MPH5040 S2 2019: Critical Appraisal of Osteoarthritis Study Report

This report provides a critical appraisal of a study investigating the association between reproductive history, hormone replacement therapy, and the risk of total knee replacement (TKR) and total hip replacement (THR) due to osteoarthritis. The study, conducted by Hellevik et al. (2017), aimed to determine if factors like age at menarche, parity, menopausal status, use of oral contraceptives, and hormone replacement therapy are linked to the risk of joint replacements. The report critically evaluates the study's methodology, including the use of the HUNT Study and the Norwegian Arthroplasty Register, and assesses the strengths and weaknesses of the study design, data collection, and analysis. It examines the impact of non-response bias, the exclusion criteria used, and the validity of the exposure data. The report also explores the study's findings, such as the association between age at menarche, HRT use, and TKR risk, and evaluates the confounding factors considered. Furthermore, the report provides a detailed analysis of the study's results, including the calculation of cumulative incidence rates and the interpretation of hazard ratios. The appraisal also considers the limitations of the study and suggests areas for future research, such as the use of randomized control trials to further investigate the relationship between hormonal therapies and osteoarthritis. Overall, the report offers a comprehensive evaluation of the study's contribution to the understanding of the risk factors associated with osteoarthritis and joint replacement.

Running head: CRITICAL APPRAISAL
CRITICAL APPRAISAL
Name of the student
Name of the University
Author Note

1
CRITICAL APPRAISAL
Question 1:
1a) Authors conducted the study in order to understand and obtain knowledge about whether the
risk of total knee replacement (TKR) and total hip replacement (THR) due to osteoarthritis are
associated with reproductive history, hormone replacement therapy, oral contraceptives use or
not. Moreover, age, age at menopause, menopausal status and age at menarche are also
associated with the risk of the TKR and THR or not is also assessed in this study by the authors
(1). The study also focused on the understanding about the confounding factors for the risk of the
TKR or THR situations among the women. The study is mostly based on the negative factors
affecting the health condition of the women with above mentioned issues. However, the authors
could not be able to conclude the study as the findings were not validated and there are different
factors affecting the results as well.
1b) The study is based on the risk factors and the association of several aspects with the TKR
and THR. Hence, the study would be helpful in the knowledge development about the causal
factors of TKR or THR (1). It can be highlighted as the advantage of the study. On the other hand
the study took a long period of time which can be highlighted as the disadvantaging factor.
Moreover, the sample size, prospective population based design and objective measurements are
the advantageous factors as well for this study. On the contrary, large number of causal effects,
validation error of the osteoarthritis and homogenous population are the disadvantageous effects
of the study. Thus, it can be stated that the large sample size is the prominent aspect that can be
highlighted as the advantage of the study whereas, the not validated data and also the large
number of variables are the negative factors that can be termed as the disadvantage of the study.
1c) The non-response can affect the study results in various ways. The primary effect of the non-
response is the increase of the survey bias or the response bias (2). The non-response can also lead

2
CRITICAL APPRAISAL
to the increase in the variance as the expected sample size and response would be reduced.
Moreover, the risk ratio and the odds ratio of the study would be changed as well. Thus it can be
stated that the statistical values of the survey would be affected due to the reduction of the
sample size. On this context it can also be stated that the result would be biased due to the non-
response of the participants. Moreover, the context of the research should follow the research
guidelines and also try to maintain the participant number in order to develop accurate result
without any bias.
1d) The HUNT2 and HUNT3 study populations helped in the knowledge development about the
reproductive history and the association of this factor with the risk of the TKR and THR (1). The
covariates such as the HRT also evaluated with the questionnaire for interviewing these two
populations. However, the large amount of covariates affected the result as the association could
not be identified properly with the age, parity, hormonal therapy and use of OC. On the other
hand after excluding 1183 people from the sample there are participants with missing
information as well which can also be termed as a disadvantage in terms of this large sample
population (1). This missing data affected the result as well. Hence, the advantages are the idea
development about the association between the reproductive history and other confounding
factors and the TKR and THR. On the contrary the disadvantage is the unidentified association
between the age, parity, hormonal therapy and use of OC due to large number of covariates.
Question 2:
2a) The 1183 women undergone joint replacement before selection has been excluded as there
are presence of missing data (1).
2b) Bilateral oophorectomy in premenopausal women induces premature menopause. Women
with ovarian preservation if undergo hysterectomy induce the risk of premature menopause to

3
CRITICAL APPRAISAL
almost double comparing with the women having uteri (1). Hence, reducing the risk of the
premature menopause among the participants was the cause of excluding 3710 participants (1).
2c) Norwegian Arthroplasty Register (NAR) is the record of the people undergone primary
osteoarthritis (1). The record has been maintained with the 11 digit personal identification
number. Hence, the identification of the persons, who have undergone TKR and THR was easy
and the data has been authentic (1). The advantage of using NAR is obtaining the data about the
people undergone primary osteoarthritis and also identifying the cases easily. On the contrary the
disadvantages are identification of cases among large amount of data and the factor of the recent
data would not be obtained from the database. Moreover, according to Danish Hip Arthroplasty
Registry the NAR data is not absolutely valid for some cases. On this context it can also be
stated as the disadvantage of the NARA data.
Question 3:
3a) Parity, age at menarche, years of menstruation, menopausal status, and age at menopause,
use of oral contraceptives and use of HRT are the exposures studied by the authors(1).
The data has been collected based on the exposures by reviewing the NAR data and
interviewing the participants (1).
3b) The authors found that the NAR data have not been validated and the Danish Hip
Arthroplasty Registry reported 85 per cent positive predictive value in comparison of NAR (1).
Hence, the exposure data can be termed as not fully valid as the data is not clear about many
aspects considered in the study. If the data is not properly valid then the exposure classification
can be misclassified. The direct acyclic graph (DAG) analysis also highlighted slight difference
in data regarding the confounders. Hence, the misclassification of data can be obtained from this
study (1).

4
CRITICAL APPRAISAL
4a) Yes, the differences among women with different levels of reproductive and hormonal
factors are marked distinctly in the table of the article with different markings.
Yes, differences in women who received a total knee replacement or total hip
replacement are marked distinctly in the table of the article with different markings.
These differences are marked for the sampling of the participants and the identification of
the data. Moreover, the inclusion and the exclusion of the participants would be depending on
the marking of these differences.
4b) Based on the table 2 of the article about the cumulative incidence of total knee replacement
=TKR/All women
= 430/30289
= 0.014
Based on the table 2 of the article about the cumulative incidence of total hip replacement
= THR/All women
= 675/30289
= 0.022
Hence, based on the calculation cumulative incidence rate of total knee replacement and
total hip replacement is 0.014 and 0.022 respectively.
4c) 6253 women have missing data on hard physical activity that is one of the exposures or
parameter data is missing among this amount of participants. Hence, it can be stated that the risk
of bias would be present in the study and also the desired result could not be obtained. Moreover,
the study could not be concluded properly.
5) Total person day of the 1st participant that is the person followed up from April 25, 1995 to 31
December 2014 = 1X7190 = 7190 days.

5
CRITICAL APPRAISAL
Total person day of the 2nd participant that is the person followed up from June 30, 1997 to 31
December 2014 = 1X6393 = 6393 days.
Total person day of the 3rd participant that is the person followed up from June 30, 1997 to 31
December 2014 = 1X2948 = 2948 days.
6) Smoking, current HRT user, past HRT user, years of OC use, menopausal status, age at
menopause, age at menarche, BMI, diabetes, hard physical activity, mean age and parity are the
confounding factors adjusted for the study (1).
Based on the study the authors considered most of the factors in the hormonal therapies
and the reproductive history and the TKR and THR. However, the occupation and habits of
physical stresses should also be considered. Moreover, the psychological factors and other
chronic disease condition should be considered as well for better assessment of the condition and
the confounding factors of the condition.
7) The findings of the study highlighted that the older age at menarche decreases the risk of the
TKR and past or systematic use of HRT increases the risk of TKR. Moreover, the parity is not
associated with risk of TKR or THR (1).
On the other hand the nulliparous group showed 25 THR and 17 TKR cases which relates
to the fact of the indecisive condition development for the analysis (1). On this context it can be
stated that the analysis was not at per as the validation of the finding was not up to the mark.
Moreover, the findings highlighted that the education level of the participants also play a crucial
role in the reduction of the risk of TKR or THR. Hence, the findings contributed in the
knowledge development about this context of different confounding factors of the risk of TKR
and THR along with the possible causes of the conditions regarding joint replacements.

6
CRITICAL APPRAISAL
However, there are scopes for the future research as the study did not properly find all the factors
or associations.
8) “Past users and users of systemic HRT were at higher risk of TKR compared to never users
(HR 1.42 (95% CI 1.06 - 1.90) and HR 1.40 (95% CI 1.03 -1.90), respectively)” this sentence
means that the for 95 per cent of the participants the hazard ratio is true that is the 1.42 and the
1.40 hazard ratio respectively for the past or systematic user of HRT and never user of HRT
respectively among the participants (3).
9) ‘Hormonal therapies may affect the risk of osteoarthritis’ is a hypothesis which can be
determined through a clinical trial by implementing the process of the experiment and control
group for the assessment by means of comparison of data. Hence, it can be stated that the
randomised control trial would be implemented for the study of the association determination
among hormonal therapy and risk of osteoarthritis (4).
This would provide strong evidences as the experimental study would provide the data
about the process and the comparison of both the control and the experiment group will highlight
the effects and the causes as well. However, the evidence would be obtained through the specific
study and specific parameters. Thus the evidence would be different for different parameters. On
this context the study evidence could not be termed as strong.
As per the study conducted by the authors highlighted that the past hormonal therapy or
systematic hormonal therapy leads to the risk of the TKR (1). Hence, among the post menopausal
women the effect of the hormone therapy can also increase the risk of the development of TKR
or THR. Thus hormonal therapy is not recommended as the intervention for the post menopausal
condition.

7
CRITICAL APPRAISAL
References
1. Hellevik AI, Nordsletten L, Johnsen MB, Fenstad AM, Furnes O, Storheim K, Zwart JA,
Flugsrud G, Langhammer A. Age of menarche is associated with knee joint replacement due to
primary osteoarthritis (The HUNT Study and the Norwegian Arthroplasty Register).
Osteoarthritis and cartilage. 2017 Oct 1;25(10):1654-62.
2. McGovern M, Canning D, Bärnighausen T. Accounting for Non-Response Bias using
Participation Incentives and Survey Design. Centre for HeAlth Research at the Management
School (CHaRMS); 2018 Apr.
3. Windmeijer F, Liang X, Hartwig F, Bowden J. The Confidence Interval Method for Selecting
Valid Instruments. Technical report, University of Bristol; 2018 Aug 16.
4. Hemming K, Haines TP, Chilton PJ, Girling AJ, Lilford RJ. The stepped wedge cluster
randomised trial: rationale, design, analysis, and reporting. Bmj. 2015 Feb 6;350:h391.