Journal Logo

Research Paper

Association of chronic pain with comorbidities and health care utilization: a retrospective cohort study using health administrative data

Foley, Heather E.a,b,*; Knight, John C.b,c,d; Ploughman, Michellee; Asghari, Shabnamc,f; Audas, Richardb

Author Information
doi: 10.1097/j.pain.0000000000002264

Abstract

Using current methodology in Canadian health administrative data, it was observed that having chronic pain was significantly associated with multimorbidity and higher health services utilization.

1. Introduction

Effective treatment for chronic pain is historically elusive.12,32 Although research and treatment techniques continue to evolve,19,114 most chronic pain conditions remain poorly understood and managed with negative impact on all facets of a sufferer's life.67,105,122 Prolonged moderate to severe pain heightens a person's desperation for relief resulting in increased utilization of various and increasingly specialized health services.21,38 Overall annual direct health care costs for chronic pain were estimated up to €32 billion in Europe,39,89 $261 to 300 billion in the United States,31 and CAD$15.1 to 17.2 billion in Canada.43 Indirect annual societal costs are even higher estimated at $299 to 335 billion in the United States31 and CAD$23.2 billion in Canada.43 Attempts to effectively treat chronic pain contributed to the unintended public health and societal fallout from opioid diversion and problematic opioid use,36 with 2014 Canadian costs estimated at CAD$313.1 million in health care and CAD$1.83 billion in lost productivity.18,33,75 Mitigating the devastating cost of chronic pain to individuals, families, communities, and society at large is imperative.

Multiple risk factors are postulated to influence the overall cost escalation of chronic pain.2,49,64 From an epidemiological perspective, chronic pain prevalence globally (2%-54%15,39,112,118) and in Canada (6.5%-44%11,23,92,98,110) represents a high volume of people impacted. From a behavioral perspective,2 overall cost amplification occurs when this high prevalence is multiplied by the high utilization of health care resources (eg, hospital admissions, emergency department visits, and health practitioner encounters) by people with chronic pain.31,39,44,58,113 From a needs perspective,2 people with chronic pain are consistently measured to have higher chronic comorbidity prevalence (eg, mood disorder, cardiovascular disease, and chronic pulmonary disease41,45,115) and higher odds of multimorbidity65 increasing the likelihood of frequent physician consultations.117 These 3 cost drivers (high prevalence, high per person health care utilization, and multimorbidity) represent measurable indicators for evaluating any public health policy or health care practice change.

Chronic pain advisory groups were struck in Canada in 2019 to survey how pain is being managed and provide recommendations for service improvement.42 An important first step is determining baseline epidemiological and health care utilization statistics.13,23,116 Other jurisdictions achieved this by analyzing data sources such as surveys,31,107,113 electronic medical records,97 and administrative data sets.39,44,93,94,96 Health administrative data, an economical data source with wide coverage, present considerable potential to identify people with complex conditions such as chronic pain, determine epidemiological distribution, and quantify health care utilization in any jurisdiction.7,54 Consequently, an algorithm to identify chronic pain cases in provincial health administrative data was validated30 and used to calculate provincial incidence and prevalence (Foley HE, Knight JC, Ploughman M, Asghari S, Audas R, unpublished data, November 2020). This study sought to characterize and compare chronic pain–identified members to non-chronic pain–identified members of the provincial cohort with respect to comorbidity prevalence and 2009/2010 publicly funded health care utilization. The study hypothesized that being identified as having chronic pain would be (1) strongly associated with being identified with other comorbid conditions and (2) strongly associated with having higher publicly funded health care resource utilization.

2. Methodology

2.1. Design, setting, and population cohort

A retrospective cohort study design using health administrative data was performed in the province of Newfoundland and Labrador (NL), Canada, which had a population of 516,729 in 2009.81 All residents identified as eligible for Medical Care Plan (MCP) benefits for the 2009/2010 fiscal year (April-March) were included in the provincial cohort, comprising approximately 98% of the NL population for that year (Canadian Armed Forces personnel, Royal Canadian Mounted Police members, and international students were ineligible for benefits and, therefore, excluded).99 Provincial cohort follow-up was based on MCP eligibility status that is released once each fiscal year, rather than birth/death or migration. Thus, physician visits and hospital admissions for each provincial cohort member were followed for 1 fiscal year from April 1, 2009, to March 31, 2010, with no assumed loss to follow-up.

2.2. Health administrative data sources

Each province and territory in Canada administers universal health plans that cover most hospital and physician services to nearly all of their residents.16 Fee-for-service is one of 2 methods typically used by provincial/territorial governments to remunerate for physician services (the other is alternate payment plan, such as salary, that does not have record-level information systematically captured in NL).17 The health administrative data generated from fee-for-service physician claims and hospital discharge abstracts are used to extract annual population-based estimates on distribution, trends, and direct health care costs of various medical conditions in Canada.88 The 3 administrative data sources used in this study were previously described (summarized in Table S1, supplementary digital content, available at https://links.lww.com/PAIN/B331).1,30 The data from all 3 data sets were organized through each resident's unique health insurance number.1,78 The data sources were (1) the MCP Fee-for-Service Physicians Claims Database File to identify cases of chronic pain, identify cases of comorbid conditions, and determine the number and type of physician service visits per person; (2) the Provincial Discharge Abstract Database (PDAD), the NL component of the Canadian Institute of Health Information national Discharge Abstracts Database, to identify cases of comorbid conditions and to determine admissions per person (number, type, and most responsible reason); and (3) the MCP Beneficiary Registration Database to determine benefits eligibility and demographics of the provincial cohort. All required record-level administrative data from January 1, 1999, to March 31, 2010 (the latest available data at the time of study initiation), were obtained from these data sets.

The MCP Claims File and PDAD data are regularly used for research and surveillance of multiple injuries and disease states.88 Data in the MCP Claims File are considered complete because of its collection for service remuneration.77 Rigorous quality control procedures are applied to the PDAD on an annual basis.53,76,77 The MCP Beneficiary Database has minimal missing data (missing residential data on 0.1% of the cohort members and missing age/sex data on 1 cohort member) because of regular checks made by the administrators that ensure completeness and accuracy of information.79

2.3. Data linkage

The MCP Claims File, the PDAD, and the MCP Beneficiary Database are held at the NL Centre for Health Information. The health insurance numbers (MCP numbers) of the provincial population cohort were linked to the MCP Beneficiary File, the PDAD, and the MCP Claims File. Analysts at the NL Centre for Health Information performed all data extraction, linkage, cleaning, and deidentification before provision of the linked data set to the research team for analysis.

2.4. Chronic pain case identification

The exposure of interest in this study was the presence of chronic pain as defined by a validated health administrative data algorithm (Chronic Pain Algorithm) applied to the MCP Claims File for the 1999 to 2009 fiscal years.30 The Chronic Pain Algorithm identifies chronic pain cases from residents attending fee-for-service physician encounters for pain-related conditions in NL. Development and validation of the Chronic Pain Algorithm were previously described and had 70.3% sensitivity, 66.8% specificity, 40.8% positive predictive value, and 87.4% negative predictive value when validated against a primary care electronic medical records data audit of an NL general population sample.30 Assessing the strength of association between NL health administrative data-derived variables and the presence of chronic pain as defined by the Chronic Pain Algorithm was considered an appropriate use of the Algorithm. The Chronic Pain Algorithm was defined as (1) a single claim date with an anesthesiologist recording a chronic pain–related provincial MCP procedure billing code (Table S2, supplementary digital content, available at https://links.lww.com/PAIN/B331) OR (2) 5 or more claim dates with any physician recording any pain-related diagnostic code (Table S3, supplementary digital content, available at https://links.lww.com/PAIN/B331) in a 5-year period with more than 183 days separating at least 2 pain-related claim dates.30

Since a prevalence-based approach was used in this study,26,44 all provincial cohort members identified by the Chronic Pain Algorithm from 1999 to 2009 was counted as a chronic pain case and formed the Chronic Pain Group (CPG). All members of the provincial cohort not identified by the Chronic Pain Algorithm from 1999 to 2009 formed the No Chronic Pain Group (NCPG).

2.5. Independent variables

2.5.1. Demographics

Information regarding sex, age, regional health authority, and rural/urban residential classification was obtained for the 2009/2010 fiscal year from the MCP Beneficiary file. Sex classification (male/female) was determined by the cohort member's registration in the MCP Beneficiary file. Age as of September 1, 2009, was classified into 4 categorical age groups representing children and youth (0-24 years), young adults (25-44 years), older adults (45-64 years), and seniors (65 years and older). There were 4 health authority regions of residence, and the Department of Health and Community Services of the NL Government defined the catchment area for each. They were the Eastern, Central, Western, and Labrador-Grenfell Regional Health Authorities. For the purposes of this study, individuals were considered to have an urban residential location if their community of residence had a population of 4000 or more people in the 2011 Statistics Canada Census, while those from communities with less than 4000 people were considered rural. A cutoff of 4000 was used because it better represented community level access to health services in NL compared with the Statistics Canada population centre cutoff of 1000 or census agglomeration cutoff of 10,000.24,103

2.5.2. Comorbid conditions

The presence of mental illness, mood and anxiety disorders, hypertension, diabetes, ischemic heart disease, chronic obstructive pulmonary disease, asthma, heart failure, acute myocardial infarction, stroke, hip fracture, epilepsy, dementia, Parkinsonism, and multiple sclerosis was determined by applying the Canadian Chronic Disease Surveillance System (CCDSS) administrative data case definitions to the cohort's 1999 to 2009 MCP Claims File and PDAD data (summarized in Table S4, supplementary digital content, available at https://links.lww.com/PAIN/B331).87 These 15 case definitions are coding algorithms used by the Public Health Agency of Canada to provide annual federal and provincial/territorial estimates on chronic disease distribution for health care resource and policy planning.88 These chronic disease case definitions were chosen for this study because, at the time of data analysis initiation, they were 15 of 22 case definitions validated for use in claims and discharge abstract data in the Canadian provinces (including NL) that were known to be associated with chronic pain and had diagnostic codes not included in the chronic pain case definition.60,88 The CCDSS case definitions for mental illness and mood and anxiety disorders annually identify individuals who used health services for but are not necessarily diagnosed with these mental health conditions.4 Comorbid condition case status for mental illness and mood and anxiety disorders for the purposes of this study was defined as any provincial cohort member identified by the corresponding CCDSS case definition within the 1999 to 2009 data period. The presence of cancer as another comorbid condition related to chronic pain was determined by the recording of 1 or more entries in any 1 or more of the 16 allowable diagnostic codes per admission in the PDAD of 1 or more cancer International Classification of Disease—10th Revision (Canadian) (ICD-10-CA) diagnostic codes124 published by the Public Health Agency of Canada.

Since all Public Health Agency of Canada codes and coding algorithms are published on the public domain of the Government of Canada website,87,124 permission was not required to use them for noncommercial activities as long as the Public Health Agency of Canada source was cited (personal communication). Comorbid case status was determined before the health care utilization observation period of the 2009/2010 fiscal year to evaluate its impact on health care utilization. Since a prevalence-based approach was used for this study,26,44 all cohort members identified by the CCDSS case definitions from 1999 to 2009 were counted as a comorbid condition case.

2.6. Dependent variables

2.6.1. Physician claims–related health care utilization

The MCP Claims File data for the provincial cohort were searched for all visits made from April 1, 2009, to March 31, 2010. For the purposes of this study, a visit was defined as any assessment, intervention, or procedure billed to the MCP by a fee-for-service physician in NL. Frequency of visits by physician type was captured based on the physician specialty code and was classified as (1) family physician only and (2) specialist only (those not identified as a family physician). Frequency of visits by reason for encounter was captured based on the associated International Classification of Disease—9th Revision (ICD-9) diagnostic code and was classified as (1) all-cause, and (2) pain-related (Table S3, supplementary digital content, available at https://links.lww.com/PAIN/B331). The diagnostic codes used to classify a visit as pain- vs non-pain-related were validated as part of the Chronic Pain Algorithm.30 Frequency of diagnostic imaging visits was captured based on MCP Provincial Procedure Billing Codes (Table S5, supplementary digital content, available at https://links.lww.com/PAIN/B331) and was classified as (1) general radiograph, (2) computed tomography scan, and (3) magnetic resonance imaging scan.

2.6.2. Hospital admission–related health care utilization

The PDAD for the provincial cohort was searched for all hospital admissions occurring from April 1, 2009, to March 31, 2010. Frequency of admissions by service type was captured and was classified as (1) day surgery admission (defined as the admission and discharge dates being the same day indicating no overnight stay) and (2) inpatient admission (defined as the admission and discharge dates not being on the same day indicating an overnight stay of at least 1 night). Frequency of admissions by most responsible diagnosis (ie, reason for admission) was captured based on the associated ICD-9 (up to March 31, 2001) or ICD-10-CA (April 1, 2001, and later) diagnostic code and was classified as (1) all-cause or (2) pain-related (Table S3, supplementary digital content, available at https://links.lww.com/PAIN/B331). The diagnostic codes used to classify a visit as pain- vs non-pain-related were validated as part of the Chronic Pain Algorithm.30

2.7. Analysis

The 2009/2010 fiscal year characteristics for the CPG and the NCPG were described by calculating mean and SD for age, and frequency and percentage for age group, sex, regional health authority, and rural/urban residential location strata. Frequency and percentage for 7 pain condition categories (other conditions associated with chronic pain, arthritis and musculoskeletal pain, back and neck pain, headaches, musculoskeletal trauma and related conditions, painful neuropathy, and bone disorders) were calculated for the CPG and the NCPG. Pain condition case status was determined by an individual having at least 1 visit in the MCP Claims File data or at least 1 hospitalization in the PDAD (all 16 allowable diagnostic codes per admission considered) with a diagnosis from the pain condition diagnostic code grouping (Table S3, supplementary digital content, available at https://links.lww.com/PAIN/B331) from 1999 to 2010. Cohort members could be counted as a case for multiple pain condition categories. The diagnostic code groupings used to identify a case in each pain condition category were informed by the literature,5,8,10,35,39,40,55–57,63,86,100,109,125 used for descriptive purposes only, and comprise pain-related diagnostic codes validated as part of the Chronic Pain Algorithm.30 However, the diagnostic code groupings did not undergo a specific validation process. Between-group differences were tested using the t test for mean age and the chi-squared test for categorical variable proportions (statistical significance defined as P < 0.05).

Prevalence of each comorbid condition in the CPG and the NCPG was calculated. Prevalence of being in the CPG for comorbid condition cases and noncases was calculated. Since chronic pain and comorbid condition case status was determined by March 31, 2009, the unadjusted odds ratio (95% confidence interval [CI]) between the CPG and the NCPG was calculated and reported for each comorbid condition. Each odds ratio was adjusted for the covariates of sex, regional health authority, and rural/urban residential location using a logistic regression model. Age group was not included as a covariate in the logistic regression because of collinearity with the case status of several comorbid conditions (age was an algorithm inclusion/exclusion criterion for several comorbid conditions. See Table S4, supplementary digital content, available at https://links.lww.com/PAIN/B331). The number of comorbid conditions was classified into groupings of 0, 1, 2, or 3 or more. Since the codes for the use of health services for mood and anxiety disorders coding algorithm were included in the codes for the use of health services for mental illness coding algorithm, mood and anxiety disorders were not counted as a separate comorbid condition for the purpose of these groupings.4,87 Odds of having 1, 2, or 3 or more comorbid conditions and the unadjusted odds ratio (95% CI) between the CPG and the NCPG were calculated and reported. Each odds ratio was adjusted for the covariates of sex, regional health authority, and rural/urban residential location using a logistic regression model.

Risks of having a physician visit, a diagnostic imaging visit, and a hospital admission and the unadjusted relative risk ratio (95% CI) between the CPG and the NCPG were calculated and reported. The dependent variable was binary (yes/no); therefore, each relative risk ratio was adjusted for the covariates of age group, sex, regional health authority, rural/urban residential location, and comorbid condition grouping using a robust Poisson regression model with log link function of the generalized linear model family (determined to be superior to the log binomial regression model in providing unbiased estimates of relative risk ratios).20 Statistical significance was defined by P < 0.05 that being in the CPG was predictive of risk of using each health service as measured by the relative risk ratio while controlling for the measured covariates.

The mean rates of physician visits, diagnostic imaging visits, and hospital admissions per 100 person-years and the unadjusted rate ratio (95% CI) between the CPG and the NCPG in 2009/2010 were calculated and reported. Each rate ratio was adjusted for the covariates of age group, sex, regional health authority, rural/urban residential location, and comorbid condition grouping using a negative binomial regression model of the generalized linear model family. It was considered the most parsimonious model with good performance when assessed for predicted vs observed probabilities of each dependent variable count value.62,84 The negative binomial model performed superior to the Poisson and zero-inflated Poisson models and closely comparable/superior to the zero-inflated negative binomial model. Statistical significance was defined by P < 0.05 that being in the CPG was predictive of the mean annual rates of visits and admissions per 100 person-years as measured by the rate ratio while controlling for the measured covariates.

The categorical covariates for the regression analyses were age group (reference category: 0-24 years), sex (reference category: male), regional health authority (reference category: Eastern Regional Health Authority), rural/urban residential location (reference category: urban), and comorbid condition grouping (reference category: 0 comorbid conditions). Cohort members with missing age, sex, regional health authority, or rural/urban category data were omitted from the regression models.

Statistical Package for Social Sciences version 25 by IBM and StataIC16 by StataCorp were used for the data analysis.

2.8. Ethics approval and consent to participate

The Health Research Ethics Board of the Health Research Ethics Authority of NL provided full approval of the study protocol (HREB Ref#2017.273). The Secondary Uses Committee of the NL Centre for Health Information and the Research Proposals Approval Committee of the Eastern Regional Health Authority also reviewed and approved the study protocol after Health Research Ethics Board approval. Since the health administrative data analyzed were part of routine data collection and normal operations of the NL Centre for Health Information, and the data were then deidentified, individual patient and/or NL resident consent was not required.

3. Results

3.1. Provincial cohort characteristics

The provincial cohort comprised 504,693 people (or 97.7% of the 2009 Census Canada reported NL population81) with a mean (SD) age of 42.4 (21.1) years, and of which 50.9% were female, 24.2% were 24 years or younger, 15.4% were 65 years or older, 55.1% lived in urban locations, and 58.9% lived in the Eastern Regional Health Authority catchment area. With respect to health care utilization, each member of the provincial cohort was followed from April 1, 2009, to March 31, 2010, for a total of 504,693 person-years and no assumed loss to follow-up. The CPG comprised 184,580 people or 36.6% of the provincial cohort (Table 1). Proportions in the female, 45 years and older age groups, urban residential location, Eastern Regional Health Authority residential location, and each pain condition group stratum were all significantly higher in the CPG compared with the NCPG (P-value < 0.001 for all comparisons). The mean age was significantly higher at 50.8 years (18.1 years SD) in the CPG vs 37.5 years (21.1 years SD) in the NCPG (P-value < 0.001). Increasing proportions with increasing age in both sexes was evident when prevalent chronic pain proportions in the provincial cohort was further stratified by age and sex (Table S6, supplementary digital content, available at https://links.lww.com/PAIN/B331). The proportions of prevalent chronic pain cases in females vs males diverged at the 18 to 24 years' age group (with higher proportions in females) that continued with increasing age.

Table 1 - 2009/2010 characteristics of provincial cohort in Newfoundland and Labrador, Canada (N = 504,693).
Demographic characteristics Chronic Pain Group* No Chronic Pain Group* P
Ngroup = 184,580 Ngroup = 320,113
N (% of Ngroup) N (% of Ngroup)
Age group
 0-24 y 17,353 (9.4) 104,433 (32.6) <0.001
 25-44 y 46,937 (25.4) 93,326 (29.2) <0.001
 45-64 y 78,930 (42.8) 86,112 (26.9) <0.001
 65 years and older 41,360 (22.4) 36,241 (11.3) <0.001
 Missing category data 1
Sex
 Female 110,024 (59.6) 146,742 (45.8) <0.001
 Male 74,556 (40.4) 173,370 (54.2) <0.001
 Missing category data 1
Rural/urban
 Urban 109,202 (59.2) 168,720 (52.7) <0.001
 Rural 75,258 (40.8) 151,193 (47.2) <0.001
 Missing category data 120 (0.1) 200 (0.1)
Regional health authority§
 Eastern 122,433 (66.4) 174,681 (54.6) <0.001
 Central 33,033 (17.9) 60,766 (19.0) <0.001
 Western 25,701 (13.9) 53,429 (16.7) <0.001
 Labrador-Grenfell 3293 (1.8) 31,037 (9.7) <0.001
 Missing category data 120 (0.1) 200 (0.1)
Pain condition category
 Other conditions associated with chronic pain 165,674 (89.8) 138,976 (43.4) <0.001
 Musculoskeletal/arthritis 147,211 (79.8) 89,511 (28.0) <0.001
 Back/neck 124,385 (67.4) 62,762 (19.6) <0.001
 Headaches 94,183 (51.0) 54,618 (17.1) <0.001
 Musculoskeletal trauma 56,633 (30.7) 31,276 (9.8) <0.001
 Neuropathic 50,496 (27.4) 22,323 (7.0) <0.001
 Bone disorders 21,885 (11.9) 7535 (2.4) <0.001
*Selection by the Chronic Pain Algorithm applied to 1999 to 2009 provincial cohort NL MCP Fee-for-Service Physician Claims File data determined Chronic Pain Group or No Chronic Pain Group membership. The Chronic Pain Algorithm was defined as (1) a single encounter date with an anesthesiologist recording a chronic pain-related provincial MCP procedure billing code (Table S2, supplementary digital content, available at https://links.lww.com/PAIN/B331) OR (2) 5 or more encounter dates with any physician recording any pain-related diagnostic code (Table S3, supplementary digital content, available at https://links.lww.com/PAIN/B331) in a 5-year period with more than 183 days separating at least 2 pain-related encounter dates.
Statistical significance was defined as P < 0.05 through the χ2 test.
Urban residential location was defined by the community of residence having a population of 4000 or more people in the 2011 Statistics Canada Census, while communities with less than 4000 people were considered rural.
§Regional health authority residential classification was defined by the community of residence being in 1 of 4 of the NL Department of Health and Community Services-defined regions.
Inclusion in the pain condition group was defined as an individual having at least 1 encounter in the MCP Claims File data or at least 1 admission in the Provincial Discharge Abstract Data (any one of the 16 allowable diagnostic codes per admission) recording a diagnosis from the pain condition diagnostic code grouping (Table S3, supplementary digital content, available at https://links.lww.com/PAIN/B331) from 1999 to 2010 (cohort members could be counted as a case for more than 1 pain condition category).
MCP, Medical Care Plan; N, total population of group; n, number selected in stratum; NL, Newfoundland and Labrador.

3.2. Comorbid conditions

The prevalence of each comorbid condition as identified at any time from 1999 to 2009 by the CCDSS case definitions was significantly higher in the CPG than in the NCPG (Fig. 1). The prevalence of being in the CPG was significantly higher for cases compared with noncases of each comorbid condition (Fig. 2). The odds of having each comorbid condition were significantly higher for the CPG compared with the NCPG as determined by the adjusted odds ratio (95% CI) (Table 2). The adjusted odds ratio ranged from 1.40 (95% CI: 1.36-1.43) to 4.27 (95% CI: 3.55-5.14).

Figure 1.
Figure 1.:
Comorbid condition prevalence by chronic pain case status in Newfoundland and Labrador, Canada. X-axis: estimated prevalence of each comorbid condition in the Chronic Pain Group and No Chronic Pain Group (as determined by the Chronic Pain Algorithm). Y-axis: comorbid conditions (as determined by the Canadian Chronic Disease Surveillance System case definitions).
Figure 2.
Figure 2.:
Chronic Pain Group proportion by comorbid condition case status in Newfoundland and Labrador, Canada. X-axis: estimated percentage of the cases and noncases of each comorbid condition that was identified by the Chronic Pain Algorithm. Y-axis: comorbid conditions (as determined by the Canadian Chronic Disease Surveillance System case definitions).
Table 2 - Association between comorbid conditions* and chronic pain in Newfoundland and Labrador, Canada.
Comorbid condition Unadjusted odds ratio (95% CI) Adjusted odds ratio (95% CI) P§
All mental illness 4.05 (4.00-4.10) 3.62 (3.58-3.67) <0.001
Mood/anxiety disorders 3.96 (3.91-4.02) 3.51 (3.46-3.56) <0.001
Hypertension 3.40 (3.36-3.45) 3.25 (3.21-3.30) <0.001
Diabetes 2.31 (2.26-2.35) 2.35 (2.30-2.40) <0.001
Ischeamic heart disease 2.89 (2.83-2.96) 3.20 (3.12-3.29) <0.001
Chronic obstructive pulmonary disease 3.53 (3.43-3.63) 3.67 (3.57-3.78) <0.001
Asthma 1.53 (1.50-1.57) 1.40 (1.36-1.43) <0.001
Cancer 2.71 (2.62-2.80) 2.75 (2.65-2.84) <0.001
Heart failure 2.66 (2.56-2.76) 3.15 (3.02-3.27) <0.001
Acute myocardial infarction 2.04 (1.95-2.13) 2.31 (2.20-2.41) <0.001
Stroke 3.11 (2.94-3.29) 3.30 (3.12-3.50) <0.001
Hip fracture 3.76 (3.41-4.15) 3.31 (3.00-3.66) <0.001
Epilepsy 2.21 (2.03-2.41) 2.27 (2.08-2.48) <0.001
Dementia 3.14 (2.84-3.47) 2.89 (2.61-3.20) <0.001
Parkinsonism 3.96 (3.31-4.74) 4.27 (3.55-5.14) <0.001
Multiple sclerosis 4.10 (3.33-5.04) 4.00 (3.22-4.95) <0.001
*Selection by the corresponding Canadian Chronic Disease Surveillance System case definition (Table S4, supplementary digital content, available at https://links.lww.com/PAIN/B331) applied to the 1999 to 2009 provincial cohort NL MCP Fee-for-Service Physician Claims File and Provincial Discharge Abstract Data determined comorbid condition case status.
Selection by the Chronic Pain Algorithm applied to 1999 to 2009 provincial cohort NL MCP Fee-for-Service Physician Claims File data determined chronic pain case status. The Chronic Pain Algorithm was defined as: (1) a single encounter date with an anesthesiologist recording a chronic pain-related provincial MCP procedure billing code (Table S2, supplementary digital content, available at https://links.lww.com/PAIN/B331) OR (2) 5 or more encounter dates with any physician recording any pain-related diagnostic code (Table S3, supplementary digital content, available at https://links.lww.com/PAIN/B331) in a 5-year period with more than 183 days separating at least 2 pain-related encounter dates.
Adjusted for sex, regional health authority, and rural/urban residential location using a logistic regression model.
§Statistical significance was defined as P < 0.05.
CI, confidence interval; MCP, Medical Care Plan; NL, Newfoundland and Labrador.

An estimated 74.7% of the CPG were identified to have at least 1 comorbid condition between 1999 and 2009, and 16.9% were identified to have at least 3 comorbid conditions. The CPG members were identified to have a mean (SD) of 1.41 (1.30) comorbid conditions (range 0-10), compared with the NCPG members who were identified to have a mean (SD) of 0.58 (0.93) comorbid conditions (range 0-9). The likelihood of being identified as having 1, 2, or 3 or more comorbid conditions was significantly higher in the CPG than in the NCPG as determined by the adjusted odds ratio. The adjusted odds ratio between the CPG and the NCPG of having 1 comorbid condition from 1999 to 2009 was 1.56 (95% CI: 1.54-1.58), of having 2 comorbid conditions was 2.87 (95% CI: 2.83-2.93), and of having 3 or more comorbid conditions was 4.25 (95% CI: 4.16-4.34).

3.3. Health care utilization

In the 2009/2010 fiscal year, 73.0% (95% CI: 72.9%-73.2%) of the overall provincial cohort had at least 1 family physician visit, 58.3% (95% CI: 58.1%-58.4%) had at least 1 specialist visit, 37.3% (95% CI: 37.2%-37.5%) had at least 1 general radiograph assessment visit, 9.3% (95% CI: 9.2%-9.4%) had at least 1 computed tomography scan visit, and 2.2% (95% CI: 2.2%-2.3%) had at least 1 magnetic resonance imaging scan visit. In the 2009/2010 fiscal year, 11.0% (95% CI: 10.9%-11.1%) of the provincial cohort had at least 1 day surgery admission and 7.1% (95% CI: 7.1%-7.2%) had at least 1 inpatient admission.

As determined by the adjusted relative risk ratio, there was a significantly higher likelihood (Table 3) in 2009/2010 of CPG members than NCPG members to have a physician visit (94.6% [95% CI: 94.5%-94.7%] vs 74.5% [95% CI: 74.3%-74.6%], adjusted relative risk ratio: 1.12 [95% CI: 1.12-1.13]), a diagnostic imaging visit (62.5% [95% CI: 62.3%-62.7%] vs 35.0% [95% CI: 34.8%-35.1%], adjusted relative risk ratio: 1.41 [95% CI: 1.40-1.42]), or a hospital admission (23.0% [95% CI: 22.8%-23.1%] vs 12.7% [95% CI: 12.6%-12.8%], adjusted relative risk ratio: 1.40 [95% CI: 1.38-1.42]). The relative risk ratio for visits (2.14 [95% CI: 2.12-2.16]) and admissions (3.07 [95% CI: 2.92-3.23]) for pain-related conditions (based on the diagnostic code associated with the visit or admission being classified as pain-related [Table S3, supplementary digital content, available at https://links.lww.com/PAIN/B331]) between CPG members and NCPG members was higher than the relative risk ratio for all-cause visits (1.12 [95% CI: 1.12-1.13]) and admissions (1.40 [95% CI: 1.38-1.42]).

Table 3 - Risk to utilize health services in Newfoundland and Labrador, Canada, in 2009/2010.
Health service type Chronic Pain Group* No Chronic Pain Group* Unadjusted relative risk ratio (95% CI) Adjusted relative risk ratio (95% CI) P
N = 184,580 N = 320,113
% risk (95% CI) % risk (95% CI)
All-cause reason
 Any family physician visit 90.2 (90.1-90.4) 63.1 (63.0-63.3) 1.43 (1.42-1.43) 1.20 (1.20-1.21) <0.001
 Any specialist§ visit 75.9 (75.7-76.1) 48.1 (47.9-48.2) 1.58 (1.57-1.59) 1.28 (1.28-1.29) <0.001
 Any day surgery admission 16.3 (16.2-16.5) 8.0 (7.9-8.1) 2.05 (2.02-2.09) 1.55 (1.52-1.57) <0.001
 Any inpatient admission 9.3 (9.2-9.5) 5.9 (5.8-5.9) 1.59 (1.56-1.62) 1.20 (1.18-1.23) <0.001
Pain-related reason
 Any family physician visit 52.9 (52.7-53.1) 19.1 (19.0-19.3) 2.77 (2.74-2.79) 2.21 (2.19-2.23) <0.001
 Any specialist visit 16.1 (15.9-16.3) 4.2 (4.2-4.3) 3.81 (3.73-3.88) 2.74 (2.68-2.80) <0.001
 Any day surgery admission 1.9 (1.9-2.0) 0.5 (0.5-0.5) 4.03 (3.80-4.27) 3.24 (3.04-3.46) <0.001
 Any inpatient admission 1.3 (1.2-1.3) 0.4 (0.3-0.4) 3.65 (3.41-3.92) 2.89 (2.68-3.13) <0.001
Diagnostic imaging
 Any general radiograph assessment 53.2 (52.9-53.4) 28.2 (28.1-28.4) 1.88 (1.87-1.90) 1.46 (1.45-1.47) <0.001
 Any computed tomography scan 15.2 (15.1-15.4) 5.9 (5.8-5.9) 2.60 (2.55-2.65) 1.72 (1.69-1.76) <0.001
 Any magnetic resonance imaging scan 3.7 (3.6-3.8) 1.4 (1.3-1.4) 2.66 (2.57-2.77) 2.05 (1.96-2.14) <0.001
*Selection by the Chronic Pain Algorithm applied to 1999 to 2009 provincial cohort NL MCP Fee-for-Service Physician Claims File data determined Chronic Pain Group or No Chronic Pain Group membership. The Chronic Pain Algorithm was defined as (1) a single encounter date with an anesthesiologist recording a chronic pain-related provincial MCP procedure billing code (Table S2, supplementary digital content, available at https://links.lww.com/PAIN/B331) OR (2) 5 or more encounter dates with any physician recording any pain-related diagnostic code (Table S3, supplementary digital content, available at https://links.lww.com/PAIN/B331) in a 5-year period with more than 183 days separating at least 2 pain-related encounter dates.
Adjusted for the covariates of age group, sex, regional health authority, rural/urban residential location, and number of comorbid conditions using a robust Poisson regression model with log link function.
Statistical significance was defined as P < 0.05.
§Specialist defined as any physician not identified as a family physician.
Presence of a pain-related diagnostic code (Table S3, supplementary digital content, available at https://links.lww.com/PAIN/B331) with the physician claim or as the Most Responsible Reason for admission.
CI, confidence interval; MCP, Medical Care Plan; N, total number in group; NL, Newfoundland and Labrador.

In 2009/2010, the mean all-cause visit rates per 100 person-years for the overall provincial cohort was 478 (95% CI: 476-480) family physician visits, 418 (95% CI: 415-420) specialist visits, 92 (95% CI: 91-92) general radiograph assessment visits, 21 (95% CI: 21-21) computed tomography scan visits, and 6 (95% CI: 6-6) magnetic resonance imaging visits. The mean all-cause admission rates per 100 person-years in 2009/2010 for the overall provincial cohort were 15 (95% CI: 15-15) day surgery admissions and 10 (95% CI: 10-10) inpatient admissions. Pain-related visits/admissions comprised 17.4% of all family physician visits, 6.9% of all specialist visits, 9.9% of all day surgery admissions, and 7.6% of all inpatient admissions.

In 2009/2010, 58.8% of all physician visits, 57.6% of all diagnostic imaging visits, and 54.2% of all hospital admissions were attributed to the CPG. Proportion of all visits/admissions attributed to pain-related conditions was higher for the CPG compared with the NCPG, comprising 21.5% vs 11.6% of all per group family physician visits, 8.7% vs 4.4% of all per group specialist visits, 13.1% vs 5.6% of all per group day surgery admissions, and 10.4% vs 4.9% of all per group inpatient admissions.

As determined by the adjusted rate ratio, the mean all-cause rates per 100 person-years (Table 4) in 2009/2010 were significantly higher for the CPG compared with the NCPG for physician visits (1440 [95% CI: 1432-1448] vs 582 [95% CI: 579-584], adjusted rate ratio: 1.63 [95% CI: 1.62-1.65]), diagnostic imaging visits (260 [95% CI: 258-262] vs 110 [95% CI: 109-111], adjusted rate ratio: 1.64 [95% CI: 1.62-1.66]), and hospital admissions (36 [95% CI: 36-37] vs 18 [95% CI: 18-18], adjusted rate ratio: 1.50 [95% CI: 1.47-1.52]). The adjusted rate ratio of physician visits (3.28 [95% CI: 3.24-3.32]) and admissions (3.72 [95% CI: 3.52-3.93]) for pain-related conditions between CPG members and NCPG members was higher than the adjusted rate ratio for all-cause physician visits (1.63 [95% CI: 1.62-1.65]) and admissions (1.50 [95% CI: 1.47-1.52]).

Table 4 - Health service utilization rates in Newfoundland and Labrador, Canada, in 2009/2010.
Health service type Chronic Pain Group* No Chronic Pain Group* Unadjusted rate ratio (95% CI) Adjusted rate ratio (95% CI) P
N = 184,580 N = 320,113
Mean rate (95% CI) per 100 person-years Mean rate (95% CI) per 100 person-years
All-cause reason
 Family physician visits 770 (766-775) 309 (308-311) 2.49 (2.47-2.51) 1.71 (1.70-1.72) <0.001
 Specialist§ visits 669 (664-675) 272 (270-273) 2.46 (2.43-2.48) 1.57 (1.56-1.59) <0.001
 Day surgery admissions 23 (23-23) 10 (10-10) 2.32 (2.28-2.36) 1.68 (1.65-1.72) <0.001
 Inpatient admissions 13 (13-14) 8 (8-8) 1.71 (1.67-1.75) 1.24 (1.20-1.27) <0.001
Pain-related reason
 Family physician visits 166 (165-167) 36 (35-36) 4.64 (4.59-4.70) 3.34 (3.30-3.38) <0.001
 Specialist visits 58 (57-59) 12 (12-12) 4.83 (4.71-4.96) 3.12 (3.03-3.21) <0.001
 Day surgery admissions 3 (3-3) 0.6 (0.5-0.6) 5.39 (5.05-5.75) 4.17 (3.88-4.49) <0.001
 Inpatient admissions 1 (1-1) 0.4 (0.4-0.4) 3.64 (3.39-3.91) 2.94 (2.71-3.19) <0.001
Diagnostic imaging
 General radiograph assessment 145 (143-146) 61 (61-61) 2.37 (2.34-2.39) 1.71 (1.70-1.73) <0.001
 Computed tomography scans 34 (34-35) 13 (13-13) 2.60 (2.54-2.67) 1.68 (1.64-1.73) <0.001
 Magnetic resonance imaging scans 10 (10-10) 4 (4-4) 2.69 (2.55-2.84) 2.13 (2.01-2.26) <0.001
*Selection by the Chronic Pain Algorithm applied to 1999 to 2009 provincial cohort NL MCP Fee-for-Service Physician Claims File data determined Chronic Pain Group or No Chronic Pain Group membership. The Chronic Pain Algorithm was defined as (1) a single encounter date with an anesthesiologist recording a chronic pain-related provincial MCP procedure billing code (Table S2, supplementary digital content, available at https://links.lww.com/PAIN/B331) OR (2) 5 or more encounter dates with any physician recording any pain-related diagnostic code (Table S3, supplementary digital content, available at https://links.lww.com/PAIN/B331) in a 5-year period with more than 183 days separating at least 2 pain-related encounter dates.
Adjusted for the covariates of age group, sex, regional health authority, rural/urban residential location, and number of comorbid conditions using a negative binomial regression model.
Statistical significance was defined as P < 0.05.
§Specialist defined as any physician not identified as a family physician.
Presence of a pain-related diagnostic code (Table S3, supplementary digital content, available at https://links.lww.com/PAIN/B331) with the physician claim or as the Most Responsible Reason for admission.
CI, confidence interval; MCP, Medical Care Plan; NL, Newfoundland and Labrador.

4. Discussion

The Chronic Pain Algorithm performed comparably with previous chronic pain survey case definitions by identifying (1) 36.6% prevalent chronic pain cases in the NL, Canada, provincial cohort (Canadian surveys reported 11%-44% prevalence),11,14,23,72,98,111 (2) higher chronic pain proportions in females (Canadian surveys reported 13%-66.7% prevalence) vs males (Canadian surveys reported 8%-57.1% prevalence),14,68,91,92,98,101 and (3) increasing chronic pain proportions with increasing age (Canadian and global surveys reported similar observations).11,14,23,52,73,85,98,101,108,111 Population-based chronic pain comparator groups were identified using this validated algorithm applied to NL province-level health administrative data sets,30 and chronic comorbidity prevalence and 2009/2010 hospital and fee-for-service physician utilization were quantified generating 3 main findings. First, the CPG accounted for significantly higher total utilization of publicly funded physician visits (58.8%), diagnostic imaging visits (57.6%), and hospital admissions (54.2%). Second, even after controlling for potential confounding variables, the CPG had about 2 to 4 times the odds of having comorbid chronic conditions, most notably mental health, cardiovascular, and neurodegenerative conditions. Finally, physician visits and hospital admissions for pain-related conditions formed an unexpectedly small percentage of the total measured utilization even for the CPG.

4.1. Chronic pain and excess health care utilization

When adjusting for demographics and comorbid conditions, the CPG had a 24% to 113% higher rate of observed health service use in 2009/2010 than the NCPG. CPG members had particularly high usage of expensive services, such as specialist assessments, hospital admissions, and specialized imaging tests.64,90 Although exact comparisons to other studies were not possible,9,44,49,50,59,113 this study adds to the body of knowledge that chronic pain presence is significantly associated with increased health care utilization.44,58,59 Some cited factors contributing to this excess service use include provider practice patterns, higher pain-related interference, multimorbidity, prescribed opioid use, and lower socioeconomic status,49,64,70,73,95 many of which were beyond the scope of this study's data to investigate. Although adequate chronic pain management may have necessitated the increased family physician and specialized service visits,84 it is possible that lack of satisfactory pain control and reduced quality of life may have played a more significant role.74,84 Surveying health service provision in NL to ensure coordinated chronic pain management that focuses on self-management and reduced pain-related interference may help foster more appropriate acute and specialized health care utilization.42,64,84,104,113

4.2. Chronic pain and comorbid conditions

The aging demographics, high chronic disease rates, and poor population health indicators in NL compared with other Canadian jurisdictions highlight the importance of measuring the association between chronic pain and comorbid conditions for the purposes of public health initiatives and resource planning.25,80 CPG members had up to 4 times the odds of having any single comorbid condition and 4 times the odds of having 3 or more comorbid conditions than NCPG members. The findings from the NL data support other studies describing the strong association between chronic pain and other chronic diseases, despite varying methods of case ascertainment and data sources.65,83,107

There is commonality in the complex biological, psychological, social, cultural, and genetic processes involved in the development of chronic comorbid conditions and chronic pain.33,37,46,115,116 Regardless of whether chronic pain was the primary or secondary chronic disease diagnosis, its effect on stress levels, physical activity, and overall quality of life negatively impacts a person's ability to maximize recovery and effectively manage chronic disease in the long term.29,34,51,69,82 Concurrent management of pain with chronic disease may help maximize clinical outcomes and mitigate the negative effects on quality of life.27

4.3. Pain-related vs non-pain–related care

Only 15.6% of the physician claims and 12.1% of the hospital admissions attributed to the CPG were for pain-related conditions (previous studies also reported <50% of encounters were for pain-related reasons).96,121 Chronic pain has a strong association with multicomorbidity. Chronic diseases—such as diabetes—necessitate more frequent medical follow-up to monitor treatment effectiveness (eg, medication or lifestyle changes) over time.6,9,64 Pain developing secondary to some chronic diseases may represent disease progression requiring more advanced care management, and diagnostic codes accompanying claims may reflect clinical care of the worsening disease rather than the pain related to it.27,55 However, delayed recognition of pain as the primary issue may result in a non-pain-related diagnosis erroneously being recorded (and treated) as the primary reason for the encounter/admission.97,102,113 People experiencing pain conditions not recognized and adequately managed by their health care providers may choose to self-medicate using over-the-counter analgesics, including codeine formulations, further contributing to hyperalgesia and pain chronification.48,71 The discord in patient needs vs care provided may delay appropriate pain treatment, prolong suffering, and influence higher acute care service utilization.28,29,47 Examining the care/administrative processes involved in this observation may foster more effective individualized pain and chronic disease management practices.

4.4. Strengths, limitations, and generalizability

There were 2 main strengths to this study. First, NL health administrative data sources had wide coverage and regular data quality checks76,77 and were not subjected to the recall bias, sampling errors, or low response rates that can plague survey data.3,120 Second, the chronic pain and comorbid condition coding algorithms were validated in the same data sets and were subjected to the same data limitations increasing internal reliability of results.30,60

There were 7 main limitations in this study. First, a chronic limitation for secondary use of health administrative data is the dependence of its data accuracy on entry at source.7,66,123 Second, nondifferential misclassification bias between the chronic pain and the comorbid condition case definitions in the NL claims data was present when assessing their strength of association (potential coding errors, undiagnosed chronic conditions, and 1 code entry per claim limit contributed to the bias).1,7,61 Third, the strength of association between chronic pain and comorbid conditions was, in part, influenced by medical surveillance bias where regular health care encounters to manage 1 condition increased the likelihood of identifying presence of another.106 This strength of association was also possibly influenced by age as a confounding variable (collinearity precluded its inclusion as a covariate). Fourth, the unavailability of data collected from visits to pharmacies, emergency departments, salaried physicians, allied health professionals, and those funded by third-party payers likely negatively impacted chronic pain and comorbid condition case ascertainment and health care utilization capture, particularly in rural and/or non-Eastern Regional Health Authority areas.17,44,58,61,90 Fifth, bias may have occurred in quantifying comorbidity presence and health care utilization in prevalent vs incident chronic pain cases since longer pain duration may be associated with more severe negative health outcomes.58 Sixth, adjusting for the potential bias to the chronic pain exposure measurement introduced by the Chronic Pain Algorithm (potential overascertainment30) and the non–fee-for-service physician data unavailability (potential underascertainment61) would be complex and require access to variables and data sets outside the scope and funding of this study.22,61 Therefore, the chronic pain disease burden measures presented here should be interpreted with caution. Finally, subjective data describing important factors impacting health care utilization, such as self-reports of pain severity/interference,104 were not captured by the administrative data sources.

The health care utilization rates reported in this study should not be generalized to the Canadian population because of potential differences in regional practice and remuneration patterns.17,119 Although assessing the strength of association between NL health administrative data-derived variables and the presence of chronic pain was considered an appropriate use of the Chronic Pain Algorithm, its performance on selection accuracy testing precluded its utility in assessing causation (considering chronic pain as the exposure or outcome).30 The comorbid condition presence and health care utilization estimates provided are representative for the NL population up to the 2009/2010 fiscal year, which was the latest available data at the study initiation. Given that patterns of disease may have shifted in the past decade, this study provides a baseline against which to compare future estimations using the presented methodology. Future study is recommended to improve the validity and utility of the Chronic Pain Algorithm by incorporating flexible algorithms based on age- and/or condition-specific health service use/codes, using a more restrictive diagnostic code list, and validating the algorithm in non-NL population health administrative data.7,30 Similarity in the structure of health service delivery and physician claims/hospital discharge data sets across Canadian jurisdictions increases the generalizability of the chronic pain/chronic disease case ascertainment and health care utilization quantification methods presented in this study.60,88

5. Conclusion

This study provides, for the first time in NL, estimates of comorbid condition prevalence and publicly funded health care utilization of people identified as having chronic pain from health administrative data. When controlling for measured confounders, being identified as having chronic pain had modest to strong association with having 1 or more chronic comorbid conditions, and modest association with having higher utilization of publicly funded physician, hospital, and diagnostic imaging services. However, about 84% of physician and 88% of hospital care of CPG members were for conditions not related to pain; a further indication of the complexities to address when managing a chronic pain condition. Deeper examination of the patient-level, practitioner-level, and system-level factors driving higher health care utilization by NL residents may foster more effective personalized care of individuals with chronic conditions, including pain.

Conflict of interest statement

H.E. Foley reports grants from the Physiotherapy Foundation of Canada and grants from the Health Care Foundation of the Eastern Regional Health Authority, during the conduct of this study. R. Audas has nothing to disclose. J.C. Knight has nothing to disclose. M. Ploughman has nothing to disclose. S. Asghari reports grants from Canadian Institutes of Health Research (CIHR), the Government of Newfoundland and Labrador, Eastern Health, Memorial University of Newfoundland, and the Trinity Conception Placentia Health Foundation, grants from International Grenfell Association, Mitacs and Memorial University of Newfoundland, outside the submitted work.

Preliminary results from this manuscript were presented in a teleconference to the Newfoundland and Labrador Provincial Pain Management Advisory Council on June 11, 2019, and as an oral presentation at the 2019 PriFor Conference in St. John's, Newfoundland and Labrador, on June 28, 2019. A previous version of this manuscript was submitted to the School of Graduate Studies, Memorial University of Newfoundland, on September 16, 2020, as part of a Doctoral Thesis titled “Epidemiology and Health Care Utilization Associated with Chronic Pain in Newfoundland and Labrador, Canada: A Population-Based Study Using Health Administrative Data” authored by H.E. Foley and was successfully examined on December 16, 2020. The final version of the Doctoral Thesis was submitted to the Memorial University Library on January 7, 2021, and a thesis publication embargo until January 1, 2022, was requested.

Appendix A. Supplemental digital content

Supplemental digital content associated with this article can be found online at https://links.lww.com/PAIN/B331.

Acknowledgements

The authors thank Dr James Flynn and Elsie Thistle for serving as clinical advisors in chronic pain management and in pain-related diagnostic and procedure code selection and Dr Jason McCarthy for assistance in grant applications.

This research was supported by the 2013 B.E. Schnurr Memorial Fund Research Grant and administered by the Physiotherapy Foundation of Canada awarded to H.E. Foley. This research was also supported by the 2012 Eastern Health Research Grant administered by the Health Care Foundation of the Eastern Regional Health Authority awarded to H.E. Foley and Jason McCarthy. The study sponsors were not involved in any stage of the study from initial design to publication.

The data that support the findings of this study are securely held at the NL Centre for Health Information and the Eastern Regional Health Authority but restrictions apply to the availability of these data, which were used under data-sharing agreements for the current study, and so are not publicly available. Data are however available from the NL Centre of Health Information if prespecified criteria are met following its information request procedures (https://nlchi.nl.ca/index.php/quality-information/information-requests). The study author (H.E.F.) may be contacted for the data set creation plan.

Authors' contributions: All authors read and approved the submitted version of this manuscript. All authors agreed to be personally responsible for their own contribution and ensure that questions related to the accuracy or integrity of any part of the work, even ones in which the author was not personally involved, are appropriately investigated, resolved, and the resolution documented in the literature. H.E. Foley drafted the work and substantively revised it and made substantial contributions towards the study design, data acquisition, analysis, and interpretation. R. Audas substantively revised the work and made substantial contributions to the study design and data acquisition, analysis, and interpretation. J.C. Knight substantively revised the work and made substantial contributions towards study design, data acquisition, analysis, and interpretation. M. Ploughman substantively revised the work and made substantial contributions towards study conception and design and data acquisition and interpretation. S. Asghari substantively revised the work and made substantial contributions towards study design, and data acquisition, analysis, and interpretation, and writing of the manuscript.

References

[1]. Alaghehbandan R, Macdonald D, Barrett B, Collins K, Chen Y. Using administrative databases in the surveillance of depressive disorders-case definitions. Popul Health Manag 2012;15:372–80.
[2]. Andersen RM. Revisiting the behavioral model and access to medical care: does it matter? J Health Soc Behav 1995;36:1–10.
[3]. Aschengrau A, Seage GR. Essentials of epidemiology in public health. 3rd ed. Sudbury: Jones & Bartlett Learning, 2014.
[4]. Bancej C, Ellison J, Gao Y, Klaver K, Loukine L, McRae L, O'Donnell S, Onysko J, Pelletier C, Pelletier L, Robbins G, Shane A, Vanderloo S, Waters C. Report from the Canadian Chronic Disease Surveillance System: Mental Illness in Canada, 2015. Ottawa: Public Health Agency of Canada, 2015. Available at: https://www.canada.ca/en/public-health/services/publications/diseases-conditions/report-canadian-chronic-disease-surveillance-system-mental-illness-canada-2015.html. Accessed July 5, 2020.
[5]. Beaudet N, Courteau J, Sarret P, Vanasse A. Prevalence of claims-based recurrent low back pain in a Canadian population: a secondary analysis of an administrative database. BMC Musculoskelet Disord 2013;14:151.
[6]. Beehler GP, Rodrigues AE, Mercurio-Riley D, Dunn AS. Primary care utilization among veterans with chronic musculoskeletal pain: a retrospective chart review. Pain Med 2013;14:1021–31.
[7]. Benchimol EI, Manuel DG, To T, Griffiths AM, Rabeneck L, Guttmann A. Development and use of reporting guidelines for assessing the quality of validation studies of health administrative data. J Clin Epidemiol 2011;64:821–9.
[8]. Berger A, Dukes EM, Oster G. Clinical characteristics and economic costs of patients with painful neuropathic disorders. J Pain 2004;5:143–9.
[9]. Bernfort L, Gerdle B, Rahmqvist M, Husberg M, Levin LA. Severity of chronic pain in an elderly population in Sweden-impact on costs and quality of life. PAIN 2015;156:521–7.
[10]. Birgenheir DG, Ilgen MA, Bohnert ASB, Abraham KM, Bowersox NW, Austin K, Kilbourne AM. Pain conditions among veterans with schizophrenia or bipolar disorder. Gen Hosp Psychiatry 2013;35:480–4.
[11]. Birse TM, Lander J. Prevalence of chronic pain. Can J Public Health 1998;89:129–31.
[12]. Bonica JJ. History of pain concepts and pain therapy. Mt Sinai J Med 1991;58:191–202.
[13]. Bouck Z, Pendrith C, Chen XK, Frood J, Reason B, Khan T, Costante A, Kirkham K, Born K, Levinson W, Bhatia RS. Measuring the frequency and variation of unnecessary care across Canada. BMC Health Serv Res 2019;19:446.
[14]. Boulanger A, Clark AJ, Squire P, Cui E, Horbay GLA. Chronic pain in Canada: have we improved our management of chronic noncancer pain? Pain Res Manag 2007;12:39–47.
[15]. Breivik H, Collett B, Ventafridda V, Cohen R, Gallacher D. Survey of chronic pain in Europe: prevalence, impact on daily life, and treatment. Eur J Pain 2006;10:287–333.
[16]. Canada's Health Care System. Government of Canada. 2016. Available at: https://www.canada.ca/en/health-canada/services/canada-health-care-system.html. Accessed December 29, 2017.
[17]. Canadian Institute for Health Information. Profiling physicians by payment program: a closer look at three provinces. 2010. Available at: https://secure.cihi.ca/free_products/physicians_payment_aib_2010_e.pdf. Accessed January 10, 2020.
[18]. Canadian Institute for Substance Use Research and the Canadian Centre on Substance Use and Addiction. Canadian Substance Use Costs and Harms Scientific Working Group: Canadian substance use costs and harms (2007-2014). 2018. Available at: https://www.ccsa.ca/sites/default/files/2019-04/CSUCH-Canadian-Substance-Use-Costs-Harms-Report-2018-en.pdf. Accessed June 18, 2020.
[19]. Chapman CR, Vierck CJ. The transition of acute postoperative pain to chronic pain: an integrative overview of research on mechanisms. J Pain 2017;18:359.e351–359.e338.
[20]. Chen W, Qian L, Shi J, Franklin M. Comparing performance between log-binomial and robust Poisson regression models for estimating risk ratios under model misspecification. BMC Med Res Methodol 2018;18:63.
[21]. Choiniere M, Dion D, Peng P, Banner R, Barton PM, Boulanger A, Clark AJ, Gordon AS, Guerriere DN, Guertin MC, Intrater HM, Lefort SM, Lynch ME, Moulin DE, Ong-Lam M, Racine M, Rashiq S, Shir Y, Taenzer P, Ware M. The Canadian STOP-PAIN project—part 1: who are the patients on the waitlists of multidisciplinary pain treatment facilities? Can J Anaesth 2010;57:539–48.
[22]. Chubak J, Pocobelli G, Weiss NS. Tradeoffs between accuracy measures for electronic health care data algorithms. J Clin Epidemiol 2012;65:343–9.
[23]. Crook J, Rideout E, Browne G. The prevalence of pain complaints in a general population. PAIN 1984;18:299–314.
[24]. Department of Finance. 2016 census information and statistics. Government of Newfoundland and Labrador, 2016. Available at: https://stats.gov.nl.ca/Statistics/Census2016/PDF/Pop_CSD_Alphabetical_2016.pdf. Accessed February 15, 2020.
[25]. Department of Finance. Population projections Newfoundland and Labrador, high scenario, April 2014, Total Province. Government of Newfoundland and Labrador, 2014. Available at: http://www.economics.gov.nl.ca/pdf/Popbyagehigh-web.pdf. Accessed May 10, 2014.
[26]. Doshi JA, Cai Q, Buono JL, Spalding WM, Sarocco P, Tan H, Stephenson JJ, Carson RT. Economic burden of irritable bowel syndrome with constipation: a retrospective analysis of health care costs in a commercially insured population. J Manag Care Spec Pharm 2014;20:382–90.
[27]. Dworkin RH, Malone DC, Panarites CJ, Armstrong EP, Pham SV. Impact of postherpetic neuralgia and painful diabetic peripheral neuropathy on health care costs. J Pain 2010;11:360–8.
[28]. Ehde DM, Jensen MP, Engel JM, Turner JA, Hoffman AJ, Cardenas DD. Chronic pain secondary to disability: a review. Clin J Pain 2003;19:3–17.
[29]. Engel JM, Jensen MP, Hoffman AJ, Kartin D. Pain in persons with cerebral palsy: extension and cross validation. Arch Phys Med Rehabil 2003;84:1125–8.
[30]. Foley HE, Knight JC, Ploughman M, Asghari S, Audas R. Identifying cases of chronic pain using health administrative data: a validation study. Cdn J Pain 2020;4:252–257.
[31]. Gaskin DJ, Richard P. The economic costs of pain in the United States. J Pain 2012;13:715–24.
[32]. Gatchel RJ, McGeary DD, McGeary CA, Lippe B. Interdisciplinary chronic pain management: past, present, and future. Am Psychol 2014;69:119–30.
[33]. Gerrits MM, van Oppen P, van Marwijk HW, Penninx BW, van der Horst HE. Pain and the onset of depressive and anxiety disorders. PAIN 2014;155:53–9.
[34]. Gerrits MM, Vogelzangs N, van Oppen P, van Marwijk HW, van der Horst H, Penninx BW. Impact of pain on the course of depressive and anxiety disorders. PAIN 2012;153:429–36.
[35]. Goff SL, Feld A, Andrade SE, Mahoney L, Beaton SJ, Boudreau DM, Davis RL, Goodman M, Hartsfield CL, Platt R, Roblin D, Smith D, Yood MU, Dodd K, Gurwitz JH. Administrative data used to identify patients with irritable bowel syndrome. J Clin Epidemiol 2008;61:617–21.
[36]. Government of Canada. Opioids and the opioid crisis—get the facts. 2020. Available at: https://www.canada.ca/en/health-canada/services/substance-use/problematic-prescription-drug-use/opioids/get-the-facts.html - a5. Accessed May 15, 2020.
[37]. Graham H. Social determinants and their unequal distribution: clarifying policy understandings. Milbank Q 2004;82:101–24.
[38]. Guerriere DN, Choiniere M, Dion D, Peng P, Stafford-Coyte E, Zagorski B, Banner R, Barton PM, Boulanger A, Clark AJ, Gordon AS, Guertin MC, Intrater HM, Lefort SM, Lynch ME, Moulin DE, Ong-Lam M, Racine M, Rashiq S, Shir Y, Taenzer P, Ware M. The Canadian STOP-PAIN project—part 2: what is the cost of pain for patients on waitlists of multidisciplinary pain treatment facilities? Can J Anaesth 2010;57:549–58.
[39]. Gustavsson A, Bjorkman J, Ljungcrantz C, Rhodin A, Rivano-Fischer M, Sjolund KF, Mannheimer C. Socio-economic burden of patients with a diagnosis related to chronic pain-register data of 840,000 Swedish patients. Eur J Pain 2012;16:289–99.
[40]. Harrold LR, Yood RA, Andrade SE, Reed JI, Cernieux J, Straus W, Weeks M, Lewis B, Gurwitz JH. Evaluating the predictive value of osteoarthritis diagnoses in an administrative database. Arthritis Rheum 2000;43:1881–5.
[41]. Häuser W, Schmutzer G, Hilbert A, Brähler E, Henningsen P. Prevalence of chronic disabling noncancer pain and associated demographic and medical variables: a cross-sectional survey in the general German population. Clin J Pain 2015;31:886–92.
[42]. Health Canada. Chronic pain in Canada: laying a foundation for action. A report by the Canadian Pain Task Force. 2019. Available at: https://www.canada.ca/en/health-canada/corporate/about-health-canada/public-engagement/external-advisory-bodies/canadian-pain-task-force/report-2019.html#intro. Accessed October 4, 2019.
[43]. Health Canada. Working together to better understand, prevent, and manage chronic pain: what we heard. A report by the Canadian Pain Task Force. 2020. Available at: https://www.canada.ca/en/health-canada/corporate/about-health-canada/public-engagement/external-advisory-bodies/canadian-pain-task-force/report-2020.html. Accessed December 22, 2020.
[44]. Hogan ME, Taddio A, Katz J, Shah V, Krahn M. Incremental health care costs for chronic pain in Ontario, Canada: a population-based matched cohort study of adolescents and adults using administrative data. PAIN 2016;157:1626–33.
[45]. Hong J, Reed C, Novick D, Happich M. Costs associated with treatment of chronic low back pain: an analysis of the UK General Practice Research Database. Spine 2013;38:75–82.
[46]. IOM (Institute of Medicine). Relieving pain in America: a blueprint for transforming prevention, care, education, and research. Washington, DC: The National Academies Press, 2011. Available at: https://www.nap.edu/catalog/13172/relieving-pain-in-america-a-blueprint-for-transforming-prevention-care. Accessed May 7, 2019.
[47]. Jensen MP, Hoffman AJ, Stoelb BL, Abresch RT, Carter GT, McDonald CM. Chronic pain in persons with myotonic dystrophy and facioscapulohumeral dystrophy. Arch Phys Med Rehabil 2008;89:320–8.
[48]. Johnson JL, Rolan PE, Johnson ME, Bobrovskaya L, Williams DB, Johnson K, Tuke J, Hutchinson MR. Codeine-induced hyperalgesia and allodynia: investigating the role of glial activation. Transl Psychiatry 2014;4:e482.
[49]. Jonsdottir T, Jonsdottir H, Lindal E, Oskarsson GK, Gunnarsdottir S. Predictors for chronic pain-related health care utilization: a cross-sectional nationwide study in Iceland. Health Expect 2015;18:2704–19.
[50]. Kay C, Wozniak E, Bernstein J. Utilization of health care services and ambulatory resources associated with chronic noncancer pain. Pain Med 2017;18:1236–46.
[51]. Khan F, Pallant J. Chronic pain in multiple sclerosis: prevalence, characteristics, and impact on quality of life in an Australian community cohort. J Pain 2007;8:614–23.
[52]. King S, Chambers CT, Huguet A, MacNevin RC, McGrath PJ, Parker L, MacDonald AJ. The epidemiology of chronic pain in children and adolescents revisited: a systematic review. PAIN 2011;152:2729–38.
[53]. Knight JC, Mathews M, Aubrey-Bassler K. Relation between family physician retention and avoidable hospital admission in Newfoundland and Labrador: a population-based cross-sectional study. CMAJ Open 2017;5:E746–52.
[54]. Lacasse A, Cauvier Charest E, Dault R, Cloutier AM, Choiniere M, Blais L, Vanasse A. Validity of algorithms for identification of individuals suffering from chronic noncancer pain in administrative databases: a systematic review. Pain Med 2020;21:1825–39.
[55]. Lacasse A, Ware MA, Dorais M, Lanctot H, Choiniere M. Is the Quebec provincial administrative database a valid source for research on chronic non-cancer pain?. Pharmacoepidemiol Drug Saf 2015;24:980–90.
[56]. Lachaine J, Beauchemin C, Landry PA. Clinical and economic characteristics of patients with fibromyalgia syndrome. Clin J Pain 2010;26:284–90.
[57]. Lachaine J, Gordon A, Choiniere M, Collet JP, Dion D, Tarride JE. Painful neuropathic disorders: an analysis of the Regie de l'Assurance Maladie du Quebec database. Pain Res Manag 2007;12:31–7.
[58]. Lalonde L, Choiniere M, Martin E, Berbiche D, Perreault S, Lussier D. Costs of moderate to severe chronic pain in primary care patients - a study of the ACCORD Program. J Pain Res 2014;7:389–403.
[59]. Langley PC. The societal burden of pain in Germany: health-related quality-of-life, health status and direct medical costs. J Med Econ 2012;15:1201–15.
[60]. Lix LM, Ayles J, Bartholomew S, Cooke CA, Ellison J, Emond V, Hamm NC, Hannah H, Jean S, LeBlanc S, O'Donnell S, Paterson JM, Pelletier C, Phillips KA, Puchtinger R, Reimer K, Robitaille C, Smith M, Svenson LW, Tu K, VanTil LD, Waits S, Pelletier L. The Canadian chronic disease surveillance system: a model for collaborative surveillance. Int J Popul Data Sci 2018;3:5.
[61]. Lix LM, Yao X, Kephart G, Quan H, Smith M, Kuwornu JP, Manoharan N, Kouokam W, Sikdar K. A prediction model to estimate completeness of electronic physician claims databases. BMJ Open 2015;5:e006858.
[62]. Long SJ, Freese J. Predicted probabilities for count models. The Stata J 2001;1:51–7.
[63]. MacKay C, Canizares M, Davis AM, Badley EM. Health care utilization for musculoskeletal disorders. Arthritis Care Res (Hoboken) 2010;62:161–8.
[64]. Mann EG, Johnson A, Gilron I, VanDenKerkhof EG. Pain management strategies and health care use in community-dwelling individuals living with chronic pain. Pain Med 2017;18:2267–79.
[65]. Mann EG, Johnson A, VanDenKerkhof EG. Frequency and characteristics of healthcare visits associated with chronic pain: results from a population-based Canadian study. Can J Anaesth 2016;63:411–41.
[66]. Marrie RA, Yu N, Blanchard J, Leung S, Elliott L. The rising prevalence and changing age distribution of multiple sclerosis in Manitoba. Neurology 2010;74:465–71.
[67]. Melzack R. From the gate to the neuromatrix. PAIN 1999;82(suppl 6):S121–126.
[68]. Millar WJ. Chronic pain. Health Rep 1996;7:47–53, 51–58.
[69]. Morales-Espinoza EM, Kostov B, Salami DC, Perez ZH, Rosalen AP, Molina JO, Gonzalez-de Paz L, Momblona JM, Areu JB, Brito-Zeron P, Ramos-Casals M, Siso-Almirall A. Complexity, comorbidity, and health care costs associated with chronic widespread pain in primary care. PAIN 2016;157:818–26.
[70]. Morasco BJ, Yarborough BJ, Smith NX, Dobscha SK, Deyo RA, Perrin NA, Green CA. Higher prescription opioid dose is associated with worse patient-reported pain outcomes and more health care utilization. J Pain 2017;18:437–45.
[71]. Morlion B, Coluzzi F, Aldington D, Kocot-Kepska M, Pergolizzi J, Mangas AC, Ahlbeck K, Kalso E. Pain chronification: what should a non-pain medicine specialist know? Curr Med Res Opin 2018;34:1169–78.
[72]. Moulin DE, Clark AJ, Speechley M, Morley-Forster PK. Chronic pain in Canada-prevalence, treatment, impact and the role of opioid analgesia. Pain Res Manag 2002;7:179–84.
[73]. Nabalamba A, Millar WJ. Going to the doctor. Health Rep 2007;18:23–35.
[74]. Nakamura M, Nishiwaki Y, Ushida T, Toyama Y. Prevalence and characteristics of chronic musculoskeletal pain in Japan: a second survey of people with or without chronic pain. J Orthop Sci 2014;19:339–50.
[75]. National Academies of Sciences, Engineering, and Medicine. Pain management and the opioid epidemic: balancing societal and individual benefits and risks of prescription opioid use. 1st ed. Washington, DC: The National Academies Press, 2017. doi: 10.17226/24781. Available at: https://www.nap.edu/catalog/24781/pain-management-and-the-opioid-epidemic-balancing-societal-and-individual. Accessed August 15, 2017.
[76]. Newfoundland and Labrador Centre for Health Information. Clinical Database Management System (CDMS) data users guide V.1.0. St. John's, NL: Newfoundland and Labrador Centre for Health Information, 2014. Available at: https://www.nlchi.nl.ca/images/PDFs/2014-04-15%20CDMS%20UG%20%20v%201%200%20%20Final.pdf. Accessed October 13, 2017.
[77]. Newfoundland and Labrador Centre for Health Information. Electronic medical record and electronic health record proof of concept study. St. John's, NL: Newfoundland and Labrador Centre for Health Information, 2013.
[78]. Newfoundland and Labrador Centre for Health Information. 2014. Available at: http://www.nlchi.nl.ca/. Accessed November 12, 2014.
[79]. Newfoundland and Labrador Centre for Health Information. NL electronic health record client registry user guide V2.0. St. John's, NL: Newfoundland and Labrador Centre for Health Information, 2013. Available at: https://www.nlchi.nl.ca/images/PDFs/Client%20Registry%20Data%20User%20Guide.pdf. Accessed October 13, 2017.
[80]. Newfoundland and Labrador Medical Association. Forum on health facilities & services review—presentations & videos. 2017. Available at: https://www.nlma.nl.ca/Advocacy/Forum-on-Health-Facilities--Services-Review/. Accessed May 1, 2020.
[81]. Newfoundland and Labrador Statistics Agency, Department of Finance. Population and demographics. Government of Newfoundland and Labrador, 2018. Available at: https://www.stats.gov.nl.ca/Statistics/Statistics.aspx?Topic=population. Accessed March 31, 2018.
[82]. O'Donnell MJ, Diener HC, Sacco RL, Panju AA, Vinisko R, Yusuf S. Chronic pain syndromes after ischemic stroke: PRoFESS trial. Stroke 2013;44:1238–43.
[83]. Ohayon MM, Stingl JC. Prevalence and comorbidity of chronic pain in the German general population. J Psychiatr Res 2012;46:444–50.
[84]. Outcalt SD, Yu Z, Hoen HM, Pennington TM, Krebs EE. Health care utilization among veterans with pain and posttraumatic stress symptoms. Pain Med 2014;15:1872–9.
[85]. Patel KV, Guralnik JM, Dansie EJ, Turk DC. Prevalence and impact of pain among older adults in the United States: findings from the 2011 National Health and Aging Trends Study. PAIN 2013;154:2649–57.
[86]. Power JD, Perruccio AV, Desmeules M, Lagace C, Badley EM. Ambulatory physician care for musculoskeletal disorders in Canada. J Rheumatol 2006;33:133–9.
[87]. Public Health Agency of Canada. Public Health Infobase: Canadian Chronic Disease Surveillance System (CCDSS). Government of Canada, 2018. Available at: https://health-infobase.canada.ca/ccdss/data-tool/. Accessed February 27, 2019.
[88]. Public Health Agency of Canada. The Canadian Chronic Disease Surveillance System—an overview. Government of Canada, 2018. Available at: https://www.canada.ca/en/public-health/services/publications/canadian-chronic-disease-surveillance-system-factsheet.html - fn6-0-rf. Accessed January 14, 2019.
[89]. Raftery MN, Ryan P, Normand C, Murphy AW, de la Harpe D, McGuire BE. The economic cost of chronic noncancer pain in Ireland: results from the PRIME study, part 2. J Pain 2012;13:139–45.
[90]. Rampersaud YR, Power JD, Perruccio AV, Paterson JM, Veillette C, Coyte PC, Badley EM, Mahomed NN. Healthcare utilization and costs for spinal conditions in Ontario, Canada—opportunities for funding high-value care: a retrospective cohort study. Spine J 2020;20:874–81.
[91]. Reitsma ML, Tranmer JE, Buchanan DM, VanDenKerkhof EG. The epidemiology of chronic pain in Canadian men and women between 1994 and 2007: results from the longitudinal component of the National Population Health Survey. Pain Res Manag 2012;17:166–72.
[92]. Reitsma ML, Tranmer JE, Buchanan DM, Vandenkerkhof EG. The prevalence of chronic pain and pain-related interference in the Canadian population from 1994 to 2008. Chronic Dis Inj Can 2011;31:157–64.
[93]. Ritzwoller DP, Crounse L, Shetterly S, Rublee D. The association of comorbidities, utilization and costs for patients identified with low back pain. BMC Musculoskelet Disord 2006;7:72.
[94]. Ritzwoller DP, Ellis JL, Korner EJ, Hartsfield CL, Sadosky A. Comorbidities, healthcare service utilization and costs for patients identified with painful DPN in a managed-care setting. Curr Med Res Opin 2009;25:1319–28.
[95]. Roberts MH, Mapel DW, Hartry A, Von Worley A, Thomson H. Chronic pain and pain medication use in chronic obstructive pulmonary disease. A cross-sectional study. Ann Am Thorac Soc 2013;10:290–8.
[96]. Roberts MH, Mapel DW, Thomson HN. The impact of chronic pain on direct medical utilization and costs in chronic obstructive pulmonary disease. Clinicoecon Outcomes Res 2015;7:173–84.
[97]. Romanelli RJ, Shah SN, Ikeda L, Lynch B, Craig TL, Cappelleri JC, Jukes T, Ishisaka D. Patient characteristics and healthcare utilization of a chronic pain population within an integrated healthcare system. Am J Manag Care 2017;23:e50–6.
[98]. Schopflocher D, Taenzer P, Jovey R. The prevalence of chronic pain in Canada. Pain Res Manag 2011;16:445–50.
[99]. Segovia J, Edwards AC, Bartlett RF. Newfoundland panel on health and medical care-adult health survey. Can J Public Health 1999;90:412–16.
[100]. Sinnott PL, Siroka AM, Shane AC, Trafton JA, Wagner TH. Identifying neck and back pain in administrative data: defining the right cohort. Spine (Phila Pa 1976) 2012;37:860–74.
[101]. Stanford EA, Chambers CT, Biesanz JC, Chen E. The frequency, trajectories and predictors of adolescent recurrent pain: a population-based approach. PAIN 2008;138:11–21.
[102]. Stanos S, Brodsky M, Argoff C, Clauw DJ, D'Arcy Y, Donevan S, Gebke KB, Jensen MP, Lewis Clark E, McCarberg B, Park PW, Turk DC, Watt S. Rethinking chronic pain in a primary care setting. Postgrad Med 2016;128:502–15.
[103]. Statistics Canada. Population centre (POPCEN). 2015. Available at: https://www12.statcan.gc.ca/census-recensement/2011/ref/dict/geo049a-eng.cfm. Updated November 27, 2015. Accessed January 19, 2020.
[104]. Stockbridge EL, Suzuki S, Pagan JA. Chronic pain and health care spending: an analysis of longitudinal data from the Medical Expenditure Panel Survey. Health Serv Res 2015;50:847–70.
[105]. Sturge-Jacobs M. The experience of living with fibromyalgia: confronting an invisible disability. Res Theor Nurs Pract 2002;16:19–31.
[106]. Szklo M, Nieto FJ. Understanding lack of validity: bias. In: Johnson M, Reilly T, eds. Epidemiology: Beyond the Basics. Burlington, MA: Jones & Bartlett Learning, 2014. p. 109–47.
[107]. Takura T, Ushida T, Kanchiku T, Ebata N, Fujii K, DiBonaventura M, Taguchi T. The societal burden of chronic pain in Japan: an internet survey. J Orthop Sci 2015;20:750–60.
[108]. Thomas E, Peat G, Harris L, Wilkie R, Croft PR. The prevalence of pain and pain interference in a general population of older adults: cross-sectional findings from the North Staffordshire Osteoarthritis Project (NorStOP). PAIN 2004;110:361–8.
[109]. Tian TY, Zlateva I, Anderson DR. Using electronic health records data to identify patients with chronic pain in a primary care setting. J Am Med Inform Assoc 2013;20:e275–80.
[110]. Tonelli M, Wiebe N, Fortin M, Guthrie B, Hemmelgarn BR, James MT, Klarenbach SW, Lewanczuk R, Manns BJ, Ronksley P, Sargious P, Straus S, Quan H; for the Alberta Kidney Disease Network. Methods for identifying 30 chronic conditions: application to administrative data. BMC Med Inform Decis Mak 2015;15:31.
[111]. Tripp DA, VanDenKerkhof EG, McAlister M. Prevalence and determinants of pain and pain-related disability in urban and rural settings in southeastern Ontario. Pain Res Man 2006;11:225–33.
[112]. Tsang A, Von Korff M, Lee S, Alonso J, Karam E, Angermeyer MC, Borges GL, Bromet EJ, Demytteneare K, de Girolamo G, de Graaf R, Gureje O, Lepine JP, Haro JM, Levinson D, Oakley Browne MA, Posada-Villa J, Seedat S, Watanabe M. Common chronic pain conditions in developed and developing countries: gender and age differences and comorbidity with depression-anxiety disorders. J Pain 2008;9:883–91.
[113]. Tumin D, Drees D, Miller R, Wrona S, Hayes D Jr, Tobias JD, Bhalla T. Health care utilization and costs associated with pediatric chronic pain. J Pain 2018;19:973–82.
[114]. Vallejo R, Bradley K, Kapural L. Spinal cord stimulation in chronic pain: mode of action. Spine (Phila Pa 1976) 2017;42(suppl 14):S53–s60.
[115]. van Hecke O, Hocking LJ, Torrance N, Campbell A, Padmanabhan S, Porteous DJ, McIntosh AM, Burri AV, Tanaka H, Williams FM, Smith BH. Chronic pain, depression and cardiovascular disease linked through a shared genetic predisposition: analysis of a family-based cohort and twin study. PLoS One 2017;12:e0170653.
[116]. van Hecke O, Torrance N, Smith BH. Chronic pain epidemiology and its clinical relevance. Br J Anaesth 2013;111:13–18.
[117]. Vedsted P, Christensen MB. Frequent attenders in general practice care: a literature review with special reference to methodological considerations. Public Health 2005;119:118–37.
[118]. Verhaak PF, Kerssens JJ, Dekker J, Sorbi MJ, Bensing JM. Prevalence of chronic benign pain disorder among adults: a review of the literature. PAIN 1998;77:231–9.
[119]. Virnig BA, McBean M. Administrative data for public health surveillance and planning. Ann Rev Public Health 2001;22:213–30.
[120]. Von Korff M, Wagner EH, Dworkin SF, Saunders KW. Chronic pain and use of ambulatory health care. Psychosom Med 1991;53:61–79.
[121]. White AG, Birnbaum HG, Mareva MN, Henckler AE, Grossman P, Mallett DA. Economic burden of illness for employees with painful conditions. J Occup Environ Med 2005;47:884–92.
[122]. Wilkie R, Tajar A, McBeth J. The onset of widespread musculoskeletal pain is associated with a decrease in healthy ageing in older people: a population-based prospective study. PLoS One 2013;8:e59858.
[123]. Williamson T, Miyagishima RC, Derochie JD, Drummond N. Manual review of electronic medical records as a reference standard for case definition development: a validation study. CMAJ Open 2017;5:E830–3.
[124]. Xie L, Semenciw R, Mery L. Cancer incidence in Canada: trends and projections (1983-2032). Health Promot Chronic Dis Prev Can 2015;35(suppl 1):2–186.
[125]. Yawn BP, Wollan P, St Sauver J. Comparing shingles incidence and complication rates from medical record review and administrative database estimates: how close are they? Am J Epidemiol 2011;174:1054–61.
Keywords:

Chronic pain; Chronic comorbidities; Health care utilization; Health administrative data

Supplemental Digital Content

Copyright © 2021 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the International Association for the Study of Pain.