Inequity in access to guideline-recommended colorectal cancer treatment in Nova Scotia, Canada André Maddison MSc, Yukiko Asada PhD, Robin Urquhart PhD(c), Grace Johnston PhD, Fred Burge MD MSc CAHSPR Conference May 11, 2011
Colorectal cancer (CRC) care continuum CRC care Screening Diagnosis Preoperativoperative Surgery Post- therapy therapy Follow-up care
Colorectal cancer All physician and hospital services covered under the provincial funding plans However. Lower income associated with longer wait times (Winger et al 2008) Rural residency associated with longer wait times (Singh et al. 2010) Older individuals half as likely to receive treatment (Cree et al. 2009)
Observations from the current literature (1): Inequality vs. inequity Inequality: variations in access (every individual does not have identical access) Inequity: variations in access that are ethically problematic Identification of inequity Variations due to need factors: equitable Variations due to non-need factors, after adjustment for need factors: inequitable
Observations from the current literature (2): Inequality indices Yet to take advantage of inequality indices Concentration Index Increasingly used for population-level equity studies Makes comparison easy between studies, points of service, and over time
Study objectives 1. To clearly define inequity in access to CRC care by adjusting for clinical practice guidelines and patient need for care 2. To calculate the degrees of income-, age-, sex-, and distance-related inequities in access to chemotherapy and radiotherapy for patients diagnosed CRC in Nova Scotia
Data ACCESS Cohort Nova Scotia Cancer Registry identified all individuals diagnosed with CRC between 01/01/2001 31/12/2005 Linked to 14 administrative databases, including: Oncology Patient Information System Hospital Discharge Abstracts Physician billings Capital Health and Cape Breton palliative care data 2001 Census of Canada
Study population individuals who were diagnosed with: stage II or III rectal cancer or stage III colon cancer, and who underwent surgical resection of their tumour n=1094
Dependent variables 1. Receipt of chemotherapy and/or radiotherapy according to Canadian clinical practice guidelines (yes/no) Stage III colon cancer (Figueredo et al. 2003) Adjuvant chemotherapy Stage II/III rectal cancer (Figueredo et al. (2004) Adjuvant chemotherapy Neo-adjuvant or adjuvant radiotherapy 2. Wait time from radiation oncology referral to consultation within 14 days (yes/no)
Independent variables Need variables: Non-need variables Previous cancer diagnosis No, Yes Co-morbid conditions 0, 1, 2+ Income ($) <30,000, 30,000-44,999, 45,000+ Distance to the nearest cancer centre (Km) Sex 0-14.99, 15-74.99, 75-124.99, 125+ M, F Age at diagnosis (years) <60, 60-69, 70-79, 80+ Disease site Tumour stage Colon, rectal II, III Year of diagnosis 2001, 2002, 2003, 2004, 2005 Region of Nova Scotia (1) CDHA, (2) CBDHA, (3) Annapolis valley, S. Shore, and S.W. Shore, (4) Colchester East Hants, Pictou county and Guysborough Antigonish strait
Measure of inequity The Concentration Index A measure of variation Always measures variation of access by X (e.g., income) Index values between -1 and 1 Inequitable variations Variations after adjustment for need i.e., removing the influence of need
Analysis 1. Model observed use according to guidelines 2. Estimate need expected use according to guidelines 3. Calculate guideline-need-standardized use
1. Model observed use according to guidelines Run logistic regression models to identify factors that were associated with observed health care use according to clinical practice guidelines Observed use according to guidelines Factor 1 Factor 2 Factor 3
2. Estimate need expected use according to guidelines Observed use according to guidelines Factor 1 Factor 2 Factor 3 Distinguish factors into need and non-need Observed use according to guidelines Need factors Non-need factors Ethically justifiable or socially unproblematic Ethically unjustifiable
2. Estimate need-expected use according to guidelines Purge the influence of non-need factors, and predict needexpected use based only on need factors Need expected use according to guidelines Need factors Non-need factors Holding constant
3. Calculate guideline-needstandardized use Guideline-needstandardized use Observed use according to guidelines Need expected use according to guidelines Population mean
Key descriptive results Radiotherapy variables n (1094) % Disease site Colon 591 54.0 Rectum 504 46.0 Age at diagnosis <60 277 25.2 60-69 265 24.2 70-79 316 28.9 80+ 236 21.6 Sex Men 605 55.3 Women 489 44.7 Co-morbidities 0 670 61.2 1 207 18.9 2+ 217 19.9
Key descriptive results Radiotherapy variables n (1094) % Disease site Colon 591 54.0 Rectum 504 46.0 Age at diagnosis <60 277 25.2 60-69 265 24.2 70-79 316 28.9 80+ 236 21.6 Sex Men 605 55.3 Women 489 44.7 Co-morbidities 0 670 61.2 1 207 18.9 2+ 217 19.9
Key descriptive results Radiotherapy variables n (1094) % Disease site Colon 591 54.0 Rectum 504 46.0 Age at diagnosis <60 277 25.2 60-69 265 24.2 70-79 316 28.9 80+ 236 21.6 Sex Men 605 55.3 Women 489 44.7 Co-morbidities 0 670 61.2 1 207 18.9 2+ 217 19.9
Key descriptive results Radiotherapy variables n (1094) % Disease site Colon 591 54.0 Rectum 504 46.0 Age at diagnosis <60 277 25.2 60-69 265 24.2 70-79 316 28.9 80+ 236 21.6 Sex Men 605 55.3 Women 489 44.7 Co-morbidities 0 670 61.2 1 207 18.9 2+ 217 19.9
Key descriptive results Radiotherapy variables n (1094) % Disease site Colon 591 54.0 Rectum 504 46.0 Age at diagnosis <60 277 25.2 60-69 265 24.2 70-79 316 28.9 80+ 236 21.6 Sex Men 605 55.3 Women 489 44.7 Co-morbidities 0 670 61.2 1 207 18.9 2+ 217 19.9
Results Key adjusted Odds Ratios (OR) Use according to guidelines Receipt of clinically recommended treatment Wait time from radiation oncology referral to consultation within 14 days OR (95% CI) OR (95% CI) Age at diagnosis <60 1.0 1.0 60-69 0.64 (0.42, 0.97) 70-79 0.20 (0.13, 0.30) 80+ 0.022 (0.01, 0.04) Sex Male 1.0 1.0 Distance to the nearest cancer centre (KMs) Female 0.61 (0.45, 0.83) 0-14.99 1.00 15-74.99 0.74 (0.47, 1.16) 75-124.99 0.63 (0.34, 1.18) 125+ 0.46 (0.24, 0.89) 0.95 (0.36, 2.51) 0.72 (0.27, 1.87) 0.92 (0.36, 2.36) 0.91 (0.55, 1.50) 1.00 1.14 (0.57, 2.27) 0.97 (0.36, 2.59) 0.89 (0.30, 1.90)
Results Key adjusted Odds ratios (OR) Use according to guidelines Receipt of clinically recommended treatment Wait time from radiation oncology referral to consultation within 14 days OR (95% CI) OR (95% CI) Age at diagnosis <60 1.0 1.0 60-69 0.64 (0.42, 0.97) 70-79 0.20 (0.13, 0.30) 80+ 0.022 (0.01, 0.04) Sex Male 1.0 1.0 Distance to the nearest cancer centre (KMs) Female 0.61 (0.45, 0.83) 0-14.99 1.00 15-74.99 0.74 (0.47, 1.16) 75-124.99 0.63 (0.34, 1.18) 125+ 0.46 (0.24, 0.89) 0.95 (0.36, 2.51) 0.72 (0.27, 1.87) 0.92 (0.36, 2.36) 0.91 (0.55, 1.50) 1.00 1.14 (0.57, 2.27) 0.97 (0.36, 2.59) 0.89 (0.30, 1.90)
Results Key adjusted Odds ratios (OR) Use according to guidelines Receipt of clinically recommended treatment Wait time from radiation oncology referral to consultation within 14 days OR (95% CI) OR (95% CI) Age at diagnosis <60 1.0 1.0 60-69 0.64 (0.42, 0.97) 70-79 0.20 (0.13, 0.30) 80+ 0.022 (0.01, 0.04) Sex Male 1.0 1.0 Distance to the nearest cancer centre (KMs) Female 0.61 (0.45, 0.83) 0-14.99 1.00 15-74.99 0.74 (0.47, 1.16) 75-124.99 0.63 (0.34, 1.18) 125+ 0.46 (0.24, 0.89) 0.95 (0.36, 2.51) 0.72 (0.27, 1.87) 0.92 (0.36, 2.36) 0.91 (0.55, 1.50) 1.00 1.14 (0.57, 2.27) 0.97 (0.36, 2.59) 0.89 (0.30, 1.90)
Results Key adjusted Odds ratios (OR) Use according to guidelines Receipt of clinically recommended treatment Wait time from radiation oncology referral to consultation within 14 days OR (95% CI) OR (95% CI) Age at diagnosis <60 1.0 1.0 60-69 0.64 (0.42, 0.97) 70-79 0.20 (0.13, 0.30) 80+ 0.022 (0.01, 0.04) Sex Male 1.0 1.0 Distance to the nearest cancer centre (KMs) Female 0.61 (0.45, 0.83) 0-14.99 1.00 15-74.99 0.74 (0.47, 1.16) 75-124.99 0.63 (0.34, 1.18) 125+ 0.46 (0.24, 0.89) 0.95 (0.36, 2.51) 0.72 (0.27, 1.87) 0.92 (0.36, 2.36) 0.91 (0.55, 1.50) 1.00 1.14 (0.57, 2.27) 0.97 (0.36, 2.59) 0.89 (0.30, 1.90)
Inequity in access to CRC care 0.25 Receipt of clinically recommended treatment Wait times Specialist use Allin S, 2008 Concentration Index 0.2 0.15 0.1 0.05 Pro-advantaged Pro-rich Pro-young Pro-male Pro-close 0-0.05 Income Age Sex Distance Pro-disadvantaged
Inequity in access to CRC care 0.25 Receipt of clinically recommended treatment Wait times Specialist use Allin S, 2008 Concentration Index 0.2 0.15 0.1 0.05 Pro-advantaged Pro-rich Pro-young Pro-male Pro-close 0-0.05 Income Age Sex Distance Pro-disadvantaged
Inequity in access to CRC care 0.25 Receipt of clinically recommended treatment Wait times Specialist use Allin S, 2008 Concentration Index 0.2 0.15 0.1 0.05 Pro-advantaged Pro-rich Pro-young Pro-male Pro-close 0-0.05 Income Age Sex Distance Pro-disadvantaged
Inequity in access to CRC care 0.25 Receipt of clinically recommended treatment Wait times Specialist use Allin S, 2008 Concentration Index 0.2 0.15 0.1 0.05 Pro-advantaged Pro-rich Pro-young Pro-male Pro-close 0-0.05 Income Age Sex Distance Pro-disadvantaged
Inequity in access to CRC care 0.25 Receipt of clinically recommended treatment Wait times Specialist use Allin S, 2008 Concentration Index 0.2 0.15 0.1 0.05 Pro-advantaged Pro-rich Pro-young Pro-male Pro-close 0-0.05 Income Age Sex Distance Pro-disadvantaged
Summary of results Indications of pro-young, pro-male, and pro-close distance inequity in receipt of treatment for CRC in Nova Scotia No statistically significant inequity in wait times for services based on income, age, sex, or distance Age-related inequity is of greatest concern
Strengths Limitations Application of the need-standardization approach for clinical data and its expansion with the use of clinical practice guidelines Tumour stage and comorbidity data to approximate patient need for care Need vs. Non-need factors No gold standard Crude co-morbidity measure Inability to capture patient choice
Take home messages For cancer researchers It is important to clearly distinguish inequity from inequality The use of need-standardization methods will enhance comparability of results For policy-makers Even with universal health insurance and clinical guidelines, some CRC care may be inequitable
Contact information André Maddison MSc, MD Candidate 2014 Faculty of Medicine, Dalhousie University andre.maddison@dal.ca
Inequity in access to CRC care Receipt of clinically recommended treatment Radiation oncology wait times Equity stratifier Concentration index (95% Confidence interval) Income 0.01 (-0.02, 0.04) 0.00 (-0.06, 0.06) Age 0.22 (0.19, 0.25) -0.01 (-0.07, 0.05) Sex 0.05 (0.02, 0.08) 0.02 (-0.04, 0.08) Distance 0.05 (0.02, 0.09) 0.05 (-0.01, 0.11)
Ease of comparability by Inequality index Example: Inequality in access to care Study 1 Odds ratio Study 2 Odds ratio Poor 1.00 $ <25,000 1.00 Middle 1.2 $ 25,000-49,999 1.3 Rich 2.0 $ 50,000-74,999 1.5 Income-related inequality Concentration Index $ 75,000-99,999 1.9 $ >100,000 2.1 0.1 0.3 Concentration Index
Concentration Index % share health care acces SES SES
Equity stratifiers and interpretation of the Concentration Index Equity stratifier Better access for: Disadvantaged Advantaged -1 Concentration Index <0 0> Concentration Index 1 Income Poor Rich Age Old Young Sex Women Men Distance Far Close Inequity