“Lumpers” and “splitters”were terms originally used to describe scientistswhoapplied centripetal or centrifugal forces, respectively, to the evolving species taxonomy and other nosology debates in the 1800s. The first use of these 2 terms is attributed to Charles Darwin: “Thosewhomakemany species are the ‘splitters,’ and those who make few are the ‘lumpers.’”1(p463) Indeed, evidence fromcognitive science suggests that we have individual differences in the way we view information2; some of us process information in a way that leads us to perceive the “whole,” or similarities, in our research, practice, and teaching, whereas others of us process information in a way that leads us to perceive the “parts,” or the distinctions that can inevitably be found. These opposing information-processingandgeneral tendenciesareplayingout inour approach to entities asdifferent as themedical care that we provide, themedical reimbursement systemunderwhich wepractice, and the types of researchdesigns, questions, and approaches that we use in our science. For example, the AffordableCareAct seeks topromote the “lumper-friendly”Primary Care Medical Home in which it is intended that 1 team of clinicians will deliver comprehensive care for multiple chronic diseases, conditions, and symptoms rather than split this care among a disparate group of specialists across different health systems. Yet, at the same time, we continue to see “splitting” tendencies inmedical care, as subspecialties continue to flourish,3 and the frequency and proportion ofmedical care that arises from referrals continues to escalate.4 Whereas theremaybeanormaldistributionof lumping tosplitting tendencies in the general population, scientists tend as a group to be splitters (as a gross stereotype). The impact of this splitting tendencyontheresearchquestions that scientists find interesting and the research designs that they accept as rigorous and worthy of funding is interesting to contemplate. It is also interesting to contemplate how lumpers (think “generalists” here) might view the usefulness and importance of research findings createdand fundedby splitters.Dependingon the manner in which you characteristically process information, you are going to either be enchanted with or have grave concerns about one of the randomized clinical trials reported in this issue of JAMA Internal Medicine.5 In a single-blind, randomized clinical trial of 183 patients hospitalized for acute coronary syndrome,heart failure, or arrhythmiaandwith symptomsofdepression, generalizedanxiety, or panic, Huffman and colleagues5 tested whether telephone-based collaborative mental health care treatment, deliveredby a socialworker and a teamof psychiatrists, could improvemental and physical health across 6months of treatment comparedwith usual care. They found statistically and clinically significant improvements in mental health–related quality of life—the trial’s primary outcome. Patients who received the treatment also experienced significant improvements in depressive symptoms, physical health–relatedquality of life, self-reportedmedication adherence, and functional capacity; improvements in anxiety symptoms or cardiac readmissions were not observed. Strengths of this trial include the low-cost, low-intensitynatureof the intervention, the large effect sizesor improvementsobserved formultiple endpoints, and the use of cognitive behavioral therapy workbooks that may facilitate dissemination of this intervention. Limitations include theuse of a single socialworker to deliver thepsychotherapy, the absence of cost data or formal cost-effectiveness analyses, and theuseof screeningmeasures toprovideprobabilistic but not definite diagnoses of clinical depression, anxiety disorders, and panic disorders. Such a trial is unusual for a number of reasons. Scientists rarely think that 1 riskmarker, or exposure, operates the same way indifferent diseases and thusoften frownon lumpingpatientswithdifferentdiseases together in theabsenceof apriori analytic plans to statistically determine the appropriateness of such lumping. The authors of this study, however, included patients with different forms of cardiac disease. Similarly, scientists rarely lump different risk markers together— even if they are highly comorbid, as is true of depression and anxiety—into 1 treatment focus.Whenresearchershavesought to do so, as was the case with the creation of the “metabolic syndrome” concept, extensive criticism of such a fusing (or lumping)approach isvoiced.6Theauthorsof this study lumped together patients with symptoms of 3 commonmental disorders, each of which has been associated with excess cardiovascular risk. Finally, scientists almostnever concurrently test whether a single treatment is effective at reducing the adverse health effects of multiple risk factors. This approach is essentiallywhatHuffmanand colleagueshave attempted.Although cognitive behavioral therapy is an evidence-based intervention for all 3 sets of psychiatric symptoms targeted in this trial, we rarely (if ever) test its effects onmultiple conditions concurrently. Finally, to complete their coup d’etat of a lumping exercise, Huffman and colleagues employed a generalist socialworkerwithminimalcognitivebehavioral therapy training to provide psychotherapy in this trial. It is common in mental health fields to have advanced training in treating only 1 or 2 of the mental conditions targeted in this trial. Of course,most trainees receivebasic clinical training in the treatmentof all 3, but as inmedicine, therehasbeenadistinct splitRelated article page 927 Research Original Investigation TheMOSAIC Randomized Controlled Trial