Data Analytics and Reporting in Pharma

Background and Challenges

In a high­ly regu­la­ted indus­try like phar­maceu­ti­cals, com­pa­nies have to docu­ment and mana­ge their pro­duc­tion pro­ces­ses and accor­ding data con­ti­nuous­ly, as well as put them to use in order to opti­mi­ze their own pro­ces­ses and pro­ducts. This also appli­es to CSL Beh­ring. The glo­bal­ly acti­ve com­pa­ny deve­lo­ps bio­the­ra­pies from plas­ma – in eight pro­duc­tion sites world­wi­de, with more than 25.000 employees. Con­cer­ning data manage­ment and ana­ly­tics the past years included the fol­lo­wing chal­lenges:
  • Imple­men­ta­ti­on of a paper­less coll­ec­tion, con­ver­si­on and espe­ci­al­ly vali­da­ti­on of data at various points within the pro­duc­tion pro­cess.
  • Bet­ter and fas­ter ana­ly­sis of data from dif­fe­rent source sys­tems: get the data to employees in a way that they always have the rele­vant infor­ma­ti­on at the right time and place in order to gene­ra­te addi­tio­nal value from it.
  • Com­pli­ance with con­ti­nuous­ly gro­wing regu­la­to­ry requi­re­ments in the indus­try – con­cer­ning appli­ca­ti­ons, ana­ly­ti­cal models and the respec­ti­ve data.
A typi­cal pro­blem in an orga­ni­cal­ly grown struc­tu­re: Employees in dif­fe­rent are­as of pro­duc­tion coll­ect data manu­al­ly in tables or spe­ci­fic data­ba­ses. But the data is not available for other users within the com­pa­ny. It is hid­den and the­r­e­fo­re not con­side­red nor used to opti­mi­ze pro­ducts and pro­ces­ses. Many of the data sources in a com­pa­ny are not elec­tro­ni­cal­ly con­nec­ted, manu­al copy-and-pas­te-pro­ces­ses are in place. The­se work­flows are ela­bo­ra­ti­ve and error-pro­ne. The struc­tures make report­ing sys­tems in pro­duc­tion very com­plex and infle­xi­ble, ana­ly­tics as a basis of opti­miza­ti­on is limi­t­ed. Sce­na­ri­os like the­se have to be a thing of the past in data-dri­ven digi­tal world.

The Solution

To opti­mi­ze the data manage­ment within the com­pa­ny along the abo­ve-men­tio­ned cri­te­ria, CSL Beh­ring is con­ti­nuous­ly estab­li­shing a com­pa­ny-wide data warehouse, in order to sim­pli­fy access to all data sources. Along­side, CLS Beh­ring is using TIBCO® Sta­tis­ti­ca™ Enter­pri­se (part of TIBCO® Data Sci­ence) as a uni­ver­sal data sci­ence plat­form sin­ce 2013. This deve­lo­p­ment is dri­ven by a clo­se part­ner­ship with Stat­Soft, seve­ral vali­da­ted appli­ca­ti­ons and other ana­ly­ti­cal pro­jects. Addi­tio­nal­ly, many exis­ting ana­ly­ti­cal pro­ces­ses were repla­ced and auto­ma­ted within the plat­form. Among the coll­ec­tively imple­men­ted pro­jects so far the­re are auto­ma­ted reports rela­ted to vali­da­ti­on and sta­bi­li­ty. The­se report pro­ces­ses were to some ext­ent known for their high expen­dit­u­re of time and error rate, due to the manu­al data coll­ec­tion. With the sup­port of Stat­Soft in the deploy­ment of new appli­ca­ti­ons based on TIBCO® Sta­tis­ti­ca™ Enter­pri­se , lean pro­ces­ses were deve­lo­ped.

Such as: auto­ma­ted import and pro­ces­sing of data, inclu­ding auto­ma­ted audit trails and aut­ho­riza­ti­ons (elec­tro­nic signa­tu­re) as well as auto­ma­ted report­ing, e.g. (encrypt­ed) pdf docu­ments. Thus, CSL Beh­ring can com­ply with regu­la­to­ry requi­re­ments, having fast and trans­pa­rent pro­ces­ses. Due to the auto­ma­ti­on of data pro­ces­ses from ana­ly­tics to report­ing, thou­sands of work hours were freed. This capa­ci­ty is now used for inten­si­ve ana­ly­ti­cal work which again gene­ra­tes addi­tio­nal value. A pro­per and important step towards digi­tal pro­cess opti­miza­ti­on. Con­cer­ning the IT, addi­tio­nal eco­no­mi­zing was pos­si­ble, becau­se the IT depart­ment now mana­ges one plat­form ins­tead of various tools and lar­ge amounts of code. The com­bi­na­ti­on of increased effi­ci­en­cy and cost reduc­tion equ­als CSL’s invest­ment mul­ti­ple times.

 

The Formula for Success: StatSoft as a catalyst and supporter of a progressive data analytics team

For the deve­lo­p­ment of ana­ly­ti­cal pro­jects, Stat­Soft usual­ly sup­ports their cli­ents as a spar­ring part­ner. This also appli­es to CSL Beh­ring. With a pro­of of con­cept as start­ing point, some func­tion­a­li­ties of the TIBCO® Data Sci­ence™- plat­form were put to use in prac­ti­ce. Over the past years, various data sci­ence appli­ca­ti­ons were deve­lo­ped in addi­tio­nal pro­jects. The core of the long­time coope­ra­ti­on is the ite­ra­ti­ve pro­cess bet­ween experts on both sides. Pro­jects are deve­lo­ped through a dyna­mic exch­an­ge of ide­as, trig­ge­red for exam­p­le by ques­ti­ons, new­ly available data or new regu­la­ti­ons. This agi­le approach, based on mutu­al trust, com­bi­ned with the indus­try know-how of the cli­ent and the tech­ni­cal and ana­ly­ti­cal know-how of Stat­Soft experts leads to fas­ter suc­cess than long spe­ci­fi­ca­ti­on sheets and clas­sic water­fall methods.

Over the years, a sym­bio­tic team deve­lo­ped – con­sis­ting of CSL spe­cia­lists from the Data Ana­ly­tics Cen­ter of Excel­lence (DACE) and a long-term-acti­ve Stat­Soft team of experts – that works accor­ding to the client’s indi­vi­du­al cir­cum­s­tances, e.g. with a spe­ci­fic pro­ject method. The small team is able to imple­ment com­pa­ra­b­ly many pro­jects in data ana­ly­tics effi­ci­ent­ly and thus gene­ra­tes addi­tio­nal value for many users. That is why the coope­ra­ti­on is high­ly valued within the orga­niza­ti­on of CSL Beh­ring. Ano­ther cru­cial point of suc­cess are sys­te­ma­tic and repe­ti­ti­ve trai­nings on the use of the plat­form for CSL employees. With the help of a train-the-trai­ner-approach, more than 300 users on dif­fe­rent levels are curr­ent­ly using various appli­ca­ti­ons, only taking into account the CSL-site in Bern.

 

The platform: TIBCO® Statistica™ Enterprise

As a uni­ver­sal ana­ly­tics plat­form TIBCO® Sta­tis­ti­ca™ Enter­pri­se inte­gra­tes into the digi­tal struc­tures at CSL Beh­ring. Data from various inde­pen­dent sources are gathe­red in the data warehouse, com­ple­te­ly paper­less. From the­re they sup­p­ly various ana­ly­ti­cal pro­jects.

TIBCO® Sta­tis­ti­ca™ Enter­pri­se sim­pli­fies the­se pro­ces­ses with its gate­ways to various sys­tems and ther­eby estab­lishes a basis for suc­cessful ana­ly­ti­cal work. At the same time, neces­sa­ry aut­ho­riza­ti­on pro­ces­ses are inte­gra­ted into the sys­tem. New data sources can easi­ly be inte­gra­ted into this infra­struc­tu­re.

The platform’s func­tion­a­li­ties enable two ways of gathe­ring insights from the data wit­hout dif­fi­cul­ty:

  • Gene­ra­te infor­ma­ti­on based on par­ti­cu­lar data
  • Prepa­re and ana­ly­ze matching data to a cer­tain user sce­na­rio

TIBCO® Sta­tis­ti­ca™ Enter­pri­se enables the deve­lo­p­ment of appli­ca­ti­ons for dif­fe­rent pur­po­ses and user groups, step by step and within one plat­form. This is not only (cost)efficient but enables also the incre­asing auto­ma­ti­on of data ana­ly­tics. Miss­ing func­tion­a­li­ties can easi­ly be added in the pro­cess. All this faci­li­ta­tes data manage­ment, for exam­p­le con­cer­ning the manage­ment of dif­fe­rent ver­si­ons, docu­men­ta­ti­on and vali­da­ti­on of data. With its low code-approach the plat­form is attrac­ti­ve to broad user groups and the ana­ly­ti­cal results can be shown to various users more easi­ly, e.g. through mul­ti­ple opti­ons of visua­liza­ti­on.

Seve­ral vali­da­ted appli­ca­ti­ons on the ana­ly­ti­cal plat­form ori­gi­na­ted from the clo­se coope­ra­ti­on of the data experts at CSL and Stat­Soft. Inclu­ding various auto­ma­ted reports like Month­ly Trend Reports (MTR) or Annu­al Pro­duct Reviews (APR), Impu­ri­ty Pro­fi­le Reports, Con­ti­nuous Pro­cess Veri­fi­ca­ti­on Reports (CPV). What should be men­tio­ned spe­ci­fi­cal­ly is that Vali­da­ti­on Pro­to­cols as well as Sta­bi­li­ty Stu­dies and Trend Ana­ly­ses or Release Reports are now com­ple­te­ly digi­ti­zed. This includes the vali­da­ti­on, elec­tro­nic review of the data up to the com­pi­la­ti­on of reports in an encrypt­ed pdf for­mat. The­re are addi­tio­nal appli­ca­ti­ons: a com­pa­ri­son of pro­cess con­trol para­me­ters and qua­li­ty attri­bu­tes, inclu­ding seve­ral pro­duc­tion sites or the increase of the yield in pro­duc­tion through the appli­ca­ti­on of boos­ted regres­si­on tree algo­rith­ms.

 

Project Assessment

StatSoft Perspective

CSL Beh­ring most sys­te­ma­ti­cal­ly pushes the use of data on the plat­form for­ward and uses the oppor­tu­ni­ties of TIBCO® Sta­tis­ti­ca™ Enter­pri­se in the are­as of vali­da­ti­on and appr­oval like no other Stat­Soft cli­ent. Over­all, the opti­miza­ti­on of its data manage­ment is the basis on which CSL Beh­ring is able to meaningful­ly expe­ri­ment with cur­rent hype topics such as Arti­fi­ci­al Intel­li­gence (AI) and Machi­ne Lear­ning.

The long­term per­so­nal col­la­bo­ra­ti­on is based on gre­at mutu­al trust and open dia­lo­gue. The plat­form is deve­lo­ped in clo­se exch­an­ge, accor­ding to the client’s spe­ci­fic needs. The cli­ent knows what they need and we sup­port them with our know-how in the are­as of tech­no­lo­gy, data pro­ces­sing and data sci­ence. With this clo­se exch­an­ge and the ite­ra­ti­ve approach, we achie­ve their goals fas­ter. We are prac­ti­cal­ly the cata­lyst for the work of CSL. We lower the acti­va­ti­on effort for the deve­lo­p­ment of solu­ti­ons based on the TIBCO® Sta­tis­ti­ca™ Enter­pri­se-plat­form and with other com­mer­cial and open-source-tools.

Micha­el Busch, Lead Data Engi­neer, Stat­Soft

 

Client Perspective

In order to imple­ment the main goal – paper­less pro­ces­ses – other sys­tems would have work­ed equal­ly well. But other tools do not neces­s­a­ri­ly offer data manage­ment and ana­ly­ti­cal capa­bi­li­ties too. TIBCO® Sta­tis­ti­ca™ Enter­pri­se brings all this tog­e­ther.

Sta­tis­ti­ca gets us on the right track in our digi­tal trans­for­ma­ti­on. The clo­se coope­ra­ti­on with and the sup­port of Stat­Soft leads us to quick suc­ces­ses. When they have deve­lo­ped some­thing, we are able to ask ques­ti­ons right away, learn from that and are later able to imple­ment many things our­sel­ves. This know­ledge trans­fer is worth a lot and the saved work hours and gene­ra­ted addi­tio­nal values speak for them­sel­ves.

Jani­ne Kur­at­li, Mana­ger Data Ana­ly­tics, CSL Beh­ring

Continuous new projects

Tog­e­ther with Stat­Soft, new pro­jects in the are­as of data pre­pa­ra­ti­on, data ana­ly­tics and report­ing, which also includes a sus­tainable docu­men­ta­ti­on, are deve­lo­ped con­ti­nuous­ly. The plat­form and its appli­ca­ti­ons are super­vi­sed by a glo­bal team of CSL Beh­ring with peo­p­le working from the US, Aus­tra­lia and Euro­pe sin­ce 2019. The onboar­ding of new admi­nis­tra­tors abroad is hand­led prag­ma­ti­cal­ly through online trai­nings from Stat­Soft. This was a major pre­con­di­ti­on for the glo­bal roll­out of appli­ca­ti­ons.

Fur­ther­mo­re, work is done on the auto­ma­ti­on of addi­tio­nal solu­ti­ons in order to make the plat­form and its appli­ca­ti­ons effi­ci­ent­ly usable for even more CSL employees and sim­pli­fy pro­ces­ses fur­ther. Final­ly, the pre­vious auto­ma­ti­on steps and their mas­si­ve reduc­tion of working time are the basis for CSL’s incre­asing enga­ge­ment in more com­plex appli­ca­ti­ons in the are­as of Arti­fi­ci­al Intel­li­gence and Machi­ne Lear­ning.



Categories
Latest News
Your contact

If you have any ques­ti­ons about our pro­ducts or need advice, plea­se do not hesi­ta­te to cont­act us direct­ly.

Tel.: +49 40 22 85 900-0
E-mail: info@statsoft.de

Sasha Shiran­gi (Head of Sales)