A Comparative Study of the Use of Statistical Process Control in Monitoring Health Care Delivery

Transcription

1 Iteratoal Joural of Iovato ad Scetfc Research ISSN Vol. 14 No. 2 Apr. 2015, pp Iovatve Space of Scetfc Research Jourals A Comparatve Study of the Use of Statstcal Process Cotrol Motorg Health Care Delvery P. E. Ogutude 1, O. A. Odetumb 1, ad O. Ojo Oluwadare 2 1 Departmet of Mathematcs, Coveat Uversty, Ota, Ogu State, Ngera 2 Departmet of Statstcs, Federal Uversty of Techology, Akure, Odo State, Ngera Copyrght 2015 ISSR Jourals. Ths s a ope access artcle dstrbuted uder the Creatve Commos Attrbuto Lcese, whch permts urestrcted use, dstrbuto, ad reproducto ay medum, provded the orgal work s properly cted. ABSTRACT: To cotrol ad mprove o the qualty of products ad servces s a mportat busess strategy. Hece, the yeargs for qualty servces all sector s ot just a call from the cosumers, clets or patets but a act to mata the tegrty of the maufacturer or servce provders. I ths research, the mportace of statstcal process cotrol to health care delvery s beg re-terated wth applcato to a real data set. The p-chart s beg used ad results compared wth other results/fdgs the lterature. KEYWORDS: Health, Patets, P-chart, Qualty, Servces, Statstcal Process Cotrol. 1 INTRODUCTION I the Uted States, the health care dustry s ow oe of the largest dustres employg over 13 mllo workers wth a projected crease of over 3 mllo ew jobs by the year 2014 [15]. Reverse s probably the case Ngera where the usustaable dowward tred the health sector s worrsome both to the govermet ad to the populace. I Ngera for example, the reasos for the deterorato health care servces our ow vew ca be attrbuted to two thgs. O the part of the govermet, t turs out to be that the oly laguage the Ngera govermet uderstads s strke acto order to lste to the demads from sectors; the cessat strke actos by the Doctors s a vald proof. Hece, patets are aturally forced to leave govermet owed hosptals (though, well equpped) to seek for treatmets elsewhere. O the part of the health workers, the watg tme for a patet to get just hs/her fle from the record offce s worrsome especally Geeral Hosptals. Our govermet ad health sttutos ca smply embrace techology ad employ the use of statstcal qualty cotrol to motor ths; bearg md that t s ot oly the frst mpresso that lasts loger, the last mpresso lasts too. We have observed that whe a patet expereces a momet of mystery due to lack of professoalsm or poor servces o the part of the health workers, he/she losses cofdece ad trust ad thereby may wat to seek for better servces elsewhere. Hece, there s a eed to cotuously motor ad mprove the delvery of health care servces. Statstcal qualty cotrol s a techology wdely used maufacturg to mprove qualty of products ad productvty of workers [6]. It was developed 1924 by W. A. Shewhart. The Statstcal Process Cotrol (SPC) s a collecto of tools desged so that we ca cotuously mprove process performace ad reduce varablty. The SPC ts cepto was used maufacturg dustres but recet tmes, t s beg creasgly appled to servce dustres cludg health care. Examples of otable applcato of SPC to health care wth the ad of cotrol charts clude that of [7], [8], [9], [10], [12], [17] amog others. Correspodg Author: P. E. Ogutude 154

2 P. E. Ogutude, O. A. Odetumb, ad O. Ojo Oluwadare There are dfferet cotrol charts as avalable the lteratures, see [5], [14] ad [4]. [9] also provded cases where the four commoly used cotrol charts; c-chart, p-chart, u-chart ad xmr-charts (also kow as the dvdual charts) ca be used health care wth detaled examples. Selectg the approprate cotrol chart to be used s grossly depedet o the type of data set at had. A report by [2] gave a example of a msuse of cotrol charts health care where a author used the c-chart where c-chart s ot actually approprate to aalyze a gve data set (though, the correct cotrol chart to be used was ot stated). The report further metoed that the p-chart may probably ot be applcable because of a codto ot satsfed. Whlst stll gvg credt to [8] who aalyzed the data by calculatg the hazard rato ad treatg the hazard rato data as cotuous data before plottg a xmr chart based o the meda (ot mea) cotrol lmts. The same data set shall be used ths artcle usg the p-chart to kow the exact cocluso we ca arrve at comparso to that of [1], [2] ad [8]. 2 THE P-CHART The p-chart s a attrbute cotrol chart; hece, t s bascally used to motor the varato heret a dscrete attrbute data set. It plots the proporto of ocoformg uts a sample whch the sample szes must ecessarly ot be the same for the data set. The attrbute/characterstc of the data must be dchotomous ature ad t s usually descrbed the form of good/bad, dead/alve, defectve/o-defectve, e.t.c. The data set used by [1] ad [8] as data of surgeo specfc mortalty rates followg colorectal cacer surgery shall be used ths artcle. The data provdes the umber of colorectal cacer surgery cases atteded to by thrtee dfferet surgeos ad the umber of deaths recorded. The summary of the data s provded Table 1. Table 1: Surgeo specfc mortalty rates followg colorectal cacer surgery (Source: [8]) Surgeo Number of Cases Number of deaths A B 66 8 C 58 9 D 52 7 E F 46 5 G 38 3 H I 36 5 J 34 7 K 32 4 L 21 2 M 21 3 The am s to verfy whether varato the mortalty across the surgeos s cosstet wth commo cause of varato. We shall make use of the proporto of deaths recorded per surgeo. Followg Mohammed [9], the procedure for plottg the p-chart s gve below; Let Number of cases represet Number of deaths represet The Cetre Le (CL) s gve by; p = = 1 = 1 x 95 = (1) ISSN : Vol. 14 No. 2, Apr

3 A Comparatve Study of the Use of Statstcal Process Cotrol Motorg Health Care Delvery The Lower Cotrol Lmt (LCL) s calculated by; p 1 p LCL = p 3 (2) The Upper Cotrol Lmt (UCL) s calculated by; p 1 p UCL = p+ 3 (3) Accordg to [4], a set of decso rule was suggested [13] that a process s out of cotrol f ether 1. Oe pot plots outsde the three-sgma cotrol lmts. 2. Two out of three cosecutve pots plot beyod the two-sgma warg lmts. 3. Four out of fve cosecutve pots plot at a dstace of oe-sgma or beyod from the ceter le. 4. Eght cosecutve pots plot o oe sde of the ceter le. The result for the data Table 1 s gve below; Table 2: Table of Result Surgeo Number of Number of Proporto of LCL CL UCL Cases ( ) Death (x ) Deaths (x/) A B C D E F G H I J K L M ISSN : Vol. 14 No. 2, Apr

4 P. E. Ogutude, O. A. Odetumb, ad O. Ojo Oluwadare The correspodg p-chart s as show below; 0,35 0,3 Proporto of Deaths 0,25 0,2 0,15 0,1 0,05 Sample Pots UCL CL=0.16 LCL Surgeos Fg. 1: P-Chart for the Surgery Mortalty Rate The cotrol chart Fgure 1 shows that there are sample pots that fall outsde the cotrol lmts whch dcate that the process s out of cotrol. Hece, vestgatos ad correctve actos are requred to fd ad elmate the assgable cause or causes resposble for ths behavor. It s also good to ote that the cotrol lmts; LCL ad UCL are ot gve straght les (as show Fgure 1), ths s as a result of the dffereces the samples szes (as show Table 2). 2.1 LIMITATIONS OF THE RESULT PROVIDED 1. Accordg to [11], the approxmato used above s good as log as > 5 codto s ot fully satsfed by the data used, that, from Surgeos A to H, < 5 M, < 5. ad 0.1 p 0.9. The frst whle from Surgeos I to 2. Based o the argumets of [2], usg p-chart, we mght be volatg the thrd assumpto gve by [5] that Let p be the probablty that a tem has the attrbute; p must be same for all tems a sample (e.g., the probablty of a partcpat meetg or ot meetg the requremets s the same for all partcpats) 3 CONCLUSION The p-chart Fg. 1 shows that the process s out of cotrol ad there s a presece of assgable causes of varatos whch efforts must be made by the hosptal or health sttuto to kow the causes ad elmate as approprate. The result gve by [1] meas that the sample pot correspodg to the frst surgeo, Surgeo A dcates a out of cotrol stuato ad cocluded that the reasos for the hgh mortalty rate of Surgeo A compared to other surgeos should be vestgated by the cocered health sttuto. I cotrast to that, we shall rather coclude that based o our results ths artcle, together wth that of [2] ad [8]; Surgeo H has the hghest mortalty rate (takg to cosderato the umber of cases atteded to by each of the Surgeos ad ot just the umber of deaths recorded). The xmr cotrol chart gve by [8] shows that oly oe sample pot s outsde the upper cotrol lmt (.e Surgeo H) whle the p-chart ths artcle reveals that three sample pots (.e Surgeos E, H ad J) are outsde the upper cotrol lmt. [8] added that Surgeos wth hazard ratos more tha 1 are cosdered hazard to the patets, we shall add that Surgeos correspodg to sample pots outsde the cotrol lmts show Fg. 1 are hazardous to the patets. Besdes, they are hazardous to the health sttuto where they work for ISSN : Vol. 14 No. 2, Apr

5 A Comparatve Study of the Use of Statstcal Process Cotrol Motorg Health Care Delvery because the mage of a orgazato s mportat. Bad servces should ot be tolerated ad competet hads should be allowed to hadle cases. ACKNOWLEDGMENT We lke to apprecate the efforts of the aoymous revewers who took tme to thoroughly revew ad commet order to mprove the qualty of ths paper. REFERENCES [1] A. A. Olatude. O the Importace of Statstcal Process Cotrol Health Care. Research Joural of Medcal Sceces. 3 (2): 87-90, [2] BPI Cosultg, LLC Bullet (2009); [3] C. S. McArdle, D. Hole. Impact of varablty amog surgeos o postoperatve morbdty ad mortalty ad ultmate survval. BMJ 1991; 302: [4] D. C. Motgomery. Itroducto to Statstcal Qualty Cotrol, 6th Edto. New York, NY: Joh Wley & Sos, [5] D. J. Wheeler, D. S. Chambers. Uderstadg Statstcal Process Cotrol, 2d Edto. Koxvlle, TN: SPC Press, [6] G. B. Wetherll. Samplg Ispecto ad Qualty Cotrol. 2d Edto. New York: Joh Wley & Sos, Ic., [7] J. C. Beeya, R. C. Lloyd, P. E. Statstcal Process Cotrol as a Tool for Research ad Healthcare Improvemet. Qualty & Safety Health Care 2003; 12: [8] M. A. Mohammed. Usg Statstcal Process Cotrol to Improve the Qualty of Health Care. Qualty & Safety Health Care 2004; 13: [9] M. A. Mohammed, Worthgto P, Woodall WH. Plottg Basc Cotrol Charts: Tutoral Notes for Healthcare Practtoers. Qualty & Safety Health Care 2008; 17: [10] M. Daz, D. Neuhauser. Pasteur ad Parachutes: Whe Statstcal Process Cotrol s Better tha a Radomzed Cotrolled Tral. Qualty & Safety Health Care 2005; 14: [11] M. Xe, T. N. Goh, V. Kuralma. Statstcal Models ad Cotrol Charts for Hgh Qualty Processes. Kluwer Academc Publshers, [12] P. Adab, A. Rouse, M. A. Mohammed, Performace League Tables: The NHS Deserves Better. BMJ 2002; 324: [13] Statstcal Qualty Cotrol Hadbook, (1956) AT & T., (Wester Electrc), Idaapols, IN. [14] T. P. Rya. Statstcal Methods for Qualty Improvemet, Secod Edto. New York: Joh Wley & sos, [15] Uted States Bureau of Labor Statstcs (BLS). (2007). Occupatoal Outlook Hadbook. VA: BLS. [16] W. A. Shewhart. Ecoomc Cotrol of Qualty of Maufactured Product. New York, NY: D Va Nostrad Compay, 1931 (Reprted by ASQC Qualty Press, 1980). [17] W. H. Woodall. The Use of Cotrol Charts Health-Care ad Publc Health Survellace (Wth Dscusso). Joural of Qualty Techology 2006; 38: ISSN : Vol. 14 No. 2, Apr