Quality of End-of-Life Cancer Care in Canada: A 12-Year Retrospective Analysis of Three Provinces’ Administrative Health Care Data Evaluating Changes over Time

Round 1

Reviewer 1 Report

Dear Dr.Khan and Dr.Barbera,

Thanks again for your remarkable and magnificent work during the pandemic.

Would you please double-check all your calculations? It is a severe problem.

Please answer the questions below to improve the quality of the manuscript.

1. for the tables and figures,

Please add a flow chart for the Population and Study Design part  (Line 67) with the number of patients in each province and the cases with missing data.

This flow chart should be listed before your Table A1.

Figure1 in the article below is an example.

 

 

 

 

 

 

 

 

 

 

 

 

Kao, Yee-Hsin, and Jui-Kun Chiang. "Effect of hospice care on quality indicators of end-of-life care among patients with liver cancer: a national longitudinal population-based study in Taiwan 2000–2011." BMC palliative care 14.1 (2015): 1-10.

 

2. Please double-check all the calculations in the tables. The data of Table 1 from Score on the Dayo-Charlson comorbidity index (from -24 months to -12 months before death) section (Line 348) is incorrect. The overall present cannot be lower than each province.

 

 

 

 

 

 

3. Since this is a cross-sectional study, an AHRQ table is used to evaluate the study's quality.

• Please write a part to explain whether it is a blinded experiment or not.

  Here is the example for the data collection part.

Data Collection

Participants’ characteristics were assessed at baseline (before random assignment). Data on outcome measures (LST preferences, QoL, anxiety symptoms, and depressive symptoms) and time-varying covariates were collected at baseline and every 3 to 4 weeks afterward until participants withdrew or died. Data collectors were blinded to participants’ group assignment.

Tang, Siew Tzuh, et al. "Advance care planning improves psychological symptoms but not quality of life and preferred end-of-life care of patients with cancer." Journal of the National Comprehensive Cancer Network 17.4 (2019): 311-320.

• Please write a part of how you handle the missing data. We knew you did it, but please note it in the method part. (not just the discussion part)

Comments for author File: Comments.pdf

Author Response

Please see attachment thank you.

Author Response File: Author Response.docx

Reviewer 2 Report

In this manuscript, “ Quality of end-of-life cancer care in Canada: a 12-year retrospective analysis of three-provinces’ administrative health care data evaluating changes over time,” the authors would like to describe the trend of palliative care using end-of-life cancer care quality indicators from the administrative database. The paper was written well, and I just have some minor questions as below,

1. The authors used the forest plot to present the ORs in Figure3. This figure is good, but I suggested that the authors should make it easily readable. The scale of OR may be changed to show the readable estimated points with 95% CI. In addition, the reference group should be noted.
2. The authors presented the age group in your study population. I suggested the Figure3 should also present age as a categorical variable, not continuous variable.
3. The authors show the different cancer types, could the authors explain why lung cancer was selected as reference group in multivariable logistic regression models? It could be described in the methods.
4. I suggested the authors could merge Figure3 and the Appendix Table3 and Table4 as two figures. These could make this valuable information easy to understand.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

very clear manuscript

The aim of this manuscript was to analyzes end-of-life care for people 62dying of cancer across three Canadian provinces, over a 12-year time period, to determine whether there are changes in palliative care practices over time in several of the 64validated indicators.

Quality of the end of life is very important topic and few papers analyze the evolution of this quality over time

Generally, published data’s is cross sectional analysis and the 12 years time analysis reported on this analysis is very relevant

Regarding the methodology, the methodology use is adequate and there is a large literature validation the quality indicators analyzed

Conclusion is supported by the results section, the references, tables and figures are adequates……

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

Article presents a retrospective cohort study of cancer decedents over a period of more than 10 years (2004–2015), focusing on end-of-life cancer care quality indicators (QIs) in 3 Canadian provinces. Choice of these indicators is supported by a robust literature review published by L.A. Henson and colleagues. As far as I'm concerned, analysis has been properly conducted. The sample collected is extensive and therefore many data has been available. Since it is not common to have such an interesting series of data, I sincerely congratulate the authors for the good work produced.

The main research question addressed by the paper is about the quality of cancer care provided by 3 Canadian provinces in 12 years (2004-2015). It could be: In comparative terms, have there been changes in the quality of end-of-life care for cancer patients in three Canadian provinces? What is the quality of end-of-life care for cancer patients in Canada today?

Lines 62-64 fully express aim of the paper: “This paper analyzes end-of-life care for people dying of cancer across three Canadian provinces, over a 12-year time period, to determine whether there are changes in palliative care practices over time in several of the validated QIs outlined by Henson et al.”

As I already wrote in the comment I’ve sent, it is not common to analyse and present results of historical series data of 12 years on more than 50% of Canadian population (2.1 Population and Study Design paragraph). In fact, population data on EOL are not publicly available. Canada, nevertheless, as UK, US and few other countries, collect systematically data on EOL giving us the opportunity to understand the effects of EOL policies. The coexistence of highly urbanized areas and extensive rural territories enriches the analyzes, making them informative also for other countries. Furthermore, the systematic application of QIs and the publication of performances is necessary for the international comparison of the state of the art of health systems in the field of palliative care and EOL care.

As I have already pointed out, the contribution adds knowledge on the basis of population, on the performance of the health system on EOL care in one of the most advanced countries in palliative care. It could be considered an update of previous works, such:

• Grunfeld E, Lethbridge L, Dewar R, Lawson B, Paszat LF, Johnston G, Burge F, McIntyre P, Earle CC. Towards using administrative databases to measure population-based indicators of quality of end-of-life care: testing the methodology. Palliat Med. 2006 Dec;20(8):769-77. doi: 10.1177/0269216306072553. PMID: 17148531; PMCID: PMC3741158.
• L. Barbera et al., “Factors associated with end-of-life health service use in patients dying of cancer,” Healthc. 416 Policy, vol. 5, no. 3, p. e125, 2010, doi: 10.12927/hcpol.2013.21644.

Authors used validated indicators, as correct methodological approaches suggest and they performed multivariable analysis to detect factors “associated with a patient receiving either aggressive or supportive care” (line 120-121). Analysing administrative and non-ad-hoc data is challenging, as known. Indeed, authors could not detect changes in all dimensions and Qis homogeneously since data were not available in all three provinces. But, as the authors have openly stated the limits of their work in the specific section, the lack of some dimensions can be considered as an indication for the control bodies to improve and update the detection systems. We also experience similar problems in Italy regarding the completeness of the information collected in the healthcare sector, despite the progressive digitization of the sector.

One of the challenges today and in the future are palliative care and EOL care for non-cancer patients. It would be interesting to compare, through the same indicators, the population affected by cancer who benefited from EOL care and the population affected by other non-oncological diseases.

Conclusions are indeed consistent with arguments. The only suggestion I would express is about the link between policies adopted by the governments of the Canadian provinces and results on QIs. The causal relationship should be modeled considering different variables and more complex approaches, but since the aim of the paper is mainly descriptive, considerations made in conclusion section are welcome.

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Dear Dr. Dear Dr.Khan and Dr.Barbera, 

Thank you for reply my questions and modify the manuscipt. 

Congratulations, I really like it and want to accept it. 

Before we finalize the articles, Could you double check the calculations again? For example the percentile of missing cases from NS (where you highlight the line 357) is incorrect. Please double check it again. 

Moreover, there are so much blinded retrospective study online. 

A randomized blinded retrospective study: the combined use of micro-needling technique, low-level laser therapy and autologous non-activated platelet-rich plasma improves hair re-growth in patients with androgenic alopecia - PubMed (nih.gov)

If you did not do it, just leave it there now. If someone wants to include your study in their meta-analysis, they will evaluate it. 

Thank you for your reply and please modification it again. 

Congratulations!!!

 

Author Response

Thank you reviewer 1 once again for your keen eye! We brought in a data analyst from our institution to double check our calculations and percentages. We had erroneously calculated some of the percentages as you pointed out and now each percentage is based on the in-group analysis and should be correct. We regret this error and appreciate your very excellent feedback!! We also have been unfamiliar with blinded retrospective studies until now and we will consider this format for our future analyses! Thank you!