How to Display Data- P16:The best method to convey a message from a piece of research in health is via a fi gure. The best advice that a statistician can give a researcher is to fi rst plot the data. Despite this, conventional statistics textbooks give only brief details on how to draw fi gures and display data. | Reporting study results 67 outcome and it is said to have four cells 2 X 2 see also Chapter 3 . The most appropriate comparative summary measure for thesedatais thedif-ference in proportions healed between the two groups. The important contrastis betweenthe 20 healed in the intervention group comuaerd tolO m alit control cuEnb-iiltsctiptieadt naturally from tah toriohi ititracoromendedthatthe datofor treatment groups is in tha column in odder that dihteeehcee between oeoupt caa 1st compared side by aide ot shoitrn hr Tabte . Another cdvanlage oU the fomiatiti Table seaoral outcomes we caoplacr ths iatafrethe Clffioedt autaomeotmbeenedOOeadh other in seporutorows. Ooe nxa ilOr Tedc the ulcer healing rates at 3 and 12 monihs. Tloi foemot. eviao thd .touts in rhe cola mat d eis o tioss-ldi elat oe o. treatment groupSin columnsfvt. ouitome in rows. uIcoo dealei tain ot healed at 12 weeks n 206 1 Cramp Intervention n 106 Control n 100 Outcome Healed 20 21 1 5 16 Not healed 1u nt1 84 84 Table Ulcer healing rates at3 and 12 months follow-upby treatment group maximum n 206 1 Group Difference in percea ta2 nsa 95 Cl P-valueb Relaive Ri nc 95 Cl Intnrnhntinn Co itrol Outcome Healed at 3 20 16 4 months 21 106 tl 6 1001 -7 to 14 Healed at 12 52 42 10 months a 42 81 Ot 9 -5 to 25 nu Cl Confidence interval. aA positive dia e e uit kiates that 0U intervenOu group doe11 Outer than the control group . bP-values from chi-squared test. CA relative riskol ieditatet tfiat the intervention group does better than the control group. 68 How to Display Data favoured by leading medical journals such as the British Medical Journal. Note that no decimalplacesarereported for the percentages of ulcers healed or the difference which makes the table clearer. The denominator is presentedfor alltheoutcomes and thusit isclearthat thesamplesize is lower for the h2-month ccmpsison. Also presunSed cs a column wilh
