Updating Surveillance System: Assessing the Need for Updating Comparative Effectiveness Reviews The University of Ottawa Evidence–based Practice Center The RAND Southern California Evidence-based Practice Center Alexander Tsertsvadze MD, MSc Date: September 11, 2012

Outline Background Why update systematic reviews (SRs)? What is an update of a SR? When/how to update a SR? AHRQ updating surveillance project Methods Overview of the process The Ottawa/RAND EPC methodology Findings and implications Brief overview

Background Why update SRs? SRs play a central role in comparative effectiveness research With the emergence of new evidence over time, conclusion(s) of any given SR may become out of date and sometimes misleading Changes in the evidence can have significant implications for clinical practice guidelines and for clinical and consumer decision-making

Background What is an update of SR? “A discrete event with the aim to search for and identify new evidence to incorporate into a previously completed systematic review” An update must involve a search for new studies Moher and Tsertsvadze. Lancet 2006; 367:881-3 Cochrane Hand Book section 3.2.2

Background Shojania 2007 Sets of quantitative/qualitative signals indicating need for updating tested on 100 SRs; within 2 yrs, 23% of SRs had signals indicating the need for updating; the median time to the emergence of signal indicating the need for updating was 5.5 years French 2005 Of updated SRs (n=254), 9% (n=23) had changes in conclusions within 4 yrs Garritty /103 of respondents estimated >50% of their respective SRs were likely out-of-date

Background One SR found very few methods and strategies used for updating SRs The authors identified the need for standardized reliable/efficient method(s) for updating ( Moher and Tsertsvadze 2007; Tsertsvadze 2011)

Background In , Ottawa and RAND Evidence-based Practice Centers (EPCs) within the Agency for Healthcare Research and Quality (AHRQ) Effective Health Care Program developed a methodology for determining the need of updating Comparative Effectiveness Reviews (CERs) Abbreviated searches for new evidence Identification of signals indicating the need of updating by applying a priori set criteria to the new evidence Expert opinion on validity of CER conclusions Topic-specific safety alerts from FDA and Health Canada Rules/guidance for determining the updating status of conclusion(s) for each key question (KQ) within CER and updating priority for each CER (low, medium, high) (Shojania 2007; Shekelle 2009)

Background Objectives Recognizing the importance of keeping CERs up-to-date, the AHRQ commissioned the RAND and Ottawa EPCs in collaboration with ECRI Institute to implement an updating surveillance system project This collaborative project would use the previously developed methodology to assess already published CERs (n=42) for the need of updating ___________________________________________________ - CERs unlike SRs contain several key questions/conclusions - Updating vs. assessing the need for updating

Original CER Abbreviated searches for new evidence using the same strategy as in the CER Screening and extracting data from relevant studies Identifying qualitative/quantitative signals Signal detection completed FDA/Health Canada alerts Expert opinion Deciding on updating status of conclusion(s) for each KQ within CER and determining updating priority for each CER Low priorityHigh priorityMedium priority Refereed for updating Goes to 2 nd cycle of assessment 6 months later Contacting expertsECRI surveillance Start of review End of review 0 3 rd month Updating surveillance process

Methods Abbreviated search for new evidence Search strategies – the same as used in the original CER but limited to 5 general medicine journals (Ann Intern Med, BMJ, JAMA, Lancet and NEJM) 5 specialty journals (specific to a topic of CER; most frequently cited journals in CER)

Original CER Abbreviated searches for new evidence using the same strategy as in the CER Screening and extracting data from relevant studies Identifying qualitative/quantitative signals Signal detection completed FDA/Health Canada alerts Expert opinion Deciding on updating status of conclusion(s) for each KQ within CER and determining updating priority for each CER Low priorityHigh priorityMedium priority Refereed for updating Goes to 2 nd cycle of assessment 6 months later Contacting expertsECRI surveillance

Methods Contacting content experts Conclusions from CER (executive summary) Is the conclusion(s) in this CER still valid? (Yes/No/Don’t know) Are you aware of any new evidence that is sufficient to invalidate the finding(s) in CER? (Yes/No/Don’t know) If yes, please provide references Comments Key Question # 1: Conclusion for key question # 1: Key question # 2: Conclusion for key question # 2: Key question # 3: Conclusion for key question # 3: Key Question # 4: Conclusion for key question # 4: Abbreviations: Expert Opinion CER-specific clinical experts (content experts, technical expert panel members, expert peer reviewer) Other clinical experts (external, local) Feeds into the decision on updating status of KQ specific conclusion(s) in each CER

Original CER Abbreviated searches for new evidence using the same strategy as in the CER Screening and extracting data from relevant studies Identifying qualitative/quantitative signals Signal detection completed FDA/Health Canada alerts Expert opinion Deciding on updating status of conclusion(s) for each KQ within CER and determining updating priority for each CER Low priorityHigh priorityMedium priority Refereed for updating Goes to 2 nd cycle of assessment 6 months later Contacting expertsECRI surveillance

Methods Safety surveillance alerts FDA/Health Canada alerts on CER topics (monthly) On CERs assigned to RAND EPC On CERs assigned to Ottawa EPC Feeds into the decision on updating status of KQ specific conclusion(s) in each CER

Original CER Abbreviated searches for new evidence using the same strategy as in the CER Screening and extracting data from relevant studies Identifying qualitative/quantitative signals Signal detection completed FDA/Health Canada alerts Expert opinion Deciding on updating status of conclusion(s) for each KQ within CER and determining updating priority for each CER Low priorityHigh priorityMedium priority Refereed for updating Goes to 2 nd cycle of assessment 6 months later Contacting expertsECRI surveillance

Methods Identifying signals for updating Qualitative signals Potentially invalidating change in evidence [pivotal trial/MA] - Opposing findings in effectiveness - Substantial harm - A superior new treatment Major change in evidence [pivotal trial/MA] - Important changes in effectiveness short of opposing findings - Clinically important expansion of treatment - Clinically important caveat [non-pivotal trial/MA] - Opposing findings in effectiveness/harms

Methods Identifying signals for updating Quantitative Change in statistical significance significant vs. non-significant non-significant vs. significant At least 50% change in effect size

Original CER Abbreviated searches for new evidence using the same strategy as in the CER Screening and extracting data from relevant studies Identifying qualitative/quantitative signals Signal detection completed FDA/Health Canada alerts Expert opinion Deciding on updating status of conclusion(s) for each KQ within CER and determining updating priority for each CER Low priorityHigh priorityMedium priority Refereed for updating Goes to 2 nd cycle of assessment 6 months later Contacting expertsECRI surveillance

Methods Rules for assessing updating status of conclusion(s) for each KQ Up to date No new evidence Only confirmatory evidence All responding experts: conclusions as still valid Some new evidence A minority of responding experts: having new evidence that might change the conclusion Possibly out of date Substantial new evidence Majority of responding experts: having new evidence that might change the conclusion Probably out of date New evidence that rendered the CER conclusion out of date or no longer applicable E.g. withdrawal of a drug or surgical device from the market, a black box warning from FDA, etc Out of date

Methods Rules for assigning the updating priority to a CER high, medium, or low How many conclusions? Probably out of date Possibly out of date Out of date How out of date are conclusions? Magnitude/ direction of changes in estimates Potential changes in practice or therapy Drugs withdrawn from the market /black box warning Availability of a new treatment

Methods A brief report summarizing the findings from the surveillance process – Title page lists the final priority for updating (“low,” “medium,” or “high”) – Description of methods and results – Summary table by conclusion: findings of the literature search findings from FDA and Health Canada expert opinion assessment of the degree to which the conclusion may be out-of-date – Evidence table

Findings 1 st year of surveillance 14 CERs were assessed for updating status and were assigned updating priorities as follows: 2 (14.2%) = ‘high priority’ 3 (21.4%) = ‘medium priority’ 9 (64.4%) = ‘low priority’

Implications We have established a Surveillance Program that has evaluated 14 CERs over the course of 12 months We have shown that a program for regular and active surveillance of CERs is feasible

Implications To our knowledge, the surveillance assessment project is the first large scale effort that has applied a set of methods to assess the currency of evidence-based reports (or systematic reviews) in a structured and standardized manner The application of this methodology proves to be a relatively feasible, efficient, and at the same time, comprehensive approach for assessing the need for updating individual CERs across a wide range of health interventions

References Moher D, Tsertsvadze A. Systematic Reviews: When is an Update an Update? Lancet 2006; 367: Garritty C, Tsertsvadze A, Tricco AC, Sampson M, Moher D. Updating systematic reviews: an international survey. PloS one 2010; 5(4):e9914. French SD, McDonald S, McKenzie JE, Green S. Investing in updating: how do conclusions change when Cochrane systematic reviews are updated? BMC Med Res Methodol 2005;5:33. Moher D, Tsertsvadze A, Tricco AC, Eccles M, Grimshaw J, Sampson M et al. A systematic review identified few methods and strategies describing when and how to update systematic reviews. J Clin Epidemiol 2007; 60(11): e1. Shojania KG, Sampson M, Ansari MT, Ji J, Doucette S, Moher D. How quickly do systematic reviews go out of date? A survival analysis. Ann Intern Med 2007;147:224e33. Shekelle P, Newberry S, Maglione M et al. Assessment of the need to update comparative effectiveness reviews: Report of an initial rapid program assessment ( ) [Internet] Sep 10. Shekelle PG, Newberry SJ, Wu H, Suttorp M, Motala A, Lim Y-W, Balk EM, Chung M, Yu WW, Lee J, Gaylor JM, Moher D, Ansari MT, Skidmore R, Garritty C. Identifying Signals for Updating Systematic Reviews: A Comparison of Two Methods. Rockville (MD): Agency for Healthcare Research and Quality. June Methods Research Report. AHRQ Publication No. 11- EHC042-EF. Tsertsvadze A, Maglione M, Chou R, Garritty C, Coleman C, Lux L et al. Updating comparative effectiveness reviews: current efforts in AHRQ's Effective Health Care Program. J Clin Epidemiol 2011; 64(11):

THANK YOU!