratings under the GRADE system. Further, Dold et al. (2013) used Cochrane Collaboration risk of bias tool to access the quality of every study. Following a successful evaluation of the research articles, eight of the fourteen studies evaluated employed the right randomization process while only three of the fourteen studies had adequate concealment of allocation. The attrition levels for the studies indicated that four studies were high, six studies were low, and four studies were moderate. Sources of bias were identified where one study used six weeks as a minimum trial period for SSRI treatment at adequate before randomization. Further, a gender imbalance was also evident. The conclusions of the methodological quality of all the studies were found to be high. Veale et al. (2014) and Dold et al. (2013) used funnel plots in order to evaluate the risks posed by publication bias. In their findings, Veale et al. (2014) found that asymmetries were present, however, they advised of caution in interpreting them since the number of articles used was few. Based on this reasoning, Sterne, Sutton, Ioannidis, Terrin, Jones, Lau, and Tetzlaff, (2011) have found asymmetries to result from heterogeneity among studies which may report chance and/or bias. After undertaking a visual examination of their plots, outliers were spotted outside the 95% confidence limits. Despite the fact that Dold et al. (2013) failed to report on the funnel plot, they undertook Egger regression intercept test and found no publication bias. As a matter of fact, they found that 160 negative trials that are not published need to be present to disapprove the statistical significance recorded in the primary outcome which should include a mean change in the Y-BOCS score among the groups. Sterne et al. (2011), however, finds that high sample size and a high number of studies are imperative towards achieving a desirable conclusion. For that reason, a lack of homogeneity in the data demonstrated methodological weaknesses. Therefore, caution should be taken on the data in drawing conclusions based on funnel plot results. There was a disagreement in the overall quality of the studies as indicated in the reviews. However, this can be assumed to have lesser impact provided that the methodologies used in rating the two articles are different, hence, the results may vary considerably.
For that matter, Kavanagh (2009) has criticized the GRADE system for lack of reliability based on evidence. Based on the evaluation of Dold et al. (2015), it is evident that the authors have received grants and have worked for pharmaceutical companies. This indicates that the author’s interests may be biased when we compare this fact to Veale, (2009) who has little favor towards pharmacological treatment strategies. For Veale et al. (2014) and Wheaton et al. (2015), both have been advocating on the use of CBT as a treatment choice, both have published books on CBT strategies for anxiety disorders and are CBT therapy consultants in many specialist anxiety hospitals. The evaluation of the two studies indicates that both performed a meta-analysis. Both studies employed Y-BOCS score as outcome measures. In the analysis, it is evident that some data may have been missing; however, this was compensated from the analysis undertaken. Differences exist in how the studies have been compared. As such, Dold et al. (2015) note that differences exist in the definition of treatment-resistance, length of trials, antipsychotic dosage, clinical settings and differences in the inclusion criteria for participants with comorbid. Based on the criteria for NICE guidelines, Veale et al. (2014) note that it is essential for the criteria employed in treatment resistance to incorporate at least one CBT trial that has failed. However, these differences have been compensated in the analysis of covariance.
Sayyah et al. (2012) have achieved their primary outcome based on the results at the end of the study. In their findings, they note that there exists a significant difference that favors experimental group, on the contrary though, they don’t report on the effect size. They find that there is a reduction of 29.5% in the mean Y-BOCS score in the aripiprazole group. This results in a mean score of 15.42. Based on FaraoRe (2008) assumptions, a reduction in the mean exceeding 35% is considered as an indication of treatment response, further, Y-BOCS score which is less than or equal to 14 is considered as the cutoff for wellness in OCD. Apart from the primary outcomes, secondary outcomes indicated that there was no significant difference between the two groups which was inclusive of side effects. As a result, the authors conclude that aripiprazole augmentation reduces OCD symptomology without leading to a significant increase in side effects. Sayyah et al. (2012) have used the Mann-Whitney U test in investigating their hypothesis with a 95% confidence interval and established a 0.05% significance level in probability. These standards are considered for the determination of statistical significance, Flechner, and Tseng (2011) note. Based on the 0.05 significance level, it implies that there is a probability of 5% of the results being due to chance; therefore, it justifies the rejection of the null hypothesis (Flechner & Tsengi, 2011). The P value for the primary outcome is less than 0.0001; this confidence level indicates that there is a clinical significance. As such, Flechner and Tseng (2011) note that confidence levels should not have a wide range of overlap since this implies that results cannot be relied upon. Despite the fact that the authors failed to report on the confidence intervals, they were calculated at 95% based on the sample size. The endpoint means and standard deviations were recorded as 20.72-25.52 for placebo and 12.82-18.02 for aripiprazole. Further, clinical significance is supported by 99% confidence interval without overlap at (19.92-26.32) for placebo and (12.02-18.82) for aripiprazole.
You buy food to a poor school kid whenever you buy these crafts
Based on the results stated herein, the meta-analysis indicates that there is a high efficiency of antipsychotic augmentation. Veale et al. (2014) have used weighted mean differences called ‘metan’ in analyzing data. The authors use a weighted mean difference by collecting data from all antipsychotics against placebo and then individually from every antipsychotic. Following the evaluation, a mean difference in the change of Y-BOCS score was reported at 2.l34 points between antipsychotic and placebo. The overall effect size is d=0.4 which they state that is an equivalent to a decrease in 10% in the Y-BOCS score. Further, a mean Y-BOC reduction of 3.1 points for aripiprazole where (d=1.11) and 3.89 points for risperidone where (d = 0.53). There were no differences obtained between the experimental and placebo group for quetiapine and olanzapine. Categorical analysis was carried on responders versus non-responders indicating that participants who received risperidone were 3.1 times likely to respond to treatment in accordance to Y-BOCS reduction of 25% for two studies and 35% for three studies. According to Citrome and Ketter (2013), an acceptable figure of 4.65 was attained. However, FaraoRe et al. (2008) note that there must be a lot of caution in adopting these values since two of the five studies have a response to treatment of 35% which is lower than the minimum recommended. On the other hand, Dold et al. (2015) have used Hedges’ to determine the effect sizes of change in Mantel-Haenszel risk ratios and Y-BOCS score for categorical data. Y-BOCS score is greatly reduced among the antipsychotic groups compared with the placebo (Hedges ‘ g =-64.95% CI:-0.87-0.41; N=14, n=478; p=<0.1). There were significant differences in the antipsychotic group for haloperidol (g=-0.82), risperidone (g=-0.59) and aripiprazole (g =1.35), however, there no significant differences for quetiapine, olanzapine, and paliperidone. Veale et al. (2014) included only one study of paliperidone with risperidone. Storch, Goddard, Grant, Goodman, Mutch, and Murphy, (2013) note that this is essential since paliperidone is the active metabolite of risperidone. A significant reduction in Y-BOCS score starting from the baseline to the endpoint was recorded in an individual test. However, the differences were not significant in group analysis. Fletcher and Tseng, however, note that in such scenarios where individual differences are noted, caution should be taken in generalizing the results. Most importantly, some tests of heterogeneity indicated a rejection of the null hypothesis. However, this is acceptable since not all data can be assumed to originate from homogenous samples. Sedgwick (2012) notes that homogeneity is an essential indicator in the meta-analysis as it determines the permissible variation in a sample population. I square and chi-square were both used in the determination of homogeneity. Dold et al. (2015) state that no heterogeneity was found. This, however, contrasts Veale et al. (2014) for failing to reject the null hypothesis for the subanalysis of quetiapine and olanzapine and antipsychotic meta-analysis. The authors, however, find that failing to reject the null hypothesis is unreliable for the sub-analysis as they involve a small number of studies (quetiapine, n=5: aripiprazole, n=2: risperidone, n=5: olanzapine n=2). For that reason, homogeneity was indicated for aripiprazole and risperidone.
The results obtained from this study are to be generalized for adults as opposed to children and adolescents. This is because both studies have used an adult population with Dold et al. (2015) using a mean age of 37.15 years with a standard deviation of 3.03 and Sayyah et al. (2012) using a mean age of 39 years for the aripiprazole group with a standard deviation of 3.17 and 37 years for the placebo group with a standard deviation of 3.53. Despite the fact that Dold et al. (1994) failed to report on age, one of the studies had a mean age of 35.76 years with a standard deviation of 10.55. Indeed, based on the ages presented, children and adolescents may be exempt from generalizations since age may affect pharmacology and side effects accordingly. On the other hand, it would be essential to undertake a similar study on all age brackets to come up with validly acceptable results. However, among all the researchers, none used the NICE criteria for treatment resistance. Further, as Wegmann (2012) notes, when research involves subjects with age differences, therefore, the results cannot be generalized to children since age affects medication metabolism. Among the studies evaluated, 2 failed to establish a significant effect of antipsychotic (Diniz et al., 2010 and Shapira, Ward, Mandoki, Murphy, Yang, Blier, and Goodman, 2004) in an 8 week SSRI and a randomized non-responders and partial-responders into placebo or olanzapine groups. These studies used fluoxetine as the SSRI with a dose of 40mg as opposed to 60mg being the maximum dosage under BNF standards (BNF, 2017). On the other hand, Dinits et al. (2009) have used a Y-BOCS score >14 to categorize non-response to treatment in the pre-randomization stage instead of reducing from the baseline. Dinitz et al. (2009) have however failed to find a significant effect of the antipsychotic in comparison to the control group in Y-BOCS score from the start to the end of the randomization period. Veale et al. (2014) reported on findings that participants underwent CBT with 3 studies finding the rate at 30% or below, 2 studies reported a 100% of the participants had undergone at least a single trial of CBT and 5 studies failing to report on anything. For that reason, based on the varying levels in the delivery of CBT, it’s hard to know whether the therapy would compare to that of NHS.
Based on the evaluation of secondary data, some information for consideration emerged to be lacking. In this case, there was no detailed reporting on the adverse effects of the drugs, there was no follow-up on the participants to determine long term side effects or benefits and qualitative data regarding patients was not availed. The authors plainly reported on Y-BOCS without undertaking a qualitative evaluation of the participants. Iczi et al. (2016) described symptomology in five case studies prior to the augmentation of clomipramine with aripiprazole. On the contrary, though, there were no details given on acceptability and tolerability of augmentation treatment, rather, the authors reported an anxiety scale post-treatment ranging from 4-6 weeks and Y-BOCS. This study, however, failed to meet the exclusion criteria since it lacked a control group. The level of subject attrition is a good indicator of the acceptability of treatment. Dold et al. (2015) did not find any significant differences in attrition in the placebo and antipsychotic groups. This can be significantly high side effects in the combined group. Dold et al. (2015), however, did not report on differences on psychotics regarding breakdown. Such loopholes make it hard to determine the tolerability of different antipsychotics.
You buy food to a poor school kid whenever you buy these crafts
A high dropout rate resulting from side effects was reported for all the studies. For that reason, it may be attributed to time constraints which may have led to faster titration than would be in natural settings. In respect to this fact, Parrott, Morlnan, Moss, and Scholey, (2004) note that it is recommended to start with a low dose elevating the levels as patients adapt to antipsychotics. In effect, this minimizes the side effects. Dold et al. (1994) did not find the correlation between dose and the effect size. As such, they found that a low risperidone dose resulted in greater effect size compared to moderate dosage. This finding could not be replicated for quetiapine though.
The length of the trial in clinical tests is a very important parameter. Agid, Kapur, Arenovich, and Zlpursky (2003) note that antipsychotics can have a therapeutic effect as early as the first week, however, this has always contrasted with delayed therapeutic onset theory. Veale et al. (2014) however did not find any relation between the length of the trial and effect size. However, analyzing the SSRI monotherapy taking less than 8 weeks, there was a greater reduction in Y-BOCS score resulting from antipsychotic augmentation compared to the sample exceeding 8 weeks. In conclusion, this was attributed to continued SSRI improvement both in the sample and the antipsychotic group.
A cost-benefit analysis is essential in determining the devastating nature of OCD with respect to the cost of treatment in non-responding cases and its side effects. As Veale et al. (2014) note, there is no clearly identifiable predictor of reliable response in antipsychotics. On the other hand, though, Dold et al. (2015) have found more effective antipsychotics for compulsive or obsessive symptomology. However, Veale et al. (2014) note that caution should be taken prior to the prescription of antipsychotics to OCD and rather a small atypical antipsychotic dose should be administered and monitored. Overall, it is worth noting that antipsychotics have not been licensed to treat OCD and the use of antipsychotics can induce psychosis.
Based on the facts reported towards Y-BOCS score, it is evident that antipsychotic is effective in easing OCD symptoms in adults with treatment-resistant OCD. A meta-analysis indicates that moderate to large effect size is attainable for aripiprazole, haloperidol, and risperidone, Moreover, the moderate effect size is attainable for combined antipsychotic. Overall, Veale et al. (2014) note that lower dosages of risperidone may be effective than moderate doses, despite this, more research should be done to determine optimal doses. It is of great importance that maximum caution is taken before conclusions can be drawn from this research. As Veale et al. (2014) and Dold et al. (2013) note, the quality of current evidence can be challenged. Further, none of the research undertaken meets NICE standards and there is a lack of a universal description of treatment resistance (Joober, Schmitz, Annable, & Boksa, 2012). On the other hand, the failure to undertake a follow up on the participants is a great shortcoming since long term effects cannot be determined. For that reason, the minimum duration for effective treatment can also not be determined for long term efficiency. It will, therefore, be essential that such shortcomings are compensated in further research studies to make the findings of this research practical.
Agid, O., Kapur, S., Arenovich, T., & Zipursky, R. B. (2003). Delayed-onset hypothesis of antipsychotic action: a hypothesis tested and rejected. Archives of general psychiatry, 60(12), 1228-1235.
Benatti, B., Camuri, G., Dell’Osso, B., Cremaschi, L., Sembira, E., Palazzo, C & Carpiniello, B. (2016). Which factors influence onset and latency to treatment in generalized anxiety disorder, panic disorder, and obsessive–compulsive disorder? International clinical psychopharmacology, 31(6), 347-352.
Benatti. B, Camuri. G, Dell’Oss. B, Cremasehi. L, Sembira. E, Palaazo. C and Altarnur. A.C. (2016). NICE Clinical Knowledge Summaries (CKS).Retrieved from https://www.nice.org.uk/!scenariorecommendation:1CriticalAppraisalSkillsPrsgramne%7B?017).f
BNF. (2017). Nice British National Formulary (BNF). Retrieved from https://www.nice.org.uk/bnf-bnfc-uk-only
CASP. (2017). Critical Appraisal Skills Program. Retrieved from https://casp-uk.net/
Citrome, L., & Ketter, T. A. (2013). When does a difference make a difference? Interpretation of number needed to treat, number needed to harm, and likelihood to be helped or harmed. International journal of clinical practice, 67(5), 407-411.
Diniz, J, Shavitt, R.G, Fereira. C.A.B, Houniel, A.., PimeRtel, 1, LM KorEn, M and Miguel E. C. (2009). Antipsychotic augmentation of serotonin reuptake inhibitors in treatment-resistant obsessive-compulsive disorder: An update meta-analysis of double-blind, randomized, placebo-controlled trials. Retrieved from https://journals.sagepub.com/doi/10.1177/0269881108099423
Dold, M, Aigner M, Lanzenberger R and Kasper S. (2013). Antipsychotic augmentation of serotonin reuptake inhibitors in treatment-resistant obsessive-compulsive disorder: an update meta-analysis of double. PubMed-NCBI. Retrieved from: https://www.ncbi.nlm.nih.gov/pubmed/25939614
Dold, M., AiBner, M., Lanzenberger, R., & KasBer S. (2015). Antipsychotic augmentation of serotonin reuptake inhibitors in treatment-resistant obsessive-compulsive disorder: An update meta-analysis of double-blind, randomized, placebo-controlled trials. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4576518/
FaraoRe, S. V. N. (2008). Meeting abstracts – Academy of managed care pharmacy nexus 2017. – PubMed – NCBI. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/28967301
Flechner, L, & Tseng, T. Y. (2011). Understanding results: P-values, confidence intervals, and number need to treat. Indian Journal of Urology doi: 10.4103/0970-1591.91447
Joober, R., Schmitz, N., Annable, L., & Boksa, P. (2012). Publication bias. What are the challenges and can they be overcome? Journal of Psychiatry & neuroscience, 37, 149-152.
Kavanagh, B. P. (2009). The GRADE system for rating clinical guidelines. PLoS medicine, 6(9), e1000094.
McDougle, C.J., Goodnran, W.K., Leckman, J.F., Lee, N.e., Heninger, G.R., & Price, L.H. (1994). Haloperidol addition in fluvoxamine-refractory obsessive-compulsive disorder. A double-blind, placebo-controlled study in patients with and without tics. Archives General Psychiatry, 551, 382-308.
NICE (2006). Obsessive-compulsive disorder: Core interventions in treatment.
NICE (2013). Obsessive-composure disorder: Evidence Update September 2013.
Parrott, J., Morinan, A., Mols, M,, & Scholey, A. (2004). Understanding drugs and behavior. West Sussex: Wiley.
Sayyah, M., Boostani, H., Ghaffari, S. M., & Hoseini, A. (2012). Effects of aripiprazole augmentation in treatment‐resistant obsessive‐compulsive disorder (a double blind clinical trial). Depression and anxiety, 29(10), 850-854.
Sedgwick, P. (2012). Meta-analyses: tests of heterogeneity. BMJ: British Medical Journal (Online), 344.
Shapira, N. A., Ward, H. E., Mandoki, M., Murphy, T. K., Yang, M. C., Blier, P., & Goodman, W. K. (2004). A double-blind, placebo-controlled trial of olanzapine addition in fluoxetine-refractory obsessive-compulsive disorder. Biological psychiatry, 55(5), 553-555.
Simpson, H. B., Foa, E. B., Liebowitz, M. R., Huppert, J. D., Cahill, S., Maher, M. J., … & Weaver, J. (2013). Cognitive-behavioral therapy vs risperidone for augmenting serotonin reuptake inhibitors in obsessive-compulsive disorder: a randomized clinical trial. JAMA psychiatry, 70(11), 1190-1199.
Sterne, J. A., Sutton, A. J., Ioannidis, J. P., Terrin, N., Jones, D. R., Lau, J. & Tetzlaff, J. (2011). Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. Bmj, 343, d4002.
Storch, E. A., Goddard, A. W., Grant, J. E., De, A. N., Goodman, W. K., Mutch, P. J., … & Murphy, T. K. (2013). Double-blind, placebo-controlled, pilot trial of paliperidone augmentation in serotonin reuptake inhibitor-resistant obsessive-compulsive disorder. The Journal of clinical psychiatry, 74(6), e527-32.
Veale, D., & Willson, R. (2009). Overcoming Obsessive Compulsive Disorder: A self-help guide using cognitive behavioural techniques. Hachette UK.
Veale, D., Miles, S., Smallcombe, N., Ghezai, H., Goldacre, B., & Hodsoll, J. (2014). Atypical antipsychotic augmentation in SSRI treatment refractory obsessive-compulsive disorder: a systematic review and meta-analysis. BMC psychiatry, 14(1), 317.
Wegmann, J. (2012). Psychotropic Medication Use in Pregnancy, Older Adults, and the Young. Wisconsin, USA: CMI Education institute.
Wheaton, M. G., Rosenfield, D., Foa, E. B., & Simpson, H. B. (2015). Augmenting serotonin reuptake inhibitors in obsessive–compulsive disorder: What moderates improvement? Journal of consulting and clinical psychology, 83(5), 926.