TEN KEY FACTORS...continued from page 11Understanding the key factors and how they relate tosample size may help establish clear sample goals, planmore efficient studies, and negotiate less costly CIAs.July 200112A large population will influence the 9. Sampling for rare events.sample size when it is comprised of heterogeneousgroups, such as a collection events, such as a rare error, may requireSampling for infrequently occurringof records from multiple departments large samples. The ultimate sizeacross multiple facilities involving depends upon rarity of the event anddozens of DRG codes. The larger diversitydrives up the sample size.determine whether errors exist,the goal of the sample. If the goal is toenough8. Sampling unit. The sampling unit records must be sampled so there is adefines what is contained in each re c o rd good chance of finding at least oneto be sampled. The sampling unit could error in the sample. This could requirebe a paid amount for an individual CPT a sample of several hundred to over onecode, or it could be a summary of all thousand depending upon the sampleCPT paid amounts for a part i c u l a r specifications and assumptions.patient. The sampling unit affects samplesizes when it influences the va r i a b i l i- If the goal is to estimate the error ratety of the data. Paid amounts by individualCPT codes will have less va r i a t i o n may be required in order to observefrom the sample, even more recordsthan paid amounts summarized by enough occurrences to produce a reasonablyprecise estimate. If the dollarpatient. T h e re f o re, a sample selected byCPT code could be smaller than a sampleselected by patient.the sample requirement may be largervalue of the errors is to be estimated,still.However, it is often logistically moreconvenient to pull and review an entire 10. Design information inadequacies.patient chart rather than select and conducta review by CPT code. Sometimes assumptions based on the population,Sample size calculations are made usingan issue, such as lab unbundling, is easierto address by patient rather than When there are too few probe items toprobes, and/or historical information.individual CPT codes.design a complex sample, or when thepopulation definition changes after theLogistical convenience and the review probe is completed, the design informationmay be inadequate for accurateprocess should be weighed against thesample size and cost. Some CIAs specify sample size calculations. When thisthe sampling unit. This is an important occurs, it is prudent to use conservativeconsideration before committing to an assumptions, such as a larger variance, ainconvenient or inefficient sampling more extreme error rate, or more stringentconfidence and precision unit.requirements.However, these solutions lead tolarger sample sizes than would be necessarywith more accurate design information.SummaryWhen preparing for a statistical sample,bear in mind the factors that may affectthe sample size and cost of the study.There are key planning decisions thataffect the ultimate sample sizes. Theseinclude the sampling unit and designspecifications for the confidence andprecision. Specifying absolute or relativeprecision will influence sample sizes aswill the number of subgroups requiringseparate estimates with confidence andprecision requirements.Know what is to be estimated from thesample because total dollars, ratios, anderror rates all have different sample sizerequirements. Also, different quantitieshave different variances and thereforedifferent sample size needs.The size of the population usually onlyplays a minor role in sample size determination,unless the population is verysmall. However, variation in the populationdata will heavily influence thesample size.Complex samples are generally smallerthan simple random samples, especiallywhen the population is diverse. However,complex samples require more statisticaland subject matter expertise todesign and produce properly weightedestimates.Sample size is determined by many factors.There is no “one size fits all” samplesize. ■
featurearticleMeet Steven W. OrtquistEditor’s note: This is an interview conductedby Odell Guyton, Corporate <strong>Compliance</strong>Officer, University of Pennsylvania, andmember of the HCCA Board with SteveOrtquist, Director of Corporate <strong>Compliance</strong>for Rush Presbyterian St. Luke’s MedicalCenter in Chicago. Odell may be reached at215/573-4806. Steve may be reached at312/942-8123.OG: Good morning, Steve.What is your position?<strong>SO</strong>:I am Director of Corporate<strong>Compliance</strong> at Rush Presbyterian St.Luke’s Medical Center in Chicago, IL.OG:<strong>SO</strong>:What is your background?I have worked in the healthcare industry for nearly 15 years. Forseveral years I worked in financial managementroles with a national physicianorganization. In the mid-1990’s, I leftthese roles to attend law school andthen practiced law with a boutiquehealth care firm in Michigan, where Iassisted several clients with reimbursementand fraud and abuse matters aswell as with establishing complianceprograms.In 1998, Rush offered me my currentrole as its Director of Corporate<strong>Compliance</strong>. I actually came to Rushspecifically to take the compliance role.I had not worked with Rush in anycapacity prior to doing this. However, Ihad worked with several clients on<strong>Compliance</strong> Officer for RushPresbyterian St. Luke’s Medical Centercompliance programs or compliancematters as an attorney. I enjoyed thiswork and thought, correctly, that an inhouseposition more focused on compliancewould be enjoyable as well.OG: And what is the major thrustof Rush’s business?<strong>SO</strong>:Rush is an academic medicalcenter on Chicago’s West Side. It is alsothe lead hospital in the Rush Systemfor <strong>Health</strong>, a health care system thatincludes four other Chicago area hospitals.Rush’s compliance program appliesonly to Rush-Presbyterian-St. Luke’sMedical Center itself. The other hospitalsin the Rush system are operatingtheir own compliance programs. Rush-Presbyterian also includes a 150 bedfreestanding skilled nursing facility, ahome health agency, about 350employed physicians, and RushUniversity.Rush also is a significant researchinstitution. In terms of fundingreceived, in 1999 it ranked near the topof the second quartile among all recipientsof Federal research funding forhealth care related research, and is typicallyamong the top ten hospital recipientsof Federal research funds.OG:And how large of a staff doyou have in your compliance department?<strong>SO</strong>:Currently, the staffing inRush’s compliance office equals about12 FTEs [full time employees]. Thecompliance office is actually structuredin an interesting way that has workedvery well for Rush.Both the Director of Internal Auditand myself are Assistant <strong>Compliance</strong>Officers at Rush. The Chief <strong>Compliance</strong>Officer is Cathy Jacobson,Associate Vice President for ProgramEvaluation and Special Assistant tothe President.Cathy spends about 25% of her timeon compliance matters and the remainderon a variety of other duties. Thedirector of internal audit and I tend todo most of the day-to-day workContinued on page 1413July 2001