Study of Quality of Care in Public and Private Sectors Research Symposium Quality of Healthcare in Sri Lanka Hosted by SLMA, CMASL and IHP 30 August 2013 Study Organization *Study Team – Isurujith K Liyanage, Janaki Jayanthan, Nilmini Wijemanne, Sanil de Alwis, Sarasi Amarasinghe, Indika Siriwardana, Shanti Dalpatadu, Prasanna Cooray, R P Rannan-Eliya Study Objective – To assess levels and differences in quality of care in public and private medical sectors in Sri Lanka Funding – World Bank contract to IHP, IHP Public Interest Research Fund Grant PIRF-2012-03, IDRC Grant 106439-003 Ethical Review – Ethical review and clearance of study design and survey instruments by IHP Ethical Review Committee (IHP ERC Approval Nos. 2012/006A, 2012/006B) 2 Outline • Study Organization • Background • Methodology • Findings • Conclusions 3 Background - Importance of measuring quality - Background on Sri Lanka 4 Why did we measure quality? • How to optimise the contribution of the private sector within Sri Lanka’s mixed health system? – Who they treat – Relative costs – Quality of care provided • Research questions – Does clinical quality differ between the public and private sectors in Sri Lanka – Does interpersonal quality differ between the public and private sectors in Sri Lanka 5 Concepts What is quality? • “The degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge.” – Institute of Medicine 1991 Dimensions of quality (Donabedian 1980) • Structure – Whether providers have correct inputs, equipment, training, etc. • Process – Whether good practices are followed • Outcomes – Impact of medical services on patients, including health outcomes and patient satisfaction 6 Measuring process quality • No widely accepted approach • Methods differ – Developing countries: • Single conditions • Little relevance to Sri Lanka (tuberculosis, HIV, malaria) – Developed countries • Broader methods (range of conditions) 7 Utilization of healthcare services CBSL CFS 2003/04 Inpatient Outpatient 4% 45% 55% 96% 8 Key indicators of hospitals in Colombo, Gampaha, Galle (2011) Inpatient spending Category Hospitals Beds Admissions per admission n n n rupees !"#$%&'()*+%,-$* ./ /01232 456'7%$$%)8 401/2. !9%:-,;'()*+%,-$* 63 /1<2= 05/'7%$$%)8 231643 Sources: Management Development and Planning Unit (2011) and Institute for Health Policy (2013) 9 Inpatient quality of care Comparison between public and private sectors Methodology - Study design - Tracer indicators & inclusion criteria - Sampling 11 Study design – inpatient care Overview • Study object – Process quality, i.e., what providers actually do • Approach – Retrospective review of inpatient medical records – Analysis of care using tracer conditions 12 Inpatient tracer conditions Criteria for selection • Conditions should be relatively frequent • Feasible quality indicators should exist with support in literature • Should be representative of a range of conditions and patient populations Tracer conditions (initial) 1. Acute Myocardial Infarction (AMI) (1% of discharges) 2. Acute Asthma (4% of discharges) 3. Childbirth (6% of discharges) 13 Selecting quality indicators • Identified possible quality indicators – Quality clearing houses – Quality assessment agencies – Studies: developed and developing countries – Clinical guidelines • Review by panel of doctors • Subsequent review of RAND quality assessment tool – Identification of additional tracer conditions and quality indicators – Choice of method to aggregate quality indicators 14 Inpatient Quality Indicators 55 Quality Indicators Resource Drug Condition Clinical area Intensity prescribing All drug Assessment and Asthma (7) Low (12) prescribing (36) investigations (10) Medium DVT AMI (15) prophylaxis Management (39) (33) (9) Childbirth Surgical High (4) antibiotics Outcome (6) (10) (4) Other (TIA, Dengue, COPD, Heart Failure, Pneumonia) 15 Inpatient quality indicators – examples Resource Indicator Condition Clinical area Drug prescribing intensity Neonatal APGAR score Childbirth Assessment / Low - recorded Investigations Prophylactic antibiotics Childbirth Management Medium All drug prescribing, given during LSCS surgical antibioitics Live discharge, AMI AMI Outcome - - AMI patient underwent PCI AMI Management High - / stenting Oxygen saturation Acute asthma Assessment / High - measured in acute asthma Investigations 16 Sampling Distribution of sampled facilities Sampled facilities Hospital type Colombo Gampaha Galle Total Public Large 2 1 1 4 Intermediate 2 1 1 4 Obstetric 1 0 1 2 Paediatric 1 0 0 1 Other specialist 0 0 0 0 Total 6 2 3 11 Private Large 3 1 0 4 Intermediate / small 1 2 2 5 Obstetric 1 0 0 1 Paediatric 0 0 0 0 Other specialist 0 0 0 0 Total 5 3 2 10 17 Gampaha Geographic distribution of sampled providers Galle Colombo 18 Distribution of private hospitals, Sri Lanka 2012 19 Data collection and processing Patient sampling • Systematic sample of patient records from 2011 discharges • Supplementary samples of tracer conditions Data collection • Data extraction and entry by pre-intern medical graduates using Apple iPads. Drug name entry using pre-coded listing of brand and generic names. Data analysis • Diagnoses coded to ICD-10 by physician. • All analysis using Stata 12.0. 20 Data collection and processing Quality and satisfaction scores Adapted method used by RAND quality study (McGlynn et al) • Quality instance = each opportunity a patient could potentially receive recommended care !"!#$%&'()*+%",%!-(*.%+*/"((*&0*0%/#+*% !""#$"%&$' 1#.%2-3*&%,"+%*#/4%5'#$-!6%-&.!#&/* 7 ()*#$ &'()*+%",%5'#$-!6%-&.!#&/*. 21 Data collection and processing Quality scores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indings - Inpatient quality of care 23 Characteristics of patient sample Weighted, standardized Standardized Characteristic Public Sector Private Sector (n = 2,523) (n = 1,815) p value Average age, years 36.8 37.0 1.0 Male sex, % 47.9 47.8 1.0 Discharge diagnoses Asthma, % 1.0 1.2 0.7 AMI, % 0.6 0.7 0.8 Childbirth, % 6.9 6.2 0.9 Average length of stay, days 3.6 3.0 0.1 24 Quality by condition 100 89 89 90 80 69† 68 68 70 63 60 50 40 30 Percentageof patients (%) 20 10 0 Asthma AMI Childbirth Public Private Resource Condition Drug prescribing Clinical area limitation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esource Condition Drug prescribing Clinical area limitation 26 Quality by resource limitation 100 94** 90 81 80 70 60 58* 50 48 40 31** 30 Percentageof patients (%) 20 10 10 0 Low Medium High Public Private Resource Condition Drug prescribing Clinical area limitation 27 Drug prescribing 80 68* 70 68* 60 60 58 50 40 30 20 Percentageof patients (%) 10 0 1 0 Drug indicators (all) DVT prophylaxis Perioperative antibiotics Public Private Resource Condition Drug prescribing Clinical area limitation 28 Difference in scores by clinical area !*"+,$- !"#$#%&"'&()& !""#""$#%&'(%)'*%+#"&*,(&*-%" ./#(&$#%&'(%)'$(%(,#$#%& 01&2-$#'*%)*2(&-/" 0+#/(33 4567' 4867' 67' 867' 567' 967' .(#/&0)''''''''''''''''''''''''''''''''.+1"#%' Resource Condition Drug prescribing Clinical area limitation 29 Quality scores by hospital size Is quality better in small/intermediate hospitals than large hospitals? !"#$#%&"'&()& '''''''*+,"#% '''''''''''''''''''''''''*(#7&/) -#.)&.)'%&/)01(2 !"#$%& !'( )$*+,-*.#$ 3).1+(%)'"#4#/&/#1$ /01 '2,*3% 4*5$ -(+0'#$5#%&/1(. 6.3578.2"9.*-*:5 6;<78.08$=+&>*" ?3.5*9&+7&:#*-*0#*978.2"9.*-*:5 !"#$#%&"'&()& !""2""%2:#7&:,7*:@2"#*5&#*0:" <.2&#%2:#7&:,7%&:&52%2:# A3#90%27*:,*9&#0." 67)(&"" BCDDE7 BFGE7 BGDE7 BHGE7 HGE7 GDE7 FGE7 CDDE7 BCDDE7 BFGE7 BGDE7 BHGE7 HGE7 GDE7 FGE7 CDDE7 -#8)()$%)'#$'.%1()'9(14'"&(0)':1.;#/&".' -#8)()$%)'#$'.%1()'9(14'"&(0)':1.;#/&".' 30 Quality scores by hospital size and resource limitation Difference in scores compared to large hospitals !"#$%&'()#%*+ ,+#"-./+)(%0%&'&%"1 )))))))))2-3(%/ ))))))))))))))))))))))))))))))))2.%4'&+ !"#$$%&%'()*+"*,-#)* ./0 1*,-2" 3-45 54+.'(( 6789% 6:79% :79% 789% 6789% 6:79% :79% 789% 6%7+.+1/+)%1)#/".+)8."0)('.9+):"#$%&'(#) 6%7+.+1/+)%1)#/".+)8."0)('.9+):"#$%&'(#) Resource Condition Drug prescribing Clinical area limitation 31 Conclusions 32 Key findings Public vs. private • Quality is fairly similar, although the public sector is slightly better • Private sector performs better in indicators that are resource intensive • Smaller hospitals tend to do