SURVEYDESIGNANDMETHODS In 2017, weightswerenotadjustedusingthenonresponseadjustmentprocessdescribedinpreviousyears’methods. Asin pastyears,Kaiser/HRETconductedasmallfollow-upsurveyofthosefirmswith3-49workersthatrefusedtoparticipateinthe full survey. Based on the results of a McNemar test, we were not able to verify that the results of the follow-up survey were comparabletotheresultsfromtheoriginalsurvey. In2010and2015,theresultsoftheMcNemartestwerealsosignificantand wedidnotconductanonresponseadjustment. Between2006and2012,onlylimitedinformationwascollectedonconventionalplans. Startingin2013,informationon conventionalplansiscollectedunderthePPOsectionandtherefore,thecoveredworkerweightisrepresentativeofallplan typesforwhichthesurveycollectsinformation. Thesurveycollectsinformationonphysicianofficevisitsforeachplantype. Differentplantypesatthesamefirmmayhave differentcost-sharingstructures(e.g.,copaymentsorcoinsurance). Becausethecompositevariables(usingdatafromacrossall plantypes)arereflectiveofonlythoseplanswiththatprovision,separateweightsfortherelevantvariableswerecreatedin ordertoaccountforthefactthatnotallcoveredworkershavesuchprovisions. Asdiscussedbelow,changesinthe2017survey havereducedthenumberofvariable-specificweightsused. Toaccountfordesigneffects,thestatisticalcomputingpackageRandthelibrarypackage“survey”wereusedtocalculate 10 11 standarderrors. , All statistical tests are performed at the .05 confidence level, unless otherwise noted. For figures with multipleyears,statistical tests are conducted for each year against the previous year shown, unless otherwise noted. No statistical tests are conducted for years prior to 1999. Statistical tests for a given subgroup (firms with 25-49 workers, for instance) are tested against all other firm sizes not included in that subgroup(all firmsizesNOTincludingfirmswith25-49workers,inthisexample). Testsaredonesimilarlyforregionand industry; for example, Northeast is comparedtoallfirmsNOTintheNortheast(anaggregateoffirmsintheMidwest,South, andWest). However,statisticaltestsforestimatescomparedacrossplantypes(forexample,averagepremiumsinPPOs)are testedagainstthe“AllPlans”estimate. Insomecases,wealsotestplan-specificestimatesagainstsimilarestimatesforother plantypes(forexample,singleandfamilypremiumsforHDHP/SOsagainstsingleandfamilypremiumsforHMO,PPO,andPOS plans); these are noted specifically in the text. The two types of statistical tests performed are the t-test and the Wald test. The smallnumberofobservationsforsomevariablesresultedinlargevariabilityaroundthepointestimates. Theseobservations sometimescarrylargeweights,primarilyforsmallfirms. Thereadershouldbecautionedthattheseinfluentialweightsmay result in large movementsinpointestimatesfromyeartoyear;however,thesemovementsareoftennotstatisticallysignificant. StandardErrorsformostkeystatisticsareavailableinatechnicalsupplementavailableatwww.kff.org/ehbs. 2017SURVEY In 2017, wecontinuedtomakerevisionstohowthesurveyasksemployersabouttheirprescriptiondrugcoverage. Inmost cases, information reportedinPrescriptionDrugBenefits(Section9)isnotcomparablewithpreviousyears’findings. Overtime, planshavedevelopedmorecomplexbenefitdesigns. Inordertobettercaptureinformationonspecialtydrugs,weelectedto askaboutthesedrugsseparatelyfromthecostsharingonothertiers. Wemodifiedthequestionaboutthenumberoftiersa firm’scost-sharingstructurehastoaskspecificallyabouttiersthatdonotexclusivelycoverspecialtydrugs. Averagecopayment andcoinsurancevaluesarestillreportedamongworkerswiththreeormoretiers,twotiers,orthesamecostsharingregardless of drugclass, but noneofthesetiersincludescostsharingfortiersthatexclusivelycoverspecialtydrugs. Forty-fivepercentof firmswithdrugcoveragecoverspecialtydrugsbutdonothaveatierthatonlycoversthisclassofdrugs. Inthesecases,cost sharingamongspecialtydrugsisstillcapturedwiththeplan’sotherdrugclasses. Figures9.1and9.2reportthedistributionofcost-sharingstructuresincludinganytiersforspecialtydrugs. Thisanalysisadds thenumberoftiersthefirmreportedbyanytierstheymayhaveforspecialtydrugs. Therefore,afirmwithtwotiersandatier exclusively for specialty drugs is considered a three tier plan in this analysis, but a two tier plan when reporting average cost sharingvalues. Evenifafirmhasmultiplespecialty-onlytiers,wecollectinformationononlyone. Similar to 2016, we no longer require that a firm’s cost-sharing tiers be sequential, meaning that the second tier copayment washigherthanthefirsttier,thethirdtierwashigherthanthesecond,andthefourthwashigherthanthethird. Asdrug 10 Analysisofthe2011surveydatausingbothRandSUDAAN(thestatisticalpackageusedpriorto2012)producedthesameestimatesandstandarderrors. 11 AsupplementwithstandarderrorsforselectestimatescanbefoundonlineatTechnicalSupplement: StandardErrorTablesforSelectedEstimates, http://ehbs.kff.org The Kaiser Family Foundation and Health Research & Educational Trust / Page 22