Seasonality of service provision in hip and knee surgery: a possible contributor to waiting times? A time series analysis.

Abstract

The question of how best to reduce waiting times for health care, particularly surgical procedures such as hip and knee replacements is among the most pressing concern of the Canadian health care system. The objective of this study was to test the hypothesis that significant seasonal variation exists in the performance of hip and knee replacement surgery in the province of Ontario.
We performed a retrospective, cross-sectional time series analysis examining all hip and knee replacement surgeries in people over the age of 65 in the province of Ontario, Canada between 1992 and 2002. The main outcome measure was monthly hospitalization rates per 100,000 population for all hip and knee replacements.
There was a marked increase in the rate of hip and knee replacement surgery over the 10-year period as well as an increasing seasonal variation in surgeries. Highly significant (Fisher Kappa = 16.05, p < 0.01; Bartlett-Kolmogorov-Smirnov Test = 0.31, p < 0.01) and strong (R2Autoreg = 0.85) seasonality was identified in the data.
Holidays and utilization caps appear to exert a significant influence on the rate of service provision. It is expected that waiting times for hip and knee replacement could be reduced by reducing seasonal fluctuations in service provision and benchmarking services to peak delivery. The results highlight the importance of system behaviour in seasonal fluctuation of service delivery.

Full text

Page 1

BioMed Central
Page 1 of 5
(page number not for citation purposes)
BMC Health Services Research
Open Access
Research article
Seasonality of service provision in hip and knee surgery: A possible
contributor to waiting times? A time series analysis
Ross EG Upshur*1,2,3,4, Rahim Moineddin2,3, Eric J Crighton1 and
Muhammad Mamdani4,5,6
Address: 1Primary Care Research Unit, Sunnybrook and Women's College Health Sciences Centre, 2075 Bayview Ave, #E-349, Toronto, ON,
Canada, 2Department of Family and Community Medicine, University of Toronto, 256 McCaul Street, 2nd Floor, Toronto, ON, Canada,
3Department of Public Health Sciences, University of Toronto, McMurrich Building, 12 Queen's Park Crescent West, Toronto, ON, Canada,
4Institute for Clinical Evaluative Sciences, 2075 Bayview Avenue, Toronto, ON, Canada, 5Health Policy Management and Evaluation, University
of Toronto, McMurrich Building, 2nd Floor, 12 Queen's Park Crescent West, Toronto, ON, Canada and 6Faculty of Pharmacy, University of
Toronto, 19 Russell Street, Toronto, ON, Canada
Email: Ross EG Upshur* - ross.upshur@sw.ca; Rahim Moineddin - rahim.moineddin@utoronto.ca; Eric J Crighton - eric.crighton@sw.ca;
Muhammad Mamdani - muhammad.mamdani@ices.on.ca
* Corresponding author
Abstract
Background: The question of how best to reduce waiting times for health care, particularly
surgical procedures such as hip and knee replacements is among the most pressing concern of the
Canadian health care system. The objective of this study was to test the hypothesis that significant
seasonal variation exists in the performance of hip and knee replacement surgery in the province
of Ontario.
Methods: We performed a retrospective, cross-sectional time series analysis examining all hip and
knee replacement surgeries in people over the age of 65 in the province of Ontario, Canada
between 1992 and 2002. The main outcome measure was monthly hospitalization rates per 100
000 population for all hip and knee replacements.
Results: There was a marked increase in the rate of hip and knee replacement surgery over the
10-year period as well as an increasing seasonal variation in surgeries. Highly significant (Fisher
Kappa = 16.05, p < 0.01; Bartlett-Kolmogorov-Smirnov Test = 0.31, p < 0.01) and strong (R2Autoreg
= 0.85) seasonality was identified in the data.
Conclusion: Holidays and utilization caps appear to exert a significant influence on the rate of
service provision. It is expected that waiting times for hip and knee replacement could be reduced
by reducing seasonal fluctuations in service provision and benchmarking services to peak delivery.
The results highlight the importance of system behaviour in seasonal fluctuation of service delivery.
Background
Waiting times for important health interventions such as
hip and knee replacements, cataracts, cancer surgery, and
coronary artery bypass surgery have assumed increasing
prominence in publicly funded health care systems glo-
bally. In Canada, waiting times have been identified as
one of the crucial challenges facing the health care system,
have featured in debates during federal and provincial
Published: 01 March 2006
BMC Health Services Research2006, 6:22doi:10.1186/1472-6963-6-22
Received: 23 August 2005
Accepted: 01 March 2006
This article is available from: http://www.biomedcentral.com/1472-6963/6/22
© 2006Upshur et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Page 2

BMC Health Services Research 2006, 6:22 http://www.biomedcentral.com/1472-6963/6/22
Page 2 of 5
(page number not for citation purposes)
elections and have served as the basis for a challenge to
the Canada Health Act at the level of the Supreme Court.
Prolonged wait times for crucial health interventions raise
concerns about the quality of service that can be delivered
by such systems. These issues figure prominently in policy
debates internationally. A recent OECD report identified
waiting times as a significant policy issue in 50% of all
OECD countries, including among them Canada, the
United Kingdom, Australia, Finland, Sweden, Denmark
and the Netherlands, all of which have publicly funded
health care systems [1]. Hip and knee replacements are
among the highest priority conditions requiring reduction
of waiting times, particularly in light of an aging popula-
tion.
The determinants of waiting times are complex and likely
multi-factorial in origin. A recent report in Ontario exam-
ined waiting times for hip and knee replacements and rec-
ommended that management of waiting times can be
improved by more precisely understanding their determi-
nants [2]. This report examined temporal trends in wait-
ing times on an annualized basis showing a progressive
increase in need over time. However, there is reason to
suspect that annual rates alone may obscure significant
seasonal influences in the delivery of care. Therefore, the
objective of this study was to test the hypothesis of
whether significant seasonal variation exists in the per-
formance of hip and knee replacement surgery among a
population over the age of 65 in the province of Ontario.
Methods
We conducted a retrospective, cross-sectional time series
analysis. The Canadian Institutes of Health Information
Discharge Abstract Database was used to obtain monthly
hospitalization data on hip and knee replacements for
patients over the age of 65 beginning the first week of Jan-
uary 1992 to the last week of March 2002. All hip and
knee replacements were extracted from the database. (Pro-
cedure codes used were as follows: ICD 9 codes: 9340,
9341, 9351 to 9353, 9359, 9365 to 9368).
This database records discharges from all inpatient hospi-
tal stays in Ontario acute-care hospitals, documenting a
scrambled patient identifier, date of admission and dis-
charge, up to 16 diagnoses as coded by the International
Classification of Diseases, Ninth Revision, Clinical Modi-
fication (ICD-9-CM), and up to 10 procedures. All trans-
fers from within one acute care hospital to another within
this study group were excluded from the analysis.
Researchers using this database have found that diagnoses
are coded with a high degree of accuracy [3]. There is very
little missing information in the Ontario database; other
province level studies have similarly found that less than
1% of the basic information on patients is missing [4-6].
The reliability of the coding of data collected by the Cana-
dian Institute for Health Information is 74% to 96% for
the ICD-9 diagnosis.
The primary outcome measure was month of procedure.
The numerator consisted of all hip and knee replacements
analyzed in aggregate. The procedures were summed for
each month in the study period and plotted over time. The
denominator was created from annual census data for res-
idents of Ontario which were provided by Statistics Can-
ada. Monthly population estimates were derived through
linear interpolation. Using this data, monthly hospitaliza-
tion rates per 100,000 population were calculated.
Monthly average rates for the study period were also cal-
culated.
To assess the statistical significance of seasonal pattern in
the monthly rates of the hip and knee surgeries, spectral
analysis was conducted. Spectral analysis detects periodic-
ity in time series. The Fisher Kappa (FK) Test and the Bar-
tlett Kolmogorov Smirnov (BKS) Test were used to detect
significant departures from a random series. R2Autoreg was
calculated to assess the strength of seasonality [7].
Results
Table 1 shows the descriptive statistics. In the study period
there were 112,168 hip and knee replacements, 43,291 in
males, 68,877 in females. Figure 1 shows the rates of hip
and knee replacement as expressed as monthly procedure
rates per 100,000 persons. The seasonal fluctuation is
apparent and the monthly peak trough ratio increases
over time. The overall procedure rate increases with time,
almost doubling during the period from a low of 43 pro-
cedures per 100,000 in January of 1993, to a high of
approximately 100 procedures per 100,000 in October of
2001.
Table 1: Counts of knee and hip hospitalizations in Ontario:
1992–2001
Year Male Female Both
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
3051
3398
3592
3966
4192
4482
4681
5227
5274
5428
5199
5309
5896
6452
6752
7180
7336
8068
8006
8679
8250
8707
9488
10418
10944
11662
12017
13295
13280
14107
Total43,29168,877112,168

Page 3

BMC Health Services Research 2006, 6:22http://www.biomedcentral.com/1472-6963/6/22
Page 3 of 5
(page number not for citation purposes)
Figure 2 shows the monthly average number of proce-
dures performed over the time period. There are marked
seasonal variations in procedures. Peak capacity occurs in
the fall where procedures peak at rates in excess of 83 per
100,000 population. There are substantial reductions of
services corresponding to the summer months, around
the Christmas and New Year's season, and to a lesser
extent the March break, where procedures are seen to
trough at rates as low as 56 per 100,000.
The data series was de-trended using first order differenc-
ing prior to conducting spectral analysis. Two tests for the
null hypothesis that the series is strictly white noise were
conducted. The Fisher Kappa (FK) Test was 16.05 (p <
0.01) and the Bartlett-Kolmogorov-Smirnov (BKS) Test
was 0.31 (p < 0.01), indicating highly significant season-
ality in the data. The R2Autoreg value for the series was 0.85
providing evidence of strong seasonality.
Discussion
Our results indicate a marked increase in the rate of hip
and knee replacement surgery over a 10-year period as
well as an increasing seasonal variation in surgeries. The
provision of services decreases during holiday times. Peak
performance occurs in the fall months. Despite the
increase in the provision of hip and knee replacements,
the waiting times for these services have increased dramat-
ically. Ontario data indicates that fewer 50% of eligible
patients receive a joint replacement within 6 months. [2]
What explains this seeming paradox? We hypothesize that
seasonal fluctuation in service provision contributes to
the problem.
There appears to be two major driving factors that can
contribute to increasing waiting times. The segment of the
population over the age of 65 has increased by approxi-
mately 23% percent over the years in this study
(1,172,620 to 1,446,227). Secondly, there is marked sea-
sonal variation in the provision of surgery indicating that
the provision of surgery is not constant. Peak service
occurs only in the months of October, November, and to
a lesser extent, May and June. This may reflect both pro-
vider and recipient preferences not to have surgery during
holiday times. Although holidays exert an influence on
the provision of service, they do not affect the demand
which relates to osteoarthritis, the predominant indica-
tion for surgery. This has not changed during the study
period. Holidays exert a significant influence on the rate
of service provision as the troughs correspond to the sum-
mer holidays, winter holidays and March school breaks.
Hip and knee replacement surgery rates per 100,000 population by month and year in Ontario (1992–2002)
Figure 1
Hip and knee replacement surgery rates per 100,000 population by month and year in Ontario (1992–2002).

Page 4

BMC Health Services Research 2006, 6:22http://www.biomedcentral.com/1472-6963/6/22
Page 4 of 5
(page number not for citation purposes)
As well, it is possible that utilization caps in the province
of Ontario explain the lower rates of service provision dur-
ing January-March. Until 2005, the salary of orthopedic
surgeons was capped at a certain level adding financial
disincentives to perform surgery for which they would not
be full remunerated. The fiscal year runs from April 1 –
March 31, so the January – March rates corresponds to the
time that surgeons would be approaching the cap. As the
caps have been lifted in 2005, their effect on rates will be
testable in the years to come.
In a simple schemata, waiting lists can be understood
operationally in terms of the relationship between inputs
(increasing number of people needing surgery), through-
puts (including seasonal fluctuation and system features),
and outputs (completed surgeries). It can be hypothesized
that waiting times for hip and knee replacement could be
reduced with more constant supply. Eliminating seasonal
fluctuations in service provision and benchmarking serv-
ices to peak delivery months would increase capacity by
about 400 to 500 procedures per annum. The cyclic peaks
and troughs likely exacerbate waiting times as the cycle is
repeated annually while the need increases steadily.
The results highlight the importance of system behaviour
in seasonal fluctuation of service delivery. The OECD
report identified several determinants of waiting lists
including physician availability and remuneration but did
not consider time of surgery as an important factor in the
creation of waiting times [1]. Recent reports on waiting
times for hip and knee replacements in Canada have
examined geographical variation, but did not examine
rates as dynamic over time despite recognizing time of
surgery as a crucial component of waiting times [2].
Although it is well known that seasonal increases of
health service demand are associated with viral patho-
gens, such as influenza, it is less well appreciated the
extent to which human behaviour influences health serv-
ices delivery. Attempts to rectify waiting lists without con-
sidering human behaviour will likely fail to address many
issues determining waiting lists.
There are certain caveats to this analysis. First, we lack an
estimate for the prevalence of need for surgery. We have
assumed that the prevalence of need has increased over
time in this analysis using an aging population as a proxy
for surgical demand. Secondly, this analysis is limited to
Hip and knee replacement surgery rates per 100,000 population aggregated by month in Ontario (1992–2002)
Figure 2
Hip and knee replacement surgery rates per 100,000 population aggregated by month in Ontario (1992–2002).

Page 5

Publish with BioMed Central and every
scientist can read your work free of charge
"BioMed Central will be the most significant development for
disseminating the results of biomedical research in our lifetime."
Sir Paul Nurse, Cancer Research UK
Your research papers will be:
available free of charge to the entire biomedical community
peer reviewed and published immediately upon acceptance
cited in PubMed and archived on PubMed Central
yours — you keep the copyright
Submit your manuscript here:
http://www.biomedcentral.com/info/publishing_adv.asp
BioMedcentral
BMC Health Services Research 2006, 6:22 http://www.biomedcentral.com/1472-6963/6/22
Page 5 of 5
(page number not for citation purposes)
the province of Ontario as the unit of aggregation. There-
fore, it will overlook important geographic variations
within the system which may play a role in waiting list
dynamics. Such factors as OR time and availability, and
the number of surgeons available to perform such surger-
ies, are obviously important determinants of waiting
times. We restricted our analysis to this population as it is
the population with by far the greatest need for the proce-
dure.
Conclusion
The results of this study indicate the importance of under-
standing the temporal trends and seasonal variation of
service provision. A similar analysis for other conditions
plagued by waiting lists such as cataracts and cancer sur-
gery would be informative to determine if the patterns
demonstrated with hip and knee surgeries similarly exist.
These results could be verified in other OECD jurisdic-
tions.
Competing interests
The author(s) declare that they have no competing inter-
ests.
Authors' contributions
REGU initiated the idea for the study, interpreted the data,
wrote the first draft of paper, and contributed to all subse-
quent drafts. RM conducted the statistical analysis, con-
tributed to the design of the study, helped interpret the
findings, and contributed to the writing of the paper. MM
and EJC helped design the study, interpret the findings
and contributed to the writing of the paper. All authors
have read and approved the final manuscript.
Acknowledgements
The authors would like to thank Shari Gruman for her expert assistance in
formatting the manuscript.
This study was funded by an operating grant No. MOP-57928 from the
Canadian Institutes of Health Research.
Dr Upshur is supported by a New Investigator Award from the Canadian
Institutes of Health Research and a Research Scholar Award from the
Department of Family and Community Medicine, University of Toronto.
References
1. Siciliani L, Hurst J: Explaining waiting times variations for elective surgery
across OECD countries Paris: Health Publication Service, OECD; 2003.
2.Bourne R, DeBoer D, Hawker G, et al.: Chapter 5 Total Hip and
Knee Replacement. In Access to Health Services in Ontario: ICES Atlas
Edited by: Tu J, Pinfold S, McColgan P, Laupacis A. Toronto; 2005.
3. Williams J, Young W: Appendix I: A summary of studies on the
quality of health care administrative databases in Canada. In
Patterns of Healthcare in Ontario. The ICES Practice Atlas 2nd edition.
Canadian Medical Association; 1996.
4. Davidson W, Molloy DW, Somers G, Bedard M: Relation between
physician characteristics and prescribing for elderly people
in New Brunswick. CMAJ 1994, 150:917-921.
5.Rawson N, Malcolm E: Validity of the recording of cholecystectomy and
hysterectomy in the Saskatchewan health care datafiles Saskatoon: Phar-
macoepidemiology Research Consortium; 1995.
Thiessen BQ, Wallace SM, Blackburn JL, Wilson TW, Bergman U:
Increased prescribing of antidepressants subsequent to beta-
blocker therapy. Arch Intern Med 1990, 150:2286-2290.
Moineddin R, Upshur RE, Crighton E, Mamdani M: Autoregression
as a means of assessing the strength of seasonality in a time
series. Popul Health Metr 2003, 1:10.
6.
7.
Pre-publication history
The pre-publication history for this paper can be accessed
here:
http://www.biomedcentral.com/1472-6963/6/22/prepub