Science Culture & Accountability Plan

Transcription

1 Duke Department f Radiatin Onclgy Duke University Schl f Medicine Science Culture & Accuntability Plan Duke University is cmmitted t maintaining the highest quality and integrity f all its scientific enterprises. Because f this cmmitment, the Schl f Medicine is required t have mechanisms t guarantee the respnsible management and critical review f scientific data. The Schl f Medicine has pted t allw individual departments t adpt their wn plicies and prcedures related t scientific accuntability and integrity. Fr this reasn, the Department f Radiatin Onclgy is cmmitted t ensuring that plicies and prcedures are in place t reflect the highest prfessinal cnduct and t prmte a culture in which scientific results are critically reviewed with accuntability fr data integrity. Each labratry maintains a Science Culture & Accuntability Plan that is reviewed annually by the chair f the Department. In the Department f Radiatin Onclgy, we recgnize this requires the active participatin f all parties in the research missin, all students, pstdctral fellws, visiting schlars and staff are respnsible fr the dcumentatin f all data generated in the prcess f perfrming research in their respective lab. In additin t being essential t the generatin f scientific knwledge, prper data management is a labratry practice that ensures reliability and reprducibility in all wrk. Principles Recgnizing this, the Department f Radiatin Onclgy has prepared the fllwing Science Culture & Accuntability Plan. This plan prmtes a culture that encurages respnsible data management and prduces data f the highest integrity and als reflects these imprtant principles: 1. We fster an envirnment where scientific integrity is the highest pririty. 2. We emphasize high-quality, reprducible data and results. 3. We value cnstructive critiques f research. 4. We allw pen discussin f any cncerns regarding research cnduct r integrity. Each and every member f the Department f Radiatin Onclgy faculty, trainees, staff and administratrs is expected t reflect and pursue these values. Ours is a shared cmmitment t the highest standards f scientific activity. Recmmended practices fr imprving the culture f scientific accuntability within an individual labratry: It is expected that each faculty member will review the steps taken t cmply with this plicy at the annual review meeting with the chair f the Department. As a principle investigatr, yu must set the example fr yur team thrugh hnest and pen discussin f results and thrugh yur emphasis n scientific integrity and data quality ver psitive results. D nt, in any way, encurage r put pressure n lab persnnel t btain specific results. Make it clear at all times that yur highest pririty is t btain the true result f all studies, irrespective f the effect such a result may have n the verall prject, grant submissin, r manuscript. Make it clear that yu have a zer-tlerance plicy with respect t data manipulatin, alteratin, r falsificatin.

2 High-quality research begins with careful planning and study design. Engage apprpriate cllabratrs, statisticians, and ther relevant team members fr cnstructive input befre actual experiments r clinical studies begin. Having well-defined study gals prtects against fraud and imprves the quality f results. Frame yur research questins in ways that test a hypthesis rather than cnfirm a result r mdel. This allws negative and psitive results t be interesting and useful; that is, refrain frm expectatins that ne type f result is mre valuable than anther. Plan fr multiple methds, techniques r analytic appraches t test the same hypthesis r ask the same questin, which will strengthen cnclusins by cmparing results using mre than ne apprach fr the experiments. Mst f us rely n ur trust f and judgment in thers t ensure the integrity f ur data. Hwever, this alne is nt sufficient. Examples exist where even the mst seemingly trustwrthy peple have manipulated data. While yu shuld cntinue t put yur faith in thers, yu shuld reinfrce this with specific practices and institute prcesses t ensure that yur data is managed respnsibly. Crss-train lab persnnel whenever pssible s that ne lab member can independently verify the results f anther. Primary data must be stred in a manner s that ther scientists can independently analyze and generate data, s that scientific cnclusins d nt have t rely n a single persn prviding the analysis. Implement a plicy f best practices with respect t research recrds, including basic labratry ntebks r clinical research recrds. When pssible, utilize electrnic recrding slutins that autmatically recrd date and timestamps f entries and data changes. Ensure that entries are being made in a way that cnfrms t lab plicy. Make this plicy clear t all lab persnnel and enfrce it. We recmmend yu review peple s lab ntebks any time yu meet with them t discuss data. Additinally, cnsider perfrming peridic audits f labratry ntebks t ensure that a third-party reviewer wuld be satisfied with the level f dcumentatin prvided fr an experiment. With regards t clinical research, maintain all apprpriate dcuments in accrdance with all regulatry and IRB requirements. Cnsider the implementatin f cmpetency training specific t individual lab r clinical research tasks. Rather than creating yur wn individual clinical research r labratry database, cnsider leveraging institutinal resurces such as the Duke Bibank, REDCap databases, Pedigene r ther similar centralized infrastructure that limits users, directly btains input frm surce elements, and tracks all changes in the data elements r samples. Fr any data that is generated by instruments, such as plate readers, scintillatin cunters, cameras, flw cytmeters, etc., require persnnel t include the specific instrument, its lcatin, and the date and time the analysis was perfrmed in their lab ntebk. This will make it easier t match the lab ntebk with instrument-generated raw data. In additin, maintain frequent mnitring, calibratin, and validatin f all labratry equipment. Fr any instrument-generated data, implement a prcedure by which the raw instrumentgenerated data is permanently archived in a secure lcatin that can nly be accessed by yu r a specific delegate. This may include digital data transferred t a secure server, data burned t CDs, r hard cpy printuts archived in a place ther than individual lab ntebks. Ideally, this data wuld be stred in a read-nly frmat that culd nt be altered nce it is depsited. The Department f Radiatin Onclgy will prvide server space fr investigatrs t permanently maintain all raw/surce data that becmes part f any published manuscript t be. The U.S. Department f Health and Human Services requires that all prject data be retained fr at least 3 years after the funding perid ends. Develp a plan with cllabratrs t ensure data integrity. Fr example, yu may request a cpy f the raw data generated by yur cllabratrs fr archiving in yur wn lab. Similar t

3 data analysis perfrmed fr yur labratry-generated data, when pssible, perfrm an independent analysis f data generated by cllabratrs t verify accuracy. T the extent pssible, independently replicate study results. Fr example, have different peple perfrm experimental prcedures (such as treating grups f mice with a drug) and experimental readuts (such as determining the phentype f the treated mice). It is als gd practice t require that persns assaying readuts be blinded t the experimental grups. Yu r a designate shuld re-analyze all critical studies, such as thse included in grant r manuscript submissins, starting with the archived raw data. Yu may cnsider including a persn with the apprpriate expertise utside yur lab. Fr clinical r translatinal studies, this can be accmplished by having study results independently analyzed by a statistician separate frm the investigative team. Develp a prcess in which yu dcument critical results, the date yu learned f them, yur interpretatin f these results, and cnclusins r discveries that these results imply. Fr example, when yu meet with a lab member t review new data, the lab member can prvide yu with a dated versin f the experiments and results, which yu can keep in a separate ntebk fr each lab member. This ntebk will serve t dcument yur wn review f the raw data. These recrds will nt nly allw yu t dcument the intellectual prgress f specific prjects t develp future hyptheses r research plans, but als when needed, t supprt intellectual prperty claims. Furthermre, if questins f data integrity were t arise, such a ntebk wuld serve t dcument that yu reviewed all critical studies as the data were generated t the full extent pssible. All individuals engaged in research will be required t cmplete nline training mdules, similar t thse required by IACUC and the IRB, that emphasize these principles. (The Department and SOM will cllabrate t develp these new mdules). In additin, individuals shuld als take advantage f prgrams ffered thrugh the SOM that are designed t address research integrity. Additinal cmpetency training fr investigatrs and their research staff can include the current cre curriculum frm Duke Office f Clinical Research, i.e. Infrmed Cnsent Prcess, Study Dcumentatin, Data Integrity and Security, and annual Human Subject Research Overview fr clinical investigatrs. At a faculty level, participatin in institutinal prgrams such as LEADER will prvide pprtunity fr further educatin in research management and versight. If yu have cncerns abut the integrity f smene s data, be it in yur lab r smene else s, yu shuld feel cmfrtable vicing yur cncerns. This is true whether yu think a certain analytic methd needs t be better validated r if yu suspect scientific miscnduct. Raising cncerns abut data integrity is nt the same thing as accusing smene f scientific miscnduct. The Department f Radiatin Onclgy wishes t prmte a culture in which all aspects f scientific findings are critically reviewed. This includes all steps in the scientific prcess, frm study design t data acquisitin t methds f analysis t the frmulatin f cnclusins. Raising and respnding t questins abut data integrity shuld be a rutine part f the critical review prcess in ther wrds, it need nt be reserved slely fr cases f suspected scientific miscnduct. It is thrugh this prcess that we all can wrk tgether t ensure the highest pssible quality f science at Duke. Have a mechanism in place within yur labratry r research team t prevent miscnduct and handle research-related cmplaints, reflecting a n-tlerance plicy related t falsifying data, deceptive advertising, r enrllment f subjects nt qualified t be in studies. If yur lab engages in translatinal r 'mics research, discuss with yur staff invlved in these studies the challenges inherent t this kind f research. Review the existing and new resurces available acrss Duke, especially thse t make yur research mre reliable and

4 transparent. Recmmended safeguards t ensure data integrity within yur lab The Department f Radiatin Onclgy believes that the prper accuntable unit fr ensuring data integrity is each individual labratry. Fr this reasn, all labs within the Department f Radiatin Onclgy shuld fllw the key principles and steps utlined in this Science Culture & Accuntability Plan. Althugh this dcument utlines the best practices with regards t scientific accuntability, we als recgnize that within each lab nt all steps, r the same set f activities, may be apprpriate. Ultimately, the Department requires that the Faculty maintain evidence f cmpliance, as suggested belw. Labs must ensure that all divisinal investigatrs and persnnel engaged in scientific research cmplete all required institutinal training mdules in respnsible data management. These will be develped and mnitred fr cmpliance thrugh central SOM resurces, similar t the IRB mdules required fr studies invlving human subjects. Labs must ensure that all investigatrs implement and maintain plicies fr respnsible data management within their labs r research grups. Labs must als establish mechanisms by which the validity and integrity f critical data generated can be cnfirmed. These mechanisms shuld nt place an undue burden n the investigatrs, but shuld ensure that investigatrs are adhering t plicies f respnsible data management Labs must ensure that all investigatrs regularly present their research findings t ther investigatrs utside their wn lab r research grup in a frum that allws pen and critical discussin f the data and its analysis. Examples include participatin in ther lab meetings, regularly scheduled multi-disciplinary grup meetings thematically rganized arund cmmn research interests, r at the annual retreat f the Radiatin Onclgy & Imaging Prgram. Review and fllw best practices fr data integrity and manuscript preparatin as required by the tp jurnals in yur field(s). Labs must identify a Labratry Integrity Crdinatr, t whm any cncern r questin abut research data management can be taken by any individual inside r utside the lab. Lab managers are expected t cnfirm that each investigatr is aware f the varius schl, department and divisin resurces that supprt researchers and their activities, including access t statistician input and review f study design, analysis and publicatin. Labs are encuraged t require that faculty and staff cmplete the Data Integrity and Security class presented by DOCR r similar n-line mdule training, as apprpriate. Encurage lab persnnel t discuss any cncerns they may have with the primary investigatr r ther investigatrs at Duke. Departmental effrts t prmte a culture f scientific accuntability. The Department f Radiatin Onclgy leadership will als take steps t supprt, guide and ensure a culture f scientific integrity, including the actins listed belw. Outline a chain f persns available t address research integrity cncerns. These individuals will be available t assist faculty and staff within the labs and will be ready t address ptential cncerns raised. Cncerns shuld be raised and addressed initially t the departmental vice chair fr basic and translatinal research, but in case f perceived cnflict f interest, cncerns may be raised directly with the chair.

5 Wrk with the SOM t develp institutinal plicies and mdules in training f respnsible cnduct f research and ensuring data integrity. Develp a mentrship prgram fr new investigatrs that includes an emphasis n the principles utlined abve. Strengthen the understanding that adhering t the SCAP principles is relevant t hw we prvide patient care and useful t future studies in humans. Institute unannunced research reviews t allw investigatrs t explain and demnstrate their prcesses and prcedures fr data integrity and analysis cnfirmatin. Allw this t be a learning experience fr the investigatr, and share lessns learned with the rest f the department as apprpriate. Cnfirm that each clinical research staff takes the Human Subjects Research at Duke training curse and annual recertificatin, as currently mandated by the SOM. Require each investigatr with active IRB prtcl(s) t cmplete rle-based training in Infrmed Cnsent Prcess, Data Integrity and Security and Study Dcumentatin, as required fr CITI certificatin. Develp and refine educatinal mdules based n staff requests and findings frm audits and IRB reviews, including a new series f Hw T vignettes as well as wrkshps and financial management training. Develp an annual cmpetency evaluatin f specific skills required fr research rles. Educate the faculty and staff f the Department abut available resurces and reprting mechanisms fr scientific accuntability, such as: The NIH Office f Research Integrity ( Guidelines fr the Prper Handling f Digital Image Data ( Online Learning Tl fr Research Integrity and Image Prcessing ( The Cmpliance and Fraud Htline: T annymusly reprt a suspected cmpliance vilatin r cncern, the Cmpliance and Fraud Htline at Duke is at