The reported incidence of several infectious disease pathogens/syndromes decreased after the first case of COVID-19 was reported in Pakistan in late February 2020. The decrease occurred for all 8 infectious disease pathogens/syndromes in most of the provinces/regions included in this study: ARI, pneumonia, viral hepatitis, diarrhea, typhoid fever, measles, malaria and NNT. The only exceptions were malaria in KPK and Balochistan, as well as NNT in KPK, Balochistan and AJK & GB, which demonstrated no difference in IRRs between the pre-pandemic and pandemic periods—after seasonality was accounted for. A potential reason that these 2 particular infectious diseases did not demonstrate a significant difference in all 5 provinces/regions is that the risk factors, clinical manifestations, and diagnosis of these 2 diseases are, generally, different enough from that of COVID-19 so as not to cause confusion and lead to frequent misclassification or misdiagnosis (18, 19). Presumably, there was sufficient priority and interest in the control and prevention of malaria and NNT in Pakistan, to ensure that resources were not diverted from them to combat the COVID-19 pandemic (20, 21) in every province. Moreover, in respect to NNT specifically, case numbers in every province were relatively limited at the outset (i.e., in the tens to hundreds), likely as a result of vaccination efforts over the years. As an illustration, there were no cases of NNT in AJK & GB during the study period.
The incidence of several of the diseases such as malaria are very affected by seasonal fluctuations, likely related to the life cycle of its mosquito vectors (22) and other weather-related factors. It is therefore important to include seasonality in the model when analyzing the association between reported COVID-19 cases and incidence of these diseases. This is illustrated by the fact that when using the Wilcoxon signed-rank test (which does not take the varying trend of COVID-19 cases into account), the reported incidence of the 8 diseases was associated with the reported incidence of COVID-19 in less provinces/regions than in the regression models that included seasonality.
Monthly incidence of most of the infectious disease pathogens/syndromes (ARI, pneumonia, viral hepatitis, diarrhea, typhoid fever, and measles) decreased with the advent of COVID-19 in late February 2020 overall in the entire country (Fig. 1) as well as in all 5 provinces/regions separately (Figs. 2–9). A potential explanation is that after the first COVID-19 cases were reported in late February 2020, precautionary measures such as the closure of schools, work from home policies, public health restrictions and lockdowns were implemented (23), which adversely affected surveillance and data reporting of these diseases. In addition, the outpatient treatment departments (OPDs) of all major hospitals across the country were closed for patients during the initial few weeks of the pandemic (23, 24) leading to a substantial effect on the inflow of patients to the hospitals (potential data points for surveillance) for reasons other than COVID-19. This explains a sudden drop in reported cases of almost all major ongoing infectious diseases in all provinces. Subsequently, different provinces adopted different strategies to re-open under their respective administrative controls. With the onset of pandemic, a sudden increase in the “fear factor” associated with forced isolation and perceived stigma around a novel disease, largely spread by hyper-reactive and sensational media reports (25), also negatively affected the public’s willingness to seek medical attention through qualified doctors and hospitals, replacing this instead with a surge in self-prescription and home remedies for various health problems (26, 27). This trend is likely 1 of the main reasons for a sudden decline of reported incidence of various diseases with the escalating pandemic in many developing countries such as Pakistan. Alternatively, or perhaps additionally, several of these infectious disease pathogens/syndromes have presentations which overlap with those of COVID-19, particularly considering the purely clinically based case definitions used by the Pakistan IDSR, and may have been mistakenly diagnosed as such (16).
Some interesting findings became clear when modelling these infectious disease pathogens/syndromes. It is not surprising to see an inverse relationship between the number of COVID-19 cases and the number of ARI cases (Fig. 2), given the overlap in symptoms between these 2 conditions. The diagnosis of ARI is made clinically and is essentially defined as an upper respiratory tract infection (URTI), whereas upper respiratory symptoms, including sore throat, cough and rhinorrhea, are well documented symptoms of COVID-19 (28). As such, given the overlap in symptoms and lack of clear diagnostic testing for ARI, it is highly likely that COVID-19 could have been mistaken for ARI and vice versa, especially in a developing setting with less access to diagnostic testing. A similar argument could be made for pneumonia (Fig. 3), given the overlap in symptoms between these 2 conditions. Pneumonia due to COVID-19 is well described (29) and, thus, there is ample opportunity for misclassification between COVID-19 pneumonia and pneumonia due to other pathogens, particularly in a developing setting with less access to distinct and differential diagnostic testing.
There was a clear inverse relationship between COVID-19 cases and the number of reported cases of viral hepatitis (Fig. 4), believed to be mainly food and waterborne hepatitis due to hepatitis A and E. As the number of cases of COVID-19 in Pakistan increased, peaked in June 2020 and then declined, the number of reported cases of viral hepatitis steadily decreased, plateaued and then began to increase and approach the expected incidence again by September 2020 in most province/regions. We believe that the inverse relationship in this case may be primarily due to a diversion of resources, including diagnostic testing capacity, from viral hepatitis towards COVID-19 testing and management, although it may also be due to an overlap in symptoms during the pre-icteric phase of viral hepatitis, resulting in misdiagnosis and misclassification.
Similarly, the reported incidence of diarrhea in Pakistan declined with the onset of COVID-19 cases in late February 2020 (Fig. 5), perhaps because diarrhea is a common manifestation of COVID-19 (16, 28). As expected, as COVID-19 case numbers decreased in July and August 2020, reported cases of diarrheal illness continued to increase and peaked in September 2020 (Fig. 6). Conversely, typhoid fever is mainly transmitted through exposure to contaminated food and water, the latter of which is more likely to be consumed in larger quantities during the hot summer months. Regardless, there was a clear negative impact of the onset of COVID-19 cases in late February 2020 on the reported incidence of typhoid fever, and a similar phenomenon occurred in October through December 2020. The general inverse correlation between typhoid fever and COVID-19 was likely due to both a diversion of resources from the former to the latter as well as an overlap in symptoms such as fever, weakness, diarrhea and cough—resulting in misdiagnosis and misclassification (30).
The incidence of reported measles cases also decreased after the advent of COVID-19 cases in late February 2020, whereas reductions in COVID-19 cases since that time were associated with increased reporting of measles cases (Fig. 7). This was attributed to a diversion of resources during the COVID-19 pandemic previously put towards measles diagnostic testing, tracing and management. Of interest, the impact of the pandemic on the reporting of measles cases was significant during the first wave of the pandemic, where all provinces had a notable decline in case numbers, whereas in Balochistan and AJK & GB the same reduction in incidence did not occur during the second wave. This may have been due to some level of adjustment and improvement in measles diagnostic testing and tracing that were in place by this point in time in those provinces/regions.
The NNT case numbers were low at the outset (Fig. 8), and apparently least affected by the pandemic of all diseases/syndromes that we explored, demonstrating a small reduction during the first wave of the pandemic in Punjab and during the second wave of the pandemic in Sindh, but not otherwise. Malaria case numbers were impacted by the onset of the pandemic in some provinces/regions more than others, with a notable decrease in case numbers in Punjab, Sindh and AJK & GB (Fig. 9). As alluded to, this infectious disease was the most impacted by seasonal variations, as demonstrated by the difference between our Wilcoxon signed-rank test and our negative binomial regression model.
Despite a growing body of literature attempting to speculate, define and describe the impact of the COVID-19 pandemic on other ongoing infectious diseases outbreaks in many developing settings, to our knowledge our study is the first original research paper to do this using infectious diseases surveillance data from the MoNHSR&C. This had several advantages: we were able to utilize a set of national level data categorized by province/region and our interrupted time series study design allowed us to implement data that were objectively defined, easy to collect, simple to present graphically and straightforward to interpret.
Our study also had several weaknesses, the most important of which is that we were only able to determine correlations and associations between the onset of the COVID-19 pandemic and each of these other infectious disease pathogens/syndromes. Based on our study design, we were unable to ascertain causality. Moreover, several of these infectious disease pathogens/syndromes overlap to varying degrees with symptoms and manifestations of COVID-19 and could easily have led to misclassification bias during data collection. There also may have been both differences in data collection resources in various parts of the country at some time points, as well as national changes in data collection methodologies over time, as the Pakistani infectious disease surveillance system has evolved and improved over the past few years from completely paper-based to primarily electronic (31–33). Our surveillance was, therefore, not consistently robust and comprehensive throughout the entire reporting interval, potentially resulting in imperfect capture of monthly incidence for some pathogens/syndromes. Fortunately, there were no missing data during the pandemic period and, thus, no need was ascertained to impute data.
Notably, it is difficult to gauge the complete effect of lockdowns, restrictions, masking, hand hygiene, and compliance with these measures by the public. Moreover, we worked on aggregated data and did not have access to case-based data; therefore, were unable to cater for confounder and effect modifiers at study design or analysis levels. Finally, but importantly, the disease definitions we used, though based on Pakistani national surveillance protocols, are by their nature non-specific and somewhat difficult to compare with 1 another—some representing specific pathogens and others representing syndromes.
Nevertheless, it was evident that the incidence, surveillance, and reporting of the studied diseases were affected considerably with progression of the pandemic. Whatever the situation was, it can be assumed that it resulted in adverse consequences with respect to assessing true infectious disease burden and capturing all outbreaks. As a result, many cases of these endemic infectious disease pathogens/syndromes have likely been present in Pakistan over the past 24 mo, but not captured and mitigated by the health system, resulting in uncontained transmission to susceptible populations. Ultimately, such a situation is likely to result in further outbreaks, hyper-endemic pathogens, or even other epidemics.