Document NGVgJZrXJn6YvgqQ9kO9pyGNQ

I 4 4 * ,7ZZr- r& ,/ Letters to the Editor Frequency of Record-Availability, Incorrect Diagnosis, Among Kodak and Rochester Area Brain Tumor Cases Identified Kodak Cases Rochester Area Casas Readers are invited to submit letters for pub lication in this department. Submit to Doris Total No. % Total No. % p (2-sided) Flournoyt Publisher, Journo! of Occupation al Medicinet 1845 W. Morse, Chicago, IL 60626, Letters should be typewritten, double spaced andshould be designated "For Publication. '* No hospital records available Metastatic cancer or - non-cancer ',CC / ni r 64 3 4.7 84 15 17.9 0.036 61 7 11.4 69 10 14.4 0.798 Cases with records available, 61 3 45 69 7 10.1 0.430 Conclusions Questioned in Brain clinical diagnosis Tumor Excess Study Cases with records available, 61 51 83.6 69 52 75.4 0.345 To the Editor -- We are greatly con pathologic confirmation cerned about errors in the recent article: "Diagnostic Sensitivity Bias-An Epidemiologic Explanation for an odak employees as for Rochester to only 82.7% for all cancer sites Apparent Brain Tumor Excess" by Dr. Peter Greenwald et al (JOM 23:690694, 1981). These errors, we feel, invalidate the conclusions drawn and suggest alternative explanations. Since the authors correctly state that dem onstration of this type of bias could have far-reaching effects on all occupa tional epidemiology, proof of its demonstration must be scrutinized in detail. The basic problem appears in Table 3 where the frequency of record availability, incorrect diagnosis, clini cal diagnosis, and pathologic confirma tion is compared between brain tumor Area residents (11.4% versus 14.4%). The data show that the only major dif ference between a death certificate indicating brain tumor for a Kodak employee and one for a Rochester Area or New York State resident is that the Kodak employee was more likely to have obtainable hospital records. This fact itself may be ex plained by the special relationship of the study authors to Kodak employees and perhaps, their physicians. Further down in Table 3, the fre quency with which certain diagnostic techniques were used is compared. Brain scans and pneumoencephalo combined.1 If the diagnostic sensitivity bias really exists as a problem in occu pational epidemiology, it was not demonstrated by this study. ,M Peter Gann, M.O., M.5. Chief, Occupational Medicine Kenneth 0. Rosenman, M.O. Director, Occupational and Environmental Health Services New jersey Department of Health Trenton, New jersey 08625 Reference 1. Percy C, Stanek E, and Gloeckler L: Accuracy of cancer death certificates and its effect on cancer mortality statistics. Am J Pub Health 71:242-250,1981. i, 1... Authors' Response cases involving Kodak and non-Kodak employees identified by death certifi cate. Hospital records were not avail able for nearly 18% of the Rochester Area tumor comparison group, as grams were used more often for the Kodak employees. We suggest that these differences are irrelevant when the frequency of craniotomy and autopsy, the definitive procedures for Substracting study subjects not having hospital records from the denominator in the upper part of Table 3 Increases the difference be tween cases and the Rochester tumor opposed to 4.7% of the Kodak group. diagnosing brain cancer, were per comparison group with respect to Cases for whom records were unavail formed at similar rates in Kodak em misdiagnosis and clinical diagnosis. able were then included in the denom ployees. An alternative explanation for The effect is he opposite of that inator while calculating frequency of incorrect diagnosis, clinical diagnosis these data is that patients with com prehensive medical insurance are sub .assertedJq^Gann and Rosenman. This can be seen by comparing the'per and pathologic confirmation. This jected to a greater number of diag centages in their table to our Table 3, amounts to assuming that all cases nostic procedures. reproduced on the next page. for whom records could not be ob Finally, the article's assertion that The Gann ;, and Rosenman distor tained either did not have pathologic ' confirmation or did not die of brain cancer. We do not find this to be a reasonable assumption. '' Recalculation of these frequencies and p values when record-not-available cases are omitted shows that'all sig nificant differences between the Ko brain tumors are difficult to diagnose relative to other fatal disease must be questioned, except perhaps when referring to the very elderly. In the population under study, which was all presumably under retirement age, there is little reason to believe that brain tumors remain undiagnosed. The tions of the'death certificate denomin ator for purposes of calculating rates of pathological confirmation are inval id. The total number of death certifi cates noting brain tumors is the correct denominator for all four upper rows of Table 3, since the basic'question was , about coding on death'certificates. dak group and the tumor comparison groups disappear. Only results for the Rochester Area group are shown in the proportion of death certificates in this study, confirmed by records,,or by pathology.- was in accord |with^dati " Verification of pathologic confirma- ly does, not depend on SWMMBbi*** "***&'. table, but results for the. statewide... fronfThe Third National Cancer Sur- group were similar. veyHvhich showed that 89% of death Thus, a death certificate diagnosis certificate brain tumors were con files as .well'as hospital records were was just as likely to be incorrect for firmed by hospital records as opposed searched in this study. j?-,- 428 Letters to the Editor UCC 096392 It is suggested by Gann and Rosenman that we, perhaps because of "a special relationship" with employees and physicians, were in a better posi tion to obtain hospital records of Kodak employees than controls. This implication is incorrect. The study was designed so that the investigators could not have affected the record retrieval process and therefore the availability of records. The hospital records were sought directly by the New York State Department of Health by persons unaware of cases/ tumor comparison group status. Study cases and controls were dispersed in many different hospitals in Upstate New York. Drs. Gann and Rosenman are in error when they presume the study group was all under retirement age. In fact, approximately 25% of the cases were retirees and thus cross the spec trum of "the very elderly," a group in which our critics agree that brain tumors are difficult to diagnose and thus are vulnerable to the diagnostic sensitivity bias. The potential for such a bias is clearly shown in a recent report of experience in Rochester, Minn., where the mean annual inci dence of primary central nervous system neoplasms in individuals over age 55 is dramatically influenced by autopsy results - with most cases being lirst dijgnoscd alter death.1 The total excess in our study was in the population over age 55, with cases being diagnosed up to age 81. During the study period, brain scans and pneumocncephalograms were important diagnostic procedures, with reported diagnostic sensitivity of 82% and 94% respectively.2 We agree that craniotomy with biopsy and autopsy often may be definitive. We agree that craniotomy with biopsy and autopsy often may be definitive. We note again that all eight procedures were more commonly performed in cases than in Rochester area controls and that seven of eight were more commonly performed in cases than in statewide controls. This pattern of 94% (15/16) of the diagnostic com parisons showing higher degrees of testing for Kodak employees provides evidence of a diagnostic sensitivity bias. Gann and Rosenman cite the Percy et al3 reference, but neglect to des cribe these investigators' comments about death certificate coding of brain tumors. To quote from Percy et al, "Word choice is critical for determin ing the correct code. For example, when malignant brain tumors such as astrocytoma or glioma were diagnosed in the hospital, the physician fre quently signed out the death certifi cate as 'brain tumor ' This cause of death is coded to category 238.1, neoplasm of unspecified nature of the brain and is therefore not recorded as a cancer death." It is difficult to see how our critics missed this message, since even the brief article abstract states: "Other misclassification prob lems were found for cancer of the uterus, brain, and buccal cavity, in cluding most of its sub-sites." Because of this potential bias, in our study the 238.1 category was included in death certificate searches. In the Percy et al study, cancers metastatic to the brain were the most common group misclassified as pri mary brain tumors (personal commun ication from C. Percy). Thus, coding rules and difficulties that coders have in applying them in a uniform and consistent manner also can lead to artifacts in brain tumor rates. We hope this clears up any confu sion that may result from the Gann and Rosenman letter. The possibility of a diagnostic sensitivity bias remains an important consideration in this type of investigation, particularly where the study population has liberal access to high quality diagnostic testing, employee insurance coverage, Contlnad on page 432 Table 3. -- Frequency of Pathology Confirmation and Diagnostic Procedures on Cases and Tumor Comparison Groups. Coded as Brain Tumor on Death Certificate No hospital records available Not cancer or metastatic cancer Clinical diagnosis (no pathology) Pathologically con- tlrmedt Pathological or Clinical Diagnosis o( Brain Tumor Craniotomy, brain surgery or biopey Brain scan Angiogram EEG Pneumoencephalogram Autopsy Skull x-ray Eailmao kodak Caaaa Total No.* % 64 3t 4.7 64 7 10 9 64 3 4.7 64 51 79.7 54 54 54 54 54 54 54 48 889 33 61 1 41 75.9 20 37 0 19 35 2 28 51 9 50 92.6 Aochntir Area Tumor Cantperlten Breupa TeUI NO.* S P Statewide Turner Comparison Groupi Total No.* K P 84 15 17.9 0.036 84 10 11.9 0.623 84 7 8.3 0 381 84 52 61.9 0.019 100 13 13.0 0 166 100 17 17.0 0 170 100 10 10 0 0 219 100 60 600 0 008 59 49 83.1 0.373 59 25 42.4 0 046 59 40 67 8 0 338 59 20 33 9 0 727 59 9 15 3 0 014 59 28 47 5 0.782 59 50 84.7 0 191 70 53 75.7 0 061 70 21 30 0 0,0005 70 45 64 3 0 163 70 23 32 9 0 627 70 12 17 1 0 021 70 24 34 3 0 049 70 67 95 7 0 455 `Number that hospital records indicated procedure was perlormed t Two ot these are included with the 54 pathologically or clinically diagnosed cases lor tne study ot exposures, giving the 56 casts shown m Table i (see text for details) tOI the pathologically confirmed cases, slides could be obtained lor review on 51 cases. 46 Rochester area matched controls, and 44 Statewide matched controls as shown in Table 2 p - Significance levels lor dillerences in proportions having procedures between Kodak cases and comparison groups Journal of Occupational Medicine/Vol. 24, No. 6/June 1982 429 UCC 096393 P nstel (mefenamlc add) Betoie presenting please set WIpfescrifiinQjnfwmatiofl A Brief Summary idtows INDICATIONS AND USAGE: PonsiH is indicated tor me relief of moderate pain wfien merapy will not exceed one wee* Ponstei -s also indicated lor me treatment ot primary dysmenorrtiea Studio m cnndrcn under t4 yeau of age nave been madervaie to evaluate the safety and effectiveness of Ponsiel CONTRAINDICATIONS: Ponstei should not be used in patients who have previously exhibited hypersensitive ity tpil Because the potential exists for uoss-svisttivity to aspirin or other nonsteroidal antiinflammatory drugs Ponstef should not be given to patients in whom these drugs induce symptoms ot txonchospasm allergic mi nors, or urticaria Ponstef is cdnframdtcated t pafrertts with active ulceration or chrome inflammation of either the upper or tower gastrointestinal tract Pon$te< should be avoided m patients with preexisting renal disease WARNINGS: in patients to a history oi utceranon or chrome inflammition of the upper or tower gastrointesti nal trad. Ponstei should be given under close supervision and only after consulting the Adverse Reactions Section it diarrhea occurs the dosage should be reduced or temporarily suspended (see Adverse Reactions and Dos age and Administration) Certain patients who develop diarrhea may be unable to tolerate (he drug because of recurrence cf me symptoms on subsequent exposure PRECAUTIONS: H rash occurs administration Of the drug should be stooped A lalse-Oosrtive reaction lor urinary bile, using the dtuo tablet lest may result alter mefenarmc acid admin istration it biliurta is suspected, other diagnostic procedures, such as me Harrison spot test should be performed in chrome animal toxicity studies of Ponstei al doses 7 to 28 times the recommended human dose, rats had minor microscopic renal papillary necrosis dogs had edema and blunting of the renal papilla, and monkeys had renal papillary edema Normal human volunteers had mild BUN elevations with prolonged administration at greater than therapeutic doses The significance of these linings is unknown Hovwver, since Ponsiel <s elimi nated primarily through the kidneys, the drug should not be administered to patients with significantly impaired renal function tnlormetlew tr Patients: Patients should be advised mat ! rash ckarmea or other digestive problems arise, they should slop the drug and consult their physician Patients m whom aspmn or other nonsteroidal antiinflammatory drugs induce symptoms of bronchosoasm allergic rhinitis or urticaria should be made awv* mat the potential exists lor aoss-sensmvity to Ponstef The long-term effects, tl any. of intermittent Ponstei therapy tor dysmenorrhea are not know Mtomen on such therapy should consult their physician it they Should decide to become pregnant Oitef lettractions: Ponsld may prolong prothrombin time Therefore, when the drug <s Mnmistered to patients receiving oral anticoagulant drugs. Irequtd monitoring of prothrontyn tune is necessary Use ia Pregnancy: Pregnancy Category C Reprofrchon studies have been performed m rats, rabbits and dogs Rats given up to times the human dose showed decreased tertikty delay m parturition, and a decreased rate survival to weaning Rtfbrts m2 5 txno the human dose showed an merwem the rwmper of reso^f' lions There were no letat anomalies observed m these shidtes norm dogs at to Q tune me human doM There are no adequate and wetkorwoiied stoics m prignartwomMt Becra jftmairvroduaon stoics are not always aeiciiv# d human response, frs drug should be used only it ctearfy needed The use d Pon$ut m late pregnancy <s not recommended because tf the rheas on toe leal cartemtescuter system d drugs of this class Naratag Mgtoft: tacc amounts gt Ponstei nvy be present m breast met and transmuted to the nursing intent, thus PonsW should not be taken by re nursing mofhsr because d me eftees on me intern carrtamuu- iM system of drugs M this class Use ia CMldfta; Safety and dtoctrveness m eft*cn below toe age d to n*t not been attttob: ADVERSE REACTIONS: Gnlistotoiltoil. The most frequently ivoned atom rodions asocnwd with me use d Ponstei involve me gasPointestmat tract to controlted stoats tor up to etfu motto, me following disturbances woe reported m decreasing ordM oi frequency oarma (aporoumately i% d PMtosi. naus* with or witiout vomiting other gastromtestmat symptoms and abdominal pam tn certain patients, the diarrhea was d stekcteflt seventy to require oicommuMton d medcauon The oca* ince d the darmea * usually do* retend. genetally subsides on reduction d dosage and fto*dy dsappears on termination d therapy Other gastrointestinal feature less trewendy riported were anoieua. pyrosrs. ttatotence and eonsnpM>on Gastromtesimai ulceration to and twmou! hemorrhage has been rooted Herntteftottc: Cases d autoimmune nemoiytte anm have been associated wilt mt continuous aonwi- istranon d Ponstef tor 12 months or longer to sudi cases me Coomps test reubs are posiirve to evidence d both accelerated R8C production and R8C destrudon The process is reversible upon lenrunabon or Ponstei administration Decreases m hematocrit have been noted m 2-8% d paberes and primarily >n mose to hive receded pro- itmgeo therapy leukopenia eosinoteutu thrombocytoperw purpura, agrartotocytons. pancytopenia. and bone marrow hypo- ptas'a have also been isorted on ocroqn Nervous System: Drowsiness dttimss. rwousness. hotocne tewed vision and msomma have occulted lategumewtary: umcana. rash, and tacto tom have been reported Rout: A$ w<m other npnsteroidM anuWtemmMMy agents, rend tutoe induang ptotery negons has Wen reported in rtOMiypauente renal tattofenaottwadaiiM faking Ponted tor 2-6 weeks The renal damage may nor be completely reversible Hemaimi and dymeia have also been rtomdwim Ponstei Other Eye irritation mrpam penoKton rmWNpa* toafy andnteimiw need tof unto m a daPto have been reported There hive been rare reports d pupato dyspnei and revembte toss d cotor vmon OVEROOlAOt: Although doses up to 6000 mg/ttty have bewigrven no socctec idormation is avaitebte on tm management* sort rmssiveoimdouQi SbteitoaEadrtd wddosageoccte me stomach should be emted(* inducing vnoav by erto goto totegeiitotoDy to ateito*strjtandicMrtdtir^ Laboratory stuaes mdcate mat PontMl toted be adsorbed from me gasuameshnai traa by activated cnarcoat Vital tundiom should be monifired and AMponM SecMemteensiiKacidandiismeteboktesaretirmhi bound to ptasme protons, tenoomu and pernomsi ddyvs may be d ktee *atut DOSAGE AMO HPHtMffTRATWM. Aanuesuton is by me om route preferably to rood The recommended regunan m acute pun tv adults and dvidren over to you d age is 500 mg as an imtur dose tdtoto by 250 mg every in hours as needed usually not to exceed one For me treatment d primary dysmenorrhea me recommended dosage >s 500 mg as an imiiet dose idtowed by 250 mg every 6 hours wrong wr me onset d bffedng and associaied symptoms Ctovtar jivdis mdcaie inat eMectme treatment can be initiated wim tne ttan d menus ano should noi be necessary tor more man 2 to 3 day* t?ttt3AOO/iO PARKE-DAVIS Dtv o* Wamer-LamOert Co 432 Moc/iS Plains NJ 07950 USA Letters -- continued occupational medical service referrals, and the impact of neurologic and neurosurgical specialists in a univer sity school setting. Peter Greenwald, M.D. National Cancer Institute Bethesda, MD 20205 Barry R. Friedlander, M.D. Charles E. Lawrence, Ph.D. Terry Hearne M.S. Eastman Kodak Company Rochester, NY 14650 Kenneth Earle, M.D. Armed Forces Institute of Pathology Washington, DC 20306 References 1. Annegers JF, Schoenberg BS, Okaiaki H, et al: Epidemiologic study of primary intracranial neoplasms. Arch Neurol 38:21 7219,1981. 2. Witcofski RL, Maynard CD, and Roper T|: A comparative analysis of the accuracy of the technetium-99m pertech* netate brain scan: followup of 1000 pa tients. / Nud Med 8:187-196,1967. 3. Percy C, Stanek E, and Gloeckler L: Accuracy of cancer death certificates and its effect on cancer mortality statistics. Am / Pub Health 71:242-250,1981. Best Medical Care at Reasonable Cost To the Editor: I was distressed to read the political advertisement (JOM 24:191, 1982) entitled "Think of Your Prescription as a Vote." If this was paid advertising, it would be clearly marked as such and it should contain the advertiser's name. If this was an editorial, I am appalled. The patent laws of this country give the "innovator" 17 years without competition to reap a fair profit from an innovation. The insinuation that one cannot rely on generic drugs is out rageous. All drugs, brand name and generic, must meet the same federal standards for quality. As physicians, our first duty is to our patients. The drug manufacturers can make a fair profit under the pres ent laws. Our concern should be that our patients receive the best medical care available at reasonable cost. Katharine C. Rathbun.M.D., M.P.H. Director of Flealth City of Pasadena Pasadena, TX 77502 Editor's Note: The indicated mate rial was sponsored by Rolf Werner Rosenthal, Inc., an advertising agency, representing a number of pharmaceuti cal companies. Letters to the Editor UCC 096394