iHEA Boston 2017 Congress, Boston Massachusetts, USA 8-11 July 2017 A Review of the Quality of Tuberculosis Costing Studies Session: Introducing a Reference Case for Costing Global Health Interventions Lucy Cunnama Health Economics Unit, University of Cape Town, RSA (on behalf of the GHCC team)

What is the aim of this talk? To present a review of TB provider studies that have been assessed using a quality framework looking at the reporting standards and variation in methods used by authors

Using cost data produced by others

Did any of the below limitations adversely impact your efforts to extrapolate or adjust the cost data?

Electronic databases searched Pubmed, EMBASE, Medline, Econlit The National Health Service Economic Evaluation Database The Cost-effectiveness analysis Registry Electronic databases The European Association for Grey Literature Exploitation (EAGLE) The System for Information on Grey Literature in Europe (SIGLE) Documents and meeting reports from the World Bank and WHO websites Literatura Latino Americana em Ciencias da Saude or Latin American and Caribbean Health Sciences Literature (LILACS) Others Grey literature SIGLE is a bibliographic database

Search strategy Category Key terms 1. Cost Cost* or economic or finance AND 2. Tuberculosis TB or tuberculosis or MDR#TB or XDR#TB or multi?drug or “resistant tuberculosis” or “strain resistance” or “mycobacterium tuberculosis” 3. Treatment treatment or management or drugs or medication or DOTS or “directly observed treatment” or “health system*” or “hospital care” or epidemiology or “government hospital setting” or “community based care” or “patient* perspective” or “isoniazid preventive therapy” or IPT or prevention Dates of studies from 1990-2016 (28th of July) No restrictions on country or language

Exclusion process Direct exclusion criteria Evaluate direct exclusion criteria Does not meet exclusion criteria: Considered for inclusion Criteria reported: INCLUDE Can easily infer (see inference rules): INCLUDE Take to senior reviewer Cannot infer: Agree enough information to infer: INCLUDE Not enough information: contact the author Author responds with necessary information: Author does not respond or does not have necessary information: EXCLUDE Meet exclusion criteria: EXCLUDE Direct exclusion criteria All modelled or secondary data sources. (No empirical data collected regarding prices and quantities)) High income countries (world bank)

Flowchart CEA Registry n=42 Cochrane library n=1 098 Pubmed n=5 967 EMBASE n=8 893 NHS EED n=279 Econlit n=69 Records identified through database searching n=16 348 Additional records identified through other sources n=4 945 Records after duplicates removed n=15 161 Records screened n=15 161 Used our search terms in the databases… Cleaned Looked to see if they were eligible Records Excluded n=14 457 Full-text articles assessed for eligibility n=704 Full-text articles excluded, n=497 Studies included n=207 No access, n=21 High income setting, n = 169 No (relevant) data reported, n = 272 Correspondence/editorial/ commentary/news/protocol, n = 35 Studies extracted (healthcare payer costs) n= 75

Overall description To date, we have extracted 75 studies reporting TB provider costs Multi-field data extraction tool for both methodology and costs of studies This data set is a ‘work in progress’ Published studies (peer reviewed) Any provider cost for any TB intervention Reporting standards for each item: whether explicitly mentioned, or inferred if there was enough information to make this assessment Snapshot Patient costs will be assessed later

Studies per year (for extracted data) 14 in 2015 13 in 2013 8 in 2016 & 2012 23 in 2015, 19 in 2013 and 9 in 2016 97 substudies

Reporting - Scope

Reporting – Sampling

Reporting - Methods

Reporting – Inclusion of Costs

Reporting – Valuation

Economies of Scale, Comparison by Subgroup

Methods - Type of Costing, Above Service Level Costing

Methods – Timing, Sensitivity Analysis

Methods – Discount Rate, Measurement of cost

Conclusions Cost data for TB from a providers perspective is available, however not all recent Reporting varies greatly – some aspects reported well others not reported at all – examples are overheads and allocation of costs which make it hard to extrapolate or adjust cost data Among those aspects reported, there is variation in methods – such as discount rate, and use of sensitivity analyses Encouraging researchers to be explicit and transparent in how they costed interventions, Reference Case will help with this Not yet sure of the representativeness of the interventions costed in the 207 studies The final (larger) data set will assist us in assessing the current quality as well as establishing future guidance on costing Time lag with papers – 2016/15/14 papers What is getting neglected? Making sure high burden countries are represented Both top- down bottom- up separately Being explicit about

Future work Building a TB provider cost repository Formulating guidelines for TB costing Use of those guidelines By GHCC members By others in the field